Anda di halaman 1dari 285


Contributions to Narrative Theory

Edited by
Fotis Jannidis, Mat´ıas Mart´ınez, John Pier
Wolf Schmid (executive editor)

Editorial Board
Catherine Emmott, Monika Fludernik
Jose Angel Garc´ıa Landa, Peter Hühn, Manfred Jahn
Andreas Kablitz, Uri Margolin, Jan Christoph Meister
Ansgar Nünning, Marie-Laure Ryan
Jean-Marie Schaeffer, Michael Scheffel
Sabine Schlickers, Jörg Schönert


De Gruyter
From Concept to Narrative Construct:
A Reader

Edited by
Jan Christoph Meister
Wilhelm Schernus

De Gruyter
ISBN 978-3-11-022208-1
e-ISBN 978-3-11-022718-5
ISSN 1612-8427

Library of Congress Cataloging-in-Publication Data

Time : from concept to narrative construct : a reader / edited by Jan

Christoph Meister, Wilhelm Schernus.
p. cm.
Includes bibliographical references and index.
ISBN 978-3-11-022208-1 (acid-free paper)
1. Time in literature. 2. Narration (Rhetoric) 3. Time — Philoso-
phy. I. Meister, Jan Christoph, 1955— II. Schernus, Wilhelm.
PN56.T5T56 2011

Bibliographic information published by the Deutsche Nationalbibliothek

The Deutsche Nationalbibliothek lists this publication in the Deutsche
Nationalbibliografie; detailed bibliographic data are available in the Internet

© 2011 Walter de Gruyter GmbH & Co. KG, Berlin/Boston

Printing: Hubert & Co. GmbH & Co. KG, Göttingen
f» Printed on acid-free paper
Printed in Germany

Acknowledgements VII

Foreword IX

The Tenses of Verbs 1

Time Experience and Personhood 13

Constituting Time through Action and Discourse 29

Time, Tense and Topology 49

The Significance of Time in Narrative Art 67

The Timelessness of Poetry 85

The Time References of Narration 101

Time Structure in the Contemporary Novel 109

Story-time and Fact-sequence-time 143
VI Contents

The Temporality Effect
Towards a Process Model of Narrative Time Construction 171

The Flow of Time in Narrative
An Artificial Intelligence Perspective 217

Bibliography: A Guide to Further Reading 237

Subject Index 253

Name Index 257


The editors are grateful for permission to translate and reproduce the
following material:

Peter Bieri: 'Zeiterfahrung und Personality,' in: Heinz Burger (ed.):

Zeit, Natur und Mensch. Beitrage von Wissenschaftlern zum Thema
"Zeit". Berlin: Berlin Verlag Arno Spitz 1986, 261-281, by Peter Bieri

Alfonso de Toro: 'Versuch eines erweiterten Modells fur die Analyse

von Zeitverfahren nach G. Genette,' in: Alfonso de Toro: Die Zeit-
struktur im Gegenwartsroman am Beispiel von G. Garcia Marquez'
Cien anos de soledad, M Vargas-Llosas La casa verde und A. Robbe-
Grillets La maison de rendez-vous. Tubingen: Gunter Narr Verlag
1986, 26-52, by Alfonso de Toro

Kate Hamburger: 'Die Zeitlosigkeit der Dichtung,' in: Deutsche Vier-

teljahrsschrift fur Literaturwissenschaft und Geistesgeschichte 29
(1955), 413-426, by Deutsches Literaturarchiv Marbach

Roland Harweg: 'Erzahlte Zeit und Sachverhaltsfolgezeit,' in: Folia

Linguistica 25 (1991), 41-73, by Roland Harweg

Peter Janich: 'Die Konstion der Zeit durch Handeln und Reden,' in:
Kodikas/Code 19.1-2 (1996), 133-147, by Peter Janich

Eberhard Lammert: 'Die Zeitbeziige des Erzahlens,' excerpt from

Eberhard Lammert: Bauformen des Erzahlens. Stuttgart: Metzler
[1955] 81990, 19-24, by Metzler

Robin Le Poidevin: 'Time, Tense and Topology,' in: The Philosophic-

al Quarterly 46 (1996), 467-481, by Robin Le Poidevin

Inderjeet Mani, orginal contribution to this volume

VIII Acknowledgements

Jan Christoph Meister, original contribution to this volume

Giinther Mullen Die Bedeutung der Zeit in der Erzahlkunst. Bonner

Antrittsvorlesung 1946. Bonn: Universitatsverlag Bonn 1947, by Ange-
la Martini

Hans Reichenbach: 'The Tenses of Verbs,' in: Hans Reichenbach: Ele-

ments of Symbolic Logic. New York: The Free Press 1951, § 51, 287-
298, by Maria Reichenbach
Time only belongs to the existent
J. Ellis McTaggart

In this anthology, we present eleven texts on 'time': an existentially
omnipresent, but philosophically evasive concept and phenomenon. As
the title indicates, we have tried to assemble a range of contributions
which, as an ensemble, relate the philosophical perspective onto time
to literary theory's attempts to clarify how time appears, functions and
is construed in and by narratives. Interrelating these distinct disciplin-
ary approaches toward time, our anthology's aim is to contribute, by
way of grouping seminal texts in a fresh context, to a new approach in
which a philosophical, a narratological and a computational perspect-
ive onto narrative-based time construction might be fruitfully integ-
rated. Read in sequence, the eleven contributions therefore also impli-
citly unfold a theoretical argument: if over time we have come to ac-
knowledge that time is indeed a cognitive construct, then building a
functional model that makes the process of construction more transpar-
ent might offer new insights into how time 'works'. And in this respect
the process of narrative time construction offers a particularly fascinat-
ing object of study.
The intricate relationship between time and narrative has met the in-
terest of various disciplines, a fact accounting for the variety of meth-
odological perspectives represented in studies on the subject. Among
the most influential that appeared during the latter half of the 20* cen-
tury count Bakhtin's essay on Chronotopos (1975),1 which investigates
the semiotic interrelation of time and space in fiction and drama, and
Paul Ricoeur's ambitious three-volume investigation Temps et recit
(1983-1985),2 which postulates narration as the privileged mode of hu-
man conceptualization and experience of time. A quarter century later,
the philosophical relevance of the narrative-based conceptualization of
time is re-emphasised by Mark Currie in his recent About Time. Nar-
rative, Fiction and the Philosophy of Time (2007). Curne's Husserl-in-
Bachtm, Michail M. (2008). Chronotopos. Frankfurt/M.: Suhrkamp.
Ricoeur, Paul (1983-1985). Temps et recit, t. 3. Paris: Semi, transl.: Time and
Narrative, 3 vols. Chicago: U of Chicago P 1984-1988.
X Foreword

spired (if not subtly Heideggenan) thesis contra torrentem and contra
Ricoeur is that narrating constitutes less a phenomenon of retrospection,
but rather and primarily one of prospection and projection.3 Its unique
function, according to Currie, lies not in the objectification of memory
and in the backward-oriented re-presentation of the past that enables
the positioning and self-affirmation of ourselves within a biographic
and historic continuum. Rather, what counts is the peculiar dialectic of
an anticipated retrospection upon the present. Phrased in Genettian
terms, the thesis thus proclaims not analepsis, but prolepsis to consti-
tute the differentia specifica of narrating. Moreover, in Currie's per-
spective the philosophical relevance of narration stems from its power
to reflect, on a structural level, the anticipatory mode of being that has
become the signature of modern human society and existence. As Cur-
rie argues, the acknowledgment of this fact is bound to have far-reach-
ing consequences not just for philosophy, but also with a view to estab-
lishing a tense-based narratology that 'takes as its starting point the
possibility of inferring a metaphysics of time from the temporal struc-
ture of narrative.'4
Not everybody will underwrite such far-reaching conclusions and
support the attempt to relate a philosophical definition of time to the
historical and the aesthetic domain, and vice versa. Perhaps the most
outspoken opponent to the idea of approaching the concept of time via
the phenomenology of narration is Mark Currie's namesake Gregory
Currie. In 'Can There Be a Literary Philosophy of Time?' (1999),5 he
reaches a decidedly negative verdict, stating that literature per se can-
not contribute anything of philosophical relevance to our understanding
of time in that whatever is represented and mediated by literary narra-
tion is, by force of its illusionary potential, always experienced as a
present occurrence. However, to narcologists and structuralists this
analytic counter-argument will probably seem reductionist, for in their
perspective what matters about narratives is in any case the how of nar-
rating, rather than the what of the narrative's illusionary story world.
The fruitfulness of an analysis of the former has been clearly demon-
strated by, among other, Genette's narratological theory with its sys-
Currie, Mark (2007). About Time. Narrative, Fiction and the Philosophy of Time.
Edinburgh: Edinburgh UP 2007. - On Currie's book also see: Meister, Jan
Christoph (published 25.01.20092009). 'Erzahlen als vorweggenommene
Rilckschau. Uber Mark Curries About Time; in: IASLonline, <http://www.-> (30.11.2010).
Ibid., 151.
Currie, Gregory (1999). 'Can There Be a Literary Philosophy of Time?' in: The
Arguments of Time, ed. by J. Butterfield. Oxford: Oxford UP, 43-63.
Foreword XI

tematic distinction of duration, frequency and order as the key paramet-

ers for narratonal control and manipulation of how readers experience
time in and through narratives.6
The process aspect of representation is also prominent in the Artifi-
cial Intelligence perspective onto time and in its attempt to reconstruct
the human faculty of handling time information in a bottom-up mode,
rather than deducing it from high-level philosophical premises. Cognit-
ive science's approaches such as Poppel's analysis of Time Perception
(1978)' have systematically laid a foundation for more speculative
models like that presented in Allen's 'Towards a General Theory of
Action and Time' (1984).8 In addition, Al-based computer science has
made significant progress in the practical domain. Here, the goal is to
enable computers to automatically identify, extract and interpret time
information from expository texts in order to support secondary pro-
cesses, such as decision making.'
If computers can be trained to build adequate models of time lines
from factual texts, can they not also be trained to help us identify how
time is structured and 'works' in fiction? This possibility has fascin-
ated practitioners of narrative theory as well as computer scientists for
a decade at least. For example, Burg, Boyle and Lang (2000) attempted
to unravel the notoriously complicated and paradoxical time construct
in Faulkner's A Rose for Emily using constraint logic programming-
Drucker and Nowvieskie (2003) conducted their Temporal Modelling
Project with the aim to 'create a visual scheme and interactive tool set
for the representation of temporal relations in humanities-based or
Genette, Gerard (1972). 'Discours du recit. Essai de methode,' in: G. Genette:
Figures III. Pans: Semi, 65-282, transl: Narrative Discourse: An Essay in Method.
Ithaca: Cornell UP 1980.
Poppel, Ernst (1978). 'Time Perception,' m: R. Held et al. (eds.): Handbook of
Sensory Physiology. Vol. 8: Perception, ed. by R. Held et al. Berlin, Heidelberg,
New York: Spnnger, 713-729. Also see: Poppel, Ernst (1997). 'A Hierarchical
Model of Temporal Perception,' m: Trends in Cognitive Science 1: 56-61.
Allen, James (1984). 'Towards a General Theory of Action and Time,' m: Artificial
Intelligence 23: 123-154.
See Ahn, D., Rantwyk, J., Rijke, M. de (2007). 'A cascaded machine learning
approach to interpreting temporal expressions,' in: Proceedings of Human
Language Technologies: The annual conference of the North American chapter of
the Association for Computational Linguistics (NAACL-HLT 2007), ed. by C.L.
Snider et. al., 420^127. <
Burg, J., Boyle, A., Lang, S. (2000). 'Using Constraint Logic Programming to
Analyze the Chronology in A Rose for Emily; in: Computers and the Humanities
34: 377-392.
XII Foreword

qualitative research, with particular emphasis on the subjective experi-

ence of temporality';" Meister (2005) presented an approach in which
manually tagged time information in literary texts is used to simulate
the Temporality Effect, i.e. the reader's step-by-step construction of a
comprehensive time-line for a fictional world.12 In contrast to these pre-
dominantly human intelligence based approaches, Mam's recent The
Imagined Moment. Time, Narrative, and Computation (2010)" presents
a state-of-the-art AI approach toward machine generation of time lines.
Although the relevant time expressions in his initial corpus of docu-
ments are also manually identified and annotated, the scope of Mam's
project extends further and aims at a combination of machine learning
and corpus narratology. More on Mam's approach can be found in his
chapter concluding this book.
The current volume is one of a number of anthologies devoted to
time that have been published during the last two decades." The motive
to compile such an anthology normally is to open up a field for discus-
sion before devoting oneself to it, critically or analytically, in a more
defined way. As for the present book, this claim cannot be made.
Rather, the anthology which you now hold in hand is the by-product of
a research project that dealt with time construction primarily, if not ex-
clusively, from the perspective of cognitive narratology and reception
theory. Its aim was to explore an approach to narrative time construc-
tion that would open up new perspectives in the empirical analysis of
individual texts as well as of text corpora. More particularly, our inten-
tion was to develop a model that would amount to more than just an-
other partly descriptive, partly interpretive inventory of time-related
phenomena in narratives, such as flashbacks, ellipses, recursion and it-
eration, acceleration or 'stretching' of time, all of which are generally
conceived to be consequences of the transformation of the (logically)
primary story into the secondary phenomenology of discourse. These
effects of transformations of the what, the content of narratives with its

Drucker, J., Nowvieskie, B. (2003). 'The Temporal Modelling Project.'
< (01.12.2010)>.
Meister, Jan Christoph (2005). 'Tagging Time in Prolog: The Temporality Effect
Project,' m: Literary and Linguistic Computing 20: 107-124 (Suppl.).
Mam, Inderjeet (2010). The Imagined Moment. Time, Narrative, and Computation.
Lincoln and London: U of Nebraska P.
Among these are: Mam, Inderjeet, Pustejovsky, James, Gaizauskas, Rob, eds.
(2005). The Language of Time: A Reader. Oxford: Oxford UP; Butterfield, Jeremy,
ed. (1999). The Arguments of Time. Oxford: Oxford UP; Zimmerli, Walter Ch.,
Sandbothe, Mike, eds. (1993). Klassiker der modernen Zeitphilosophie. Darmstadt:
Foreword XIII

'hard-wired' immanent chronology, by the how of narrating, the manip-

ulation of (among other) chronology by techniques of representation,
have received particular attention in contemporary narratology, most
prominently in Genette's structuralist taxonomy of chronological phe-
nomena already mentioned above.
Most structuralist theories in the vein of Genette's consider narrat-
ives as a product rather than a process. With regard to time, this focus
on the outcome of narrative discourse makes it hard to grasp the intric-
ate mechanisms of the genuinely narrative mode of time construction.
Unlike the process of time construction by which we structure the real-
world phenomena presented to us via sensory channels, narrative time
construction is mainly symbol based and controlled. Moreover, it is a
dynamic process: the mental activity of narrative based time construc-
tion has a temporal logic in itself that impacts on the time construct
which we generate for a narrated world. The model of narrative time
construction which we developed paid tribute to this double-layer pro-
cess logic; furthermore, we also made an attempt at simulating the
complex dynamic of a reader's time construction by computational
methods.15 Meanwhile, some aspects of our approach have been super-
seded by Mam's more rigidly formalised approach. However, as he
rightly observed, our project was in any case not 'aimed at evaluating
state-of-the-art tools for automatically building timelines; rather, it in-
volves the use of computers in developing substantial theoretical exten-
sions to particular (mainly structuralist-inspired) approaches.'-
It is primarily with a view to this theoretical interest that the current
anthology was labeled a 'by-product': for in the course of our investig-
ations, we came to realise that two of the major discourses on time dur-
ing the 20th century, that of philosophy of time and that of narrative
theory, have for the most part been conducted in parallel, rather than in
dialogue. This disciplinary insulation from one another is of course
partly owed to a need for methodological focus and consistency in sci-
entific thought. So-called 'literary theories' of philosophical issues are
as fashionable as they are problematic, and literary theories of time, as
already mentioned, form no exception here." Still, the desire for meth-
odological stringency should not necessarily preclude philosophers and
literary theorists from engaging in dialogue. Also, part of the lack
thereof is probably owed to contingent factors and more trivial circum-
stance, and in particular to the inaccessibility of some important 20th"
century German language contributions. These texts, of which some
Meister: Tagging Time (Fn 12).
Mam: The Imagined Moment (Fn 13), 205.
See Currie: Can There Be a Literary Philosophy of Time? (Fn 5), 43-63.
XIV Foreword

clearly bear the signature of time philosophical thought in the phe-

nomenological tradition of Husserl, remained obscured to the English
speaking world for the mere reason that they were never translated.
Against this background, the current anthology, which presents sev-
en German language contributions in English translation for the first
time, attempts to open up a new historical perspective.18 In addition, we
hope that the compilation will also convince its readers that dialogue
across disciplines is indeed fruitful. Time, we believe, is of a twofold
interest to man: (a) as an existential and (b) as a representational phe-
nomenon. Considered in this vein, our motto - McTaggart's dictum
Time only belongs to the existent - should be read with and against the
grain: If time, in the philosophical perspective, is indeed an exclusive
attribute of the existent, then the attribution of time to any content of
our consciousness will necessarily have an ontological consequence.
Now, we obviously do respond to narratives and their contents as
something that has temporality - and if McTaggart is right then this af-
fective and cognitive response to narratives amounts to an implicit at-
tribution of time which elevates not only factual, but also and in partic-
ular fictional representational contents to the status of the existent. In
other words, the temporality effect and the reality effect go hand in

Philosophy of Time

The first four articles presented in this book deal with time from a
philosophical perspective, investigating it firstly in relation to time ex-
pressions in natural language, secondly as a category intricately related
to our sense of self-awareness and identity, and thirdly as a life-world
phenomenon which structures and influences the encounter of our hu-
man consciousness with the world.
Hans Reichenbach's 'The Tenses of Verbs' (1947) proceeds from
the observation that tenses of verbs determine time with reference to
the time point of the speech act, differentiating accordingly between
the point of speech, the point of the event, and the point of reference.
He then demonstrates the specific usage of different tenses in different
languages (with a focus on English) to express time relations and ar-
In order to serve this purpose, we have decided on a translation approach which at-
tempts to preserve the original German document's argumentative and rhetonc style
as far as tenable in order to honour its historicity. Admittedly, given the well-
known means of German grammar to support complex hypotactic structures, this
decision is not unproblematic. In addition, one of the consequences of our approach
to translation is to accept terminological inconsistency across contributions.
Foreword XV

rives at thirteen possible forms of ordering the three time points. Nine
of these are termed fiindamental forms, noting at the same time that in
English, for example, only six recognised grammatical tenses exist and
that the tenses for which a language has no established forms are there-
fore expressed by paraphrase. Reichenbach draws the conclusion that
'logical categories were not clearly seen in the beginnings of language
but were the results of long developments; we therefore should not be
astonished if actual language does not always fit the schema which we
try to construct in symbolic logic.'1'
Such faculty of symbolic representation of time presupposes the
faculty of its mental representation. The most fundamental implication
of our human ability to represent time mentally is discussed in Peter
Bien's 'Time Experience and Personhood' (1986).™ The concept of
personhood, as Bien shows, is dependent on time consciousness,
which, in turn, is founded upon self-awareness. Self-awareness and
time consciousness enable us to experience the present against the
background of an appropriated past, and in the light of a project for the
future, thus creating a sense of identity. In a causal perspective, the ap-
propriation of our past and the conceptualization of our future, as Bien
shows, are necessary pre-conditions for the emergence of a sense of
identity, and both are facilitated by an integration of past experiences
and future projections into a coherent self-narrative: 'We expect from a
person that his projects match the explanatory story that constitutes his
present identity. I have to plan what I want to be in the future in the
light of what I have become in the past.'21
Peter Jamch's 'Constituting Time through Action and Discourse'
(1996)- investigates how we actively constitute time through symbolic
as well as concrete acting, raising the question of primacy. His ap-
proach is to look at how we talk about time and temporality; his con-
clusion is that time consciousness is indeed a function of acting, which
always comes first. In Jamch's view, time-statements are actually of a
meta-linguistic order; they are 'merely a shorter, more elegant way of
describing temporal statements' which in themselves remain subject to
the 'methodical primacy of the ability to act over the constitution of a

Reichenbach, 11 in the current volume.
Translated from the German original: Bien, Peter (1986). 'Zeiterfahrung und
Personahtat,' m: H. Burger (ed.): Zeit, Natur und Mensch. Beitrage von
Wissenschaftlern zum Thema "Zeit". Berlin: Berlin Verlag Arno Spitz, 261-281.
Bien, 26 m the current volume.
Translated from the German ongmal: Jamch, Peter (1996). 'Die Konstitution der
Zeit durchHandeln und Reden,' in: Kodikas/ Code 19.1-2: 133-147.
XVI Foreword

discourse which allows us to identify the uniqueness of incidents in

their chronological sequence by using concepts.'"
In 'Time, Tense, and Typology' (1996) Le Poidevin discusses a
problem initially raised by McTaggart whose 1908 article 'The Unreal-
ity of Time' counts among the most influential contributions to modern
time philosophy. McTaggart had introduced the distinction between
what he termed the 'A-senes' of past, present, future, and the 'B-ser-
ies' of before-after relations among facts. He then argued that a sub-
jective and 'tensed' conception of time based on the A-senes descrip-
tion is logically inconsistent with an untensed and objective B-senes
conception - and because of this logical paradox, as McTaggart con-
cluded, time is logically non-real. In addition to this provocative
thought, McTaggart also tried to prove that our human notion of time is
in all likelihood based on the primacy of the A-senes over the B-senes,
and that B-senes facts can be reduced to A-senes facts. This second
idea of McTaggart's is the basis of the so-called 'tensed' theory of
time. By conducting two thought expenments on 'disumfied time' and
'branching time', Le Poidevin demonstrates that while the former ex-
penment does not put the 'tensed' theory of time into question, the lat-
ter expenment casts 'doubt on the doctrine that the B-senes is redu-
cible to the A-senes. And if we can cast doubt on this doctrine, then we
weaken the plausibility of the tensed theory of time.'- In other words:
while the reality of 'tensed' time might be hard to prove, we cannot lo-
gically rule out the possibility for the existence of an objective, 'un-
tensed' time.

Nanative Theory of Time

The articles in the second group consider time as it appears and func-
tions in the context of literary narratives; in doing so, they attempt to
elucidate how effects of time and temporality influence and are them-
selves shaped by narration. All of the five contnbutions contained in
this section have been translated from the onginal German and are now
available in English for the first time.
Giinther Miiller's 1946 inaugural lecture The Significance of Time
in Narrative Art* introduces the fundamental opposition of time of
narration vs. narrated time (Erzahlzeit / erzahlte Zeit), a distinction

Jamch, 47 m the current volume.
Le Poidevin, 65 in the current volume.
Translated from the original German: Miiller, Gunther (1947). Die Bedeutung der
Zeit in derErzahlkunst. Bonn: Umversitatsverlag.
Foreword XVII

that has become a tenet of all later theories of narrative irrespective of

methodological brand. Miiller proceeds from the observation that nar-
ration is always and necessarily a re-presentation of something that has
passed and is absent. Also, no narration represents everything: selec-
tion cannot be avoided, for otherwise no narration could ever bridge
the temporal gap between the representational content, which is located
in the past, and the process of narration, which has its own temporality
that extends from a point in time in our existential present. Therefore,
narrated time is necessarily marked by contraction; moreover, the
strength of this contraction vanes: 'All narrating is a narrating [..] of
something that is not narration, but a process in life'; something that
happens in the 'spatio-temporal world, even if it is an inner experience
of the soul. If a narration contracts clock time such that the spatio-tem-
poral condition is contracted away, then it makes an obvious choice
with an interpreting effect' and which comes out of a signifying 'atti-
tude to reality.'-
In 'The Timelessness of Poetry' (1955)," Kate Hamburger points to
the essential a-temporality of representational content. Three observa-
Miiller, 69 in the current volume.
Translated from the original German: Hamburger, Kate (1955). 'Die Zeitlosigkeit
der Dichtung,' in: Deutsche Vierteljahrsschrift fur Literaturwissenschaft und
Geistesgeschichte 29: 413^126. Note that this article appeared immediately pnor to
Hamburger's seminal Die Logik der Dichtung. Stuttgart: Klett ([1955] 21957),
transl: The Logic of Literature. Bloommgton: Indiana UP. In the article (which is,
pro forma, a reply to a polemic criticism by Herbert Seidler) Hamburger outlines
her influential definition of fiction as a form of discourse practice that is based on a
fundamentally different referential logic, and she does so by way of detailing the
particularly 'illogical' use of tense and time in fictional literature. - A remark on a
translation issue that exemplifies the difficulty to render adequate translations of
German texts from the mid-1950s: Similar to its English counterpart the German
term 'Dichtung' used in Hamburger's title refers to literature in general, i.e., to
poetry in the wider sense, as well as to lync poetry specifically. However, while the
author does occasionally use 'Dichtung' in the former, relatively neutral sense of
literature at large, her argumentative focus is almost exclusively on epic poetry, and
more particularly, on third person narratives as the only 'pure' form of fictional
narration. Against the backdrop of our contemporary understanding of 'literature'
as a genenc category for fictional as well as factual texts the clear-cut translation of
'Dichtung' as 'literature' (which the translators of Hamburger's book opted for) is
thus somewhat a-histoncal in that it tends to downplay the aspects of referential
autonomy and authorial control, in short, the specificity of literature qua poetry
which were of key concern to Hamburger. Moreover, if applied consistently across
the translation of the current article, the use of 'literature' in conjunction with genre
adjectives would result in (if nothing else) stylistically dubious formulations such
as 'lync/dramatic literature' for 'lynsche/dramatische Dichtung'.
XVIII Foreword

tions substantiate this claim: one, what is represented as a (fictional or

factual) event must in logical terms necessarily be positioned outside of
any empirical time line. Two, novels need not explicitly narrate time -
and if they fail to do so, then time is in fact simply not present as rep-
resentational content. This potential for time neutrality is in significant
contrast to our real-world experience, which is always embedded in
and affected by time. Three, fiction may be presented using verbs - yet
when used in fictional mode verbs loose their function to denote tem-
porality, and their tense is reduced to an abstract grammatical feature:
for example, the simple past used in a fictional narration is normally in-
terpreted as referring to a fictional present. Hamburger's conclusion is
that time, like space, turns into something ideal when encountered un-
der the conditions of fictional mimesis - both are no longer forms of
something that we experience to exist, but forms of something that we
Our translation of Eberhard Lammert's 'The Time References of
Narration'- presents a short excerpt from his book Bauformen des
Erzahlens, one of the most important pre-structuralist systematic tax-
onomies of narrative forms in which the author paid particular attention
to compositional features such as flashbacks, flash-forwards, iteration
etc. This 1955 doctoral thesis, currently available in its 9th unaltered
edition, is not only one of the most often republished books by a Ger-
man Humanities scholar: it has also had an obvious influence on
Genette's analysis and taxonomy of phenomena of 'order'. In the ex-
cerpt contained in this anthology, Lammert identifies succession as the
most general compositional principle which narrative art initially
shares with every linguistic utterance. However, in order to overcome
the monotony of this base level sequentiahty, and in order to also high-
light the aesthetic and semantic intentionality of artful representation as
a defined whole the narrator has to apply the forming principle of
omission. This unavoidable and necessary 'pausing, contraction and
omitting on the part of the narrator not only accentuates specific phases
in the sequence of events; it is precisely this activity which helps the
entire narrated contents to emanate from the monotony of pure succes-
sion as something newly created:™
The thrust of de Toro's 'Time Structure in the Contemporary Nov-
el' (1986) is indicated by its descriptive subtitle 'Suggestion of an En-

Translated from the original German: Lammert, Eberhard ([1955] 91990). 'Die
Zeitbezilge des Erzahlens,' m: E. Lammert: Bauformen des Erzahlens. Stuttgart:
Metzler, 19-24.
Lammert, 106 in the current volume.
Foreword XIX

hanced Model of Time Usage Analysis according to G. Genette'.30 By

expanding Genette's terms ordre (translated into 'time arrangement' by
de Toro), duree (duration) and frequence (frequency), the author sug-
gests a more differentiated terminological apparatus that enables the
analyst to describe the different forms of time usage. With the aid of
this expanded taxonomy, de Toro demonstrates how time structure in-
fluences the constitution of meaning within a literary text (intra-textual
function) as well as how aspects of time influence the reception of the
reader (extra-textual function). Neither is time construction a mere ex-
ercise in building a mimetic representation, nor is time manipulation a
mere exercise in creating aesthetic form - both have a significant se-
mantic effect and hermeneutic consequence: 'The question of the func-
tion of time usage [...] is seen here as an instrument with an extra-textu-
al and an intra-textual function, e.g. irony or perspectivation, and not
just as a mere sequence of actions in time (there was . . . and then).
From the intra textual point of view, we consider each type of time or-
ganization as important and as having far-reaching consequences to the
interpretation of the text.'31
In 'Story-time and Fact-sequence-time' (1991), Harweg suggests to
extend Miiller's traditional dichotomy of narrated time and time of
narration into a four-level model that distinguishes between (a) fact-
sequence-time, (b) narrated time, (c) discourse-time (the equivalent of
Miiller's time of narration), and its logical counterpart in terms of the
reception process, (d) observation-time. In his subsequent delibera-
tions, Harweg focuses on the first two categories. With regard to the
former, the author argues item fact-sequence-time constitutes an ontolo-
gical as well as an epistemological sub-phenomenon in which the op-
positions of material vs. formal and subjective vs. objective play a role.
With regard to the second category, Harweg points out that authors of
narrative texts do not narrate fact-sequence-time, but narrated time.
The latter is distinguished from the former in the main by two criteria,
namely selection and orientation. The emulation of fact-sequence-time
in narrated time is termed longitudinal; the point-by-point co-presenta-
tion of parallel fact-sequences which results in spatial distribution is
called latitudinal, and the partial co-presentation of such sequences

Translated from the original German: de Toro, Alfonso (1986). 'Versuch ernes
erweiterten Modells filr die Analyse von Zeitstrukturen nach G. Genette,' m: A. de
Toro: Die Zeitstruktur im Gegenwartsroman am Beispiel von G. Marcia Mdrquez'
Cien anos de soledad, M. Vargas Llosas La casa verde und A. Robbe-Grillets La
maison de rendez-vous. Tubingen: Gunter Narr, 26^17. In our translation we have
amended the chapter title to reflect its contextualization in de Toro's book.
de Toro, 110m the current volume.
XX Foreword

that evokes a scalar gradation from concrete to abstract semantic rela-

tions among facts is termed altitudinal. Further variations can be
achieved in the form of progredient and regredient arrangements
among fact-sequence-time and narrated time.

Towards Computational Models of Narrated Time

The two concluding contributions appear for the first time here. Both
focus on what might be considered the underlying - albeit not always
explicit - assumption shared by all texts contained in this volume:
time, in the form that we experience it, is not directly accessible to our
perception, but a construct of the human mind. And if that is indeed the
case, then narratives and the activity of generating as well as decoding
narrative representations ought to be modeled as dynamic processes.
This is where computers come into play.
Meister's contribution 'The Temporality Effect. Towards a Process
Model of Narrative Time Construction' tries to address a methodolo-
gical lacuna mentioned earlier on, namely the hitherto insufficient at-
tempts to relate cognitive and computational models of text-controlled
time construction jointly and evenly to, on the one hand, philosophy of
time and, on the other hand, to narrative theory. Whether the article
and the decision to base such a model on McTaggart's and Husserl's
time philosophy as well as on narcological categories is a step into
the right direction, remains for the reader to judge.
The concluding chapter of this anthology presents Mam's insightful
'The Flow of Time in Narrative. An Artificial Intelligence
Perspective'. Time, as the author reminds us, does not appear alone in
narrative; it is wound up with events, and involves relationships that
hinge on modality and point-of-view. This accounts for a subjective no-
tion of time, where the position of an event in time is dynamic and
changes relative to the speaker. This notion must be reconciled with
another, so-called objective notion of time where the events are ordered
in a fixed, static fashion. Against this background, Mam introduces an
AI perspective that integrates these concepts within a coherent frame-
work. As he demonstrates, AI models based on narcological theory
and findings from cognitive sciences are able to tell us what sorts of
reasoning about time and events an intelligent agent can carry out.
These findings, the author suggests, should be taken further in two
ways: one, by conducting psychological experiments to learn more
about how humans construct time representations; two, by scaling up
the corpora in which time lines are identified. Mam considers his in-
vestigations 'only a first step [... ] in a much richer examination of time
Foreword XXI

in narrative, within the framework of an empirical discipline of corpus

narratology, where multimillion-word collections of narrative text are
analysed using sophisticated Al-based timelining tools.'"


This anthology may have come into being as the side-product of a re-
search project. However, it quickly outgrew the status of a stepchild,
both in terms of the challenges which it presented, and in terms of the
intellectual reward that it gave. The intellectual reward we hope to be
able to share; as for the challenges one thing is clear: the editors would
not have been able to meet them without the generous support that we
enjoyed. We thank the German Research Council (DFG) for the finan-
cial grant awarded to the 'Temporality Effect' project of which this an-
thology forms an integral part, and we thank our institution, the Uni-
versity of Hamburg, for providing the material, personal and adminis-
trative infrastructure on which academic research work is dependent.
A particularly big 'Thank you' is owed to the team of graduate stu-
dents and assistants who helped with the editing of the manuscript:
Lisa Griinhage, Lena Schiich and Rike Lohmann. We also thank our
translators, Ingnd Launen and Alexander Starntt, for their remarkable
Finally, we would also like to express our gratitude to the col-
leagues whose intellectual support and engagement in our project was
extremely motivating and helpful - the members of the Interdisciplin-
ary Center for Narratology (ICN) at the University of Hamburg, and
those of its precursor, the DFG 'Narratology Research Group' (FGN).
Among these are two colleagues we would like to mention by name:
Rolf Krause, a steadfast and critical ally in matters equally narrative,
digital and humanist, and Giinter Dammann who, among other things,
introduced us to narratology and who enabled the project as a whole.
Both played a decisive role in discussing and selecting the articles and
in conceptualizing the anthology which we have edited and which you
now hold in your hands.

Hamburg, December 2010

Mam, 235 in the current volume.

The Tenses of Verbs

A particularly important form of token-reflexive symbol is found in the
tenses of verbs. The tenses determine time with reference to the time
point of the act of speech, i.e., of the token uttered. A closer analysis
reveals that the time indication given by the tenses is of a rather com-
plex structure.
Let us call the time point of the token the point of speech. Then the
three indications, 'before the point of speech', 'simultaneous with the
point of speech', and 'after the point of speech', furnish only three
tenses; since the number of verb tenses is obviously greater, we need a
more complex interpretation. From a sentence like 'Peter had gone' we
see that the time order expressed in the tense does not concern one
event, but two events, whose positions are determined with respect to
the point of speech. We shall call these time points the point of the
event and the point of reference. In the example the point of the event
is the time when Peter went; the point of reference is a time between
this point and the point of speech. In an individual sentence like the
one given it is not clear which time point is used as the point of refer-
ence. This determination is rather given by the context of speech. In a
story, for instance, the series of events recounted determines the point
of reference which in this case is in the past, seen from the point of
speech; some individual events lying outside this point are then re-
ferred, not directly to the point of speech, but to this point of reference
determined by the story. The following example, taken from W.
Somerset Maugham's Of Human Bondage, may make these time rela-
tions clear:

But Philip ceased to think of her a moment after he had settled down in his
carriage. He thought only of the future. He had written to Mrs. Otter, the
massiere to whom Hayward had given him an introduction, and had in his
pocket an invitation to tea on the following day.

The series of events recounted here in the simple past determine the
point of reference as lying before the point of speech. Some individual
events, like the settling down in the carriage, the writing of the letter,
2 Hans Reichenbach

and the giving of the introduction, precede the point of reference and
are therefore related in the past perfect. Another illustration for these
time relations may be given by a historical narrative, a quotation from

In 1678 the whole face of things had changed ... eighteen years of misgov-
ernment had made the ... majority desirous to obtain security for their liber-
ties at any risk. The fury of their returning loyalty had spent itself in its first
outbreak. In a very few months they had hanged and half-hanged, quartered
and emboweled, enough to satisfy them. The Roundhead party seemed to
be not merely overcome, but too much broken and scattered ever to rally
again. Then commenced the reflux of public opinion. The nation began to
find out to what a man it had intrusted without conditions all its dearest in-
terests, on what a man it had lavished all its fondest affection.

The point of reference is here the year 1678. Events of this year are re-
lated in the simple past, such as the commencing of the reflux of public
opinion, and the beginning of the discovery concerning the character of
the king. The events preceding this time point are given in the past per-
fect, such as the change in the face of things, the outbreaks of cruelty,
the nation's trust in the king.
In some tenses, two of the three points are simultaneous. Thus, in
the simple past, the point of the event and the point of reference are
simultaneous, and both are before the point of speech; the use of the
simple past in the above quotation shows this clearly. This distin-
guishes the simple past from the present perfect. In the statement 'I
have seen Charles' the event is also before the point of speech, but it is
referred to a point simultaneous with the point of speech; i.e., the
points of speech and reference coincide. This meaning of the present
perfect may be illustrated by the following quotation from Keats:

Much have I traveled in the realms of gold,

And many goodly states and kingdoms seen;
Round many western islands have I been
Which bards in fealty to Apollo hold.

Comparing this with the above quotations we notice that here obvi-
ously the past events are seen, not from a reference point situated also
in the past, but from a point of reference which coincides with the point
of speech. This is the reason that the words of Keats are not of a narrat-
ive type but affect us with the immediacy of a direct report to the read-
er. We see that we need three time points even for the distinction of
tenses which, in a superficial consideration, seem to concern only two
The Tenses of Verbs 3

time points. The difficulties which grammar books have in explaining

the meanings of the different tenses originate from the fact that they do
not recognize the three-place structure of the time determination given
in the tenses.1
We thus come to the following tables, in which the initials ' £ ' , 'R\
and '5" stand, respectively, for 'point of the event', 'point of
reference', and 'point of speech', and in which the direction of time is
represented as the direction of the line from left to right:

In some tenses, an additional indication is given concerning the time

extension of the event. The English language uses the present participle
to indicate that the event covers a certain stretch of time. We thus ar-
rive at the following tables:

In J.O.H. Jespersen's excellent analysis of grammar (The Philosophy of Grammar.
New York: Holt 1924) I find the three-point structure indicated for such tenses as
the past perfect and the future perfect (256), but not applied to the interpretation of
the other tenses. This explains the difficulties which even Jespersen has in
distinguishing the present perfect from the simple past (269). He sees correctly the
close connection between the present tense and the present perfect, recognizable in
such sentences as 'now I have eaten enough'. But he gives a rather vague definition
of the present perfect and calls it 'a retrospective variety of the present'.
4 Hans Reichenbach

The extended tenses are sometimes used to indicate, not duration of the
event, but repetition. Thus we say 'women are wearing larger hats this
year' and mean that this is true for a great many instances. Whereas Eng-
lish expresses the extended tense by the use of the present participle, oth-
er languages have developed special suffixes for this tense. Thus the
Turkish language possesses a tense of this kind, called muzari, which in-
dicates repetition or duration, with the emphasis on repetition, including
past and future cases. This tense is represented by the diagram

An example of this tense is the Turkish word 'goriirum', translatable as

'I usually see'. The syllable 'gor' is the root meaning 'see', 'iir' is the
suffix expressing the muzari, and the 'iim' is the suffix expressing the
first person T. 2 The sentence 'I see' would be in Turkish 'goruyorum';
the only difference from the preceding example is given by the inflec-
tion 'iiyor' in the middle of the word, expressing the present tense. The
Greek language uses the aorist to express repetition or customary oc-
currence in the present tense. The aorist, however, is originally a
nonextended past tense, and has assumed the second usage by a shift of
meaning; in the sense of the extended tense it is called gnomic aorist?
German and French do not possess extended tenses, but express such
meanings by special words, such as the equivalents of 'always', 'habitu-
ally', and so on. An exception is the French simple past. The French lan-
guage possesses here two different tenses, the imparfait and the passe
deflni. They differ in so far as the imparfait is an extended tense, whereas
th& passe deflni is not. Thus we have

Turkish vowels with two dots are pronounced like the German vowels ' o' and 'ii'.
This shift of meaning is explainable as follows: One typical case of the past is
stated, and to the listener is left the inductive inference that under similar
conditions the same will be repeated in the future. A similar shift of meaning is
given in the English 'Faint heart never won fair lady'. Cf. W.W. Goodwin: Greek
Grammar. Boston: Ginn 1930, 275.
The Tenses of Verbs 5

We find the same distinction in Greek, the Greek imperfect corres-

ponding to the French imparfait, and the Greek aorist, in its original
meaning as a past tense, corresponding to the French passe deflni. Lan-
guages which do not have a passe deflni sometimes use another tense
in this meaning; thus Latin uses the present perfect in this sense (histor-
ical perfect).
We may add here the remark that the adjective is of the same logical
nature as the present participle of a verb. It indicates an extended tense.
If we put the word 'hungry', for instance, in the place of the word 'see-
ing' in our tables of extended tenses, we obtain the same extended
tenses. A slight difference in the usage is that adjectives are preferred if
the duration of the event is long; therefore adjectives can often be inter-
preted as describing permanent properties of things. The transition to
the extended tense, and from there to the permanent tense, is seen in
the examples 'he produces', 'he is producing', 'he is productive'.
When we wish to express, not repetition or duration, but validity at
all times, we use the present tense. Thus we say 'two times two is
four'. There the present tense expressed in the copula 'is' indicates that
the time argument is used as a free variable; i.e., the sentence has the
meaning 'two times two is four at any time'. This usage represents a
second temporal function of the present tense.
Actual language does not always keep to the schemas given in our
tables. Thus the English language uses sometimes the simple past
where our schema would demand the present perfect. The English
present perfect is often used in the sense of the corresponding extended
tense, with the additional qualification that the duration of the event
reaches up to the point of speech. Thus we have here the schema

In the sense of this schema we say, for instance, T have known him for
ten years'. If duration of the event is not meant, the English language
then uses the simple past instead of the present perfect, as in T saw him
ten years ago'. German and French would use the present perfect here.
When several sentences are combined to form a compound sen-
tence, the tenses of the various clauses are adjusted to one another by
certain rules which the grammarians call the rules for the sequence of
6 Hans Reichenbach

tenses. We can interpret these rules as the principle that, although the
events referred to in the clauses may occupy different time points, the
reference point should be the same for all clauses - a principle which,
we shall say, demands the permanence of the reference point. Thus, the
tenses of the sentence, 'I had mailed the letter, when John came and
told me the news', may be diagramed as follows:

(1) 1st clause: Ei — Rl — S

2nd clause: R2, E2— S
3rd clause: R3, E3— S

Here the three reference points coincide. It would be incorrect to say, 'I
had mailed the letter when John has come'; in such a combination the
reference point would have been changed. As another example, con-
sider the compound sentence, 'I have not decided which train I shall
take'. That this sentence satisfies the rule of the permanence of the ref-
erence point is seen from the following diagram:

(2) 1st clause: Ei — S, R,

2nd clause: S, R2 — E2

Here it would be incorrect to say: 'I did not decide which train I shall
When the reference point is in the past, but the event coincides with
the point of speech, a tense R — S, E is required. In this sense, the form
'he would do' is used, which can be regarded as derived from the
simple future 'he will do' by a back-shift of the two points R and E. We
say, for instance, 'I did not know that you would be here'; this sentence
represents the diagram:

(3) 1st clause: R b E, — S

2nd clause: R2 — S, E2

The form 'I did not know that you were here' has a somewhat different
meaning; it is used correctly only if the event of the man's being here
extends to include the past time for which the T did not know' is
stated, i.e., if the man was already here when I did not know it. Incid-
entally, in these sentences the forms 'would be' and 'were' do not have
a modal function expressing irreality; i.e., they do not represent a con-
ditional or a subjunctive, since the event referred to is not questioned.
The nonmodal function is illustrated by the sentence 'I did not know
The Tenses of Verbs 7

that he was here', for which the form 'that he were here' appears incor-
When a time determination is added, such as is given by words like
'now' or 'yesterday', or by a nonreflexive symbol like 'November 7,
1944', it is referred, not to the event, but to the reference point of the
sentence. We say, 'I met him yesterday'; that the word 'yesterday'
refers here to the event obtains only because the points of reference and
of event coincide. When we say, 'I had met him yesterday'; what was
yesterday is the reference point, and the meeting may have occurred the
day before yesterday. We shall speak, therefore, of the positional use
of the reference point; the reference point is used here as the carrier of
the time position. Such usage, at least, is followed by the English lan-
guage. Similarly, when time points are compared by means of words
like 'when', 'before', or 'after', it is the reference points to which the
comparison refers directly, not the events. Thus in the above example
(1) the time points stated as identical by the word 'when' are the refer-
ence points of the three clauses, whereas the event of the first clause
precedes that of the second and the third. Or consider the sentence,
'How unfortunate! Now that John tells me this I have mailed the letter'.
The time stated here as identical with the time of John's telling the
news is not the mailing of the letter but the reference point of the
second clause, which is identical with the point of speech; and we have
here the schema:

(4) 1st clause: S, Rl5 Ei

2nd clause: E2 — S, R2

For this reason it would be incorrect to say, 'Now that John tells me
this I mailed the letter'.
If the time relation of the reference points compared is not identity,
but time sequence, i.e., if one is said to be before the other, the rule of
the permanence of the reference point can thus no longer be main-
tained. In 'he telephoned before he came' Rl is said to be before R2;
but, at least, the tenses used have the same structure. It is different with
the example, 'he was healthier when I saw him than he is now'. Here
we have the structure:

(5) 1st clause: R b E, — S

2nd clause: R2, E2— S
3rd clause: S, R3, E3
8 Hans Reichenbach

In such cases, the rule of the permanence of the reference point is re-
placed by the more general rule of the positional use of the reference
point. The first rule, therefore, must be regarded as representing the
special case where the time relation between the reference points com-
pared is identity.
Incidentally, the English usage of the simple past where other
languages use the present perfect may be a result of the strict adherence
to the principle of the positional use of the reference point. When we
say, 'this is the man who drove the car', we use the simple past in the
second clause because the positional principle would compel us to do
so as soon as we add a time determination, as in 'this is the man who
drove the car at the time of the accident'. The German uses here the
present perfect, and the above sentence would be translated into 'dies
ist der Mann, der den Wagen gefahren hat'. Though this appears more
satisfactory than the English version, it leads to a disadvantage when a
time determination is added. The German is then compelled to refer to
the time determination, not to the reference point, but to the event, as in
'dies ist der Mann, der den Wagen zur Zeit des Ungliicksfalles
gefahren hat'. In such cases, a language can satisfy either the principle
of the permanence of the reference point or that of the positional use of
the reference point, but not both.
The use of the future tenses is sometimes combined with certain de-
viations from the original meaning of the tenses. In the sentence 'Now I
shall go' the simple future has the meaning S,R — E; this follows from
the principle of the positional use of the reference point. However, in
the sentence T shall go tomorrow' the same principle compels us to in-
terpret the future tense in the form S—R,E. The simple future, then, is
capable of two interpretations, and since there is no prevalent usage of
the one or the other we cannot regard one interpretation as the correct
one.4 Further deviations occur in tense sequences. Consider the sen-
tence: 'I shall take your photograph when you come'. The form 'when
you will come' would be more correct; but we prefer to use here the
present tense instead of the future. This usage may be interpreted as
follows. First, the future tense is used in the first clause in the meaning
S — R,E; second, in the second clause the point of speech is neglected.
The neglect is possible because the word 'when' refers the reference
point of the second clause clearly to a future event. A similar anomaly
is found in the sentence, 'We shall hear the record when we have

The distinction between the French future forms je vais voir and je verrai may
perhaps be regarded as representing the distinction between the order S, R — E
and the order S — R, E.
The Tenses of Verbs 9

dined', where the present perfect is used instead of the future perfect
'when we shall have dined'.5
Turning to the general problem of the time order of the three points,
we see from our tables that the possibilities of ordering the three time
points are not exhausted. There are on the whole 13 possibilities, but
the number of recognized grammatical tenses in English is only 6. If
we wish to systematize the possible tenses we can proceed as follows.
We choose the point of speech as the starting point; relative to it the
point of reference can be in the past, at the same time, or in the future.
This furnishes three possibilities. Next we consider the point of the
event; it can be before, simultaneous with, or after the reference point.
We thus arrive at 3 • 3 = 9 possible forms, which we call fundamental
forms. Further differences of form result only when the position of the
event relative to the point of speech is considered; this position, how-
ever, is usually irrelevant. Thus the form S — E — R can be distin-
guished from the form S, E — R; with respect to relations between S
and R on the one hand and between R and E on the other hand, how-
ever, these two forms do not differ, and we therefore regard them as
representing the same fundamental form. Consequently, we need not
deal with all the 13 possible forms and may restrict ourselves to the 9
fundamental forms.
For the 9 fundamental forms we suggest the following terminology.
The position of R relative to S is indicated by the words 'past',
'present', and 'future'. The position of E relative to R is indicated by
the words 'anterior', 'simple', and 'posterior', the word 'simple' being
used for the coincidence of R and E. We thus arrive at the following

5 In some books on grammar we find the remark that the transition from direct to
indirect discourse is accompanied by a shift of the tense from the present to the
past. This shift, however, must not be regarded as a change in the meaning of the
tense; it follows from the change in the point of speech. Thus T am cold' has a
point of speech lying before that of T said that I was cold'.
10 Hans Reichenbach

Structure New Name Traditional Name

E—R— S Anterior past Past perfect

E,R — S Simple past Simple past
R — E—S -i
R — S,E L Posterior past —
E — S, R Anterior past Present perfect
S,R,E Simple present Present
S,R — E Posterior present Simple future
S—E — R~}
S,E — R I Anterior future Future perfect
£ — S — RJ
S — R,E Simple future Simple future
S—R—E Posterior future —

We see that more than one structure obtains only for the two retro-
gressive tenses, the posterior past and the anterior future, in which the
direction S — R is opposite to the direction R — E. If we wish to dis-
tinguish among the individual structures we refer to them as the first,
second, and third posterior past or anterior future.
The tenses for which a language has no established forms are ex-
pressed by transcriptions. We say, for instance, 'I shall be going to see
him' and thus express the posterior future S — R — E by speaking, not
directly of the event E, but of the act of preparation for it; in this way
we can at least express the time order for events which closely succeed
the point of reference. Languages which have a future participle have
direct forms for the posterior future. Thus the Latin 'abiturus ero' rep-
resents this tense, meaning verbally 'I shall be one of those who will
leave'. For the posterior past R — E — S the form 'he would do' is
used, for instance in 'I did not expect that he would win the race'. We
met with this form in an above example where we interpreted it as the
structure R — S,E; but this structure belongs to the same fundamental
form as R — E — S and may therefore be denoted by the same name.
Instead of the form 'he would do', which grammar does not officially
recognize as a tense,6 transcriptions are frequently used. Thus we say,
'I did not expect that he was going to win the race', or, in formal writ-
ing, 'the king lavished his favor on the man who was to kill him'. In the
It is sometimes classified as a tense of the conditional mood, corresponding to the
French conditional. In the examples considered above, however, it is not a
conditional but a tense in the indicative mood.
The Tenses of Verbs 11

last example, the order R — E — S is expressed by the form 'was to

kill', which conceives the event E, at the time R, as not yet realized, but
as a destination.
Incidentally, the historical origin of many tenses is to be found in
similar transcriptions. Thus 'I shall go' meant originally 'I am obliged
to go'; the future-tense meaning developed because what I am obliged
to do will be done by me at a later time.7 The French future tense is of
the same origin; thus the form 'je donnerai', meaning T shall give', is
derived from 'je donnerai', which means T have to give'. This form of
writing was actually used in Old French.8 The double function of
'have', as expressing possession and a past tense, is derived from the
idea that what I possess is acquired in the past; thus 'I have seen'
meant originally T possess now the results of seeing', and then was in-
terpreted as a reference to a past event.9 The history of language shows
that logical categories were not clearly seen in the beginnings of lan-
guage but were the results of long developments; we therefore should
not be astonished if actual language does not always fit the schema
which we try to construct in symbolic logic. A mathematical language
can be coordinated to actual language only in the sense of an approxim-

In Old English no future tense existed, and the present tense was used both for the
expression of the present and the future. The word 'shall' was used only in the
meaning of obligation. In Middle English the word 'shall' gradually assumed the
function of expressing the future tense. Cf The New English Dictionary, Oxford
1914, Vol. VIII, Pt. 2, S-Sh, p. 609, col. 3.
This mode of expressing the future tense was preceded by a similar development of
the Latin language, originating in vulgar Latin. Thus instead of the form 'dabo',
meaning the future tense T shall give', the form 'dare habeo' was used, which
means T have to give'. Cf. Ferdinand Brunot: Precis de grammaire historique de
la langue francaise. Paris: Masson 1899, 434.
This is even more apparent when a two-place function is used. Thus T have
finished my work' means originally T have my work finished', i.e., T possess my
work as a finished one'. Cf. The New English Dictionary. Oxford 1901, Vol. V, Pt.
I, H, p. 127, col. 1-2. The German still uses the original word order, as in Teh habe
meine Arbeit beendet'.

Time Experience and Personhood

We regard ourselves and other people as persons. The term person has
many facets. There are many conditions of personhood that interact in
a complicated way. One of these conditions is that a being has to have
a certain relation to the time span during which it lives in order to be
considered as a person. In the following, I will deal with this aspect of
personhood. In the course of the analysis, the interrelation of person-
hood with other classical aspects will come to the fore.
My reflections are divided into two parts. In the first part, I explore
the question of the formal or general conditions for a being not only to
factually live in time, but also to have a consciousness of the temporal
dimensions of its existence. In the second part I attempt to develop the
idea that persons are capable of creating a particular kind of identity by
rendering a special kind of unity to their own lives stretched out over

Formal Conditions for Time Consciousness

What are actually the conditions for a being to have an entity, a con-
sciousness of time? In order to answer this question, I will begin with a
minimal model of a relation to time, and will then enrich this model
step by step until we can recognise our full time consciousness in it.
That means I will do something similar to what Strawson calls 'de-
scriptive metaphysics'1: to make general conditions of our experience
visible by reflecting on what would happen if these conditions were not
Let us begin with the examination of objects. For persons are phys-
ical entities and, in this sense, objects, although the point of the term
person is, that they are not just objects. Objects have a history; they are
subjected to change. In order to describe their changes, it suffices to
name the time structure that is given by the relation 'earlier than' or
'later than': the object has different attributes at a later point in time
P.F. Strawson: Individuals. London 1959; 'Analysis, Science, and Metaphysics,'
in: R. Rorty (ed.): The Linguistic Turn. Chicago 1967, 312-320.
14 Peter Bieri

than at an earlier. In order to describe this simple and elementary rela-

tion of an object to time, we only need a continuum of time coordinates
- tl, t2 - to whom we can allocate the description of state of the object.
We do not need to speak of another, richer time structure, of past,
present and future.
When we speak of the history of an object, we want to indicate that
it is one and the same object that has this history. In order to speak of
the numerical identity of an object throughout time, we need two
things: (1) we have to know its movement through space; (2) we have
to be able to tell a causal story about the object; to specify to which
causal influences the object has been exposed, and how it has changed
because of this exposure. This assertion of the identity of an object
goes together with a causal explanation of its change. Assertion of
identity and explanatory history are counterparts, they go hand in hand.
When I speak of causality here, I mean the relation between the causal
dependence between events, and as these, I define the relation of the
contra factual dependence. If A was the cause of B, then at least it can
be presumed that if there was no A, then B would not have happened.
This explication allows a liberal interpretation of the relation between
causality and time: causal relations are also possible between simultan-
eous events, and in principle, even 'backwards causality' is possible
(although I believe it is contra intuitive).2
We sometimes call traces what are actually the results of causal ef-
fects on an object. With that, we say that we are reminded by certain
conditions of an object to earlier influences, or reminded that they rep-
resent certain earlier events. But then, we already speak of us, and we
are not only objects. Traces on an object that is only an object do not
carry any memory for the object and do not represent anything to the
object. If we only look at the object itself and leave ourselves out of the
game - so that there is nobody who can draw conclusions - we cannot
even speak of traces, as this is a potentially epistemic term. They are
simply conditions of an object that are the causal results of earlier ef-
Consequently, objects do have a history. But they do not have a
consciousness of this history. What do we have to add to an object in
order for it to be able to generate such a consciousness? My first step is
to provide an object - e.g., an organism - with representations of its
environment. I use this term to define intentional consciousness. To
represent means to be in an intentional condition: to relate to some-
Cf. J.L. Mackie: 'Causes and Conditions,' in: E. Sosa (ed.): Causation and
Conditionals. Oxford 1975, 15-38; D. Lewis: 'Causation,' in: The Journal of
Philosophy 70 (1973), 556-567.
Time Experience and Personhood 15

thing as something. Representations are representations of something.

The ability to make representations is the ability to reference. I use the
term of representation as a very wide, formal term: I leave open if all
representation has to be conceptual representation, if all conceptual
representation is proposition^ representation and if all prepositional
representation presupposes a natural language. It is only decisive that
we are dealing with intentional conditions whose description is inten-
tional (non-extensional), because they are referentially opaque.' Famil-
iar cases of intentional conditions - representations - are: to perceive,
to think, to mean, to wish, to imagine etc.
By equipping an object with the faculty of representation, did we
automatically allocate to it a time consciousness? No. It is certain that
the intentional conditions of an object are temporarily organised in the
same way as its other conditions, namely by the relation 'earlier -
later'. Its representations form a temporal sequence. But this is only
something that is factually the case. It is not yet something for the ob-
ject. A sequence of representations is not yet a representation of a se-
quence. The order of the representation that we have up to now, is
simply the order of its causal genesis. And this order alone does not yet
allow for what time consciousness is: the representation of a temporal
In the next step, let us endow the object with the ability to allocate
time indices to events represented by it. For example, let us assume that
a computer equipped with sensors has intentional conditions and give it
the ability to attach indices to the processes which it has registered. It
then can print out, according to its 'inner clock': Atl, Bt2, Ct3, etc.
With this, did we give the computer the ability to represent events as
temporally organised events? The answer is, again, 'no'. It is true that
the consciousness of such a being can be illustrated by a linear con-
tinuum that, to us, seen from outside, is a continuum of coordinators of
time. But seen from inside, it is not yet a consciousness of a time struc-
ture. To us, this continuum is a temporal continuum, as to us, it is
defined by the temporal relation 'earlier - later'. And in this, we recog-
nise a specifically temporal logic to which, e.g., belongs the rule that if
x precedes y, then y cannot precede x, and if x y and y precedes z, then
x precedes z. To the computer, however, the continuum simply means
an abstract relational structure constituted by a transitive, asymmetric
relation. To the computer, the logic of this structure is not yet a tem-
On the intentional contexts, cf. W.V. Quine: Word and Object. Cambridge, Mass.
1960, §§ 30-32.
Here and below, I owe very much to Jay F. Rosenberg's analyses in One World and
Our Knowledge of It. Dordrecht 1980, 57-86.
16 Peter Bieri

poral logic, even if we let it use the letter 't\ Accordingly, the same is
true for a sequence of representations without indices: it is a de facto
temporal order, that is to say, the order of its causal genesis. But it is
not yet a consciousness of the world as existing in a temporal order.
What is missing in this consciousness is the representation of time
modes. To represent the world order as genuinely temporal means to
represent it from a temporal perspective that is continuously dislocat-
ing itself, or from a temporal point of view that is continuously chan-
ging. It means to represent certain events as present or happening now,
and others as past resp. in the future. And it means to have a conscious-
ness and knowledge that events which are in the future from a certain
point in time, are the present at a later point in time, and can be in the
past at an even later point in time.
But what do we have to supply to something that is equipped with
the ability to represent, in order to enable it to have this temporal per-
spective? Obviously, it is not enough to simply hand over the vocabu-
lary of time modes to our computer, without adding something else. It
then could print out: Atl now I, Bt2 now I, etc. And we could under-
stand these utterances as expressions of consciousness of the present
time. But this is not true for the computer. In order to understand that
the expression 'is now' is an expression of a time mode, one has to un-
derstand the contrast between 'is now', 'was' and 'will be'. And the
computer lacks this understanding, so that 'now' is a mere word to it.
What is it that is still missing in order to arrive at an understanding
of time modes? What, for example, is missing in order to distinguish
between present and past events? One is tempted to say that the com-
puter will be able to achieve this as soon as we provide it with traces of
earlier representations. Certain representations that belong to a being
at a given point in time are causal consequences, are after effects of
earlier representations, and these traces of the history of representa-
tions of a being - one could perhaps argue - constitute its conscious-
ness of the past. They constitute its memories. In this way, the memory
theorists of British empiricism concluded: there are active, living rep-
resentations that constitute our consciousness of the presence, and
there are faded representations that make, as traces of earlier represent-
ations, a reference to the past possible. This is the assumption that a be-
ing can read from the content of its representation - e.g., from its in-
tensity - whether it is dealing with the representation of a present or of
a past event. But this is an error: our computer, for example, will alloc-
ate the same time index to the content - the intentional object - of a
perception, and to a representation that we can understand as memory
if both representations appear at the same time. Differences in the rep-
Time Experience and Personhood 17

resented content alone do not yet constitute a temporal perspective as it

characterises our consciousness of the time modes. What we have giv-
en to the computer are only traces of earlier representations and not yet
real memories.
The problem is that a being as described above is able to situate in a
certain sense events of the world in time, but not itself. And exactly that
ability is what we need for a consciousness of the time modes. In order
to have a temporal perspective, i.e., a perspective from which an event
appears as present at a certain point of time, and also can be seen as
past from the next point in time, one has to represent oneself within
permanently changing temporal relations to the events of the world. In
other words, one has to have a form of self-awareness.
What exactly does that mean? At first, it means that I can make a
distinction between the history of the world - the sequence of its events
- and the history of my representations of the world. Both histories
proceed in one and the same time, and what I do by representing the
difference between past, present, and future, is to bring my representa-
tion of the world into a temporal relation to the events of this world.
Thus, to represent an event in the world as present (now), means to rep-
resent it simultaneously as one that I represent now, for example by
perceiving it. To represent an event as past - to remember it - means to
represent it as an event that occurred at the same time as one of my
earlier representations, i.e., as an event that was formerly experienced
by me as happening in the present. And to relate to a. future event - to
expect it - means to relate to an event that will be at the same time as
one of my later representations in time. Thus, in order to have a genu-
ine time consciousness, a being has to be able to make a difference
between the history of the world and the history of its encounter with
the world. And the continuously changing perspective I have spoken
about is nothing else but the continuous process to connect these two
sequences of events in one representation of a single, consistent time.5
This analysis also explains the apparent paradox that one time mode
can fall into the area of another. For example, we speak of past and fu-
ture presence. If one remembers something, one does not only remem-
ber something that was, but something that was present, and if one ex-
pects something one does not only relate to something that will be, but
to something that will be present. This can now be explained: a past
event is one that was simultaneous with one of my earlier representa-
tions, and as concurrence with a representation is called presence, we
are dealing with a past present. And the case of the future present is to
be explained by analogy. The specific logic of the time modes is ex-
Cf. my Zeit und Zeiterfahrung. Frankfurt/M. 1972, 79-120.
18 Peter Bieri

plained by the connection between the history of the world and the his-
tory of my representation.
Time consciousness pre-supposes self-awareness. In order to not
just organise the events in the world - like the computer - according to
an abstract relation, but to represent the events as a story, I need the
ability to represent my representation of the world themselves, my
autobiography of representations. Consequently, I need representations
of a second order, or, as it is called, metarepresentations. In the con-
sciousness of the time modes, I do not just use my representations, I
also mention them. And I quote them as temporally organised, that is as
organised among themselves as well as organised in relation to what is
represented in them - the world events. This is what Kant meant when
he said that time is the form of the outward as well as the inward mean-
Metarepresentations are representations of representations, that is to
say, they are representations whose intentional object is a representa-
tion itself. In a metarepresentation, I quote three aspects of my repres-
entations: (1) their temporal position relative to the events of the
world, and relative to other representations; (2) their intentional object;
(3) their mode: if we are dealing with a perception, an imagination or
already with a memory. In this way, I can produce a continuum of in-
terlaced memories, as described by Husserl in the case of retention.7
Now I can represent an earlier representation (remember) that is a
memory in itself, in other words, a representation of an even earlier
representation that, again, can be a memory whose intentional object is
also quoted, etc. In our consciousness, this iteration factually breaks off
at one point. But logically, there is no limit to the iteration of metarep-
resentations, as there is no limit to the iteration of metalanguages.
When we provided traces of earlier representations to our computer,
it seemed at first as if we had already endowed it with genuine memor-
ies. But then it transpired that it was only traces. We can now under-
stand why that is so: memory is the term for a representation in which
not only an earlier world event, but also an earlier representation of
this event is being represented. And the computer - the way I described
it - did not have this ability to quote itself.
At this point, we have reached an intermediary result: as physical
beings we are objects whose elementary relation to the world is charac-
terised by the fact that we change with time and consequently, have a
history. As persons, in addition, we are beings that have a conscious-
I. Kant: Critique of Pure Reason, B 49ff.
E. Husserl: On the Phenomenology of the Consciousness of Internal Time.
Dordrecht 1991, §§ 10-12.
Time Experience and Personhood 19

ness in the sense that we can represent our environment. This con-
sciousness - according to its causal genesis - is in fact organised tem-
porally. Furthermore, as persons, we have a consciousness of this order
as a temporal order, as we can represent time modes and thus gain a
temporal perspective. And we have this time consciousness because we
have self-awareness: because we have metarepresentations and can
quote our own history of representations by connecting it to the history
of the world.
This result has an important epistemological consequence. If in
terms of our time consciousness we have to distinguish between the
history of the world and the history of our representations, then we
have to distinguish between the happening of an event, or the being of
an object in this world, on one side, and their being represented on the
other side. And this means no less than that in order to have a con-
sciousness of time, we have to be epistemological realists. By defini-
tion, this realism includes the conviction that the being of an entity in
the world is something different from its being represented by us: that
its esse is not percipi, or, more general, repraesentari. And whoever is
convinced of this, also believes in the following two assumptions
which are fundamental to our common sense conception of the world:
(1) there is a diachronic continuous existence of entities in this world
even when they are not represented. (Note that this assumption does
not hold for certain phenomena of the inner world, like, for example
pain: pain that is not felt does not exist). (2) There can be a difference
between the way we represent an entity, and the way the entity is. In
other words, a realistic consciousness makes a distinction between ap-
pearance and reality. (A distinction that makes no sense with pain:
pain is the way it feels). In order to represent time, we have to present
the world as a realistic world. In other words: an idealistic conscious-
ness (one that conceptualises all entities according to the logic of pain)
cannot be a real consciousness of time. And as our consciousness is a
true time consciousness, idealism is wrong. Persons are necessarily
realists. To them, it is no option to be an idealist. To understand one-
self as a person, and at the same time to claim that one is an idealist is
an incoherent position.
As realists, it is our epistemic goal to distinguish between appear-
ance and reality. Firstly, this is true for the presence. Persons try to find
out how objects they are confronted with really are. They attempt to
make a difference between the perhaps deceptive aspects of their rep-
resentation and the real attributes of the objects. They also try to do the
same in their encounter with the past: we want to distinguish between
the way the past really was and the way that we remember it - how it
20 Peter Bieri

seems to have been. And here, one can discover a second connection
between having a time consciousness and the way how we have to rep-
resent the world8: the attempt to distinguish between appearance and
reality in the past can only be successful if we organise the world
events according to causal dependencies, which I understand here, as I
said before, simply as contra factual dependencies. In order to recon-
struct the real course of the past and to disclose errors of memory, I
have to be able to make use of causal inferences that tell me how the
event sequences of the past have to have been. For let us assume I do
not have causal inferences at my disposal. Then I have the belief foaL
there could be a difference between the past events and my representa-
tions of them. But I cannot work with this difference. It is true that I
have representations of present and past. But I can use none of these
representations as an evidence for another. I have the impression that
the past was so and so. But I cannot verify this impression. I want to
make a distinction between appearance and reality of the past, but I
cannot do so. I have a concept for which there is no set of rules to ap-
ply: I have an incoherent consciousness.
However, if I have at my disposal the idea of causal dependencies
between events, and thus of causal inferences, I can construct an ob-
jective time order within the past. For example, if I know the causal
connection between fire and ashes, then I can know that in a certain
past context the ashes cannot have existed before the fire, even if it
may appear to me like that in my memory. And now I can establish an
objective time order not only within the events in the world but also
within my past representations. When I assume that my representations
- perceptions - of fire and ashes depend causally on the existence of
fire and ashes, then I can know that my representation of ashes cannot
have anteceded the representation of fire, even if I might present it that
way in my metarepresentations. In other words, if, and only if, I think
of the realistic world as causally structured, and if I regard myself as
causally implicated in this world, then I can reach my epistemic goal
and distinguish between the real and the apparent image of the past.
Then, and only then, I can distinguish between experienced and object-
ive time. And, of course, this is nothing else than the argument of
Kant's'Second Analogy'.9
At this point I can finally return to the topic of 'self-awareness'. I
argued above that a genuine consciousness of time is only possible be-
For the following, cf. Jonathan Bennett: 'Analytical Transcendental Arguments,' in:
P. Bieri, R.-P. Horstmann, L. Krttger (eds.): Transcendental Arguments and
Science. Dordrecht 1979, 45-63.
I. Kant: Critique of Pure Reason, B 233ff.
Time Experience and Personhood 21

cause we can situate ourselves in time, and I tentatively explained this

idea by introducing the concept of metarepresentations. Yet self-
awareness is not completely described by this. It is obvious that the
representations that I quote are factually my own. However, this does
not mean that I already have a description of myself as a being that has
a history - a description of myself as a subject. I can only describe my-
self as something that is a history, that is to say, as a sequence of rep-
resentations that are in a temporal relation to the events of the world.
But when I quote representations, I do not only want to quote their ex-
istence or their appearance; I want to represent the fact that it was me
who represented earlier. I want to represent the representations as my
own. Examples for this form of a metarepresentation are 7 represent,
that / represented', or 7 remember, that / perceived'. In both cases the
assumption is that T relates to the same subject, and therefore, we
speak of 'consciousness of self identity in time'.
A question often raised in this regard throughout the history of the
debate concerns the referential object of T : to what does this expres-
sion relate? From the outside - from the perspective of others - it is
clear that we are dealing with a physical person that remains the same
through time. But from the inside - from my own perspective - it may
appear that I need not necessarily have to understand the consciousness
of my identity as the consciousness of the identity of my physical per-
son - as if I could abstract from my physical aspects or parenthesise
them and see myself as merely a mental subject that remains the same
But that is not possible. For we have seen that we want to distin-
guish between our representations of the past and the real past and we
can only do that if we embed the history of our representations causally
into the history of the world. Yet this we can only do, if, so to speak,
we give the world a chance to come into causal contact with our repres-
entations - to cause them. And therefore, we have to represent
ourselves as physical persons on which the world can have an impact.
In other words, we can only have the concept of an objective past if we
are able to represent our past as a past of physical persons.
With this, I will close my piece of descriptive metaphysics of time
consciousness. In these somewhat transcendental reflections I have
tried to show that there are four closely connected conditions which
must be met if a being factually living in time shall also have a con-
sciousness of the temporal dimension of its existence: (1) in order to at-
tain a temporal perspective on the world, it has to have metarepresenta-
tions at its disposal; (2) in order to achieve this, it is necessary that it
represents the world as a realistic world; (3) in order to be able to dis-
22 Peter Bieri

tinguish the objective time of this world from the experienced time, it
has to represent the world as causally organised; (4) and in order to be
able to do this, it has to understand its identity as an objective identity.
These four conditions together constitute the very formal structure of
the temporality of persons.

Identity of the Person in Time

However, there are other aspects of this temporality. Persons do not
simply have a consciousness that they have a past, a present, and a fu-
ture. They can also develop certain attitudes towards these temporal di-
mensions of their life. And it belongs to our concept of personality that
they are supposed to develop certain attitudes. The concept of a person
does not only have descriptive, but also normative aspects. With the
idea of personhood, we connect certain normative expectations or
ideals, and some of these ideals concern the relation of a person to its
past, present, or future.
Let us begin with the past of a person. Persons do not only have a
past; it is, as we have seen, essential to persons that they can represent
this past. And we have further seen that they necessarily represent it,
when they represent it as past, as their own past. Next to this com-
pletely formal aspect of belonging, there is another aspect: we expect
from persons that they also take possession of their past. It is not
enough, this ideal says, that I simply state vis-a-vis my past that certain
things happened to me. As a person, I should try to take them into pos-
session and to understand them as parts of myself, so that they do not
appear as alien to me. In other words, this ideal means that I should try
to create a unity in my past life. This past life, though, does already
have a unity that has materialised without me: the events in my story
stand in relations of causal dependence to one another and therefore
form a causal unity. The unity that is referred to in the concept of ap-
propriation, however, is a unity that I have to produce in the first place.
It derives from my need to understand the past. To the degree that I
succeed in understanding my past, I unify it. I want to name this unity
hermeneutic unity.
In order to be able to understand my past I have to be prepared to re-
member it in a way that I can live through it again. In other words, it is
a necessary condition for appropriation that I do not suppress my past
in the sense that it is not accessible any more to my deliberate and dir-
ected memory. But this alone is not enough. Memories without inter-
pretation are in a certain sense blind. They do not yet constitute com-
prehension. To create understanding means to see a remembered exper-
Time Experience and Personhood 23

ience in the context of earlier and later experiences in my history. And

the relations that I create here are explanatory relations. Understanding
is explaining. The creation of hermeneutic unity is an explanatory pro-
cess that means an epistemic achievement.
Appropriation of one's own past means to narrate an explanatory
story about this past. What are the general characteristics of such a
story?10 At first, we have here the postulation that memory on which a
story is based is mostly correct, that is, true. If we had to assume that
most memories of somebody are wrong, we would not only not speak
of appropriation of the past, but would also hesitate to attribute
memory to them at all. But correctness of memory is not enough. I can
appropriate my memory only to the degree that I succeed in making
sense of it. Doing this, I have three aspects of my story at hand: (1) the
story of my intentional attitude: my history of actions; (2) the story of
my needs, wishes, and emotions: my emotional history; (3) the story of
my opinions, of my world of thoughts: my ideological history. Thus, to
create a hermeneutic unity of my past means to merge these three stor-
ies into one single explanatory story. The formal structure of this story
is: I have done certain things because I believed, wished, and felt cer-
tain things. My story of appropriation consists of intentional explana-
tions of my past actions, in which intentional conditions like opinions,
wishes, and emotions are quoted as antecedents.
Intentional explanations construct the person they are dealing with
as a rational being - as a being that acts in accordance with what it
wishes, feels, and believes. Rationality is a parameter in the attribution
of intentional conditions, and, in this sense, it is a condition for person-
hood. Accordingly, it is a construction principle of the story of my past
that I can represent my story as the story of a rational character.
Whatever I concede to myself, ex post, are imperfections like neglect,
ignorance, lack of information, etc., but not blank irrationality. For if I
saw myself as largely irrational, I would not be able to understand the
actor of my past as a person.
Thus, we rationalise our past. This is an undertaking that contains
the danger of retrospective distortion, for, of course, we were not al-
ways rational. Rationality is also an ideal of personhood. But explana-
tions of our past that are oriented at this ideal are our only chance to
produce a hermeneutic unity. And perhaps some things that may appear
irrational and bizarre in our past are so only at the surface. Psychoana-
lysis, for example, impresses us, among other things, because it allows
In the following, I apply Daniel C. Dennett's theory of intentional systems to
memory. Cf D.C. Dennett: 'Intentional Systems and Conditions of Personhood',
in: D.C.D.: Brainstorms. Hassock 1978, 3-22, 267-285.
24 Peter Bieri

us, with the explanatory assumption of subconscious wishes, feelings,

and opinions, to see ourselves as rational persons to a much higher de-
gree than when we only stick to what is conscious to us (in the sense
that we can report on its existence at any time). Consequently, the ex-
planatory assumption of a subconscious can enlarge the chance to ap-
propriate our past.
When we try to produce, in the described way, a hermeneutic unity
in our past life, then we create, as we say, an identity. I make who I am
understandable to myself by describing who I have become, and by ex-
plaining to myself, why I became who I am. When I tell such a story, I
operate selectively: firstly, in what I remember at all, and what I forget
or suppress; secondly, in how I value and assess what I remember, i.e.,
what I see as being essential to me and not just accidental. In a story
about how my presence derived from my past, I cut the past to size: I
give a form to it or render it a profile. I distinguish certain events that
happened to me, and certain experiences, as a key to other events, as
turning points, etc. Retrospectively, I act as the dramaturge of my life.
And by doing this, I mostly succumb to the lie of all classical dramat-
urgy: by overlooking or just not accepting that the course of my life
was simply also determined by a multitude of coincidences, I give an
unavoidability to my life that it did not have in reality.
The criterion for a story of appropriation is explanatory coherence:
the ability of a story to make all available data on my past understand-
able is decisive. But our experience is that there always seems to be
several, similarly coherent stories. Therefore, we sometimes have the
impression that we could choose among various stories - various pos-
sible identities. And sometimes we change our story. If we do that, we
'change our identity'; we 'lose' an old identity and 'win' a new one.
This can, for example, happen during a long period of psychoanalysis.
We then have the impression that we discover a new identity for
ourselves and can re-write our past. That is the case when, for ex-
ample, censorship is loosened and as a result new memories become
available. Sometimes we may also have the impression as if we could
invent or create a new identity for ourselves. In this case we do not
mean a fictitious identity, but we speak of a new explanatory potential
that is revealed by data that were known to us earlier but now appear in
a new light. What we invent are new explanatory hypotheses about us
that enlarge the coherence of our history of appropriation.
When we successfully design a new hermeneutic unity of our past,
which we believe to be more integrative than the former one, our iden-
tity will change. This is a continuous process, and this kind of change
is essential to persons. In this process, the effort to come ever closer to
Time Experience and Personhood 25

an ideal of the personhood that appropriates the past as completely as

possible is reflected. And it is important to see that the sequence of
changing stories of appropriation is all we have as an answer to the
question 'Who am I?'. There is no stable core in a person, no position
outside these stories, from where I could ask: 'And who am I really in-
dependently from these stories?'
Finally, I will mention a moral aspect of this idea that I am sup-
posed to appropriate my past. I am held responsible for my past ac-
tions. To me, this only makes sense if I can accept my actions as my
own. Insofar as I can remember my actions through metarepresentation,
I represent them already as mine. Nevertheless, the readiness to take re-
sponsibility intuitively includes more: it only really makes sense to me
if I can accept the responsibility for my actions under a description of
myself to which these, my actions, apply. Therefore, a normative ex-
pectation to persons can be recognised in the effective rule that I have
to take responsibility in every case - pathological cases aside: we ex-
pect from persons that they search for an identity that makes an integra-
tion of their past actions possible and allows them to stand by these ac-
Let us regard now the relation that a person can have to their future.
As we have seen, to have a consciousness about the future means to be
able to relate to a future presence. If a being only has this purely formal
ability, then it can happen that it simply lets its future happen to itself.
If existence is this way, one simply stumbles passively into the future.
But we as persons can do otherwise. To me, as a person, it is specific
that I can be concerned and care for my future. By doing this, I am con-
scious firstly about what I do in fact do and what I can achieve, but
secondly also about who I will be. These two aspects of my concern for
my future can be called my project.
Since a consciousness of the past without an explanatory story is
blind, a consciousness of the future without projects is blind as well.
To have a project means to conceptualise oneself as a certain person,
as a person with a certain identity, into the future. By doing this, I take
care that not only my future presence makes something out of me but
that I also make something out of it. Within such a concept, I care
about my future history of action from the perspective of my present
wishes, emotions, and opinions. However, I myself cm also care about
these wishes, emotions, and opinions: I can decide which of the wishes,
expectations, hopes, etc. that factually exist in me I indeed want to ap-
propriate and regard as mine in an emphatic sense.11 Doing that, I devel-
Cf. H. Frankfurt: 'Freedom of the Will and the Concept of a Person,' in: The
Journal of Philosophy 68 (1971), 5-20.
26 Peter Bieri

op wishes or preferences on a second level. I wish to be somebody,

who, in the future, will be guided by these-and-these wishes and prefer-
ences. In this way, as a person, I can take an attitude of reflective self-
evaluation, and the result of my self-evaluation will be my project for
the future.
When I go into my future under the influence of such a project, I
lead my life in the emphatic sense of this word. Part of it is that there is
continuity between my projects and my appropriation of my past. We
expect from a person that his projects match the explanatory story that
constitutes his present identity. I have to plan what I want to be in the
future in the light of what I have become in the past. For my projects
are supposed to afford a realization of myself, the realization of a cer-
tain person who, because of a certain past, is the one that he is. When
my projects do not match at all to what I have become, then I am in
danger of breaking up as a person: I am giving away the chance of real-
izing an identity that embraces my past as well as my future. To devel-
op a project means, in other words, to strive for the unity of the person
that I am.
This observation also has a moral component. The condition of the
continuity of my projects with my past is supposed to secure that I do
not accept obligations that I cannot possibly fulfill. This is an important
point towards others as well as towards my own self. Others have to be
able to rely on me, and I am not allowed to commit myself (e.g.,
through a promise) to something that I cannot honor. In that case, I
would appear to myself as an incoherent person and in the end even
lose the status of a person in my own perspective.
What, finally, can be said about the relation of a person to its pres-
ence? To not experience the present blindly means to experience it in
the context of an appropriated past and a project for the future. Both di-
mensions are necessary to make sense of my presence in reality. At the
same time the consequence is that I can only turn certain sectors of the
presence that I can potentially experience into my own presence. The
first of these constraints derives from the story about my past and the
self-image constructed by this story. I have become somebody who can
only feel, wish, and do that-and-that in a present reality which, object-
ively seen, may harbour to others completely different opportunities to
experience and act.
This kind of constraint can result in the fact that I might miss a
present reality that is a completely possible presence. I miss it because
I, as one says, live 'too much in the past' in that I concede to the past
an undue power over my present. Three aspects of this phenomenon
can be distinguished: (1) I cling so much to one of my appropriation
Time Experience and Personhood 27

stories that I, as it were, freeze the identity that it suggests. In that case,
my presence - the way and how I feel and think now - reproduces in a
certain sense only past presences - the way and how I felt and thought
in the past. In other words: I do not give the present reality the chance
to change me and to correct me in my self-image. (2) The conscious-
ness of the presence is covered, flooded by memories into which I sink
and where I remain. This can, for example, happen with the death of a
person who was very closely intertwined with our identity. We then
substitute the past for the present and have the perception 'that life
stands still'. (3) We suppress our past and separate it from us, so that it
is no longer accessible to us in our memory. Under this condition it re-
mains causally effective, and has its considerable causal effects espe-
cially because it is suppressed and defies appropriation and accom-
plishment. In such a case, one can say that we do not live enough in the
past. But in the end, as Freud has taught us to see, it ends up being the
same: our present is so much under the dictation of the past that real
consciousness of the present is no longer possible.
In the same way as I can miss my present because I live - in one of
these ways - too much in the past, I can also miss it by living too much
in the future and by having the perception that life can always only be-
gin the day after tomorrow. In this case, I substitute the present with
the future, so that the present never really counts. I am only interested
in my future presence, how it appears in my projects and how I imagine
it in my anticipatory imagination. My identity, projected into the fu-
ture, quasi suffocates my present. And that carries a subtle danger: the
future person that I wish to be is one that has grown into the future
identity by allowing itself to be changed by the proceeding moments of
presences. If I, however, as it were misappropriate or bracket my life in
the meantime because of a fixation on the future present moments, then
I will be the old (same) person when the future point in time arrives,
i.e., has become present. And I will not be able to slip into the identity
which was imagined for me in my project.
One misses one's presence if one only lives the present moments as
a past or future and never as a present presence. What does it mean to
avoid this danger? I only know very vague descriptions of this ideal
that we have here. One of them is that one should be open to the
present reality and get involved with it. Then there is talk about the
ability or the preparedness to let oneself be - temporarily - over-
whelmed by the present and get lost in it. Another thought that we find
in the philosophical tradition, mainly with Kierkegaard, is the idea that
everything depends on living in the present moment: in a way, as if life
would come to an end in the next moment. I, myself, believe that this is
28 Peter Bieri

what it is all about in the end: in the same way, as consciousness of the
past without explanatory story is blind, consciousness of the present is
blind without implicit connection to an appropriation story and without
a projection of my identity into the future. At the same time, these two
dimensions that I need in order to embed also the present into a her-
meneutic unity of my life harbor the danger that I miss my present
presence. Therefore, we have to find a balance towards our presence
between blindness and lost chances. This balance is an aspect of what
we call equanimity: the security that I can, whatever happens to me,
place these happenings into a hermeneutic unity of my life, and at the
same time I can have the strength to not allow my life to be suffocated
either by memory or by my worries about my future. We lose this bal-
ance again and again and have to search for it repeatedly. And if we re-
gard this balance as an ideal of personhood this means that we lose the
state of being a person again and again and have to regain it again and
again. This is a basic fact of our conscious life.

Constituting Time through Action and Discourse

0. Introduction
When it comes to the classical topic of time, it seems philosophers'
clocks move differently. Insights brought forward in the last one hun-
dred years in the context of the linguistic and pragmatic turn of philo-
sophy are noticeably ignored: the demarcations of problems have be-
come increasingly blurred, the answers less clear. But above all, it ap-
pears that which or whose problems should be solved by a philosophy
of time is out of sight. However humble these solutions might be, and
with whatever relative reservations they are suggested - the solutions
brought forward have to at least claim transsubjective plausibility.
In the following, some exemplary remarks intend to point out the
sense in which contemporary philosophy seems to have lost its purpose,
and lead to a second step in which an assumption about the root causes
of this development is made: not only the linguistic distinctions of time
itself but also the epistemological distinctions of how to describe, ex-
emplify or explain time lack sufficient explication and reflection on the
conditions of plausibility, and thus these very reflections and answers
have lost their place in life. In the next chapter, a methodological recon-
struction of our temporal distinctions in the framework of social action
and discourse will be given. By doing this, a life-experience related
(lebensweltliche) base for scientific and philosophical theories of time
will be brought forward as well. This will finally lead to a conclusion
concerning realistic and ontological concepts of time. The point of ref-
erence remains the question why we as human beings need 'time' in our
actions and in our discourse - and accordingly seek to provide those
linguistic creations which are also object to philosophical debate.
30 Peter Janich

1. Philosophy of Time without Orientation -

a Short Criticism of Language
Far from any claim to give an overview of contemporary philosophy of
time or even to just list the most important articles, some well known
concepts, and such concepts which have been distinguished by frequent
quoting, will be singled out as examples. This allows a remarkable dia-
gnosis. When reading programmatic texts of the 'linguistic turn',1 the
problem of time seems to be especially appropriate to discuss philo-
sophical problems as problems of their linguistic phrasing. From the
Pre-Socratics and Aristotle's concept of time to the then contemporary
days of the Phenomenologists, philosophical tradition had yielded an
abundance of texts, giving linguistic philosophy, with its purposes of
language analysis and criticism of metaphysics, a lot to reflect on. Fur-
thermore, it delivered an important example of a relativistic revision of
classical philosophy and of the distinct theory of relativism that philo-
sophers of the linguistic turn adhered to. Here, the substitution of epi-
stemological theory and nature philosophy through scientific theory
could be exemplified. Especially attractive was the inclusion of the time
modes of past, presence and future into the language of physics within
the world of Minkowski, and the apparent introduction of scientific,
measured time into the individual time consciousness of an observer.
This was to be seen in the framework of a relativistic theory of physics,
which now had to be reconstructed syntactically and semantically by
the philosophers of logical empiricism.
Assertions of existence (Existenzaussagen) already played a promin-
ent role in explanations within concepts of the linguistic turn - regard-
less if 'exist' concerned classical teleological arguments2 or basic prob-
lems of arithmetic,3 and if the word 'to exist' was denied the character
of the predicator,4 or if the problem of existence was approached by a
See R. Rorty (ed): The Linguistic Turn. Recent Essays in Philosophical Method.
Chicago, London 1967, see also the texts quoted in note 2 to 7.
See e.g. W. Kneale: 'Is Existence a Predicate?', in: H. Feigl, W. Sellars (eds.):
Readings in Philosophical Analysis. New York 1949, 29^13.
See G. Frege: Die Grundlagen der Arithmetik. Eine logisch mathematische
Untersuchung iiber den Begriff der Zahl. Breslau 1884, Introduction (trans.: The
Foundations ofArithmetic. A logical-mathematical investigation into the concept of
number 1884. New York 2007) and G. Frege: Begriffsschrift, eine der arith-
metischen nachgebildete Formelsprache des reinen Denkens. Jena, Preface of Dec.
18, 1878 (trans.: 'Concept Script. A formal language of pure thought modelled
upon that of arithmetics', in: J. van Heijenort (ed.): From Frege to Godel A source
book in mathematical logic, 1879-1931. Cambridge, MA 1967.
Like in the essay mentioned in note 2.
Constituting Time through Action and Discourse 31

theory of the markers of the subject position in sentences.5 It was within

the tendency of philosophical thought of that time when, in 1908, Mc-
Taggart attempted to prove the unreality of time through an analysis of
terminology of modal and ordinal time lines and their relationship.6
The literature discussing his article is almost limitless. Here is not
the place to comment on this literature or the assumption of McTaggart
himself. As far as the deficits of contemporary philosophy of time are
concerned, the following diagnosis seems to be sufficient: there is not
the least attempt at terminological explanation, let alone at defining
temporal words or epistemological words utilised as central points of
reference. The comprehensibility of, e.g., 'past', 'present', and 'future',
of 'earlier than' and 'later than' are taken for granted in the same way
as that of the noun 'time' itself. But also expressions like 'the true
nature of time', 'unreal', 'realities which can be found in a time line',
'object of our belief respectively our imagination', and 'belief in the ex-
istence of a plurality of timelines' remain as unexplained as expressions
like 'direct perception' or 'purely subjective'. It can be regarded as un-
disputed that the first, the temporal words, are known to us from every-
day language, and that the latter, the epistemological ones, only appear
in the form of completely undefined idioms which have come down
from philosophy and science - rendering such words unusable for the
purpose of a philosophical clarification without further explanations. It
may be asked which standard and which hope for new insights is at the
basis of such philosophical thought, if, on one side, a comparison of
linguistic systems of discrimination is introduced (McTaggart's A- and
B-line as modal and ordinal time lines), and on the other side, no men-
tion of what constitutes the meaning and the use of these discrimina-
tions is made.
Some assumptions of this much-quoted article are even presented
below a generally accepted level of scientific analysis, i.e., without a
distanced presentation of philosophical statements, in the form of per-
sonal confessions ('I believe that time is unreal'). This is a level on
which non-realists and realists manifest their personal preferences. The
question of the purpose of the research in the sense of a philosophical
or scientific problem that can be transsubjectively communicated, and
See B. Russell: 'On Denoting', in: Mind 1905, from R.C. Marsh (ed): Logic and
Knowledge, Essays 1901-1950. London 1956, 41-56.
J.E. McTaggart: 'The Unreality of Time', in: Mind 17 (1908), 457^174. [Janich's
erroneous reference to John McTaggart Ellis McTaggart in the plural has been
corrected in our translation. Despite its unusual form, the name refers to one
individual only. In the following, all references to McTaggart have therefore been
transformed into the singular.]
32 Peter Janich

to which the claim of a transsubjectively comprehensible answer can be

attached, neither exists explicitly, nor can such a purpose be recognised
This situation is different in the article by Bertrand Russell, 'On the
Experience of Time', of 1915,7 which is the second of the texts dis-
cussed here. Also within the framework of the programmatic orienta-
tions of the linguistic turn, albeit with a noticeable relationship to Ernst
Mach's Analysis of Perceptions, Russell pursues the question of the re-
lation between the non-scientific experience of time and physical time.
Sensation und memory and their pre-scientific roles in the constitution
of time in experience are reconstructed in their concurrence. 'Defini-
tions' are explicitly sought, e.g., in order to make sense of the meaning
of discourse of 'one (momentary) total experience'. Also, the temporal
relaters 'earlier' and 'later' resp. the adjectives for the three time modes
are tentatively defined, although as 'purely verbal definitions' ('We say
that A is earlier than B if A is succeeded by B') against the background
of a realistic epistemological theory: 'There is no logical reason why
the relations of earlier and later should not subsist in a world wholly
devoid of consciousness.'
There is a clearly recognizable concern in Russell's approach: to
construct the physical out of the mental time, and it shows an awareness
of the linguistic character of the theory of time. However, a certain idea
concerning definitions becomes virulent in regard to the discriminations
of time concepts without being made explicit: the explanation of in-
stances of language usage according to their empirically controllable
adequacy in concrete circumstances. Ultimately, it is 'perceptions'
which provide the material that has to be organised into a theory of
time in the sense of the early logical positivism - that is, limited to the
means of logics and empiricism. In the absence of any terminological
explanation, one can assume that in this case the mention of 'percep-
tions' points to Mach's heritage.
About thirty years before Russell's essay, in 1886, William James
had published his essay 'The Perception of Time',8 in which he ex-
plores the perception of time in accordance to experimental psychology
since Wilhelm Wundt and interprets it along hypotheses concerning the
temporal function of the brain in relevant processes of consciousness.
James contests Kant's transcendental philosophy pointing out that 'the
assumption' of time 'must be due to a permanently present cause. This

B.Russell: 'On the Experience of Time', in: The Monist 25 (1915), 212-223.
W. James: 'The Perception of Time', in: Journal of Speculative Philosophy 20
(1886), 374^107.
Constituting Time through Action and Discourse 33

cause - probably the simultaneous presence of brain-processes of dif-

ferent phases - fluctuates.'
As far as our considerations are concerned, we have to note again
that all distinctions of time are given by James in an unexplained lan-
guage that is assumed to be sufficiently evident, including the linguistic
means used to present results of experimental psychology as well as of
their epistemological interpretation. All distinctions are used naively.
Similar to the two authors mentioned above, James looks for the prob-
lem of time in the area of reception, perception, the sensory, that is, in
the realm of passivity. If one leaves out a completely different tradition
of debate on the problem of time, namely the tradition of the older (E.
Husserl) and newer (H. Schmitz) phenomenology and the special phe-
nomenological concept of Heidegger's existential propositions on time
- a tradition that would necessitate a totally different essay -, and if one
concentrates on the mainstream of analytical-empirical concepts includ-
ing their modern variations, then one should not be amazed that the de-
ficits in terms of linguistic explanations, and of explanations of purpose
and means provide a humus on which, a hundred years later, metaphys-
ics and ontology are again thriving.
Instead of partaking in the lessons of the linguistic, and later of the
pragmatic turn of philosophy, and of striving to overcome its remaining
weaknesses, modern studies strengthen the old errors and - unfortu-
nately - lose the old merits. Just to mention, more recent examples of
this development are - in chronological order - the books by P. Bieri
(Zeit und Zeiterfahrung, 1972), W. Deppert (Zeit. Die Begrundung des
Zeitbegriffs, seine notwendige Spaltung und der ganzheitliche Charak-
ter seiner Teile, 1989) and K. Mainzer (Zeit. Von der Urzeit zur Com-
puterzeit, 1995).
In its subtitle, Bieri's book9 promises the 'exposition of a problem
area'. It discusses 'the traditional problem of reality, objectivity and
subjectivity of time in the context of contemporary theory of time' - as
the blurb aptly asserts. Bieri's very detailed and, in detail, subtle ana-
lysis of the unreality argument of McTaggart comes to the conclusion
that the consciousness of time should, with 'the greatest neutrality pos-
sible', be understood as the 'fact', 'that our conscious data present
themselves as chronologically aligned'.l0 The explanation of this dia-
gnosis amounts to the assumption that the 'phenomenon' of time per-
ception resp. of our time consciousness can only be understood if we

P. Bieri: Zeit und Zeiterfahrung. Exposition eines Problembereichs. Frankfurt/M.
P. Bieri, I.e., 182 [our translation].
34 Peter Janich

take it to refer to a real time that exists outside of consciousness -

whatever that is supposed to mean.
In the course of his reasoning, Bieri adopts from his philosophical
sources the deficits diagnosed above completely and without any modi-
fication: chronological distinctions are neither explained nor defined
but assumed as self-evident. Whenever there is a reference to Physics, it
is considered neither in terms of its empirical validity nor in terms of its
semantic viability. The purpose of the study is not mentioned - if we
presume that it is not the rehabilitation of the old metaphysics of time -,
and is nothing more than a contribution to an actual debate in academic
philosophy. It does not seem to be possible to establish whose problems
are supposed to be solved with which means - no daring argument, as
such a task is probably not the concern of the author.
W. Deppert reaches a similar result, though from a different starting
point and in a completely different way. Historically, he goes much fur-
ther back and confronts aprioristic and empirical theories of time -
with the reconciliatory result that both have something to contribute to
the understanding of time. He asserts that time does have an 'epistemo-
logical-logicaV and an 'ontological aspect'," and thus can only be un-
derstood holistically if both aspects are considered. The reason for the
study is, as with almost all studies on the philosophy of time, the
(seemingly) paradoxical nature of time12 - a topos dealt in since St. Au-
gustine - as well as the competition of modern natural sciences with the
philosophical tradition.
But such a motif does not yet render a philosophical approach pur-
poseful, and therefore does not provide a parameter enabling one to de-
cide if the suggestions made are successful or not. If one does not in-
tend to solve a problem, one cannot be criticised for not having solved
one. The 'explicit purpose' of Deppert's work,13 to provide 'stimulation
for a mutually beneficial cooperation between the various individual
scientific disciplines and philosophy' - could probably also be reached
by intelligently provoking but false theories. Therefore, it may be al-
lowed to assume that the true - instead of the pretended - purpose of
the author lies more in the rehabilitation of an emphatic discourse of the
'being', more precisely, of the 'being of the recognizable' and the 'be-
ing of the recognising',14 or in short, it lies in the revival of ontological
W. Deppert: Zeit. Die Begriindung des Zeitbegriffs, seine notwendige Spaltung und
der ganzheitliche Charakter seiner Telle. Stuttgart 1989, e.g. 247f. [our translation].
W. Deppert, I.e., 12f., 250 [our translation].
W. Deppert, I.e., 233 [our translation].
W. Deppert, I.e., 240 [our translation].
Constituting Time through Action and Discourse 35

As far as chronological distinctions and epistemological instruments

are concerned, largely the same can be said of Deppert's as of Bieri's
book: everything is somehow given, and is assumed as sufficiently
known and reliable. Thus, Deppert's epistemological attention to the
topic of time complies with what R.J. Hankinson ironically states about
epistemology under the title 'The Basic Questions of Philosophy-. 'The
only thing that you really have to know about epistemology is that it
does not contribute anything to knowledge.'15 Ultimately, Deppert's
study is an attempt of reconciliation and appeasement in the dispute or
the competition between philosophy and science - except that this dis-
pute has ended long ago due to a largely legitimate lack of interest on
behalf of the natural sciences, and due to conflict aversion and lack of
ideas on philosophy's side.16
The most recent of the books on time mentioned here, K. Mainzer's
text which was published in a paperback series bearing the title 'Know-
ledge',17 obviously assigns itself the task to give an overview on philo-
sophical and scientific theories of time. This book gains a 'special pro-
file' by the 'new perspective' that science and arts are 'mutually de-
pendent on each other'.18 It does so by way of a sweeping survey from
the Ancient World through classical and relativistic physics, to
quantum physics and thermo dynamics to the scientific analysis of the
phenomena of life - from Darwin's evolutionary biology to contempor-
ary neurobiology including its models of mental processes - and finally
in a short alibi-like appendix on historical philosophy and cultural stud-
ies, in which culture is strained - as would be said in the language of
cooks - through the finely-woven sieve of scientific terminology. The
culture of human achievements is not mentioned nor is its capacity not
only to describe but also to actively organise life and culture itself (in-
cluding science) by way of a tried-and-tested system of time distinc-
Here, too, it is a continuing effect of the analytical-empirical tradi-
tion that the issue of language in the philosophy of time was not taken
seriously - against programmatic commitments. The purposes of an un-
dertaking such as the philosophical clarification of problems of time
have to be named explicitly beforehand in order to be able to later judge
the means offered by the philosopher. Mainzer's book, too, suffers
R.J. Hankinson: Bluff your Way in Philosophy. Horsham 1985.
Here: towards the specialised sciences. There is no claim that conflict avoidance
exists towards competing philosophical theories. See note 23.
K. Mainzer: Zeit. Von der Urzeit zur Computerzeit. Miinchen 1995 (series
K. Mainzer, I.e., 7 [our translation].
36 Peter Janich

from the fact that no nameable, clearly defined purpose is being fol-
lowed. Hovering above everything is what might have beruled for dec-
ades as Logical Empirism's slogan: 'The philosophical tradition has
failed. The modern mathematical sciences are right, they are beautiful
and good. Let us develop a new philosophy out of their analysis.'
In no way should the merits of analytical-empirical philosophy be
belittled. (By the way, traditional Hegelians, for example, would have
no problem including the suggestions of the following chapters among
their 'analytical philosophy'). Also, on the other hand, the defects of
the philosophical tradition - which are multiplied by the defects of a
historical-philological hermeneutics without orientation - should not be
denied. But the way the 'problem of time' is treated in Mainzer's lu-
cidly written overview gains the character of a new paradox, that of an
un-philosophical philosophy. Not only does the purpose and means of a
philosophy of time remain unmentioned but epistemological and sci-
entific theoretical rationales, criticisms, refutations, reconstructions and
judgments are avoided. With this, philosophy as purest doxography
from a philosophical distance appears as the latest form of philosophy
of time. Against this background it becomes an urgent task to attempt
any new systematic suggestion in philosophy of time only after having
revived the awareness for its problems, an awareness that demands
more than yet another gush into the waters of existing theories of time.

2. The Differentiations of Time within the Lebenswelt

When time becomes explicit as an issue, every competent speaker of an
ordinary language already commands a multitude of distinctions of
time. Whether one deals with the experience of time or the making of a
calendar and the measuring of time, with philosophical, anthropologic-
al, scientific or cultural problems - everyone, any time, speaks about
time as a matter of course and especially makes use of the noun 'time'
('Zeif) - which is, by the way, the case in all civilised languages where
philosophical and scientific statements on time exist.
The availability of the noun 'time' already hints to tacit and unre-
flected linguistic preconditions of every philosophy of time which
makes use of this noun and discusses problems such as whether time
exists, whether it is infinite or perhaps has its origin in the big bang,
whether time is paradox or inconceivable, or whether in the end it can
be reduced to our sense of time and our perception of time that is
caused by procedures and rhythms in our organism. Prima facie, these
and countless other questions owe their logical-grammatical expression
to the fact that the noun 'time' is at hand.
Constituting Time through Action and Discourse 37

Substantives, in elementary German school grammar also called

Dingworte (words of things), suggest in a misleading way that there is a
thing, a substance, that is dealt with, and whose attributes can be
defined. Substantives are a linguistic means of substantiation, quasi
without a warning signal to the origin of this substantiation. A short re-
flection on the role of the word 'time' in the linguistic-logical sense
when it appears as the subject of a declarative sentence, however,
already serves to exclude some potential interpretations. In the first
place, time cannot be a proper noun (in the logical sense), it does not
stand for a natural or artificial object, be it a thing or an event. Ex-
amples of such proper nouns would be 'Eiffel Tower' and 'Vesuvius'
for a natural resp. artificial thing, and 'the first world exhibition' resp.
'the first eruption of Vesuvius' are proper nouns for an artificial resp. a
natural event. In this context, such proper nouns, if we reconstruct their
appearance in instructions or statements logically, play the role of re-
placing indicator words. Indicators render every speech dependent on
its situation of utterance.
Proper nouns, on the other hand, make a discourse independent of
situations of utterance. When the before mentioned proper nouns are
available, you do not have to be able to refer to the Eiffel Tower or
Vesuvius resp. the world exhibition or the volcanic eruption in the re-
spective context of utterance, and thus do not have to limit yourself to
the situation of the immediate presence of the objects referred to. If
'time' were a proper noun it should be possible to state in which con-
texts of utterance it substitutes which indicatory linguistic expressions,
so that we may then formulate sentences of an instructing or asserting
character that are invariant to situations of utterance. One only has to
test the various options for ordinary language formulations to see that
no combination with the word 'time' meets this condition.
The same applies to a predicative character of time. Predicators have
to be defined in an exemplary way with examples and counter examples
before rendering them more precise in further terminological standard-
ization. But it should be difficult to name examples and counter ex-
amples to which we are used to, or tend to assign or deny the predicator
time. In short: 'time' is no predicator.
It was Aristotle who realised with great clarity that 'time' can be
used neither in a nominatory nor in a predicative mode. Therefore, he
determined time to be something attached to motion. In a modern way,
it can be said that we define time by reference to motions (and, as the
examples that Aristotle gives suggest, more precisely: by referring to
comparisons of motions).19 Taking up Aristotle's insight as a modern
For an interpretation of Aristotle's time doctrine, see P. Janich: Die Protophysik der
38 Peter Janich

way of representation, we can characterise and define the word 'time'

as a term of reflection. According to this, to speak about time does not
mean anything else but to use distinctions of time. Therefore, to use the
noun 'time' means not to refer to an object of utterance as its contents
but to classify the discourse itself, for example in order to distinguish
the time distinctions from others (e.g., most commonly perhaps, spatial
The choice of the word 'term of reflection' (Reflektionsterminus)20
appears appropriate because with its aid, the linguistic means them-
selves are reflected, even when, in a misleading way, objective lan-
guage seems to be spoken. Terms of reflection are defined in such a
way that the noun 'time' is determined by the adjective Ltemporal\
which, in turn, is used to classify words (and with them, distinctions) as
a meta-linguistic predicator. Lists of temporal words can easily be as-
sembled for ordinary speech but also for the language of philosophy
and science, like now, then, past, present, future, earlier, later, duration,
point of time, eternity, second, and many more. All of these may be
classified by way of the meta-linguistic predicate 'temporal'.
Accordingly, to talk about 'time' means nothing more than to use
such temporal words and, which amounts to the same, to form temporal
sentences. Consequently, time is no new object that is indicated or ap-
pears by the use of a noun, but the use of the noun is only a more eleg-
ant way to express a meta-linguistic classification. If we ask, for ex-
ample, whether time is an object of nature or of culture, it simply
means to ask for the basis of the definition and validity of temporal
statements. If, for example, somebody claims that time is one dimen-
sional or focused, then these are no natural or artificial attributes of a
given object time. Rather, what is at stake is the analysis of temporal
statements with regard to the question whether they are characterised
by a one dimensional structure or a direction. Therefore, statements
about 'time as such' do not possess new criteria of validity exceeding
those of the respective temporal statements of ordinary speech, of sci-
ence and philosophy; they do not represent additional knowledge or ad-
ditional understanding (e.g., of an 'ontology'). They are merely a short-
er, more elegant way of describing temporal statements.
Obviously, this definition of 'time' as a term of reflection is not mo-
tivated by anything more than the original concern of the fathers of the
linguistic turn: language (here the substantival character of 'time') mis-
Zeit. Konstruktive Begriindung und Geschichte der Zeitmessung Frankfurt/M.
See P. Janich: [Art.] 'Reflektionsterminus', in: J. MittelstraB (ed): Enzyklopiidie
Philosophic und Wissenschaftstheorie. Stuttgart 1995, vol. 3, 528-529.
Constituting Time through Action and Discourse 39

leads with its suggestion of a reification or objectification. Therefore,

we do not discuss philosophical problems here as if one had to accept
the imposition of phrases and questions by a philosophical or other tra-
dition, but they are deconstructed by language analysis: How could they
have been phrased at all as a question or problem, and consequently,
how did they come into existence? In the history of ideas, not every
grammatically correct phrasing of a question or a problem forces the
searcher committed to truth to look for an immediate solution or answer
to his efforts. The grammatical possibility of pseudo problems and
pseudo questions demands that we begin with a prior methodical step
and confirm that we are not dealing with the latter.
The characterization and definition of 'time' as a term of reflection
involves the consideration of the access to temporal distinctions that
are used and needed whenever time is mentioned - with whatever in-
tention and in whatever context. With this, we have reached the chal-
lenge to 'methodically reconstruct those words which in studies on
time are always taken as intelligible.
Methodical reconstruction does not necessarily mean an explicit
definition, which, on its part, is technically and methodically only pos-
sible in the context of sufficiently established language use of other
words. Reconstruction does not mean that something new has to be in-
vented, 'constructed', but that existing, time-honored speech is recon-
structed in its use, in its intentions and achievements that establish this
speech as functional. To proceed Lmethodically only means to avoid in
the course of a definition such anticipations which are impossible to
make during a course of action (e.g., the impossibility to eat a cake be-
fore it is baked), or to explicitly or implicitly allow assertions of se-
quentially that would lead to failure in action (e.g. the predication the
painting and thereafter the carving of a wooden bloc will result in a
painted statue). Methodical reconstruction of temporal distinctions has
to specify the teachability and the learnability of temporal words in
paradigmatic, conventionalised contexts of action in such a way that
the process of definition is free of circular arguments, gaps and anti-
cipations and leaps of reasoning.

3. Action and Discourse

Even though we will not be dealing in the following with an empirical
or empirical-constructive approach (e.g. in the sense of Piaget), it is
heuristically helpful for a methodical reconstruction to ask how we
achieve our first temporal distinctions as children in first language ac-
quisition, or, more precisely, to identify the prototypical standard situ-
40 Peter Janich

ations in the context of which first temporal distinctions can be learned.

It is self-evident that neither scientific nor philosophical theories of
time will play a role in this, and that the discourse on time is neither
taught nor learned through explicit definitions, and in most cases not
even with terminological support.
Mother tongue acquisition does not occur at a distance, like certain
linguistically orientated philosophies prefer to have it - according to
whom language serves primarily to describe the preexisting world -
but in and for the organization of everyday practical life, and that
means: it is primarily prescriptive. In first language acquisition, dis-
course and action are still largely unseparated aspects of the adaption to
a communal (as a rule familiar) praxis. Linguistic proficiency is trained
together with proficiency of action, in connection with an order to do or
not to do something, guided by positive or negative examples of actions
and supported by sanctions like praise and blame. In a - now already
methodologically-philosophically stylised - sequence, it is not the as-
sertive but the appellative speech that marks the methodical beginning.
(Here the methodical suggestion takes leave of a tacit general premise
of the linguistic turn according to which philosophers are dealing
mainly with logical analysis and reconstruction of assertive speech that
describes the world. This error, which is based on an assumption con-
cerning the purpose and means of scientific understanding of the world,
also found its way into the linguistically oriented scientific theory of
Logical Empiricism).
When we are dealing with the acquisition of proficiency of action
and discourse, we understand by 'action' the following terminological
specifications which are supported by examples and counter examples
well known to every normal speaker of [German]21 everyday language:

- actions can be prompted in a meaningful way;

- actions can be performed or not be performed;
- actions can be successful or they can fail in that they will realise
or miss their goal.

As far as a methodical-constructive theory of action is concerned, I

have to refer to the relevant literature.22 The distinction between action
Editorial note: In this instance, the article makes particular reference to the German
language. However, as the argument as such is not language specific, we have
bracketed the reference.
See D. Hartmann: Naturwissenschaftliche Theorien. Wissenschaftstheoretische
Grundlagen am Beispiel der Psychologic Mannheim, Leipzig, Wien, Zurich 1993;
P. Janich: Grenzen der Naturwissenschaft. Miinchen 1992; P. Janich: 'Beobachtung
Constituting Time through Action and Discourse 41

and 'mere behavior' - that which 'happens' to us like other happenings

that are no attitude, no movement of the own body (like, e.g., to be hit
by a ball) - concerns an achievement which is principally indispens-
able for the training of a normal social competence in humans: it is the
distinction between what is attributed to a human being and what is not.
Praise and blame, and every other form of sanctions are always directed
only to that which is the blame or the achievement of the educated, ex-
cept with malicious or mentally deficient 'educators'. But it does not
make sense to praise or blame what is not within the capacity of a hu-
man being. The point of this remark is that every normally minded and
normally competent human being has, at least implicitly, a concept of
action which makes a difference between what is in his/her command
and what is not, and in which success and failure can be experienced.
If we try to reconstruct how, embedded in communicative praxis,
competences of action and discourse lead to first temporal words, then
this step must be preceded, methodologically, by an acknowledgement
that what can be lived, in reconstructing retrospect, cannot be described
without temporal words. A prompt like 'Give me your ball' can only
make sense when the person addressed is not already engaged in the act
of giving the ball to the person requesting it. That means, every request
aims, in a kind of extended presence of the communication, to some-
thing that lies in the future, as compared to the prompting. As prompt-
ing is also an action which, like all action, is focused on purposes and
goals, it aims at something contra-factual, that is, toward a state of facts
that will have to be established or maintained by the action prompted
for. That means action itself has a temporal modal structure insofar as
it happens in the present and is focused on the future. And whoever un-
derstands prompts and proves this by following them, converts a future
oriented prompt into the present.
In this elementary stage of reconstruction, the past only comes onto
the stage much later. Only when, for example, the person that prompts
cannot control compliance or non-compliance with his/her order but
has to rely on a report, then there is a discourse on past actions or,
when it comes to conditional action prompts, on past, conditional ac-
tions (D. Hartmann substantiated this aspect in terms of theory of action
as a transition from prescriptive to predicative speech).23

und Handlung', in: Hans Poser (ed.): Erfahrung und Beobachtung. Erkenntnis-
theoretische und wissenschaftstheoretische Untersuchungen zur Erkenntnis-
begriindung. Berlin 1992, 13-34; P. Janich: Erkennen ah Handeln. Von der
konstruktiven Wissenschaftstheorie zur Erkenntnistheorie. Erlangen, Jena, 1993.
See D. Hartmann: Konstruktive Fragelogik. Vom Elementarsatz zur Logik von
Frage undAntwort. Mannheim, Wien, Zurich 1990, esp. 37, 38.
42 Peter Janich

The linguistic instruments described so far are still working without

explicit temporal words of their own. However, in reports on complied
or non-complied orders, the grammatical past tense might come into
play, so that, implicitly, we have to learn the temporal distinction of the
modalities linguistically in the form of temporal conjugation of verbs.
First temporal words are incorporated explicitly into the context of
prompts when these are in fact orders to perform chains of actions, and
the order of the partial actions is relevant to the person issuing the
prompt. You would say, e.g., do 'first' this and 'then' that, and if the
person addressed does not first do that and then this, there will be neg-
ative sanctions. Or you request to do something when or as soon as a
certain thing happens, and where appropriate initiate a dialogue of (pos-
itive or negative) sanctions that is subject to the conditions stated as a
temporal order.
Without question, the acquisition of first temporal distinctions does
always happen in an actual communication process that refers to the
communication situation and to joint action itself. It is the shared pres-
ence of shared action and discourse which secures the self-evident ref-
erences of temporal indicators. At first, all temporal indicators are de-
pendent on a situation as the deictic term 'now' exemplifies. Every
prompt with the indicator 'now' is bound to an actual situation of utter-
ance and makes this prompt's interpretability dependent on this refer-
ence. The same is true of the indicatory use ofLearlief and 'later' -
different from the predicative 'earlier than' resp. 'later than'. Methodic-
ally, therefore, the modes of time have to be learned primarily, nor-
mally in the grammatical forms of presence, future and past of the con-
jugation of verbs, later as adjectives 'present', 'future' and 'past', and
even later as the respective substantives.
A main extension of human discourse beyond what is available in
the situation of utterance concerns what is absent in this situation. Prop-
er names (in the logical sense) substitute deictic gestures and indicatory
words, rendering modal as well as ordinal aspects linguistically com-
prehensible with regard to the time order. In a fictional text like a fairy
tale or a detective story, we can, in this way, characterise the described
agents with past sequences of events even in the context of their acting
and talking. And we can learn to use the past perfect and the future cor-
rectly as it is provided most clearly in the Latin grammar as a verbal
form, quasi as the time modes of a speaker who, on his part, is seen in
the past or the future by a speaking observer. On the other hand, with
two-place mutually converse predicators like 'earlier than' and later
than' it is possible to express the order of events independent of the
temporal positioning of the speaker with regard to these events. For the
Constituting Time through Action and Discourse 43

user of everyday language, this case also does not present any philo-
sophical problems. We learn in actual performances of action, e.g.
when opening a bottle and pouring out of it, the words earlier and later
predicatively, and we can then in the future also apply them to new ex-
amples of sequences of actions and events in a predicative way.
(Consequently, the immense philosophical literature that followed
McTaggart's essay on A-sequences and B-sequences as the modal and
ordinal time sequences, which was quoted above, is merely owed to a
hurdle of reflection in analytical language philosophy: in the methodic-
al reconstruction given here, the inability to map the two rows onto
each other simply results from the fact that all utterances which refer to
the situation of utterance itself constitute a methodically prior way of
speaking compared to the predicative and ordinal mode in which one
can speak about the order of events independent of the situation of ut-
terance. Thus, the reason for the inability to map the rows mutually lies
only in the strict logical disjunction of a discourse that is dependent and
a discourse that is independent of its situation of utterance.)
If linguistic instruments for the organization of our actions exist in
modal and ordinal regard, we can introduce the discourse of duration
as a methodical third. In this way, one and the same scheme of action
can be actualised at different celerity, e.g. dramatically in the case of a
race. But archaic distance indications like a day's journey, area indica-
tions in a farmer's day's work or speed indicators when navigating
ships are also pre- and non-scientific examples for the fact that we
judge our actions and experiences and, generally, movements according
to the faster or the slower and thus, actions and events according to
their (estimated) relative duration.
As we live in a civilization that is dictated by clock time, speed and
tempo of life, it may be difficult to imagine a way of life in which not
even a roughly measured clock time played a dominating role in the
daily routine. For example, the complaint of an Egyptian from the time
of about 2,500 A.D. has been preserved in which he regrets the intro-
duction of water clocks because now he has to eat whenever it is time
and he is no longer allowed to eat when he is hungry.24
On the history of calculation of time (according to natural events) and me asuring of
time (according to artificial events) see P. Janich: Die Protophysik der Zeit.
Konstruktive Begrundung und Geschichte der Zeitmessung Frankfurt/M. 1980,
221-245. W. Deppert impressively proves which errors occur even in
historiography of time measuring when the own linguistic means in respect of time
have not been reconstructed methodologically. Deppert speaks on water clocks in
their reception within history of science: "Whoever [...] takes our present concept
of time for granted - a concept, however, that is not before our eyes in full clarity -
44 Peter Janich

In cultural history, calendrical time measures were established on

the basis of natural processes like the daylight and the season, method-
ically preceding the use of clocks without the need of technical aides.
These were then controlled by instruments like sun dials and astrolabes
marking observed incidents in the 'rotation' of the sky. Much speaks
for the assumption, however, that the sun dials that determined the time
of day followed the use of water meters, that means, artificial proced-
ures to measure time were introduced by a change in the use of sun di-
als (determination of the season by comparison of the highest solar alti-
tude, that is at noon).25
Time measuring, in the strict sense of a comparison of the duration
of events with the aid of artificially controlled incidents like the running
of water out of a basin, later the retarded fall of a weight in a pendulum
clock, developed entirely independently of the use of time measuring in
the sciences and also of scientific aids for a technology of time measur-
ing. Even the direction of rotation that we call 'clockwise' refers to the
cannot gain any inspiration from history because he will draw a distorted,
ultimately incomprehensible image of history. Unfortunately, a large part of
historical accounts on time measuring suffer from this flaw. Almost unanimously,
these authors arrive at the conviction that the shadow- and water clocks mentioned
here were only insufficient instruments for measuring time, for lack of sufficient
technology and mathematics with archaic men [...]. Even Vercoutter is of the
opinion, therefore, that the water clocks of Ancient Egypt did not have a constant
outflow speed because the Egyptians did not have the necessary mathematics to
calculate the respective Klepsydra-formula. And Peter Janich feels obliged to
emphatically agree to this verdict. All these authors presume a concept of time that
is bound to consistently proceeding events. But, as has been shown, there is nothing
like that in myth', (I.e., 145 [our translation]). - Obviously, Deppert did not fully
read the respective quotation in Janich. It says there: 'Here, Vercoutter seems to
have overlooked that, methodically seen, it was not lack of mathematics that denied
the Egyptian clock makers the solution of the problem. For, in order to be able to
approach the problem mathematically in the first place, the dependency of the
outflow speed to the liquid level has to be known. But only Torricelli (1646) made
the statement that the outflow speed is proportional to the radix of the liquid level
above the outflow opening. But it was already known to Heron of Alexandria that
the outflow speed is so high as if the liquid has passed the complete height in free
fall' (P. Janich: Die Protophysik der Zeit, 295-96). It should not only be critically
remarked here that Deppert did obviously not read far enough in the text that he
criticises. Up to criticism, above all, is the missing awareness for the problem in
regard to his own linguistic means, when he does not even distinguish
'continuously proceeding events' from 'continuously repeatable events' and 'events
with a constant speed'. Because of lack of conceptual clarity in his own language,
Deppert is not able to comprehend and criticise the proto theory of time adequately.
See P. Janich: "Was messen Uhren?', in: alma materphilippina 1982, 12-14.
Constituting Time through Action and Discourse 45

simulation of the course of the sun by the speed of the clock hand - that
is in western culture and in the northern hemisphere, from east to west,
and with double speed when it comes to 12-hour clock faces. Clocks
for everyday use are nothing else but instruments that copy a natural
event, which conveniently is the rotation of the earth, even in a modern,
globalised technology. Only Kant dethroned the rotation of the earth as
time norm with his indication of the slowdown of the earth rotation be-
cause of the tidal friction. Only after this argument, which proved the
exemplary reference of time measuring to a natural event as being open
to criticism, do the sciences have the epistemological problem of defin-
ing the motion of a clock in a way that it can be controlled and technic-
ally reproduced without going back to empirical insights which can
only be found with the aid of clocks. But the sciences did not recognise
this, neither in the classical nor in the relativistic physics.
It is true that Albert Einstein's special theory of relativity rightly at-
tacked the deficits in the definition of an absolute concept of time in
classical mechanics, and it rightly demanded programmatically that
physicists should not stick to 'metaphysical monsters' (Ernst Mach) but
to the observable. But the enthusiasm of the physicists for the relativist-
ic physics (and today, probably also the cult around the person of Ein-
stein) cover up how strongly the relativization of simultaneity and time
order to an observer include a naturalistic misunderstanding: again nat-
ural events are being chosen as events that measure time, and not events
that man has artificially produced with the explicit purpose of generat-
ing universal reproducibility. With this, physics of special relativity,
and, by the way, contemporary physics in general, also do not solve the
problem that is decisive in physical but also in every other scientific re-
search: the difference between useful and defective clocks. Defective
clocks are also subject to so-called natural laws, and cosequently, the
aspect of naturalness or the laws of nature cannot define which clock
would be suitable (failure-free) in order to sample empirical data. Only
human purposes and, in the sciences, explicit norms which define the
purposes of the measuring of time, can achieve a distinction between
malfunctioning and interference-free clocks. These insights were ac-
quired and founded in the protophysics of time.26
Consequently, a philosophy of the time cannot simply adapt the lan-
guage usage of physics resp. of the sciences, which is shaped by scient-
ism and naturalism. Rather, physics, in its discourses on measured time,
owes its semantics - against its naturalistic self-misunderstanding - to
its technical fundament, more precisely, to a highly developed art of
See P. Janich: 'Hat Ernst Mach die Protophysik der Zeit kritisiert?', in: P. Janich
(ed.): Protophysik heute, special issue of Philosophia Naturalis 1 (1985), 51-60.
46 Peter Janich

clock making which knows how to realise the purposes of time meas-
urements for use in physics.
With this we have explained, at least in an overview, how the lin-
guistic instruments for temporal distinction of the three aspects of time,
the modal, the ordinal and the durative, can be defined step by step and
in a methodical order, explicitly and without circular argument or a gap
in reasoning. At no point do we have to formulate or accept a hypothes-
is of the kind 'there is time outside human consciousness'. In fact, all
discourse in terms of temporal distinctions is embedded in our actions
which create the respective temporal structures, in our daily lives
(Lebenswelt) as well as in the natural sciences which emanate from it.
In short, we would not be able to speak about the ability to act (which is
assumed as an acquired ability) if we were not able to always do this in
temporal distinctions, and we could not do so if we were not able to act
in a temporally structured way.
Up to now, other aspects of time, e.g. psychological or cultural his-
torical have not been mentioned. But these cannot be primarily con-
stitutive for the reconstruction of time language because in order to de-
scribe them, you already have to make use of temporal distinctions of
everyday language - as well as in experimental psychology as in all
cultural theory. Furthermore, wherever psychologists perform their ex-
periments, they even need availability of a scientific kind of time meas-

4. Against the Seduction of Philosophers by St. Augustine

With his famous figure of thought according to which we know what
time is, as long as we are not asked to define it, but when asked find
ourselves inable to give an answer, St. Augustine has become the prin-
cipal witness of a frequently quoted yet equally absurd assumption: that
'time' cannot be conceptualised. Sure enough, this assumption is
caused, on one side, by a lack of a clear specification of the reasons for
clarification, and on the other side, by a missing set of instruments in
logic, language philosophy and theory of action. Augustine cannot be
blamed for not having it at his command but modern philosophers of
time can be. They did not take into account the conclusions of the lin-
guistic and the pragmatic turn of philosophy. To show this was one
purpose of this study. Because of that, they have fallen victim to a se-
duction which can be identified as a doctrine based on ontology. To
show this is a second purpose of the methodical reconstruction of our
discourse on time.
Constituting Time through Action and Discourse 47

Whoever assumes, like Russell in his article quoted above, that

something like a set of temporarily organised events 'exists', and that
they can claim their own individuality and uniqueness by virtue of the
time sequence, has, under a epistemological perspective, already lost
his way while taking his first step. For the assumption of a unique,
chronological sequence of events cannot be expressed other than by lin-
guistic means. Yet by doing so, you invariably have to speak in trans-
subjectively comprehensible distinctions about the relative temporal
position of one incident in comparison to another. That means, the dis-
course on time must already be available. And certainly it cannot be-
come available if the return of the same (same incidents and same ac-
tions) is supposed to be created quasi by way of a comparison of events
or actions at different times.
(Frequently, this problem can be found in literature as the assump-
tion: 'Basically there is no method to compare the courses of a clock
which follow each other. For you cannot transport the later time dis-
tance back and put it against the earlier.'27 This assumption proves to be
an error, as it looks for a definition in the domain of objects for some-
thing that lies, metaphorically speaking, in the domain of subjectivity,
and that means, in human action. If the intention of the assumption of
temporal non-comparability of incidents at different times were true,
then we could not even, in the final analysis, identify our own actions
as actualizations of action schemes and therefore, we could not act.)
The methodical sequence of the creation of terms cannot begin with
the 'uniqueness of every different event' and thus explain a 'repetition
of same events' (and under these, same actions).28 Rather, as has been
shown above, the simplest indicatory and predicative utterance about
actions and incidents is already of such a nature that it cannot make do
without the execution of repetitions of actions. But with this, the lin-
guistic disposability of temporal distinctions in general is based on the
axiomatic condition of the human ability to train actions as action
schemes, as well as on the ability to actualise these schemes when
needed. In other words, there is a methodical primacy of the ability to
act over the constitution of a discourse which allows us to identify the
uniqueness of incidents in their chronological sequence by using con-
cepts - like e.g. in calendrical measurements.
Methodically, time as the object of knowledge can only then come
on the stage when sequences of actions are linguistically structured.
See e.g. K. Mainzer, I.e., 41 [our translation].
See P. Janich: 'Einmaligkeit und Wiederholbarkeit. Ein erkenntnistheoretischer
Versuch iiber die Zeit', in: P. Rohs, J. Kuhlmann (eds.): Zeiterfahrung und
Personality Frankfurt/M. 1992,247-263.
48 Peter Janich

Even an 'ontology' that speculates on the being of time before the re-
cognition of time cannot avoid this condition. Temporary conditions of
a modal, ordinal or durative kind are not only, to put it in Aristotelian
terms, something connected to motion but, phrased in a modern way,
something that is created purposefully by language and acting in the
motion of action and differentiation itself: a construct. Whoever wants
to claim something different would have to achieve the impossible, to
speak about time without speaking about time.

Time, Tense and Topology

1. Introduction
The central question of this paper is how we should represent the rela-
tionship between past, present and future on the one hand, and the rela-
tions of temporal precedence and simultaneity on the other. In particu-
lar, should we think of the 'tenseless' relations between events, such as
today's breakfast being before tomorrow's tea, as dependent upon, or
determined by, 'tensed' facts about those events, such as today's break-
fast being past and tomorrow's tea being future? I am going to explore
this issue by considering two thought-experiments about time. These
experiments are of a topological nature: that is, they concern what we
might call the shape of time. Another, related, question concerns the le-
gitimacy of using topological thought-experiments in this context.
Both questions are prompted by a well known discussion of McTag-
gart's, in which he attempts to prove the unreality of time. I shall begin,
then, with his treatment of the topic.

2. McTaggart on Time and Topology

McTaggart drew attention to a distinction so commonplace that its sig-
nificance had eluded earlier writers, but which has come to dominate
discussions of time. As he put it, positions in time may be distin-
guished as past, present or future, or they may be distinguished by the
fact that a given position is earlier than some positions and later than
others. The A-series is the series of positions which runs from the dis-
tant past, through the near past, present, and near future, to the distant
future. The B-series is the series which runs from earlier positions to
later ones. The relations of earlier than, simultaneous with and later
than may therefore be described as B-series relations.
What is the relationship between the A-series and the B-series?
Does the existence of one necessitate the existence of the other?
Mc Taggart's answer is that there cannot be a B-series without an A-
series. That is, for there to be B-series relations between moments,
50 Robin Le Poidevin

those moments must also have A-series positions. The existence of a B-

series, in other words, is dependent on the existence of an A-series.
We need not go into McTaggart's reasons for thinking this, as they
do not directly impinge on the central issue of this paper, but we should
note that it is a crucial premise of his argument for the unreality of
time. For McTaggart goes on to argue that the notion of the A-series is
incoherent, and that in consequence any attempt to defend the reality of
time would require a conception of a B-series which is not dependent
on the A-series. What is of more immediate concern, however, is how
he deals with one objection to his contention that there cannot be a B-
series without an A-series. He considers the possibility of 'several real
and independent time-series', i.e., series which each have an internal
time-order, but whose moments bear no temporal relation to moments
of other series. McTaggart's imaginary objector now puts this diffi-

every time-series would be real, while the distinctions of past, present and
future would only have a meaning within each series, and would not, there-
fore, be taken as absolutely real. There would be, for example, many
presents. Now, of course, many points of time can be present. In each time-
series many points are present, but they must be present successively. And
the presents of the different time-series would not be successive, since they
are not in the same time. And different presents, it would be said, cannot be
real unless they are successive. So the different time-series, which are real,
must be able to exist independently of the distinction between past, present
and future.1

Here, then, is our first topological

thought-experiment, which we
may call the possibility of disuni-
fied time.1 In so far as we can
represent this - and this will turn
out to be a bone of contention -
Flgure l
we may do so by drawing lines
on paper (parallel so that they
never meet), each line representing a different time-series (see Fig. 1).

J.E. McTaggart, The Nature of Existence (Cambridge UP, 1927), ch. 33,
'Time', repr. as 'The Unreality of Time' in: R. Le Poidevin and M. MacBeath
(eds.), The Philosophy of Time (Oxford UP, 1993), 23-34, 30.
A discussion of this and other non-standard topologies for time can be
found in W.H. Newton-Smith, The Structure of Time (London: Routledge & Kegan
Paul, 1980).
Time, Tense and Topology 51

How does this pose a problem for the A-series? As McTaggart rep-
resents it, the assumption on which the objection rests is that an A-ser-
ies must be unique, in that reality can only contain one A-series. But
there cannot be a unique A-series if time is disunified.
In presenting the objection, McTaggart does not explain why any-
one should suppose that the A-series is necessarily unique, and in dis-
missing the objection he simply denies the supposition (ibid.):

No doubt in such a case, no present would be the present - it would only be

the present of a certain aspect of the universe. But then no time would be
the time - it would only be the time of a certain aspect of the universe. It
would be a real time-series, but I do not see that the present would be less
real than the time... if there were any reason to suppose that there were sev-
eral distinct B-series, there would be no additional difficulty in supposing
that there should be a distinct A-series for each B-series.

McTaggart, then, is happy to allow that, if time were real, it could be

disunified. Such a possibility would not, according to him, imply that
we could conceive of a B-series independently of an A-series.
Later writers have felt that McTaggart should not have been so tol-
erant of the idea of disunified time, and I shall examine their reasons in
§ 4 below. I want first to explore an implication of the thought-experi-
ment which McTaggart does not consider, and that is that if we allow
disunified time as a possibility, we have to give up the idea that the B-
series is reducible to, or determined by, the A-series.

3. Reductionist Programmes
When, in The Nature of Existence, he first introduces the distinction
between the two kinds of time-series, McTaggart does not talk (at least
explicitly) of reducibility: he simply says that the existence of the B-
series depends on that of the A-series. Later in the book (p. 271), when
he returns to the subject of time, he is more specific:

The term P is earlier than the term Q if it is ever past while Q is present, or
present while Q is future.

The B-series, then, is to be analysed in terms of the A-series, and not

vice versa.
The question of the reducibility of one series to the other is central
to the contemporary debate which McTaggart's discussion has gener-
ated. I shall now briefly outline this debate.
52 Robin Le Poidevin

According to the tensed theory of time, McTaggart's A-series is

real. That is to say, there exists in reality a non-relational distinction
between past, present and future. The term 'non-relational' conveys the
idea that the A-series is not reducible to the B-series. Being present, for
example, is not reducible to being simultaneous with some event. Fur-
ther, A-series facts make true tensed statements such as 'The party is
over'. In the modern jargon, the truth-conditions of tensed statements
are themselves irreducibly tensed.3 Let us call this part of tensed theory
the'existential thesis':

Existential thesis: There are in reality A-series facts, and these

facts make true tensed assertions or beliefs.

In addition to this affirmation of an irreducible A-series, tensed theor-

ists propose, with McTaggart, that the B-series is reducible to the A-
series. For example, the fact that i is present and the fact that j is
present jointly determine the fact that i and; are simultaneous. By con-
sidering all the possible A-series facts which could determine simultan-
eity, we could give a reductive account as follows:

x is simultaneous with y if and only if x is present and y is present,

or x is n units past and y is n units past, or x is n units future and y is
n units future.

This is not the only reductive account available, as we can see by com-
paring the above with McTaggart's schema. I shall consider three pos-
sible analyses in this paper. But whatever the details of the reduction,
tensed theory seeks to reduce B-series facts to A-series facts. This part
of tensed theory I shall call the 'reductionist thesis':

Reductionist thesis: B-series facts are wholly determined by A-ser-

ies facts.

We might note that the existential thesis and the reductionist thesis are
logically independent of each other. The reductionist thesis need not be
read as asserting the existence of A-series facts. We could read it as a
conditional: if there are B-series facts, then they are determined by A-
For proponents of this view, see, e.g., A.N. Prior, Past, Present and Future
(Oxford: Clarendon Press, 1967), and 'The Notion of the Present', in: Studium
Generate 23 (1970), 245-248; A. Loizou, The Reality of Time (Aldershot: Gower,
1986); J.R. Lucas, The Future (Oxford: Basil Blackwell, 1989); Q Smith, Language
and Time (Oxford UP, 1993).
Time, Tense and Topology 53

series facts. It may turn out that there are no A-series facts (perhaps, as
McTaggart argued, the notion of such a fact involves a contradiction),
in which case the reductionist thesis entails that there cannot be any B-
series facts either. So we could affirm the reductionist thesis and reject
the existential thesis. This is McTaggart's own position. Conversely,
one could affirm the existential thesis and reject the reductionist thesis.
If this implies, however, that A-series and B-series facts exist inde-
pendently of each other, or that there are in reality A-series facts but no
B-series facts, this is not a very promising line to take. That there is
some relationship between the A-series and the B-series can hardly be
denied. And the reductionist thesis is a very plausible one. If we say, of
three events, that e, is past, e2 present and e3 future, then we surely im-
ply that d is earlier than e2 and both are earlier than e3. The two theses,
then, make natural companions. Without the one, the other looks much
less safe.
Both theses are denied by the tenseless theory of time.4 According
to this theory, A-series, or tensed, statements are true in virtue of B-ser-
ies, or tenseless, facts. Here is atypical schema:

Any token u of the type 'It is presently the case that/?' is true if and
only if « is simultaneous with its being the case that/?.

On this view, there is no non-relational distinction between past,

present and future. Any such distinction must be relative to some event
or position in time. As it is sometimes put, there are no tensed facts.
The A-series does not exist in reality, but is merely a feature of our rep-
resentation of reality.
It is tempting to characterize the difference between the tensed and
tenseless theories of time simply by saying that, for the tensed theorist,
the B-series is reducible to the A-series, whereas for the tenseless the-
orist, the A-series is reducible to the B-series. But the theories are not
mirror-images of each other. The reductionism of the tenseless theory
is more complicated, involving as it does reference to representations
of reality, and the view that the A-series is only a feature of those rep-
resentations and not of reality itself. For the tenseless theorist, the A-
series only reflects our perspective on time. Tensed theorists, in con-
trast, are neither concerned to deny the reality of the B-series, nor make
it a feature merely of our representation of reality. The notion of rep-
For proponents of this view, see, e.g., J.J.C. Smart, 'Time and Becoming', in: P.
van Inwagen (ed.), Time and Cause (Dordrecht: Reidel, 1980); D.H. Mellor,
Real Time (Cambridge UP, 1981); L.N. Oaklander, Temporal Relations and
Temporal Becoming (Lanham: Univ. Press of America, 1984).
54 Robin Le Poidevin

resentation need not enter into the tensed theory at all. The B-series is a
feature of reality for both sides of the debate. What they disagree about
is its relation to the A-series.
This asymmetry between the tensed and tenseless theories is reflec-
ted in the different strategies their proponents have traditionally adop-
ted in defence and attack. For tensed theorists, the focus of the attack
has been the tenseless reductionist programme. For tenseless theorists,
the focus of the attack has nearly always been the existential thesis of
tensed theory. In this paper, the spotlight is turned on the reductionist
thesis. It is this thesis, I suggest, which makes the A-series harder to re-
concile with the possibility of disunified time than McTaggart sup-
Let us say that, if there is in reality an A-series, then what it is to be
present is to have the monadic property of presentness (and similarly
for being past and future). Now what the hypothesis of disunified time
shows, if it is coherent, is that an attempt to reduce simultaneity to the
monadic properties of presentness, pastness and futurity will fail. To
return again to the reductionist schema:

x is simultaneous with y if and only if x is present and y is present,

or x is n units past and y is n units past, or x is n units future and y is
n units future.

In disunified time, an instance of one of the disjuncts of the right hand

side could obtain although the left hand side did not obtain (see Fig. 2).
If being present is simply to have a monadic property, then it is coher-
ent to suppose that in one time-series i is present, and in another time-
series j is present. But they are not
simultaneous. Indeed they bear no
temporal relation to each other.
We could, in order to get around
this difficulty, insist that each in-
stance of 'and' on the right hand
side of the reductionist schema Figure 2
means 'and simultaneously, but
then, of course, the reduction fails.
There are two moves which the tensed theorist could make at this
eductionist schema to make it co point. The first is to revise the insist-
ent with disunified time. The second is to deny the possibility of dis-
unified time.
Let us look at the first of these possible moves. An alternative
schema, suggested by McTaggart's analysis, would go as follows:
Time, Tense and Topology 55

x is simultaneous with y if and only if, whenever x is present, y is


The temporal content of 'whenever' is made more explicit in the fol-

lowing formulation of the right hand side of the schema:

(VO (x is present at t -»• y is present at t)

If this is to be a genuine tensed analysis, the times quantified over must

be positions in the A-series. For if they were positions in the B-series,
'present at f would not define a non-relational property; 'x is present
at 4 p.m. on 1st April 1999' is no more tensed than 'x occurs at 4 p.m.
on 1st April 1999'.
The reduction of simultaneity in this alternative analysis does deal
with the problem of disunified time, for the right hand side could not
be satisfied by two events which belonged to different time-series.
However, it is not an analysis which would be endorsed by all tensed
theorists. It depends on the existence of compound, or iterated, tenses,
such as 'present five minutes ago', and the acceptability of such iter-
ated tenses has been called into question by Jonathan Lowe.5 He rejects
such tenses as incoherent, and argues that it was the illegitimate intro-
duction of iterated tenses which McTaggart relied on in his proof of the
unreality of the A-series. I shall not discuss iterated tenses here, but
simply note that they are controversial and that the controversy corres-
pondingly infects the analysis of simultaneity above.
A third analysis, which avoids the use of iterated tenses, goes as fol-

x is simultaneous with y if and only if x is as many units of time past

as j .

The condition would be satisfied if both x and y were present, for then
both would be zero units of time past. And if they were both future,
they could still be the same negative number of units of time past. The
analysis appears to be non-circular, because although it introduces
tensed properties as relations, these are not obviously B-series rela-
tions. Again, the right hand side could not be satisfied by two events
which belonged to different time-series. So here is an analysis which

E.J. Lowe, ' The Indexical Fallacy in McTaggart's Proof of the Unreality of Time',
in: Mind 96 (1987), 534-538.
56 Robin Le Poidevin

appears to meet the requirements. But something has been lost, namely
the simplicity and plausibility of the original tensed analysis. In addi-
tion, the worry remains that B-series relations have after all been
smuggled into the right hand side. In any case, I shall argue that the
second and third analyses face another topological problem, which I
shall present in the final section.
What of the second move, that of denying the possibility of disuni-
fied time?

4. Arguments from Tensed Theory to the Essential Unity of Time

It will not do for the tensed theorist to deny the possibility of disunified
time simply on the grounds that it conflicts with the reduction of the B-
series to monadic A-series properties. For precisely what the thought-
experiment calls into question is the viability of such a reduction. What
is needed is some independent reason for rejecting disunified time. To
show that it conflicted with some other feature of the A-series would
be an acceptable, non-question-begging, way of dismissing the prob-
John Lucas suggests that the connection between tense and unity
has to do with one's own perspective on time:

it would be unintelligible for me to offer a frame of temporal reference

within which I could not refer to the date at which I was then speaking. It is
part of the concept of time that it is connected to us, whereas it is not abso -
lutely necessary ... that space should be connected to us.... The essential
egocentricity of time is reflected in the ineliminability of tenses.6

There are two ideas in play here. The first is that when we use tensed
terms such as 'now', 'yesterday', 'next week', we locate ourselves vis
a vis the events we are talking about. That is, tensed terms reflect our
temporal perspective. The second is that we can only talk about time in
such tensed, and hence perspectival, terms. It follows that we cannot
coherently talk of a time-series which is unrelated temporally to the
one we are in, for we could not coherently say, of some event in that
other time-series, that it was present. To describe it as such would im-
ply that we had a location in that time-series, which ex hypothesi we do
In Dummett's discussion of McTaggart, the same two ideas are in
play, although no conclusion is drawn on the issue of time's unity. Of

J.R. Lucas, A Treatise on Time and Space (London: Methuen, 1973), 280.
Time, Tense and Topology 57

McTaggart's contention that the A-series involves a contradiction,

Dummett writes:

McTaggart is taking it for granted that reality is something of which there

exists in principle a complete description. I can make drawings of a rock
from various angles, but if I am asked to say what the real shape of the rock
is, I can give a description of it as in three-dimensional space which is inde -
pendent of the angle from which it is looked at. The description of what is
really there, as it really is, must be independent of any particular point of
view. Now if time were real, then, since what is temporal cannot be com-
pletely described without the use of token-reflexive expressions, there
would be no such thing as the complete description of reality.7

Where time is concerned, then, we can only describe it in perspectival

terms, terms which reflect our own location. Now this seems to me not
at all a happy way of putting things if one believes, as Lucas and Dum-
mett appear to, that reality contains an A-series, and that the A-series is
irreducible to a B-series. For to describe a term as perspectival, or de-
pendent upon the observer's position, implies two things. One is that
the term picks out a relation between our representation of reality and
reality itself it does not pick out something which is independent of our
representation of reality. The other is that, although the term may ap-
propriately be used from one perspective, there are other equally legit-
imate perspectives from which the term would not be legitimate. These
are certainly the implications of describing terms such as 'here' and
'on the left' as perspectival. The perspective from which something is
appropriately described as 'here' or 'on the left' is just one of a number
of perspectives. Such terms do not pick out intrinsic features of objects
or the spaces which they inhabit, but rather spatial relations between
the observer and the object or place in question. Moreover, these rela-
tions can be described in non-perspectival terms. Now if we cannot de-
scribe the temporal relations between the observer and the object or
time in question except in tensed terms, then such terms cannot be per-
spectival in the way that spatial indexicals are perspectival.
Defenders of the A-series should be particularly concerned, I think,
to reject the suggestion that there are other equally legitimate perspect-
ives from which, for example, your reading this paper is past. If time
consists of a non-relational past, present and future, then there is only
one legitimate perspective, namely, the present one. And if there is

M.A. Dummett, 'A Defence of McTaggart's Proof of the Unreality of Time', in:
Truth and Other Enigmas (London: Duckworth, 1978), 356.
58 Robin Le Poidevin

only one legitimate perspective, then it is not, strictly speaking, a per-

spective at all.
If 'here' does not provide the appropriate model of the perspectival
for our present purposes, what does? Perhaps aesthetic judgements.
'Beauty is in the eye of the beholder' is an elegant way of saying that
aesthetic qualities are irreducibly perspectival: they can only reflect a
point of view. But what is only in the eye of the beholder is not there in
the face of the beheld. The mind-dependence of aesthetic qualities robs
them of objective reality. If this is the right model of the perspectival,
then, as Adrian Moore has pointed out, to insist that time could only be
described in perspectival terms is to imply that time is simply a feature
of our representation of reality, and not a feature of reality itself8
Moore offers (p. 2) the following helpful characterization of a per-
spective, or point of view: 'When a representation of the world is from
a particular point of view, then that point of view makes an indispens-
able contribution to its content'. Another way of putting Moore's point
is that the point of view must be part of the truth-conditions for the rep-
resentation. This is precisely what we want to say about space:

Any token of 'It is the case that p here', tokened at place s, is true if
and only if it is the case that/? at s.

Now if this is our account of what it is to be a perspectival representa-

tion, that the context in which it occurs is part of its truth-conditions,
then the natural account of the A-series is surely non-perspectival. That
is, we might naturally offer the following kind of schema:

Any token of 'It is presently that case that p' is true if and only it is
presently the case that/?.

Graham Priest offers precisely this characterization of tensed theory.9

On such a view the context of the token simply does not feature at all
in the truth-conditions. The consequence of this is that whether or not a
token of 'It is presently the case that p' is true is quite independent of
when that token occurs. That is, the truth-value of such tokens is cap-
able of changing over time. Priest is quite willing to accept such a con-
sequence. For example, suppose on a dry and sunny Tuesday morning I
say 'It's raining'. That token, when uttered, is false. But the clouds

A.W. Moore, 'Points of View', in: The Philosophical Quarterly 37 (1987), 1-20.
G. Priest, 'Tense and Truth Conditions', in: Analysis 46 (1986), 162-166.
Time, Tense and Topology 59

gather and by four o'clock in the afternoon it is pouring down. At that

time, my token is true.10
This account of the truth-conditions of tensed tokens is quite com-
patible with the possibility of disunified time. Let us suppose that there
is a time-series which bears no temporal relation to the one in which
we are, and that 'The F' uniquely picks out some event in that time-ser-
ies. (Never mind how we could know that there was such an event.
Those are the facts.) Then if we say

The F is present,

our statement is false in the context in which we utter it, viz., our posi-
tion in this time-series, since the truth-conditions of the statement do
not obtain in this time-series. But the truth-conditions of the statement
could obtain in the other time-series (indeed, do obtain at some time in
that series), and so our statement - that very token - could be true in
the other time-series. Or, if the idea of a token having a truth-value at a
time other than the one in which it is uttered is too odd, we could say
that the proposition expressed by the token could have a different truth-
value at different times and at different time-series.
But now suppose that 'The G' uniquely picks out some other event
in that time-series which is not related to ours, and that it occurs after
the F. Then if we say

The F is earlier than the G,

what we say is true even in the context in which we utter it. Provided,
of course, that 'is earlier than' is tenseless. So statements of this kind,
attributing B-series locations to events, can be true in contexts where
no A-series statement about those events could be true. Hence, if dis-
unified time is a possibility, the B-series cannot be reduced to the A-
series. Now someone will complain that 'is earlier than' is not genu-
inely tenseless, but to say this is to take up a position on the relation
This, apparently, was the mediaeval theory of propositions. Peter Geach tells us: For a
scholastic, 'Socrates is sitting' is a complete proposition, enuntiabile, which is
sometimes true, sometimes false; not an incomplete expression requiring a
further phrase like "at time t" to make it into an assertion' (Critical Notice of Julius
Weinberg's Nicolaus of Autrecourt, in: Mind 58 (1949), 238-245, quoted in Prior,
Past, Present and Future, 15). The availability of such an account shows, I think,
that I am not begging the question against the tensed theorist by assuming
that there is a complete description of reality whose truth is not dependent on
our perspective on that reality.
60 Robin Le Poidevin

between the A-series and the B-series, and so does not provided a non-
question-begging objection to the topological thought-experiment.
So far we have not found a convincing argument from the existence
of an A-series to the impossibility of disunified time. But there is an-
other consideration which we have so far overlooked, and that is the
connection between tense and existence claims. Arthur Prior is one
writer who insists on this connection, and he employs it (Post,
Present and Future, p. 198-99) in casting doubt on McTaggart's

If, as I would contend, it is only by tensed statements that we can give the
cash-value of assertions which purport to be about 'time', the question as to
whether there are or could be unconnected time-series is a senseless one.
We think we can give it a sense because it is as easy to draw unconnected
lines and networks as it is to draw connected ones; but these diagrams can-
not represent time, as they cannot be translated into the basic non-figurative
temporal language. If we try so to translate them, we produce contradictions
like 'There are things going on which neither are going on, nor will be go-
ing on, nor have been going on'.

The crucial move is in the last sentence. The statement in inverted

commas is self-contradictory if'There are' is significantly tensed. If, as
Prior supposed (see especially 'The Notion of the Present'), reality is
temporally restricted, so that the future is not part of reality at all, then
to say that something is part of reality is to imply that it is not future
(and perhaps also that it is not past either). To say that something is
real is to locate it in the A-series, so the suggestion that reality contains
a series which is not part of the (i.e., our) A-series is to contradict one-
self. Here, it seems, is a quite legitimate objection to the possibility of
disunified time, one which makes no presuppositions about the redu-
cibility of the B-series to the A-series.

5. Branching Time
To summarize the discussion so far: the suggestion that time might be
disunified need not, it seems, disturb those who regard B-series facts,
such as the simultaneity of two events, as reducible to A-series facts.
Tensed theorists who are willing to countenance the possibility of dis-
unified time can produce analyses of B-series facts in terms of the A-
series which are compatible with that possibility. And those tensed the-
orists who are not willing to countenance such a possibility can appeal
to Prior's objection to the very coherence of talk about disunified time,
Time, Tense and Topology 61

that assertions of existence are tensed, and no true tensed assertion can
be made of a temporal series other than the one in which the assertion
is made.
But now I want to present a second thought-experiment, similar to
the first, but one in which these
manoeuvres are ineffective. In this
experiment, time is hypothesized to
be branching. Let us consider, for
example, a branching past (see Fig.
3). Here, two time-series, which
were not related to each other tem- Figure 3
porally, joined up to form a single
Of events in different branches, it is not true to say either that they
are simultaneous, or that one is earlier than the other. Such events,
however, will be earlier than some event in what we might call the
post-fusion time-series. There is no direct inconsistency between the
branching-past hypothesis and the use of tensed expressions to describe
it. Indeed, such a topological structure can be characterized tense-logic-
ally. Let us consider the following theorem:

[P„ (p) & P„ (q)] -^Fn(p&q)

The theorem states that if it was the case n units ago that p and it was
the case n units ago that q, then it was the case n units ago that/? and
q. In branching time, however, the possibility in Fig. 4 arises. In this
case, although p was the case n units ago and q was the case n units
ago, it was never the case that (p
& q), given that the conjunction
implies simultaneity. So the theor-
em fails. Branching time can
therefore be characterized as a to- (Now)
pology in which that particular
tense-logical theorem fails
n units
(though the temporal implications
of the conjunction make this
tense-logical characterization only
impurely tense-logical). We can Figure 4
also characterize branching time
tense-logically by some rather strange propositions. For example: 'It
was the case n units ago that/?, although there was a moment, m units
62 Robin Le Poidevin

ago (where m is greater than n), at which it was not going to be the
case that/?'.
It is important that we do not confuse branching time with branch-
ing possibility. We could think of the past (or, more plausibly, the fu-
ture) as a series of possibilities branching out from the present actual-
ity. But these possible pasts (or futures) all share a common time-ser-
ies. One way of bringing out the difference between branching time
and branching possibility is to say that nothing could occupy two time
branches, but everything occupies any number of possible pasts or fu-
tures. In thinking what we might have done, we contemplate ourselves
in some possible past, but that possibility is still located in the actual
time-series. We think what we might have done during that summer of
1978, not in some summer which has no temporal relation to the re-
membered 1978."
What this case illustrates is that the A-series positions at which p is
the case and q is the case, respectively, do not determine the B-series
relation between them. For although p's being the case and q's being
the case are both n units past, they are not simultaneous. So neither our
first suggested analysis,

x is simultaneous with y if and only if x is present and y is present,

or x is n units past and y is n units past, or x is n units future and y is
n units future,

nor our third suggested analysis,

x is simultaneous with y if and only if x is as many units of time past

as j ,

succeed in capturing the sufficient conditions for 'x is simultaneous

w i t h / . For if branching time is a possibility, it is also possible that the
right hand side of either of the biconditionals is satisfied when the left
hand side is not.
What of the second analysis? This went as follows:

I suspect that one of the reasons why Prior is more tolerant of branching time than
of disunified time is that he conflates branching time with branching
possibility. This is not to accuse him of confusion since for him time and
possibility are intimately related: see Past, Present and Future, pp. 50-1.
Lucas, who adopts a large part of Prior's conception of time, nevertheless makes
the distinction clear: see The Future, p. 101.
Time, Tense and Topology 63

x is simultaneous with y if and only if (V0(x is present at t -»• y is

present at f).

The right hand side quantifies over all A-series positions - that is, the
A-series positions events now occupy. In our branching-time example,
there is only one A-series position in which p's being the case is
present, namely n units ago. But at that A-series position, q's being the
case is also present. So q and/? satisfy the right hand side of the bicon-
ditional, but not, of course, the left hand side.
Something very similar to the second analysis, with one crucial dif-
ference, is suggested by Geach.12 His analysis (p. 98) is of temporal
precedence, but we may adapt it for the case of simultaneity as follows:

x is simultaneous with y if and only if it either is or was or will be

the case that both x is present and y is present.

The crucial difference between this and our second analysis is the in-
troduction of the word 'both'. The right hand side of Geach's analysis
would not be satisfied by our branching-time example. We cannot say
truly that it was the case that both p is present and q is present. But, I
would argue, the introduction of 'both' in an attempted reduction of the
B-series to the A-series is illegitimate. Where 'both' has temporal im-
plications, they are purely B-series implications: 'both/? and q' is equi-
valent to 'simultaneously/? and q\
So the first tensed strategy against disunified time, that of providing
alternative analyses of simultaneity, fails against branching time. Fur-
ther, the attempts to cast doubt on the coherence of disunified time do
not appear to unseat the notion of branching time. For, in describing
the branching past, we can locate ourselves temporally vis a vis those
branches, and I can truly say that it was the case that there existed two
time-series. So the fact (if it is one) that we cannot but adopt a per-
spective on time will not conflict with the possibility that the past con-
tains disparate time-series. Further, when we say 'There existed two
disparate time-series', what we say is clearly tensed, so branching time
does not fall foul of Prior's objection to disunified time.
It may be suggested (it has been, by Peter Simons) that all we have
established is simply that the usual tensed analyses of B-series facts
proceed on the (surely anodyne) assumption that time is linear. Recog-

P.T. Geach, Truth, Love and Immortality: an Introduction to McTaggart's
Philosophy (London: Hutchinson, 1979).
64 Robin Le Poidevin

nizing this, the tensed theorist could make the reduction sensitive to
different topologies. For linear time, the reduction goes like this, for
disunified time, like that, and so on. Even if this were the right ap-
proach, however, we would still need to come up with an analysis com-
patible with branching time, and I am not at all clear that this can be
done. I suspect strongly, in fact, that it cannot be done. But in any case,
the truth-conditions of a statement like '/ is simultaneous withy' cannot
be a purely contingent matter: statements have their truth-conditions as
a matter of necessity. So topological sensitivity is no virtue in a tensed
analysis of B-series facts.
I shall end this discussion by considering an objection (which I owe
to John Lucas) which is specifically aimed at branching time: that such
a topology would open the door to causal anomalies. In particular, the
objection goes, problems arise once we allow determinism into the pic-
Let R and S be two disparate time-series that fuse at some point to
form a single series T. Let t be a particular moment in T, and r and s
particular moments in R and S respectively such that both r and s are n
units earlier than t, for some n. Suppose that the universe in R and the
universe in S are both deterministic: that is, the conjunction of the laws
of nature together with the state of the universe at a particular time
makes possible only one series of states at subsequent times. Suppose
further that r makes it physically necessary that p will obtain n units
later, and that s makes it physically necessary that ~p will obtain n
units later. This seems to be a perfectly coherent possibility, for surely
what obtains in R is entirely independent of what obtains in S. It fol-
lows from these assumptions that (p & ~p) obtains at t, which is, of
course, a contradiction.
What does this establish? That the thesis that time is branching, the
thesis of determinism and some added assumptions form an inconsist-
ent set. It does not show that the notion of branching time per se is in-
coherent. It is possible to tell an inconsistent story about branching
time. But not all stories about branching time need be inconsistent. Per-
haps worlds in which the past contains two or more branches are inde-
terministic worlds. Or perhaps determinism is not well defined for
branching-time worlds.13 Alternatively, we could argue as follows: if
the laws of nature are time-reversal invariant - i.e., if, for any process
they permit, they also permit the reverse of that process - then determ-
inism is symmetrical. That is, in a deterministic universe, the state of
Jeremy Butterfield, in 'Substantivalism and Determinism', International Studies in
the Philosophy of Science, 2 (1987), 10-31, notes (17-18) the difficulties of defining
determinism consistently with non-standard topological structures for space-time.
Time, Tense and Topology 65

the universe at a given time physically necessitates a unique series of

states both subsequent to and prior to that time. In the case described
above, what obtains in T, the post-fusion series, thus determines what
obtains in the two branches. What obtains in R is therefore not inde-
pendent of what obtains in S.
It is perhaps true that talk of branching time invites the postulation
of causal anomalies, rather as talk of time-travel invites the idea of
travelling back in time to prevent one's own conception. But opening
the door to paradox is not necessarily to entail paradox. We just have
to be careful what we say about branching-time worlds, just as we have
to be careful what we say about time-travel worlds.
To conclude: I have tried to defend the idea of a branching past as a
logically possible topology for time. Such an idea is not inconsistent
with the notion of an A-series per se, nor with the related idea that ex-
istence assertions are irreducibly tensed. So it is not an unfair move to
point to branching time as casting doubt on the doctrine that the B-ser-
ies is reducible to the A-series. And if we can cast doubt on this doc-
trine, then we weaken the plausibility of the tensed theory of time.14

14 Earlier versions of this paper were read to the Philosophical Society at Oxford and
the History and Philosophy of Science Seminar at Leeds. I am very grateful to those
present, and especially to Richard Swinburne, John Lucas, Peter Simons,
Jonathan Hodge, Geoffrey Cantor and Anna Maidens, for their comments. I
am also grateful to anonymous referees for The Philosophical Quarterly for
encouragement to clarify the argument at various points.

The Significance of Time in Narrative Art1

Narrating is representing, a re-presenting of events which are not sensu-
ally perceivable to the listener. All poetry, all art, 'makes present'
(vergegenwartigt), embodies (verkorpert). This characteristic must be
considered in a wider context. The reality in which we live and which
co-determines our lives is in no way exclusively objective, or only sub-
stance based and effective in its immediate and physical presence.
Rather, it is shaped to a high degree by the impalpable (Ungreijbares),
by the absent. Philosophical and scientific doctrines about the world as
a system of invisible forces may be seen as mere theories, but very
simple reflections show how strongly these forces rule our reality of
life. To a large extent, we only know from hearsay about men and
events who affect our actions in an uplifting or destructive way. Few
knew the statesmen and generals personally and from an immediate
presence whose decisions and actions decisively influenced the fate of
millions. As for those great agents, they in turn were familiar with just
a fraction of these millions whom they used, and the dead whose works
of hundreds and thousands of years ago continue to shape our existence
and our conditions are even more inaccessible to us. What is true for
the historical agents is also true for artists and scientists, for inventors
and merchants, for the great religious men and the great teachers of eth-
ics. Contemplations such as these make us aware of invisible human
forces and systems of power which, albeit absent and completely indir-
ectly, touch and pervade us in a decisive way.
There are many ways to represent what cannot be grasped with our
senses, the absent, the non-substantial (Undingliche). One of these
ways is that of art. Painting and sculpture represent incarnate characters
to the eye; music brings movements in their internal cross-references
before the ear. When it comes to poetry, the variety of genres is imme-
diately apparent. Lyricism visualises the I or the we as they are affected
by emotions, drama enacts the interplay of contradictions or tensions,
and narration renders the course of events. In all of them, in different
This lecture anses from the draft of a morphological poetics in my Die Gestaltfrage
in derLiteraturwissenschaft. Halle: Niemeyer 1944.
68 GuntherMiiller

ways, something that in itself is not immediately present turns into a

sensually perceivable appearance, and, as such, more or less, infuses
and affects our lives.
In narration, the visualising effect (vergegenwdrtigende Wirkung)
can easily be realised (if not as easily understood). For even about the
closest friends or/and foes, the lover as well as the hater acquires a lot
of knowledge, sometimes even most of it, through narrative accounts;
and even when somebody communicates his insights, the ideas which
he formed and the decisions which he made, he always narrates about
himself. Not everyone will become aware of this to the extent that
Goethe was, who often sent to his confidants a new piece of poetry in
the format of a circular letter about events in his own life, which he did
for example with his Elective Affinities (Wahlverwandtschaften). But
this inevitably happens nevertheless. Once we have become conscious
of the fact that our life space (Lebensraum) is to a high degree shaped
by narration and consists of it, then we can easily lose the ground under
our feet before, however, winning a new foothold, less solid perhaps
but also closer to the truth.
These simple observations tell us more about the basic principles of
narrative art than if we attempted to derive the variants of the high art
of poetry from so-called primitive narration. The primordial form of
narration is not the temporally first one. The latter cannot be determ-
ined at all, and when asking such questions, one should keep in mind
what Thomas Mann said, in his introduction to the novel Joseph and
his Brothers {Joseph und seine Bruder), about the continuously reced-
ing backdrops of time. The primordial form of narration is a type just
like the primordial, the archetypal plant (Urpflanze).
Walzel2 and Petsch3 did not really succeed in making sense of their
many fortunate and detailed observations concerning narrative art be-
cause they did not aim to identify a type in the Goethean sense. That is
why they did not find a consistent comparative relation in their indi-
vidual observations. On the other hand, the most distinguished contri-
bution to the understanding of narrative art, the Theory of the Novel
(Theorie des Romans) by Georg Lukacs,4 obstructs its own view on the
poetic forms proper. For before the poetic forms as such can appear,
O. Walzel (1926). Das Wortkunstwerk Mittel seiner Erforschung. Leipzig: Quelle
& Meyer, 125-259.
R. Petsch (1942). Wesen und Form der Erzahlkunst. Halle: Niemeyer.
G. Lukacs (1920). Theorie des Romans. Ein geschichtsphilosophischer Versuch
uberdie grofien Formen derEpik. Berlin: Cassirer [English transl: The Theory of
the Novel. A historical-philosophical essay on the forms of great epic literature.
Cambridge: MIT P].
The Significance of Time in Narrative Art 69

and before they can allow the question for their structure, they are bur-
ied [in Lukacs' theory] under the historical-philosophical construction
of ages of the epos that are coherent and meaningful, and of eras of the
novel that belong to fragmented cultures lacking essential meaning and
a totality of being. The classifications of the epos and diverse forms of
the novel with their specific cultural-sociological conditions are tre-
mendously subtle, but they refer rather to historical-philosophical rela-
tions than to morphological features, and, when it comes to individual
considerations, predetermined yardsticks are used indiscriminately. In
these observations, the poetic works of Balzac, Jacobsen, Flaubert,
Goethe, and Tolstoy are dealt with more thoroughly than others, but
their formative and creative achievements are only considered incident-
ally. Consequently, this approach studiously turns away from the
simple fact that the epos as well as the novel does narrate something,
and that therefore, here like there, a basically common relation to time
does exist, even if it might be considerably modified in the various
forms of narration. If one wants to identify the design principles
(Gestaltgesetze) of narrative art, or dares to attempt a typology of its
forms, one can neither compile unrelated details, nor can one deduct
types of poetry from preconditions that are alien to poetry. Rather, one
must start with observations that will form a series, observations of
what in narrative art is comparable to the skeleton in vertebrate anim-
als, namely narrating.5
Narrating visualises (vergegenwartigt); it represents something that
has passed, makes present something that is absent. It renders the ab-
sent present without itself creating it. For all narrating is narrating
about something, of something that is not in itself narrating. In narra-
tion it is transformed and yet still represented in such a way that the
traits that have been changed will contribute to the concrete representa-
tion of the event that is in a certain sense real, but absent. Narrating
does not give a self, but it gives a rendition. While the narration that is
formed by the narrating process does of course exist as such it is some-
thing different from the narrated. Wallenstein's life, but also the life of
Jean Vatican, is something quite apart from the narration of these lives
as rendered by Alfred Doblin or Victor Hugo. The narration is a lan-
guage body (Sprachleib) with its own process and rules of becoming
As methodologically important the following should be mentioned: R. Ingarden
(1931). Das literarische Kunstwerk. Eine Untersuchung aus dem Grenzgebiet der
Ontologie, Logik und Literaturwissenschaft. Halle: Niemeyer [Engl, transl: The
Literary Work of Art. Evanston: Northwestern UP]; E. Staiger (1939). Die Zeit als
Einbildungskraft des Dichters. Zurich: Niehans. For the history of genre theory cf
I. Bekrens (1940). Die Lehre von der Einteilung der Dichtkunst. Halle: Niemeyer.
70 GiintherMuller

(Werdevorgang und Werdegesetz) that are completely different from

those inherent in the narrated event. And still, this difference is of such
a nature that the language body is completely designed to make the nar-
rated vivid, to realise it. The narrated, however, is a temporal process.
In art, each genre has its own manifestation of time. Especially V. v.
Weizsacker has given us insights into the difference between physical
and biological time that could be made fruitful for the morphology of
poetry.6 The poetic time, that is, the time that rules poetry, is more
closely related to biological time, and that is not amazing, as 'art is an-
other nature' according to Goethe's conviction. This is also true of nar-
rative poetry, of the linguistic representation of events. But here, addi-
tionally, we have a close relation to physical time, to the time of the
clock, for what is represented here is a time process in space. Even
where emotional processes are described in their movements beyond
time and space, a proportional relation to the 'outward', to the physical
time, establishes itself, whether explicitly or not. This becomes clear,
for example, in the sense that all this or so little happens within so few
or so many minutes and hours. And this relation, this interacting of
emotional, biological, and physical time not only really happens; it also
constitutes the main force of every narration: its rhythm. Admittedly,
many other factors are also effective here, for example the empathy or
distanced observation of the narrator, the passion or aloofness of the
narrated human beings, continuous or erratic development of the
events, the alteration between the powers of guidance and accumulation
{Fuhrkraft und Schwellkraft).
Like all poetry and music, also pure lyric poetry moves within phys-
ical time when it is read or spoken. But whatever it renders present has
no relation to this time, rather, it is removed from time. Considering
Sappho's poems or Goethe's 'An den Mond' ('To the Moon'),
Petrarca's sonnets, Holderlin's 'Patmos' hymn or Shelley's 'Ode to the
West Wind', it is senseless to ask about the 'when'. Although the stan-
zas of those poems are not exchangeable, their organising principle is
not time. They are bound in their sequence by other strings of pro-
cesses, for example by the transition of one wave into the other, the
emergence of an imagination, a thought, one out of the other. In pure
drama, the sequence of time is indeed formative, but it serves essen-
tially as the unavoidable form in which human contradictions that are
not time-bound are expressed. By comparison, one may say that time in
fact does belong to drama like it belongs to a game of chess. On the
other hand, true narrative deals with the sequence of events as some-

V. v. Weizsacker (1942). Gestalt und Zeit. Halle: Niemeyer.
The Significance of Time in Narrative Art 71

thing temporal, it deals with with time as something that is filled with
events and in turn eventuates them, with the temporality of life.
Lukacs also recognised this in connection with a special type of the
novel, the Sentimental Education (L'Education sentimentale), but he
wanted to limit the observation to this type. However, in the narrative,
the unstoppable flow of life-time, whose timeless experience can mater-
ialise in lyric; this flow itself is what is represented, in the Iliad as well
as in the Aithiopika, in the Jerusalem Delivered (La Gerusalemme
liberta), the Simplicissimus, the Elective Affinities (Wahlver-
wandt-schaften), in The Magic Skin (Le Peau de chagrin), in the Green
Henry (Der Grune Heinrich), in Niels Lyhne, in the ^ e x - n o v e l s , in
In Search of Lost Times (A la recherche du temps perdu), in Of Time
and the River. It could almost be stated that the more time-bound life is,
the purer the epic will be. From this point of view, the primordial form
of epic is the 'and then', a formula that sharply illustrates the difference
of epic poetry to lyric and drama. The 'and then', however, is not yet an
answer, but it points to the essential questions of composition. Among
these questions, that of narrative time will be examined more closely in
the following as morphologically, it has an especially visible and com-
prehensible significance.
Fielding, the founder of the coming-of-age novel, raised the ques-
tion of narrative time and answered it with serene clarity. In Tom
Jones, he elaborates:

Though we have properly enough entitled this our work, a history, and not
a life; nor an apology for a life, as is more in fashion; yet we intend in it
rather to pursue the method of those writers who profess to disclose the re-
volutions of countries, than to imitate the painful and voluminous historian
who, preserve the regularity of his series, thinks himself obliged to fill up
as much paper with the details of months and years in which nothing re-
markable happened, as he employs upon those notable eras when the
greatest scenes have been transacted on the human stage. [...].

Now it is our purpose, in the ensuing pages, to persue a contrary method.

When any extraordinary scene presents itself (as we trust will often be the
case), we shall spare no pains nor paper to open it at large to our reader; but
if whole years should pass without producing anything worthy his notice,
we shall hasten on to matters of consequence, and leyve such periods of
time totally unobserved. [...].

My reader is then not surprised if, in the course of this work, he shall find
some chapters very short, and others altogether as long; some that contain
72 GuntherMiiller

only the time of a single day, and others that comprise years; in a word, if
my history sometimes seem to stand still, and sometimes to fly.7

Consequently, in the headlines of his seven books, he states exactly the

periods of time:

Book II: [...] scenes of matrimonial felicity in different degrees of life; and
various other transactions during the first two years after marriage
between Captain Blifil and Miss Bridget Airworthy
Book III: [...] from the time when Tommy Jones arrived at the age of four-
teen, till he attained the age of nineteen. [...]
Book IV: [...] the time of a year
Book V: [...] a portion of time somewhat longer than half a year
BookVI:[...] about three weeks
Book VII: [...] three days
Book VIII: [...] above two days
Book IX: [...] twelve hours
BookX:[...] about twelve hours
BookXI:[...] about three days
Book XII: [...] the same individual time with the former
Book XIII: [...] twelve days
Book XIV: [...] two days
BookXV:[...] about two days
Book XVI: [...] five days
Book XVII: [...] three days
Book XVIII: [...] about six days

However, nothing is thereby said about the internal relation between the
expanses of clock time, and the mode of defining time is rather ironic
and seems to make some fun of the reader as well as of the author. Yet
the humorous tone is characteristic of Fielding's work in general and
does not at all weaken its deeply humane seriousness. The other theor-
etical introductory chapters to each of his books, too, have considerable
weight, in spite of their humorous ease of expression. With references
to the alternation of standstill and flow of narration, of disregard and
detailed description they touch upon a basic form of all narration, the
very primordial form that is closely related to the rhythmic flow of life
time and experienced time.
Research on narrative forms has paid little attention to this basic
form - presumably because of the strong challenges which especially
the epos, the novel and the novella pose in terms of their themes and

H. Fielding (2008). Tom Jones. Ed. by J. Bender & S. Stern. Oxford: Oxford UP,
book II, chap. I, 67f.
The Significance of Time in Narrative Art 73

subject matter. This is not the place to attempt a compilation of all the
relevant statements taken from the history of narrative theory. Let me
emphasise one reference only. Thomas Mann, who, since The Magic
Mountain {Der Zauberberg), is magnetically attracted by the problem
of time, mentions this basic form of narrating several times in his
Joseph-novel, using the expression 'gap' ('Aussparung').
It is true that the Joseph-novel provides a special case of structuring
time, for while Joseph's short account of the Old Testament is strictly
adhered to, the emergence of this course from hitherto silent domains is
developed in breadth and width. However, as will be shown later, this
is also just one special case of narrating in general, and Th. Mann's re-
marks concerning the specific character of the time of narration are ob-
viously of an instructive and illuminating nature.
In the beginning of the fourth main chapter of volume IV, it is ex-

Der Lakonismus des bisher davon Uberlieferten [des Gesprachs zwischen

dem Pharao und Joseph] geht bis zur ehrwiirdigen Unwahrscheinlichkeit.
DaB nach Josephs Traumdeutung und seinem Ratschlag an den Konig, sich
nach einem verstandigen und weisen Mann, einem Mann der Vorsorge um-
zusehen, Pharao ohne weiteres geantwortet habe: "Keiner ist so verstandig
und weise wie du; dich will ich iiber ganz Agyptenland setzen!" und inn in
wahrhaft enthusiastischer - man kann schon sagen: ziigelloser Weise mit
Ehren und Wiirden uberschiittet habe, - das schien uns immer der Abkiir-
zung, Aussparung und Eintrocknung zu viel: wie ein ausgenommener, ge-
salzener und gewickelter Uberrest der Wahrheit erschien es uns, nicht wie
ihreLebensgestalt; [...].

The laconic nature of the tradition up till now [i.e. of the conversation
between Pharaoh and Joseph] almost makes it, however venerable, uncon-
vincing. For instance upon Joseph's interpretation and his advice to the
King to look about for a wise and knowledgeable and forethoughted man,
Pharaoh straightway answers: "Nobody is so knowledgeable and wise as
you. I will set you over all Egypt." And overwhelms him on the spot with
the most extravagant honours and dignities. There is too much abridgement
and cutting-away about this, it is too dry, it is a drawn and salted and em-
balded remnant of the truth, not truth's living lineaments.8

(In another passage, he explicitly speaks of 'mummified truth'.) Sub-

sequently, the narrative gap is discussed:

Th. Mann (1990). Joseph der Ernahrer (Joseph und seme Bruder. Gesammelte
Werke m dreizehn Banden). Frankfurt/M.: Fischer, 1478; (1944). Joseph the
Provider {Joseph and his Brothers). New York: Knopf, 229f
74 GuntherMiiller

Wohlverstanden, wir haben nichts gegen die Aussparung. Sie ist wohltatig
und notwendig, denn es ist auf die Dauer vollig unmoglich, das Leben zu
erzahlen. Wohin sollte das fuhren? Es ffihrte ins Unendliche und ginge
iiber Menschenkraft. Wer es sich in den Kopf setzte, wiirde nicht nur nie
fertig, sondern erstickte schon in den Anfangen, umgarnt vom Wahnsinn
derGenauigkeit. [...]
Was ware aus uns geworden ohne Aussparung, als Jakob diente bei Laban,
dem Teufel, sieben und dreizehn und ffinf, kurz: funfundzwanzig Jahre
lang - von denen jedes winzigste Zeitelement ausgefullt war mit genauem,
im Grande erzahlenswertem Leben? Und was sollte jetzt aus uns werden
ohne jenes vernunftige Prinzip [der Aussparung], da wiser Schifflein, vom
mafiig gehenden Strom der Erzahlung dahingetrieben, wieder einmal an
den Rand eines Zeit-Katarakts bebt von sieben und sieben geweissagten

Of course, there is really nothing against the cutting-away in itself. It is

useful and even necessary. In the long ran, it is quite impossible to narrate
life just as it flows. What would it lead to? Into infinite. It would be beyond
human powers. Whoever got such an idea fixed in his head would not only
never finish, he would be suffocated at the outset. Entangled in a web of
delusory exactitude, a madness of detail. [...]
What would have become of us, for instance, when Jacob was serving with
the devil Laban, seven and thirteen and five - in short, twenty-five years,
of which every tiniest time element was full of a life-in-itself, quite worth
telling? And what would become of us now without that reasonable prin-
ciple [of cutting-away], when our little bark, driven by the measuredly
moving stream of narration, hovers again on the brink of a time-cataract of
seven and seven prophesied years?9

This all sheds some light from a slightly different point of view onto
these same traits which we already observed. Only the expression that
life once narrated itself may need an explanation in our context. Voiced
by the great narrator, he transfers its representation on that which he has
represented. In this case, it is an incident that was firstly narrated in the
First book of Moses, that is, as an incident that happened outside the
narration, in the time and life of the arch fathers. Every historical novel
that uses historical sources displays such a doubling effect, and like-
wise every epos that, like the Iliad or the Nibelungen-sa&i, refers back
to older poems that are closer to the time of the events. But in any case,
to claim that the incidents given in the first narration are something that
has happened in the temporal space of real life and have thus been nar-
rated by life itself, can only be taken as a simile. Life does not narrate
itself, life lives itself. Life does not leave anything out, but it is com-
Th. Mann (1990), I.e., 1479; (1944), I.e., 230f.
The Significance of Time in Narrative Art 75

plete down to the most delicate movement of each individual cell. Like-
wise, every incident that has not been distilled from historical sources
but is created by the imagination is also designed as a complete interre-
lation of events. On the other side, Th. Mann's Joseph-novel itself also
shows the significance of imagination in the historical novel. Therefore,
in academic research, the expression that life narrates itself can only be
used with strong reservations. Reservations, because the difference
between narrating and narrated should not be concealed. This differ-
ence does exist, no matter if the events are historically confirmed or
created by imagination. In every narration, events are represented not as
narrated but as lived incidents of life living itself. Consequently, it is
the difference between the narrated and the narrating that is decisive to
the individual forms and the morphology of the art of narration. With
its paradoxical oneness of separation and relation, this distinction is the
root soil for a 'natural form' of poetry.
One aspect of this polar unification is, as Th. Mann stresses, the im-
possibility to narrate life in the way it once happened. It is an important
insight to point out that in reality every tiny element of time is and was
filled with a detailed life that is basically worth being narrated. Even
James Joyce and his school are not able to completely represent this
saturation of every tiny element of time. Gaps cannot be avoided, even
when complete accuracy is intended indiscriminately.
With Thomas Mann, cutting-away obviously means a meaningful
gap. I would like to suggest the term 'time contraction' ('Zeitrqffung').
For, in short, the narrator does not simply leave something out, he con-
tracts the narrated time continuously, but to a different degree.
Up to now, narrating has been repeatedly distinguished from the
narrated. This includes the relation between the time of narration and
narrated time. Now, this should be scrutinised more thoroughly. -
Grellmann in his encyclopedia entry, 'Roman' ('Novel'),10 also distin-
guishes between the narrator and the narrated. Yet, he only pays atten-
tion to the subjective appearance resp. his objectivity and the ideologic-
al conviction of the narrator and does not see what is basic to all narrat-
ing: namely, the intertwining of different courses of time.
In order to narrate a story, the narrator needs a certain span of phys -
ical time. Its measurement by a clock is characteristic of the peculiarity
of this time, for the clock allows a projection of time into space and
thus a spatial measurement of time - for measurement of time proper
does not exist. But when a star or the hand of a clock has covered this
and that spatial distance, then time has run so and so 'long'.
H. Grellmann (1928/29). [Art.] 'Roman,' m: Reallexikon der deutschen Ltieratur-
geschichte. Ed. by P. Merker & W. Stammler. Berlin: de Gruyter, vol. 3, 62-72.
76 GihitherMiiller

Therefore, there is no basic difference between counting the time of

narration in minutes or in the number of printed pages of a certain
work, and if a page norm existed, we would have a unit as useful as a
normative clock. The measurable time of a narration should not be con-
fused with the time that is needed for the creation of a narrative work to
be completed. The time of creation may take many years, like the Iliad
only received its present form, and thus its present time of narration, in
the course of many centuries. And anyway, the time of creation of a
narration is necessarily longer than the time of narration, meaning the
time during which a work that has now been completed by the poet nar-
rates its story - in cases in which an oral narration comes first and is
then fixated in written form, like with Brentano's 'fairy tales'
(Mdrchen), there are certain variations which we may neglect here.
Also the fact that various readers read or recite the same story during a
longer or shorter period of time belongs to a different time domain, to
that of the interpreting presentation of the language body in a space of
time. This space of time is caused by a marriage of the recipient with
the gestalt of the work, and the individual ways of these recipients ne-
cessarily influences their own reading speed. But above all, it is import-
ant to note that the time of narrating has a different extension, continu-
ity, and repeatability, that this time is of an essentially different kind
than the narrated time.
Let us first remain in the area of temporal extension. In parts, the
time of narrating may be sometimes shorter, sometimes longer than the
narrated time, but altogether, it is almost always shorter than the latter
because of the inevitable gaps, the time contraction. In The Sorrows of
Young Werther {Die Leiden des jungen Werthers) and in Elective Affin-
ities (Wahlverwandtschaften), the narrated time stretches roughly over
one and a half year, in Wilhelm Meister's Apprenticeship (Wilhelm
Meisters Lehrjahre) roughly over ten years, in the classical baroque
novel Argenis (Histoire de Foliarque et d'Argenis) from a summer to
the following autumn, the time in Sentimental Education {^Education
sentimentale) covers twenty-seven years, in the Buddenbrooks fortytwo
years, in the Iliad almost two weeks, in The Egoist by Meredith one
week. The time of narrating, however, can be counted in hours in all of
them. And even when, in works of the school of Joyce, the time of nar-
rating and the narrated time are almost identical, one cannot ignore the
fact that contractions and gaps have been created even here. As said be-
fore, it is virtually impossible to narrate a life span completely and ex-
haustibly because, otherwise, the movement of every individual cell,
every muscle, every metabolic system had to be narrated. Here, the in-
scrutability of life manifests itself quietly and impressively, and the ob-
The Significance of Time in Narrative Art 77

server is pointed to the difference of the two mutually related time con-
cepts in a narration as well as to the different measurements of time
contraction possible in this constellation. At the same time, it provides
an illustration for the paradoxical saying that all creative forming is an
Very generally, we can distinguish between three main sorts of nar-
rative time contractions. One is the simple skipping of time spans. That
can be done explicitly ('several years after', in Kleist's Bettelweib von
Locarno) or without special reference and can apply to hours as well as
to years. This way of time contraction can be especially well observed
in the structure of Jiirg Jenatsch. Again and again, larger periods of
time are cut out, whereas the narrated phases are broadly represented.
Secondly, there is the contraction of time in large steps or main
achievements in the way of 'vem, vidi, vici'. This emerges very ab-
ruptly in the The Brothers Karamazov, of which the first book narrates
almost thirty years in strongly simplified main lines, while the follow-
ing books develop three days, in the way, that the time of narrating and
narrated time are for long periods as largely congruent as they are in
Joyce's Ulysses, although, here and there, the style of representation is
completely different - in Dostoyevsky's work, it is the consequence of
the extensive part that dialogues play in the narration. In a third variant,
the individual movements and incidents are contracted into the general
traits of a transitional conditionally - here, one could speak of iterative
and durative traits, in analogy to the terms for verbal forms of action,
for example, when we read: 'Now, he rode out daily' or 'For weeks he
could not free himself from the idea'. To a large extent, Stifter's Witiko
is considerably shaped by a low degree of contraction of iterative and
durative actions.
It is obvious that one or the other way of time contraction has a
formative effect on the rhythm and the structure, and consequently on
the gestalt of the work as a whole. But to avoid misunderstandings, one
has to keep in mind that in each work, there are numerous, not to say
countless, contractions and that these do not necessarily have to be of
the same kind. Decisive are not just the most numerous kinds of time
contractions in a work, but the ones that leave the strongest mark, or, to
be more precise, what is most important is the relation between the fre-
quency and the impact of the various time contractions in a work. Em-
phasised periods of time that are portrayed as especially meaningful in
a work of poetry are generally less contracted, although, for example,
the two murders in the Jew 's Beech {Die Judenbuche) are not explicitly
narrated, they are in fact omitted in the narrated time.
78 GihitherMiiller

Another difference adds to this. There are narrations in which the in-
dividual progressions are very clearly defined in order of their place in
a certain period of time. It is narrated how many days or hours have
passed between this and that event; hour, day, and year of an event are
explicitly given. Thus, the narrated time is fixed in a quasi calendncal
way. The narrative works of Herzog Anton Ulnch of Brunswick have
this tendency, but also Goethe's; Bel ami, Buddenbrooks, and the For-
syte Chronicles are strikingly exact in a calendncal way. On the other
side, there are narrations that narrate almost nothing about a relation to
clock time. Not only the Hellenistic novels with a good part of their fol-
lowers belong into this category, but also many fairy tales, and among
the works of high art such works as Henry of Ofterdingen (Heinrich
von Ofterdingen) by Novalis. In the same way, in the Artus epics, it
would be difficult to limit the individual events to a certain time of the
day. Now, there are countless transitions between the two border cases
of clock-time accuracy and clock-time independence, and even in one
work, generally, the level of accuracy is not consistently the same. For
example, with E.T.A. Hoffmann or Meyrink, the change between
clock-time independent periods and those that do have exact clock time
is caused by the narrated. However, when one realises how exact in
calendncal terms time is determined in Werther, for example, one will
be cautious to make conclusions offhand about attitudes toward life that
are denved from the kind of time of nanating, and to interpret
calendncal accuracy as an expression of a view of life that is determ-
ined by physics. Without a doubt, only so much can be said, namely
that here and there, a different expenence of time by the poet is influen-
tial, that connected to it is a different relation of life to the spatio-tem-
poral world, and that, in fact, a different mode of creation, a different
type of form, a different basic law of becoming will be related to it.
How to judge this individually can only be decided case by case.
This is confirmed by the following observation. All nanating is a
nanating of something that is not nanation, but a process in life; this
process occurs in the spatio-temporal world, even if it is an inner exper-
ience of the soul. If a nanation contracts clock time such that the spa-
tio-temporal condition is contracted away, then it makes an obvious
choice with an interpreting effect and as the result of a sense-giving at-
titude towards reality. This is no less the case when the clock time is in
fact not contracted away, but narrated precisely.
But not only these special contractions add significance, every con-
traction does so. To skip a sequence of events, that means, to contract it
away, to make it denser or more expansive, actually means to give sig-
nificance to it or to remove it from some point of view. What Achilles
The Significance of Time in Narrative Art 79

does in detail while he angrily stays away from the fight, what Parsifal
does in detail while he angrily keeps away from God - in Homer, in
Wolfram, it has been contracted away. But, from a different point of
view, this may be especially worthwhile to narrate. The months spent
painfully by Wilhelm Meister after the loss of Marianne, the months
spent storming through during the war by Eduard in Elective Affinities
(Die Wahlverwandtschaften) could very well be narrated in elaborated
detail instead of durative-iterative contraction or in short reflections, if
there was not a rule of value and consequently, an interpretative inten-
tion that led to just this contraction of the narrated time.
With this, an important phenomenon comes into sight, that not just
the explicitly evaluating opinion turns the narration into an interpreta-
tion. Rather, it is a basic structural law of narrative art that already one
of its elementary forming processes, the representation of time periods
in a tension between time of narrating and narrated time, has an inter-
preting effect. Remarks by Goethe are helpful to grasp this more
clearly. He says in his 'Introduction to the German Gil Bias' (Geleit-
wort zum Deutschen Gil Bias) that the novel as an ethical phenomenon
of art is rightly expected to have an 'inner consequence, which, even if
we are led through so many labyrinths, appears again and closes
everything as a whole in itself [our translation]. This is in accordance
with the interpretation of catharsis in Goethe's 'Afterthought to Aris-
totle's Poetics' CNachlese zu Aristoteles' PoetiV), in which the famous
controversial quotation is translated: 'Tragedy is the mimesis of a signi-
ficant and completed action, which [... ] after a period of sympathy and
fear, closes its business with the balance of such passions' [our transla-
tion]. The 'Geleitwortzum Deutschen Gil Bias' continues: 'But human
life, faithfully recorded, never presents itself as a whole; after the most
wonderful beginnings, bold progresses follow, then accident interferes,
man recovers, he begins, perhaps on a higher level, his old play, which
was agreeable to him, then he either disappears early or disappears
slowly without every knot that was tied having been untied' [our trans-
lation]. Here, an important point of view, although in no way the only
one, is given under which a contraction of incidents, and that means
contraction of narrated time, may happen.
When dealing with Diderot's essay on painting (?Versuch tiber die
Maleref), signification as a reason for contraction is even more obvi-
ous, and whatever was observed on the relation between life time and
narrated time, gains clearer contours here. There, Goethe says: 'Nature
seemingly acts on behalf of itself, the artist acts as a human being on
behalf of other human beings. In life, we only barely select the desir-
able, the enjoyable. Whatever the artist offers to man, should all be
80 GihitherMiiller

comprehensible and comfortable to the senses, all be enjoyable satisfy-

ing, all be inspiring, educating and elevating to the spirit; and con-
sequently, the artist, grateful to nature that also created him, gives a
second nature back to it, but an experienced nature, a reflected nature, a
nature perfected in a human sense' [our translation]. Shortly before,
Goethe writes: 'Nature organises a living, indifferent being, the artist a
dead, but a meaningful being, nature a real being, the artist an apparent
being. The beholder himself has to add, significance, emotion, reflec-
tions, affect, and effect on the mind to the works of nature. In the work
of art, he will and has to find all this already present' [our translation].
In regard to the narrated time, the following can be concluded: the
poet narrates a sequence of events which has happened independently
of the act of narration, outside the narration. He re-presents it, but not
exactly in the way as it is in itself, but poorer in one regard, richer in
the other. What Goethe says about nature is also true of narrated events:
these events are nature, they are perceived or represented as nature by
the poet as well as by the listener outside in nature, and consequently,
they initially appear to the observer as 'indifferent', as inarticulate pro-
cesses, even alien to meaning.
To make this clearer by an example: at first, the incidents represen-
ted in the Aeneid, in Fortunatus, in The Red and the Black (Le rouge et
le noir) or the ever newly narrated Joseph-story take place in front of
the eyes of the poet, and there, they are indifferent events in nature.
They happen like the lightbeams emitted by the sun, the bonding and
separation of chemical elements. The Hymn 'On the Devine' ('Das
Gottliche') announces: 'For the realm of nature is unfeeling' [our trans-
lation]. A diversity offerees of nature act together so that events may
happen and proceed. Human emotions, reflections and movements of
the will also belong to these forces of nature. But these differ from each
other, in the first place, only by the degree of their contribution toward
the emergence of the sequence of events, not by their rank, although
man comments emotionally, intentionally and reflecting on the course
of nature to which he belongs, and tries to influence it in an escalating
or defensive way. These incidents themselves say as little about a tran-
scendental meaning, about significance, as the direction of the Gulf
Stream, the advancement or retreat of a glacier, the stormy movements
around an atomic nucleus, the cleavages of a fertilised egg cell. They
are incidents like any other incidents of the sublime unfeeling, unthink-
ing, unwilling nature, even if they can be perceived with the human
eyes of the body or the mind. If they have a time, it is the time of the
metronome, applying a technical measure with an agreed measuring
unit to them.
The Significance of Time in Narrative Art 81

This is what Goethe's grim sentence - nature organises a living, in-

different being - means. And this, above everything else, turns this
reality into the adequate elemental material for artistic creation, for rep-
resentation in poetic narrating. For it is the narrator who adds articula-
tion, value and degradation to the 'indifferent being'. It is he who de-
velops out of the senseless sequence of events 'significance, emotion,
thought, effect, impression on the mind'. He makes them 'nurturing,
educating and elevating to the mind', and by giving significance to a
symbolic case of life, he awards human meaning to the world; to wit:
his meaning, the meaning that he believes in, the meaning he has exper-
ienced, achieved, and suffered. The basic structure of meaningful artic-
ulation, however, results from the explicit narrating of those time se-
quences, which are important to the creation of meaning, and from the
cutting-away of those, which are unimportant to this purpose, through
time contraction.
Under the forming hands of the poet, the relentlessly silent natural
process turns into a moved and moving narration about the deeper hu-
man meaning of human life, even where despair about meaninglessness
is articulated in the work, for then the longing for meaning and signific-
ance is the humane that continues to exist. Even where renunciation of
meaningful interpretation is the result, something is awarded to the nat-
ural event, namely, the character of senselessness, which is no less alien
to the natural event as such than the character of the significant and the
The narrator may only rarely be conscious of these interrelations. In
most cases, he might form the narrated event with judgmental empathy.
But that he is unable to avoid these morphological conditions, is an in-
escapable rule, like metamorphosis in organic natures. But the poet cre-
ates his work, which creates human meaning, through forming, and the
structuring of the narrated by time contraction is one of the leading
forces of creation.
Time contraction articulates. It leaves out, more or less, those of the
meaningless events that are unimportant or incidental in the meaningful
context of the given work. It intensifies and relaxes the time periods, re-
leases from calendar time or subjects to it. It can represent the beat of
the biological pulse, the swing of the physical pendulum, the curve of
the spiritual wave, together or individually, as the decisive force. It can
contract time with respect to the most personal, the least personal, or
those incidents beyond the personal; it can show, all in all, only one, or
in multi-voiced exchange, many kinds offerees or incidents as effect-
ive. It can relate the narrated, the natural event, as closely as possible to
its original course, but it can also anticipate what comes later and catch
82 GihitherMiiller

up with what was earlier by virtue of the polar relation of the time of
narrating and narrated time. It can give a voice only to the event but
also to the sympathy of the narrator, to his empathy and his reflection,
to his distant statements, his confidence, his distress, his humor.
All this and much more that could not be specified here, gives hu-
man meaning, human significance to the natural event which is alien to
meaning. It co-builds the creative construction of the narration in a de-
cisive way, the shape of its sound and the beat of its pulse. This, how-
ever, is the formation-reformation of all poetry. It corresponds to the
formation-reformation of organic natures, up to the metamorphosical
growth from the mono cellular to the multi cellular entity. The language
body (Sprachleib) of poetry is shaped by the movement of the emer-
gence (Werdebewegung) of narrating. Like every other language, the
language body of poetry not only articulates itself but it also speaks
about something else which it causes to appear by its representation.
Regardless how important theoretical statements and the reported indi-
vidual events may be in a narration, it is the language of the complete
movement of emergence, of the sounds, the sequence of images, shaped
from all inventory and forces of the work, it is this language that makes
poetry become poetry. Knowledge and action have their own lan-
guages. The language of art, of poetry, of narration, is form, is gestalt.
Time, as has been shown, has a basic significance to the formation of
narrative art. It is the one formative force that allows a comparison
between all narrative works with respect to this decisive trait of the
common form and thus makes the identification of morphological lines
and groups possible. This, however, is the indispensable condition in
order to arrive at a typology of narrative art to which we ought to as-
Finally, an unintended result of these observations has to be
touched. Art belongs to the signifying forces of human life that creates
meaning, and at the same time belongs to those entities, which are cre-
ated by a natural force, when the appropriate conditions arrive, without
being asked for and without a purpose, like plant and animal. It sup-
ports the shaping and the liberation of the sense of being, of the experi-
ence of the world. It carries us away, as is said in Schubert's song, into
a better world - a higher, deeper, wider world, so to speak, a world that
is more humane because it is meaningful. Through this, it is able to
strongly, sometimes decisively, influence the lived human life. But it is
not able to change the unfeeling world that is alien to meaning, to abol-
ish or to alter its cycles of physical or biological nature that are indiffer-
ent towards human interpretation. Rather, within the enormous and in-
comprehensible edifice of nature, it acts like a natural force in itself. As
The Significance of Time in Narrative Art 83

a world with its own meaningful, significant way, it rotates in its own
orbit according to its own rules, which are analog but not identical to
the growth-oriented world of life and only touches the physical-chemic-
al-biological world insofar, as mind and spirit of man is connected to it.
It can radiate its vitalizing meamngfulness to the mind and spirit of
man, and, from there, also influence bio-chemical events to a certain
degree. It is true, this still harbors enough dangers, but the grand nature
does not know fear, and as an apparition of nature with its own rules,
art plays its part in the diversified fabric of the human domain to
achieve the sublimation of life.

The Timelessness of Poetry*

I was asked by the editors to comment on Herbert Seidler's polemics
against my theory of the epic preterit.1 Kindly allow me to do this as
follows. Instead of dealing with all of Seidler's individual attacks,
which derive from a point of view that is, as one might say, more mo-
tivated by an 'emotional stylistic' than by a view deriving from the lo-
gics of discourse, I only would like to discuss and elucidate, if possible,
some principal problems. For the rest, I would like to refer to the book
that I hope soon to present: Die Logik der Dichtung {The Logic of Lit-
erature). Furthermore, may I consider the object of my - by necessity -
polemic response as a general one, by dwelling on some of the views
and problems mentioned in the discussion, but without quoting the indi-
vidual works. What is at stake is one point, the problem of time which
has impinged upon the theory of fiction, and which has, from the begin-
ning - already with Lessing, already with Goethe and Schiller -, invited
an erroneous approach that misinterprets the 'mode of being' ('Seins-
weise') of fiction. As I was made to understand, it can only be elucid-
ated and corrected by stringent logical proofs. Namely, by a method
that was, up to now, hardly practiced, or was not practiced at all: the
comparison of the function of discourse that generates fiction, to non-
fictional discourse.
The sources of errors from which obviously, directly or indirectly,
confusion slipped into the current debate on the problematic of time
with regard to poetry, have their origin in Lessing's theory of Laocoon
on one hand, and in Goethe's and Schiller's view of epic and dramatic
poetry on the other hand. The source of errors in the theory of Laocoon
is Lessing's assumption that poetry is an art of time. He bases this as-

* Editorial note: Footnotes 1 and 3 were added by the editors. We have decided to
translate the title's German term 'Dichtung' as 'poetry' rather than 'literature' for
reasons detailed in our foreword, footnote 27.
H. Seidler (1955). 'Dichtensche Welt und epische Zeitgestaltung', in: Deutsche
VierteljahrsschriftfurLiteraturwissenschaft und Geistesgeschichte 29: 39CM13.
86 Kate Hamburger

sumption on the view that it is an art of speech and that the action of
discourse, by necessity, takes place within time. For this reason, he ar-
gues, the poet (he refers to the epic poet) can only 'represent' ('ab-
bilderf) such phenomena of reality that take place in time, namely
events or actions. And if the poet wants to describe objects, he can only
do so by conversing them into actions. When Homer wanted to describe
the Shield of Achill, he could only succeed by showing how Hephaes-
tus produced it, thus dissolving it into the actions of Hephaestus.2
Already this example shows how this, Lessing's theory, leads to com-
pletely wrong interpretations, not only of the illustrative art of Homer
but of epic means of representation in general. It is true that Homer dis-
integrates the images of the Shield, like the conflict between the two
cities, into action, event, moving life, but he does not disintegrate them
into the manual actions of Hephaestus. Indeed, the indication, which is
repeated again and again, that Hephaestus produced and created does
mean the contrary, namely, the indication that we are dealing with im-
ages here, something that may be easily forgotten due to the lively de-
But Lessing's theory has been kept in mind and, in modified form,
has become effective again as the discrimination between time of narra-
tion (Erzahlzeit) and narrated time (erzahlte Zeit). But how does that
work? Firstly, the problem will be considered with a simple Homeric
Agamemnon, woken up from the dream that was sent by Zeus,
raises, dresses, and rushes towards the ships:

Jetzo erwacht vom Schlaf, noch umtont von der gottlichen Stimme,
Setzte sich aufrecht hin, und zog das weiche Gewand an,
Sauber und neugewirkt, und warf den Mantel dariiber;
Unter die glanzenden FiiB' auch band er sich stattliche Sohlen;
Hangte sodann um die Schulter das Schwert voll silberner Buckeln;

'Homer namhch malet das Schild mcht als ein fertiges, vollendetes, sondern als ein
werdendes Schild. Er hat also auch hier sich des gepnesenen Kunstgnffes bedient,
das Coexistierende seines Vorwurfs in em Consecutives der Handlung zu
verwandeln und dadurch aus der langweiligen Malerei ernes Korpers das lebendige
Gemalde emer Handlung zu machen. Wir sehen mcht das Schild, sondern den
gottlichen Meister, wie er das Schild verfertiget' (Laokoon, XVIII) - 'Homer, that
is to say, pamts the Shield not as a finished and complete thing, but as a thmg in
process. Here once more he has availed himself of the famous artifice, turning to
the co-existing of his design mto a consecutive, and thereby making of the tedious
pamtmg of a physical object the living picture of an action. We see not the Shield,
but the divme artificer at work upon it' (Laocoon, XVIII).
The Timelessness of Poetry 87

Nahm auch den Konigsstab, den er erbeten, ewiger Dauer,

Wandelte dann zu den Schiffen der erzumschienten Achaier.

He woke from sleep, the god's voice

Eddying around him. He sat upright,
Pulled on a silky shirt, threw on a cloak,
Laced a pair of sandals on his shinning feet,
And hung from his shoulder a silver-worked sword.
And he held his imperishable, ancestral staff
As he walked through the ships of the bronze-kilted Greeks.

This incident is described in verses Iliad II, 41-47. It takes a measur-

able amount of time to read these verses or listen to the recitation. But
does the incident described happen within a certain time? Is it described
as a progression of time? If the incidents are depicted in more detail, if
the clothes, the movements of Agamemnon are specified, then the poet
would have needed to make more verses, and the reader would have
needed more time to read them. But the incident described would not
present itself as a longer incident in time. The time of narration would
be longer but this time of narration, which is needed for the poetic pro-
duction of the incident, and appears again in the auditory or reading re-
ception of the description, is as little associated with the incident de-
scribed as the time that I spend regarding a painting is associated with
the painting itself. A time of narration to which the narrated events can
be related does not exist. I can measure the time that the reading of the
verses requires but I am not able to measure the time that Agamemnon
needs to wake up, to dress, and to rush to the ships. I know nothing
about this time, a fictitious time, because it is not specified and there-
fore, does not 'exist'. But even if it was indicated by some time spe-
cification, this representational indication would have nothing to do
with the real time, which is, for example, represented by the extension
of the verses. Here, a further point has to be considered.
I concede that Lessing was right when he claimed that a real act -
and that includes a discursive act, 'articulated sounds in time'
('artikulierte Tone in der Zeif; Laocoon, XVI) - does take place within
time. In the same way, every real object is, by being in space, also loc-
ated in time, according to the theory of relativity. But relevant to the
theory of poetry is the idealistic doctrine of Kant, which is still valid,
co-existing with the theory of relativity. Kant teaches that time and
space are both categories of perception in which we experience the spa-
tio-temporal world. This doctrine should not be disregarded in the the-
ory of fiction precisely as it has no validity in poetry. What Hegel ex-
88 Kate Hamburger

pressed in his Aesthetics. Lectures on Fine Art (Vorlesungen fiber die

Asthetik) applies to it, to its 'mode of being' ('Seinsweise'): 'Thinking,
however, results in thoughts alone; it evaporates the form of reality into
the form of the pure concept, and even if it grasps and apprehends real
things in their particular sphere into the element of the universal and
ideal wherein alone thinking is at home with itself ('Das Denken aber
hat nur Gedanken zu seinem Resultat; es verfluchtigt die Form der
Realitdt zur Form des reinen Begriffs, und wenn es die wirklichen Bin-
ge in ihrer wesentlichen Besonderheit und ihrem wirklichen Dasein
fafit und erkennt, so erhebt es dennoch auch dies Besondere in das all-
gemeine ideelle Element, in welchem allein das Denken bei sich selber
isf).3 This quotation of Hegel is not only valid for the poetic but also
for the non-poetic, the 'prosaic' reflective-imaginative and thus concep-
tual-linguistic representation of reality. But with special epistemologic-
al persuasiveness it can now be applied to the epic and dramatic, i.e., to
the fictional or mimetic genre. While reality is the 'material' of this
genre, it 'evaporates' into the mode of being 'mimesis', in other words,
a reproductive appearance, but an appearance that does not present it-
self, like in fine art, in representational form (albeit in different materi-
als), but in the form of perception, concept, idea. Here, the forms of
representative reality are not valid, neither in time nor in space. And if
Seidler, as the basis of his proof, says that 'the construction of the fic-
tional world concurs with the construction of reality' (393), because it
is 'in an essential reference to reality' (393), it would be difficult for
him as well as for anybody to go for a walk in a garden that is described
in a novel. In the same way, he might hardly have the impression that a
real garden is by virtue of language - 'it is based in language that real
and fictional world are ever subjected to the same rules of construction'
(395 [our translation]) - a garden 'that has been constructed by men'.
Here the transcendental-phenomenological epistemological theory is
envisioned. But it has been overlooked that only reality, the material
world proper, exposes itself in the form of appearance, of phenomena,
to this subjectively limited perception. This is not the case with an in-
vented reality, no matter to which extent it is shaped and invented ac-
cording to the pattern of reality. The conditions of the perception of
reality are not its conditions; the forms of assertions of reality are not
the forms of reality proper. And Kant's doctrine of time as inner form
of intuition (Anschauung) cannot be applied to the events narrated in a
novel, in the same way, as the doctrine of space as outer intuition can-
not be applied to the space that is described in a novel. For space and
G.W.F. Hegel (1975). Aesthetics. Lectures on Fine Art. Oxford: Clarendon P, 976,
resp. Vorlesungen iiber die Asthetik III (Werke 15). Frankfurt/M: Suhrkamp, 244.
The Timelessness of Poetry 89

time are forms of experience of the existing, but not of the imaginary. A
train accident happening in reality takes, with everything that comes
with it, a certain time. The newspaper report on it 'contracts' this time
by conceptual time specifications of all kind: at five o'clock, after
twenty minutes, after that, etc. When reading this report, we cannot ex-
perience the time during which the accident happened as we cannot ex-
perience the real accident itself, but we only conceptually take cogniz-
ance of the accident, that it happened and took so and so long. The con-
tracting of time, that is, the specifications that denominate time concep-
tually, are not only given in fiction but in every report on reality. As
well in the newspaper report as perhaps in the narrative of a novel, the
train accident is displaced from any time experience because it is dis-
placed from any progress of time.
However, during the last decades the interest of literary theory has
shifted towards the construction (Gestaltung) of time in a novel, and
thus on the problem of time in general. This interest was motivated by
novehstic poetry itself in which time and the experience of time became
topical, with Proust, Joyce, Virginia Woolf, Thomas Mann, etc. By
their treatment of time, these poets made a partly epistemological,
partly constructive 'transcendental' turn, the turn from 'naive' presenta-
tion of events to a 'critical' reflection of the mode of the conditions of
their sequence. This turn is of course associated with the profound ex-
ploration of psychological-existential depths that characterises modern
narrative and dramatic art. These poets attempted to construct time it-
self, to narrate it, like space is poetically constructed and narrated,
either in a more reflexive way (like Thomas Mann's chapters
'Ewigkeitssuppe' ('Soup-Everlasting') and 'Strandspaziergange' ('By
the Ocean of Time') in The Magic Mountain and similar philosophies
of time that are applied to the larger periods of time in the Joseph-nov-
el), or by strongly stressed hints to the course of time in which the nar-
rated events 'happen', or rather, are imagined as happening. In these
works, the time unit of a day became especially popular, most famous
in Joyce's Ulysses, or in Virginia Woolf s Mrs. Dalloway. Here, time is
made recognizable, narrated like the space. But, time cannot be as eas-
ily constructed as space, and it cannot, ultimately, even when extended
into larger spaces of time, be presented differently as in every other his-
torical account, and as in natural changes in life that are, of course, pro-
duced by time. Here, we can see one of the causes of the fact that a
small unit of time like a day was preferably chosen for the poetic con-
struction of time. By this, the infiniteness of the inner world which is
not time-bound could be constructed most strikingly as a contrast to the
limitedness of the outer existence (this is one of the levels of sigmfica-
90 Kate Hamburger

tion of Ulysses, and also of Mrs. Dalloway). But here, it is important to

see what this construction of time means to the narrative technique. Be-
cause, this exactly can be an argument against the absolute value, the
'cultic' value that literary analysis has occasionally given to the prob-
lem of time and the construction of time. The novel can narrate time
but it does not have to. This can be well demonstrated by a comparison
to the construction of space in painting. Painting can construct space
with the aid of perspective. But in no way does it construct space all the
time. A portrait, a still life does not construct space. The novel can con-
struct time as the works named above do. But this is the exception.
Generally, a novel narrates events, situations, human beings with their
reflections and problems without recollecting that 'in reality' events,
life do happen within time. In Mrs. Dalloway, when every hour the
chime of Big Ben resounds, this means that time is being narrated, that
the course of time is intended to remain conscious. If in a novel, there is
a recurrence to the prehistory of a character, then this is another, albeit
less distinct way of constructing time. In indications like: the next day,
one year later, the time had come, etc., the construction of time is even
less distinct, even less dramatic, but given with the narrative course of
events. Therefore, in such cases, overemphasis of the 'structure of time'
may lead to similar misinterpretations, similar confusion of the relevant
with the irrelevant, as Lessing's interpretation of Homer's illustrative
style of narration does. If time is not topical, not constructed, not nar-
rated, then it 'is' not in a work of poetry. It is not there nevertheless,
like in reality - in the same way as space is not in a painting that does
not construct space - even when we can 'forget' about it there, that is,
we are not prepared to experience it there. For in reality, time is the 'in-
ner form' of the course of life, and when we forget about it, perhaps be-
cause we are preoccupied with something else, then we can remember it
again at any time. We can become aware of its course by looking at the
clock, or at the beginning darkness, or at the change of seasons. But if
in our quote of Agamemnon, the time Agamemnon needs to dress him-
self and rush to the ships, is not indicated by some indicator - even if it
is the 'rose fingry Eos' - I would not be able to remember that some
time has to pass with such an action. If time is not indicated by a term
or image, then it is not in the narration. For in poetry, only what is nar-
rated is existent. Only then narrated time exists, when it is explicitly
narrated, like painted space only exists when it is painted. What is true
for the short quotation from Homer is true for the most extensive narra-
tion of life or events. If in such a narration, time has not been included
as a constructed and constructive factor by the poet, the interpreter is
not allowed to use this factor for his interpretation, be it, that he stresses
The Timelessness of Poetry 91

conceptual time indicators that are following the course of time or be it

that he interprets a imagined time into the novel from his side.4 The -
possible - construction of time does not distinguish the novel from oth-
er constructed materials of reality, and therefore, should not be inter-
preted as a special underlying 'transcendental condition' of the narrated
courses of life, just because, in reality, time (and space) do exist. By the
way, it was this problem that led Fielding in his Tom Jones to under-
stand the peculiarity of fictional writing, and from there, he incorpor-
ated his 'new' novel theory into this novel.


The second correlation of fictional genres with natural meanings of

time and tense that provides a source of errors in the analysis of time in
poetry, goes back mainly to Goethe, less so to Schiller. It is basically
connected to the first one: that narrative poetry presents action as past
and therefore in the preterit tense. Seidler's attack is directed against
my indication of this source of error. He accuses me of having
'burdened the creation of a fictitious world on one single form of verb'
(390 [our translation]) and, at another place, expresses it in the way that
was supposed to make 'the preterit present the fiction'. As a matter of
fact, I cannot be accused of having written such a senseless formula-
tion, but I said - and showed from the grammatical 'attitude' of fiction-
al narration - that the preterit loses the original function it had in the
statement of reality, in 'historical narration', to denote something that
has passed. But setting aside this misunderstanding of Seidler, notice
when he conducts his own stylistic interpretation about the space of the
past, etc. in the narration, he himself does not use anything else but just
this single verb form. And why should he not? In the sentence, in the
oral and written communication, it is the finite verb that expresses
something of the 'mode of being' ('Seinsweise') of people and things:
about their being or non-being, their not-yet-being, and their no-longer-
For example, there are many time indicators in Fontane's novels - 'At ten o'clock,
the night tram left from the Station FnedrichstraBe. Even before mne o'clock, one
was there m full travelling equipment [...] In the morning, just after five, the tram
arrived in Schmiedeberg, from there it was just one small hour to Adamsdorf [...]'
Those dates have nothing to do with structure or meaning but have their source m
Fontane's casual, realistic, style of narration that does not distinguish between
essential and non-essential information. The time mdicators should only be used,
on one level with other mdifferent indicators, as a symptom for the non committed
gossipy chattiness of this style. With observations of this kind, characteristics of the
bourgeois-naturalistic style of the epoch can be made accessible.
92 Kate Hamburger

being, that is, their place and time, and therefore their place in reality.
And consequently, it is the preterit that expresses, in a statement in real-
ity, that, for example, a man does not live any more. If I read or learn
from a statement in reality: 'The merchant N. was a rich and generous
man', then I know that now, when I come to know this, he does not live
any more. If I read this sentence in a novel, then I am not informed that
he was a rich and generous man when he was still alive, but that he is
such a man. And if this sentence in a novel is written in present tense, I
am made to know exactly the same: Not that he still lives right now as
this rich and generous man, but that he lived - and behold! The imper-
fect tense was or lived does not have any meaning any more for
whatever I learn from this sentence in the novel. I am not oriented to-
wards the temporal, the preterit meaning of this verb, but towards the
meaning-content signified by it, and I am not realizing the present tense
form of a historical present tense either. The preterit loses its function
and meaning of 'past'. But saying this, I did not indicate that 'the
preterit presents the fiction'. In fact, fiction is presented in a different
way, it is created, albeit with the aid of verbs, yet not with their tempor-
al form, but with their meanings. I will show how this works with an
example especially suited to demonstrate that emotional stylistics is not
sufficient to understand the highly intricate functions used by language
when it does not make statements on reality but wants to construct a po-
etic fictional reality - functions which are subconscious to the poet in
as much as the grammatical forms in which he expresses himself are
subconscious to the ordinary speaker and writer.
The frame narrative of Keller's Zurich Novellas (Zuricher Novellen)
begins with the following sentence:

Gegen das Ende der achtzehnhundert zwanziger Jahre, als die Stadt Zurich
mit weitlaufigen Festwerken umgeben war, erhob sich an einem hellen
Sommermorgen mitten in derselben ein junger Mensch von seinem Lager,
der von den Dienstboten des Hauses bereits Herr Jacques genannt und von
den Hausfreunden einstweilen geihrzt wurde, da er fur das Du sich als zu
groB und fur das Sie sich noch als zu unbetrachtlich darstellte.

Towards the end of the eighteen hundred twenty years, when the city of
Zurich was surrounded by extended fortifications, in the centre of the
same, on a bright summer morning, a young man rose from his bed, who
was already called Mr. Jacques by the servants of the house and addressed
with 'Ihr' [2nd person pi.] by the friends of the house, as he appeared to be
too big for the 'Du' [2nd person sing.] and still too insignificant for the
'Sie' [3rd. person pi.; our translation].
The Timelessness of Poetry 93

Like no other, this text seems to confirm explicitly that a plot of a novel
is imagined as having passed and is therefore told in preterit tense.
When does it happen? At the end of the twentieth year of the nineteenth
century. But let us continue to ask: what happened then? A young man
rose from his bed. If we ask the reverse question: when did the young
man rise from his bed?, then we have to reply: towards the end of the
eighteen hundred and twenty years, on a bright summer morning. Giv-
ing these answers, we realise that they are inadequate. The question:
when did this happen? somehow does not match the verb in relation to
which we ask this temporal question. We do not use verbs like to rise
(from a bed, a chair etc.), to go, to sit, to have a restless night (for he
had had a restless night, as is said immediately after this), when mak-
ing statements about points in time that reach far back in time and are
indefinite. I can say: Yesterday, or one week ago, Peter rode his bike to
town, but I usually do not say: Ten years ago or at the beginning of this
century, he rode his bike to town, or even: He rose from a chair. In a
statement of fact (Wirklichkeitsaussage), I use such situation verbs
(Situationsverben) in imperfect only in relation to points of time which
have passed not long ago. Why? Because these verbs signify a situation
that is concrete and can still be remembered and overlooked by me, the
person who makes the statement here and now. A sentence like the one
by Keller cannot appear in a statement of fact (Wirklichkeitsaussage).
Here, a statement that links a young man rising from his bed to the in-
formation that the city of Zurich - where this happened at the end of the
eighteen hundred twenty years - is surrounded by extended fortifica-
tions, would be impossible. When we read this text without knowing
the context it is taken from, we still know immediately that this is not a
reality statement. The first verb that we come across, he rose from his
bed, teaches us that we are dealing with a fictional narrative. At the
same time, this verb does even more. It eliminates the time marker in its
function as marker of the past tense. It does so although it is given in
imperfect, that is, according to a strict grammar, it indicates an action in
the past, even in a given time. But what does the situation verb really
do? It does the opposite: it makes time and space present, it leads to a
(fictitious) situation here and now in that our young man did not rise
but does rise. And in spite of the imperfect and in spite of the definite
time statement, I really cannot ask when the young man rose from his
bed. I am not supposed to learn that he rose from his bed at a given time
on a day of the 19* century but that he rose. And as it is this factum that
matters, the preterit form becomes insignificant and disappears, it be-
comes de-emphasised and can be substituted by a present tense as well,
94 Kate Hamburger

which likewise would not be emphasised.5 But what happens to the pre-
cise statement of time which was already far in the past of the poet
Keller? It loses its function as a historical statement of the past; it
merely sets the stage for the coming narration which we have entered,
the image of the city of Zurich that was still surrounded by fortifica-
tions at this time. The situation verb eliminates the indication of the
past that both time marker and preterit form have in a reality statement,
and it creates a fictitious present, which immediately constitutes itself
more intense and clearer in all further moments of the narration. Let us
read on:

Herrn Jacques' Morgengemiit war nicht so lachend wie der Himmel, derm
er hatte eine unruhige Nacht zugebracht, voll schwieriger Gedanken und
Zweifeliiber seine Person.

The morning mind of Mr. Jacques was not as laughing as the sky, as he
had spent a restless night, full of difficult thoughts and doubts of his own
person [our translation].

The reader experiences in the same way as the poet who wrote this,
only that Mr. Jacques' morning mind is not laughing - in the fictitious
moment of being of this fictitious character. Therefore, the decisive ele-
ment that creates fiction in this text is the situation verb, which already
has the power to eliminate the indication of the past tense in tenses and
time markers. Situation verbs are always tools of fictionahzation but
they are not decisive for the character of epic fiction, as they also ap-
pear in statements of fact, as for example, in eyewitness accounts
(which have only really taken shape in their precise and typical form in
modern radio reports, like those covering sport events, state funerals,
coronations, but also in travel accounts in which an eyewitness reports
on events in the moment when they are happening, a modern form of
the Mauerschau or teichoscopy in drama). The decisive criteria for nar-
However it may be argued that, when reading a historical work, I am likewise
expecting not the temporal, but the significant meamng of the verbs. When reading
in Ranke's work Die groflen Machte {The Great Powers): 'In this moment (1740)
of an obviously true danger to the German fatherland, Fredenc II appeared, Prussia
rose' [our translation], then the appearance of Fredenc II, the rise of Prussia, would
be decisive. Nevertheless, in my reading expenence, I have the knowledge, what
the verbs express did happen, and accordingly has happened at a certain time, not
only the rise of Frederic but also that he rose at that time. Elsewhere, I will discuss
in detail the decisive meaning of the context, to which we have to turn for the
elucidation of the language system in such cases when the use of words itself does
not enlighten the place of a written work, or even of a sentence.
The Timelessness of Poetry 95

rative fiction, however, are the verbs of inner processes, as I have

shown in detail in my article on the epic preterit.
To say it even more clearly now, the preterit tense form is the sub-
strate in which the narrative can proceed because it is not possible to
narrate without a certain definite verb form. Again, a comparative look
at paintings can be revealing. A painting cannot be painted into the air;
it has to have a base: a wall or a screen. But wall and screen as a base
lose their original material. Not that it dissolves, but in the way that it
has no material value for the painting, for the work of art. The screen is
a screen outside the painting, once it is a painting it does no longer have
any meaning as a screen. This is the way the preterit tense acts in fic-
tion: outside of fiction, it is real preterit tense and establishes a relation
to the past. In fiction, it is the colorless substrate of the referential con-
tents of the verbs which alone is meaningful. Therefore, it can be
treated, grammatically and linguistically, in a way that otherwise would
not be possible: in connection with deictic pronouns, etc., as can be
read in my essay.6 There, I attribute the aesthetic significance of the
a) Here, I would like to respond to another one of Seidler's arguments which, for
some reason, possibly has been taken up in other circumstances. He criticises my
proof that only in fiction, but not in a statement of fact, deictic time adverbs can
appear in connection with the imperfect. Against my quotation of a sentence from a
novel, 'Heute abend wollte der Komg Flote spielen' ('This evening, the king
wanted to play the flute'), he put the example: 'Man kann in einer Gesellschaft
sagen: eigentlich wollte ich (bzw. er) morgen spielen, aber ich tue es doch mcht
(bzw. er ist erkrankt)' ('You can say at a party: originally, I [resp. he] wanted to
play tomorrow, but I do not do it [resp. he has fallen sick] (Fn 37)' [our
translation]. Here, the modal auxiliary verb want (wollen) is misleading. Indeed,
among the auxiliary verbs, only want and in one of its meamngs also the verb shall
(sollen), can apparently appear in a statement of fact in its pretent tense form in
connection with a deictic future tense verb. For this verb, as such, is directed
towards an action that will be conducted in the future and can be substituted by the
expression: to have the intention to do, etc. Its pretent tense has the meaning of a
pre-existing intention which could not or cannot be realised. So the meaning of
Seidler's sentence is: I say today, now, at the party, where I am presently, that
yesterday, I (still) had the intention to play tomorrow - tomorrow from the point of
my today, the day after tomorrow from the point of my yesterday - but already
today I know that I will not do it. This means that, from my present, I look back
into my past of yesterday as well as forward into my future of tomorrow, and I can
only make the grammatical connection wanted tomorrow {wollte morgen) because
of the special meaning of the verb. Any other modal verb will resist this connection
in a statement of fact. I cannot say: tomorrow I could or had to play. But in the
novel, the sentence on Fredenc's flute play can be: tomorrow he had to or he could
play, and had to or could have the same present tense meamng as wanted to. For in
this case wanted to does not mean that somebody knowing something about the
96 Kate Hamburger

epic preterit tense (especially compared to the historical present tense),

and its function generally, to the character effectuality that is attached
to it. I believe that the concept of substrate that is explained above, is
one step further on the way of its semantic elucidation, even more, it is
the final significant elucidation that solves the problem in a satisfactory
way. With this, the moment effectuality has not been turned invalid but
merges into the concept of the substrate.
Consequently, the situation is as follows: the fictional narrative is
not fictional because the preterit tense loses its temporal function, but
the preterit tense loses its function because the narrative is fictional,
meaning that it creates fictitious characters with means that historical
narration cannot make use of, like verbs of inner processes, free indirect
discourse, monologue, or dialogue. As far as I can see, the mundane
fact that dramatic poetry creates fictitious characters has not been in-
cluded into the definitions of epic and dramatic poetry. But it is only

king states that yesterday, or at any other past point in time, the king had the
mtention to play. Rather, this wanted to signifies the fictional momentary inner
situation of the kmg's will, not a past experienced from his side, but his fictitious
present, from where'this coming evening' is imagined.
b) Following this argument, I would like to clear a further misunderstanding of
which, apparently, my own argument is not innocent. Seidler argues that my proofs
of the absence of the I-Origo of a narrator in a novel cannot be correct, as I myself
had shown that this is not true for the desenption of milieus in a novel. Now I
founded my proof, for the purpose of showing the genesis of a fictional narration,
on the entrance of Stifters Hochwald, where, actually, the account of the milieu
does not yet belong to the novel as such (something that I will show in more detail
in my forthcoming book). The present tense in which it is narrated is no histoncal
present tense, but marks a kind of historical eye witness report that only passes
over into fictional narration with the preterit tense. However, with common
introductory descriptions of milieu in pretent tense, like for example at the
beginning of Jiirg Jenatsch, the situation is different. Here, the concept of context
is important, a concept that I had not yet introduced in the essay on the preterit
tense. A mere description of milieu in which the characters of the novel do not yet
appear does not or does not need to contain those structural elements that identify
them as a novel's description. But if I know, for example from the title, that I have
just started to read a novel and no travel account, I already experience an
introductory milieu desenption as the setting of a novel and relate to the fictitious
characters whose appearance I can surely expect because I read a novel. In this
case, the imperfect of this description immediately loses its function as a marker of
the past, and the impression that is given by the first sentence of Jiirg Jenatsch:
'The midday sun shone over the bare peak that was surrounded by rock heads, of
the Mierpass in Biinden county' ('Die Mittagssonne stand iiber der kahlen, von
Felshauptern umragten Hohe des Julierpasses im Lande Biinden') is that of the
temporal present of the expected plot of a novel.
The Timelessness of Poetry 97

this fact that informs us about the linguistic and representational rules
of this, the fictional genre. This fact clarifies that fictional narration is
something categorically different from historical narration which is an
account of reality. But wherein lies the structural root cause of this cat-
egorical difference? It lies in the fact that the contents of epic poetry
(which as a genre restricts us to lead our proof by reference to the nar-
rative) will only exist because they are narrated, whereas the contents
as the object of a statement of fact (Wirklichkeitsaussage) exist whether
or not they are narrated.7 A narrated fiction is not the object of the state-
ment of a 'narrator' or an 'I-narrator', but it is a junction of the process
of narration which in itself is not a 'person', but just a generative func-
tion. When, in our Keller-text, it is said: 'Mr. Jacques' morning mind
was not as laughing as the sky' ('Herrn Jacques' Morgengemut war
nicht so lachend wie der HimmeV), it does not mean that somewhere a
Mr. Jacques exists and somebody relates this about him, but this char-
acter is generated by the words chosen as somebody with a bad mood
in the morning. Narration is the function that generates fiction like the
colored brush of the painter generates the painting. It is the mimetic
function that as the narrative function can be reduced to zero: at this
point, the dramatic form emerges where the narrative function is partly
substituted by the scenic representation, or from the poetic point of
view by the dialogical system. It serves to elucidate and give contour to
the poetic genres if we reserve the concept of the 'narrator' for the au-
thor of a narrative work of poetry, but does not personify the narration
itself to become a narrator. The terms narrator, dramatist, lyricist,
identify the kind of art that is performed by these artists, in the same
way as the terms painter and sculptor identify the kind of their art. This
is not a mere play with terminology but, in the case of epic and lyrical
poetry, leads to insights into the logic and thus to the phenomenology
of the poetic genres. When today particularly effort is taken to avoid
the identification of the 'lyrical I' with the poet, then there is less reas-
on for such qualms then in the case of calling the narration the narrator.8

A more detailed analysis of this statement that will also dissolve the seemingly
paradoxon of unreal contents of a statement of fact, I will elaborate in the future.
In my article that is discussed here (which was already written in 1952 ['Das
epische Praetentum', m: Deutsche VierteljahrsschriftfurLiteraturwissenschaftund
Geistesgeschichte 27, 1953, 329-57]), this was not stressed with full terminological
clearness. The term 'non-fictitious factor' {nichtfiktiver Faktor) that I used there at
one point and which was criticised by Wolfgang Kayser (in his article 'Die
Anfange des modernen Romans rm 18. Jahrhundert und seme heutige Krise', m:
Deutsche Vierteljahrsschrift fiir Literaturwissenschaft und Geistesgeschichte 28,
1954, 417^16), is admittedly not adequate and could lead to misunderstandings, as
98 Kate Hamburger

The difference also proves to be important to the phenomenology of the

Concerning the I-narration, Seidler accuses me with special vehe-
mence of daring to say that this concept avoids the intrinsic task of the
epic poet to create humans. I am ready to withdraw the expression
'avoid' (ausweichen) in this sentence (which appears only as a paren-
thetic relative clause in my work), but not the fact itself. To prove that
this is not incorrect I can refer to a great principal witness, to Aristotle,
who reprimanded whenever an epic poet 'himself spoke'. Because then,
as he remarks shrewdly, he is not an 'imitator' (/ui/uriTrig; Poetics 1460).
Although it is not correct to translate the 'avxov ... Set ... Aeyeiv' of
this quote with 'to speak in first person (= I)', as it sometimes happens,
but by implicitly making the same difference as was stressed above
between the 'epic' that is the narrator, and the 'mimetic' narration, this
also becomes a hint to the special form of the I-narration. Indeed, it is
no autochthonous epic form. It is not 'mimesis' but a - in which way
ever - fictitious statement of fact, where we have to stress that fabric-
ated is not the same as fictitious. And if you look deeper into the struc-
tures of poetry, the sense of an I-narration is not the creation of a ficti-
tious world but a self-representation of the I-narrator himself, whatever
amount of world and man he may integrate into the realm of his self-ac-
The logic of poetry, which has, to those who care about poetry, a
stricter regularity than may be visible on first sight, tells us something
about its aesthetic phenomenology. It is part of the logic of thinking it-
self, just as poetry is part of the general system of language, in which
this logic manifests itself in the form of grammar. The laws that rule
language when it generates poetry can only be discovered when it has
been, in careful observation of its characteristics, distinguished from
those of the general language system. Then, however, it shows that fic-
tional poetry is separated from the general system of expression by an
insurmountable boundary whereas lyrical poetry has its place inside it,
albeit in the most delicate way.
It may be true when Seidler claims that the 'unity of poetry is torn
apart' with this. But where is it written that poetry has to be a unity?

it creates the impression as if the 'narrator' was identified with the author. Already
at that time, the argument was that narration itself is a factor that does not belong to
the fictitious world proper, but at that time, I had not discovered the functional
constitution of this factor that cannot be named a 'fictitious' one - as again, Kayser
would like to do - for the exact reason that itself generates the fiction. In the book I
referred to, I try to give a more detailed analysis of the character of the narrative
The Timelessness of Poetry 99

There is certainly one unity to which it belongs with all its genres and
varieties: it is the unity of art itself which embraces poetry like all the
other arts. It is separated from the mode of being of reality (Seinsweise
der Wirklichkeit) by the fact that what it creates not only exists but also
signifies, meaning that it has the existence of a symbol and with this ne-
cessarily also of an idea. In the system of art, however, poetry takes a
special, even a precarious place. Because it is the very art that exists
through a 'material' that is not only the material of art, of the poetic ex-
pression and creation of reality, but also of the extra-poetic, the 'prosa-
ic'. And with his sharp eye, Hegel saw and said of poetry that it is 'the
special art in which art begins to dissolve itself and merges with the
prose of scientific thinking'. This insight already contains that a logic of
poetry exists because there is a logic of thought.

The Time References of Narration

The most general compositional principle which narrative art initially
shares with every linguistic enunciation is the concept of succession in
which it can be presented and received only. Therefore, the gradual
process of 'becoming' characterises language's work of art (Sprach-
kunstwerk) as a whole, as well as its specific, individual forms in a
much more essential way than it effects the whole and the parts of an
artwork of painting.1 At least since Lessing, but ultimately since Aris-
totle, the principle of 'succession in time' {Laocoon, XVIII, 65; 'Zeit-
folge\ Laokoon, XVIII, 371) has been acknowledged as a fundamental
prerequisite of poetic enunciation. This principle of language-based
presentation had led Lessing to the far-reaching conclusion that the
poet also had to arrange the subject matter of his poetry successively;
indeed, that merely those objects were worthy of poetry that could be
illustrated in a temporal sequence. Such objects, Lessing further con-
cludes, 'are generally termed actions. Consequently, actions form the
proper subject matter of poetry' [our translation; cf Laocoon, XVI,
In his 'First Critical Forest', Herder sharply criticised this view and
rightly accused Lessing of an inappropriate generalisation. Of course,
he also confirms: 'The successive nature of the process is and remains
the crux of the matter' [Herder 2006: 148] ('Das Nacheinander werden
(der Dichtung) ist und bleibt der Knoten'; [Herder 1878:] 148). How-
ever, it would not necessarily follow from the first principle of succes-
sion in language that temporal succession had to be the inherent pnn-
Editorial note: The numbering of footnotes differs from that in the original. Also
note that the bibliography at the end of this translation was generated from the
comprehensive bibliography contained in Lammert's book and augmented where
necessary. - Based on this premise, Medicus [1930] has already drawn considerable
conclusions in regard to the problem of form in his ' Vergleichende Geschichte der
Kunste' ('Comparative history of the arts') (see esp. 194ff, 203, 213), whereas
Walzel [1923] levels these distinctions in different ways.
Editorial note: Lammert does not quote directly from Lessing's chapter XVI but
from the annex (= Laokoon, Anhang III, 434) which is not included in the English
102 EberhardLammert

ciple of poetry's content matter. According to Herder, Lessing's con-

clusions would cause a terrible 'bloodbath' among 'poets both ancient
and modern' (155). - Undoubtedly, Herder partly overplayed Lessing's
fruitful insights with his combative polemics but he also pointed out
the weak link in Lessing's chain of reasoning. In doing so, as will be
shown below, Herder actually harnessed Lessing's theory as one that
pertains to narrative art (Erzahlkunst).
Herder argues that Lessing made his observations by using the ex-
ample of Homer and, based on the realization that Homer represents
'nothing but progressive actions' - rightly so, Herder already considers
this an exaggeration - 'immediately subjoining the main proposition:
Poetry represents nothing but progressive actions' (152). Here, Herder
makes an important distinction first:

Homer composes his poem as a narrative: 'It happened! There was!' In

Homer, then, everything can be action and must rush to action. This is the
direction in which the energy of his Muse strives; wondrous, heart stirring
events are his world; he has it in his power to utter the words of divine cre-
ation:'There was!'(151)

However, Herder positions Pindar side by side with Homer as the cre-
ator of 'large lyrical paintings', and Pindar's word of creation: "I sing!'
Thus, narrative work of art has its energy source in a world of
events (Begebenheiten), which art creates and organises into an action
line (Handlung). The lyric poetry's point of convergence, by contrast,
lies in the soul of the poet and may, indeed must orientate its enunci-
ations toward an emotional condition, without being bound to the suc-
cession of real processes. It may tell of condition, current spirit and
timeless thoughts: it simply has to sing. Lyric poetry is bound to the
first principle like any other enunciation but not any longer to the
Herder already goes as far as questioning the absolute necessity of
progressing in time - and he nearly would also have turned his criti-
cism to the 'basic concepts' ('Grundbegriffe') of the epic and the lyric-
al, which everywhere - and even with Homer - shape literature jointly
by succeeding one another. Anyhow, urging us to let each kind of po-
etry retain its own principle of blending, he states that as far as the epic
poet Homer and his epos is concerned, 'the essence of his poem' lies in
its forceful (!) progression; the successions 'are the body of the epic ac-
tion' (150).
The Time References of Narration 103

Here, one can feel the basic affinity that exists despite of all separa-
tion between the 'basic concepts' and the 'main genres', in this case,
between 'epic' [adject, form] and 'epic poetry'. An epos in general
must possess the basic epic strength of the progression of events in or-
der not to miss its genre. In addition, it may also feature - and it actu-
ally requires this for its artistic forming - lyrical or discursive traits,
even for relatively long periods. The discourse may, as Herder also em-
phasised, expand into the spatial dimension and shape it in the illustrat-
ive fashion of 'painting'. However, the scaffolding (Gerust) of the nar-
rative work has to be the progressing, indeed the energetically pro-
gressing action that is characterised by a striving force!
In one story or the other, the description of milieu, the creation of
human character sketches, the presentation of ideas and, finally, the
educational purpose may overshadow the mere sequence of events and
thus render it meaningful in the first place. But we are, in case we want
to deal with the preconditions and typical features of all narratives, re-
ferred back to the events taking place in time which E.M. Forster called
their atavistic armature.
For this reason, the established formula 'there was' ('es ward'),
which Herder calls the narrator's word of creation, characterises the
peculiarity of poetry more precisely than the formula 'once upon a
time' ('es war einmaV), which before and after Petsch has been con-
sidered as the origin of the epic account. The formula 'once upon a
time' does not yet express any intention of eventfulness; consequently,
it can therefore only be thought of as the primal schema of narrative
exposition, as the entrance gate to the fictitious world that is only nar-
ratively shaped when the static 'it was' evolves into an 'it became' or
'it happened'.
Following Giinther Miiller, 'the primordial form of epic is the "and
then", a formula that sharply illuminates the difference of epic to lyric
and drama' (in this volume p. 71). There was... and then... - if one
adds the two together, one has the ideal outline of the narrated, the pat-
tern of unreeling events triggered by the impetus of a first incident.
So our first conclusion is: it is the poet's task to transform his ideas
and opinions as well as his visions of space and characters into tempor-
al processes, into events, or at least to embed them within these if he
wants to make them narratable (erzdhlbar). In Herders terms this
meant for the epos that every part 'must rush toward the action'; put
into the language of the narrator Jean Paul it reads: 'The entire inner
chain or the chain of reasoning must disguise itself as the flower chain
of time' [our translation]. And Jean Paul assumes: 'This is the most
104 EberhardLammert

difficult task' ('Dies ist das Schwerste'; Jean Paul, § 68, 176, resp.
We have deliberately chosen the findings of Lessing and Herder as
the basis of our deliberations because especially Herder receives far
too little attention for his fundamental achievements in the science of
literature, unlike his contribution in literary history which has charac-
terised German philology to this date. The basic conditions of narrating
were already outlined in Herder's criticism [of Lessing], and, as far as
Homer is concerned, their consequences for the narrative presentation
of the world had already been realised.
However, the insight that a piece of art constituted by language and
characterised by defined movements as well as examinable contours is
not the result of a chance correspondence of the two categories of time,
but rather the result of their continuously changing tension did not
cross Herder's mind yet, although he had already accepted Shaftes-
bury's concept of 'forming forms' and derived his understanding of the
forming powers (Bildekrafte, energeia) of language from it.
The tension between the simulated real events and their narrative
mastery is initially based on the tension between empirical and lan-
guage-based reality, i.e. intentional reality per se.3 Language cannot im-
itate objects and processes but it only may indicate and bring them to
mind to the extent that is necessitated by the relevant purpose of the
message. But by indicating, the speaker presents the whole issue from
his perspective. With his selection from the unlimited ensemble, which
is available to him in real or fictitious form, he constructs a limited
whole, according to the principle that all forming, and especially the
forming by human hand, is an omission.4
If these principles of indication and selection are obligatory in every
kind of linguistic representation of the world then they must also be
equally identifiable in all phenomena of narrative art. This provides us
with a point of departure from which we can swiftly overcome the very
old debate on theories of mimesis, from which even Lessing could not
escape yet. For the double nature of the narrated process and the pro-
cess of narration, better than any other phenomenon of narrating, al-

The indicative mode in the language based work of art is intentional because it
points to a trans-literary entity! About intentionahty in detail: Ingarden Das
literarische Kunshverk {The Literary Artwork) - Stenzel 'Philosophic der Sprache'
('Philosophy of language'), esp. 35ff. - Recently, Hamburger [1953] accentuates
the distinction between real and literary reality with her thorough analysis of the
fictional character of narrative works.
About 'indication' Flemmmg [1925], 9f About 'selection' Lugowski [1932], 185;
more in particular Milller [1944] and in this volume, 67-84.
The Time References of Narration 105

lows us to grasp the principle of indication and selection positively.

Therefore, the comparison of narrated time (erzahlter Zeit) and time of
narration (Erzahlzeit)5 in terms of observation and evaluation, even if
it is just one among many other means, is initially the safest way to de-
scribe the relation between narrated reality and linguistic representa-
tion. Here, the general principles of indication and selection in terms of
contraction (Raffing) and omission (Aussparung) manifest themselves
so clearly that in many instances they even do not evade exact meas-

Using the terms 'ideal time' (action time) and 'real time' (time of speech), Heusler
[1902] already conducted research on 'the dialogue in the older German narrating
poetry' in 1902, and at the same time, Zielmski [1901] analysed the problem of
'Homeric succession' by examining the tension between the course of events and
the presentation of events.
Older than these scholarly approaches, and even older than the analyses of Herder
and Lessing, are the remarks about the specific difficulty of time in narrative art by
the authors themselves, e.g. by Fielding, Sterne, and then Jean Paul. These remarks
are interwoven reflectively into their works Tom Jones, Tristram Shandy and Titan.
As far as we can see, Hut [1923: 10] is the first one who contrasts the terms 'time
of narration' QErzahlzeif) and 'time of plot' QHandlungszeif) but who disregards
their specific tension by concentrating on the unity of time in drama. - In an epoch
that is characterised by a sensitiveness to time, time relations in narrative texts have
become one of the most important aspects of interpretation in the works of Muir
[1928/1949] and E.M. Forster [1927/1947] as well as in a number of special
research papers mostly based on Jean Pomllon's Temps et roman [1946] in which
he analyses the philosophical premises of time phenomena. Important in terms of
criticism: Gaetan Picon [1947]. - With regard to the theory of drama, 'time' has
always been in the focus of interest. An overview can be found in Junghans' work
Zeit im Drama [1931] in which he takes the terms 'expansion of time'
CZeiterstreckung') and 'management of time' ('Zeitbewaltigung') as a basis. But in
the course of his work, he increasingly misjudges his literary subject in favour of
philosophical and psychological theorems despite the excellent individual
considerations. - The structure of Petsch's chapters about time in Wesen und Form
derErzahlkunst [1934/1942] and Wesen und Form des Dramas [1945] is obviously
influenced by this book. Indeed, the tension between time expansion of 'the process
of narration' ('Erzahlvorgang') and the 'narrated event' Qerzahlten Vorgang') has
caught Petsch's attention before ([1930:] 266, see [1934/1942:] 163), but he also
turns to the 'experience of time' (<Zeiterlebnis') and the psychology of 'duration'
QDauef) - something we would like to prevent hereafter. Thiebergers work Der
Begriffder Zeit bei Thomas Mann [1951] shows for example that there are a lot of
interesting, literary historical tasks in this field, in which the intellectual historical
backgrounds are made available in clear reference to Bergson. - 'Die Bedeutung
der Zeit in der Erzahlkunst' ('The significance of Time in the Art of Narration') as
an indicator of structural relations and narrative style has been newly brought to
mind by G. Milller during the last decade. He also invented the two terms 'time of
106 EberhardLammert

urement. Furthermore, this opens up the possibility to look at narrative

works of all times and all languages from a common point.
The tensions between the temporal succession of narrating (Zeit-
folge des Erzahlens) and the temporal succession of narrated events
(Zeitfolge des erzahlten Geschehens) are already obvious from the fact
that the 'history' ('Geschichte') of several generations may be recoun-
ted in a few hours. It would be hard to assume that these contractions
affect each part of the sequence of events to a similar degree: so, the re-
lation of the time of narration to the narrated time has to change con-
tinuously throughout the narrative. The fact that this change not only
affects the extents but also the different modes of narrating is one of
Giinther Muller's fundamental findings that we will be sufficiently oc-
cupied with later.
This pausing, contraction and omitting on the part of the narrator
not only accentuates specific phases in the sequence of events; it is pre-
cisely this activity which helps the entire narrated contents to emanate
from the monotony of pure succession as something newly created. By
the hand of the narrator, the succession of events attains a structure and
the sequence of the whole segments itself into very different elements
of narration - elements that constitute not merely pieces of the whole
but, due to their forceful and directed growing apart, phases in the cre-
ation of it. Moreover, the narrator is at liberty to use the means of split-
ting, reorganizing and suspending the chronology in order to create a
narrative composition that intends to foreground connections that are
defying plain succession.
We want to bring the method of our analysis in line with the way in
which the unbiased reader and observer of an individual narrative work
approaches his object from an aesthetic as well as scientific point of
view. We will start with a rough outlining of the larger contexts and
will then, step by step, descend towards smaller forms and individual-
phenomena. In the course of this we will see that different phenomena,
especially those relating to the formation of phases and the modes of
connection, always recur in an analogue fashion. Nevertheless, they
each find their expression in different forms of narration. It will there-

narration' QErzahlzeif) and 'narrated time' {'erzahlte Zeif), denved from

drscourse and events, as heuristic root terms for an interpretation of narrative
lrterature in terms of rts essential qualrties ([1944, thrs volume, 67-84, 1948, 1948a,
1950, 1951, 1953, 1953a], addrtronal lrterature esp. rn [1950] and [1953]).
Recently, H. Meyer extended the methodrc guidelines by spatial representations
whrch, based on reflections by Th. Mann, also deal with the constellation of rdeas
rn narrative works ([1950] and [1953]). In dorng so, he also focuses on the artifrcral
nature of structural phenomena.
The Time References of Narration 107

fore be possible to clearly distinguish between major and minor forms

of narrating not only with regard to their dimension, but also with re-
gard to their essence.


Flemmmg, Willi (1925). Epik undDramatik. Karlsruhe: Braun.

Forster, Edward M. (1927/81947). Aspects of the Novel. London: Arnold.
Hamburger, Kate (1953). 'Das epische Praetentum,' in: Deutsche Vierteljahrsschrift
furLiteraturwissenschaftundGeistesgeschichte 27: 329-357.
Herder, Johann Gottfried (1878). 'Kntische Walder. Oder Betrachtungen, die Wissen-
schaft und Kunst des Schonen betreffend, nach Maasgabe neuer Schnften. Erstes
Waldchen [1769]', in: Samtliche Werke, Band III. Ed. by Bernhard Suphan. Berlin:
Weidmann, 1-188.
Herder, Johann Gottfried (2006). 'Critical Forests, or Reflections on the Art and Sci-
ence of the Beautiful: First Grove', in: J.G. Herder: Selected Writings on Aesthetics.
Translated and edited by Gregory Moore. Pnnceton, Oxford: Princeton UP, 5 1 -
Heusler, Andreas (1902). 'Der Dialog in der altgermamschen erzahlenden Dichtung'
in: ZeitschriftfiirdeutschesAltertum 46: 189-284.
Hilt, Ernst (1923). Das Formgesetz der epischen, dramatischen und lyrischen Dich-
tung. Leipzig: Teubner.
Ingarden, Roman (1931). Das literarische Kunstwerk. Halle: Niemeyer.
Jean Paul (1935). Vorschule der Aesthetik. Weimar: Bohlau (= Jean Pauls Samtliche
Werke. Erste Abteilung. Elfter Band).
Jean Paul (1973). Horn ofOberon. Jean Paul Richter's School of Aesthetics. Introduc-
tion and translation by Margaret R. Hale. Detroit: Wayne State UP.
Junghans, Ferdinand (1931). Die Zeit im Drama. Berlin: Eisner.
Leasing, Gotthold Ephraim (1935). 'Laokoon, oder Uber die Grenzen der Malerei und
Poesie [1766]', m: G.E. Lessing. Werke. Vollstandige Ausgabe in funfundzwanzig
Teilen. Ed. by Julius Petersen and Waldemar von Olshausen. Berlin: Bong, 275-
Lessing, Gotthold Ephraim (1949). 'Laocoon or: The Limits of Painting and Poetry
[1766]', in: Laocoon, Nathan the Wise, Minna von Barnhelm. New York: Dutton,
Lugowski, Clemens (1932). Die Form der Individuality im Roman. Berlin: Junker &
Medicus, Fritz (1930). 'Das Problem einer vergleichenden Geschichte der Kunste,' in:
E. Ermatinger (ed). Philosophie derLiteraturwissenschaft. Berlin: Junker & Dilnn-
haupt, 188-239.
Meyer, Herman (1950). 'Zum Problem der epischen Integration,' m: Trivium 8: 299-
Meyer, Herman (1953). 'Raum und Zeit in Wilhelm Raabes Erzahlkunst,' in: Deutsche
Vierteljahrsschriftfilr Literaturwissenschaft und Geistesgeschichte 27: 236-267.
108 EberhardLammert

Miiller, Gilnther (1944). Die Gestaltfrage in der Literaturwissenschaft und Goethes

Morphologic Halle: Niemeyer.
Miiller, Gilnther (1948). 'Erzahlzeit und erzahlte Zeit,' in: Festschrift fur Paul Kluck-
hohn undH. Schneider. Tubingen: Mohr, 195-212.
Miiller, Gunther (1948a). Gestaltung - Umgestaltung in Wilhelm Meisters Lehrjahren.
Halle: Niemeyer.
Miiller, Gunther (1950). 'Uber das Zeitgeriist des Erzahlens,' in: Deutsche Viertel-
jahrsschriftfilr Literaturwissenschaft und Zeitgeschichte 24: 1-32.
Miiller, Gunther (1951). 'Goethes Morphologie in rhrer Bedeutung fur die Dichtungs-
kunde', m: Goethe und die Wissenschaft. Frankfurt/M: Klostermann, 23-35.
Miiller, Gunther (1953). 'Le pere Gonot und Silas Marner. Erne vergleichende Aufbau-
studie,' m: Archiv fur das Studium der neueren Sprachen 189: 97-118.
Milller, Gunther (1953a). 'Aufbauformen des Romans,' m: Neophilologus 37: 1-14.
Muir, Edwin (1928/51949). The Structure of the Novel. London: Hogarth.
Petsch, Robert (1930). 'Die Analyse des Dichtwerks', in: E. Ermatinger (ed.). Die Phi-
losophie der Literaturwissenschaft. Berlin: Junker & Dunnhaupt, 240-276.
Petsch, Robert (1934/21942). Wesen und Form derErzahlkunst. Halle: Niemeyer.
Petsch, Robert (1945). Wesen und Form des Dramas. Allgemeine Dramaturgie. Halle:
Picon, Gaetan (1947). 'Erne Philosophie des Romans,' in: Die Umschau 11: 524-531.
Pouillon, Jean (1946). Temps etroman. Paris: Galhmard.
Stenzel, Julius (1934). Philosophie der Sprache. Munchen, Berlin: Oldenbourg (=
Handbuch der Philosophie. 4).
Thieberger,Hans (1952). Der Begriff der Zeit bei Thomas Mann. Baden-Baden: Verlag
fur Kunst und Wissenschaft.
Walzel, Oskar (1923). Gehalt und Gestalt im Kunstwerk des Dichters. Berlin-Babels-
berg: Athenaion.
Zielmski, Thaddaeus (1901). 'Die Behandlung gleichzeitiger Ereigmsse im antiken
Epos,' in: Philologus 8.3, suppl, 405^50.

Time Structure in the Contemporary Novel

1.2 Suggestion of an Enhanced Model of Time Usage Analysis

According to G. Genette
Genette's terminological instruments of analysis of time usage form the
basis of this study. He distinguished between three parts, ordre, duree,
and frequence. We will supplement them with the introduction of cat-
egories that allow further development of the various types and forms:

- within the category of ordre (which we will call 'time

arrangement'), we make the distinction between 'explicit' and 'im-
plicit' anachronies based on the manner of communication to the
reader; further, we consider time phenomena - next to the types of
analepses and prolepses that Genette distinguishes1 - such as the ex-
plicit/implicit permutation of time, the explicit/implicit overlapping
of time, the explicit/implicit interdependence of time, the
explict/implicit synchrony, simultaneity and circularity.
- within the category of duree (duration) we make a further formal
distinction between time summary, ellipsis and expansion of time
that is partly based on Lammert and Ricardou. However, for two
reasons, the phenomenon of duration will only play a limited role in
the analysis: firstly, because duration as a phenomenon only touches
the phenomenon of time arrangement marginally, secondly, because
the analysis of duration is limited to a mere quantificational listing.2
Only when forms of duration are relevant to interpretation, will they
be included in the study.

The same is true for the phenomenon of 'frequency'. The various types
of repetition on the level of language and story constitute a phenomen-
In traditional language use 'flashback' {Ruckwendungen) and 'foreshadowing'
{Vomusdeutungen); on these and further terms see below page 116ff
This would mean a return to the descriptive-quantificational method of Miiller's
Bonn morphological school. On criticism of Miiller's theory see Lammert (51972:
23, 33, 82; Jauss (21970: 15f); Genette (1972: 77f).
110 Alfonso de Toro

on that is, in our opinion, only partly time specific. Repetitions like 'X
eats every day at 12 o'clock' or 'X comes today, X comes today, X
comes today' are either dealing with the story or the language but not
with time usage. These linguistic repetitions or repetitions of story
units may be connected to time usage but they do not have to be and
will, therefore, be dealt with individually. Finally, the category of fre-
quency also includes the 'concretisation of time', in other words, at
what point and in what manner do time indicators or similar data ap-
pear. Here, we will similarly ask about their function.
Finally, to be more concrete and to depict the phenomena of time as
clearly as possible, Genette's model is complemented by the use of
time diagrams. The time diagrams are constructed with consideration
of the story levels D I [Discourse I] and D II [Discourse II].3
Compared to Genette, the problem of the study will also be expan-
ded to include:

a) The question of the function of time usage will be central to our

consideration. It is seen here as an instrument with an extra-textual
and an intra-textual function, e.g. irony or perspectivation, and not
just as a mere sequence of actions in time (there was . . . and then).
From the intra textual point of view, we consider each type of time
organisation as important and as having far-reaching consequences
to the interpretation of the text. Consequently, time usage is seen as
a 'sign', as a 'message' and it can be observed that time usage is a
means to guide the reception of the text.
b) In order to be able to describe the effect of time usage on the reader,
we refer to the concept of the implied reader.4 With this, to us, re-
ception is not the subjective reception by the individual reader but a
procedure textualisation.
c) The procedures of time usage are also analysed in connection with
the 'narrative situation' (Booth 101973), for it can be observed that
in texts with a certain narrative situation often a certain type of time

Following Stierle (1966: 138-147), Todorov (1966: 138-147) and Genette (1972:
75) we differentiate between two discourse levels. The former 'Discourse level T (=
D I) constitutes the 'deep level discourse' equivalent to the rhetorical dispositio, i.e.
the arrangement that includes precedures of temporal structuring. The second
'Discourse level IT (= D II) denotes the 'text of the narrative' corresponding to the
rhetorical elocutio, the process of narrating which subsumes the narrative situation
and modes. The term 'narrative situation' refers to the techniques of point of view,
the term 'modes' refers to ways of narrating such as narration (fr.) or telling and
representation or showing respectively.
See Iser (1972, 1975).
Time Structure in the Contemporary Novel 111

usage occurs. For example, in texts with an 'auctorial' narrative

situation, numerous prolepses, analepses, or a linear-circular usage
of time can be expected, and in texts with a 'personal' narrative
situation, mainly procedures like time permutation or time overlap-
ping can be expected. Of course, also in an auctorial narrative situ-
ation, typical time usages of a personal situation may occur and the
other way round; but this in no way invalidates the observation that
is described here.
d) The paradigmatic and syntagmatic construction principles of artistic
texts have also to be seen in a certain relation to the temporal pro-
cedures in the organisation of the plot (Handlungsorganisation).
Therefore, it can be assumed that texts constructed according to a
paradigmatic principle lean toward radical anachrony - because of
the disentanglement of the syntagmas - but mainly to achrony,
while texts constructed on the basis of syntagmatic procedures of
textualisation lean towards 'chronology', to certain forms of 'ana-
chrony'and to circularity.
1.21 Time Structure as Part of the 'Message':
'Selection'/'Combination' and Time Usage
It was stated above that texts are constructed on the basis of paradig-
matic and syntagmatic procedures of textualisation. Independent of the
emphasis on one or the other procedure, these two principles are trans-
ferred here to the usage of time because they are operational on all tex-
tual levels.
The paradigmatic selective operation aims, at first, at the control,
manner, and emphasis of the given temporal forms of organisation, at
the syntagmatic and the interrelating operation, at the way of temporal
distribution and the temporal combination of story units. Both organ-
isations do not occur by chance but are functional interferences of the
author in order to textualise his message.
The writer acts similar to the composer who has a group of sounds
available from which he makes a selection and combines them in a cer-
tain way. Instead of sounds, he has to select action sequences and to
combine them in accordance to time.
It can be assumed that an author has at his disposal the action se-
quences A, B,C, D, E... n with respectively five action segments, and
he selects and combines the action sequences A to E. Then he has a
multitude of possibilities of temporal combination: he can e.g. first
present the action sequence A with its five segments chronologically
and then the four others, also chronologically, and then also chronolo-
112 Alfonso de Toro

gically the four last ones, or he can connect individual action sequences
with respective segments in an achronological way:

Al,2,3,4,5 Bl,2,3,4,5 Cl,2,3,4,5 etc.


A3 E2 — C5 — B2 D1 B1 etc.

It should be stressed here that the temporal combination/distribution of

segments is a phenomenon of time and does not, initially, have any-
thing to do with the segmentation of the story5 as Propp, Bremond, To-
dorov, or Greimas practice it.6 Actions like: the hero leaves home, the
hero goes through a series of adventures (= overcoming of obstacles),
the hero liberates a princess whom he then marries, and finally he
comes to the throne, or the hero performs an action or not, with success
or not, are no time phenomena, but possibilities of connection of action
elements: in other words, such an analysis moves on the level of the
story and not on the level of Discourse I.
Finally, we would like to remark that the reason for the statement
above (a temporal order always has a textual intention resp. a function)
is also founded in the necessary selection and combination of certain
procedures of time usage. For each selection of a set of n elements im-
plies that they are chosen with a certain purpose. In art, all chosen ele-
ments are meaningful, and this means nothing more than that they have
an intention and a function.7

1.22 Forms of Time Usage

When we introduce analysis instruments for time usage in narrative
texts, we are certainly aware of the methodological problem consisting
in the fact that the concrete text with its manifold specific demands is
always necessarily more complex than an ideal-typical construct.
Therefore, one can and should not expect to find all procedures of time
usage in one specific text, nor will they occur as ideal-typical as they
can be described theoretically. We can, accordingly, understand this set
Lammert, e.g., does not always clearly distinguish between procedures of temporal
arrangement and those of the syntagmatic, non-temporal connection established
according to the logic of action. See the criticism of Lammert by Janik (1973: 10)
Bremond (1964, 1966, 1970, 1973); Greimas (1966, 1967); Propp (1972).
Lotman (1973: 35); the point is here that so-called formal elements (also time
usage) are being semantisised, that is, they contribute to the construction of
meaning in a text: here see Jakobson (1973: 219-233); Muller (21974: 276).
Time Structure in the Contemporary Novel 113

of instruments as a preliminary heuristic construct that aims to facilit-

ate a precise and substantiated description of phenomena of time in
narrative poetic texts.8
In an analytical model of time usage, a distinction between external
and internal time first has to be made.9
We will not deal with external time here, as it is not the object of
our study. It can be defined as the time outside the text, as the empiric-
al, historical time of the author, of the real reader, and as the time dur-
ing which the text was written.10
The internal time is the time within the text that is constituted
through 'act time' (Aktzeit) and 'text time' (Textzeit)}1

See here Miiller (21974: 300): 'When we attempted below, by example of some
prominent cases, to clarify the experience of time and the construction of time
within the work that is arranged by the poet (...), then this had to be done in a
strongly simplifying, schematising way that did not do justice to the often mutual
requirements in an individual work, and that was also in a certain way true for the
arbitrary choice of the "cases'" [our translation]. No matter how voluminous a
certain work is designed, this basic problem that is inherent to all models, will
persist; see, in a different context, also Pfister (1977: 15f).
S. Ducrot & Todorov (1972: 400).
The extra-textual reference of the time to which we refer here should not be
confused either with the extra-textual function of time, that is, with its effect on the
reader, or with the internal/external analepses/prolepses. Our term 'external time'
also partly corresponds to the one of Hristo Todorov (1968: 41^19); he understands
by it the real time of the communication partners, that is, a time that is located
outside the text. His definition of temps interne (as temps simule) corresponds also
only partly to ours, insofar as he does not make the distinction between real and
fictional'act time'.
We adopt the term 'text time' {Textzeit) - in slightly altered form - from Weinrich
(21971: 56), the term 'act time' (Aktzeit) from Wunderlich (1970: 31). Ricardou
(1967: 161-170) speaks of temps de la narration and de la fiction in the sense of
'Textzeit and 'Aktzeit in our language use. Rossum-Guyon (1970: 215-227) uses
the term temps de I 'ecriture in the sense of temps de la narration, as does Ducrot &
Todorov (1972: 400), however, she substitutes this term with the one of temps du
sujet, and defines temps de la narration (= de I'ecriture) as 'reading time'
(Lektiire); see also Ducrot & Todorov (1972: 400), whose term of temps externe is
congruent to our term of external time, but whose term temps interne only complies
partly with our term of internal time, because the authors also include, next to
temps de I 'histoire (or temps de la fiction, temps raconte, temps represent*?) and the
temps de Vecriture (or de la narration ou racontant) the temps de la lecture. Apart
from the fact that the time of reading {Lektiire) would belong, taxonomically
(should such a distribution be possible), as well to the internal as to the external
time, which is why we regard this category as not suitable for text analysis because
of its variability and non-verifiability (each reader reads differently), Todorov
114 Alfonso de Toro

'Act time' is the multidimensional, chronological time of the rep-

resented events, that belongs to the level of the story and can be meas-
ured in minutes, hours, days, weeks, months and years. It has three dif-
ferent temporal dimensions: the dimension of the present, of the past,
and of the future. With 'act time', we have to distinguish between real
and fictional time.12 The real 'act time' is the time of the historical ac-
count or of newspaper reports, a time with an external time reference.
As is well known, a historical text is pragmatically linked and pos-
sesses a rough isomorphic link between its configuration, its course of
action and time and its description on one side, and real processes that
can be described scientifically, on the other side.13 Its link to pragmatics
sets out with the chaining of the literary production to the standard cal-
culation of time or chronometry. Consequently, in a historical text, the
temporal difference is determined by the relation between the proced-
ure of writing and the temporal situation of the written. Accordingly,
the hie et nunc-deixis refers to the empirical chronometric time and to
the historically definable space. We can define the real act time as fol-

'Real act time' = a time that is pragmatically linked to empiric-

al historical time.

Fictional act time is the time of poetic texts: of the novel, the short
story, the drama etc. Different from real time, fictional time has to be
determined within the fiction, the poetic text, and is part of the creation
of the situation. Fictional time does not know the link to pragmatics. It
is neither characterised by the chaining of the literary production to
empirical real time, nor by a rough isomorphic link between its config-
uration, its course of action and time and its description on one side,
(1966) and Ducrot & Todorov (1972) do not make a distinction between real and
fictional 'Aktzeif.
See also Mendilow (1952: 65) who speaks of fictional time (= fictional act time), as
well as Ricardou (1967: 161-170) who speaks of temps de la fiction. Rossum-
Guyon (1970: 215-227) uses the terms temps narre, de Faction, de Vaventure and
also de la fiction, and Genette (1972: 77) the terms of temps de la chose-racontee,
du signifie; Miiller (21974: 247ff.) uses the term of 'narrated time' (erzahlte Zeit)
that is only partly congruent to our 'act time': Kayser (151971: 207ff) uses the term
'objective time' (objektive Zeit) in the sense of 'time of narration' (Erzahlzeit) by
Miiller (21974: 247ff), and like him, understands it as 'time of representation'
{Darstellungszeit; see here Pfister (1977: 327-381), but he warns us not to believe
that the 'objective' time could become congruent with 'poetic' time (= narrated
time) in narrative texts (see also Miiller 21974: 258; 307f).
See Link (1974: 286f).
Time Structure in the Contemporary Novel 115

and real processes that can be described scientifically, on the other

The hie et nunc-deixis only refers to itself, that is, to the time that is
constituted immanently in the poetic text, and to the immanently con-
stituted space:

'Fictional act time' = a self-referential time that is immanently con-

stituted within the poetic text and not prag-
matically linked.

Consequently, in the case of poetic texts one has to speak of a fictional

present, past and future.
Text time, as Genette explicitly states, is a pseudo-temps, as the dis-
course does not have a real time at all.15 With this term, we refer to the
position in which a certain event appears on D I. As is well known, the
signifier of the text is linear but the events may be differently organised
and arranged in time. For example, a text may begin with the chronolo-
gical end of the story and end with the chronological beginning. The
term 'text time' is certainly not satisfying but we could find no better.
If the term is chosen here, then only to avoid any equalsation between
it and Miiller's term 'time of narration' (Erzahlzeit) which derives from
a completely different issue.16 Text time can be defined as follows:

'Text time' = the position, in which a story segment P of a

story sequence P appears on discourse level I.

See (ibid. 293-297).
See Genette (1972: 77f).
On Miiller's theory see (21974: 225-246; 247-268; 299-314; 388^118; 556-570;
571-590). Again, it should be stressed that our term 'text time' (Textzeit) is not
congruent with Miiller's 'time of narration' (Erzahlzeit), and that the instruments of
analysis serve to describe the procedures of temporal arrangements and not to
quantify and measure them. By 'time of narration' (Erzahlzeit), Miiller understands
the extension of the text (pages and lines that are needed for a certain extension of
time), although this term is also defined as the time of reading or the time of the
play. With the term 'narrated time' (erzahlte Zeit), he refers to the extension of a
narrated story in minutes, hours etc. The issue of the Miiller school in relation to
the usage of time results from the definition of these pairs of terms: Miiller is
concerned about the confrontation of the extension of text and the extension of
time, a phenomenon that, in our model, will be placed, following Genette, in the
field of duration.
116 Alfonso de Toro

1.221 Time Arrangement17

The analysis of the temporal sequence of events on the level of the
story and its disposition on D I is restricted to the description of the re-
lation between text time and act time that can be discordant or concord-
The time phenomenon of chronology (= 'temporally organised
presentation of a story') exists if the temporal sequence of the events
on the level of the story and their disposition on the D I tend to concur.
Chronology is possible in groups of texts which are narrative or per-
forming. In these texts, the concordance may only be seen as an ap-
proximate value, not as an exact symmetry.19 In Western narrative tradi-
tion, the use of concordance is less common than the use of discord-
ance.20 The time phenomenon of discordance, which we would like to
name with Genette's term of 'anachrony', can be found wherever a
temporal sequence of events on the level of the story, and its disposi-
tion on the D I, does not concur. In its traditional form of analepses and
prolepses with a supplementing function, discordance between text
time and act time can be traced back to Homer.21 Nevertheless, the phe-
nomenon of anachrony remains worth analysing because of its further
development in which it returns with new forms and respective new
1.2211 Chronology

The attempt to produce a chronological sequence of segments of ac-

tions enables one to demonstrate at which level the temporally organ-
ised sequence has been invalidated, and which temporal transforma-
tions were necessary resp. which have been performed in order to aban-
don the chronology. In the removal of the chronology of a sequence of
actions, the fictional character of poetic texts is most clearly presented:

Within time arrangements two functions can be distinguished: the intra text
function with far reaching consequences for the constitution of the story, and the
extra text function, which facilitates a certain guidance of reception, and transmits
the message by the author, without the use of the omnipotent narrator.
See Genette (1972: 77ff.) who uses the term ordre.
Ibid., 79.
Ibid., 79ff; complex forms of time arrangement can already be found in
HeUodorus'E/fao/Hca, see Kayser (151971: 210) and Nolting-Hauff (1974: 440ff).
See Genette (ibid); the difference in the use of anachronies in the ancient, older,
and modern literature lies in the gain of complexity of time usage, in the functional
changes, and in the inclusion of the levels of consciousness in time usage.
Time Structure in the Contemporary Novel 117

'Chronological sequences of actions' = chronometric, that is, well-

organised linear disposition of the
segments of actions of one or sever-
al sequences of actions within time
resp. the tendency of concurrence
between text time [hereafter: TT]
and act time [hereafter: AT].

The following scheme is supposed to clarify the tendency of concur-

rence between text time and act time. The hyphens ' - ' signify the tem-
poral progression from one segment of action to the other - no matter if
chronological or not. The letters represent the various segments of ac-
tion of a sequence of action that could look as follows: A = the hero
leaves the house - B = the hero liberates the princess - C = wedding -
D = accession to the throne. On the level of the act time, the numbers
(right) function as indication of time, as they give the position of the
segments of action within the chronology (level of story). With this
simple example, the letters suffice as indication of time but this is not
the case with more complex structures of the story, as we will see be-
low. With text time, the exponent numbers indicate the position in
which the action units occur on D I. This somewhat laborious indica-
tion has the significant advantage that it represents the temporal struc-
ture of the arranged story in the way it actually appears in the text.
While with the AT we always represent the course of the action se-
quence in a chronologically reconstructed form (which obviously is es-
pecially true for a-chronological sequences of action), the arrangement
on the level of the TT actually shows how the AT is arranged and how
the reader actually reads the story. In this example, AT and TT are con-
gruent. The characteristics of D II are always given when they are rel-
evant to the usage of time, [...].

Story level :AT:A1-B2-C3-D4


DI : TT: A 2 1 - B 2 - C3 3 - D4 4

DII: Narrator's account

dialogue etc.
118 Alfonso de Toro

1.2212 Anachrony
Genette uses the term 'anachrony' in view of analepses and prolepses.
In regard to the contemporary Latin American novel, however, it
proves to be useful to distinguish sub-types of anachrony.
1.22121 Explicit Anachrony
Explicit anachrony can be subdivided into five types: explicit 'time
permutation', explicit 'time overlap', explicit 'time interweaving', into
'time circularity' and explicit synchrony. Where anachrony is explicit,
analepses (flashbacks) and prolepses (flash-forwards) constitute two
time levels in the text:22 a level of time I (= TL I), the present into
which the anachrony has been inserted, and a level of time II (TL II) -
subordinated to TL I -, which is created by the time of anachrony itself
and which is constituted by a past or future, by a deeper past or deeper
future. Therefore, to distinguish these two time levels, we will speak of
TL Hi, and of TL II2. A good example to illustrate these two time levels
can be found in Flaubert's Madame Bovary.* Arrival and daily life of
Charles and Emma in Tostes constitute TL I; the analepsis, in which
Emma's life in the convent school is shown constitutes TL II.
By distinguishing these two time levels, Genette is led to another
distinction: the distinction between 'extension' (amplitude) and 'scope'
(portee).1* By 'extension', he understands the time section which is
covered by anachrony (analepsis or prolepsis) in TL II. 'Scope' is the
term for the temporal distance between the events contained in ana-
chrony (= TL I), and those events that happen in the present time (= TL
II). In other words: 'scope' is the temporal distance between events in
past or future and those in the present time:

'Temporal extension' = temporal reach of a story segment p of a

time TL II which is inserted in a time TL

'Temporal scope'= temporal distance between a story seg-

ment p of a time TL II and a story segment

See Genette (1972: 78-90); Lammert (51972: 100-194) distinguishes between two
main types of discordance: 'flashback' {Ruckwendung; = analepsis) and 'fore-
shadowing' {Vorausdeutung; = prolepsis).
Flaubert (1971: Chapt. vi, 36^1).
See Genette (1972: 89f).
Time Structure in the Contemporary Novel 119

In the example above, Emma's memories of her arrival at the convent

school, her release and return home constitute the analepsis. The exten-
sion of this anachrony stretches over several years in the convent
school up to Emma's return home; the 'scope' of the analepsis, the
temporal distance between Emma leading a married life with Charles in
Tostes and the Emma in the convent school adds up to several years
(there is no information in the novel about the number of years in
between).25 Explicit anachrony can be defined as follows:

'Explicit anachrony' = Temporal permutation, overlapping, inter-

ference and synchronisation of story seg-
ments and/or story sequences as well as
the circular organisation of story se-
quences that occur due to the influence of
an intra textual instance of communication
(of a narrator or a character) and lead to
the constitution of two (or more) time
levels that relate to each other, where TL
II is always subordinated to TL I.

In this definition, the status of the different time levels is also ex-
pressed because analepses and prolepses belonging to TL II are always
inserted from TL I. It also happens that analepses constitute themselves
within other analepses and/or within prolepses; also, prolepses within
other prolepses and/or prolepses within analepses may appear; this in-
terlacing of prolepses and analepses, we call anaprolepses.26
The difference between explicit and implicit anachrony can be seen
among other in the fact that the reader does not necessarily have to re -
construct a story sequence that has been intercepted by an explicit ana-
chrony, because the omnipotent narrator resp. the character as a guar-
antor for the temporal order remains openly present. With implicit ana-
chronies, this is not the case.

1.221211 Explicit Time Permutations

We subsume Genette's categories of analepse and prolepse under the
term 'explicit time permutation':

Indication of time is not necessary, as Emma Bovary herself knows when the events
that she recalls took place. Here, the mentioning of time indicators would reveal
Emma's perspective and the perspective of the narrator.
This term is also used by Dallenbach (1977: 76, note 1).
120 Alfonso de Toro

'Explicit time permutation' = analeptic or proleptic rearrangement

of story segments and/or story se-
quences leading to the constitution
of two (or more) interdependent time
levels, where TL II is always subor-
dinated to TL I.
1.2212111 Analepses

'Analepsis' = temporal transfer by a narrator or a character of a

story segment p of a story sequence P from the past
to the pre sent leading to the constitution of two
time levels, where TL II is always subordinated to

According to Genette, analepses can be subdivided into three main


a) Internal analepses are those whose complete extension remains

within TL I, in other words, it reaches back only shortly after the
beginning of the text:
aa) Within internal analepses, Genette distinguishes further
between heterodiegetic analepses that remain within TL I but
whose story segment has a different content the one of TL I,28
and between
ab) homodiegetic analepses, whose story segment has the same
content as the story sequence of TL I. Further, with Genette,
two types of homodiegetic analepses can be distinguished:
aba) The completing (completives) analepses with a completing
function. That means they subsequently close a temporal gap
that was produced by an 'ellipsis'. In this gap, temporally un-
See Genette (1972: 90ff.); Ducrot & Todorov (1972: 401) use, instead of temporal
permutation, the term inversion without distinguishing between an explicit and an
implicit one. Lammert (51972: 100) calls the analepses 'Ruckwendungen' ('flash-
backs'). Internal and external analepses/prolepses should not be equated with
internal resp. external time, for both types of analepses and prolepses belong to the
internal time, in addition to the fact that the distinguishing criterion is different.
Likewise, these analepses and prolepses should not be confused with the internal
resp. external usage of time although both types of analepses and prolepses have
both options to function.
Lammert (51972: 112) uses the term 'Riickschritt' (literally translated, a step
Time Structure in the Contemporary Novel 121

clear or only insinuated events can be placed. Besides the el-

lipsis, there are other gaps like e.g. the 'paralipsis'. Here, the
narrator does not, like with an ellipsis, skip a story segment
but leaves certain important story moments in the dark which
are then occasionally mentioned later. These omissions do not
refer to the chronological sequence of events and rather relate
to the content.
abb) The second type of the homodiegetic analepsis is the re-as-
suming resp. repetitive (repetitive) type that we would like to
call with Lammert an analepsis that is connected to the
present (TL I); it appears again and again in order to narrate a
section of the past in an additional or comparative way.29 This
analepsis confronts two events with each other that can mutu-
ally interpret one another. From this, it may be possible for
past events to be interpreted in a new way, to receive a new
meaning, or even to have a given meaning be taken away. A
special form of the reassuming analepsis is the 'enigma' (en-
igme).30 The narrator hints at something that will later on re-
ceive a meaning. These analeptic hints are very popular in de-
tective stories and can either confuse the implied reader or ac-
tually help him to follow the entangled events. The enigma is
not temporally fixed; it is a creator of suspense and challenges
the implied reader to combine certain events or moments with
each other.

With Genette, we would like to distinguish between 'explicit' and 'im-

plicit' enigmas: enigmas of the first type can be found when the narrat-
or relates whatever is hinted at in the enigma by a retrospect to the hin-
ted object, place, or indicated person. Here, the enigma is solved by the
narrator himself. Nevertheless, the implied reader must make the effort
to relate the past and the present. With the implicit enigma, the narrator
does not give a signal that enables the reader to find a connection
between the enigma and the indicated object, place, or character. In this
case, the implied reader has to reverse gears and look back into the past
himself in order to solve the enigma:

b) External analepses are characterised by the fact that their total ex-
tension - contrary to internal analepses - remains outside the tem-
poral level I, that is, before the beginning of the text. Out of external

Lammert (ibid., 122) speaks of 'Ruckgriff (flashback) in this case.
Genette (1972: 97).
122 Alfonso de Toro

analepses, episodes may emerge which have no relation to the story

sequence, or which are directly connected to them.31

Two types of external analepses can be distinguished:

ba) the incomplete (partielles) external analepses that deal with

an isolated past event and whose extent reaches back to a
point before the beginning of the text.
bb) the complete (completes) external analepses whose extent
reaches back to the beginning of the text itself.

c) The mixed analepses are defined by Genette as those whose scope

lies outside the temporal level I, but whose extent stretches to a
point before the beginning of the text. In a stricter sense, mixed
analepses are mainly those whose scope is equal to their extent.

1.2212112 Prolepses
In contrast to analepses, prolepses occur less often in narrative texts,
and from the second half of the 19th century onwards, a clear decrease
can be observed, after Flaubert had postulated the impassibilite of the
narrator as a desirable aim.32 However, even in the 20* century, pro-
lepses can be found in several texts by authors like Proust, Thomas
Mann, Garcia Marquez etc.

'Prolepsis' = temporal transfer of a story segment p of a story se-

quence P from the future to the present, leading to
the constitution of two time levels, where TL II is
always subordinated to TL I.

Also, according to Genette, prolepses can be subdivided into three

main types: internal, external, and mixed.33

a) Internal prolepses are those whose total extent remains within TL I,

that is, reach until shortly before the end of the text. Here - like
with the 'analepses' - a distinction can be made between

See Genette (1972: 90f.). Lammert (51972: 112ff.) counts the internal and external
analepses under the term 'Ruckschritf (step backward).
See Flaubert's Correspondence in G. Bolleme (1963: 95; 9 decembre 1852).
See Genette (1972: 105ff.); Lammert (51972: 143ff.) uses here the term
'Vorausdeutung' (foreshadowing).
Time Structure in the Contemporary Novel 123

aa) heterodiegetic prolepses that remain within the time level TL

I but whose story segment have a different content from the
story sequence TL I, and
ab) homodiegetic prolepses whose story segments have the same
content as the story sequence of TL I. With the homodiegetic
prolepses, Genette distinguishes further between
aba) completing (completives) prolepses that fill in advance future
temporal gaps that are produced by ellipses,34 and
abb) Pre-emptive resp. anticipative (repetitives) prolepses which
can be regarded as those that explicitly announce future events
that will later on be extensively narrated. The formulas of
such announcements (annoncesT read e.g. 'as we will see
later', 'you will see, that...', 'many years later' etc. Next to
the explicit pre-announcements, there are 'indicators'
(amorces);'6 that is, insignificant structures that have no pro-
leptic, but an allusive function. Often, the reader only recog-
nises their meaning a posteriori, after the second or third read-
ing of the text.

b) The external prolepses are characterised by the fact that their exten-
sion remains outside the temporal level I, that is, they reach further
than the end of the text. All those prolepses that have their place
after the chronological end of the story, in other words, when the
hero has died or escaped from the world of events, can be under-
stood as external.

As with the analepses, a distinction is made between

ba) partial (partielles) prolepses containing an isolated event that

lies in the future and whose extent stretches beyond the end of
the text, and
bb) complete (completes) prolepses whose extent reaches to the
end of the text.37

34 See Genette (1972: 105ff.), Lammert (51972: 171ff.) uses the term 'erganzende
parallele Vorausdeutung' (an augmenting parallel foreshadowing for this type.
35 Here Lammert speaks of 'Phaser? (phases) and 'Ausgangsvorausdeutungen' (initial
36 Genette (1972: 112).
37 Lammert (51972: 154-159) names these prolepses as ' Vorausdeutungen der
Endsituation' (foreshadowings of the final situation), and the complete prolepses as
' Vorausdeutung des Endzustands' foreshadowing of the final state).
124 Alfonso de Toro

c) Mixed prolepses can only be defined for heuristic reasons, that is, as
prolepses whose extent lies before the end of the text but whose
scope goes beyond the end of the text.

Finally, it is important to point out that analepses as well as prolepses

have to be differentiated from a communication theory resp. a perspect-
ive related point of view. Mainly in texts with an auctorial narrative
situation, it can be observed that information, which is given in retro-
spect from the point of view of the narrator, will constitute for the read-
er an information given in advance if he or she did not previously hear
anything about it. When a narrator begins a novel with the sentence
'On his deathbed, X will remember his childhood', then the first part of
the sentence is clearly a prolepsis, the second part is an analepsis. For
the reader, however, both have a proleptic value. Furthermore, in this
case, the prolepsis has to be attributed to the narrator, the analepsis at
least indirectly to the character. Further, a distinction is made between
mentioned and executed analepses/prolepses. For example, the narrator
announces: 'X did not know that in Rome the police waited for him'. If
this sentence appears again and again without further information, then
we are dealing with a mentioned prolepsis; but if the narrator accounts
why the police is waiting for him in Rome and how he is arrested, then
we are dealing no longer with an anticipated information, but with an
anticipated story segment, consequently, with an executed prolepsis.

1.221212 Explicit Time Overlap and Explicit Time Interweaving

With a massive use of analepses and/or prolepses, an explicit time
overlap and an explicit time interweaving occurs.
The case of explicit time overlap exists when a character creates a
second continuous level of time - perhaps through memory - next to
his/her present immediate action (= TL I). This second time level can
either belong to the past (TL II) or it may exist within the mind (in the
form of imagined situations) and then be timeless. Examples can be
found in Proust's A la recherche, Joyce's Ulysses, Robbe-Grillet's Le
Voyeur or Fuentes' La muerte de Artemio Cruz.
Time overlaps can thus catch up with events but this does not neces -
sarily have to be the case, as they can have a new content. In the
second case, they nevertheless belong to the explicit anachronies, as
they represent temporal distortions by way of a communicative in-
stance. They can also be simple or complex, depending if they form
two or more TLs, or if two or more story sequences are overlapped.
Time interweaving exists if a narrator presents various story seg-
ments from various story sequences in chronological and/or a-chrono-
Time Structure in the Contemporary Novel 125

logical order, so that temporal confusion may happen38. Examples can

be already found in the tradition of the heroic Gallant Novel (as a fur-
ther development of the Hellenistic Novel), and in newer novels like
Radcliffe, The Italian, or Sue, Les Mysteres de Paris. However, in
these texts, the story sequence is clear because of the leadership of an
auctorial narrator and does not reach the complexity that characterises
the modern novel.

1.221213 Time Circularity

With the procedure of circularity, we have to make a distinction
between 'temporal circularity' (on the level of discourse I), and 'story
circularity' (on the level of the story). Time circularity happens when
analepses and prolepses are used similarly. From a point of time X
within the fiction, the narrator mentions an event A in the future from
this time, then an event B in the past, and then, from there, narrates lin-
early, until he reaches A again and gives a detailed account of A. In a
graph, this looks as follows:39

The circles may be simple or complex. A simple circle is e.g. the one
mentioned above; but if it contains more circles that have formed them-
selves during the narration of past events or during the passage from V.
to Z., then you have a complex circle:

Our term corresponds only partly to the one that Todorov uses (Ducrot & Todorov
1972: 402), of histoires enchassees, as he assumes a chronological order of events,
but this is only one possibility of the interdependence of temporal segments.
Accordingly, circularity is a special form of anaprolepses; see Vargas Llosa (1971:
545ff); Segre (1973: 152-193).
126 Alfonso de Toro

However, when similarly structured situations that existed at the begin-

ning of a novel and seem to be overcome in the course of the novel,
nevertheless re-emerge at the end, then you have story circularity:

Situation: A -> Situation: B -> Situation: A (resp. A')

An example is Flaubert's novel Bouvard etPecuchet, where the heroes

begin as copistes and end as copistes.

1.221214 Explicit Synchrony

Events that are arranged in a parallel chronological order by a narrator
or a character may be called synchronic, according to the formula
'while x happens in A, y happens in B and q in C etc. Synchronic
story segments or story sequences may occasionally show an equal or a
similar structure of story and semantics but do not necessarily have to.

1.22122 Implicit Anachronies

Like the explicit anachronies, implicit anachronies are temporal re-

arrangements but differ because of their lack of an internal communica-
tion instance. They also comprise four types: the implicit time permuta-
tion, the implicit time interweaving, the implicit time overlap, and the
implicit synchrony. Additionally, there is simultaneity that derives
from this procedure.

'Implicit anachrony' = temporal permutation, interweaving, over-

lap and synchronising of story segments
and/or story sequences that do not emerge
because of the influence of an intra-textu-
al, but of an extra-textual communication
instance and lead to the constitution of
two (or more) mutually independent time

With this type of anachrony, the reader does not immediately realise
the temporal distortion but is surprised by it, as it is caused neither by
the narrator nor by a character, but by the author who is not deictically
1.221221 Implicit Time Permutation and Implicit Time Interweaving
The 'implicit time permutation' is the achronological, surprising dislo-
cation of a story segment p forming part of a story sequence P from a
Time Structure in the Contemporary Novel 127

temporal position tx to a position ty within the same story sequence P

without an internal communication instance. The 'implicit time inter-
weaving' is a chronological or a-chronological, surprising dislocation
of a story segment p forming part of a story sequence P from a position
t x to a position ty e.g. Q, without an internal communication instance.
Of course, such dislocations also have analeptic or proleptic character
in regard to the respective main story sequence - in case such a main
story sequence can still be identified. But they have to be distinguished
from the analepses and prolepses proper, for the latter are, according to
our definition, always dependant on the point of time of the dislocation
and are introduced through the intervention of an internal textual com-
munication instance (narrator or character). In addition, permutations
and interweavings are not omissions but dislocations and therefore
should not be confused with ellipses (see below).
An example for implicit time permutation: we begin with a story se-
quence consisting of four story segments A, B, C, D which appear in
the following temporal order:

Story level AT: A1 - B2 - C3 - D4

DI TT: B!2 - C23 - D34 - A41

Up to now we have, in case of a single story sequence, used one re-

spective letter to mark the story segments. In our above example (p.
139), the departure from the house by the hero, A, the liberation of the
princess, B, the wedding, C, and the accession to the throne, D, were
named. Here the letters received an exponential digit as a mark of their
course on the DI and a digit to the right to indicate the chronological
progression. Now, this indication proves to be insufficient if we have
several interrelated story sequences. Therefore, it is necessary to keep
one letter for the whole story sequence but with an additional digit in
order to distinguish the various story segments. The story sequence
mentioned above should now be indicated as follows jA - 2A - 3A - 4A
instead of A - B - C - D. In order to describe the phenomenon of time
interweaving, another story sequence can be imagined: a second hero
also courts the princess; departure from the house jB; on his way, he is
met by several obstacles: obstacle 1: Fight with monsters: 2B; obstacle
2: Fight with the villain: 3B and finally belated arrival and disappoint-
ment: 4B.
As case of a simple time interweaving, in other words, as one in
which the story segments progress chronologically, these two story se-
quences can be presented graphically as follows:
128 Alfonso de Toro

Story level AT: jAl - jB2 - 2A3 - 2B4 - 3A5 - 3B6 - 4A7 - 4B8

DI TT: jA11 - jB22 - 2A33 - 2B44 - 3A55 - 3B66 - 4A77


More complex time interweavings may appear when the story segments
proceed achronologically, as in the following example:

Story level AT: jAl - jB2 - 2A3 - 2B4 - 3A5 - 3B6 - 4A7 - 4B8

DI TT: 2A!3 - 2B24 - 3B36 - jA4l - 4B58 - tf6! - ,A77

- 2A83

Here, the story segments A and B are not only temporally interwoven
but the respective story segments of A and B are, at the same time, per-
m i t t e d temporally.
As we have seen in the graphs, the exponential digits of the story-
segments resp. -steps were retained (e.g. TT: iA4 - iB22; resp. TT:
2A!3 - 2B24. This was necessary because the steps do not follow each
other causally (B does not follow A). In this case, the distinguishing
exponential digits have to be retained. Otherwise, the segments tempor-
ally succeeding each other in an arbitrary way would be declared as
causal. In our example above, steps A and B are of the same nature (in
both cases a respective hero leaves the house). It would only be pos-
sible to do without the action-related digits if in a text all story seg-
ments followed each other according to the logic of action as then these
digits would be congruent with those of the chronology.
1.221222 Implicit Time Overlap
Implicit time overlap represents a third type of implicit anachrony. It
consists of the overlap of two (or more) time levels, without an internal
textual communication instance. Within the individual story sequence,
or within several story sequences, we distinguish between simple and
complex time overlap. A simple time overlap exists when two different
story sequences can be found on respective time levels, e.g. one in the
present and the other in the past or in the future. A complex time over-
lap exists when at least three different story segments/story sequences
take place on respectively different time levels, one e.g. in the present,
the second in the past, the third in a more remote past or in the future
or in a more remote future. We can transfer this to a graph as follows:
simple time overlap within a given story sequence A happening in the
Time Structure in the Contemporary Novel 129

present (TL I); story segments also occur (p,q,r), which could be situ-
ated in the past or the future:

With two story sequences A and B:

A complex time overlap within a given story sequence A that exists in

the present, TL I; the story segments p, q, r occur that belong to the
past, TL n b and the segments s, t, u belonging to the future, TL II2:

With three story sequences A, B, C

130 Alfonso de Toro

1.221223 Implicit Synchrony

Parallel events running chronologically and communicated by the in-
fluence of an extra-textual communication instance can be called impli-
cit synchrony. At times, they may show an identical or similar semantic
structure, but not necessarily. As with other time phenomena, you can
assume here that synchronies do not happen incidentally but are con-
sciously intended and have a certain function, as can already be seen in
the case of Flaubert.40 Synchrony does not have anything in common
with the use of analepses and prolepses, as in the first case, when both
story segments (and -sequences) are time independent, but in the
second case, they are not.41
1.221224 Simultaneity
Because of the linearity of the discourse mentioned above, there is no
real simultaneity in literary and especially in poetic texts. In narration,
simultaneity is simulated with the aid of procedures of various kinds.
At least three procedures can be listed:

a) Simultaneity can be achieved by procedures of time usage (belong-

ing to D I), here by implicit anachronies. The a-chronological use of
the AT of several story sequences resp. the incorporation of several
independent time levels give an impression to the reader that
everything narrated is happening at the same time.42
b) Simultaneity may also be achieved when a story is presented
(through procedures belonging to D II) by the use of several narrat-
ors or personal media whose perspectives and speech appear closely
connected and can often hardly be separated. This procedure is be-

In Flaubert's L'Education sentimentale (1964: 145ff; 280-284) and also in
Bouvard et Pecuchet (1965: 227-333). All these synchronies set the events into
contrasting relations and substitute an auctorial commentary. Here, the implicit
reader is asked to draw his conclusions; see de Toro (1987: 9-31) and (1987a: 121-
In opposition to Lammert (21972: 102) who regards analepses ('Ruckwendungen')
as synchronisations.
Ducrot & Todorov (1972: 403) define simultaneity as dedoublement que le temps
de Vecriture projette dans sa succession. Often the use of analepses and prolepses
is referred to as simultaneity. But this is wrong, as the precondition for a situation
of simultaneity - belonging to the implicit anachronies - is the lack of an intra-
textual communication instance. As soon as such a communication instance exists,
two interdependent TL are created, and the different TLs always remain obvious to
the reader.
Time Structure in the Contemporary Novel 131

ing used especially in descriptions and within the framework of

stream of consciousness.
c) Finally, a specific typographical segmentation can contribute to
simultaneity, so that story segments of one or several story se-
quences are torn apart so that in every typographical segment, dif-
ferent story segments or parts thereof appear.43

Although the procedures b) and c) to create simultaneity are not of a

temporal nature, they serve, as a rule, to strengthen the impression of
1.2213 Achrony
Genette calls a story segment sans date et sans age an achrony. We
distinguish between weak and strong achrony:

a) Weak achrony always exists when story segments can be roughly

temporally determined by their content (characters, space, etc.)
b) Strong achrony, however, exists when the temporal succession can-
not be reconstructed at all.

As we will see below, strong achrony is common within texts that have
a tendency to be sujetless. It will be seen that in such texts, story seg-
ments may partly be placed into a certain temporal line on the basis of
assumptions, but partly they may not be, as there is not only a lack of
time specification, but the chronology is destroyed to a degree that tem-
poral organisation no longer has meaning.
1.222 Duration
Up to now, the relation between text time and act time was discussed
under the aspect of anachrony, limited to an analysis of the relation
between text time and act time. With the analysis of 'duration', it is not
the relation between act time and text time but the relation between act
time and the length of the text (= 'LoT') that is being analysed, i.e., the
temporal longitude or brevity, the duration of the story segments. We
do not attribute any principal meaning to 'LoT', it only serves as an
empirical starting point for a comparison of the duration between the
story segments.

See Pfister (1977: 122ff.), who does not only consider the procedures of time usage
for the creation of an illusion of simultaneity, but also the use of channels of
various kinds.
Genette (1972: 119).
132 Alfonso de Toro

Brevity and longitude do not provide absolute criteria: the duration

of a story segment p is short or long only in comparison with a story
segment q or r. In the same way as time constitutes itself in the text, the
criterion of duration constitutes itself (immanently) within the text.45
The relation act time/text extension can be an-isochronal, that is it is
deviant (a story segment has an act time of one day, for which time e.g.
a text length of 100 pages is needed) or isochronal, that is, there is a
tendency of congruence,46 a consistent rhythm between act time and
text length; there is always the same or a similar text length dedicated
to story segments with similar time extents. This is the case, e.g., with
dialogues. Here, the question for purpose and effect of 'anisochrony'
and 'isochrony' becomes important.
Every narrative text is based on a counter proportionality of time: a
text can, on the level of the story, show great ellipses and/or very great
pauses. To give an exaggerated example: in a text, the whole life of a
character can be concentrated within one line, or an event that took
only one hour, can be represented over a thousand pages.
Following Genette, we distinguish three types of an-isochronies:
'Pause', 'time summary', 'ellipsis', and a type of isochrony: 'scenic
presentation' or 'time congruence'.47
1.2221 Pause
Pause or 'time expansion' (pause descriptive) exists when the LoT is
much larger than the AZ.
The 'pause' can be found wherever the story comes to a total stop,
i.e., where the narrative flow is interrupted by interventions of an auct-
orial narrator who reports from a great distance to the narrated, as e.g.
in a description or a commentary, or when catching up with narrated
events. In this context, it is important to state that not every pause
means a description, and not every description does mean a pause.

This part of the model constitutes - as has been mentioned before - the core subject
of Muller's theory.
See Genette (ibid: 1972: 122f); Ricardou (1967: 164ff); Weinrich (21971: 57) and
Genette (1972: 128ff); Lammert (51972: 84ff); Ducrot & Todorov (1972: 402f).
Todorov uses the terms analyse and digression (for our term of pause), but they
should be understood rather as a gradual distinction than as a necessary
differentiation for the phenomenon itself. With Todorov, the difference between the
two terms is not very clear. He calls time summaries resumes and consequently
does not make a difference between different types of summaries. Ellipses, he calls
escamotages, and congruence of time, style direct.
Time Structure in the Contemporary Novel 133

When a description is given by a narrator, the narrator interrupts the

story (= static description); but when a description is given by a char-
acter who is simultaneously an actor, or by a narrator who uses a per-
sonal medium (= dynamic description), then this is not the case. The
story sequence is only insignificantly affected in its flow.
Finally, 'pauses' can be created by the static speeches of a character
like monologues or stream of consciousness,™ as these forms of expres-
sion do not interrupt the flow of the story less than digressions of the
1.2222 Scenic Presentations or Time Coverage49
Genette speaks of 'scenic presentation' (scene) or 'time coverage'
when referring to the generally isochronic relation between 'LoT' and
'AT' that may occur in narratorial discourse or character discourse, on
the condition that the discourses do not interrupt or slow down the
events in any way. Texts dominated by scenic presentations come close
to the dramatic form, as the mediating instance is strongly reduced. The
reduction of the mediating inner system of communication, which ap-
pears mainly in texts with a 'personal' narrative situation, creates the
illusion in the reader of experiencing the events directly and immedi-
ately: here, the narrator wants to create a real, objective and almost em-
pirical reflection of reality.
But scenic presentation may also - different from Genette's theory
- lean towards anisochrony, if, in dialogical parts, the 'LoT' becomes
much longer than the 'AT' without interference from the narrator and
without the appearance of any ellipsis; the isochrony is disturbed, when
the discourse of a character does not lead to action but interferes with
the course of events. As mentioned, an extreme example is the mono-
1.2223 Time Summary
'Time summary' (Zeitraffung; sommaire) can be defined as the prepon-
derance of the 'duration' of the 'AT' over the 'LoT'.
Time summary can come close to 'time congruent narration' or, in
extreme cases, lead to a complete omission or ellipsis. According to the

On this term Humphrey (81972). Groundbreaking were the novels by V. Woolf,
The Waves (1931) and by J. Joyce, Ulysses (1922).
Lammert (51972: 84); Stanzel (61972: 43ff.) speak of 'szenische Gestaltung' and
Todorov (1966: 146) according to Lubbock (31960) of 'style panoramique' resp.
'style scenique'.
134 Alfonso de Toro

type of time summary, a distinction can be made between 'successive'

and 'iterative-durative' time summary50:

a) The successive time summary is a line-up of events proceeding in

the same direction as the act time. The linguistic basic formula of
this time summary is: 'then...and then...' Within the successive
summary, a distinction can be made, according to the intensity of
the summarisation, between:
aa) 'Leap summary', characterised by a narrator who narrates hur-
riedly in big steps in the style of vera vidi vici; here another
distinction has to be made between
aaa) a simple leap summary like: 'Many years later, X came back'
aab) a proleptic leap summary like: 'Many years later, X would
find himself in a different position.' Here, the summary points
to the future. Furthermore, leap sum mary can border on the
ab) 'step summary', which is continuous and comes close to time
congruent narration. Here, also two additional types of step
summary have to be distinguished:
aba) pure step summary like 'on the first day,...on the second
day..., a few day later...'and
abb) mixed step summary, like: 'on the first day, several days later,
after two months' where small leap summaries occur.

b) The iterative-durative time summary condenses a more or less large

period of time through indication of individual, regularly repeated
events (iterative) or general conditions which last throughout the
whole period of time (durative). Often, these two forms appear
closely intertwined so that they can be presented together here.51

Their basic linguistic formula is: 'In this time it once happened...' or
'So it happened for example....', 'Again and again in this time...' or
'Through all this time...'etc. Summaries always appear when the nar-
rator has to present certain events or indications of a character in order
to elucidate the presently running story. These events and indications
do not represent the focus of his interest; they provide secondary and
50 See Petsch (1978: 47); Milller (21974: 259); Lammert (51972: 83f.); we only speak
of 'time summary' to separate it from other types summaries like 'spatial' and
'topical', which, however, is difficult, like in the case of iterative-durative time
summaries; see Lammert (51972: 85f).
51 Lammert (ibid: 83f).
Time Structure in the Contemporary Novel 135

additional information, subordinated to his/her interest. Especially in

the case of a leap summary, the distance of the narrator is very large, as
he reports with a previous knowledge of characters and events, which
is shared neither by the actors nor by the reader. With summaries, the
connecting communication system becomes evident.

1.2224 Ellipsis
When the 'LoT' is 'much smaller' than the AT, we speak of an ellipsis.
We distinguish three types of ellipses:

a) Ellipses on the level of the story, that is, ellipses that leave out a
certain time span of the story. We distinguish here between three
forms of omissions52:
aa) the explicit 'ellipses'. These are explicitly mentioned by the
narrator. The ellipsis can be presented in a definite manner
('Two years have passed') or an indefinite way ('Many years
have passed), or may at first not be indicated by the narrator
but only be marked at the beginning of a new chapter;
ab) the implicit ellipses. In this case, the omitted time is not indic-
ated. The reader can only a posteriori, after attentive reading,
realise that there is a temporal gap in the diachrony;
ac) the hypothetical ellipses. These cannot be determined within
the diachrony. Thus, they are 'timeless' ellipses. Sometimes,
the reader can help himself with elements within the content,
like characters, places, motives etc. to define their temporal

A classical example of the ellipsis in general can be found in Flaubert's

L 'Education sentimentale, in chapter III between part V and part VI
(Frederic returns from Nogent to Paris and then takes to travel; 418-419):

Un hurlement d'horreur s'eleva

de la foule. L'agent fit un cercle
autour de lui avec son regard; et
Frederic, beant, reconnut Senecal.

A yell of horror arose from the crowd. The police-officer,

with a look of command, made a circle around him; and Fre-
derick, gazing up at him in open-mouthed astonishment, re-
cognised Senecal.

See also Genette (1972: 139-141).
136 Alfonso de Toro


II [Frederic] voyagea
II connut la melancolie des paquebots ...
II revint.
II frequenta le monde ... etc.53

He [Frederick] travelled.
He knew the melancholy of the steamboat...
He came back.
He frequented society ....

Here, the summarisations // voyage, il connut la melancolie des paque-

bots... border on ellipses as they bridge large temporal gaps.

b) Ellipses on a typographic level are those like the transitions between

chapters, parts or sub-parts, or an empty page like in Robbe-Grillet' -
s Le Voyeur.5" Here the narrative flow is interrupted.
c) Finally, Ricardou distinguishes ellipses on the D II which constitute
an interruption of the narrative flow but not of the story sequence,
as is the case in Ricardou's L 'Observatoire de Cannes.55

The three 'an-isochronies', pause, time summary and ellipsis, and the
isochrony, the scenic presentation, can be presented as follows56:

Flaubert (1964: Chapt. vi, 419).
Ricardou (1961).
See also Ricardou (1967: 161ff.).
Time Structure in the Contemporary Novel 137

1.223 Frequency
1.2231 Repetition on 'D IF and Frequency of the Story
In this chapter, following Genette, the realisation of D II and the story
will be discussed under the point of view of frequency.57
Actions like 'x ate every day at 12 o'clock' as well as sentences like
'X ate every day at 12 o'clock; x ate every day at 12 o'clock' etc. may
be repeated; the frequency is a characteristic of D II, of the language
and the story, a quantifiable ratio. For this reason, the text frequency
can be defined as the frequency of D II (= RD [repetitions of dis-
course]) and of the story (= RS [repetitions of story]).5*
However, as explained above, the phenomenon of frequency will
not be regarded in detail. For our purposes, the phenomenon of fre-
quency as such is important, as are its functions and its relation to time
usage, but not its different forms.
According to Lotman, RD II and RS have two basic functions in

a) to reveal differences or highlight certain elements within the same

or similar elements,
b) to lower the semantic relevance of the repeated element and reveal
the principles of arrangement within the same or similar elements.59

It has to be determined from case to case which occurs when. This de-
pends on the number of RD II and/or RS, i.e., whether the repetitions
occur at a high or low frequency. In addition, one has to take into ac-
count whether these repetitions are of an identical or equivalent nature.
The repetition of equivalent linguistic elements or action segments
may contribute to represent the same object from different perspect-
ives, wherby its semantic or structural meaning changes with each re-
petition. For this reason, repetitions receive a structural significance
and they consequently can be included with semantics so that they are
not considered as mere, so-called formal elements.60

Genette (1972: 145-182).
One may also consider such structures as forms of enonciation if analysed
See Lotman (1973: 139; 187-212).
To my knowledge, Jakobson (1973: 219-233) is one of the first authors who stated
the inclusion of so-called formal elements as part of the content. With this, he
acknowledges the semantisation of form and overcomes the traditional separation
of content and form.
138 Alfonso de Toro

The repetition is associated with the time arrangement in different

ways, e.g. with analepses and prolepses. The narrator or a character
may continually repeat an analepsis in order to evoke specific, preced-
ing events for the agents (i.e. intra-textual) or for the reader (i.e. ex-
tra-textual), or in order to compare preceding and similar or equal actu-
al events. Furthermore, every repetition of the actual event may be of
analeptic (or maybe proleptic) character by evoking preceding similar
events or predicting them. The repetitions that function as intensifiers
for the recurrence of similar and equivalent elements are closely related
to the circularity of action and time: to the circularity of action in the
respect that an equivalent or similar condition reoccurs, even if it is
caused by different agents. The circularity of time is concerned insofar
as the narrator starts from the point of time ti and returns to ti.
1.2232 Selective vs. Non-selective Concretisation of Time
One last aspect of time usage is the concretisation of the passing of
time by way of time indicators.
The presence or absence of dates may be relevant or irrelevant, de-
pending on whether they are employed with a communicational value
or not. Time passing by may manifest itself in a selective or non-select-
ive way.
Selective concretisation of time is defined here as an exact, almost
chronometric, temporal fixation of an event; non-selective time con-
cretisation is its vague, metaphorical positioning.

1.22321 Forms and Functions of Selective Concretisation of Time

Some forms of selective time are:

After three weeks, the first month, for two nights, she is only fifteen
years old, it is six o'clock.

The functions of these different forms cannot be stated in general, but

again they also have to be analysed and extracted from each text, and
within the text, from each context. They may serve these functions:

a) Function of alienation: there are parts in texts with contradicting

dates which confuses the implied reader.
b) Negating function: in this case, dates do not have the function to
provide the reader with temporal orientation but to negate that time
elapses. This is achieved by mentioning an exact date in a certain
passage that does not contain any other dates indicating that time
Time Structure in the Contemporary Novel 139

passes by, without the narrator having called the reader's attention
to the time's passing by in any other way.
c) Summarising function: the selective concretisation of time is one of
the possibilities that allows to shorten the presentation of events.
d) Function of temporal orientation.
e) Elliptic function: the selective concretsation of time is generally
used if a narrator withholds an entire part of the story.
f) Relativising function: The point in time at which certain events took
place is often put into question by supplying large amounts of time
1.22322 Forms and Functions of Non-selective Concretisation of Time
There are two different forms of non-selective concretisation of time;
the implicit and the explicit form:

a) Within the non-selective, implicit concretisation of time, the

elapsing of time may be expressed by the following forms:

- by day: 'the midday sun'; by night: 'The moon was covered by

- by describing the physiognomy of one character: 'X's hair was
beautiful, now she is grey';
- by the change of townscape, technological advance and clothing
of the characters: 'We only used to have horse-drawn carriages
here, now we go by underground';
- finally, the elapsing of time may be expressed by the way of life
of one of the characters: 'Once, X was very poor, now he has
become very rich.'
b) Non-selective, explicit concretisation of time manifests itself in the
form of time indicators that express the elapsing of time in hours,
days, months and years in a very vague and undefined way. Some of
these expressions are: 'many years have passed', 'several weeks'

The non-selective implicit concretisation of time is mainly used in an

iterative type of narration. One of its functions is to produce the illu-
sion of simultaneity. However, its main function is to lift the boundar-
ies between the fictional time (internal time) and the time of the reader
(external time). The reader is taken into the fiction by a non-selective,
implicit concretisation of time and he is distracted from his own time-
bound existence.
» auctorial
• narrative situation *—•personal

A •» report
modes jg:_—~~— •
free indirect discourse

4 explicit permutation of

ypographic level

' explicit overlapping of t

explicit anachrony
"""""* explicit interdependenc
" circularity of time
anachrony "explicit synchrony

r implicit permutation of ti
* implicit anachrony v~" —•implicit overlapping of t
• time arrangement ^
* implicit interdependence
, strong "implicit synchrony
••weak "simultaneity

»Pause • leap summa

>chrony - —•time summary *• step summar
• def
'*• ind
n story level - • implicit
• time usage * hypothetical
"ellipsis n DII and on story level

>chrony c presentation/time congrv.

»• D II
• frequency o
*• story

r characters
\ r implicit
concretization of time <."_._ •• explicit
tory level - • plot
Time Structure in the Contemporary Novel 141

Bolleme, G., ed. (1963). Extraits de la correspondence ou preface a la vie d'ecrivain.
Paris: Seuil.
Booth, W.C. (101973). The Rhetoric of Fiction. Chicago: U of Chicago P.
Bremond, C. (1964). 'Le message narratif,' in: Communications No. 4, 4-32.
Bremond, C. (1966). 'La logique des possibles narratifs,' in: Communications No. 8,
Bremond, C. (1970). 'Morphology of the French Folktale,' in: Semiotica 2.3: 247-276.
Bremond, C. (1973). Logique du recit. Paris: Seuil.
Dallenbach, L. (1977). La recit speculate. Essais sur la mise en abyme. Paris: Seuil.
de Toro, A. (1987). 'Flaubert precurseur du roman moderne ou la releve du systeme
romanesque balzacie: Le pere Goriot et L'education sentimentale,' in: A. de Toro
(ed.). Gustave Flaubert. Procedes narratifs et epistemologiques. Tubingen: Gunter
Narr, 9-31.
de Toro, A. (1987a). 'Bouvard et Pecuchet: description du niveau de l'histoire,' in: A.
de Toro (ed.). Gustave Flaubert. Procedes narratifs et epistemologiques. Tubingen:
Gunter Narr, 121-149.
Ducrot, O. & Todorov, T. (1972). 'Temps et modalite dans la langue/Temps du dis-
cours,' in: O.D. & T.T. Dictionaire encyclopedique des sciences du langage. Paris:
Flaubert, G. (1971). Madame Bovary (1856). Paris: Gamier Freres.
Flaubert, G. (1964). L Education sentimentale (1869). Paris: Gamier: Freres.
Genette, G. (1972). Figures III. Paris: Seuil.
Greimas, A.J. (1966). Semantique structurale. Recherche deMethode. Paris: Larousse.
Greimas, A.J. (1967). 'La structure des actants du recit (essai d'approche generative),'
in: Word 23: 221-238.
Humphrey, R. (81972). Stream of consciousness in the modern novel. : University of
California P.
Iser, W. (1972). Der implizite Leser. Kommunikationsformen des Romans von Bunyan
bis Beckett. Miinchen: Fink.
Iser, W. (1975). 'Die Appellstruktur der Texte,' in: R. Warning (ed.). Rezeptionsasthet-
ik. Theorie undPraxis. Miinchen: Fink, 228-252.
Jakobson, R. (1973). 'Poesie de la grammaire et grammaire de la poesie,' in: R.J. Ques-
tions dePoetique. Paris: Seuil, 219-233.
Janik, D. (1973). Die Kommunikationsstruktur des Erzahlwerks. Ein semiologisches
Modell. Bebenhausen: Rotsch.
Joyce, J. (1932). Ulysses (1922). Hamburg, Paris, Bologna: Odyssey Press.
Jauss, H.R. (21970). Zeit und Erinnerung in Marcel Prousts 'A la recherche du temps
perdu'. Ein Beitragzur Theorie des Romans. Heidelberg: Universitats-Verlag.
Kayser, W. (151971). Das sprachliche Kunstwerk Ein Einfuhrung in die Literaturwis-
senschaft. Bern: Francke.
Lammert, E. (51972). Bauformen des Erzahlens. Stuttgart: Metzler.
Link, J. (1974). Literaturwissenschaftliche Grundbegriffe. Eine programmierte Ein-
fuhrung auf strukturalistischer Basis. Miinchen: Fink.
Lotman, J.M. (1973). Die Strukturdes kunstlerischen Textes. Frankfurt/M.: Suhrkamp.
142 Alfonso de Toro

Lubbock, P. (31960). The Craft of Fiction. New York:

Mendilow, A.A. (1952). Time and the Novel. London, New York: Viking P.
Miiller, G. (21974). 'Die Bedeutung der Zeit in der Erzahlung (1946/47),' in: G.M.
Morphologische Poetik. Gesammelte Aufsatze. Darmstadt: WBG, 247-246.
Nolting-Hauff, I. (1974). 'Marchenromane mit leidendem Helden. Zur Beziehung zwis-
chen einfacher Form und narrativer GroBform in der Literatur,' in: Poetica 6: 417-
Petsch, R. (1978). 'Zeit in der Erzahlung,' in: A. Ritter (ed.). Zeitgestaltung in der
Erzahlkunst. Darmstadt: WBG, 32-50.
Pfister, M. (1977). Das Drama. Theorie und Analyse. Mtinchen: Fink.
Propp, W. (1972). Morphologie desMarchens. Mtinchen: Hanser.
Ricardou, J. (1961). L'observatoire de Cannes. Paris: Seuil.
Ricardou, J. (1967). 'Temps de la narration, temps de la fiction,' in: J. R. Problems du
nouveau roman. Paris: Seuil, 161-170.
Robbe-Grillet, A. (1955). Le Voyeur. Paris: Minuit.
Segre, C. (1973). 'Curved Time in Garcia Marquez', in: C.S. Semiotics and literary
criticism. The Hague: Mouton, 152-193.
Segre, C. (1973). 'Curved time in Garcia Marquez,' in: C.S. Semiotic and literary criti-
cism. The Hague, Paris: Mouton, 152-193.
Stanzel, F.K. (61972). Typische Formen des Romans. Gottingen: Vandenhoeck &
Todorov, H. (1968). 'Logique et temps narratif,' in: Social Science Information 7.6:
Todorov, T. (1966). 'Les categories du recit litteraire,' in: Communications No. 8, 125-
Vargas Llosa, M. (1971). Historia secreta de una novela. Barcelona: Cuadernos mar-
ginales.Weinrich, H. (21971). Tempus. Besprochene und erzahlte Welt. Stuttgart:
Woolf, V. (21979). The Waves (1931). London: Atriad Panther Book.
Wunderlich, D. (1970). Tempus und Zeitreferenz. Mtinchen: Hueber.

Story-time and Fact-sequence-time*

1. Time Levels in Grammar and Narrative Theory

Modern narrative theory, in terms of its theory of time, distinguishes

between two time levels, discourse-time (Erzahlzeit, i.e., the time of
narration) and story-time {erzahlte Zeit, i.e., the narrated time).1 Dis-

* Editorial note: The literal translation for the German term ' Sachverhaltszeif is
'time of a state of fact'; the corresponding term central to this article is 'Sach-
verhaltsfolgenzeit', i.e. 'time of a sequence of states of fact'. In order to avoid these
unwieldy literal translations, we have decided to use the term 'fact' for 'state of
fact' (note that 'Sachverhalf in Harweg's usage refers to static situations as well as
to events), resulting in the terminological equivalents 'fact-time' for
'Sachverhaltszeif and 'fact-sequence-time' for 'Sachverhaltsfolgezeif. While the
distinction between an objective 'fact' and a 'state of fact' - i.e. a descnption or
conceptualisation of a fact, not the fact itself - is thereby partially obliterated, the
term 'fact' seems adequate in that Harweg's mam argument concerns the distinction
between representation of time in the form of 'story' and the (logically) pre-
supposed factual time that exists independently from narrative verbalisation. The
philosophical problem inherent to the notion of 'fact' is implicitly discussed in
Harweg's chapter 2.2.
The termini [i.e., discourse-time /story-time = Erzahlzeit / erzahlte Zeit] go back to
G. Milller (1947/1968, [transl. in this volume], and 1948/1968). However,
according to E. Lammert (1955: 257), the levels as such have been distinguished
earlier in literary studies by Th. Zielmski (1901), A. Heusler (1902) and E. Hirt
(1923). Lammert (1955) himself, carrying on from Milller, takes this opposition as
the focus point of his study Bauformen des Erzahlens {Forms of Narration). In the
non-German language area, B. Tomasevskij (1925/1965: 281) distinguishes
between 'temps de la narration' and 'temps de la fable', and T. Todorov (1966:
139) draws the line between 'temps du discours' and 'temps de Vhistoire\ S.
Chatman (1978/1983: 62) speaks of 'discourse-time' and 'story-time'. A.A.
Mendilow (1952: 36ff) speaks of 'pseudo-chronological or fictional time' instead
of 'story-time', and instead of 'time of the act of narrating', he speaks of
chronological time while making a distinction here between 'the reader's clock
time' and 'the writer's clock time'. Doing this, he stresses that he does not believe
that the 'time of the act of narration' or the chronological time are fictional. Or, like
me, he sees a difference between Activity as the attributes of certain sigmfier and
144 Roland Harweg

course-time is the time covering the action of narration, and story-time

is the time of what is being narrated within the framework of this ac-
This duality of time levels in narrative theory corresponds to the du-
ality of time levels in traditional grammar. In the models of tempus, a
distinction is made between time of 'discourse' and time of 'acting' or,
as I would prefer to say: between 'utterance-time' and 'fact-time'
CSachverhaltszeif). It is not surprising that in both disciplines, in lin-
guistics as well as in literary studies, two - or better: only two - levels
of time were viewed as basic, and in the main still are considered to be
so today, where the assumption appears to be widely accepted in liter-
ary studies. This duality is, after all, nothing else but a reflection of the
semiotic dichotomy of sigmfier and signified and thus a reflection of a
widespread semiotic opposition.
However, just as one occasionally realised the need to expand and
refine the dichotomist sign model in semiotic theory, various scholars
of tempus- and time levels felt a similar need to expand and refine the
dichotomy of the 'time of utterance' and the 'fact-time' - although
there is no direct relation between both expansions.2
In various articles, I myself attempted to prove that the two levels of
'utterance-time' and 'fact-time' are not sufficient to explain the German
tenses and the time levels which they indicate. Rather, an explanation
of the tenses makes it necessary to introduce a third time level: the time

signified and fictionality as the attribute of a certain relation between sigmfier and
signified. Thus, he would maintain that, in the case of fictional works (he and other
authors are only interested in those), only the 'story-time', but not the 'time of the
act of narrating' is fictitious. Also, other authors seem to understand the 'time of the
act of narrating' as a non-fictive phenomenon, measunng it, like e.g. Milller
(1947/1968: 257; in this volume 67-84), Lammert (1955: 32) and G. Genette
(1971: 99ff) into - non-fictitious - printed pages, or by understanding it as reading
time of a non- fictitious reader, like Tomasevsky (1925/1965: 281) and Chatman
(1978/1983: 62). But this is questionable from the point of view of terminology,
because, as soon as the term is taken literally - and that means it is related to the
producer - it becomes obvious that it can only be the production time of a fictitious
narrator who narrates orally, and not the time of writing of a non-fictitious author
(Milller 1947/1968: 257f; in this volume 67-84] already stresses this
emphatically). In return, the interpretation of the 'time of the act of narrating' as a
non-fictitious phenomenon was only possible because of these reinterpretations of
the term.
E.g. H. Reichenbach (1947/1966: 288; [in this volume p. 1-11]) with the
introduction of a .point of reference' (besides a 'point of speech' and a 'point of the
event'), and K. Baumgartner & D. Wunderlich (1969: 34ff) with the introduction
of a 'contemplation time' (Betmchtzeit) (next to 'speech time' and 'action time').
Story-time and Fact-sequence-time 145

of the act of viewing, i.e. 'observation-time'. For example, the differ-

ence between the New High German tenses 'perfect' and 'preterit' and
between the time levels of 'perfect past' and 'imperfect past' indicated
by them, cannot be explained only by the levels of 'utterance-time' and
'fact-time', 'utterance-time', 'fact-time' and 'observation-time' all have
to be put in relation to each other.3
Transferring this insight to the field of narrative theory, I then sug-
gested in my article 'Das Riickschaukapitel in Thomas Manns Novelle
Der Tod in Venedig oder Welche Moglichkeiten eroffnet die Stagnation
der Betrachtzeit' ('The Retrospect Chapter in Thomas Mann's Novella
Der Tod in Venedig or Which Opportunities are Opened up by the Stag-
nation of the Time of the Act of Viewing?')4 to introduce the level of
the 'observation-time' into narrative time theory. Thus, its dichotomy
of 'discourse-time' and 'story-time' would be expanded into a tricho-
tomy of 'discourse-time', 'observation-time' and 'story-time'. But
while there cannot be any doubt of the necessity for this expansion, the
final question is if this is sufficient for an adequate description of the
complex relations between time levels in narrative texts.
This question does not only occur from an empirical perspective, it
already arises when comparing somewhat more precisely both tricho-
tomies and especially, within them, both levels that are seen as mutu-
ally corresponding, 'story-time' and 'fact-time'. It occurs, so to speak,
already under a purely conceptual point of view; because conceptually,
story-time is already a linguistic level, while fact-time is still an ex-
tra-linguistic level.5 Conceptually, the exact correspondence to the level
of story-time in narrative theory is not the tense-theoretical level of the
time of facts, i.e. of fact-time, but a tense-theoretical level of the
uttered, or better still, of the expressed time.
However - a tense-theoretical distinction between 'expressed time'
and 'fact-time' may make sense conceptually, even semiologically (as
it corresponds to the general semiological difference between signified
and non-signified facts or signified and non-signified objects).6 But it is
rather questionable whether such a distinction is really necessary. For
where do we find a gaping difference between the time that is ex-
See R. Harweg 1976,1977, and 1987: 1571T.
The article is still unpublished.
Story-time rs rnsofar already a lrngurstic level, as rt rs - textually seen - the
verbalrsed fact-time.
The establrshed srgn models, however, do not make such a drstrnctron because rt rs
not a semantic drstrnctron, that rs: not a langue-drstrnction, but a drstrnctron of the
srgmfrer, that rs: of the parole. I suggested a parole-onented srgn model into whrch
thrs drstrnctron rs integrated in R. Harweg 1980: 286ff
146 Roland Harweg

pressed in a certain tense and the time of its respective fact? Certainly,
if the time of a certain fact is not expressed, then it remains mere fact-
time and does not become and is not expressed time. But this is an ob-
servation that only has conceptual implications and not also empirical
ones; because as soon as the time of certain facts is expressed in a cer-
tain tense, then the time of these facts is also the expressed time, and
even as both time units are not similar, they are, nevertheless, congru-
ent. But in the framework of tense theory - which is mainly dealing
with individual tenses - this congruence may, from the empirical point
of view, still seem to render the said distinction of time levels obsolete.
Therefore, seen in empirical terms and in the framework of tense the-
ory, it still appears to make sense if we restrict the analysis to three
levels. So I uphold my suggestion to name these three levels 'utterance-
time' (Aufierungszeit), 'observation-time' (Betrachtzeit), and 'fact-
time' (Sachverhaltszeit).
However, this does not apply to narrative theory's time model. Ac-
cording to the academic opinio communis, it still relies on a mere two
level model that deals with the layers of the time of the act of narrating,
i.e. 'discourse-time' and the narrated time, i.e. 'story-time'. But neither
this model nor a tnchonomic concept, as I suggested in my essay 'The
Retrospect Chapter in Thomas Mann's Novella Death in Venice', will
suffice. In fact, narrative theory's time model requires an extension to a
concept of four levels. This four-level concept is a theoretical equival-
ent to the four-way distinction in tense theory between 'utterance-time',
'observation-time', 'expressed time' and 'fact-time' - distinctions
which I had suggested as conceptually, but not as empirically necessary
and useful. It is a correspondency which in a way takes shape as a syn-
tagmatic expansion of those tense-theoretical levels of time. Thus, the
concept of the 'utterance-time' is syntagmatically expanded into the
concept of 'discourse-time'; the concept of an 'observation-time' con-
gruent to the 'utterance-time' extends into one congruent to dis-
course-time and/or reception-time; the concept of 'expressed time' is
syntagmatically enlarged into 'story-time'; and the concept of 'fact-
time' expands syntagmatically into 'fact-sequence-time'.
While two of the four time levels in tense theory - the levels of 'ex-
pressed time' and of 'fact-time' - which can be distinguished conceptu-
ally concur empirically, as I already mentioned, their corresponding
levels in the framework of narrative theory and narrative theory of time
do not. In other words, the levels of story-time and of fact-sequence-
time do not correspond. This is a matter of great interest to a theory of
narrative time. It is especially relevant as it implies a theoretical exten-
sion to a four-level-concept, which exceeds the extension of the preval-
Story-time and Fact-sequence-time 147

ent two-level-concept to a three-level concept that I anticipated in my

article 'The Retrospect Chapter in Thomas Mann's Novella Death in
Venice': Therefore, I will explore it more extensively below.

Some of the observations in the theory of time whrch have led to my suggestion to
drstingursh between the levels of story-time and fact-sequence-time have been
drscussed before, but always only - inadequately in my opinion - within the
framework of the traditional distinction between discourse-time and story-time.
Like the distinction itself, they are normally seen as attributes of those super
ordinate distinctions of levels in narrative theory whose elements have been
described in the English language with the terms 'story' and 'plot', with the terms
'sujet' {'sjuzet') and 'fable {'fabula') by Russian Formalists, like e.g. Tomasevskij
(1925/1965: 267ff; see also Erlich 1955), by Structuralists like Todorov (1966:
126ff) with the terms 'discours' ('discourse') and 'histoire' ('story'), by Lammert
(1955: 24ff) with the terms 'FabeV und 'Geschichte' (the first one being not on the
same side of the opposition as with the Russian Formalists), by Chatman
(1978/1983), following the French Structuralists, with the terms 'discourse' and
'story', and by Stanzel (1979: 39ff) with the terms 'Erzahlung' ('narration') and
'Erzahlung minusMittelbarkeif ('narration minus mediacy'). I cannot discuss here
in detail the question if my distinction can actually be correlated with the time
theoretical differences of these distinctions. Definitely, it has some strong points of
contact. Nevertheless, I would like to point out two aspects. First, I would like to
stress that the distinction that I suggested, contrary to the two established
distinctions, the time-theoretical and the general one, is not limited to the field of
art and fiction. Secondly, and also in contrast to the two established distinctions,
whenever my distinction is applied to this field, it always remains with both levels
in the field of fiction; none (or: not at least one) of the levels will transgress the
border towards non-fiction. Furthermore, finally, it can be assumed that the
distinctions in narrative theory of the type 'FabeF vs. 'Geschichte', 'plot' vs.
'story' or 'discours' vs. 'histoire' are not sufficiently differentiated and that they,
like the distinctions among time levels, can and have to be proved as elements of a
more differentiated system, like the time level distinctions of 'story-time' and 'fact-
sequence-time' and 'time of the act of narrating' and 'narrated time'. It seems that
Stierle (1971/1975) took an important step in this direction with his trichotomy of
'Geschehen' ('happenings'), 'Geschichte' ('story') and 'Text der Geschichte' ('text
of the story'). However, I believe that Stierle's explanation of his tnchotomy is not
concrete enough, so that I cannot recognise and judge the state of its elements.
Another tnchotomy, a distinction between 'histoire', 'recif and 'narration', has
been suggested by Genette (1983: lOf). Genette's level of the 'histoire' seems to
correspond to Stierle's level of 'Geschichte', and Genette's level of 'recif to
148 Roland Harweg

2. Fact-sequence-time
The phenomenon of 'fact-sequence-time' is not a homogeneous one in
itself, and it not only permits division into sub phenomena but it can
also be divided into different sub-phenomena from differing perspect-
ives. I divide it into different sub phenomena under two points of
views, that is, an ontological and an epistemological.

2.1 Fact-Sequence-Time from an Ontological Point of View

From the ontological point of view, a material and a formal fact-

sequence-time can be distinguished. Material fact-sequence-time is the
time that is constituted by the consequences of a material fact, and
formal fact-sequence-time is the time constituted by the consequences
of a formal fact. Material facts not only structure time but also fill it
with substance, or, in other words, these are all facts that constitute sub-
stantial time. On the other side, formal facts only structure time but do
not fill it with substance, or in other words: they only constitute empty
time. Thus, material facts are facts in the proper sense, actions, incid-
ents or situations. On the other hand, formal facts are clock and calen-
dar time, that is, facts in the broadest sense. Thus, material facts are ac-
tions such as those signified by sentences like 'Karl went to town' or
'Peter wrote a letter', incidents signified by sentences like 'It began to
rain' or 'A strong storm arose', or situations signified by sentences like
'It was very cold' or 'The population was well'. In contrast, formal
facts are facts like the one signified by sentences like 'Today is May 5'
or'We have the year 1939'.
Of course, material and formal facts are not two categories without
any interrelation. We know that the formal facts are founded, directly or
indirectly, in certain material ones, like the course of the day is founded
in the rotation of the earth around itself and the course of the year in the
rotation of the earth around the sun. But when it comes to the denota-
tions of the mere calendncal information - and, last not least, the nu-
merical components contribute or contributed to it - then the material
occurrences have been pushed to the background to a degree that makes
their characterisation as opposed to the material facts look justified and
Stierle's level of the 'Text der Geschichte'. However, Genette's level of 'narration'
as the level of the expression of the text of the story seems to go beyond Stierle's
trichotomy at the point of the expression of the text of the story, and Stierle's level
of 'Geschehen' seems to go beyond Genette's tnchotomy at the point of 'histoid.
Consequently, Genette's tnchotomy is not corresponding to Stierle's level of
'Geschehen' and Stierle's tnchotomy does not seem to have a conespondence to
Genette's level of 'narration'.
Story-time and Fact-sequence-time 149

reasonable. This is, of course, even truer for clock time statements than
it is for calendncal statements.
The level of material as well as the level of formal fact-sequence-
time consists of a multitude of strands. However, the number of strands
of formal fact-sequence-time comprises only a small percentage of the
number of strands of material fact-sequence-time; for the majority of
the strands of formal fact-sequence-time is, as is generally known, only
a function of the difference between geographical places, and for those
places that belong to the same time zone, only one single strand of
formal time is valid anyway. This aside, differences of time zones do
have a more or less clear effect on clock time and day limits, but -
since the extensive worldwide leveling of culturally founded differ-
ences in calendars - they impact only insignificantly on the year limits.
Of time differences on a higher level, only the seasonal differences
between places of the northern and places of the southern hemisphere
on earth are not accessible to the worldwide leveling, because they are
not culturally, but geographically and astronomically founded, unless
you do not define the seasons according to the state of the sun, but ac-
cording to their sequence in the course of the year, in other words, in
their sequence in the course of a year that begins or could begin at the
same time everywhere, with a difference of maximally one day.
Formal strands of fact-sequence-time are, different from the material
ones, not only few in numbers but can also, because of the equality of
the length of their units, be correlated to each other. Therefore, it would
be theoretically possible to reduce the various formal strands of fact-
sequence-time to a single one. In fact, this is exactly what has been
done in astronomy by declaring the medial time of the zero meridian of
Greenwich to be the normal time known as 'world time'. Different
from the formal strands, the various material strands of fact-sequence-
time cannot, as was stated above, be correlated to each other because
they are not composed of elements of the same length. They can also
not be correlated to each other as complete units. Aside from cases in
which different material strands of fact-sequence-time meet each other
and then partially interfere with one another - for they run, in contrast
to locally different formal strands, not always separately - and aside
from cases in which they run at the same place, we do not even know
how they could be localised in time with respect to each other as units.
Such localisation of locally different material strands of fact-sequence-
time that do not appear in the same area of perception is generally only
possible by a recourse to units of formal fact-sequence-time.
But material strands of fact-sequence-time can often, even in most
cases, not be correlated to each other. Moreover, in most cases, a mater-
150 Roland Harweg

ial strand of fact-sequence-time cannot be clearly identified at all, or,

more precisely, it is not easy to say how to define or segregate it. This
difficulty refers to the dimension of consecutiveness as well as to the
dimension of simultaneity. Nevertheless, certain material strands of
fact-sequence-time seem to exist which are able to eliminate this diffi-
culty or at least keep it within limits. These are those strands of fact-
sequence-time that are connected to a certain individual - to an indi-
vidual and not to a certain temporal allo-individual8 or a group of indi-
viduals. If we connect, by definition, strands of fact-sequence-time to
an individual instead of to a certain allo-individual of this individual, be
it a lifetime-allo-individual or a moment-allo-individual, we avoid the
difficulty of having to identify additional criteria in the dimension of
consecutiveness for a beginning and an end of the strands, in other
words, criteria for definition and segregation. Connecting the strands of
fact-sequence-time to individuals instead of groups of individuals also
allows a practical definition and segregation in the dimension of- sim-
ultaneous - coexistence of facts and sequences of facts (fact-
sequences). Of course, with the advantages of these two connections,
there are also certain disadvantages. Of these, the greatest are those
consisting in the exclusion of all facts and sequences of facts that do
not hold individuals as their elements, and these are not few, even if we
extend the concept of the individual to individual objects. But the set of
those facts and sequences of facts may be everything else but insigni-
ficant - because it not only comprises facts and sequences of facts
which are connected to diffuse groups of individuals but also those that
hold neither individuals nor groups of individuals as their elements - in
any case, the set of facts and sequences of facts connected to individu-
als is still so large that they are able to define the concept of material
fact-sequence-time in a representative way. For the time being, we will
regard such an exemplifying partial inclusion of the concept as suffi-
In the area of the formal as well as in the area of the material strands
of fact-sequence-time, there is, next to the many simultaneously paral-
lel running strings, also a group of simultaneous strings which run quasi
on top of each other. They differ by various degrees of abstraction and
corresponding inclusion. In the same way as the formal strand of fact-
sequence-time is defined by a certain geographical place and can be
measured in units of various degrees of inclusion, like e.g. hours, days,
or years, the material strand of fact-sequence-time too can be defined
I define a temporal allo-individual in a certain analogy to terms like 'allophone' or
'allomorph' as a temporal phase in the life of an individual which stands for the
individual as such. See R. Harweg 1991.
Story-time and Fact-sequence-time 151

by a certain individual, and can be deconstructed into facts of various

degrees of inclusion. To name two quite extreme ways of deconstruc-
tion: the deconstruction of a material strand of fact-sequence-time con-
nected to a certain individual as a curriculum vitae or biography, is, as
we know, considerably more extensive than the deconstruction of a di-
ary that is kept lifelong, first by the parents of the individual and then
by the individual itself. It is true that neither the curriculum vitae nor
the diary are restricted to the mere deconstruction of the material string
of fact-sequence-time - both correlate this deconstruction with the de-
construction of the accompanying formal strands of fact-sequence-time
of the respective individual. For example, they indicate in which year
resp. on which days the respective facts took place, but this correlation
does not change anything in the principal self-reliance of the correlated
The inclusive and less inclusive sequences of facts are, so to speak,
different deconstructions of one and the same substrate of a fact. It is
true that they constitute respectively different strands of fact-sequence-
time, but their difference is completely different from the difference of
simultaneously parallel running strands of fact-sequence-time, that is,
from those that - in the area of material sequences of facts - are tied
between those sequences that are connected to various individuals. For
the sequences of facts that run on top of each other represent, in the
form of the common substance, in each case a kind of indirect identity.
The sequences of facts running parallel to each other are, by lack of a
common substance, are not even indirectly identical.

2.2 Fact-Sequence-Time from the Perspectives of Epistemology

and Knowledge Theory

Other than from the ontological point of view, the phenomenon of fact-
sequence-time can also be sub-categonsed from the perspectives of epi-
stemology and knowledge theory.
The main distinction from the perspective of epistemology and
knowledge theory is the one between objective and subjective fact-
sequence-time. But can the fact-sequence-time that has been labelled as
objective really be recognised and known? Is not a fact-sequence-time
that is not perceived or, to be more precise, not perceived and experi-
enced, and more so, the fact-sequence-time that is only known in retro-
spect eo ipso always a subjective fact-sequence-time?
Yes and no. Yes insofar as the perceived and experienced fact-
sequence-time is, in fact, always subjective, if and as long as the per-
ceiving-experiencing entity is a human being, or, more precisely: a
152 Roland Harweg

mere human being, a human being with human possibilities and abilit-
ies. But not if the perceiving-experiencing and the perceived then re-
cording and then knowing entity is, in place of a human being, a film
camera or a sound recorder or, instead of the ordinary human being, a
Active super human. Both entities register, as one is inclined to assume,
objectively the fact-sequence. The objective eye of the camera has
already become proverbial. And why should a fictitious super human
who is able to perceive, like the majority of fictitious third person nar-
rators in fictional narratives who can perceive what the characters
which they narrate about think and feel, not be in a position to perceive
sequences of facts objectively and then save them objectively in their
brains? At least, each longer conversation that is reproduced literally
bears witness to their super human abilities to save the experienced and
to recall it from their memory.
To recapitulate: a fact-sequence-time that is perceived, saved and re-
called by a normal human being is subjective, but this subjective fact-
sequence-time is subjective in varying degrees. The least subjective, of
course, is a fact-sequence-time that is perceived isochromcally; all re-
membered fact-sequence-times are more subjective. At best, they can
meet the relative degree of objectivity of isochromcally perceived fact-
sequence-time on selected occasions. Of course, even on selected occa-
sions, they cannot surpass it. What they are able to surpass and occa-
sionally really do surpass is the degree of understanding because a re-
membered sequence of facts is probably often understood better than an
actually perceived one.
At certain points, a specific kind of fact-sequence-time can interfere
with the remembered fact-sequence-time that runs through all degrees
of closeness and distance to objectivity. I would like to name it 'recon-
structed fact-sequence-time'. As a rule, the reconstruction aims at elim-
inating gaps or doubts in the remembered 'fact-sequence-times' accord-
ing to certain facts. But this can also lead to a retrospective correction
of sequences that did not, at first, raise any doubts in memory. At least
in the most favorable cases, the reconstruction is able to insert compon-
ents of objective fact-sequence-time into strands of subjective fact-
sequence-time und thus to objectify them partially. However, this ob-
jectification can probably only be relative, as it can never totally elim-
inate the subjectivity of its base.
The re-constructible fact-sequence-time, or more precisely: the re-
constructible components of fact-sequence-time can be reconstructed in
detail to a varying degree. Accordingly, their share in fact-sequence-
times differs. It is the greater the less detailed it is, supposed that it is
Story-time and Fact-sequence-time 153

re-constructible, or, in other words: the more we forego details in our

attempts to reconstruct.
Not only a person who remembers attempts to reconstruct compon-
ents of past fact-sequence-time, but so, too, does a reader. Like the per-
son who remembers attempts to reconstruct components of sequences
of facts that he himself has experienced or done, the reader tries to re-
construct not only, but predominantly, those that he has not experienced
or done himself. This activity is most widely spread among the readers
of non-fictional texts, and especially among the readers of historical
documents, that is, among historians who attempt to reconstruct histor-
ical events. But of course, less spectacularly and more subconsciously
than consciously, this is also done by readers of fictional texts, and
here, the fact-sequence-time is, of course, a fictitious one.9 However,
this is sometimes understood neither by the authors nor by the readers,
laymen or philologists alike.10

3. Story-time and its Differences to Fact-Sequence-Time

Whatever the authors of narrative texts, non-fictional as well as fiction-

al ones, narrate - which the latter can only do through a medium, the
instance of the fictitious narrator - is in itself not fact-sequence-time,
but story-time. This is a form of time that is, as we mentioned before,
not only terminologically but also empirically different from the form
of fact-sequence-time. What then are these differences from the empir-
ical point of view?

The level of fictitious fact-sequence-time attributed to a fictional text is the time
level of rts fictitious world. It rs a fictitious world drfferent not only from our non-
fictitious world (within which the interpreters of fictional narrative works
occasionally saw fit to locate the level of represented events, as opposed to that of
the plot). It is also different from the level manifesting itself in terms of tables of
contents, headlines of chapters and drafts, a level on which F.K. Stanzel (1979,
39ff) believed the story of a fictional narrative work to manifest itself. For while
the facts embedded in the fact-sequence-time are at the same time also entities
which the fictitious narrator has found to exist in his fictitious world, the facts
referred to in tables of contents, headlines of chapters and drafts, are entities that
only exist in the imagination of non-fictitious authors. On tables of contents and
drafts, see besides Stanzel 1979: 39ff, also R. Harweg 1979a.
Normally, the time of which a fictitious text narrates should not be regarded as
fictitious, at least not when it is dated, using data of years.
154 Roland Harweg

3.1 The Parameter of Selection

The main parameter by which story-time, from the empirical point of

view, differs from fact-sequence-time is the parameter of selection. For
it can be assumed that story-time in general covers only part of- alto-
gether only a very small part - of the fact-sequence-time, at least, of the
objective one." Of course, this is mainly true for the totality of strands
of fact-sequence-time, for the unmanageable totality of those countless
strands of fact-sequence-time that partly follow each other, partly run
parallel, and partly cross each other in one or the other way, and many
of those are not even perceived - let alone remembered or even nar-
rated. But beyond this, it is also true for those strands of fact-sequence-
time that are actually being narrated; because of those strands, only a

The parameter of selection has been observed in literature on narrative theory
before. However, here it is inappropriately addressed within the framework of a
different opposition of time levels, that is in the traditional opposition of 'discourse-
time' (Erzahlzeit) and 'story-time' {erzahlte Zeit). An example for this treatment of
the parameter is the one by G. Genette (1971: 99ff) and following him, S. Chatman
(1978 [1983]: 68ff). They recur to the term of ellipsis, a term that refers to the
restitution of time and form, so to speak, the systematically complementary term to
selection. Chatman defines this term as a procedure of narration in which the
discourse-time is not only shorter than the story-time but almost zero. According to
him, only in this way can the discourse-time continue, if, for example, there are
three of four non-narrated hours between two chapters. These three of four hours of
story-time are then zero. But such an interpretation is obviously wrong because the
process of narration - and only this constitutes this narration - is not interrupted
between the two chapters. At least, we have no indication for the assumption that
the fictitious narrator paused at the end of a chapter or that he paused there for a
longer time, as a fictitious reader would pause and which would correspond to the
contingent of empty space in the written text. Although, under certain
circumstances, he is not refused a chance to pause there, like e.g. a fictitious reader
is not refused a chance to pause there, perhaps for several hours, but such a pause
would obviously not stand in an intrinsic relation to an accordingly long non-
narrated time, more precisely: an accordingly long non-narrated fact-sequence-time.
That shows that the correlation of story-time and discourse-time, as interesting as it
may be as a playful moment, does not make sense from the point of view of theory.
Not story-time and discourse-time, but story-time and fact-sequence time are the
levels which logically have to be compared with each other when dealing with
selection according to time of an ellipsis. For only two levels can logically be
compared to each other that, like the level of story-time and fact-sequence-time, are
related in, semiotically seen, a naturally motivated way, and not two levels that, like
story-time and discourse-time, are in an artificial-arbitrary relation to each other.
The relation - more precise: the durative relation - between story-time and
discourse-time is not of semiotic, but only of physical interest.
Story-time and Fact-sequence-time 155

small number are actually narrated. Only some punctual discontinuing

details are being narrated, at least in detail. This can be especially
clearly recognised when looking exclusively at the story-time of an in-
dividual text; for there, in many cases, the non-narrated gaps in the in-
dividual strands of fact-sequence-time are explicitly marked by certain
expressions for sections of formal time. When, for example, a new
chapter in a narration starts with the expression 'two years later', then
this expression is an indicator that two years of this strand of fact-
sequence-time, at least the material variety of it, are jumped and omit-
ted, under the condition that the same strand of fact-sequence-time is
being continued, for example, the life story of one and the same person.
If they are not made up retrospectively at a later point - something that
can only be done selectively - they remain omitted.
If you do not want to confine yourself to one individual text and if
you are ready to consult several or even as many texts as are possible
and relevant for the narration of certain strands of fact-sequence-time,
then the narrative coverage of the relevant strand of fact-sequence-time
should be less incomplete or punctual. Rather, it is very likely that the
relevant strands of story-time complement each other mutually and
consequently and fill part of the gaps within the said strand of fact-
sequence-time left by the individual strands. Certain parts of this strand
will be even narrated in several texts. However, even with a relatively
large number of texts dealing with one and the same strand of fact-
sequence-time (and whose authors have not copied each other but who
have narrated what they experienced by themselves), a considerable
part of the relevant strand of fact-sequence-time will remain un-nar-
rated and therefore, will have no correspondence on the level of the
story-time. Even when some parts of the strand of fact-sequence-time
are narrated several times, this deficit will obviously not be balanced.
The number and size of gaps of story-time in relation to the fact-
sequence-time differs strongly according to the use of written texts
only, or if oral texts (and their traces of memory) are also taken into
consideration. Accordingly, in order to fill existing gaps, historians of
contemporary history make an effort to not only make use of the writ-
ten documents that they can acquire but also, if possible, to use oral
texts that reproduce the experiences of their authors.
Sometimes, some of the gaps that cannot be closed in this way may
be closed by reconstruction in regard to certain details and to certain
stages of abstraction, but very often, partly because of contradictory
and partly because of information that cannot be interpreted in a non-
ambiguous way, the attempt to reconstruct only leads to hypotheses
with a limited degree of probability. A beautiful and perhaps - accord-
156 Roland Harweg

ing to the relation of its size to the specificity of the genre poetics of the
text - unique example for such an attempt to reconstruct, is the chapter
'Disappeared Journey' in Peter de Mendelssohn's Biography of Thomas
Mann Der Zauberer (The Magician)" It is devoted to the reconstruc-
tion of an unknown trip of Thomas Mann to Pans. In this chapter, the
biographer evaluates oral as well as written documents. The first ones,
utterances by Thomas Mann's wife Katja, state that she does not know
anything about this trip, and the latter, quotations from letters and other
texts by Thomas Mann, hint only ambiguously and more or less indir-
ectly to this trip.
The strands of story-time, belonging to various texts that are inde-
pendent of each other, but relating to one and the same strand of fact-
sequence-time, differ from it in the way that they, put together, do not
form a strand themselves. Rather, they may only be able to form one in
a re-narration that integrates them in a line. But this lack of succession
and linearity of those strands can also be found in the opposite direction
when a multitude of more or less simultaneous strands of fact-
sequence-time corresponds with one strand of story-time within a
single text. Thus, for a linguist who cares for symmetry between the
levels, this would be no reason to desist from a consideration of strands
of story-time that can be related to one single strand of fact-sequence-
time in a multitude of texts.
However, already on the level of the practitioner, it is not always
possible to recur to a multitude of texts in order to mutually integrate in
a linear fashion their respective strands of story-time so that they can be
related to one and the same strand of fact-sequence-time. Although this
can be done, at least principally, in the field of non-fictional narratives,
that is, in the non-fictional cosmos of narration (where one has to apply
this strategy for practical reasons, as for example in the case of judges
and historians), it is principally not possible to follow this approach in
the field of fictional narrative texts. To recur to other fictional texts is
not an option if we understand by fictional text, as is the general rule,
only the highest-order and most inclusive fictional narrative. In other
words, we exclude the possibility to recur to embedded conversations
or letters inside a narrative or a novel but may only use the respective
narratives and novels themselves and, in case the narratives are com-
ponents of a consistent series or and the novels components of a con-
sistent cycle, then we may recur to these series and cycles themselves.
For all these most integrative fictional texts relate to only one fictitious
world respectively. That means, for example, that the strands of story-
time in various novels which do not happen to belong to one and the
See P. de Mendelssohn 1975: 764-69.
Story-time and Fact-sequence-time 157

same cycle, always also relate to different strands of fact-sequence-

time, if not even to strands of fact-sequence-time belonging to different
fictitious worlds. Therefore, they can never be integrated into each oth-
er in a linear way. Even if it sometimes looks, because of reference to
similar places or similar dates, as if the plot is set at the same place and
at the same time - in reality, it only takes place at homologous places
and at homologous times that may have similar characteristics and bear
the same names but that belong to existentially different - fictitious -
On the other side, it is possible that the story-time of one and the
same fictional text is distributed among different strands of fact-
sequence-time; and the more voluminous the text, the more use is made
of this possibility. It provides a wide area for more detailed studies to
explore to which degree these various strands of fact-sequence-time of
a fictitious world related to such a text will run in parallel, and to which
degree they will follow after each other."
Consecutiveness and simultaneity are also the basic questions when
it comes to the strands of fact-sequence-time in one and the same ficti-
See R. Harweg 1979a and 1979b.
Also E. Lammert (1955:85) points to the simultaneity of strands of fact-sequence-
trme, albert not in the framework of the opposrtron of fact-sequence-trme to story-
time, but, as rs customary in narrative theory wrthrn lrterature analysis, rn the
framework of the drstrnctron of story-time and drscourse-trme. Srmultaneously
runnrng events are, according to Lammert, rn a specral fnctron to the
consecutiveness of the narration. But at one pornt, he writes, lrberally quoting, that
the classrcal phrlologrst Zielmski proved rn an 'interesting study' that rn the Iliad,
'events that ran srmultaneously rn realrty (...) were moved rn a way by the narrator
rn therr narrated time (!), that the "consecutive narration turns rnto a consecutive
state of facts" (Zrelrnskr 1901: 434)' (Lammert 1955: 85 [our translation]). Here
appears, at least for a short moment, the opposrtron of fact-sequence-trme and story-
time. But Lammert obvrously drd not trust thrs opposrtron. At least, rt seems that he
drd not see rt as normal; for he put an exclamation mark behrnd the expressron 'rn
therr narrated time'. However, one has also to agree to thrs exclamation mark
objectively because probably thrs case rs not the normal case rn the relatronshrp
between the 'fact-sequence-trme and story-time, and thrs especrally since the Iliad
is a fictional text. It seemingly should deal with the case in which the story-time
rncorrectly reflects the fact-sequence-time - as is the case in lies and false accounts
- rn a srgmfrcant way, that rs, not only through mere omrssrons, perspectives, and
srgnalrsed conversrons - albert possrbly not on purpose. But rn fictional texts, such
corruptions cannot be realrsed from the outsrder perspective of a non-frctrtrous
reader; at best he can only realrse contradrctrons. Thus, one rs inclined to ask how
Zielinski, after all, in his study of the Iliad, could come to the conclusion that
something that rs narrated as happemng consecutively, happened srmultaneously rn
158 Roland Harweg

tious world that are related to the strands of story-time in various em-
bedded narrative texts, for example, various conversations or letters
that occur in one and the same novel. However, the strands of story-
time of various conversations and/or letters in one and the same novel
do not have to relate necessarily to different strands of fact-sequence-
time, they can relate to one and the same strand. In this case and under
certain circumstances, they allow a linear mutual integration.
The phenomenon of narrative invention seems to be in opposition to
the phenomenon of narrative selection; for, in the same way as the phe-
nomenon of narrative selection leads to the underrepresentation of the
level of story-time against the level of fact-sequence-time, the phe-
nomenon of the narrative invention, that is, the phenomenon that seems
to basically constitute narrative texts, seems to lead to an overrepresent-
ation of the level of story-time compared to fact-sequence-time. But do
the fictional narrative texts, as implied in the term of overrepresenta-
tion, really miss the level of fact-sequence-time?
This is a question which, at a closer look, cannot be answered off-
hand in this way because, strictly speaking, the expression 'fictional
narrative text' is a contradiction in itself. Strictly speaking, it signifies a
phenomenon belonging to two different worlds, on one hand to this our
non-fictitious world, and on the other hand to a certain fictitious one -
two worlds with different creators of this phenomenon, the non-ficti-
tious author and the fictitious narrator.15 As a product of the fictitious
narrator, the text is a narrative only in his fictitious world, and there we
find, as has been stressed several times, a level of fact-sequence-time in
addition to the level of story-time. But in this our non-fictitious world,
the so-called fictional narrative text is no narrative in the strict sense
but only an invention. But if in this, our non-fictitious world, it is not
narration, then it cannot possess a level of story-time. Can it then have
a level of fact-sequence-time? Not a really secure one. At best an inven-
ted one. But should one not simply call it the level of invented time? In
any case, the phenomenon which has been defined as overrepresenta-
tion of story-time over fact-sequence-time, is, at least as far as the so-
called fictional narrative texts are concerned, a mere pseudo-phenomen-
on. It remains to be seen if this is also the case with lies,that is, in the
case of texts that cannot be counted as fictional texts although they do
represent inventions albeit combined with certain elements of reality.

See Harweg 1979a and 1979b.
Story-time and Fact-sequence-time 159

3.2 The Parameter of Direction

Apart from the parameter of selection, the level of story-time differs

from the fact-sequence-time mainly in terms of the parameter of direc-
tion; and the directions in which story-time elapses run in different di-
mensions. For the fact-sequence-time, or to be more precise, the section
of it that corresponds to the story-time of one single narrative text, is
not one-dimensional - as the term 'sequence' may suggest - but a quasi
three-dimensional phenomenon. It is three-dimensional because of a
multitude of different strands running in parallel, and also because of a
certain number of strands of fact-sequence-time that run on top of each
other. Therefore, the movements of the story-time that partially corres-
ponds to this fact-sequence-time are not limited to one dimension, that
is to the consecutive dimension - which is, however, the leading and
main dimension - but they also take place in the two other dimensions,
the parallel dimension and the dimension of one-upon-the-other.16
As already mentioned above, the parallel dimension is formed by
isochronal strands of fact-sequence-time which differ in place or per-
sons. The dimension of one-upon-the-other is formed by strands of
fact-sequence-time that are isochronal and/or refer to the same places
Similarly to the parameter of selection, the parameter of the direction of the story-
time has also not remained unreflected in narrative theory, let alone undetected -
albeit only in connection with the consecutive dimension. In the same way as the
parameter of selection, it has been discussed in the framework of an inadequate
opposition - that is the opposition between story-time and discourse-time - for it
was seen as a phenomenon on the level of discourse-time instead of a phenomenon
of story-time. E. Lammert (1955: 34 [our translation]), e.g., speaks of the
'rearrangement of parts of the story time in the course of narration', and S.
Chatman (1978 [1983]: 63) writes: 'The discourse can rearrange the events of the
story [...].' It has to be granted that the talk of rearrangements in the course of the
narration resp. of the rearrangement by the discourse can also refer to matters of
content, and probably is meant so sublimmally - even a term like the reversal of the
direction of the narration might be meant like this. But if, besides the level of story-
time, only the level of discourse-time is known, then story-time has to be seen as
the object of the rearrangement, along with Lammert, and the rearrangement itself
has to be regarded as a phenomenon of discourse-time. Also Chatman (1978
[1983]: 63) must see it that way, as he understands his sentence 'The discourse can
rearrange the events of the story [...]' as a specification of his statement -
following G. Genette -, that the arrangement ('order', W r e ' ) is a time relation
between 'story-time' and 'discourse-time'. Chatman's terms 'story-time' and
'discourse-time' correspond exactly - and so do his definitions of those as 'the
duration of the purported events of the narrative' resp. 'the time it takes to peruse
the discourse' (1978 [1983]: 62) - to the German terms ' erzahlte Zeif and
160 Roland Harweg

and persons, but differ in abstraction. As far as story-time is concerned,

it is able to connect isochronal facts of parallel running strands or
strands of fact-sequence-time that run on top of each other, in the same
way as it is able to follow facts of one and the same or of different
strands of fact-sequence-time. If it follows consecutive facts, I will call
it and the dimension of fact in which it moves, longitudinal; if it con-
nects isochronal facts of parallel strands of fact-sequence-time, I will
call it and the dimension of fact in which it moves, latitudinal, and if it
connects isochronal or better still, partly isochronal strands of fact-
sequence-time - for different degrees of abstraction only allow partial
isochrony - then I will call it and the dimension of facts in which it
moves, altitudinal.
In each of the three dimensions, and not only in the longitudinal one
- for which it is well known - but also in the latitudinal and in the alti-
tudinal dimension, two opposite directions of story-time can be identi-
fied: in the longitudinal dimension, it is the progredient and the re-
gredient, in the latitudinal dimension, it is the departing or centrifugal
and the approaching or centripetal time, and in the altitudinal dimen-
sion, the descendent and the ascendant time. The basis of the opposi-
tions of directions in the course of the story-time are asymmetries in the
area of fact-time or fact-sequence-time: their one-dimensional progres-
sion in the longitudinal dimension, their spatial perspective in the latit-
udinal dimension and the - scalar - opposition between abstraction and
concreteness in the altitudinal dimension.

3.2.1 Altitudinal Directions of Story-time

To begin with the altitudinal dimension: of the two directions in which
story-time can move in this dimension, the one which is mostly chosen
is probably the descendent direction, the direction leading from the ab-
stract to the concrete and thus an act of concretion.
This concretion - an act that consists, from the point of view of
time, in a restriction, a partialisation of the story-time - demands, from
the point of view of text grammar, a respective marker, a marker that
can localize definitely and indefinitely, the temporal place of the
product of the restriction and the partialisation in relation to its original
unrestricted abstract temporal space. The restriction of story-time with
definite explicit localisation of the temporal product of restriction can
be observed, e.g. in the sequence of phrases taken from Thomas Mann's
novella Tristan:
Story-time and Fact-sequence-time 161

(1) Herr Spinell saB der Gattin Herrn Kloterjahns bei Tische gegentiber.
Zur ersten Mahlzeit, an der die Herrschaften teilnahmen, erschien er ein
wenig zu spat in dem groBen Speisesaal im ErdgeschoB des Seitenfliigels,
sprach mit weicher Stimme einen an alle gerichteten GruB und begab sich
an seinen Plate, worauf Doktor Leander ihn ohne viel Zeremonie den neu
Angekommenen vorstellte. Er verbeugte sich und begann dann, offenbar
ein wenig verlegen, zu essen, indent er Messer und Gabel mit seinen
groiten, weiiten und schon geformten Handen, die aus sehr engen Armeln
hervorsahen, in ziemlich affektierter Weise bewegte. Spater ward er frei
und betrachtete in Gelassenheit abwechselnd Herrn Kloterjahn und seine
Gattin. Auch richtete Herr Kloterjahn im Verlaufe der Mahlzeit einige Fra-
gen und Bemerkungen betreffend die Anlage und das Klima von "Einfried"
an ihn, in die seine Frau in ihrer lieblichen Art zwei oder drei Worte ein-
flieBen lieB, und die Herr Spinell hoflichbeantwortete.

Herr Spinell sat opposite Herr Kloterjahn's wife at table. On the occasion
of the new guests' first appearance in the great dining-room on the ground
floor of the side wing, he arrived a minute or two late, murmured a greeting
to the company and took his seat, whereupon Dr Leander, without much
ceremony, introduced him to the new arrivals. He bowed and began to eat,
evidently a trifle embarrassed, and manoeuvring his knife and fork in a
rather affected manner with his large, white, well formed hands which
emerged from very narrow coat sleeves. Later he seemed less ill at ease
and looked calmly by turns at Herr Kloterjahn and at his wife. Herr Kloter-
jahn too, in the course of the meal, addressed one or two questions and re-
marks to him about the topography and climate of Einfried; his wife also
interspersed a few charming words, and Herr Spinell answered politely.17

In this sequence of phrases, the first sentence indicates the places taken
by Mr. Spinell and Mrs. Kloterjahn in the dimng-hall at mealtime over
a longer period of time. These places are obviously determined by a
table order. The text indicates the places, and at the same time, albeit in
a very abstract way, the time that they spend at these places during the
meals. This is an ambiguity according to which the time factually spent
during the meals is discontinuous, and the time filled with the table or-
der is a continuous time which, at the first meal, already includes the
seconds before the described actual place taken by Mr. Spinell occurs
and, therefore, his belated appearance in the dining hall. The sequence
of sentences that follow on the opening one then indicate only one of
the many meals during which Mr. Spinell and Mrs. Kloterjahn sit op-
posite each other, and they also indicate this meal as the first.
A restriction of story-time with indefinite explicit localisation of the
product of restriction exists everywhere, i.e. where story-time tran-
Th. Mann 1958: 225, resp. Th. Mann 1998: lOOf.
162 Roland Harweg

scends from a section of durative or iterative time to a piece of punctual

time that is localised by an indefinite singular (semelfaktiv) adverb like
once or one day within this section of time. This is the case e.g. in this
sequence of phrases taken from Thomas Mann's Death in Venice:

(2) [...] Aschenbach erwartete taglich Tadzios Auftreten, und zuweilen tat
er, als sei er beschaftigt, wenn es sich vollzog, und lieB den Schonen
scheinbar unbeachtet voriibergehen. Zuweilen aber auch blickte er auf, und
ihre Blicke trafen sich. Sie waren beide tiefernst, wenn das geschah. In der
gebildeten und wiirdevollen Miene des Alteren verriet nichts eine innere
Bewegung; aber in Tadzios Augen war ein Forschen, ein nachdenkliches
Fragen, in seinen Gang kam ein Zogern, er blickte zu Boden, er blickte
lieblich wieder auf, und wenn er voriiber war, so schien ein Etwas in seiner
Haltung auszudriicken, daB nur Erziehung inn hinderte, sich umzuwenden.
Einmal jedoch, eines Abends, begab es sich anders. [...]

[...] Aschenbach waited daily for Tadzio to make his appearance and
sometimes pretended to be busy when he did so, letting the boy pass him
seemingly unnoticed. But sometimes, too, he would look up, and their eyes
would meet. They would both be deeply serious when this happened. In the
cultural and dignified countenance of the older man, nothing betrayed an
inner emotion; but in Tadzio's eyes there was an inquiry, a thoughtful
questioning, his walk became hesitant, he looked at the ground, looked
sweetly up again, and when he had passed, something in his bearing
seemed to suggest that only good breeding restrained him from turning to
But once, one evening, it was different. [...]"

Different from the p o s t u l a t e s of text grammar - which are, however,

formulated in no grammar - there are often, especially in literary narra-
tions, restrictions of story-time that are not explicitly identified at all.
Such a restriction, for example, can be found in the sequence of phrases
- taken again from Thomas Mann's novella Tristan

(3) Das gute Wetter hielt an. [...] Der Gattin Herrn Kloterjahns ging es
leidlich in dieser Zeit; sie war fieberfrei, hustete fast gar nicht und aB ohne
allzu viel Widerwillen. Oftmals saB sie, wie das ihre Vorschrift war, stun-
denlang im sonnigen Frost auf der Terrasse. [...] Dann bemerkte sie zuwei-
len Herrn Spinell, wie er [...] sich im Garten erging. Er ging mit tastenden
Schritten und einer gewissen behutsamen und steif-graziosen Armhaltung
durch den Schnee, griiBte sie ehrerbietig, wenn er zur Terrasse kam, und
stieg die unteren Stufen hinan, um ein kleines Gesprach zu beginnen.
"Heute, auf meinem Morgenspaziergang, habe ich eine schone Frau gese-
Th. Mann 1958: 497, rsp. Th. Mann 1998: 243.
Story-time and Fact-sequence-time 163

hen ... Gott, sie war schon!" sagte er, legte den Kopf auf die Seite und
spreizte die Hande.

The fine weather continued. [...] At this period Herr Kloterjahn's wife
seemed to be in tolerably good health; she had no fever, scarcely coughed
at all, and had not too bad appetite. Often she would sit out on the terrace
for hours in the frost and the sun [...]. Sometimes she would see Herr
Spinell walking in the garden. [...] He walked through the snow with a
tentative gait and a careful, prim posture of the arms; when he reached the
terrace he would greet her very respectfully and mount the steps to engage
her in a little conversation.
'I saw a beautiful woman on my morning walk today ... Ah, dear me, how
beautiful she was!' he said, tilting his head to one side and spreading out
his hands.19

In this sequence of phrases, the story-time is, in the first phrases, partly
durative and partly iterative: durative in the case of the predicate hielt
an (continued), and in the case of adverbial adjunct in dieser Zeit (at
this period), and iterative in the case of the adverbial adjunct oftmals
(often), dann and zuweilen (sometimes), and the conjunction wenn
(when). The last phrase of the sequence of phrases, the first sentence of
a conversation between Mr Spinell and Mrs Kloterjahn, then indicates,
without being explicit, the beginning of a unique event, and that is a
conversation which is indicated as unique by content and extensiveness.
However, the quoted sequence of phrases is not only an example for
implicit restriction; it is, at the same time, an example for an act that
could be called a thinning {Ausdiinnung). In this sequence of phrases,
thinning - an act that precedes the restriction - consists of the segment-
ation of the primarily narrated sentence into iterative discontinuing ele-
ments. This is done in two steps, with the adverbs 'often' {oftmals) and
'then' {dann) at the first step, and at the second step with the adverb
'sometimes' {zuweilen) and the conjunction 'when' {wenn). These are
expressions that, together with their predicates, perform an iterative
partialising selection from the denotative domains of the expressions
'often' {oftmals) and 'then' {dann) and their predicates.
Also, the ascendant, that is, the variant of altitudinal story-time that
leads from the concrete to the abstract and thus indicates a temporal ex-
tension of the initial temporal section, does not always explicitly indic-
ate this change, in this case an extension instead of a restriction. The
extension is called explicit, e.g. in the case of ascendant story-time
which makes up the end of the second chapter of the eleventh part of
Thomas Mann's Buddenbrooks. Beginning with the rattling of the
Th. Mann 1958: 229f., resp. Th. Mann 1998: 105f.
164 Roland Harweg

alarm clock in the early morning, the chapter describes in detail the
course of a weekday in the life of the young Johann Buddenbrook,
called Hanno. The last sentence resumes: 'This was one day in the life
of the little Johann' ('Dies war ein Tag aus dem Leben des kleinen Jo-
hann'). The sentence expands the day, albeit in the most possible vague
way, explicitly into a section of time that might include - in its function
as pars pro toto - if not the whole life of the little Johann, but neverthe-
less certainly some years of the same.
On the other side, in the sentence that follows the last sentence of
sequence (1), the ascendant extension of the story-time is only implicit:

Auch richtete Herr Kloterjahn im Verlaufe der Mahlzeit einige Fragen und Be-
merkungen betreffend die Anlage und das Klima von "Einfried" an inn, in die
seine Frau in ihrer lieblichen Art zwei oder drei Worte einflieiten liefl, und die
Herr Spinell hoflich beantwortete

Herr Kloterjahn too, in the course of the meal, addressed one or two questions
and remarks to him about the topography and climate of Einfried; his wife also
interspersed a few charming words, and Herr Spinell answered politely).

The following sentence Seine Stimme war mild und recht angenehm;
aber er hatte eine etwas behinderte und schlurfende Art zu sprechen,
ah seien seine Zahne der Zunge im Wege (His voice was soft and really
quite agreeable, though he had a slightly impeded, dragging way of
speaking, as if his teeth were getting in the way of his tongue) also
refers to the situation described in the sentence before - but this is not
explicitly said.

3.2.2 Latitudinal Directions of Story-time

Both direction of story-time in the latitudinal dimension, the departing
or centrifugal dimension, or the approaching or centripetal dimension
could be met, e.g. in descriptions of locations; for descriptions of loca-
tions can - hybrid forms aside - begin with the foreground and then
proceed to the background, or begin with the background and then pro-
ceed to the foreground. However, in most cases, descriptions of loca-
tions are not indications of fact-sequences.
But diary records are reproductions of fact-sequences, at least as the
foreground is concerned, as the somewhat abstract terms of foreground
and background may be used to declare the personal-private records as
foreground and the general comments, beyond the personal notes, as
background. When a diarist like Thomas Mann, in the last days of the
Second World War, regularly records the personal-private moments as
Story-time and Fact-sequence-time 165

well as the isochromc non-personal moments, and therefore narrates

them from the perspective of the respective evening, then he can either
begin with the foreground and the centrifugal, and proceed to the back-
ground, or he can begin with the background and then, centripetal, pro-
ceed to the foreground. As a rule, Thomas Mann chooses the first, the
centrifugal way. But sometimes, when it comes to events of extraordin-
ary importance and uniqueness, it may be the other way round. For ex-
ample, Thomas Mann's diary entry of May 7*, 1945, begins with some
details on the surrender of Germany, then proceeds to personal reflec-
tions in connection with the surrender and only then notes personal is-
sues that are not connected to the surrender but at the end, in a centrifu-
gal way, he mentions again non-personal general issues. In the same
way, non-personal general issues are put in front of Thomas Mann's di-
ary entries of August 7, 9, 10 and 11, 1945. In the same way as - in the
diary entry of May 7, 1945 - it was the end of the Second World War
that motivated the author to subordinate the personal records to the
non-personal-general and thus chose the 'centripetal' direction of story-
time, it is in the entries of August 1945 marking the end of the Second
World War in the Far Eastern theatre of war: first the dropping of the
atomic bomb over Hiroshima and Nagasaki and then, as its con-
sequence, rumors and negotiations on the surrender of Japan.™
3.2.3 Longitudinal Directions of Story-time
Finally, the clearly dominating one of the two directions of story-time
in the longitudinal dimension, as is generally known, is the progredient,
that is the one which corresponds to the direction of the fact-sequence-
time. The regredient, that is, the one which is opposite to the direction
of fact-sequence-time and therefore the real story-time, which consti-
tutes the longitudinal quality of difference between the two levels of
time is not only considerably more uncommon. It is also unusually re-
stricted, mainly concerning the number of its continuous steps, that is, it
is not interrupted by steps that do not again reach the starting point of
regression. In the majority of cases, it does not step back more than one
single step. Even a continuously stepping back of the story-time by two
steps is comparably uncommon, like the one that can be observed in the
sequence from the beginning of the second chapter of Thomas Mann's
novella Death in Venice in which the story-time, at first, goes back to
the birth of the hero and then further back to his ancestors. If you want
to reach events that lie far back, you normally use one big step instead
of a multitude of smaller steps, and then narrate progrediently from

See Th. Mann 1986: 200 and 238ff.
166 Roland Harweg

there. In fact, it seems that a continuous regression of story-time ap-

pears to us as more unnatural, the smaller and the more numerous its
steps are. But nevertheless, the degree of its unnaturalness does not
seem to reach the degree that is reached by, e.g. a film that runs back-
ward, and this for two reasons: firstly, because each step in it can be
connected to a certain signal of regression, and secondly, because the
individual elements of the story-time are formed progrediently, irre-
spective of the signals of regression that mark it. These are two facts
that, complementing each other and functioning together, make sure
that the respective fact-sequence-time appears as progredient although
it is narrated regrediently, which is very different from films being
played backwards. For example, the signals of regression in the regredi-
ently narrated story-time of the sequence of phrases Karl war wieder
da. Er war tags zuvor aus Amerika zuruckgekehrt (Karl was here
again. He had returned from the United States the day before) the ad-
verbial expression the day before {tags zuvor), and the past perfect -
leave no doubt that the fact-sequence-time flows from the past into the
future - that means: progrediently - in the same way as it would have
occurred with progredient narrated story-time. And, as far as an incid-
ent is referred to by the verb return (zuruckkehren) is concerned, it is
always progredient anyway.21
Probably, how many steps the story-time can step backward con-
tinuously at a certain place, and which degree of naturalness such a
continuous regression has, does not insignificantly depend on the kind
of inner structure of the respective fact-sequence-time in a text. Without
taking into account the point of view of boredom: the most extensive
regression concerning the number of steps (and also the most natural)
would be a case of straight rows of facts, e.g. straight lines of genealo-
gies. A very extensive regrediently narrated genealogy - albeit without
signals of regression - is the genealogy of Jesus as narrated by the
Evangelist Luke.
But if sequences of facts that are to be narrated do not consist of
such rows of similar facts, then one can, for the time being, only say
about their capacity of steps and the naturalness of the continuously re-
gredient story-time that it is significantly lower than in the former.
Something more precise, especially something typologically more pre-
cise, can probably only be said when the phenomenon of the regredi-
ently narrated story-time has been submitted to comprehensive and de-
tailed research. As a small foretaste to such studies, I will try to refor-
On the difference between aspects and kinds of actions see R. Harweg 1976: 5.
Aspects are phenomena that transcendend facts, kinds of actions are imminent to
Story-time and Fact-sequence-time 167

mulate a small text that is progrediently narrated - for further studies

will have to contain such experiments. As a text section, I will choose
the beginning of Death in Venice* and exactly this section, because it
is especially suited for the experiment in terms of its narrative structure.
Although progrediently narrated, it begins with a retrospective and a
retrospective that starts - against the rule, by the way - without any
point of departure. The point of departure, the denotation of the predic-
ate erwartete [...] am Nordlichen Friedhof die Tram (waited at the
Northern Cemetery for the tram), is only revealed later, to be precise: at
the end of the progedient narration of the retrospective, i.e. where the
retrospective transforms into the parallel perspective. And this is, at the
same time, the point from where we will attempt to reformulate the ret-
rospective in a regredient way which is, in the original, narrated pro-
grediently. The result could be the following sequence of phrases:

(4) Am Nordlichen Friedhof in Miinchen erwartete der Schriftsteller Gu-

stav Aschenbach oder von Aschenbach, wie seit seinem funfzigsten Ge-
burtstag amtlich sein Name lautete, an einem Friihlingsabend des Jahres
19.., das unserem Kontinent monatelang eine so gefahrdrohende Miene
zeigte, die Tram. Er befand sich auf dem Riickweg von einem weiteren
Spaziergang durch den Englischen Garten zu seiner Wohnung in der Prinz-
regentenstralte. Er hatte das letzte Stuck des Weges bei sinkender Sonne
auiterhalb des Parks iiber die offene Flur genommen. Vorher hatte er beim
Aumeister eine kleine Weile den volkstiimlich belebten Wirtsgarten iiber-
blickt, an dessen Rand einige Droschken und Equipagen hielten. Zum Au-
meister hatten inn stillere und stillere Wege gefuhrt. In der Nahe der Stadt
war der Englische Garten zunachst voller Wagen und Spazierganger gewe-
Aufgebrochen zu dem Spaziergang war Aschenbach bald nach dem Tee,
und der Grand fur den Aufbrach war gewesen, daB der Dichter am friihen
Nachmittag den entlastenden Schlummer nicht gefunden hatte, der ihm, bei
zunehmender Abnutzbarkeit seiner Krafte einmals untertags so notig war.
Er hatte ihn deshalb nicht gefunden, weil er, iiberreizt, dem Fortschwingen
des produzierenden Triebwerks in seinem Inneren nicht Einhalt zu tun ver-
mocht hatte, das die schwierige und gefahrliche, eben jetzt eine hochste
Behutsamkeit, Umsicht, Eindringlichkeit und Genauigkeit des Willens er-
fordernde Kraft der Vormittagsstunden in Gang gesetzt hatte.

At the Northern Cemetery in Munich, the writer Gustav Aschenbach or von

Aschenbach, as he had been known officially since his fiftieth birthday,
waited for the tram on a spring afternoon of the year 19.., which showed
such a dangerously threatening face to our continent as it had not done for
months. He was on the way back from a longer walk through the Eng-
See Th. Mann 1958: 444.
168 Roland Harweg

lischer Garten to his dwelling in the Prinzregentenstralte. The last distance

of his way he had taken through the open field, with the sinking sun outside
the park. Before, he had, at Aumeister's, for a little while, looked over the
popular and lively beer garden, at whose edge some Droschken and
Equipagen were stopping. Increasingly quiet paths had let him to
Aumeister's. Close to the city, the Englischer Garten had been, at first, full
of vehicles and people out for a stroll.
Soon after tea, Aschenbach had left for the walk, and the reason for the
walk had been the fact that the poet had, in the early afternoon, not found
the relaxing slumber that was so important to him once a day, with an in-
creasingly decrease of his strengths. He had not found it, because he had
not been able to stop the swinging of a producing mechanism inside him
that had been ignited by the difficult and dangerous work of the morning
hours that even now demanded highest caution, prudence, forcefulness and
elaborateness of the will [our translation].

I will not analyse the sequence of phrases here; it shows a strand of re-
gredient story-time that continuously steps eight steps back into the an-
teriority without being interrupted by any progredient steps. But per-
haps it can be sensed that the precise analysis of such regredient re-
arrangements of progrediently narrating texts and the comparison of
these rearrangements with their progrediently narrating originals gives
us the opportunity to gain deeper insights in the possibilities and limits
of regredient narration.

4. Comparison and Combination of the Differences

between Fact-sequence-time and Story-time

In the longitudinal dimension and with regard to time direction, the dif-
ference between fact-sequence-time and story-time is restricted to the
case of regredient story-time because within the longitudinal dimen-
sion, fact-sequence-time always has a direction. In the latitudinal and
the altitudinal dimension, we can observe a difference between the re-
spective fact-sequence-time and each of the two opposing directions of
story-time, as within these, fact-sequence-time is without own direc-
tion. However, if on the level of story-time these dimensions appear in
combination - and that is not infrequent - then there always exists, lo-
gically speaking, a specific difference of direction between fact-
sequence-time and story-time. And in addition, both levels of time al-
ways exhibit, as we remember, differences in quality. For a story-time
is, at least compared to objective time - or to the variety of subjective
fact-sequence-times that comes close to objectivity - always the
product of a selection.
Story-time and Fact-sequence-time 169

Baumgartner, Klaus & Wunderlich, Dieter (1969). 'Ansatz zu einer Semantrk des deut-
schen Tempussystems,' m: Der Begriff Tempus - eine Ansichtssache? Dilsseldorf:
Schwann, 23^19.
Chatman, Seymour (1978/1983). Story and Discourse. Narrative Structure in Fiction
and Film. Ithaca: Cornell UP.
Erhch, Victor (1955). Russran Formalrsm. Hrstory - Doctrine. 's-Gravenhage: Mouton.
Genette, Gerard (1971). 'Time and Narrative in A la recherche du temps perdu; in: J.
Hillis Miller (ed.): Aspects of Narrative. New York, London: Columbia UP.
Genette, Gerard (1983). Nouveau discours du recti. Paris: Ed. du Seuil [English transl.
1988: Narrative Discourse Revisited. Ithaca: Cornell UP].
Harweg, Roland (1976). 'Aspekte als Zeitstufen und Zeitstufen als Aspekte,' m: Lin-
guistics 1*1: 5-2*.
Harweg, Roland (1977). 'Zeitstufensysteme,' m: Orbis 26: 189-224.
Harweg, Roland (1979a). 'Inhaltsentwurf, Enahlung, Inhaltswiedergabe. Zum fiktions-
theoretischen Doppelstatus fiktionaler Enahlungen,' m: W. Frier & G. Labroisse
(eds.): Grundfragen der Textwissenschaft (= Amsterdamer Beitrage zur neueren
Germamstik. 8). Amsterdam: Rodopi, 111-130.
Harweg, Roland (1979b). 'Sind Richardsons Pamela und Fieldings Shamela ein und
dieselbe Person? Ein Beitrag zum Problem der Anzahl fiktiver Welten,' in: Poetica
11: 348-368.
Harweg, Roland (1980). 'Meta-assertonsche, meta-propositionale und meta-ontologi-
sche Aussagen. Em Beitrag zur Typo- und Textologie metakommumkativer Rede,'
in: Folia Linguistica 14: 283-328.
Harweg, Roland (1987). 'Probleme der Perspektivik in der Deutung von Perfekt und
Pratentum im gesprochenen Neuhochdeutsch,' in: P. Camsius (ed.): Perspektivitat
in Sprache und Text. Bochum: Brockmeyer, 152-182.
Harweg, Roland (1991). Tndividuum, Allo-Individuum, Archi-Individuum. Bezeich-
nung, Perspektivik, Substitution', m: K. Eimermacher & P. Gnybek (eds.): Zeichen
- Text-Kultur. Bochum: Brockmeyer, 187-212.
Heusler, Andreas (1902). 'Der Dialog in der altgermamschen enahlenden Dichtung',
in: ZeitschriftfurdeutschesAltertum 46, 189-284.
Hilt, Ernst (1923). Das Formgesetz der epischen, dramatischen und lyrischen Dich-
tung. Leipzig, Berlin: Teubner.
Lammert, Eberhard (1955). Bauformen des Erzahlens. Stuttgart: Metzler.
Mann, Thomas (1958). Erzahlungen. Frankfurt/M: Fischer.
Mann, Thomas (1998). Death in Venice and Other Stories. London: Vintage Books.
Mann, Thomas (1986). Tagebucher 1944-1.4.1946, ed. by Inge Jens. Frankfurt/M: Fi-
Mendelsohn, Peter de (1975). Der Zauberer. Das Leben des deutschen Schriftstellers
Thomas Mann. 1. Teil: 1875-1918. Frankfurt/M.: Fischer.
Mendilow, AA. (1952). Time and the Novel. London, New York: Nevill.
Miiller, Giinther (1947/1968). 'Die Bedeutung der Zeit in der Erzahlkunst', in: G. Mai-
ler (1968). Morphologische Poetik. Gesammelte Aufsatze. Darmstadt: WBG, 247-
268 [English transl. in this volume].
170 Roland Harweg

Miiller, Gilnther (1948/1968). 'Erzahlzeit und erzahlte Zeit,' in: G. Milller (1968). Mor-
phologischePoetik. Gesammelte Aufsatze. Darmstadt: WBG, 269-286.
Reichenbach, Hans (1947/1966). Elements of Symbolic Logic. London: Collier-McMil-
lan, New York: The Free Press [§ 51 in this volume].
Stanzel, Franz K. (1979). Theorie des Erzahlens. Gottrngen: Vandenhoeck & Ruprecht
[English transl. 1984: A Theory of Narrative. Cambridge: Cambridge UP].
Stierle, Karlhernz (1971/1975). 'Geschehen, Geschichte, Text der Geschichte,' in: K.
Stierle (1975). Text ah Handling. Perspektiven einer systematischen Literaturwis-
senschaft. Munich: Fink, 49-55.
Todorov, Tzvetan (1966). 'Les categories du recit litterarre', in: Communications No. 8,
125-151 [English transl. 1980: 'The Categories of Literary Narrative', m: Papers
on Language and Literature 16: 3-36].
Tomasevskij, Bons (1925/1965). 'Thematique,' m: T. Todorov (ed.): Theorie de la in-
ternum. Textes des formalistes russes. Paris: Ed. du Semi 1965, 263-307 [English
transl. of this chapter from the 1928 ed.: L.T. Lemon & M.J. Reis (eds.): Russian
Formalist Criticism. Lincoln: U of Nebraska P, 61-95].
Zielmski, Thaddaeus (1901). 'Die Behandlung gleichzeitiger Ereigmsse rm antiken
Epos,' in: Philologus , suppl. 8, 405^149.

The Temporality Effect

Towards a Process Model of Narrative Time Construction


When Roland Barthes coined the term reality effect he intended to

highlight realism as a product of certain rhetorical strategies by which
narratives influence their readers to process a fictional representation
as the depiction of a 'real' world.1 One of these strategies is to supply
apparently non-topical information that seems to have no bearing on
the core issues at stake. This rhetoric 'overkill' creates the impression
that, just as in our real-world encounters, we must filter out the relevant
from the trivial in the Active world, thus obliterating the fact that
everything represented in a piece of art is of course there by design,
and not by chance.
Persuading a reader to process a fictional narrative as real rather
than made up is one thing - making a narrative evoke in us a sense of
temporality and thus experience time is, as Ricoeur and others have ar-
gued, in all likelihood an even more fundamental function of narrative
representation.2 In the following I want to lay the foundation for a pro-
cess model of narrative time construction that can make this function
more transparent. The model itself will only be sketched out rather
briefly as a more concrete application of it has already been presented

Barthes, Roland (1989). The Rustle of Language. Berkeley: U of California P.
English language contributions to narrative theory and narratology sometimes use
the terms 'fictional', 'Active' and fictitious' as synonyms. This terminological
laissez-faire, however, encourages a confusion of representational and ontological
predicates. In the following I will therefore adhere to the more consistent
definitions formulated in Schmid (2010): 'Fictive = a property of elements (time,
space, situations, characters, actions) contained in the represented world of fictional
works' and 'Fictional = a property of representations of fictive worlds'; Schmid,
Wolf (2010). Narratology: An Introduction. Berlin/New York: de Gruyter, 245.
The terminological opposites are thus fictive vs. real (ontological distinction) and
fictional vs. factual (representational distinction).
172 Jan Chnstoph Master

elsewhere.3 In this previous work it formed the conceptual basis for a

computational narratology-onented approach toward the analysis of
narrative time constructs. Correspondingly, my presentation focused
mainly on the design principles and logic of three experimental soft-
ware applications: one for marking and extracting relevant text ele-
ments in a narrative, a second for analysing this mark-up data in order
to re-construct the temporal order in narrated events by way of an al-
gorithm-driven process of analysis and recombination of textual seg-
ments, and a third for creating some visual output that simulates the dy-
namic of the human reader's construction and experience of temporal-
One shortcoming of this computational simulation was the lack of a
more substantial philosophical foundation. While it was able to demon-
strate how narratively organised and communicated information motiv-
ates and aids readers to generate the mental image of a chronologically
organised world it could not sufficiently explain the underlying notion
of time consciousness. However, though this lacuna will be the focus
of my following deliberations I must make clear at the outset that I can-
not present a discussion of time sui generis: my discussion ultimately
serves a narratological purpose, namely to explain how time works in
and by narrative texts on the level of the elementary mental images
which are represented by words, and embedded in grammatical and
textual structure.
Philosophical discussion of the thematic complex 'Time and Tem-
porality' has a long tradition; to offer an overview or only a sketch of
its entirety presents even philosophers with considerable problems in
light of the historical depth of the debate - it reaches back to before So-
crates - and the range of methodological variants.4 However, a consid-
eration of only a few selected contributions to this field already allows
us to recognise a fundamental difference between two approaches: Psy-
See Meister, Jan Chnstoph (2005). 'Tagging Time in Prolog. The Temporality
Effect Project,' in: Literary and Linguistic Computing 20: 107-124. The software
developed in the related project is available for download at http://www- [last seen: 10.11.2010]. For a significantly
more elaborated Al-based approach see Inderjeet Mam's important The Imagined
Moment. Time, Narrative and Computation. Lincoln: Nebraska UP 2010, as well as
Mam's contribution to the current volume.
An instructive overview is provided by Sklar, Lawrence.: 'Time,' in: Routledge
Encyclopedia of Philosophy (2000) <
sophy/articles/entry> (11.03.2002. Access requires a license). However,
existentialist approaches are left largely to one side. - A more comprehensive
bibliography can be found in Macey, Samuel L. (1991). Time. A Bibliographic
Guide. New York, London: Garland.
The Temporality Effect 173

chologically and empirically directed philosophy about 'time' primar-

ily pursues the question of the logic of sensory, mental and cognitive
processes, as a consequence of which we manage to perceive and ex-
perience temporal ordering. Metaphysically or logically oriented philo-
sophy, on the other hand, may not be able to neglect the aspect of how
time is perceived, but does not discuss principally how empirical per-
ception of time comes about, focusing rather on how to conceive of the
nature of time and temporality, and what it is at all possible to know
and say about it: ultimately, then, the question of what, and under what
conditions, is 'time'. 5
It seems obvious with which type of time-philosophy a given philo-
logical project will have the greater affinity: hermeneutically oriented
literary study will prefer the metaphysical-logical perspective, whereas
descriptive literary study, primarily interested in the theory of narrative
or reception, will want to link up with investigations into the processes
constitutive of human time consciousness - and the latter all the more
decisively if it is working towards a computer-supported model and
thus a formal description of narrative phenomena, which, in turn,
makes it necessary to operationalise philosophically founded categories
for the purpose of describing texts and their reception. However, when
examined more closely, this obvious-seeming systematic division does
not hold. The formal description of elements and the modelling of pro-
cesses that lead the recipient to constitute conceptions of time within
the reception of narrative texts also and especially demand, according
to our understanding, at least a rough consensus on what we even mean
when we talk about 'time'. With that, however, we are already at the
core of the problem - for it is not just in subjective opinion, but more
fundamentally in judgment that our perception and thinking are ex-
pressed. Acts of judgment, however, are in themselves subject to
'time', as Kant pointed out in his Critique of Pure Reason. In the
chapter on Transcendental Logic he comments 'On the Schematism of
the Pure Concepts of Understanding'6:

On the systems of the time-philosophical discussion, see Le Poidevm, Robin: 'The
Experience and Perception of Time,' in: E.N. Zalta (ed.): The Stanford Ency-
clopedia of Philosophy (Winter 2009 Edition), <
archives/win2009/entnes/time-expenence/ (11.11.2010)> Attempts to mediate
between the two threads of the debate can be found in Le Poidevm, Robin;
MacBeath, Murray, eds. (1993). The Philosophy of Time. Oxford: Oxford UP. A
systematic study of the connection between the expenence of time and
metaphysical concepts of time is provided by Mclnerney, Peter K. (1992). Time
and Experience. Philadelphia: Temple UP.
Kant, Immanuel: 'On the schematism of the pure concepts of understanding,' m:
174 Jan Chnstoph Master

Now one sees from all this that the schema of each category [presupposes
and conceptualises; JCM] time itself, as the correlate of the determination
of whether and how an object belongs to time. The schemata are therefore
nothing but a priori time-determinations in accordance with rules, and
these concern, according to the order of the categories, the time-series, the
content of time, the order of time, and finally the sum total of time in re-
gard to all possible objects.7

In the preceding exposition, Kant had argued that all acts of reason can
be reduced to judgments, reason thus being defined as the 'ability to
judge'. This 'judging' does not constitute a direct verdict on individual
sensory (empirical) intuition, but rather a 'cognition through concepts'
- that is, a judging on the basis of an abstract, generalising conception,
in which are gathered together the common features of all objects fall-
ing under the same conceptual definition. In view of a plate, for ex-
ample, I reach the judgment, 'This is a plate', by comparing the con-
crete, empirically given intuition with a 'concept' of a plate as an ob-
ject, which, inter alia, has the general definition 'circular'. This 'sub-
sumption of an object under a concept'8 is, according to Kant, evidently
unproblematic insofar as the analogy itself can be empirically ob-
served, as with the example of the plate, whose geometrical definition
as 'circular' does clearly come to the fore in the concrete object.
However, things are different with the judgmental application of the
so-called 'pure concepts' - this is, those a priori terms which have no
empirical analogue, but rather establish the possibility of empirical (ex-
ternal) and sensible (internal) intuition: space and time. How should

Critique of Pure Reason. First edition 1781. Quotations in the following are taken
from the Cambridge-Edition of the works of Immanuel Kant (see Fn 7).
Kant, Immanuel (1997). Critique of Pure Reason. Translated and Edited by Paul
Guyer, Allen W. Wood. Cambridge: Cambridge UP (= The Cambridge Edition of
the Works of Immanuel Kant), A 145, 275-276. Hereafter cited as Kant CPR. -
The quotation reads in full: 'Now one sees from all this that the schema of each
category contains and makes representable: in the case of magnitude, the generation
(synthesis) of time itself, in the successive apprehension of an object; in the case of
the schema of quality, the synthesis of sensation (perception) with the
representation of time, or the filling of time; in the case of the schema of relation,
the relation of the perceptions among themselves to all time (i.e., in accordance
with a rule of time-determination); finally, in the schema of modality and its
categones, time itself, as the correlate of the determination of whether and how an
object belongs to time.' - Kant CPR, A145, 275-276. The second edition of the
Critique from 1787 takes on the entire first part unaltered.
Kant CPR, 271 -A137/B176.
The Temporality Effect 175

one imagine mediation between this type of pure concept and an object
(say, the fact of the limited duration of a plate)? The 'mediating repres-
entation' required here, i.e. the abstract tertium comparationis of pure
concepts and objects of both empirical and sensible intuition is, accord-
ing to Kant, the so-called 'transcendental schema'. This schema, simil-
ar to the image of an object created in us, is not a product of the ima-
gination (so, again, 'transcendental' and not objectively given). On the
other hand, nor is it a pictorial representation, but rather the 'represent-
ation of a method' for representing 'the image itself 'in accordance
with a certain concept'.9 However, this transcendental schema, which
establishes in the first place the methodical nexus between the forms of
external and internal intuition - space and time - developed by Kant in
his Transcendental Aesthetic, is that of time-determination. This is be-
cause our reason must, in order to be able to judge at all, while simul-
taneously securing its own identity by linking distinct representations,10
necessarily undertake temporal time-determinations of itself as well as,
on the other hand, assigning time determinations to occurrences in or-
der to be able to observe them as distinct. It is therefore time-determin-
ation that is common to all three - the intuitions of objects, the con-
cepts and categories applied to these, and judgmental reason itself- as
a transcendental schema." Kant deduces:

Time, as the formal condition of the manifold of inner sense, thus of the
connection of all representations, contains an a priori manifold in pure intu-
ition. Now a transcendental time-determination is homogeneous with the
category (which constitutes its unity) insofar as it is universal and rests on
a rule a priori. But it is on the other hand homogeneous with the appear-
ance insofar as time is contained in every empirical representation of the
manifold. Hence an application of the category to appearances becomes
possible by means of the transcendental time-determination which, as the

Kant CPR, 273 -A140/B179.
This doubled basis between the capability for synthetic judgment and the identity of
the judging subject may not, however, as Kant makes explicit in his critique of the
so-called 'Paralogisms', be taken as an indicator of objective I-identity.
This simplifying summary leaves unexamined Kant's further conclusions about the
distinction between empirical and transcendental apperception. The unity of
consciousness necessary for synthetic judgments cannot, according to Kant, be
determined by time-bound empirical apperception. This unity can be founded only
in transcendental reference to a noumenal, atemporal subject; see Chamberlain,
Jane (1998). 'Thinking Time: Ricoeur's Husserl in Time and Narrative; in: Minerva
-An Internet Journal of Philosophy, Vol. 2,
176 Jan Chnstoph Master

schema of the concept of the understanding, mediates the subsumption of

the latter under the former.12

Kant's intention of confirming his so-called 'Transcendental Doctrine

of the Power of Judgment' on this conception of the transcendental
schema has been - we will examine this briefly at a later stage - criti-
cised in particular by Husserl who, with reference to 'time' and the ap-
perceiving subject, postulated an unbridgeable divide between phe-
nomena and noumena, and called Kant's postulate of the 'transcendent-
al ideality of time' into question.13 However, we will leave out this, in
the strictest sense, philosophical debate for now and ask instead a sig-
nificantly less demanding, or perhaps, more naive question: can one
take Kant's reference to the unavoidable fixedness in time of our per-
ceptions and judgments, as well as his thesis of the principally tran-
scendental composition of time itself - because of which it cannot be
made accessible to our direct intuition - into account in the examina-
tion of the subject area of narratively evoked conceptions of time? This
much seems certain: nobody will want to dispute that each individual
act of our behaviour, our perception and our cognition takes places 'in'
time and can thus be determined for us ourselves in its relationship to
preceding, simultaneous and subsequent acts. We virtually cannot ima-
gine that one of our acts would have no time-determination and would
therefore be subject to an ontology that does not know 'time' - which
also counts for the reading of literature and the formation of mental im-
ages of the things being narrated. What, however, makes the phe-
nomenon 'time' - whether from an everyday, pragmatic perspective or
from the perspective of literary-theoretical and narratological question-
ing - uncommonly difficult to describe is its abstractness, the charac-
teristic from which Kant begins when he discusses time as a 'pure

Kant CPR, 272 -A138/139/B177/178.
Kant formulates his postulate as follows: 'But, on the contrary, we dispute all claim
of time to absolute reality, namely where it would attach to things absolutely as a
condition or property even without regard to the form of our sensible intuition.
Such properties, which pertain to things in themselves, can never be given to us
through the senses. In this therefore consists the transcendental ideality of time,
according to which it is nothing at all if one abstracts from the subjective conditions
of sensible intuition, and cannot be counted as either subsisting or inhering in the
objects in themselves (without their relation to our intuition).' - Kant CPR, 164 -
A 35/36/B52.
The Temporality Effect 177

We cannot say all things are in time, because with the concept of things in
general abstraction is made from every kind of intuition of them, but this is
the real condition under which time belongs to the representation of ob-
jects. Now if the condition is added to the concept, and the principle says
that all things as appearances (objects of sensible intuition) are in time, then
the principle has its sound objective correctness and a priori universality.14

Our sensible intuitions of both the real (empirically experienced) and

the represented (narrated) world are unavoidably temporally ordered,
though time is itself the form of internal intuition. However, 'time' as
such evades empirical observation, irrespective of whether structuring
(our apperception of) a real or a Active 'world'. This is the essential
difference from space. The spatial distribution of distinct empirical ob-
jects can, as is well known, be described in more than just the abstract.
Rather, one can make it observable and thus also represent it as a mod-
el, for example in the analogy of a three-dimensional picture. The tem-
poral distribution of phenomena, on the other hand, can be intersubject-
ively communicated only via an abstract formal notation such as a mu-
sical score or the numerical coordinates of a chronological-calendn-cal
system. However, it cannot be simulated or illustrated by analogy:
'time' in itself, or indeed a specific temporal structure, can be experi-
enced only by the repeated performance of acts of perception or beha-
viour, of which music is perhaps the most convincing example.
The philosophical problems and umquity of 'time', and its differ-
ence to the form of space, become particularly distinct when we relate
them to the concepts 'event' and 'object'. Against the backdrop of a
narratological interest and thus an approach directed primarily at the
subject area of eventful narration, it is the examination of the specifics
of the temporal dependency of events that offers us a substantial way in
for our understanding. The most pertinent of these specifics are sum-
marised by Casati and Varzi in the Stanford Encyclopedia of Philo-
sophy as follows:

Although not undisputed, some standard differences between events and

physical objects are commonplace in the philosophical literature. First,
there is a difference in mode of being: material objects such as stones and
chairs are said to exist; events are said to occur or happen or take place
[Hacker 1982a]. Second, there are differences in the way objects and events
are said to relate to space and time. Ordinary objects have relatively clear
spatial boundaries and unclear temporal boundaries; events have relatively
unclear spatial boundaries and clear temporal boundaries. Objects are invi-
Kant CPR, 1 6 4 - A 35/B52.
178 JanChristophMeister

diously located in space - they occupy their spatial location; events tolerate
co-location [Quinton 1979, Hacker 1982b]. Objects can move; events can-
not [Dretske 1967]. Objects are continuants - they are in time and they per-
sist through time by being wholly present at every time at which they exist;
events are occurrents - they take up time and they persist by having differ-
ent parts (or "stages") at different times [Mellor 1980].»

Correspondingly, the phenomenology o f events provides a substantial

starting point for our literary-theoretical and narratological modelling
and theory formation on the complex 'temporality o f the narrated / o f
narration'. The standard narratological concept of event, however,
refers mainly to occurrences taking place - or rather, represented as
taking place - in a Active world. In this context we observe a qualitat-
ive distinction between unmarked events that take place, but do not
possess a pronounced thematic significance with regard to the overall
conception of the narrative on the one hand, and on the other hand a
type of marked, highlighted event that is crucial for our overall under-
standing of the narrative: i.e. an event that amounts to significantly
more than just a normal and expectable occurrence.-
Yet in the context of our current deliberations a third dimension of
event becomes particularly important, namely that of so-called 'percep-
tion events' by virtue of which both the narratively represented objects
in a world, and the processes manifested in them, enter our conscious-
ness as representations." With that, although we are taking on from
Kant the basic idea of an, in the widest sense, 'epistemological' exposi-
Casati, Roberto, Varzi, Achille: 'Events,' in: E.N. Zalta (ed.). The Stanford Encyclopedia
of Philosophy (Fall 2010 Edition), <http://plato.stanford.-edu/archives/fall2010/-
entnes/events/> (11.11.2010). The entry actually states the contrary, namely that 'events
have relatively clear [sic!] spatial boundaries and unclear [sic!] temporal boundaries.' This
is inconsistent with the previous argument and has therefore been corrected in the
For a comprehensive overview on current narratological definitions see Hiihn, Peter:
'Event and Eventfulness,' in: P. Hiihn et al. (eds.): the living handbook of narratology.
Hamburg: Hamburg University Press,
title=Event and Eventfulness&oldid=753> (10 Nov 2010).
Perception events systematically correspond to the so-called 'discourse events'
defined in Meister, Jan Chnstoph (2003). Computing Action. A Narratological
Approach. Berlin, New York: de Gruyter, 4 5 ^ 8 . In general these events concern
the transformation of the knowledge status of a consciousness observing the Active
world. Discourse events are thus, on the one hand, delimited from the fundamental
'object events' (= transformations to the state of objects in the Active world), but,
on the other hand, are positioned systematically below the 'outrageous events',
which are always identifiable first in the interpretative ascnption of relevance, in
the sense of an emphatic event concept.
The Temporality Effect 179

tion of the phenomenon of time, in reference to its constituents, we are

also, on the other hand, orienting ourselves methodologically on the
phenomenological approach of his adversary Husserl, following whom
we can leave aside the question of the ontic status of the perceived ob-
jects and the characteristics assigned to them.
Against this background the 'temporality' of narrative texts must
thus be viewed, on the one hand, from the perspective of a Active event
time, and, on the other, from that of the factual dependency on time of
literary information processing. This dualism of event time and dis-
course time does not just express, in a specific way, the capacity for re-
flexive phrasing fundamentally anchored in language in the sense of
'speaking-about-speaking', of which narrative texts in the autothematic
mode (can) make facultative use. This dualism is rather, insofar as it
refers to 'time', an unavoidable and irrefutable one, which is always
already a given. Whether the author, the narrator or some other Active
narrative figure wants to bring it up or not: by narrating, they are
already speaking to us in, and therefore about, 'temporality' - theirs as
well as ours. In a narratological perspective this necessitates a threefold

- One, how at all - that is, with which principal conditions and
prerequisites - is it possible to achieve an idea of the temporal
order in a Active world through the reading of narrative texts?
- Two, how do we achieve a consciousness of the temporality of
that mediative process itself, by virtue of which our pictorial
representation of this Active world and its internal temporal or-
der emerge?
- Three, to what extent does the temporality of the mediative
process - independent of whether it comes to our conscious-
ness or not - have a retroactive effect on our delineation of the
Active world's chronology?

'Temporality', in terms of the Active world's always already 'doubled'

being-in-timeness and the reception process (re^constructing it, thus
attains an ontic dignity and primacy over the category of space. While
a narrative text is given us materially 'in space' (e.g. in the form of a
book), our conception of the spatial dimension of its 'world' often re-
mains very abstract. Narrated space need not have a common border
with our own real, existential experience of space. By contrast, the fact
that the very process of reading, (re-)construction and interpretation ne-
cessarily takes place in real time, substantially determines our delin-
eation of a Active chronological ordering of that Active world both
180 Jan Chnstoph Master

formally and as regards content. This interdependency, particularly the

dynamic feedback effects between phenomena of the discourse-tempor-
al and object-temporal ordering, and their textual guidance, must re-
ceive special attention in a narratological model of the temporality phe-
nomenon. A decisive failing of previous approaches, which (in one
way or another) all fall under the methodological paradigm of the 'time
of narration / narrated time' (Erzahlzeit I erzahlte Zeit) distinction, is
the ontologising premise, often unreflectedly concomitant with this dis-
tinction, of a distinct temporality of Active worlds. In contrast, our per-
haps most fundamental thesis is that there can be no separate 'Active'
time in this sense - since, seen from the logic of how they are consti-
tuted, all stretches of time form an epistemological continuum.
As a narratological undertaking, our model claims relevance for a
very narrowly defined subject area: that of fictional prose.18 This is also
a genre that is decisively shaped by engagement with the themes of
'time' and 'temporality', on the levels of both form and content. 'Time'
and 'temporality' are, beyond this, epochal themes of philosophy. In
light of this combination of content-formal and general intellectual-his-
torical dignity, it is tempting for the literary theorist, in particular, to
work from the basis of time-philosophical assumptions from textual
theory and analysis, and to 'overshoot' to a general interpretation or
even to a philosophising individual thesis. The perhaps best-known ex-
ample of this is provided by Ricoeur (already mentioned above), with
his emphatic evaluation of narrating literature as the privileged medium
of human experience of time.- I will not be able to do justice to those
kinds of philosophical ambitions here; too great also would be the risk
of submitting another contribution to the dilettante 'Literary Philo-
sophy of Time', as criticised by Gregory Currie.- The method of my
The concept of narrativity itself is of course anything other than unproblematic; see
Sternberg, Meir (2001). 'How Narrativity Makes a Difference,' in: Narrative 9.2:
115-122. We will therefore formulate the following working concept: We will treat
as 'narrative' that section of prose texts whose content level is constituted
dominantly by linguistically represented 'object events' in the sense of the
definition by Meister (2003).
On the criticism of Ricoeur, see Chamberlain, Jane (1998). 'Thinking Time:
Ricoeur's Husserl in Time and Narrative; m: Minerva - An Internet Journal of
Philosophy 2 (see footnote 11).
Currie, Gregory (1999). 'Can There Be a Literary Philosophy of Time?' m: J.
Butterfield (ed): The Arguments of Time. Oxford: Oxford UP, 43-63. Curne's
criticism is aimed particularly at the attempts of Ricoeur and Bakhtin to model a
specifically literary form of time-experience. Currie argues, in opposition to them,
that there can be no special literary (or otherwise representational) form of time-
experience, but only one: the real one.
The Temporality Effect 181

approach is rather largely that of an inductive-descriptive process, at-

tempting to grasp the constituting logic of the 'temporality effect' step-
by-step and 'bottom up'. However, I will try to sketch at the outset of
our considerations the time-philosophical starting point that will de-
termine the direction of our narratological approach. We will start with
Augustine and thus with the classic of the philosophy of time. There-
after, Husserl's philosophy of time and his conception of the 'time win-
dow' will be examined. The third section is dedicated to McTaggart in
whose fundamental differentiation of two distinct time series we will
again encounter the (not only) Augustiman division of objective time
and subjective time experience. McTaggart's approach offers a possib-
ility that has hitherto been, as I believe, insufficiently taken into con-
sideration, namely to place the distinction, widespread in the theory of
narrative, between discourse time and Active narrative time onto a
more solid philosophical basis. In the fourth section, I will examine
two examples from the reception of McTaggart - Le Poidevin's at-
tempt to depict McTaggart's model in Possible Worlds semantics, and
Bien's constructive interpretation of the model (which is nonetheless
philosophically and not narratologically motivated). The fifth section
shows how McTaggart's model can be mapped onto narrative texts and
their processing; the following discusses what I term 'temporal operat-
ors' as the main devices by which narratives control our designing of
temporal constructs.
What will remain hidden in the following is the existentialist debate
on time that is particularly bound up with the name Heidegger. This is
not to deny the eminent significance of Heidegger's deliberations on
the fundamental role of time as an existential horizon for human be-
ings. It is simply that the themes broached there are outwith the remit
of primarily text-empincally oriented narratology.

. The Augustiman Paradigm

In his Confessiones Augustine formulated a simple question: 'What is

time?'- The problem turned out to be profound, and so is its discussion
by Augustine. Heidegger, for example, counted the Confessiones -

Saint Augustine (2008). The Confessions. Translated with an Introduction and Notes by
Henry Chadwick. Oxford: Oxford UP (= Oxford World's Classics). See also the
translation by A. Outler, Augustine: Confessions. <
confessions/confessions-bod.html> (11.11.201029.11.01) or: The Confessions of
Augustine. An Electronic Editon. Text and Commentary by James J. O'Donnell
<> (11.11.2010).
182 Jan Chnstoph Master

alongside the contributions of Aristotle and Kant - as one of the 'the

three groundbreaking reflections on the nature of time' in Western
philosophy.- However, in the debate on the philosophy of time, it is
not the Confessions in their entirety that are examined, but rather the
theses formulated, in the 11th book, on the relativity of human percep-
tion of time and the unreality of time. The autobiographical-theological
context of these meditations is seldom taken into account in this field.23
From the perspective of the theologically interested Augustine com-
mentators, on the other hand, the abrupt shift from autobiographical
tale to speculative exposition, at the end of the 9* book, has been seen
as problematic. This apparent change of genre made problematic the
self-contained interpretation of the Confessiones as a didactic, autobio-
graphical and confessional text."
Both selective approaches (i.e. the philosophically motivated blend-
ing out of the autobiographical element, and the theologically motiv-
ated blending out of the speculative part of the text), however, do not,
as Fischer (2000) has shown, do justice to the thoroughly sequential to-

Heidegger, Martin: 'Des hi. Augustmus Betrachtung ilber die Zeit. Confessiones lib
XI', quoted by Fischer (2000: XI). Fischer, for his part, quotes from a hitherto
unpublished copy of a lecture of Heidegger's in the Abbey of St Martin, Beuron,
26.10.1930, which will appear in volume 80 of Heidegger's complete works; cf
Fischer, Norbert (2000). 'Einleitung,' in: Aurelius Augustinus: Was ist Zeit?
Hamburg: Meiner, XI-LXIV).
It is particularly his precedence in view of existentialist philosophy of time that is
often emphasised; so for example in Brann, Eva T.H. (1999). 'Augustine As
Phenomenologist: A Time Diagram,' m: Hopkins, B.C. (ed.). Phenomenology:
Japanese and American Perspectives. Dordrecht: Kluwer. Also in Herrmann,
Fnednch-Wilhelm von (1993). 'Augustinus und die phanomenologische Frage
nach der Zeit,' m: Philosophisches Jahrbuch 100: 96-113.
Compare O'Donnell's critical meta-commentary: 'The last three books of the
Confessions are the pnncipal obstacle to the work's reputation for greatness in the
literary, as well as the psychological or theological, order. One scholar recounted
no less than nineteen different theories that had been devised to explain their
presence and their relation to the rest of the work, then proceeded to add his own.'
In: O'Donnell, James O. (2001) >
jod/twayne/aug5.html> (11.11.2010) . O'Donnell is here referring to Grotz, Klaus
(1970). Die Einheit der 'Confessiones'. Tubingen: Diss. U of Tubingen. - On this
question see also the section 'Zur Diskussion urn das hteransche Genus der
Confessiones' in Fischer's (2000: XXVff; Fn 22) introduction to the edition of the
Confessiones XI. Fischer refers, inter aha, to the example of the Wilhelm
Bornemann edition, which simply omitted books 10-12. See: Augustins
Bekenntnisse. In neuer Ubers. und mit einer EM. dargeboten von W. Bornemann.
Gotha 1888. (=Bibhothek theologischer Klassiker 12).
The Temporality Effect 183

tality of the Confessions." The apparent gap in the transition from the
ninth to the tenth and especially to the eleventh books of the Confes-
sions makes the fundamental complexity of the themes of 'time' and
'temporality' exemplanly clear, and show that Augustine had already
anticipated the basic idea of the Kantian argument: that of the unavoid-
able subjectivity of the empirical concept of time, which can be dis-
pelled only through a transcendental or theological explanation of time.
What Augustine reconstructs and relates in the first nine books of the
Confessions is not, after all, only the disinterestedly observed case of
someone else's biography, whose myriad individual moments and indi-
vidual events are ultimately smelted to a solid chain by the conscien-
tiousness and skill of the biographer. It is rather a case of his own life,
which is relating itself and thereby encounters the paradox of always
having to imagine itself standing outside that temporal concatenation in
order to be able to make it all possible to describe it own entanglement
in an ordered temporality. The tenth and eleventh books demonstrate
with the transition from the level of representation to that of reflection,
that this is a case of an, in the profoundest sense, 'time-conscious'
autobiography that has become aware of its own apona. In these cir-
cumstances, it can no longer exhaust itself in objectivising the linkage
of events represented as memory into the linearity of before - after.
Rather, Augustine understands every point in the imaginary chain he
See Fischer (2000; cf Fn 22). Fischer's general interpretation, developed from the
perspective of the eleventh chapter, sees in the autobiographical narration of Books
1-9 a practical realisation of Augustine's own fixedness in time, in Book 10 a
reflexive abstraction from the autobiography to problems of memory, which is
elucidated more thoroughly in Book 11 in terms of the problems of human
understanding of time and our own existential fixedness in time. In the central 11 *
book, Fischer sees three approaches to the clarification of the concept of time: an
attempted empirical inference failed for Augustine with the realisation that this
would lead to an unacceptable objectivising of time; the realisation that time is as a
distentio animi an effect of the 'actions of the spirit, in which the past, present and
future are realized' (ibid., p. LV) did provide, for Augustine, a plausible
explanation, but would also indicate the aporetic consequences of the temporal
'stretching' of the spirit. So Augustine came to the conviction that going beyond
the immanent distentio animi is possible only through transcendence and
unification with God. - Paul J. Archambault (1984. 'Augustine, Time and Auto-
biography as Language,' m: Augustinian Studies 15: 7-13) proposes the thesis that
Augustine 'establishes a theoretical basis for autobiographical narrative by
showing, through his excellent analysis of memory, that life is not serial but that it
can be entirely recalled in a single instant.' With that, the weighting is turned on its
heads; the Augustinian meditation on time would then be serving the
184 Jan Chnstoph Master

has created also as one which obeys the dynamics of positioning in a

second continuum - that of past-present-future. However, this recogni-
tion leads to a dilemma. Our acts of cognition are necessarily present
and take place in the 'now'; on other hand, we claim to be able to 're-
member' or even 'anticipate' ourselves: that is, to cogmtively process
something that clearly no longer has empirical existence at the time of
its thematisation. Cognition and perception thus come into conflict: if
time is unavoidable, as perception claims, then there is nothing that can
simultaneously be and not be; if, on the other hand, something can both
be and not be, as cognition demonstrates, then time has become invalid
as an organising principle.
Augustine's well-known solution consisted in concluding that all
time-determinations are unavoidably relative, and in formulating his
thesis of a 'threefold present' - a present which is not just selective, i.e.
which does not reduce itself to a given now-standpoint. Even the 'now'
is more than just one single indexical point in time; rather, it is consti-
tuted from a stretch of time formed of several such points, whose en-
tirety makes up the present. Embedded in this now-present, however,
are also mental representations, called up by memory or anticipated by
imagination, of - relative to the 'now' point in time - recent or pro-
spective stretches of time, which we believe 'objectively' to be situated
in the past or the future. So, both a remembered 'present of the past'
and an imagined 'present of the future' are immanent components of
the present we experience. The principal relativity, deduced in this
way, of all objectivising time-determinations, which should be ima-
gined as 'transcended' in a static, threefold present of the conscious-
ness, aims for Augustine at a theological point: in the fleeting moments
in which we engage reflectively with the fact of time's unreality, we
experience the best possible analogy to the principally time-transcend-
ent nature of the divine consciousness. Of course, this imagined time-
transcendence, at all comprehensible only via intense abstract reflec-
tion, delivers only a caricature of actual divine extra-temporality.
In this, Augustine does not merely argue, but rather demonstrates
with his Confessions that it is not the objectivising-histonographical
aspect of his autobiographical consciousness, but the aspect that reflex-
ively questions its own prerequisites that will be unavoidably confron-
ted with the quintessence of the theological or philosophical problem
of determinism. For him, the profound phenomenological and concep-
tual complexity of 'time' and 'temporality' thus rests on the fact that
every form of time determination necessarily contains a reflexive com-
ponent and refers back to the perceiving subject and his own fixedness
in time. Quite aside from the significance which this realisation has
The Temporality Effect 185

both theologically and philosophically, it is also fundamental to the

rather humbler narratological investigation of the constituting logic of
narratively evoked temporality effects.- The second Augustinian real-
isation also fundamental to our context is: the perception of 'time' in
terms of temporal extension cannot be reduced to the perception of dis-
tinct points in time from which a stretch of time is formed - because
this concept would call the identity of the perceiving subject into ques-
tion. If we think of our own self as remaining itself over time, we
clearly require the ability to synthetically observe distinct points in
time as a stretch or a time frame. This raises the question of by what
means and processes we perform precisely this temporal synthesis dur-
ing the reception of narrative texts, and to what extent it is dependent
upon empirically verifiable textual components.

2. Husserl's Phenomenological Approach: The 'Time Window'

The two problems raised by Augustine - firstly: how can we explain

time-determinations that retro-act upon the perceiving subject despite
being undertaken by the self?; secondly: what establishes our capacity
for the synthetic observation of distinct points in time as more compre-
hensive stretches of time or 'time frames' - are taken up in Husserl's
philosophy of time.
Husserl takes on the thesis, originating from Brentano (and indir-
ectly from Kant), that the temporal determinations undertaken by us
express nothing about the objective characteristics of the objects of
perception. Objective time and subjective time-consciousness must be
fundamentally divided, and Husserl, in keeping with his phenomenolo-
gical approach, deals primarily with the 'suspension of objective
time'" Thus, time does not interest him from the perspective of the
In contrast to the interpretation sketched here, see O'Donnell's exposition, which
raises the subjectivity of time-expenence and sees this as contrasted by Augustine
with the extra-temporal existence of God: 'Time is inherent in the created intellect,
a category for describing the apparent transience and rmpermanence of reality.
Time is not even a created thing, for it is a creation of created things. Intelligent
created beings see the world around themselves in a framework of their own
invention, which they call time. This characteristic distinguishes their expenence
from that of their creator. God as creator sees all things simultaneously in a single
vision, perceiving process and change but, freed of experiencing those things in
temporal succession, he does not expenence time. The creator lives outside created
things and therefore, a fortiori, outside time. Time cannot be, Augustine concludes,
without created being' (O'Donnell 2001; cf Fn 24).
This is already the title of the first paragraphs in the introduction to the 'Lectures
186 Jan Chnstoph Master

perception of the characteristics, external to consciousness (in

Husserl's sense, 'transcendent'), of objective items but rather from that
of their consciousness-internal sensation; in other words, as a so-called
'piece of phenomenological data'. Three main tasks present themselves
for a phenomenological investigation: firstly, 'to describe the given
naive and scientific consciousness of time according to its sense (time
is taken by us to be what? temporal relations present themselves as
what? and what sorts of relationships are meaningfully intended in the
sense belonging to the representation of time, to this intuition of
time?)';- secondly, the description of this consciousness's content,
whereby sensory (empirical) and interpretative (cognitive) contents
have to be distinguished; thirdly, however, the

the exhibition of the particular cases in which 'adequate intuition of time' is

perhaps given; the exhibition of that which is quasi-temporal (duration, suc-
cession, and the like), which is not interpreted transcendently and 'object-
ively' as reaching out beyond itself and which asserts nothing with regard
to an 'objective time.' On the contrary, the quasi-temporal is interpreted im-
manently, that is, taken simply and just as it is, and makes up the proper
material that, as content of apprehension, underlies the interpretation that
constitutes objective time.29

In light of the over-arching aims of Husserl's argument, the most ex-

plosive part is contained in what is modestly referred to as the third
subtask. An individual 'adequate intuition of time' expresses nothing
reliable about objective time - our sensation of the simultaneity of two
phenomena, for example, may appear to us to be a substantive phe-
nomenon, but does not stand up to an, in the physical sense, 'objective'
investigation. Nonetheless, there is a route that leads from time phe-
nomena as experienced to time as an inter-subjective construct:

on the Phenomenology of the Consciousness of Internal Time' in the winter

semester 1904/05 in Gottingen. See Husserl, Edmund (1991). On the Pheno-
menology of Internal Time (1893-1917). Translated by John Barnett Brough.
Dordrecht: Kluwer (= Collected Works. 4); hereafter referred to as Husserl,
PhCiT..- On the notorious editorial history of the original lecture manuscript
worked on by Martin Heidegger and Edith Stern, which was first published in 1928
in the JahrbuchfurPhilosophie und phanomenologische Forschung (Vol. 9), see
alongside the translator's introduction to vol. 10 of the Husserliana also Wiehl,
Reiner, ed. (1981). Geschichte der Philosophie in Text und Darstellung Vol. 8:
20Jahrhundert. Stuttgart: Reclam, 89f
Husserl P/iCiT, 194.
The Temporality Effect 187

The 'sensed' temporal data are not merely sensed; they are 'charged' with
characteristic modes of apprehension, and to these in turn belong certain
claims and entitlements: to measure against one another the times and tem-
poral relations that appear on the basis of the sensed data, to bring them
into this or that objective order, and to distinguish various apparent and ac-
tual orders. What becomes constituted here as objectively valid being is fi-
nally the one infinite objective time in which all things and events - bodies
and their physical qualities, psyches and their psychic states - have their
definite temporal positions, which we can determine by means of a chrono-

However, truly objective time as a 'transcendent' structure entering

consciousness directly via experience (i.e., independent of all percep-
tion and sensation) is something that lies beyond the horizon of phe-
nomenological knowledge. That is why Husserl does not undertake the
deduction of immanent consciousness of time from the a priori of ob-
jective time, but instead: 'We seek to bring the a priori of time to clar-
ity by exploring the consciousness of time, by bringing its essential
constitution to light, and by exhibiting the apprehension-contents and
act-characters that pertain - perhaps specifically - to time and to which
the a priori temporal laws essentially belong.'31 Husserl's subject is
thus 'time' neither as an a priori form, nor as an objective content - it
is time as an immanent construct of consciousness. But how do we cre-
ate time as a construct on the basis of our sensations? With this ques-
tion, Husserl again picks up Brentano's approach. Brentano had recog-
nised that the sensation of duration and succession is not necessarily
created through the objectively given continuity and sequentiality of
the stimuli, but rather via associations generated in the consciousness
itself, or, as Husserl comments:

The fact that the stimulus endures still does not mean that the sensation is
sensed as enduring; it means only that the sensation also endures. The dura-
tion of sensation and sensation of duration are two very different things.
And this is equally true of succession. The succession of sensations and the
sensation of succession are not the same.32

Ibid., 7.
Ibid., 10. By these 'temporal laws', Husserl means the features of the unlinearily
directedness of the time arrow and the unambiguous determination of individual
points in time within a continuum.
188 Jan Chnstoph Master

Husserl develops this train of thought largely on the example of the

connection between notes and melody. A melody is something we per-
ceive as a temporally ordered sequence of distinct notes, of which each
has a certain duration. This means for our consciousness that, firstly,
we do not only register the note currently being heard, but can also re-
tain the preceding ones, and, secondly, that with each new note, all pre-
ceding and remembered notes experience a continuing modification -
they move further and further into the past. Up to now, however, this
explains only how we are able to integrate the individual notes in their
succession into a melody; what remains unexplained is how we sense
the duration of individual notes and how we can then reproduce this
duration in our consciousness by integrating the notes into the complex
sensation of a melody. The problem intensifies if we do not take as our
starting point the perception of a melody that consists of different
notes, but investigate the perception of a constantly changing (i.e. tem-
porally and qualitatively regular) object. Its perception, too, does, after
all, take place in a physiologically determined way, through a sequence
of individual acts of perception. The differences inherent in the percep-
tions, however, remain without any recognisable qualitative profile and
do not establish any possibility of creating a representation of duration
on the basis of regularly clocked acts of perception. And the problem
becomes even more distinct if we consider the ongoing perception of
an unchanging object - the sequence of apparently identical acts of
perception does not in any way justify a temporality of the perceived as
a phenomenon of temporal extension. Until now, we have been consid-
ering a pure succession of perceptions, but not the perception of suc-
cession, to say nothing of duration. Our consciousness is, at best, that
of a constant sequence of momentary perceptions of an object, some-
thing which Husserl calls 'original consciousness of time'. But how do
we achieve a corresponding 'reproductive consciousness of time'33,
which remembers not only the past object but simultaneously also the
entire complex of recent perceptions in which this object was originally
Husserl solves this problem by introducing a fundamental distinc-
tion. On the one hand are continuously modified temporal determina-
tions, which are assigned to phenomena not in an objective-qualitative,
but rather a modal and therefore subjective-relational sense (now, past,
future). On the other hand, there are areas of consciousness in which
phenomena become objectivised according to the mode of their appear-
ance in the intentional, retentional and protentional consciousness. At
the start of the perception process, the note is registered intentionally
The Temporality Effect 189

as a so-called 'note-now'. Already the perception of the present must

thus be imagined not as being of a single point but as the registering of
an extended time window, in which the objects directly bordering the
point are kept present in the mode of the 'primary memory'": thus it
encompasses also the 'just-past' and the 'directly-imminent'. It is only
on the border of the window that the perception's modal-temporal de-
termination is transformed to a true 'past'. Here, the perception be-
comes the object of the retentional consciousness, which no longer per-
forms a simply 'primary' memory function, but a 'reproduction of the
past'.35 In this retention is contained not only the past perception but
also the fact of its modal designation as past. Moreover: not only this
one, but also the perception assigned to the now, both appear in a new,
conscious temporal modification. In retention, all preceding instances
of retention are stored in the sense of a retention of retentions.- So, re-
tention is not the mere imagination and storing of the contents of a
now-perception that is directly in the past - it is always also the percep-
tion of the past as being past, the pre-past and its being pre-past.
For Husserl, perception itself encompasses all three areas: intention,
retention, pretention. It is neither restricted to a Active mathematical
point on a time line, nor does it adhere to the idea of a reality consist-
ing of a singular now-point. For Husserl,

these are sheer fictions and lead to absurdities. In phenomenology we do

not have to do with objective time but with the data of adequate perception.
This requires us to consider perceptions, with their appearing now, past,
and future, as given. Reduced, they yield the evident now, past, and future,
In place of an objectivising theory of time, which presumes the abstrac-
tion of the one-dimensional now and therefore a linear conception of
time as a sequence of such points, Husserl proposes a phenomenologic-
al conception characterised by a quasi-spatial concept of time. In that,
he starts from what he considers the evident fact that our perceptions
are not limited to a single point in time, but always encompass a so-
called 'time window' - a piece of evidence that for Husserl also counts
as an essential indicator of the 'givenness' of the self that Kant had cat-

Ibid., 170.
Ibid., 171.
Here, as in the following, we are leaving to one side the entire complex of the
protentional consciousness postulated by Husserl, the functions of which can be
thought of as analogous with those of retention.
Husserl PhCiT, 174; emphases in the original.
190 Jan Chnstoph Master

egoncally and irrevocably divided into an empirical and a transcend-

ental ego.38 That is the actual point of Husserl's theory of time: in that
the retentional consciousness does not only keep the 'image' of a past
note present in the memory, but also makes conscious the fact and the
measure of its pastness, it reflexively creates a consciousness of dura-
tion. Furthermore, with this act it simultaneously constitutes itself as
retentional consciousness; it thus constitutes itself and justifies the im-
manent conception of time. Using the example of time, consciousness's
ability to give evidence of itself is proved. - It is only with reference to
this wider context that the emphasis with which Husserl sums up can
be understood:

[...] obviously it must be claimed as something absolutely given itself that a

retention, in which what is just past in its unity with the now and the always
new now comes to absolute itself-giveness, already inheres in the
perception. If we scan the flowing tree, the tree becomes given in a
temporal form. And if we hear a bit of melody, we do not hear, merely
single tones, even less moments of single tones or mathematical tone-nows,
matching the now-points that could be abstracted in thought. We rather hear
enduring tones - specifically, tones combining into a tone-formation; and
we grasp this whole tone-formation as a formation that is steadily building
up and as that which is heard. And in the unitary regard continuously
directed towards it, we grasp the unity of the total perceptual appearance of
this tone-formation as something absolutely given itself. And if the whole
tone-phase is finished, retention still apprehends the just-having-been of the
total phase that has there elapsed; and it still apprehends the total perceptual
appearance in the manner of an appearance that has just been and that no
longer contains moments of actually present perception. The evidence here
concerns the just-having-been by means of which a relation of the object to
the flowing now is co-given; and the object cannot be detached from this
now. All of this is found in phenomenological reduction under its continual
suspension of present or past natural reality.39

What perspectives does Husserl's philosophy of time - which can here

be only portrayed in abbreviation - open up for our narcological in-
vestigation? At least two aspects seem worth considering. Husserl's ap-
proach delivers a philosophical basis for the plausibility of an (in a
wider sense) 'constructive' modelling of narratively evoked temporal
effects, whose emphasis is not on a planar reconstruction of an as-
sumedly 'objective' temporal structure of the ordo naturalis and the

This interpretation of Husserl's philosophy of time, in the sense of a critical reading
of Kant, follows the theses of Chamberlain (1998; cf. Fn 11).
Husserl PhCiT, 355; emphases in the original.
The Temporality Effect 191

concomitant ontologising stratification of distinct 'times', but rather on

the dynamics of the constituting process itself. Essential in this context
is particularly the indication that merely temporally segmenting acts of
perception does not in itself necessarily imply the perception of tem-
porality. Transferred to the subject area of a narrated world: the repres-
entation of a temporal order in the Active world does not arise neces-
sarily from the trivial fact that each word denotes another object or an-
other representation and that ever new quasi-perceptions of the mental
images evoked by these words are performed successively. If Husserl is
right, temporality is constructed by our consciousness only under the
presumption of a non-linear change of the content of perception - i.e.,
there where we experience acceleration, expansion and compression.
Moreover, we can experience temporality only when true retention (or
protention) is present, where it is not a case of the primary representa-
tion of an object, but where we, remembering, generate a representa-
tion of the perception of an object. Remembering in the sense of 'rep-
resenting an object in our fantasy [...] is having an apparition in the
present that possesses the character of representation.'- We also have
to add how a narrative text, particularly with regard to time, can awake
that consciousness of the representational character of the images
evoked by it - already one suspects that, in view of this, the conven-
tionalised use of tenses also plays a role, as does the explicit designa-
tion of Active objects and events by means of time-determinations:
both establish the distance between the perceived, which is a Active en-
tity, and the perceiving real consciousness of the reader, without which
a consciousness of its representational character cannot emerge.
Following Husserl we therefore have to understand temporality -
and also narratively coded temporality - as a phenomenon of difference
in two ways. The first aspect of this quality of difference is of a qualit-
ative nature: neither a sequence of unchanging nor of constantly chan-
ging perceptions enables us to become conscious of temporality; this
can be done only through a succession of quantitatively and qualitat-
ively discrete individual perceptions - experiencing time rests on the
dynamics becoming conscious, not on mere sequentially.41 The second
Ibid., 165.
This sense is also that taken by Meir Sternberg, who wants to trace the concept of
'narrativity' itself back to the temporal dynamics that result from the dichotomy
histoire/recit - i.e. the transformation of the natural temporality of the represented
happenings through the teleological (poetical-affectively, communication-strate-
gically or situational determined) temporality of the representation happenings. If
one follows this thought further, then isochronal narratives would be neither
'narrative' nor would they possess any 'temporality'. On this, see Sternberg, Meir
192 Jan Chnstoph Master

aspect of this quality of difference is an ontic one: a consciousness of

temporality can emerge only where we have a consciousness of the
represented character of past perceptions, which are represented in
memory. It is in this doubled sense that we will have to model narrative
'temporality itself as a form of representational consciousness'42 in the
context of our narratological research.

3. McTaggarts A/B/C-senes Theory

Considered in a narratological perspective, the philosophical ap-

proaches to time and time perception which I have discussed so far
may serve to highlight individual aspects of our experience and con-
ceptualisation of time through narratives. But neither of them offers a
comprehensive model that could be fully mapped onto the phenomenon
of narratively induced temporality constructs. Ironically enough, the
philosophical model that will take us further in this regard is the one
which denies the reality of time: that of McTaggart. Beside Heidegger,
McTaggart has exercised the greatest effect on the debate in the 20th
century. Yet aside from their common theme, the difference between
Heidegger and McTaggart could hardly be larger: McTaggart's argu-
ment is logical-analytical and not existentialist;43 the development of
(1992). 'Telling in Time (II): Chronology, Teleology, Narrativity,' in: Poetics
Today 13.3: 515ff. Also Sternberg (2001; cf. Fn 18).
HusserlP/iCiT, 164.
This designation refers pnmanly to the method not the subject of his philosophy. In
this regard, McTaggart, as a Hegel specialist, is rightly classified by Jerome B.
Schneewmd under the rubric of a 'metaphysician'. See Schneewmd, Jerome B.
(1967). 'McTaggart, John McTaggart Ellis,' in: Encyclopedia ofPhilosophy, vol. 5.
London/NewYork: Macmillan, 229-231. - A companson of the essay discussed
below with McTaggart's earlier Hegel commentary shows the impulse for the
development of unreality thesis originated from the Hegel exegesis. In his Studies
in the Hegelian Dialectic (1896), McTaggart explicates in the 5* chapter ('The
Relation of the Dialectic to Time') the contrary theses of finite and infinite time;
both are criticised as contradictory. In section 146, his conclusion anticipates his
later argument: 'Since either hypothesis as to the extension of time leads us into
equal difficulties, our course should surely be, not to accept either, but to reject
both. Time must, we are told, be either finite or infinite. But there is a third
alternative. There may be something wrong in our conception of time, or rather, to
speak more precisely, there may be something which renders it unfit, in
metaphysics, for the ultimate explanation of the umverse, however suited it may be
to the finite thought of every-day life. If we ask whether time, as a fact, is fimte or
infinite, we find hopeless difficulties in the way of either answer. Yet, if we take
time as an ultimate reality, there seems no other alternative. Our only resource is to
The Temporality Effect 193

his thought, moreover, requires no monograph, but was presented in an

essay of less than twenty pages.- Nevertheless, his controversial thesis
that time is 'unreal' because its logical deduction leads into an insolv-
able paradox has kept time philosophers busy for almost one hundred
McTaggart's argument is as follows:45 temporal positions and
events- can be ordered in two ways, in the linear succession of an
'earlier - later' stretch of time, or also in the sequence of 'past-present-
future'. The first organising principle McTaggart calls the B-senes; the
second provides a so-called A-senes. Events are also organised in a
third, non-temporal way - namely in the varied forms of non-temporal
sequence, for example by the organising principle of the alphabet. This
extra-temporal sequence he calls the C-senes.
McTaggart presumes that 'time' is necessarily determined through
the concept of transition and that the direction in which it runs is unam-
biguous. From the first premise follows that 'time' exists only under
the condition that the event, first imagined in a time-neutral C-senes,
can be assigned changing A-senes predicates. For this time-determina-
tion, the B-senes predicates do not come into consideration, for two
reasons: firstly, because they themselves would have to have been ar-
ranged on the basis of a temporal order; that is, they assume what they
would have to prove. Secondly, because the B-predicates of events
cannot change at all - since what has once been 'earlier' than some-
thing else will always remain so, or else the concept of linear temporal
order would be called into question.
But what is the situation regarding the possibility of finding a basis
for temporal ordering within the C-senes? This undertaking fails at the
second premise: non-temporal organising principles, such as the alpha-
bet, do indeed determine a sequence, but no direction (I can spell from
Z to A just as well as from A to Z). Onginal and also changing temper -

conclude that time is not an ultimate reality.' - Quoted from the electronic edition,
<> (12.11.2010). Ori-
ginal edition: A Commentary on Hegel's Logic. New York: Russell & Russell
McTaggart, J. Ellis (1908). "The Unreality of Time,' in: Mind 17: 457^174.
Here we are largely taking on the portrayal in Schneewind 1967 (cf Fn 43) but are
supplementing it with the exposition of the C-senes. A usable but for short
summary is provided by Savitt, Steven: 'Being and Becoming in Modern Physics,'
in: E.N. Zalta (ed.). The Stanford Encyclopedia of Philosophy (Winter 2008
McTaggart talks about events in the narrower sense of 'transformations'.
194 Jan Chnstoph Master

al determinations of events given in a time-neutral C-senes can thus

only exist by virtue of the A-senes: 'time' is therefore an effect of the
mapping of an 'earlier - later' A-sequence onto the elements of the C-
senes.47 Without an A-senes, there is no time. But the assumption of an
A-senes leads - according to McTaggart - to an infinite regress:

1. To be able to be a valid element of the A-series, an 'event' must have

all A-features; it must be able to have the predicates 'past', 'present'
and'future'assigned to it.
2. This assignment of predicates clearly cannot occur at one and the same
point in time, since this would be contradictory.
3. The obvious solution seems to lie in the assumption that these irrecon-
cilable predicates can be assigned to the element at different points in
4. Then, however, each of these three points in time, insofar as these,
after all, must take place in one and the same temporal continuum,
would have to be defined, in turn, in a second A-series. With that, the
requirement made of the assignment event is exactly that which has
been applied to the 'event' under point 1 - and so begins the infinite

With that, McTaggart conceives of time as 'unreal' - it is a logically

self-contradictory construct of our consciousness, which cannot cones -
pond to anything in objective reality, which is non-contradictonly or-
ganised. For him, this does not, however, mean that the discussion of
time is meaningless, since it would be nonsensical to deny that we have
a subjective perception of time. He conceives of this perception as
centred on the concept (shaped by E.R. Clay) that William James- la-
belled the 'specious present' and took as the starting point of his cog-
nitive-psychological explication of the functioning of human percep-
tion of time: the 'time window' whose extension varies individually, in
which the individual perceptions encompassed by it appear to us to be
McTaggart (1908: 460): 'Changes must happen to the events of such a nature that
the occurrence of these changes does not hinder the events from being events, and
the same events, both before and after the change [...] Take any event - the death
of Queen Anne, for example - , [ . . . ] At the last moment of time - if time has a last
moment - the event in question will still be a death of an English Queen. And in
every respect but one it is equally devoid of change. But in one respect it does
change. It began by being a future event. It became every moment an event in the
nearer future. At last it was present. Then it became past, and will always remain so,
though every moment it becomes further and further past.'
James, William (1886). 'The Perception of Time,' m: Journal of Speculative
Philosophy 20: 374^107, repr. James, Wiliam (1899). The Principles of Psych-
ology, vol. 1. New York: Holt, 605-642.
The Temporality Effect 195

subjectively 'simultaneous', and which we use as the subjective anchor

for future-oriented anticipations and past-oriented memories. McTag-
gart sums up correspondingly:

Our conclusion, then, is that neither time as a whole, nor the A series and B
series, really exist. But this leaves it possible that the C series does really
exist. The A series was rejected for its inconsistency. And its rejection in-
volved the rejection of the B series. But we have found no such contradic-
tion in the C series, [...].
It is, therefore, possible that the realities which we perceive as events in
time-series do really form a non-temporal series. It is also possible, so far
as we have yet gone, that they do not form such a series, and that they are in
reality no more a series than they are temporal. But I think [...] that the
former view, according to which they really do form a C series, is the more
Should it be true, it will follow that in our perception of these realities as
events in time, there will be some truth as well as some error. Through the
deceptive form of time, we shall grasp some of their true relations.49

4. A-lB-senes, Possible Worlds and the

Formal Conditions of Time Consciousness
McTaggart's reasoning took up and juxtaposed not only the core argu-
ments of the theological (Augustine) and the epistemological-tran-
scendental (Kant) considerations - it also anticipated- the fundamental
argument on which Husserl's phenomenological approach was based:
Human perception of time is centred on the - objectively counterfactu-
al, but existentially necessary - experience of a distinct 'now', aptly
termed the 'specious present' by E. Robert Kelly already in 1882:

The present [...] is really a part of the past - a recent past - delusively
given as being a time that intervenes between the past and the future. Let it
be named the specious present, and let the past, that is given as being the
past, be known as the obvious past.51

The concept was later elaborated upon by, among other, William James
and CD. Broad. The present has no clear delineation; it is a fleeting
McTaggart 1908: 473 (cf. Fn 44).
It is a case of anticipation despite its appeanng after 1904 (Husserl's lecture in
Gottmgen) insofar as McTaggart's basic idea was already developed in his 1896
Hegel commentary (cf.Fn 43).
Anonymous (= E. Robert Kelly 1882). The Alternative: A Study in Psychology.
London: Macmillan; quoted after: James 1899: 609 (cf. Fn48).
196 Jan Chnstoph Master

phenomenon. This realisation underlies Augustine's idea of a

'threefold present' as much as it informs Husserl's reasoning: in a
philosophical perspective our perception of the present cannot be un-
derstood as the perception of something located at a defined point in
time, but only as something appearing in an elongated 'time window'
(Husserl), and which moves continuously through this field, already
possessing in it a direct before and after. Similarly to Augustine and
Husserl, McTaggart also clearly presumes in his conception of time as
such - that is, as an effect that he tries to describe by connecting the
A-/B-/C-senes - that our experience of time is to be thought of in de-
pendence upon so-called 'indexical' representations within our con-
sciousness. Indexical in this sense: both remembering and anticipating
also indicate the position of the perceiving or judging intellect; again a
thought that we have already encountered in Augustine and then in de-
veloped form in Husserl (protention, retention)."
Despite these parallels in regard to the phenomenological conceptu-
alisation of 'time', the emphasis of McTaggart's argument naturally
lies on the attempt to trace the problem or the correlation between ob-
jective time and subjective experience of time back to its metaphysical
basis. With McTaggart's thesis of the 'unreality' of time, the question
of the reality of time, already raised by Augustine, receives a negative
ruling. But McTaggart's conclusion sounds more radical than was in-
tended. As a Hegelian he does not, for example, take on the position
represented by Kant, that time is no more than a 'transcendental
schema'.53 And it must also be remembered that McTaggart was very
On the concept of indexical representation, see Mohr, Georg (2000). 'Indexahsche
Representation von Zeit und die Simultaneity von innerer und auBerer Erfahrung,'
in: C. Kupke (ed.). Zeit und Zeitlichkeit. Wiirzburg, 119-137. - Mohr examines in
particular the use of the time indicator 'now'. His conclusion: the indexical
reference is 'characterised by a specific referential two-dimensionality (121), which
corresponds to a 'double-perspectived consciousness' of the speaker. The two-
dimensionality emerges as a result of the necessary 'restriction of the subject-
centred perspective' of the subject saying 'now', with which the mearnng 'at the
point in time of my present statement' is expressed. With that, the world is
positioned as an intersubjective frame of reference in which something happens at
the point in time of the statement, and which is the object of the statement. Thus, in
the 'now' overlap the 'series of speech events' and the 'series of events that become
the object of speech' (129); the consequence if the 'simultaneity of the indicated
event with the event of indications' (129) - own translations.
McTaggart (1908: 474) contrasts Kant and Hegel as follows: 'If this view is
adopted, the result will so far resemble those reached by Hegel rather than those of
Kant. For Hegel regarded the order of time-series as a reflexion, though a distorted
in the real nature of the timeless reality, while Kant does not seem to have
The Temporality Effect 197

conscious of having refuted, with his demonstration, the 'reality' of a

specific concept of time - namely the concept based on the premises of
one-dimensionality, uni-lmeanty and reality, i.e. objectivity. For him,
other conceptions are thoroughly possible, among them a time that,
though 'merely' phenomenologically given, is nested within itself."
His conclusion is thus essentially Solomonic: time may be unreal
from a metaphysical perspective - however, it still promotes a recogni-
tion in us, because, despite its erroneousness, it also unintentionally
gives us an insight into the world's true organising principle - repres-
ented by the C-senes. With that, McTaggart's approach offers a priv-
ileged point of contact for narratological investigations, for which (at
least in the subject area 'fictional literature') the question of the 'real-
ity' of the objects and relations in the narrated world is nonsensical
anyway. If, as McTaggart claims, 'time' is unreal, but finds its meaning
in that we use it as a kind of explorative construct - where, if not in its
application to something that is already by definition 'unreal', could
the genesis and the function of 'time as a construct', from the perspect-
ive of an A-/B-/C-senes, be better understood?
Precisely this question has been investigated by Le Poidevin.55 He
examines the special case of the interpretation of the temporal structure
of fictional events, as it arises in dependence on an assessment of the
truth value of fictional propositions within the limits of Possible
Worlds semantics. Le Poidevin ultimately reaches the analogue conclu-
sion of McTaggart's: the temporal 'nature of narrative', too, is funda-
mentally called into question by our intuitive conception of 'time as
transition' in terms of the A-senes. In his line of reasoning, Le
Poidevin reformulates the McTaggart model in propositional-semantic
terms. Following this approach two types of statements about 'time'
can be distinguished:

a) indexical ones, i.e. those that refer back to the subject of the statement
as a point of reference. These are statements about the A-series;
b) non-indexical ones, i.e. those that perform a relational determining of
position in the earlier - later continuum, which happens either through
the application of an absolute chronological marking of events, or also

contemplated the possibility that anything in the nature of the noumenon should
correspond to the time order which appears in the phenomenon.'
Ibid.: 'And how are we to deal with the appearance? If we reduce time and change
to appearance, must it not be to an appearance which changes and which is in time,
and is not in time, then shown to be real after all?'
Le Poidevin, Robin (1988). 'Time and Truth in Fiction,' m: The British Journal of
Aesthetics 28: 248-258.
198 Jan Chnstoph Master

inferentially, e.g. that over which the temporal interpretation of causal

relations is inferred. These are statements about the B-series.

With McTaggart, we have to assume that the position of an event in

'time' is to be essentially determined via the A-senes. We must be able
to trace B-senes statements back to A-senes 'facts' or A-senes 'state-
ments', from which they become semantically deducible as elements in
the mass of logically valid inferences- But - and this is the decisive
question - which way leads in the opposite direction, from B-senes
statements about the events in a fictional world back to the A-senes
statements and where - i.e. on which observing subject - would these
A-senes statements be anchored?
This problem, too, had already been addressed in principle by Mc-
Taggart himself (on the example of Don Quixote); he did, however, see
it as a nonsensical question, since he was less interested in the logical
problems of truth-neutral A-senes statements than in those of A-senes
'facts'. But if we consider the fictional text in itself as an empirical A-
senes, as Le Poidevin suggests, the series of 'mental events' in the
reader's consciousness can now be taken into consideration. We then
focus on the succession of the imaginations elicited by the text, which
are generated one after the other and which thus, seen objectively, do
constitute a real B-senes, but which, evaluated subjectively, form in-
stead an empincal A-senes. However, that merely tells us something
about the temporality of the cognition process and not about the Active
events and their temporal structure as such. Alongside this orthodox ar-
gument, there is, however, a second which is formulated not in ontolo-
gical terms, but semantically and reflects the problem not in regard of
the status of the 'facts' (Active vs. real) but that of the 'statements'
(fictional vs. factual). Indeed, in this perspective we are no longer talk-
ing about afictive world, but about zfictional one; a world whose char-
actenstic lies not in what it is, but in how it is being refened to, namely
by way of fictional (rather than factual, i.e. referential) statements.57A
fictional world in the sense of possible worlds semantics (according to
which all statements about that world lie within the scope of the modal
operator of fictionality) absolutely does allow the possibility of order-
ing the individual events in the form of its 'own', disjunctive, fictional
B-senes. Yes, one can even theoretically imagine that this kind of 'pos-
sible world' also 'has' its own, disjunctive (i.e. not identical with our
real one), factual A-senes. However, what one is unable to imagine is
how, under this presumption of 'doubled' A-/B-time series, statements
Cf. ibid., 249.
On the distinction^-ve /fictional cf. Fn 2.
The Temporality Effect 199

could be made in the real time series about the ordering of the fictional

[...] if the adventures of Don Quixote take place in a possible world, then
that world will have its own B-series, not connected temporally to us. Sim-
ilarly, there can be disjoint A-series: each world may have its own past,
present, and future, unconnected to ours. The difficulty arises if we try to
combine the idea that A-series statements are semantically independent of
B-series ones with the possibility of making A-series statements about fic-
tional events.

If we treat 'In the world of fiction fit is true that [...]' as a sentential operat-
or locating the referents of the terms which follow it in a possible world
[...], then temporal expressions such as 'before' as in 'It is true in f that (p
before q)' will not locate the events referred to in the actual world. How-
ever, the trick will not work for A-series terms. That is, placing 'p is past'
within the scope of the sentential operator will lead to an incoherent state-
ment. This is because we cannot make sense of an A-series temporally un-
connected with ours which is not relativised to some event, e.g., 'p is past
relative to q'. In other words any A-series statements which are genuinely
semantically independent of B-series statements will be about our A-ser-

Put more simply: every concrete (not made within a fiction) statement
about the location of an event in a Active continuum of past-present-fu-
ture (= a Active A-senes) is unavoidably a statement about our own po-
sition in our real time - unless we were to refer this statement back to a
starting point in the Active B-senes, which would lead to an infinite re-
gress. The special case of the constitution of time in a fictionally nar-
rated possible world seems equally aporetic, if not even more problem-
atic than the 'time' perception in the real world discussed by McTag-
gart. To corroborate his argument, Le Poidevin discusses three con-
ceivable objections:

1. 'Narrative statements are largely A-series statements in the past tense;

one can undertake the required relational determination on the basis of
grammatical designations.' - This type of foundation Le Poidevin dis-
misses as purely 'surface tensing', which loses its effect if we imagine
ourselves into the world of the story.
2. 'But is it not the case that precisely this 'imagining' ourselves into a fic-
tional-immanent time defines the necessary subjective-temporal refer-
ence point for Active A-series statements?' - No: because our 'imagin-
ing' is conceived in terms external to the Active world. There are no
Le Poidevin 1988: 251; emphases in the original (cf. Fn 55).
200 Jan Chnstoph Master

textual statements that immanently refer to this imagination and thus

make it into a (fictionally claimed) part of the fictional world itself: '...
what narrative would allow us to infer that 'It is true in f that I (the read-
er) exist?'59
3. 'In fiction itself, B-series specifications are made (explicit date specific-
ations or contexts that lead inferentially to them). These refer to real,
historical dates and events, which allow a connection to our real world
to be established.' - Le Poidevin's reply: since historically counter-fac-
tual statements are also made in fiction, one has to - in order to avoid
contradictions - assume that they do not really refer to our B-series, but
to a third'parallel world'.60

None of the three 'common sense' objections has so far been able to
conclusively show how one could find a reliable fiction-immanent ref-
erence point for A-senes statements about B-senes facts, which would
not in turn be somehow dependent upon the placing of a reference
point in the real A-senes of our own 'time'. However, there does ap-
pear to be one more possibility: We assume that every fictional report
is fundamentally to be imagined as having been expressed by an im-
manent narrator. With that, the anchor point of the A-senes statements,
which our premises imply are absolutely necessary for building up a
temporal structure, is placed firmly inside the fictional world: 'The
suggestion is to interpret all fiction [...] as being (explicitly or impli-
citly) told from the first-person perspective and that perspective de-
termines the position of events in the fictional A-senes.'61

Ibid., 252. - Le Poidevin makes it too easy for himself here. These kinds of so-
called 'metaleptic' statements are fairly common, particularly in late 18* / early 19th
century Romantic and 20th century postmodern literary narratives. On the
narratological concept as such see Pier, John: 'Metalepsis,' in: P. Hilhn et al. (eds.).
the living handbook of narratology. Hamburg: Hamburg University Press.
Le Poidevin (ibid.) takes this objection even further: Let us assume that the identity
proposed by objection (3) was ^problematically given. The totality if time is then
imagined as being divided into individual sections and these individual time
sections are hypothetically thought of in terms of set-theory, as partial sets of the
total set of all events in the world. This partial set is made up of events which exist
either simultaneously (with the same B-designation) or in an ordered sequence (=
as a structured B-senes). Two 'worlds' would then be identical if it were possible to
make these kinds of event sets congruent. But that is not enough - since it is not
only the identity of the partial sets that is decisive, but also their sequence. This
kind of strong identity in the temporal structure would then raise the question of
what it is that still divides the two worlds!
The Temporality Effect 201

This assumption is one with which literary study is entirely at home;

it is not for nothing that we place such emphasis in the handling of lit-
erary texts on the distinction between author and narrator! However,
seen in terms of the philosophy of time, it leads to another apona: An
immanent narrator is, for his part, fictional and narrated and therefore
subject to the temporality of the narrated world as a whole. But how
are we to deduce this temporality? The assumption of a second narrator
who narrates the first, and so on, leads to an infinite regress again;
however, absolutising the immanent narrative entity as a fixed point re-
duces everything to the Active B-senes and precludes the possibility of
an A-senes, i.e. the observation of events from the perspective of their
In view of the narrated world, also, we are apparently unable to
evade the time-philosophical dilemma that McTaggart demonstrated -
unless we draw a conclusion as radical as the claimed 'unreality of
time'. Le Poidevin does not go quite so far - but he does support defin-
ing the dependence between the A-senes and B-senes statements in ex-
actly the opposite way when refernng to the case of a nanated Possible
World: a position, which is known in the McTaggart debate as the B-
thesis (in contrast with the A-thesis, proposed by McTaggart himself,
which argues for the pnmacy of the A-senes determination). Le
Poidevin at least suggests that his thesis requires drawing conclusions
about the more general philosophical determination of the concept of
'The second solution is to give up the idea that B-senes statements are semantically
dependent upon tensed, A-senes statements (e.g., that "p before q" entails the
disjunction, "Either p is past and q present, or p is present and q is future ..." etc.).
Instead, A-senes statements are reducible to tenseless, B-senes ones. This would go
naturally with, but is not entailed by, the doctnne that tensed statements are made
true by B-senes facts.
My own inclination is to adopt the second solution, and accept the consequential
revision of our ordinary conception of time. To do that, of course, is to say nothing
very enlightening about fiction, but it is because I suspect that this will be a very
unpopular option, that I think the puzzle could lead to some intrinsically interesting
insights into the nature of nanative' (ibid., 2511). - Although the conclusion is
here related to the special case of narrative temporality, but certainly seems to be
intended more comprehensively in the sense of a fundamental time-philosophical
argument - this question, already very briefly discussed by McTaggart, would
otherwise not have needed to have been so thoroughly treated to reach a very
similar result. In my opinion it is true that it is implied that the special case of
narrative temporality can call into question our real conception of 'time as
transition' and the A-senes foundation of time-consciousness denved from that; see
on this also the introduction in Le Poidevin & MacBeath, eds. (1993: cf Fn 5).
202 Jan Chnstoph Master

A significantly more enthusiastic proponent of the 'B-thesis' is

Peter Bieri, for whom the question of the relationship between the ex-
perience of time and of personality stands in the foreground.63 Bien's
initial thesis can be reconciled with both McTaggart's model and
Husserl's definition:

Thus, in order to have a genuine time consciousness, a being has to be able

to make a difference between the history of the world and the history of its
encounter with the world. And the continuously changing perspective I
have spoken about is nothing else but the continuous process to connect
these two sequences of events in one representation of a single, consistent

In contrast to Husserl, McTaggart and Le Poidevin, however, Bieri pos-

its the reality of time, which, in comparison with Le Poidevin, leads to
a less sceptical portrayal of the B-thesis, but does underline the 'So-
lomonic' impetus of temporal constructivism, which we encountered at
the end of McTaggart's line of reasoning. The argument can be
sketched as follows:

1. There is a diachronic continuous existence of entities in this world even

when they are not represented.65
2. There can be a difference between the way we represent an entity, and
the way the entity is.
3. However, I can correct my representations by checking them for causal
connections and consistency. This will allow me to establish an object-
ive temporal order in the past, or more precisely, in my representations
(memories) of it.

With this, Bien overall defends the realist position and the logical and
ontic primacy of the B-senes, which he positions as real and given -
and indeed independent of our representations and A-senes contextual-

Bien, Peter (1986). 'Zeiterfahrung und Personality,' in: H. Burger (ed.). Zeit,
Natur und Mensch: Beitrage von Wissenschaftlern zum Thema. Berlin: Berlin
Verlag A. Spitz, 261-281. Translated in the current volume as: 'Time Experience
and Personhood', 13-28.
Bien in the current volume, 17.
The Temporality Effect 203

isations ('metarepresentations'").67 There are three points here that re-

quire closer examination.
First, the reference to the differing status assigned to the A- and B-
senes statements is essential: B-senes statements are statements about
objectively eiven relations ('a is earlier than b'); A-senes statements on
the other hand are subjective determinations of the relation between the
datum of perception- and the subject of perception. Secondly, we have
to distinguish between the principal assumption of a primacy of the ob-
jective B-senes and the phenomenological fact that the 'primary way in
which time is represented in the consciousness [is] the A-senes data'.-
Thirdly, the assumption of the 'reality' of time is a positioning that -
independent of its philosophical tenability - has a pragmatic basis, be-

time-consciousness presumes a temporal structure as real when this pre-

sumption either offers the only possibility of explaining the phenomenon
and without which it would remain 'incomprehensible' in the sense of such
an explanation, or when it at least offers a greater chance of explanation for
the assumption of real time[...].™

On this basis, Bien also bnngs into position something he calls - to

distinguish it from Husserl's phenomenology, which he believes
provides only a pure description of the perception of time - a 'method-
ological'71 explanatory approach. This approach aims at an 'explana-
tion of time-consciousness, defining itself as the naming and specifica-

Cf. ibid., 19ff.
Bien understands this as positioning. For him, there is no logical or even
metaphysical foundation for this exception to reality, even though pragmatic
reasons argue for it. The possibility is conceded throughout that epistemological or
transcendental philosophy will force us to the realisation that, because of the
necessanly temporal constitution of the realisation, the assumption of the reality of
time represents a circular proof.
Bien takes on essential components of Husserl's approach - including the concept
of the 'datum'. However, he does criticise the so-called 'phenomenological
reduction': the principle '"suspension" of objective and real time in time-
experience' needs to be undone. See also: Bien, Peter (1972). Zeit und Zeiter-
fahrung. Exposition eines Problembereichs. Frankfurt/M.: Suhrkamp, 200;
hereafter referred to as Bien, ZuZ.
Bien, ZuZ, 184. All English quotations of this text are my own translations.
Ibid., 202; emphases in the original.
Cf. Bien, ZuZ, 207. - In the later essay, 'Time Expenence and Personhood'
(contained in the present volume) Bien also asks in a Kantian way about 'the
conditions for a being to have an entity, a consciousness of time' (14).
204 JanChristophMeister

tion of necessary and sufficient conditions for the phenomenon' and, in

conscious allusion to Kant's transcendental method, as a derivation of
the 'conditions of possibility' of something. 72 At heart, there are four
conditions that have to be fulfilled:

1. The - qualitative or compositional - transformation of the data.

2. A mere mass of data organised on a B-series relation is not sufficient;
rather, there also has to be a consciousness of this B-series order.
3. The conciousness data must also have the quality of being 'conscious'
not only as points, but for a specific duration; however, they also have
to be transformed within this duration. That means we have to be con-
scious of their transformation, which necessitates that a transformation
of their mental portrayal, their representation, also takes place.
4. The unity of the perceiving consciousness is a necessary presupposition
for conditions (l)-(3). 7 3

In view of the question of to what extent the 'reality of time' is implied

under (1) to (4), divergent findings emerge. The reflexive self-descrip-
tion or 'self-interpretation' of consciousness under (4) necessarily as-
sumes A-senes temporality, although this is not an argument for or
against the reality of the A-senes structure. Instead, it suggests 'the
suspicion that past, present and future are no more than modes of por-
traying subjectivity'.- Objective explanation of time-consciousness, on
the other hand, must presume only the reality of the B-senes as implied
by the conditions (1) to (3). So, one may not conceive of the experience
of time using the pattern of other objective experiences of reality. What
need to be distinguished are the representations of data in the perceiv-
ing consciousness, i.e. the A-senes determination, and the real order of
data in the sense of B-senes relations. The determinations thus provide
the subjective image of the relations:

Real time as a B-series presents itself in the conscious events that it orders
through A-determinations, and this is possible because the former imply the
latter as their principle of construction.75

In one of his concluding remarks, Bien examines the relationship

between time-consciousness and identity-consciousness. Like Kant, he
assumes that 'knowledge about one's own identity cannot be under-

Bien, ZuZ, 203, including his footnote 14.
Cf.Bicri, ZuZ, 205ff.
Bien, ZuZ, 211.
Bien, ZuZ, 217.
The Temporality Effect 205

stood as an identification on the basis of distinct identity criteria',"

what he suggests instead is the possibility of founding identity-con-
sciousness in the perception and consciousness of real time. In his later
essay (contained in this anthology), Bien provided a noteworthy contri-
bution to the theory, required at this stage, of an interpretation of sub-
jectivity as the representation of real time to the self." Again, this
touches upon the debate about 'time', which, although being of emin-
ent philosophical significance, is beyond the horizons of the current
narcological approach.

5. Mapping McTaggart's A-/B-/C-senes onto

Narrative Time Construction

I will now try to sketch how, in a narratological perspective, McTag-

gart's model can be mapped onto the phenomenology of narratives. In
principle my approach follows Le Poidevin's suggestion to concentrate
on the phenomenon of time relations assigned by the recipient of fic-
tional narrative texts to the events located in a possible world. How-
ever, in contrast to Le Poidevin, my focus is not on the results, already
expressed in the form of statements, of cognitive processes of recon-
struction and their conditions. An explication of these statements con-
tributes little to the elucidation of the constituents and processes that
lead to the construction of temporality in the sense of generating con-
ceptions of temporal order, and thereby also to the conception of tem-
porality in general. A further difference lies in that, unlike Le Poidevin,
my model will also take the so-called C-senes into account (whose
conceptual relevance within the McTaggart model was also accorded
little attention by Bieri). On the other hand, we want to leave entirely
to one side the question of the 'reality' of time. I believe that the prag-
matism for which McTaggart himself argues towards the end of his ex-
position, as well as Bien's phenomenologically inspired interpretation
of McTaggart, justify this decision from a philosophical perspective.
Let us briefly sketch again the three 'series' from which McTaggart
built up his model:

(1) A-series: the series in which the content of consciousness appears ac-
cording to a subjective determination in the continuum of fu-
(2) B-series: the series in which the contents of consciousness appear in the
more objective relation of earlier - later.
Bien, ZuZ, 221; emphases in the original.
Cf. Bien in the current volume, 13-28.
206 Jan Chnstoph Master

(3) C-series: the series in which the contents of consciousness are given in a
non-temporal order (e.g. as a set).

A constructive interpretation has already flowed into this portrayal as

I consciously aim to model the process of constituting temporality ef-
fects in the course of reading a text, and not an assumedly pre-existing,
quasi-real temporal structure or histoire. How do these three series now
present themselves in our subject area?

Narrative C-senes Ordering

Strictly speaking, one should imagine the C-senes as the set of textual
signs given even before any reading. However, let us be more pragmat-
ic: for us, the C-senes is formed by the chain of lexemes and lexeme
groups functioning as the bearers of representational contents that we
decode and process in the form of mental images. In their overall com-
bination they are the base elements that, if the reception is successful,
evoke in the reader's consciousness a mental representation of a Active
or real world. These entities can be quasi-realia (things, people, more
generally put: existents) or ideal (events, situations, thoughts). The or-
ganising pnnciple, as a rule, is the syntax of natural language. Newer
medial forms of presentation, such as hypertexts, show that spatial
forms of organising pnnciple are also possible. Of course, each of
these organising pnnciples is already somehow affected by temporal-
ity; in this regard, one should remember Kant's definition of 'time' as a
'transcendental schema'. But this type of 'temporality' is purely formal
and abstract; it remains a 'possibility condition' without being expen-
enced as an actual phenomenon. To summarise, the C-senes is consti-
tuted by the linear sequence of mental images that runs in parallel to
that of the matenal symbolic elements that make up the text. In a liter-
ary nanative, these elements to which the secondary mental images at-
tach are the words and phrases of a natural language.
To illustrate the first step in our nanatological mapping of McTag-
gart's model let us consider as an example sentence the beginning of
Edgar Allen Poe's 1841 A Descent into the Maelstrom which opens in
medias res: 'We had now reached the summit of the loftiest crag.' Let
us focus on the verbally represented existents in this section of a fic-
tional world only and highlight the respective words accordingly:

We had now reached the summit of the loftiest crag.

The Temporality Effect 207

As a C-senes this phrase amounts to a string of mental images that can

be numerically ordered as (1), (2), (3) where

(1) evokes the mental image of a minimum of two agents (= 'we')

(2) evokes the mental image of the highest point of an elevation (= 'sum-

(3) evokes the mental image of a pointed rock (= 'crag')

Of course the syntax, by enforcing conventionalised word order,

already induces a seemingly temporal sequentiality. But that is merely
accidental - consider the following semantically identical re-phrasing
in the passive voice:

The summit of the loftiest crag had been reached by us.78

The C-line is still a string of (1), (2), (3), but now

(1) evokes the mental image of the highest point of an elevation (= 'sum-

(2) evokes the mental image of a pointed rock (= 'crag')

(3) evokes the mental image of a minimum of two agents (= 'us')

Narrative A-senes Ordering

The narrative A-senes, too, is formed of the three dimensions of temporal

consciousness: past, present, future. Empirically, temporal consciousness is
first constituted when its dimensions are no longer empty, but when C-ser-
les elements, i.e. mental images are assigned to them and then appear in
them. It is necessary to distinguish between two modes of this appearance:
images can be directly perceived or called up from memory (or imagined
in anticipation) and thus be indirectly perceived as representations. I will
concentrate first on the mode of direct perception and then discuss repres-
entation-based perception further below.
This point could probably be made even stronger if we base our reasoning not on a
phrase grammar, but on a Tesmenan dependency grammar. Indeed, one might
argue that McTaggart's conceptualisation of the C-line as a line is in itself
contradictory and unnecessarily exposes his argument to criticism: there is no
reason why the contents of the C-line could not be conceptualised in even more
abstract terms, i.e. as an ordered set or a matrix.
208 Jan Christoph Meister

In the sequential scanning of the C-series of words and phrases, it is

initially the case that one image after another moves into the present-
dimension of consciousness. We can imagine this dimension, in refer-
ence to Husserl, as a 'now-time window', or in short, a 'now- window'.
This narratively evoked present, too, contains more narrated content
than just a singular mental image processed and foregrounded at a par-
ticular, mathematically defined point in reception time. As in real-
world perception, the narrative-based now-window rather shifts itself
successively across the symbol-evoked elements presented to us in the
pre-temporal C-series. Let us look at our example sentence again: the
'summit' and the 'crag' may evoke two distinct mental images - but
the phrase relates them closely and invites us to stretch our now-win-
dow beyond their individual borders:

Diagram 1: Now-window containing a temporally homogenous set of mental images

The images contained in this now-window form a temporally homogen-

ous set: they are initially 'timeless' in relation to one another and can-
not be distinguished by means of their temporal index. Nonetheless, the
set is not unordered insofar as it is still organised by the measure of the
C-series. But this sequential ordering is not temporal; or, in Husserl's
words: the logical succession of elements that move through the
present time-window, does not determine the perception of this succes-
sion as temporal. This is why every element contained within the now
window bears the identical temporal index t0.
Note that at present we are dealing with an isochronous narrative,
i.e., a narrative in which the representational elements are arranged in
accordance with their chronological occurrence in the (real or ima-
gined) ordo naturalis, and that we can therefore leave aside temporal
determinations that attach to the mental images evoked by the narrative
- from explicit temporal expressions such as the deictic 'now' in 'We
had now reached...' through to inferentially deduced chronological de-
terminations. In our model, temporal determination markers located on
the surface level of the text - be it verbally as in the case of a deictic
expression, grammatically as in the case of a tense shift, or structurally
The Temporality Effect 209

as in the case of a chapter border - are termed temporal operators. We

call them operators because, in a cognitivist perspective, they do more
than just denote temporal position or structure of representational ele-
ments: they actually control the processing and ordering of these ele-
ments as mental images. We will soon encounter this in practice.
The 'now window' is the only one of the three temporal dimensions
with two defined borders, namely to the future and past dimensions.™ If
the window shifts further to the right along the C-series, then on the
left an image falls out of the limits of the window. It would then re-
ceive an individual temporal index t<„. This process repeats itself con-
tinuously, and all temporal indexes - starting from the dynamically
placed value t0 - are re-calculated with every shift to the right. We have
now begun to experience subjective time, for in addition to a present
now-window, there is a dynamic past now-window extending to the left
of it. This can again be demonstrated with the help of Poe's narrative
which continues as follows:

Diagram 2: Re-indexing of mental images and A-series constitution effected by a

now window shift

As a consequence of the now-window shift to the right the mental images

of a crag, a summit and a collective agent denoted by the 'we' successively
recede into the past, which becomes discernable as the past: an A-series of
past-present-future perceptions has begun to take shape.

About how the extension of this time-window could be defined, nothing can be said
at this point. In regard of the processing of literary texts, it is presumably defined
through fundamental cognitive-psychological parameters as well as through cultural
and aesthetic conventions. This question will be addressed in the following chapter.
210 Jan Christoph Meister

Narrative B-series Ordering

Insofar as images successively fall out of the time window of present

perception and receive an individual temporal value, while new ones
enter the now-window and receive a neutral time index, we also ap-
proach an objectivising of the temporal structure and thus, in addition
to the A-series, the constitution of a B-series. The C-series provides an
absolute, pre-temporal ordering of images. In Western languages the
spatial left-to-right order of words is conventionally mapped onto a nu-
meric sequence from 1 to n; yet other ways of ordering are possible:
consider for example the tabular ordering principle in a computer gen-
erated collocation index. The A-series, on the other hand, constitutes
the 'indexical' temporal ordering of perception itself, i.e. one that re-
ilexively refers back to the observer's position. It is the B-series that
first puts us in a position to delineate a fixed temporal ordering of what
is perceived, although this is merely a purely relational temporal de-
termination in the continuum 'earlier-later'. Related to our example
narrative and put in concrete terms, we begin to envisage a world in
which an old man first climbs a lofty crag and then is 'too much ex-
hausted to speak'. And nothing will ever change this eternally fixed re-
lation of before-after, no matter how much further we decide to shift
our now- window along the narrative's C-series. We have now begun
to consciously co-experience subjective and objective time and a fully
developed model of the three co-existing series can be rendered as fol-

Diagram 3: A-series and B-series constituted on the basis of a C-series

The Temporality Effect 211

Until now, our exposition has dealt only with the case of an entirely
isochronous narrative in which the events have been portrayed in the
sequence of their natural occurrence in the narrated world, and which
the reader can reconstruct 1:1 by generating individual mental images
on the basis of the textual signs. However, in order to be able also to
represent cases of amsochronous reverse-order narratives in our model,
we have to take into account the Active temporal information contained
in the textual material, each piece of which belongs to a mental image.
Pieces of temporal information can 'belong' to an image in two funda-
mental ways: firstly, in the sense of a piece of information about the
latter's temporal position in the B-senes continuum of represented hap-
penings; secondly, in the sense of apiece of information about its posi-
tion in the B-senes continuum of perception. Within the text, it is only
the first type of information that can be realised, insofar as information
about chronological position in the continuum of perception presup-
poses that the perception (here: reading and conceiving of something)
has already been completed.- In order to grasp the temporal informa-
tion relative to its position in the Active continuum of happenings, we
will now introduce the notation index tf (= Active time-point or time
frame of occurrence). 'Image tfle3' should thus be read as: 'the third C-
senes existent successively realised as a mental image, which is located
on the B-senes at the Active point in time 1'. The following diagram
presents a case of two mental images, which are - measured against the
B-senes temporal relation of their contents - organised regressively.
As an illustration, let us re-use the re-phrasing of Poe's initial sentence
in the passive voice:

Narrated perceptions, for their part, are also events in the continuum of
happenings. Particularly difficult to model are preceding references: if, for example,
a frame narrator announces the presentation of an inner narrative. It is not the
ultimate temporal classification of what is (or is about to be) reported in the inner
story that presents a problem, but rather the fact that this announcement, outside the
fictive happemngs, also relates to the real perception happenings of reception.
212 Jan Christoph Meister

Diagram 4: Regressive B-series time

Our model's need to process these conceptual units correctly in the

reader's time-consciousness now becomes significantly more demand-
ing: it is now clearly a case of a process taking place on a meta-level
and encompassing several steps:

(1) We register that the earlier - later relation, constructed in the B-series,
between images 1 and 2 and image 3, stand in contradiction to a medi-
ated piece of temporal information within the image.
(2) The ongoing processing of new C-series conceptual units in the time-
window must therefore be temporarily suspended: if we simply kept on
processing C-series units and shifted our now-window further to the
right we will not be able to normalise the B-series any longer once our
capacity to memorise the B- versus A-series inversion has been ex-
hausted. To prevent this from happening,
(3) images 1 and 2 are summoned back out of memory and into the present
time-window in reversed order; thereafter
(4) the temporarily suspended processing of new C-series elements is re-
activated; the images 1 and 2 move to the left and out of the window,
and now take up their 'proper' positions to the right o/image 3 and thus
within a reconstructed B-series identical to the one presented in diagram

Already in this relatively simple correction, we have twice 'switched'

between the modes of direct perception and representation-based per-
ception. But merely switching is not enough for us to reach the desired
The Temporality Effect 213

outcome. The temporal consciousness must not only process fict-

ive-happemngs-onented temporal information in the bracket, but also
the process-oriented information before the bracket - and must simul-
taneously recognise that it is dealing with two distinct types of tempor-
al information in order to be at all able to solve the problem. Becoming
aware of the divergence between temporal process information and
temporal happenings information goes hand in hand with the distinc-
tion between direct and representation-based perception, but then also
with becoming aware of the 'past' as the specific dimension of tempor-
al consciousness in which images are no longer available 'in and of
themselves' as phenomena currently being experienced, but as repres-
entations of such experiences. - Our graphical portrayal can only
present the conceivably simple case of this kind of reflexive transform-
ation as it happens to two images shifted into the immediate past. How-
ever, one can imagine just how complex these kinds of revision process
can actually end up and how 'powerful' the extension of the time-con-
sciousness to be actuahsed must be: if, for example, the contradiction
appeared between the images that occur at larger C-senes intervals.
And to make things even more complicated: similar contradictions can
also be caused by reverse order B-senes, and they can also involve
mental images which in and by themselves contain not just individual
existents, but entire sequences of Active events, which again need not
be represented in the C-senes in an isochronous mode.
Alongside the revision of the B-senes order, which is tnggered by
fictive-happemngs-onented temporal information encapsulated within
the images that constitute the series, there is also a second possibility
for the temporal consciousness to switch from a 'blind' to a 'reflexive'
process mode. In this case, too, concrete, empincally demonstrable
components of the text are the trigger - namely the so-called 'indexic-
al' time expressions that provide us with information about the position
and, above all, the extension of the present time-window in which we
successively process the C-senes elements and transform them into a
B-senes. Making the imaginative process itself dynamic in this way, is
relevant, in contrast to the examples discussed about, pnmanly to the
A-senes. This can be brought about not only through temporal informa-
tion in a nanower sense, but also, inter alia, through formal or stylistic
means (change of chapters, switch in nanative perspective etc.). In this
instance, too, there is a 'switching' from direct (in Plato's sense: mi-
metic) to representational (diegetic) perception and processing of the
images. In other words, the process of reading the text and processing
the individual 'image-blocks' that contnbute to the construction of a
complex mental image of a Active world made up of entities and events
214 Jan Chnstoph Master

switches back and forth between two modes of operation: simple de-
fault isochronous decoding and construction and higher-level meta-

6. The Role of Temporal Operators in Narrative

Discourse Processing

The mapping of McTaggart's A-/B- and C-senes approach to time's

paradoxical logic onto a sequence of narrative propositions suggests
that two classes of text elements have to be distinguished in the narra-
tological modelling of temporal phenomena. Common to both is that
they influence the delineation of temporal constructs. However, they do
this in different ways and with fundamentally different effects. As was
already mentioned, I term these text elements temporal operators and
correspondingly differentiate between structural and dynamic temporal

Structural Temporal Operators

Structural operators perceive the function of chronometnc marking.
Examples are particularly absolute determinations of time such as date
stamps, and some of the temporal deictica. Three sub-classes of this
class of operators can be theoretically distinguished: chronometnc,
grammatical and inferential. Chronometnc operators assign to the indi-
vidual events an absolute temporal B-senes predicate:

tchroni(E),tchron2(E), tchronN(E)

In the statement 'first he slept, then he woke up', for example, these are
performed by the determinations 'first, then". These operators can al-
ways also effect a retrospective 'normalising' of the B-senes. In the
simplest case, this happens through the specification of an absolute
'earlierness' relation or through the selection of tense; in the most com-
plicated, through inference from valid causal relations:

W o ( E ) , tg^-^E) = 'he woke up; he had slept'.

tM-effect(E), W-cauSe(E) = 'he was happy; she loved him!'

Temporal deictica present a problem in this context, in that they are ac-
tually always already indexical expressions, marking a speaker's posi-
The Temporality Effect 215

tion. That is why we have to differentiate between immanent deixis

within the Active world - 'now, then, at twelve o'clock' - and the
deictica that already, at least in part, indicate the fact of their conceived
nature, such as 'two days later': that is, the case of a transcendent deix-
is. It may also be possible to define this difference purely quantitat-
ively: the greater the deixis' B-senes extension, the more strongly it
tends towards a transcendence of the situational context of current per-
ception. This kind of quantitative determination would not be an abso-
lute one, but would have to be defined as a deviation opposite the nor-
mal deviation already prevailing in a given text.

Dynamic Temporal Operators

Dynamic operators are text entities, which, in regard to time, draw our
attention to the imagined nature of an image or to our involvement in
the process of its imagination. The common functional determination
of these markings is that they, when successful, trigger the (reorgan-
isation and ordering of a number of individual conceptual elements into
a conceptual complex, to which is assigned an A-senes predicate, so
for example:

past (El E17); present(E18, E19); future (E19+1 E19+n)

The essential service of dynamic operators in regard of the temporality

effect consists in that they (re-)activate or re-define the reference posi-
tion, lying outside the Active B-senes, which, according to McTaggart
and Le Poidevin, is absolutely necessary in order to reach an A-senes
structure and with that the effect of 'time as transition'.

7. Conclusion

We will now stop here with this increasingly abstract portrayal. The
design of a typology of temporal operators and the concrete assignment
of the diverse forms of time definition (e.g. the context-dependent or
relational operators such as 'at the time', such deictic operators in the
narrower sense as 'now', absolute time definitions such as '12 October
1916' etc.) remains a desideratum; it is clear that we would have to
take into account grammatical and text-linguistic categories. The com-
plex of the dynamic guidance of time-onented perception constructs
(e.g. analepsis, prolepsis, formally enforced 'leaps in time' through a
change in chapters etc) would, on the other hand, clearly have to be re-
216 Jan Chnstoph Master

fleeted in the light of cognitivist theories and models concerning ques-

tions of memory function, the physiological clocking of time-percep -
tion etc.
However, even in the absence of a fully developed typology and a
more precise model of the cognitive processes involved it has hopefully
been shown why narratives provide a prime opportunity for humans to
train their skills in temporal construct design. In narrative processing
we can constantly engineer, reverse engineer, test and evaluate alternat-
ive time constructs for validity - and the omnipresence of narratives as
our contemporary societies' primary mode of representation points to
the dominant role that narratives therefore play in developing a shared
human time consciousness. In the end, the metaphysical question of
whether time really exists is perhaps as non-sensical as is the question
if anything exists, including ourselves. Taken in this light it appears
that we perhaps better read McTaggart's apodictic refusal to engage in
any speculation on the nature of narrated time as a tongue-in-cheek re-

The [...] objection rests on the possibility of non-existent time-series - such,

for example, as the adventures of Don Quixote. This series, it is said, does
not form part of the A-series. I cannot at this moment judge it to be either
past, present, or future. Indeed, I know that it is none of the three. Yet, it is
said, it is certainly a B-series. The adventure of the galley-slaves, for ex-
ample, is later than the adventure of the windmills. And a B-series involves
time. The conclusion drawn is that an A-series is not essential to time. I
should reply to this as follows: Time only belongs to the existent.81

Indeed, McTaggart seems to be right on that, too. For let us restate Le

Poidevin's conclusion that

[...] we cannot make sense of an A-series temporally unconnected with

ours which is not relativised to some event, e.g., 'p is past relative to q'. In
other words any A-series statements which are genuinely semantically inde-
pendent of B-series statements will be about our A-series.82

And so, if fictional B-senes statements are about non-existents, the

past-present-future A-senes of our imagining the narrated non-exist-
ents is even more about - us being in time.

McTaggart, McTaggart John Ellis (1927). 'The Unreality of Time,' in: The Nature
of Existence, vol. 2. Cambridge: Cambridge UP, 29.
Le Poidevm 1988: 251 (cf. Fn 55); emphases in the original.

The Flow of Time in Narrative

An Artificial Intelligence Perspective

'I wonder', said Ada. 'I wonder if the attempt to discover

those things is worth the stained glass. We can know the
time, we can know a time. We can never know Time. Our
senses are simply not meant to perceive it. It is like —'.


Time passes. The evidence for its passage is incontrovertible, and yet
that fact does not make it any easier for us to grasp what our sense of
time is. Unlike space, time is not directly perceived, though we can
measure it in various ways. Despite our perceptual blindness with re-
gard to time, or perhaps because of it, our temporal cognitions are very
rich; they include our experience of events as they appear the present,
our memory of events in the past and our visions of them in the future,
as well as our regrets for and hallucinations of those events that did not
occur. To communicate these cognitions in a narrative, human lan-
guages provide a variety of systematic mechanisms such as tense, as-
pect, and types of time expressions; these mechanisms allow narrators
and characters to express, from their viewpoint, the position and tempo
of events in time, as well as to attempt to portray time itself. Finally,
human inventiveness allows for stimulating devices and clever circum-
locutions for narrating the ordering, distance, and tempo of events in
As I have indicated, time does not appear alone in narrative; it is
wound up with events, and involves relationships that hinge on modal-
ity and point-of-view. Further, I have described a subjective notion of
time, where the position of an event in time is dynamic and changes re-
lative to the speaker; this notion will have to be reconciled with another
218 InderjeetMani

so-called 'objective' notion of time where the events are ordered in a

fixed, static fashion and sometimes grounded in terms of a calendar, al-
lowing us to reason and reckon with it. These two notions have been
well-discussed since the time of Augustine's Confessions, and corres-
pond roughly to the A- and B-series respectively of McTaggart.' In this
chapter I will introduce an Artificial Intelligence (AI) perspective that
integrates these concepts together in a coherent framework. In relation
to time, AI asks the question (1) what sort of reasoning related to time
and events can an intelligent agent carry out? It further asks (2) how
can an intelligent agent carry out this reasoning in relation to a given
narrative in natural language? A related question (3) is a cognitive one:
what sort of reasoning does a specific agent, the human reader, carry
out in relation to time and events in a narrative? This question is
answered to some extent by cognitive science, using methods from ex-
perimental psychology. The answer to (2) has not been so far guided by
the answer to (3), but I will suggest how the two can be related to each
The answer to (1) can help the narratologist in providing a formal
temporal framework for investigating patterns of narrative ordering and
tempo. The answer to (2) can potentially alter the way we do narrato-
logy, for it also allows systems to discover temporal patterns automatic-
ally from vast volumes of on-line narratives. The quality of these pat-
terns, however, depends in part on humans investing the time to teach
the AI systems about them. The answer to (2) also allows for automatic
systems to generate temporally interesting narratives, though the effic-
acy of these latter systems is limited by the vast amounts of knowledge
of the world that have to be programed into them. Finally, the answer
to (3) allows one to make predictions that can be tested with humans.
Before going further, let me introduce some informal definitions.
Let us say a narrative is an artifact (text, movie, etc.) produced by an
agent (the author); it is also interpreted by an agent (the reader or view-
er). The agents may be artificial or human. The artifact is in a particular
medium (video, text, audio, etc.) distributed in some format (podcast,
newspaper, pamphlet, book, etc.) and is expressed in one or more nat-
ural languages. In posing questions (1) and (2), AI assumes that there is
some functional mapping between an internal mental representation
that an agent has and the narrative itself. There can be two mappings, a
'production' mapping from an internal representation in the author's
mind to the narrative artifact, and an 'interpretation' mapping from the
narrative to an internal representation formed in the reader's mind.
Thus, there can be an AI system for an author and another AI system
McTaggart, John Ellis (1908). 'The Unreality of Time,' in: Mind 17: 457^174.
The Flow of Time in Narrative 219

for a reader. The mappings have to be computable, i.e., executable (and

ideally, also executed) by a computer program. This means the repres-
entation will be a specific data structure, very precisely instantiated.
Such a view has a number of inherent problems, including the lack
of biological and social grounding, which stems from the functionalism
inherent in this approach. However, its main problem bearing on time
in narrative is the nature of the internal representations. We have no
real idea, for example, of the internal representation in Nabokov's mind
of the chronology and tempo of events in Ada, even given the partially
explicatory essay on time within the book. One might also question
whether any such representation could be even said to exist in an au-
thor's mind, bearing in mind the fact that ideas are usually revised
along with the writing. In any case, any information about authorial in-
tention that we might learn from literary sleuthing (and shuffling
Nabokov's index cards) will likely fall far short of a precise representa-
tion. Similar arguments can be made about the internal representation
that the AI reader might build. In particular, different human readers
might build different representations, though they may share elements
in common.
Nevertheless, there are several arguments that one might marshal in
favor of considering the AI view as a substantial contribution to narra-
tology. First and foremost, the particular one I will describe for time
commits to a precise and formal internal representation, one that has
been used quite widely to analyze natural languages. In doing so, it al-
lows for ambiguity (multiple interpretations) as well as under-specific -
ation (the omission of temporal relationships). This is far more precise
than the various interpretations narratologists have given to the corres-
ponding notion of the fabula, which is taken to include both plot as
well as chronology. There is no reason, therefore, to jettison such a
highly expressive representation a priori when it comes to narrative. Its
adequacy as a descriptive framework may be assessed in terms of its
ability to capture the varieties of temporal and modal phenomena found
in narrative, as well as, if desired, its verisimilitude compared to hu-
mans. In fact, the current lack of understanding of how humans experi-
ence narrative can be ameliorated to some extent by testing the psycho-
logical predictions made by such formal models.
Second, let us focus on the AI reader. Since the representation the
system builds is computable, it allows the narratologist to pose various
questions to the AI reader and discover (and judge) the answers. In
Faulkner's story A Rose for Emily, at what intervals did the townsfolk,
according to the first-person plural narrator, visit Miss Emily's resid-
ence? And what sorts of temporal structures are observed in Faulkner's
220 InderjeetMani

oeuvre? In Flaubert's Madame Bovary, could Emma really have spent

Tuesday afternoons in Rouen, given the posted schedule of the
Hirondelle coach? The answer to the latter is no, as pointed out by
Steegmuller.2 As mentioned earlier, the AI representation allows us to
discover patterns couched in terms of it. These patterns may or may not
be dissimilar from the ones pointed out by classical narratology, but
they can be derived automatically from the data. The more the data, the
more powerful the inferences about patterns.

Time in AI
AI answers question (1) by constructing formal, mathematical models
that represent events and their relationships over time. These models
are used by artificial agents that write or interpret narratives.
Now, mathematical frameworks for reasoning about time have a
long history in the fields of philosophy and logic, and they have also
been exploited in computer science for practical tasks involving the
specification and verification of the behavior of concurrent systems,
i.e., where more than one computation is going on simultaneously3 The
AI models that I am going to focus on are drawn from these frame-
works, though they are adapted to the conceptualizations found in hu-
man languages. They are thus also relevant to question (2).
Unlike the purely mathematical abstraction of time as an infinite
succession of infinitesimal instants, natural languages conceptualize
time in terms of successions of events that each have a certain duration.
The AI models I consider therefore treat events as occupying finite in-
tervals of time; a character can have lunch at 3, or between 1 and 2, but
in either case the narrator is conceptualizing the event as taking time.
The axis of time itself can however be infinite if desired. Another key
aspect of natural language narratives is the subject-centered, or tensed,
view of time; an event can be spoken of at one time as being in the fu-
ture, and at another time as being in the past. In answering question (2),
these AI systems can interpret these tensed descriptions, locating the
event in time, where possible, irrespective of how it was conceptual-
ized linguistically.

Steegmuller, Francis (1991). 'Translator's Introduction to Madame Bovary, by
Gustave Flaubert'. New York: Random House/Quality Paperback Book Club.
Eminently practical examples of these tasks include the design of signaling rules on
railways, collision avoidance systems on aircrafts, safety measures in nuclear
reactors, and the checking and debugging of software and hardware.
The Flow of Time in Narrative 221

Let us turn first to answer (1). The framework used is called the
temporal interval algebra, or more commonly, the interval calculus.4
Let us assume that events are treated as time intervals. Pairs of events
A and B are ordered with respect to each other by means of seven rela-
tions, shown in Table 1.
I have simplified Allen's representation to ignore inverse relation-
ships (e.g., A AFTER B); his scheme has thus thirteen relations instead
of the seven shown here. All the relations can in fact be expressed in
terms of the relation MEETS; for example, A is BEFORE B if there ex-
ists an interval C such that A MEETS C and C MEETS B.



Table 1. Temporal Relations (simplified) in the Interval Calculus

Using this interval calculus, pairs of relations can be composed togeth-

er to make inferences. Thus, an AI system can reason that if a character
drank a glass of wine during dinner, and if her companion dipped that
glass in arsenic solution before dinner, by transitivity, the dipping pre-
ceded the drinking. (The full Allen system thus has 13x13=169 entries
in its transitivity table). Given a set of events, some of which are re-
lated by one of these seven relations, the system can try to complete a
network (called a temporal constraint network) of possible relation-
ships between all the events. A considerable body of research, e.g.,
Meiri5 has focused on developing efficient algorithms for carrying out
Allen, James (1984). 'Towards a General Theory of Action and Time,' in: Artificial
Intelligence 23: 123-154.
Meiri, Itay (1996). 'Combining Qualitative and Quantitative Constraints in
Temporal Reasoning,' in: Artificial Intelligence 87: 343-385.
222 InderjeetMani

such completions and numerous practical tools are available to do just

that. Once the algorithm has been applied, a system can answer ques-
tions about the orderings of any pair of events, even about those order-
ings which were not given.
Note that in many situations there may be pairs of events where
there is no ordering. This is not only the case for instances of achrony,
an extreme case of which is found in Robbe-Grillet's Jealousy, where
the ordering of most of the events is deliberately omitted; most texts
will consist of such partial orderings.
The temporal ordering relations between intervals can be converted
to a representation that reasons about distances in time, as discussed in
Meiri. For example, consider two events from Chekhov's story The
Lady with the Pet Dog. The event X of Gurov seducing Anna is before
the event Y of his confronting her at the theater. If we define both
events X and Y numerically in terms of start and end points along a
time line, the computer can trivially infer that (ystart-xend) results in a
positive value, where X starts at instant xstart and ends at instant xend
(likewise, Y starts at instant ystart and ends at instant yend). Here in-
stants may be viewed as bounding the intervals. Note that the instants
need not be fully anchored on the calendar. There may be even stronger
constraints, e.g., bounds such as (ystart-xend) < 5 months, since Gurov
sees her off at the station in Yalta at the beginning of Fall, several
weeks after first meeting her, and then visits her in the theater that very
As one might suspect, instants cannot be entirely dispensed with.
They are required for efficient computation as well as for expressing
situations involving continuous motion. However, one has to be careful
mixing points and intervals, for they co-exist uneasily together: for ex-
ample, viewing instants as part of intervals leads to the quandary of not
knowing which event holds at the point where two events meet.
So far, we have treated events as time intervals. Allen allows one to
distinguish between the two, using a mathematical logic that has events
occurring during an interval, either for the entire interval or for some
part of it. Allen's representation confines itself to a linear model of
time, where time does not branch towards the future (or past). This
design decision is based on modularity; a separate reasoning engine
would be required to reason about hypothetical situations in the future
or past, and as Allen argues, it would not be time-specific. However,
there are AI logics for branching time, e.g., Computational Tree Logic
(CTL).6 In CTL, it is possible to specify what can happen along paths
See Huth, Michael, and Ryan, Mark (2004). Logic in Computer Science. Modelling
and Reasoning about Systems. (Second Edition). Cambridge: Cambridge UP.
The Flow of Time in Narrative 223

in time, using operators (quantifiers) that specify that for some (or all)
paths in time there exists a future or past state (or states) where the pro -
position P is true. In effect, it allows a computer to express what Steph-
en Albert tells his assassin Hsi P'eng, in Borges' The Garden of Fork-
ing Paths, 'Time is forever dividing itself toward innumerable futures
and in one of them I am your enemy.' That would be stated precisely as
ExistsPath[Future-State[enemy(sa, hp)]]. A CTL reasoner (called a
model-checker) can verify whether a given statement (expressed in this
formally precise way) is true, when furnished with a model of the
events along the time branches.

Computable Mappings
I will now address question (2), answering it in terms of the comput-
able mapping from natural language narrative to the temporal con-
straint network. The mapping in the inverse direction will also be dis-
cussed, albeit briefly.
Natural language relies on devices for expressing the positioning of
information in time, relative to the time of the utterance (the speech
time) and the time of the event or object being described. Language
captures the when in terms of time expressions ('three o'clock','three
days a year', 'tomorrow'), systems of tense (past, present, and future)
and aspect (namely whether an event is ongoing, completed, etc.), as
well as expressions of temporal relations ('before', 'during', 'at', etc.).
With the past tense, the event occurs prior to the speech time; for the
present, it occurs roughly at the speech time; and with the future tense
(which in Germanic languages involves modal notions), the event time
is later than the speech time. The way these components work and in-
teract can vary greatly across languages; for example, the Bantu lan-
guage ChiBemba has four past tenses and four future tenses, while
Mandarin Chinese uses aspect-indicating particles instead of tense
markers. In some languages, like Burmese, there is no tense, but
particles are used to indicate whether an event is real or hypothetical.
An AI reader has to be highly cognizant of these mechanisms in the
language in which the narrative is expressed.
In order to get from a narrative to a temporal constraint network,
modern AI algorithms make use of machine learning from examples.
Humans first annotate example narratives with features, indicating the
events and times and their temporal relations. The features pertain to
the verbs and other parts of speech indicating the events of interest, and
the various time expressions. The time expressions are resolved, so
that, for example, 'last March' and 'November' are anchored to calen-
224 InderjeetMani

dar years given the speech time when those expressions were uttered,
and 'seven years earlier' are anchored with respect to earlier mentioned
reference times. Then the events are linked to the times using at most
one of the seven temporal relations in Table 1.
As an example, consider the opening sentence of a Garcia Mar-
quez's One Hundred Years of Solitude:

Many years later, as he faced the firing squad, Colonel Aureliano Buendia
was to remember that distant afternoon when his father took him to dis-
cover ice.

The time expression 'Many years later' would get marked as a period P
of a vague number X of years Y after a reference time RT, i.e.,
tl=RT+PXY. The event 'faced' would be tagged as FACE[+PAST], 're-
meaning that it is in infinitive form, that it is an intensional verb (what
is remembered is in the mind), and that it represents a state. The tem-
poral relations might include FACE being DURING tl, and its being
SIMULTANEOUS with REMEMBER, etc. These sorts of linguistic
features are recorded in a markup language called TimeML,7 which
happens to be an international ISO standard. Since humans also have
strong intuitions about how long particular events last (e.g., invasions
last longer than sneezes), it is also possible to add to the markup estim-
ates of the minimum and maximum bounds for events, as Pan et al.8
have done for TimeML (so as to cover 80% of the probable scenarios
given the text context). Humans tend to agree almost 90% of the time
on such bounds.
A visual display of the completed temporal constraint network for
the Garcia Marquez opening sentence is shown in Figure 1. The left
and right show the network before and after automatic completion, re-
spectively. Here, both events and times are treated as time intervals,
and are given subscripts indicative of narration order, whereas the or-
der of occurrence (the chronology) of the events is displayed left-to-

Pustejovsky, J., Ingria, B., Sauri, R., Castano, J., Liftman, J., Gaizauskas, R., Seteer,
A., Kate, G., and Mani, I. (2005). 'The Specification Language TimeML,' in: Mani,
I., Pustejovsky, J., and Gaizauskas, R (eds.). The Language of Time: A Reader. New
York: Oxford UP, 549-562.
Pan, Feng, Mulkar, Ritu, and Hobbs, Jerry (2006). 'Learning Event Durations from
Event Descriptions,' in: Proceedings of the 44th Annual Meeting of the Association
for Computational Linguistics (COLING-ACL'2006), Sydney, Australia, 393^00.
The Flow of Time in Narrative 225

that distant aft.. lany years lat..

f i:ed(e1)

ttol<e3) rbmembei(e2)


any ye a is I at..




Figure 1: Temporal Constraint Network, Before and After Completion

226 InderjeetMani

The dark lines with arrows indicate BEFORE relations (the interval
at the arrow's tail being BEFORE the one at its head), while the ones
with circles and squares indicate DURING and SIMULTANEOUS rela-
tions, respectively, with the interval below being DURING (or SIMUL-
TANEOUS with) the one above. The light lines will be explained later.
As 'that distant afternoon' is earlier in time than the 'many years later',
the computer infers in the right-hand diagram that Aureliano's father's
taking him to discover ice (event e3) is before his facing of the firing
squad (event el).
For a more complicated example, consider that old narratological
chestnut from Proust's Jean Santeuil discussed by Genette:9

Sometimes passing in front of the hotel he remembered the rainy days

when he used to bring his nursemaid that far, on a pilgrimage. But he re-
membered them without the melancholy that he then thought he would
surely some day savor on feeling that he no longer loved her.

(Here the completion does not add any new links). It can be seen that
the passing is SIMULTANEOUS with the sometimes, the remembering
is DURING the passing, the bringing is DURING the rainy days, the
not loving is BEFORE the savoring, etc.
To provide for automatic timelining, humans have to label hundreds
of narratives with such information, creating a database of thousands of
examples of temporal relations. These training examples are grouped
into classes, one per temporal relation, and the machine learning pro-
gram then tries to learn rules to discriminate among the different
classes. This is done by learning which combination of features best
predict a given class, based on the frequency of the features and the
class in the data. Once trained, a new example that has not been annot-
ated will be automatically classified based on its similarity, in terms of
presence of features, to the patterns in the learned rules. Such a statist-
ical learning method allows the system to infer the times, events, and
temporal relations for any new narrative. The statistics however need to
be supplemented by rules from human intuitions, as the statistical fea-
tures will not pick up patterns that are relatively rare in the training

Genette, Gerard (1980). Narrative Discourse. Ithaca: Cornell UP.
The Flow of Time in Narrative 227



r imemberedf^)

the rainy daysCt2) []






some day(t5)



Figure 2: Another Temporal Constraint Network

228 InderjeetMani

While such approaches work reasonably well on news stories10 and

have been used for a variety of languages and text genres, they have
not been applied to anything more than small fragments of literary nar-
ratives. The serious bottleneck here is the burden of creating a suffi-
cient quantity of training data. Such a detailed level of micro-annota-
tion is infeasible for a human to create for long narratives such as nov-
els; for those, perhaps annotating novel summaries might work instead.
While large-scale computer timelining of literary narratives is currently
more of a promise than a reality, once corpora of literary texts are ere -
ated and marked up with time lines, the computer can mark up time
lines for thousands of other literary narratives. This can help to gener-
ate new data, test hypotheses, and hopefully enhance the reading exper-
Given a temporal constraint network, with some or all of the text or-
der unspecified, an AI author can also produce a narrative. In such a
case, the system has to generate a coherent narrative, part and parcel of
which is selecting a narrative ordering of the events from the given
chronology. For example, the system nn, from Montfort11, is an interact-
ive fiction system that provides a narrative of events in a dynamic mi-
cro-world, representing both the 'actual world' of simulated reality as
well as what each focalizer (i.e., character from whose point-of-view
the world is experienced) knows about that world. Instead of compos-
ing a story from scratch, it is given an existing chronology of events,
which it then narrates according to various focalization strategies. Most
relevant for our purposes, it implements Genette's seven different ways
of ordering events. It implements chronicle, retrograde, and achrony by
narrating the input events in chronological, reverse chronological, and
random order, respectively. Zigzag involves interleaving between the
two time periods, while for analepsis the past event to flash back is
chosen based on heuristics such as selecting among the most salient
events that the focalizer has seen happen in the artificial world in the
past, nn also reasons about the tense to use for narration, along with
connecting words such as 'then' or 'before' to express temporal rela-
tions. While nn uses rules to decide about how much discourse time to
devote to different types of events, it does not reason as to when it is
Mani, I., Wellner, B., Verhagen, M., Lee, C. M., and Pustejovsky, J. (2006).
'Machine Learning of Temporal Relations', in: Proceedings of the 44th Annual
Meeting of the Association for Computational Linguistics (COLING-ACL), Sydney,
Australia, 753-760.
Montfort, Nick (2007). 'Ordering Events in Interactive Fiction Narratives,' in:
Proceedings of the AAAI Fall Symposium on Intelligent Narrative Technologies.
Technical Report FS-07-05, AAAI Press, 87-94.
The Flow of Time in Narrative 229

best to use any given one of the seven orderings. AI authors like nn are
limited not only by the simple logic of their micro-worlds, but by the
story content at the level of the fabula having to be pre-specified in ad-

These temporal constraint networks, more generally, are display-inde-
pendent mathematical objects called graph structures, which are used
to model relationships between entities; here the entities are events or
times, and the links between them are temporal relations. The networks
are flexible enough to represent temporal relations and distances (along
with bounds) in any narrative, irrespective of the order in which the
events are enumerated in the text. They can also represent branches and
cycles in time. However, they do reflect the static view associated with
the B-series, to which any A-series language is reduced.
Is such a temporal constraint network a suitable representation for a
narrative's temporal properties? The ordering offered in this represent-
ation captures Genette's sense of 'story' (or 'histoire'). Thus, analepses
will involve events with higher subscripts occurring earlier to the left
than those with lower numbers. However, the temporal relations being
considered in the time line are more expressive than merely precedence
and equality in 'histoire', since we allow for time intervals which are
related in seven different ways. It also distinguishes between story and
discourse, capturing the story time in the ordering of temporal relations
and partial calendar anchoring of events and times. The ordering of
events and times in the discourse is also evident. Discourse time itself
is not directly captured, though this is a matter of detail; the indices
(e3, etc.) might be easily extended to include a measure of offset into
the narrative. Once that is added, the story time of a narrative (the time
it occupies in the time line) can be compared to its discourse time, e.g.,
the number of words used to recount the event. The result of such com-
parison, which Genette classifies into isochronous, accelerated, and de-
celerated tempos, can be a valuable derivative product from the time
line. For a detailed treatment along these lines of temporal ordering,
tempo, and other time-related phenomena in fiction.12
As stated earlier, the adequacy of such a representation must be as-
sessed in terms of its ability to capture the varieties of temporal and
modal phenomena found in narrative, as well as, if desired, its
Mani, I. (2010). The Imagined Moment: Time, Narrative and Computation.
Lincoln: U of Nebraska P.
230 InderjeetMani

verisimilitude compared to humans. Turning to the latter, how are we to

deal with the variability across humans in terms of interpretations of
the time line? The key here is to admit that interpretation is not inde-
terminate, in contrast to the view of literary scholars such as Herrnstein
Smith,13 who have taken the existence of multiple narratives of a given
folk-tale as an argument in favor of indeterminacy. While interpretation
is strongly constrained by language, a human marking up the time line
of a narrative will have to adhere to a principle of conservative inter-
pretation. This interpretation is one that is carried out based on inform-
ation from the text, the reader's knowledge of language, and her know-
ledge of the real world, without appealing to guesses or 'unusual' cir-
cumstances not mentioned in the text, or relying on specific circum-
stances related to the author of the work, its creation, or the reader's
prior experience. This sort of approach has proved feasible in similar
linguistic annotation tasks requiring commonsense inferences. It is an
additional constraint on top of Ryan's 'Principle of Minimal Departure'
(PMD),14 which posits that readers fill in gaps in the text by assuming
the similarity of the fictional world to their own experienced actual
world (an idea also echoed in Eco15): in our case, the only gaps filled in
are the ones based on information mentioned in the text and the
reader's knowledge of language and the world.

Reasoning about Modality

I said at the beginning that time is wound up with events, and that it in-
volves relationships with modality and point-of-view. Here I will con-
sider modality. We touched on modality when talking about possible
futures, but now let us focus on modalities that are not strictly tempor-
The central idea in the classical treatment of modality is that an
agent knows something if it is true in all situations that she deems pos-
sible.16 These situations are called possible worlds; the worlds deemed
possible for an agent are determined by an 'accessibility' relation. The
metaphor of a world here is a literary one, but the sense of the term is
Herrnstein Smith, Barbara (1980). 'Narrative Versions, Narrative Theories,' in:
Critical Inquiry 1: 209-218.
Ryan, Marie-Laure (1991). Possible Worlds, Artificial Intelligence, and Narrative
Theory. Bloomington: Indiana UP.
Eco, Umberto (1990). 'Small Worlds,' in: U. Eco: The Limits of Interpretation.
Bloomington: Indian UP, 64-82.
Kripke, Saul (1959). 'A Completeness Theorem in Modal Logic,' in: Journal of
Symbolic Logic 24.1: 1-14.
The Flow of Time in Narrative 231

purely mathematical, involving an algebraic structure consisting of

states with properties that can be true or false, and the accessibility re-
lation that defines which states are possible for the agent given each
state the agent can be in. The algebraic properties of the accessibility
relation determine the limits of an agent's knowledge.17 If the accessib-
ility relation is reflexive, then the actual (real) world is one of the pos-
sible worlds, i.e., the agent cannot know things that are false. If it is
transitive, then the agent knows whatever she knows (i.e., she knows
all logical implications of everything she knows); if it is symmetric and
transitive, then she knows what she does not know. Further, if a group
of agents have mutual or common knowledge of something, then each
knows that the other knows that the other knows ... it. Obviously, mod-
al logics with such properties are far too omniscient in comparison
with humans; but they are perhaps adequate to represent the far greater
epistemic powers of literary characters.
The theory of possible worlds also considers other propositional at-
titudes that involve relations with mental states, such as beliefs, ima-
ginings, rememberings, and so forth. Thus, the proposition 'Buendia re-
membered that they discovered ice' is necessarily true in the actual
world if and only if 'they discovered ice' is true in all possible worlds
corresponding to his memories. Modal logics fall into a number of lo-
gical families, with plain alphanumeric names like K, T, S4, S5, etc; for
an introduction.18 There are also a variety of theorem-provers available
for these logics.19
In Figure 1, to express the fact that being taken to discover is a
memory in Buendia's mind, TimeML uses a subordinating link, shown
as a light-colored arrow, between 'remember' and 'take', and likewise
between 'take' and 'discover' (for the discovery is hypothetical; the
sentence does not indicate that the discovery took place). Likewise, in
Figure 2, Jean's bringing and thinking are subordinated to the remem-
bering, and the savoring to his thinking. These subordinating links
serve as a rough shorthand for modal notions, as well as other kinds of
subordination relations such as narration. However, they do not address
the question of the relation between individuals in these different pos-
sible worlds. For example, for the Jean Santeuil passage, Figure 2 de-
clares the rainy days that Jean remembered to be SIMULTANEOUS
with the ones that are postulated by the narrator - and referred to by
Halpern, Joseph Y, and Moses, Y (1992). 'A guide to completeness and complexity
for modal logics of knowledge and belief,' in: Artificial Intelligence 54: 319-379.
See Hughes, G. E., and Cresswell, M.J. (1982). An Introduction to Modal Logic.
London: Methuen.
For example, see
232 InderjeetMani

'then', but the simultaneity should perhaps instead hold between the
narrator's rainy days and the counterpart20 of the remembered days in
the narrator's world.
In applying possible world semantics to fiction, there has been some
narratological discussion of proposals such as the idea of Lewis21 of
treating any fictional proposition P as a counterfactual conditional, so
that the fictional proposition P by the narrator that Jean passed in front
of the hotel is true if and only if there is some possible world where (i)
both the facts in Jean Santeuil are true and P is true, and (ii) that is
closer to the actual world than every possible world where those facts
are true and P is false. Ryan22 has adopted an interesting variant of this
idea, where the possible worlds with the facts in Jean Santeuil being
true are 'actual' with respect to the text, i.e., such a possible world is a
textual actual world (TAW), namely the representation proposed by the
However, a vague appeal to similarity between worlds offers little
help to a modal reasoner. While there has been encouraging progress in
theorem-proving using the modal theory of abstract objects (including
fictional ones) by Fitelson and Zalta,23 far more research is needed on
modal theories applied to fictional discourse.
Possible world analyses of fiction nevertheless provide a valuable
critical device even when shorn of the power of automatic AI computa-
tion. In the case of post-modernist fiction, as McHale points out,24 the
existence and structure of these worlds are emphasized by the author so
as to expose the artifice of their creation; and as Eco illustrates, pos-
sible worlds are crucial in illuminating one function that fiction ap-
pears to serve: 'Fiction suggests that perhaps our view of the actual
world is as imperfect as that of fictional characters'.25

Insights from Cognitive Science

The AI perspective enumerated thus far is not informed by experimental
discoveries in cognitive science. To attempt to answer question (3), one
needs to understand what psychology and neuroscience tells us about our
human capabilities related to the understanding of time in narrative.
Lewis, David (1973). Counter/actuals. Cambridge: Harvard UP.
Ryan: Possible Worlds (cf. Fn 14).
Fitelson, Brian, and Zalta, Edward (2007). 'Steps Toward a Computational
Metaphysics,' in: Journal ofPhilosophical Logic 36.2: 227-247.
McHale, Brian (1987). Postmodernist Fiction. London, Routledge.
Eco: Small Worlds (cf. Fn 15), 74.
The Flow of Time in Narrative 233

Let us first consider the brain. Human brains have three different
systems to detect time, operating on different timescales. As a review
by Buhusi and Meek indicates,26 circadian rhythms are relatively pre-
cise and operate on a 24-hour cycle; they are involved in the regulation
of appetite and sleep. Interval timing is used in tasks such as decision-
making and multi-step arithmetic, and is less precise, but it is able to
operate in a broad seconds-to-minutes-to-hours range. Finally, milli-
second timing operates in the impressive sub-second range and used in
speech, playing music, and dancing. These three systems involve very
different brain areas. While the clock for millisecond timing is located
in the cerebellum, circadian rhythms originate from the hypothalamus
(which controls, based on the amount of light, the secretion of the tran-
quilizing hormone melatonin from the pineal gland). Interval timing is
not as localized and relies on the activation of a number of different
circuits spread across the brain. There is suggestive evidence that these
latter circuits are also involved in other processing such as the estima-
tion of quantity. All three of these systems are together involved in our
sense of time.
Consider next our ability to recognize events in the real world.
Thanks to evolution, our brains have very sensitive event detectors. An
event in the real world needs to last only three to four hundredths of a
second for it to be distinguished from another event, as shown in the
surveys by Poppel.27 The evidence comes in part from experiments by
Kanabus et al.28 where subjects have to judge the order of visual stimuli
(such as flashing green and red lights) or auditory ones (low and high
tones) and decide on whether they are before, simultaneous with, or
after one another.
To continue to perceive and act, the brain has to hold the percept it
has formed in memory, while querying the outside world as to what has
changed. This holding can only last for up to 3 seconds, after which the
brain updates the information. As Poppel (1994) indicates, when sub-
jects are made to hear computer-generated sequences of syllables such
as 'SO' and 'MA', they perceive either 'SOMA' or 'MASO', switching
to the alternative interpretation after 3 seconds. Turner and Poppel
Buhusi, Catalin V, and Meek, Warren H. (2005). "What makes us tick? Functional
and neural mechanisms in interval timing,' in: Nature Reviews Neuroscience, 6,
October 2005, 755-765.
Poppel, Ernst (1994). 'Temporal mechanisms in perception,' in: International
Review of Neurobiology 37, 185-202; Poppel, Ernst (1997). 'A hierarchical model
of temporal perception,' in: Trends in Cognitive Sciences 1.2: 56-61.
Kanabus, M., Szelag, E., and Poppel, E. (2002). 'Temporal order judgement for
auditory and visual stimuli,' in: Acta Neurobiol. Exp. 62: 263-270.
234 InderjeetMani

have speculated that the present lasts at most 3 seconds, and that this
bound corresponds to the average duration of what they call a 'line' of
verse across cultures and languages.29 As an example, Shakespeare's
opening line 'Shall I compare thee to a summer's day?' (from Sonnet
XVIII) lasts a little over 3 seconds when recited at 'normal' speed.
Now, let us turn to higher-level cognitive function in terms of under-
standing time in narrative. Zwaan shows that readers expect that suc-
cessive sentences in a narrative will describe chronologically success-
ive and temporally adjacent events.30 Deviations from this narrative
format will result in delays in processing information. Zwaan found
that sentences with a time shift between events (for example, expressed
with the phrase 'an hour later') take longer to read, and result in sub-
jects taking a longer time to digest and answer questions about whether
a particular word occurred in the story, in comparison with similar sen-
tences lacking such a shift. Events separated by a time shift were less
strongly connected in long-term memory than those that were not sep-
arated by a narrative time shift. Also, when processing a narrative se-
quence of immediately successive events, readers took longer to access
events that, although mentioned recently, were temporally remote from
the current narrative 'now'. This temporal distance effect was absent
when the text had a time shift.
Narratives that focus on time itself can be particularly difficult to
process temporally. Alan Lightman's novel Einsteins Dreams explores
temporally exotic situations: in one chapter, time is circular, in another,
time can branch back to the past; time stands still in one, and in another
there is no time. Experiments by Graesser et al.31 show that subjects
found it hard to imagine the situations in Einstein s Dreams, faring
poorly in making inferences based on them.
These findings suggest that readers build cognitive models in their
minds of the situation described by the narrative. They represent
whether events are before or after each other, and how far apart they

Turner, Frederick, and Poppel, Ernst (1980). 'The neural lyre. Poetic meter, the
brain, and time,' in: Poetry, 277-309.
Zwaan, Rolf A. (1996). 'Processing narrative time shifts,' in: Journal of
Experimental Psychology: Learning, Memory, and Cognition 22: 1196-1207.
Graesser, Arthur C , Olde, Brent, and Klettke, Bianca (2003). 'How does the mind
construct and represent stories?' in: Narrative Impact: Social and Cognitive
Foundations. Mahwah: Lawrence Erlbaum; Graesser, Arthur C , Kassler, M.A.,
Kreuz, R.J., and McLain-Allen, B. (1998). 'Verification of Statements about Story
Worlds That Deviate from Normal Conceptions of Time: What Is True about
Einstein's Dreams?' in: Cognitive Psychology 35.3: 246-301.
The Flow of Time in Narrative 235

are in time. Readers do appear to construct time lines, preferring simple

The empirical limits for how short events can be and how long the
present are grounds for treating non-instantaneous events as primitive
elements in our time lines. Since the AI time lines provide an explicit
(though under-specified) representation of event chronology and nar-
rated order, the effect of different types of anachronies and their com-
plexity on narrative understanding can be studied experimentally in
terms of specific representational elements (number of order reversals,
number of temporal links in time lines like Figures 1 and 2, etc.). More
general psychological questions can also be posed. For example, what
inventory of temporal relations seem to be involved? How do time lines
get consolidated, simplified, or blurred in memory? How does tempo
get perceived as a function of discourse time versus story time? Does
representing the time line explicitly improve the appreciation of the

I have argued here that AI models are able to tell us what sorts of reas-
oning about time and events an intelligent agent can carry out, as well
as showing how natural language narratives can be mapped to these AI
representations. In doing so, I have tried to address the many insights
on time in narrative that narratologists have arrived at. I have also sum-
marized some relevant results from cognitive science that have a bear-
ing on how humans interpret time in narrative, suggesting how such
results may be further elaborated by psychological experiments based
on the AI representations.
These investigations are only a first step, I believe, in a much richer
examination of time in narrative, within the framework of an empirical
discipline of corpus narratology, where multimillion-word collections
of narrative texts are analyzed using sophisticated Al-based timelining
tools. Such an effort has the potential to alter the foundations of narrat-
ive theory.32 Even if we can ultimately never know time, or even say
what it is, the methodologies sketched here may allow us to engage
more deeply with the manners in which time is explored and often bril-
liantly exploited in literary narrative.

Mani: The Imagined Moment (cf. Fn 12).
Bibliography: A Guide to Further Reading

The following list, though far from exhaustive, offers a sample of the
most valuable and representative literature on time. Titles are arranged
by category.


Fraser, Julius (1981). 'A Report on the Literature of Time, 1900-1980,' in:
The Study of Time IV. Papers from the Fourth Conference of the Interna-
tional Society for the Study of Time. Ed. by J.T. Fraser, N. Lawrence, D.
Park. New York, Heidelberg, Berlin: Springer, 234-266
Macey, Samuel L. (1991). Time. A Bibliographic Guide. New York, London:
Macey, Samuel L., ed. (1994). Encyclopedia of Time. New York, London:
Sandbothe, Mike in collaboration with Stephan Giirtler, Konstantin Pollok and
Nikolai von Rosen (1993). 'Auswahlbibliographie,' in: Klassiker der Zeit-
philosophie. Ed. by Walther Ch. Zimmerli, Mike Sandbothe. Darmstadt:
WBG, 299-309


Baert, Patrick, ed. (2000). Time in Contemporary Intellectual Thought. Ams-

terdam: Elsevier
Baumgartner, Hans Michael, ed. (1994). Zeitbegriffe und Zeiterfahrung.
Freiburg, Miinchen: Alber
Burger, Heinz, ed. (1986). Zeit, Natur und Mensch. Beitrage von Wissen-
schaftlern zum Thema "Zeit". Berlin: Berlin-Verlag Spitz
Helferich, Hede, ed. (1996). Time and Mind. Seattle: Hogrefe & Huber
Hess-Luttich, Ernest W.B., andBrigitte Schlieben-Lange, eds. (1998). Signs &
Time. Zeit & Zeichen. An International Conference on the Semiotics of
Time in Tubingen. Tubingen: Narr
Kablitz, Andreas, Wulf Oesterreicher and Rainer Warning, eds. (2003). Zeit
und Text. Philosophische, kulturanthropologische, literarhistorische und
linguistische Beitrage. Miinchen: Fink
Maini, Inderjeet, James Pustejovsky and Rob Gaizauskas, eds. (2005). The
Language of Time. A Reader. Oxford: Oxford UP
238 Bibliography

Middeke, Martin, ed. (2002). Zeit und Roman. Zeiterfahrung im historischen

Wandel und asthetischer Paradigmenwechsel vom sechzehnten Jahrhun-
dert bis zur Postmoderne. Wiirzburg: Konigshausen & Neumann
Oaklander, L. Nathan, and Quentin Smith, eds. (1994). The New Theory of
Time. New Haven, London: Yale UP
Peisl, Anton, and Armin Mohler, eds. (1983). Die Zeit. Miinchen, Wien: 01-
denbourg (= Schriften der Carl-Friedrich-von-Siemens-Stiftung. 6)
Ritter, Alexander, ed. (1978). Zeitgestaltung in der Erzahlkunst. Darmstadt:
Rusterholz, Peter, and Rupert Moser, eds. (1997). Zeit. Zeitverstandnis in Wis-
senschaftundLebenswelt. Bern: Lang
Simonis, Annette, and Linda Simonis, eds. (2000). Zeitwahrnehmung undZeit-
bewufitsein in derModerne. Bielefeld: Aisthesis
Stadler, Friedrich, and Michael Stotzner, eds.(2006). Time and History. Pro-
ceedings of the 28. International Ludwig Wittgenstein Symposium Kirch-
berg am Wechsel, Austria 2005. Heustenstamm: Ontos (= Publications of
the Austrian Ludwig Wittgenstein Society. New Series 1)
Zimmerli, Walther Ch., and Mike Sandbothe, eds. (1993). Klassiker der Zeit-
philosophie. Darmstadt: WBG

General and Philosophical Approaches to Time

Aichelburg, Peter C. 1997). 'Zeit im Wandel der Zeit. Zur Entwicklung des
Zeitbegriffs,' in: Zeit. Zeitverstandnis in Wissenschaft und Lebenswelt. Ed.
by Peter Rusterholz and Rupert Moser. Bern: Lang, 75-96
Alves, Pedro M.S. (2008). 'Objective Time and the Experience of Time:
Husserl's Theory of Time in Light of Some Theses of A. Einstein's Special
Theory of Relativity,' in: Husserl Studies 24: 205-209
Alweiss, Lilian (2002). 'Heidegger and "the concept of time",' in: History of
the Human Sciences 15: 117-132
Andersen, Holly and Rick Crush (2009). 'A Brief History of Time-Conscious-
ness: Historical Precursors to James and Husserl,' in: Journal of the His-
tory of Philosophy 47: 277-307
Baumgartner, Hans Michael (1994). 'Zeit und Zeiterfahrung,' In: Hans Mi-
chael Baumgartner (ed.): Zeitbegriffe und Zeiterfahrung. Freiburg, Miin-
chen: Alber, 189-216
Berlinger, Rudolph (1953). 'Zeit und Zeitlichkeit bei Aurelius Augustinus,' in:
Zeitschriftfurphilosophische Forschung 7: 493-510
Bernet, Rudolf (1987/88). 'Die Frage nach dem Unsprung der Zeit bei Husserl
und Heidegger,' in: Heidegger Studies 3/4: 89-104
Bieri, Peter (1972). Zeit und Zeiterfahrung. Exposition eines Problembereichs.
Frankfurt/M: Suhrkamp
Bieri, Peter (1986). 'Zeiterfahrung und Personalitat,' in: Heinz Burger (ed.):
Zeit, Natur und Mensch. Beitrage von Wissenschaftlern zum Thema
"Zeit". Berlin: Berlin-Verlag Spitz, 261-281
A Guide to Further Reading 239

Brentano, Franz (1920). 'Zur Lehre von Raum und Zeit. Aus dem Nachlasse
von Franz Brentano,' in: Kant-Studien 25: 1-23
Brough, JohnB. (2006). 'Husserl and the Deconstruction of Time,' in: Review
ofMetaphysics 46: 503-536
Cresswell, M.J. (2006). 'Now is the Time,' in: Australasien Journal of Philo-
sophy 84:311-332
Dodd, James (2005). 'Reading Husserl's Time-Diagrams from 1917/18,' in:
Husserl Studies 21: 111-137
Dowden, Bradley. 'Time,' in: The Internet Encyclopedia of Philosophy
Diising, Klaus (1980). 'Objektive und subjektive Zeit. Untersuchungen zu
Kants Zeittheorie und zu ihrer modernen kritischen Rezeption,' in: Kant-
Studien 11: 1-34
Dux, Giinter (2003). 'Strukturen der Zeit. Zeit der Handlung, Zeit des Systems,
Zeit der Geschichte,' in: Andreas Kablitz, Wulf Oesterreicher, Rainer
Warning (eds.): Zeit und Text. Philosophised, kulturanthropologische,
literarhistorische und linguistische Beitrage. Miinchen: Fink, 13-32
Ermarth, Elizabeth Deeds (2010). 'What If Time Is a Dimension of Events,
Not an Envelope for Them?,' in: Time & Society 19.1: 133-150
Findlay, J.N. (1941). 'Time: A Treatment of some Puzzles,' in: The Australasi-
en Journal of Psychology and Philosophy 19: 216-235
Findlay, J.N. (1976). 'Husserl's Analysis of the Inner Time-Consciousness,' in:
The Monist 59: 3-20
Fine, Kit 2006). 'The Reality of Tense,' in: Synthese 150: 399-414
Flasch, Kurt (1993). Was ist Zeit? Augustinus von Hippo. Das XL Buch der
Confessiones. Historisch-philologische Studie. Text- Ubersetzung - Kom-
mentar. Frankfurt/M.
Franck, Georg (1989). 'Das Paradox der Zeit und die Dimensionszahl der
Temporalitat,' in: Zeitschrift fur philosophische Forschung 43: 449-471
Fraser, Julius T. (1967). 'The Interdisciplinary Study of Time,' in: Annals of
the New York Academy of Sciences 138: 822-847
Fraser, Julius T. (1970). 'Time as a Hierarchy of Creative Conflicts,' in: Stadi-
um Generate 23: 597-6S9
Friedman, Laurence (1953/54). 'Kant's Theory of Time,' in: Review of Meta-
physics 7:379-388
Gadamer, Hans-Georg (1983). 'Das Ratsel der Zeit,' in: Universitas 38: 453-
Griesser, Wilfried (2008). 'Zeit und Zeitlichkeit beim spaten Paul Natorp und
bei Heidegger,' in: Philosophisches Jahrbuch 115.2: 261-287
Griinbaum, Adolf (1967). 'The Status of Temporal Becoming,' in: Annals of
the New York Academy of Sciences 138: 374-395
Hodges, Matt (2008). 'Rethinking time's arrow. Bergson, Deleuze and the an-
thropology of time,' in: Anthropological Theory 8: 399-429
Hiibner, Dietmar (2009). 'Gibt es eine objektive Gegenwart? Zur Metaphysik
der Zeit,' in: Philosophisches Jahrbuch 116.2: 269-293
240 Bibliography

Husserl, Edmund (1928). Edmund Husserls Vorlesungen zur Phanomenologie

des inneren Zeitbewufitseins. Ed. by Martin Heidegger. Tubingen (= Son-
derdruck aus Jahrbuch fur Philosophie und phanomenologische For-
schung, vol. 9)
Husserl, Edmund (1991). On the Phenomenology of the Consciousness of In-
ternal Time (1893-1917). Translated by John Barnett Brough. Dordrecht:
Kluwer (= Collected Works. IV)
Husserl, Edmund (2006). Spate Texte tiber Zeitkonstitution (1929-1934). Die
C-Manuskripte. Ed. by Dieter Lohmar. Dordrecht: Kluwer (= Husserliana,
vol. VIII)
Jammer, Max (2007). 'Concepts of Time in Physics: A Synopsis,' in: Physics
in Perspective 9: 266-280
Janich, Peter (1996). 'Die Konstitution der Zeit durch Handeln und Reden,' in:
Kodikas/Code. Ars Semeiotica 19: 133-147
Jordan, Robert (1954/55)- 'Time and Contingency in St. Augustine,' in: Re-
view ofMetaphysics 8: 394-417
Kaspar, Rudolph and Alfred Schmidt (1992). 'Wittgenstein tiber Zeit,' in: Zeit-
schriftfurphilosophische Forschung 46: 569-583
Kortooms, Toine (2002). Phenomenology of Time. Edmund Husserl's Analysis
ofTime-Consciousness. Dordrecht: Kluwer (= Phaenomenologica. 161)
Kroes, Peter (1984). 'Objective versus Minddependent Theories of Time
Flow,' in: Synthese 61: 423-446
Kiimmerl, Friedrich (1962). 'Zeit und BewuBtsein. Entwurf einer Theorie der
Erinnerung als BewuBtwerdung des BewuBtseins,' in: Zeitschrift fur philo-
sophische Forschung 16: 533-560
Le Poidevin, Robin. 'The Experience and Perception of Time,' in: Stanford
Encyclopedia of Philosophy [http://plato.stanford.-edu/entries-/timeecper-
Le Poidevin, Robin (1992). 'On the Acausality of Time, Space, and Space-
Time,' in: Analysis 52: 146-154
Le Poidevin, Robin (1996). 'Time, Tense and Topology,' in: Philosophical
Quarterly 46: 467-481
Le Poidevin, Robin (1998). 'The Past, Present, and Future of the Debate about
Tense,' in: Questions of Time and Tense. Ed. by Robin Le Poidevin. Ox-
ford: Clarendon, 13-42
Le Poidevin, Robin (1999). 'Egocentric and Objective Time,' in: Proceedings
of the Aristotelian Society n.s. 99: 19-36
Le Poidevin, Robin (2003). Travels in Four Dimensions - The Enigmas of
Space and Time. Oxford: Oxford UP
Le Poidevin, Robin (2004). 'A Puzzle Concerning Time Reception,' in: Syn-
these 142:109-142
Le Poidevin, Robin (2006). 'Memory and the A-Series,' in: Friedrich Stadler,
Michael Stotzner (eds.): Time and History. Proceedings of the 28. Interna-
tional Ludwig Wittgenstein Symposium Krichberg am Wechsel, Austria
2005. Frankfurt/M (= Publications of the Austrian Ludwig Wittgenstein
Society. New Series. 1), 31-42
A Guide to Further Reading 241

Le Poidevin, Robin (2007). The Images of Time. An Essay on Temporal Rep-

resentation. Oxford: Oxford UP
Levinson, Arnold B. (1987). 'Events and Time's Flow,' in: Mind 96: 341-353
Markosian, Ned (1992). 'On Language and the Passage of Time,' in: Philo-
sophical Studies 66: \-26
McEvoy, James (1984). 'St. Augustine's Account of Time and Wittgenstein's
Criticisms,' in: Review ofMetaphysics 37: 547-577
McLure, Roger (2005). The Philosophy of Time. Time before Times. London,
New York: Routledge
Mesch, Walter (1998). 'Augustinus als Wegbereiter der modernen Zeittheorie?
Zu den ontologischen Voraussetzungen von Confessions XI,' in: Antike
undAbendlandW: 139-162
Mohanty, J.N. (1988). 'Time: Linear or Cyclic, and Husserl's Phenomenology
of Inner Time Consciousness,' in: Philosophia Naturalis 25: 123-130
Sachs, Mendel (1996). 'Changes in Concepts of Time from Aristotle to Ein-
stein,' in: Astrophysics and Space Science 244: 269-281
Sandbothe, Mike (1997). 'Die Verzeitlichung der Zeit in der modernen Philo-
sophie,' in: Die Wiederentdeckung der Zeit. Ed. by Antje Gimmler, Mike
Sandbothe and Walther Ch. Zimmerli. Darmstadt: WBG
Sandbothe, Mike (1998). Die Verzeitlichung der Zeit. Grundtendenzen der
modernen Zeitdebatte in Philosophie und Wissenschaft. Darmstadt: WBG
Sandbothe, Mike (2000). 'The Temporalization of Time in Modern Philo-
sophy,' In: Patrick Baert (ed.): Time in Contemporary Thought. Amster-
dam: Elsevier, 17-35
Schnell, Alexander (2002). 'Das Problem der Zeit bei Husserl. Eine Untersu-
chung iiber die husserlschen Zeitdiagramme,' in: Husserl Studies 18: 89-
Schnieder, Benjamin (2009). 'Bolzanos Erklarung des Zeitbegriffs,' in: Archiv
fiir Geschichte der Philosophie 91: 42-69
Schulz, Reinhard (1997). 'Die Zeit - Phanomen oder Konstrukt? Hermeneut-
ische Reflexionen zuNatur und Geschichte,' in: Logos NF 4: 96-115
Simon, Josef (1998). 'Zeichen und Zeit und das Problem der Bedeutung,' in:
Ernest W.B. Hess-Luttich, Brigitte Schlieben-Lange (eds.): Signs & Time.
Zeit & Zeichen. An International Conference on the Semiotics of Time in
Tubingen. Tubingen: Narr, 27-40
Stegmaier, Werner (1987). 'Zeit der Vorstellung. Nietzsches Vorstellung der
Zeit,' in: Zeitschriftfurphilosophische Forschung 41: 202-228
Tarouca, Amadeo Silva (1962). 'Moment und Augenblick. Reflexionen zur
Philosophie der Zeit,' in: Zeitschrift fiir philosophische Forschung 16:
Teichmann, Roger (1995). The Concept of Time. Basingstoke: Macmillan
Teichmann, Roger (2000) 'The Complete Description of Temporal Reality,' in:
Patrick Baert (Ed.): Time in Contemporary Intellectual Thought. Amster-
dam: Elsevier, 11-15
Tooley, Michael (1999). 'The Metaphysics of Time,' in: The Arguments of
Time. Ed. by Jeremy Butterfield. Oxford: Oxford UP, 21-42
242 Bibliography

Treisman, Michel (1999). 'The Perception of Time: Philosophical Views and

Psychological Evidence,' in: The Arguments of Time. Ed. by Jeremy But-
terfield. Oxford: Oxford UP, 217-246
Wendorff, Rudolf (1986). 'Zur Erfahrung und Erforschung von Zeitphanomen-
en im 20. Jahrhundert,' in: Heinz Burger (ed.): Zeit, Natur und Mensch.
Beitrage von Wissenschaftlern zum Thema "Zeit". Berlin: Berlin-Verlag
Spitz, 17-43
Wang, Hao (1995). 'Time in Philosophy and in Physics: From Kant and Ein-
stein to Godel,' in: Synthase 102: 215-234
Wetzel, James (1995). 'Time after Augustine,' in: Religious Studies 31: 241-
Whitrow, G.J. (1970). 'Reflections on the History of the Concept of Time,' in:
Studium Generate 23: 498-508
Zwart, P.J. (1972). 'The Flow of Time,' in: Synthase 24: 133-158

Philosophy of Time

Baltzer, Ulrich (1998). 'Zeit und Indexikalitat. Warum eine Theorie der Zeit
weder "tensed" noch "tenseless' seinkann,' in: Logos NF 5: 311-333
Beer, Michelle (2010). 'Tense and Truth Conditions,' in: Philosophia 38: 265-
Braddon-Mitchell, David (2004). 'How do we know it is now now?,' in: Ana-
lysis 64: 199-203
Brogaard, Berit (2006). 'Tensed Relatios,' in: Analysis 66: 194-202
Burley, Mikel (2008). 'Should a B-Theoretic Atheist Fear Death?,' in: Ratio ns
Button, Tim (2006). 'There's no time like the present,' in: Analysis 66: 130-
Corish, Denis (2005). 'McTaggart's Argument,' in: Philosophy 80: 77-99
Corish, Denis (2006). 'Time Reconsidered,' in: Philosophy 81: 81-106
Craig, William Lane (2000). 'McTaggart's Paradox,' in: William Lane Craig,
The Tensed Theory of Time. A Critical Examination. Dordrecht: Kluwer (=
Synthese Library. 293), 169-217
Craig, William Lane (2000). The Tensed Theory of Time. A Critical Examina-
tion. Dordrecht: Kluwer (= Synthese Library. 293)
Craig, William Lane (2000). The Tenseless Theory of Time. A Critical Exam-
ination. Dordrecht: Kluwer (= Synthese Library. 294)
Craig, William Lane (2001). 'McTaggart's Paradox and Temporal Solipsism,'
in: Australasien Journal of Philosophy 79: 32-44
Diekemper, Joseph (2007). 'B-Theory, Fixity, and Fatalism,' in: Nous 41: 429-
Dummett, Michael (1960). 'A Defense of McTaggart's Proof of the Unreality
of Time,' in: The Philosophical Review 69: S. 497-504
Earman, John (1969). 'The Anisotropy of Time,' in: Australasien Journal of
Philosophy 47: 273-295
A Guide to Further Reading 243

Earman, John (1972. 'Notes on the Causal Theory of Time,' in: Synthese 24:
Earman, John (2002). 'Thoroughly Modern McTaggart Or What McTaggart
Would Have Said If He Had Read the General Thory of Relativity,' in:
Philosophers' Imprint 2.3
Fraassen, Bas C. van (1972). 'Earman on the Causal Theory of Time,' in: Syn-
these 24: 97-95
Hansson, Tobias (2007). 'The Problem(s) of Change Revisited,' in: Dialectica
Hofweber, Thomas (2009). 'The Meta-Problem of Change,' in: Nous 43: 286-
Kline, George L. (1986). '"Present", "Past", and "Future" as Categorial Terms,
and the "Fallacy of the Actual Future",' in: Review of Metaphysics 40:
Le Poidevin, Robin, and D.H. Mellor (1987). 'Time, Change, and the "Index-
ical Fallacy",' in: Mind 96: 534-538
Poidevin, Robin (1993). 'Lowe on McTaggart,' in: Mind 102: 163-170
Levison, Arnold B. (1987). 'Events and Time's Flow,' in: Mind 96: 341-353
Lofgren, Lars (2000). 'Understandings of Time in Complementaristic Lan-
guage,' in: Patrick Baert (ed.): Time in Contemporary Intellectual
Thought. Amsterdam: Elsevier, 37-51
Lombard, Lawrence Brian (1978). 'Relational Change and Relational
Changes,' in: Philosophical Studies 34: 63-79
Lowe, E.J. (1987). 'The Indexical Fallacy in McTaggart's Proof of the Unreal-
ity of Time,' in: Mind 96:62-70
Lowe, E.J. (1987). 'Reply to Le Poidevin and Mellor,' in: Mind 96: 539-542
Lowe, E.J. (1992). 'McTaggart's Paradox Revisited,' in: Mind 101: 323-326
Lowe, E.J. (1993). 'Comment on Le Poidevin,' in: Mind 102: 171-17
Lundmark, Lennart (2006). 'The Historian's Time,' in: Time & Society 2: 6 1 -
Mozersky, J.M. (2006). 'A Tenseless Acount of the Presence of Experience,'
in: Philosophical Studies 129: 441-476
Magalhaes, Ernani (2010). 'Introduction: McTaggart's Paradox at One Hun-
dred,' in: Philosophia 38: 225-227
Maudlin, Tim (2002). 'Thoroughly Muddled McTaggart Or How to Abuse
Gauge Freedom to Generate Metaphysical Monstrosities. With a Response
by John Earman,' in: Philosophers'Imprint 2.4
McTaggart, J. Ellis (1908). 'The Unreality of Time,' in: Mind 17: 457-474
Monton, Bradley (2010). 'McTaggart and Modern Physics,' in: Philosophia
Moore, A.W. (2001). 'Apperception and the Unreality of Time,' in: Time and
Memory. Issues in Philosophy and Psychology. Ed. by Christoph Hoerl
and Teresa McCormack. Oxford: Clarendon, 375-391
Needham, Paul (1972). 'The Speaker's Point of View,' in: Synthese 32: 309-
244 Bibliography

Needham, Paul (1979). 'A.N. Prior and Kit Fine, "Times, Worlds, Selves",' in:
Synthase 40: 389^108
Oaklander, L. Nathan (1976). 'Propositions, Facts, and Becoming,' in: Philo-
sophical Studies 29: M-W2
Oaklander, L. Nathan (1977). 'The "Timelessness" of Time,' in: Philosophy
andPhenomenologicalResearch 38: 228-233
Oaklander, L. Nathan (1983). 'McTaggart, Schlesinger, and the Two-Dimen-
sional Time Hypothesis,' in: The Philosophical Quarterly 33: 391-397
Oaklander, L. Nathan (1991). 'A Defence of the Tenseless Theory of Time,'
in: The Philosophical Quarterly 41: 26-38
Oaklander, L. Nathan (1996). 'McTaggart's Paradox and Smith's Tensed The-
ory of Time,' in: Synthese 107: 205-221
Oaklander, L. Nathan (2010). 'McTaggart's Paradox and Crisp's Presentism,'
in: Philosophia 38: 329-241
Ohrstrom, Peter, and Per Hasle (1993). 'A.N. Prior's Rediscovery of Tense Lo-
gic,' in: Erkenntnis 39: 23-50
Parsons, Josh (2002). 'A-Theory for B-Theorists,' in: The Philosophical
Quarterly 52:1-20
Prior, Arthur N. (1962). 'Tense-Logic and the Continuity of Time,' in: Stadia
Logica 8: 133-148
Prior, Arthur N. (1968). Papers on Time and Tense. Oxford: Clarendon
Posser, Simon (2000). 'A New Problem for the A-Theory of Time,' in: The
Philosophical Quarterly 50: 494-498
Posser, Simon (2007). 'Could We Experience the Passage of Time?,' in: Ratio
as. 20: 75-90
Roberts, Joy H. (1980). 'Statements, Sentences and States of Affaires in McTa-
ggart and in General,' in: Erkenntnis 15: 73-89
Rukgaber, Matthew (2010). 'Time and Metaphysics: Kand and McTaggart on
the Reality of Time,' in: Kant Yearbook 2: 175-194
Russell, Bertrand (1925). 'On the Experience of Time,' in: The Monist 25:
Schlesinger, George (1971). 'The Structure of McTaggart's Argument,' in: Re-
view ofMetaphysics 24: 668-677
Schlesinger, George (1976). 'The Stillness of Time and Philosophical Equan-
imity,' in: Philosophical Studies 30: 145-159
Shoemaker, Sydney (1969). 'Time without Change,' in: The Journal of Philo-
sophy 66:363-381
Smith, Quentin (1987). 'Problems with the New Tenseless Theory of Time,' in:
Philosophical Studies 52: 371-392
Smith, Quentin (1990). 'Temporal Indexicals,' in: Erkenntnis 32: 5-25
Tallant, Jonathan (2007). 'The have been, are (now), and will be lots of times
like the present in the hybrid view of time,' in: Analysis 67: 83-86
Tallant, Jonathan (2007). 'What is B-time?,' in: Analysis 61: 147-156
Tallant, Jonathan (2001). 'Time for Presence?,' in: Philosophia 38: 271-280
Thomson, Judith Jarvis (2001). 'McTaggart on Time,' in: Philosophical Per-
spectives 15: 229-252
A Guide to Further Reading 245

Tomberlin, James E. (1970). 'Prior on Time and Tense,' in: Review of Meta-
physics 24:57-81
Torre, Stephan (2009). 'Truth-conditions, truth-bearers and the new B-theory
of time,' in: Philosophical Studies 142: 325-344
Torrengo, Giuliano (2010). 'Time, Context, and Cross-Temporal Claims,' in:
Philosophia 38: 281-296
Tsompanidis, Vasilis (2010). 'Smart and Tensed Beliefs,' in: Philosophia 38:
Williams, Clifford (1988). 'A Bergsonian Approach to A- and B-Time,' in:
Philosophy 73: 379-393
Williams, Clifford (1996). 'The Metaphysics of A- and B-Time,' in: The
Philosophical Quarterly 46: 371-381
Zeilicovici, David (1986). 'Eine (Auf)l6sung des Paradoxes von McTaggart,'
in: itefio 28: 159-178
Zenardo, Alberto (2006). 'Moment/History Duality in Prior's Logics of
Branching-Time,' in: Synthase 150: 483-507

Time in the Arts (Text, Film, Music)

Alperson, Philip (1980). '"Musical Time" and Music as an "Art of Time",' in:
Journal ofAesthetics and Art Criticism 38: 407-417
Anderle, Martin (1971). 'Die Zeit im Gedicht,' in: The German Quarterly 44:
Bernstein, Ilya (2008). 'Temporal Registers in the Realist Novel,' in: Philo-
sophy and Literature 12: 173-182
Bluestone, George (1961). 'Time in Film and Fiction,' in: Journal of Aesthet-
ics and Art Criticism 19: 311-315
Campbell, John (1997). 'The Structure of Time in Autobiographical Memory,'
in: European Journal of Philosophy 5: 105-118
Currie, Gregory (1992). 'McTaggart at the Movies,' in: Philosophy 67: 343-
Currie, Gregory (1999). 'Can There Be a Literary Philosophy of Time?,' in:
The Arguments of Time. Ed. by Jeremy Butterfield. Oxford: Oxford UP,
Currie, Mar (2007). About Time. Narrative, Fiction and the Philosophy of
Time. Edinburgh: Edinburgh UP
Currie, Mark (2009). 'The Expansion of Time,' in: Narrative 17: 353-367
de Toro, Alfonso: Die Zeitstruktur im Gegenwartsroman am Beispiel von G.
Marcia Marque' Cien anos de soledad, M. Vargas Llosas La casa verde
undA. Robbe-Grillets La maison de rendez-vous. Tubingen: Nan-
Epstein, David (1983). 'Das Erlebnis der Zeit in der Musik. Struktur und Pro-
zeB,' In: Anton Peisl, ArminMohler (eds.): Die Zeit. Miinchen, Wien: 01-
denbourg (= Schriften der Carl-Friedrich-von-Siemens-Stiftung. Band 6),
246 Bibliography

Epstein, David (1988). 'Tempo Relations in Music: A Universal?,' in: Beauty

and the Brain. Biological Aspects of Aesthetics. Ed. by Ingo Rentschler,
Barbara Herzberger, David Epstein. Basel: Birkhauser, 91-116
Epstein, David (1989). 'Time and Timing in Music. Musical and Neurological
Aspects,' in: Musik - Gehirn - Spiel. Beitrage zum vierten Herbert von
Karajan-Symposium in Wien, 24. und 25. Mai 1988. Ed. by Hellmuth
Petsche. Basel: Birkhauser, 153-162
Epstein, David (1995). Shaping Time. Music, the Brain, and Performance.
New York: Schirmer
Fleming, William (1945). 'The Newer Concepts of Time and Their Relation to
the Temporal Arts,' in: Journal of Aesthetics and Art Criticism 4: 101-106
Fludernik, Monika (2003). 'Chronology, time, tense and experientiality in nar-
rative,' in: Language and Literature 12: 117-134
Frohlich, Melanie (2005). Zeitkonzeptionen: Zur Kritik der aktuellen Erzahl-
forschung aus kaltavwissenschaftlicher Perspektive. Magisterarbeit Uni-
versity Liineburg
Genette, Gerard (1971). 'Time and Narrative in .4 la recherche du temps per-
du, in: Aspects of Narrative. Selected Papers from the English Institute.
Ed. with a Foreword by J. Hillis Miller. New York and London: Columbia
UP, 93-118 (Translated by Paul de Man)
Grethlein, Jonas (2010). 'The Narrative Reconfiguration of Time beyond
Ricceur,' in: Poetics Today 31.2: S. 313-329
Gumbrecht, Hans Ulrich (2006). 'Presence Achived in Language (with Special
Reference Given to Presence of the Past,', in: History and Theory 45: 317-
Hamburger, Kate (1953). 'Das epische Praeteritum,' in: Deutsche Vierteljahrs-
schriftfurLiteraturwissenschaft und Geistesgeschichte 27: 329-357
Hamburger, Kate (1955). 'Die Zeitlosigkeit der Dichtung,' in: Deutsche Vier-
teljahrsschrift fur Literaturwissenschaft und Geistesgeschichte 29: 413-
Hamburger, Kate ([1957] 1968). Die Logik der Dichtung. Stuttgart: Klett
Hamburger, Kate (1965). 'Nocheinmal: VomErzahlen. Versuch einer Antwort
undKlarung,' in: Euphorion 59: 46-71
Hamburger, Kate (1973). The Logic of Literature. Bloomington: Indiana UP
Hamburger, Kate (1986). Logique des genres litteraires. Paris: Seuil
Hohler, Thomas P. (1987). 'Storytelling and Human Experience. [Rev.: Paul
Ricceur, Temps et recit, Vol. I-III],' in: Research in Phenomenology 17:
Hutchens, Eleanor N. (1972). 'The Novel as Chronomorph,' in: Novel 5: 215-
Ingarden, Roman (1960). Das literarische Kunstwerk 2., verbesserte und er-
weiterteAuflage. Mit einem Anhang: Von den Funktionen der Sprache im
Theaterschauspiel. Tubingen: Niemeyer
Ingarden, Roman (1964). Time and Modes of Being. Translated by Helen R.
Michejda. Springfield: Thomas
A Guide to Further Reading 247

Kavanagh, Thomas M. (1971). 'Time and Narration. Indexical and Iconic

Models,' in: MLN: Modern Language Notes 86: 823-834
Kohler, Dayton (1948). 'Time in the Modern Novel,' in: College English 10:
Kubler, George (1967). 'Style and the Representation of Historical Time,' in:
Annals of the New York Academy of Sciences 138: 849-855
Limon, Jerzy 2010). 'Theatre's Fifth Dimension. Time and Fictionality,' in:
Poetica 41: 33-54
Le Poidevin, Robin (1988). 'Time and Truth in Fiction,' in: The British Journ-
al ojAesthetics 28: 248-258
McClain, Jeoraldean (1985). 'Time in the Visual Arts: Lessing and Modern
Criticism,' in: Journal of Aesthetics and Art Criticism 44: 41-58
Mendilow, A.A. ([1952] 1972). Time and the Novel. Introduction by Professor
J Isaacs [1952]. New York: Humanities Press
Miller, J. Hillis 2003). 'Time in Literature,' in: Daedalus 132.2: 86-97
Miiller, Giinther (1947). Die Bedeutung der Zeit in der Erzahlkunst. Bonner
Antrittsvorlesung 1946. Bonn: Universitatsverlag Bonn
Miiller, Giinther (1948). 'Erzahlzeit und erzahlte Zeit,' in: Festschrift fur Paul
Kluckhohn und Hermann Schneider. Gewidmet zu ihrem 60. Geburtstag.
Hg. von ihren Tubinger Studenten. Tubingen: Mohr, 195-212
Giinther Miiller (1955). 'fiber das Zeitgeriist des Erzahlens. (Am Beispiel des
'Jiirg Jenatsch'). Erich Rothacker zum 60. Geburtstag,' in: Deutsche Vier-
teljahrsschriftfur Literaturwissenschaft und Geistesgeschichte 24: 1-31
Giinther Miiller (1955). 'Zeiterlebnis und Zeitgeriist in der Dichtung,' in: Sta-
dium Generate 8: 594-601
Niinning, Ansgar, and Roy Sommer (2002). 'Die Vertextung der Zeit: Zur nar-
ratologischen und phanomenologischen Rekonstruktion erzahlerisch insze-
nierter Zeiterfahrung und Zeitkonzeption,' in: Martin Middeke (ed): Zeit
und Roman. Zeiterfahrung im historischen Wandel und asthetischer Para-
digmenwechsel vom sechzehnten Jahrhundert bis zur Postmoderne.
Wiirzburg: Konigshausen & Neumann, 33-56
Pasler, Jann (1989). 'Narrative and Narrativity in Music,' in: The Study of
Time VI: Time and Mind. Interdisciplinary Issues. Ed. by J.T. Fraser.
Madison, 233-257
Philipowski, Katharina (2008). 'Negative Prasenz. Die gespaltene Zeit der Er-
zahlungbei Paul Ricceur,' in: Journal of Literary Theory 2.1: 71-98
Ricceur, Paul (1979). 'The Human Experience of Time and Narrative,' in: Re-
search in Phenomenology 9: 17-34
Ricceur, Paul (1980). 'Narrative Time,' in: Critical Inquiry 7: 169-190
Ricceur, Paul (1985). 'Narrated Time,' in: Philosophy Today 29: 259-272
Ricceur, Paul (1985). 'History as Narrative and Practice. Peter Kemp talks to
Paul Ricoeur in Copenhagen,' in: Philosophy Today 29: 213-222
Ricceur, Paul (1989). 'Mimesis, Reference and Refiguration in Time and Nar-
rative; in: Scripsi 5, No. 4:91-102
Ricceur, Paul (1991). 'Narrative Identity,' in: Philosophy Today 35: 73-81
Ricceur, Paul (1988/1991). Zeit und Erzahlung. 3 Bde. Miinchen: Fink
248 Bibliography

Ricceur, Paul (1984/1988). Time and Narrative. Translated by Kathleen

Blarney and David Pellauer. 3 vol. Chicago, London: Chicago UP
Ryan, Marie-Laure (2009). 'Temporal Paradoxes in Narrative,' in: Style 43:
Sesonske, Alexander (1980). 'Time and Tense in Cinema,' in: Journal of Aes-
thetics and Art Criticism 38: 419-426
Seidler, Herbert (1955). 'Dichterische Welt und epische Zeitgestaltung,' in:
Deutsche Vierteljahrsschrift fur Liter aturwissenschaft und Geistesge-
schichte 29: 390-413
Spanos, William V. (1970). 'Modern Literary Criticism and the Spatialization
of Time: An Existential Critique,' in: Journal of Aesthetics and Art Criti-
cism 29: 87-104
Todorov, Hristo (1968). 'Logique et temps narratif,' in: Social Science Inform-
ation 7.6: 41-49
Turner, Frederick (1978). 'Poiesis: Time and Artistic Discourse,' in: The Study
of Time III Proceedings of the Third Conference of the International So-
ciety for the Study of Time. Ed. by J.T. Fraser, N. Lawrence, D. Park. New
York, Heidelberg, Berlin: Springer, 614-633

Linguistic Approaches to Time

Abusch, Dorit (1997). 'Sequence of Tense and Temporal de re,' in: Linguistics
and Philosophy 20: 1-50
Aqvist, Lennart (1979). 'A Conjectured Axiomatization of Two-Dimensional
Reichenbachian Tense Logic,' in: Journal of Philosophical Logic 8: 1-45
Bach, Emmon (1986). 'The Algebra of Events,' in: Linguistics and Philosophy
9: 5-16
Ballweg, Joachim (1998). 'Zeit, Temporalitat und Sprache,' in: Ernest W.B.
Hess-Luttich, Brigitte Schlieben-Lange (eds.): Signs & Time. Zeit &
Zeichen. An International Conference on the Semiotics of Time in Tubin-
gen. Tubingen: Narr, 41-57
Beer, M. (1988). 'Temporal Indexicals and the Passage of Time,' in: Philo-
sophical Quarterly 38: 158-164
Binnick, Robert I. (2001). 'Temporality and Aspectuality,' in: Handbucher zur
Sprach- und Kommunikationswissenschaft, vol. 20.1. Berlin, New York 2,
Cooper, Robin (1986). 'Tense and Discourse Location in Situation Semantics,'
in: Linguistics and Philosophy 9: 17-36
Declerck, Renaat (1986). 'From Reichenbach (1947) to Comrie (1985) and
Beyond. Towards a Theory of Tense,' in: Lingua 70: 305-364
Dinsmore, John (1982). 'The Semantic Nature of Reichenbach's Tense Sys-
tem,' in: Glossa 15: 216-239
Dowty, David R. (1982). 'Tenses, Time Adverbs, and Compositional Semantic
Theory,' in: Linguistics and Philosophy 5: 23-55
A Guide to Further Reading 249

Dowty, David R. (1986). 'The Effects of Aspectual Class on the Temporal

Structure of Discourse: Semantics or Pragmatics?,' in: Linguistics and
Philosophy 9: 37-61
Dry, Helen (1981). 'Sentence Aspect and the Movement of Time,' in: Text 1:
Hamm, Fritz, and Michiel van Lambalgen (2003). 'Event Calculus, Nominal-
isation, and the Progressive,' in: Linguistics and Philosophy 26: 381-458
Harweg, Roland (1991). 'Erzahlte Zeit und Sachverhaltsfolgenzeit,' in: Folia
Linguistica 25: 41-73
Larson, Richard (2003). 'Time and Event Measure,' in: Philosophical Per-
spectives 17: 247-258
Lascarides, Alex, and Nicholas Asher (1993). 'Temporal Interpretation, Dis-
course Structure and Commonsense Entailment,' in: Linguistics and Philo-
sophy 16: 437-493
Lawrence, N. (1978). 'Levels of Language in Discourse about Time,' in: The
Study of Time III Proceedings of the Third Conference of the Internation-
al Society for the Study ofTimeAlpach -Austria. Ed. by J.T. Fraser, N.
Lawrence, D. Park. New York, Heidelberg, Berlin: Springer, 22-49
Leith, Miguel, and Jim Cunningham (2001). 'Aspect and Interval Tense Lo-
gic,' in: Linguistics and Philosophy 24: 331-381
McCann, Hugh J. (1979). 'Nominals, Facts, and Two Conceptions of Events,'
in: Philosophical Studies 35: 129-149
Mourelatos, Alexander P.D. (1978). 'Events, Processes, and States,' in: Lin-
guistics and Philosophy 2: 415-434
Nakhimovsky, Alexander (1988). 'Aspect, Aspectual Class, and the Temporal
Structure of Narrative,' in: Computational Linguistics 14.2: 29-43
Nerbonne, John (1986). 'Reference Time and Time in Narration,' in: Lingui-
stics and Philosophy 9: M-95
Oesterreicher, Wulf (2003). 'Zeit - Text - Sprache. Die Zeitlichkeit von Dis-
kursen und der Zeitkern von Sprachregeln,' In: Andreas Kablitz, Wulf
Oesterreicher, Rainer Warning (eds.): Zeit und Text. Philosophische, kul-
turanthropologische, literarhistorische und linguistische Beitrage.
Miinchen: Fink, 46-70
Pustejovsky, James (1991). 'The Syntax of Event Structure,' in: Cognition 41:
Reichenbach, Hans (1947). 'The Tenses of Verbs,' In: Hans Reichenbach: Ele-
ments of Symbolic Logic. New York: Free Press, § 51, 287-298
Schiffrin, Deborah (1981). 'Tense Variation in Narrative,' in: Language 57:
Smessaert, Hans, and Alice G.B. ter Meulen (2004). 'Temporal Reasoning with
Aspectual Adverbs,' in: Linguistics and Philosophy 27: 209-261
Smith, Carlota S. (1999). 'Activities: States or Events?,' in: Linguistics and
Philosophy 22: 479-508
Smith, Carlota S. (2007). 'Tense and temporal interpretation,' in: Lingua 117:
250 Bibliography

Song, Fei, and Robin Cohen (1991). 'Tense Interpretation in the Context of
Narrative,' in: Proceedings of the 9th National Conference on Artificial In-
telligence. Menlo Park, 131-136
Stempel, Wolf-Dieter (1971). 'Moglichkeiten einer Darstellung der Diachronie
in narrativen Texten,' in: Wolf-Dieter Stempel (ed.): Beitrage zur Textlin-
guistik. Miinchen: Fink, 53-78
ter Meulen, Alice G.G. (1995). Representing Time in Natural Language: The
Dynamic Interpretation of Tense and Aspect. Cambridge: MIT P
Tichy, Pavel (1980). 'The Logic of Temporal Discourse,' in: Linguistics and
Philosophy 3:343-369
Vendler, Zenon (1957). 'Verbs and Times,' in: Philosophical Review 66: 143-
Weinrich, Harald (2001). Tempos. Besprochene und erzahlte Welt. (6., neu be-
arbeiteteAuflage). Miinchen: Beck

Human Time Perception

Bedford, Felice L. (2004). 'Analysis of a Constraint on Perception, Cognition,

and Development: One Object, One Place, One Time,' in: Journal of Ex-
perimental Psychology: Human Perception and Performance 30: 907-912
Block, Richard A., and Dan Zakay (2001). 'Retrospective and Prospective
Timing: Memory, Attention, and Consciousness,' in: Time and Memory.
Issues in Philosophy and Psychology. Ed. by Christoph Hoerl and Teresa
McCormack. Oxford: Clarendon, 59-76
Brown, Gordon A., and Nick Chater (2001). 'The Chronological Organization
of Memory: Common Psychological Foundations for Remembering and
Timing,' in: Time and Memory. Issues in Philosophy and Psychology. Ed.
by Christoph Hoerl and Teresa McCormack. Oxford: Clarendon, 77-110
Brown, Scott W., and Myrily G. Boltz (2002). 'Attentional Processes in Time
Perception: Effects of Mental Workload and Event Structure,' in: Journal
of Experimental Psychology: Human Perception and Performance 28:
Carstensen, Laura L., Derek M. Isaacowitz and Susan T. Charles (1999). 'Tak-
ing Time Seriously. A Theory of Socioemotional Selevtivity,' in: Americ-
an Psychologist 54: 165-1S1
Efron, Robert (1967). 'The Duration of the Present,' in: Annals of the New
York Academy of Sciences 138: 713-729
Fetz, Uta (2006). 'Worauf bezieht sich jetzti; in: Philosophia Naturalis 43:
Friedman, William J. (2000). 'Time in Psychology,' in: Patrick Baert (ed.):
Time in Contemporary Intellectual Thought. Amsterdam: Elsevier, 295-
A Guide to Further Reading 251

Gibson, J.J. (1975). 'Events are Perceivable But Time Is Not,' in: The Study of
Time II Proceedings of the Second Conference of the International Soci-
ety for the Study of Time Lake Yamanaka - Japan. Ed. by J.T. Fraser and
N. Lawrence. Berlin, Heidelberg, New York: Springer, 295-484
Grondin, Simon (2001). 'From Physical Time to the First and Second Moments
of Psychological Time,' in: Psychological Bulletin 127.1: 22-44
Grosser, Otto-Joachim (1983). 'Zeit und Gehirn. Zeitliche Aspekte der Signal-
verarbeitung in den Sinnesorganen und im Zentralnervensystem,' in: Anton
Peisl, Armin Mohler (eds.): Die Zeit. Miinchen, Wien: Oldenbourg (=
Schriften der Carl-Friedrich-von-Siemens-Stiftung. Band 6), 79-132
Hemmes, Nancy S., Bruce L. Brown and Chris N. Kladopoulos (2004). 'Time
Perception with and without a Concurrent Nontemporal Task,' in: Percep-
tion & Psychophysics 66: 328-341
James, William (1886). 'The Perception of Time,' in: Journal of Speculative
Philosophy 20: 374-407
James, William (1902). 'The Perception of Time,' in: William James: The
Principles of Psychology, vol. 1. London: Macmillan, 605-642
Meredith, Patrick (1971). 'The Psychphysical Structure of Temporal Informa-
tion,' in: Studium Generate 24: 70-84
Michon, John A. (1970). 'Processing of Temporal Information and the Cognit-
ive Theory of Experience,' in: Studium Generate 23: 249-265
Pereira Damascene, Benito (1996). 'Time Perception as a Complex Functional
System: Neuropsychological Approach,' in: The International Journal of
Neuroscience 85: 237-262
Pockett, Susan (2003). 'How long is "now"? Phenomenology and the specious
present,' in: Phenomenology and the Cognitive Sciences 2: 55-68
Poppel, Ernst (1978). 'Time Perception,' in: Handbook of Sensory Physiology.
Volume VIII: Perception. Berlin, Heidelberg, New York: Springer, 713-
Poppel, Ernst (1983). 'Erlebte Zeit und die Zeit iiberhaupt. Ein Versuch der In-
tegration,' in: Anton Peisl, Armin Mohler (eds.): Die Zeit. Miinchen,
Wien: Oldenbourg (= Schriften der Carl-Friedrich-von-Siemens-Stiftung.
Band 6), 369-382
Poppel, Ernst (1994). 'Temporal Mechnisms in Perception,' in: International
Review of Neurobiology 37: 185- 202
Poppel, Ernst (1997). 'A Hierarchical Model of Temporal Perception,' in:
Trends in Cognitive Sciences 1: 56-61
Poppel, Ernst (2000). Grenzen des Bewufitsein. Wie kommen wir zur Zeit, und
wie entsteht Wirklichkeit? Frankfurt/M: Insel
Speer, Nicole K., and Jeffrey M. Zacks (2005). 'Temporal changes as event
boundaries: Procesing and memory consequences of narrative time shifts,'
in: Journal ofMemory and Language 53: 125-140
Turner, Frederick, and Ernst Poppel (1983). 'The Neural Lyre: Poetic Meter,
the Brain, and Time,' in: Poetry 142: 277-309
252 Bibliography

Turner, Frederick, and Ernst Poppel (1988). 'Metered Poetry, the Brain, and
Time,' in: Beauty and the Brain. Biological Aspects of Aesthetics. Ed. by
Ingo Rentschler, Barbara Herzberger, David Epstein. Basel, Boston, Ber-
lin: Birkhauser, 71-90
Waerden, J.H. (2001). 'Internal Clocks and the Representation of Time,' in:
Time and Memory. Issues in Philosophy and Psychology. Ed. by Christoph
Hoerl and Teresa McCormack. Oxford: Clarendon, 37-58

Artificial Intelligence Approaches to Time

Allen, James F. (1983). 'Maintaining Knowledge about Temporal Interval,' in:

Communications of the ACM 26.11: 832-843
Allen, James F. (1984). 'Towards a General Theory of Action and Time,' in:
Artificial Intelligence 23: 123-154
Allen, James F., and Patrick J. Hayes (1985). 'A Common-Sense Theory of
Time,' in: Proceedings of the 9* Internationalist Conference on Artifi-
cial Intelligence, Los Angeles 1985. Vol. 1. Los Altos, 528-531
Allen, James F. (1989). 'Moments and Points in an Interval-based Temporal
Logic,' in: Computational Intelligence 5: 225-238
Allen, James F. (1991). 'Time and Time Again: The Many Ways to Represent
Time,' in: InternationalJournal of Intelligent Systems 6: 341-355
Almeida, Michael J. (1995). 'Time in Narratives,' in: Deixis in Narrative. A
Cognitive Science Perspective. Ed. by Judith F. Duchan, Gail A. Bruder,
Lynne E. Hewitt. Hillsdale, Hove: Lawrence Erlbaum, 159-189
Bennett, Brandon, and Antony Gallon (2004). 'A Unifying Semantics for Time
and Event,' in: Artificial Intelligence 153.1-2: 13-48
Gallon, Antony (1990). 'A Critical Examination of Allen's Theory of Action
and Time,' m: Artificial Intelligence 42: S. 159-188
Gallon, Antony (1996). 'Time and Continuity in Philosophy, Mathematics, and
Artificial Intelligence,' in: Kodikas/Code. Ars Semeiotica 19.1-2: 101-119
Mani, Inderjeet (2010). The Imagined Moment. Time, Narrative, and Compu-
tation. Lincoln, London: U of Nebraska P
Meister, Jan Christoph (2003). Computing Action. A Narratological Approach.
Berlin, New York: de Gruyter
Meister, Jan Christoph (2005). 'Tagging Time in Prolog: The Temporality Ef-
fect Project,' in: Literary and Linguistic Computing 20, Suppl., 107-124
Moens, Marc, and Mark Steedman (1988). 'Temporal Ontology and Temporal
Reference,' in: Computational Linguistics 14.2, 15-28
Shoham, Yoav (1987). 'Temporal Logics in AI: Semantical and Ontological
Considerations,' in: Artificial Intelligence 33: 89-104
Subject Index

acceleration XII, 191 duration 1X1, XIX, 4f, 38, 43f, 105,
achrony 111, 131,222,228 109, 115, 131ff, 159, 175, 186ff,
acttime 113-117, 131ff. 190,204,220,234
action time 105,144 durative 46, 48, 77, 79, 134, 154, 162f
Aufierungszeit 146 duree XIX, 109
Aktzeit 113f. ellipsis XII, 109, 120f, 132-136,
anachrony 111, 116, 118f., 126, 128, 154f.
131 emotional time 70
analepsis X, 118ff., 124, 138,215,228 empirical time XVIII
Aussparung 73f., 105 erzahlteZeit XVI, 86, 106, 114f, 143,
Betrachtzeit 144ff. 154, 159, 180
biological time 70 Erzahlzeit XVI, 86, 105f, 114f, 143,
branching time XVI, 61-65, 222 154, 159, 180
calendrical time 44, 47, 78, 148 event time 179,223
chronological order 33, 126f., 179 expansion of time 105, 109
chronological sequence VI, 47f, 116f., experience of time IX, 32, 36, 78, 89,
121 105, 113, 173, 180,196,202,204
chronological time 114,143 experienced time 22,73
chronology XIII, 106, 111, 116f., 128, expressed time 145f.
131, 179,219,224,228,235 external time 113f, 120, 139
circularity 109, 111, 118, 125f., 138 fact-time 143-170
clock time XVII, 43, 72, 78, 143, 149 fact-sequence-time 143-170
concept of time X, 30, 44f., 57, 183, fictional time 114, 139, 143
189, 197,201 flashback XII, XVIII, 109, 118, 121
contraction of time XVIf.,77 flash-forward XVIII, 118
consciousness of duration 190 foreshadowing 109, 122f
consciousness of time 13, 19f, 34, frequence XIX, 109
186ff Handlungszeit 105
construction of time 46, 89ff, 113 frequency XI, XIX, 77, 109f, 137, 226
course of time 89ff historical time 113f
Darstellungszeit 114 ideal time 105
Dauer 106 imagined time 91,184
discourse time XIX, 143, 145ff, 154, internal time 50,113,120,139
157, 159, 179, 181, 229f, 235 interval of time 213ff, 219ff
disunifiedtime XVI, 51f, 54ff, 59-64 iteration XII, XVIII, 18
doctrine of time 88 iterative 77, 79, 134, 139, 162f
254 Subject Index

lifetime 7If., 79 recursion XII

linear time 64 reference point 2,6-9, 199f
measured time 30,45 repetition 4f,47, 109f, 137f
measurement of time 46f, 75, 77 representation of time XV, 16, 79,
narrated time XVIf, XlXf, 75-84, 86, 143, 174, 186
91, 105f, 115, 143, 146f, 154, retention 18, 188-191, 196
157, 180,216 retrospection X
narrative gap 74 Ruckgriff 121
narrative theory of time XVI, 145f. Ruckschritt 120, 122
narrative time 71, 77, 181, 234 Riickwendung 109,118,120,130
narrative time construction IX, Xllf, Sachverhaltszeit 143ff, 146
171-216 sense of time 36,171,233
notion of time XVI, XX, 172, 217f sequence of time 70
objective time 20ff, 114, 168, 181, sequentiality XVIII, 39, 187, 191,207
185ff, 196,203,218 shape of time 49
objektive Zeit 114 simultaneity 46, 49, 52, 54f, 60f, 63,
observation-time XIX, 145f. 109, 126, 130f, 139, 150, 157,
omission XVIII, 104f, 121, 127, 133, 186, 196,232
135, 157,219 sommaire 133
omitting XVIII, 77, 106 speech time 144, 223f
pause 132f, 136, 154 succession in time 101
pausing XVIII, 106 story time 143-170,235
perception of time 32, 36, 173, 182, stretching of time XII
194, 203 summary 109, 132-136
perception of temporality 191 synchrony 109, 118, 126, 130
permutation of time 109 temporal consciousness 207,213
physical time 32,70,75 temporal constraint 221,223-229
poetic time 70, 114 temporal continuum 15, 194
point in time XVII, 13, 16f, 27, 96, temporal determination 185, 188f,
139, 184,189,194,211 208,210
point of the event XIV, Iff, 9, 144 temporal distance 118f, 234
point of reference XIV, l f f , 9 f , 2 9 , temporal effect 190
144, 197 temporal extension 76, 118, 163, 185,
point of speech XIV, Iff, 5ff, 9, 144 188
position in time 53 temporal function 5, 32, 96
progression of time 87 temporal gap XVII, 120, 123, 136
prolepsis X, 109, 111, 113,116, temporal index 208f
118ff, 122-125, 127,130, 138, temporal indicator 42
215 temporal interval 221
prospection X temporality XIV, IXI, 22, 71, 171f,
Raffling 105 173, 178-180, 183f, 188, 191f,
reception-time 146,208 198,201,204,205f
reading time 113, 144 temporality effect XII, XIV, 171-216
realtime 34, 51, 87, 105, 113ff, 199, temporality of life 71
203 temporal operator 181, 209, 214ff
reality effect XIV, 171 temporal organised events 15
Subject Index 255

temporal order 16, 19,42, 112, 119, time coverage 133

127, 172f.,179f., 191,193,202, time determination 3,7f., 174ff., 184f.,
205f.,210,222,230 191, 193
temporal permutation 119f., 126 time diagram 110
temporal perspective 16f., 19, 21, 56 time expression XII, XIV, 213, 217,
temporal position 18, 42, 47, 127, 187, 223f.
193,209,211 time indicator 91, 110, 119, 138f., 196
temporal precedence 49, 63 time interval 221f., 224, 229
temporal process 70,103,213 time interweaving 118, 1254, 126ff.
temporal progression 117 time line Xif., XVIII, XX, 31, 189,
temporal relation XI, 15, 17, 21, 50, 222,228ff.,235
54, 57, 59f., 186f., 211,219, 221, time manipulation XIX
223ff.,229,235 time marker 93f.
temporal sequence 15, 101, 116, 193 time mode 16-19, 30, 32, 42, 146
temporal structure X, 15, 46, 117, 177, time of narrating 76-79, 82
190, 197f., 200, 203, 206, 210, 219 time of narration XVI, XIX, 73, 75f,
temporal succession 101,106,131, 86f, 105f, 114f, 143, 180
185 time overlap 111, 118, 124, 126, 128f
temps de la fable 143 time of representation 114
temps de la fiction 113f., 143 time of speech 105
temps de la lecture 113 time order 1, 9f, 20, 42, 45, 50, 197
temps de la narration 113,143 time point XIV, Iff, 6f, 9, 211
temps deVecriture 113 time position 7
temps de I'histoire 113,143 time process 70
temps dudiscours 165 time relation XIV, 1,11,105, 159, 205
temps dusujet 113 time span 13,66, 135
temps externe 113 time structure XIX, 13f, 109
temps interne 113 time summary 109, 132ff, 136
temps narre 114 time usage XIX, 109-116, 130f, 137f
temps raconte 113 time window 181, 185, 189, 194, 196,
temps represents 113 208ff, 212f,
temps simule 113 unity of time 56, 105
texttime 113, 115ff., 131 understanding of time X, 16, 34, 183,
Textzeit 113,115 232
time arrangement XIX, 109, 116, 138 unreality of time 31, 49ff, 182, 196,
time concept 32, 77, 89 201
time consciousness XV, 13, 15, 17-20, utterance-time 144ff
30,33, 172f., 185, 190, 195,201- Vorausdeutung 109, 118, 122
204,212f.,216 Zeitbewaltigung 105
time contraction 75-78, 81 Zeiterlebnis 105
time coordinates 14 Zeiterstreckung 105
time construction IX,XIIf., 171-216 Zeitraffung 75, 133
Name Index

Allen, James XI, 22If. Dennett, Daniel C. 23

Archambault, Paul J. 184 Deppert, Wolfgang 33ff, 44
Aristotle 30, 37, 79, 98, 101, 182 de Toro, Alfonso VII, XVHIf, 130
Augustine 34,46, 181-186, 195f.,218 Diderot, Denis 79
Bakhtin, Mikhail M. (Bachtin, Michail Doblin, Alfred 69
M.) IX Dostoyevsky, Fjodor 77
Balzac, Honore de 69 Drucker, J. Xif
Barthes, Roland 171 Ducrot, Oswald 113f, 120, 126, 130,
Baumgartner, Klaus 144 132
Behrens, Irene 69 Dummett, Michael A. 56f.
Bennett, Jonathan 20 Eco,Umberto 230
Bergson, Henri 105 Einstein, Albert 45
Bieri, Peter VII, XV, 20, 33ff. 181, Erlich, Victor 147
202-205 Faulkner, William XI, 219
Borges, Jorge Luis 223 Fielding, Henry 71f, 90, 105
Boyle, A. XI Fischer, Norbert 182f
Brann, Eva T.H. 182 Fitelson, Brian 232
Bremond, Claude 112 Flaubert, Gustave 69, 118, 122, 126,
Brentano, Clemens 76 130, 135,220
Brentano, Franz 186f. Flemming, Willi 104
Broad, CD. 195 Fontane, Theodor 92
Brunot, Ferdinand 11 Forster, Edward M. 103,105
Brunswick, Herzog Anton Ulrich of 78 Frankfurt, Harry G. 25
Buhusi, Catalin V. 233 Frege, Gottlob 30
Burg, J. XI Fuentes, Carlos 123
Butterfield, Jeremy X, XII, 64, 180 Gaizauskas, Rob XII
Cantor, Geoffrey 65 Garcia Marquez, Gabriel 122, 224
Casati, Roberto 177f Geach, Peter 59, 63
Chamberlain, Jane 1175,180,190 Genette, Gerard Xff, XVHIf, 109-
Chatman, Seymour 143, 147, 154, 159 142, 144, 147f, 154, 159,226,
Chekhov, Anton P. 222 228f
Clay,E.R. 194 Goethe, Johann Wolfgang 68ff, 78-
Cresswell, M.J. 231 81,85,91
Currie, Gregory X, XIII, 180 Goodwin, William Watson 4
Currie, Mark Ixf. Graesser, Arthur C. 234
Dammann, Giinter XXI Greimas, Algridas Julien 112
Dallenbach, Lucien 119 Grellmann,H. 35
Darwin, Charles 35 Grotz, Klaus 182
258 Name Index

Grunhage, Lisa XXI Keller, Gottfried 92ff, 97,

Halpern, Joseph Y. 231 Kelly, E. Robert 195
Hamburger, Kate VII, XVIIf., 104 Kleist, Heinrich von 77
Hankinson, Robert J. 35 Kneale, William 30
Hartmann, Dirk 40f. Krause, Rolf XXI
Harweg, Roland VII, XIX, 143, 145, Kripke, Saul 230
150, 153, 157f., 166 Lammert, Eberhard VII, XVIII, 101,
Hegel, Georg Wilhelm Friedrich 87f., 109, 112, 118, 120-123, 143f,
99, 192, 195f. 147, 157, 159
Heidegger, Martin 33, 181f., 186, 192 Lang, S. XI
Heliodorus 116 Laurien, Ingrid XXI
Herder, Johann Gottfried 101-105 Le Poivevin, Robin VII, XVI, 50, 173,
Herrmann, Friedrich-Wilhelm von 182 181, 197-202, 205, 215f
Heron of Alexandria 44 Lessing, Gotthold Ephraim 85ff, 90,
Herrnstein Smith, Barbara 230 lOlf,104f
Heusler, Andreas 105,143 Lewis, David 14,232
Hirt, Ernst 105,143 Lightman, Alan 234
Hodge, Jonathan 65 Link,Jiirgen 114
Holderlin, Friedrich 70 Lohmann,Rike XXI
Hoffmann, E.TA. 78 Loizou, Andros 52
Homer 79, 86, 90, 102, 104, 116 Lotman, Juri M. 112, 137
Huhn, Peter 178,200 Lowe, Jonathan 55
Hughes, G.E. 231 Lubbock, Percy 133
Hugo, Victor 69 Lucas, John R. 52, 56f, 62, 64f
Humphrey, Robert 133 Lukacs, Georg 68f, 71
Husserl, Edmund IX, XIV, XX, 18, Lugowski, Clemens 104
33, 175f, 179, 181, 185-192, MacBeath, Murray 50,173,201
195f,202f,208 Macaulay, Thomas B. 2
Huth, Michael 222 Macey, Samuel L. 172
Ingarden, Roman 69, 104 Mach, Ernst 32,45
Jacobsen, Jens Peter 69 Maidens, Anna 65
Jakobson, Roman 112,137 Mainzer, Klaus 33, 35f,47
James, William 32f, 194 Mani, Inderjeet VII, Xllf, XXf, 172,
Janich, Peter VII, Xvf, 31, 37f, 40f, 224,228f,235
43ff,47 Mann,Katja 156
Janik, Dieter 112 Mann, Thomas 68, 73ff, 89, 105f,
Jauss, Hans Robert 109 122, 145ff, 156, 160-165, 167
Jean Paul 103ff Martini, Angela VIII
Jespersen, Jens Otto Harry 3 Maugham, W. Somerset 1
Joyce, James 75ff, 89, 133 McHale, Brian 232
Junghans, Ferdinand 105 Mclnerney, Peter K. 173
Kanabus,M. 233 McTaggart, John Ellis IX, XIV, XVI,
Kant, Immanuel 18, 20, 32, 45, 87f, XX, 31, 33, 43, 49-57, 60, 64, 181,
173-190, 195f,203-206 192-199, 201f, 205f, 214ff, 218
Kayser, Wolfgang 97f, 114, 116 Meek, Warren H. 233
Keats, John 2 Medicus, Fritz 101
Name Index 259

MeirLItay 22If. Robbe-Grillet, Alain 136, 222

Meister, Jan Christoph VIII, X,XIIf, Rorty, Richard 13,30
XX, 172, 178, 180 Rosenberg, Jay F. 15
Mellor, David H. 53,178 Rossum-Guyon, Francoise van 113f
Mendelssohn, Peter de 156 Russell, Bertrand 3 If, 47
Mendilow, A.A. 114,143 Ryan, Mark 222
Meredith, George 76 Ryan, Marie-Laure 230, 232
Meyer, Herman 106 Sandbothe, Mike XII
Mohr,Georg 196 Sappho 70
Montfort, Nick 228 Savitt, Steven 193
Moore, Adrian W. 58 Schiller, Friedrich 85,91
Moses, Y. 231 Schmid,Wolf 171
Miiller, Giinther VII, XVIf., XIX, Schmitz, Hermann 33
103-106, 109, 113ff.,132, 134, Schneewind, Jerome B. 192
143f. Schiich, Lena XXI
Muir, Edwin 105 Segre,Cesare 125
Nabokov, Vladimir 217,219 Seidler, Herbert XVII, 85, 88, 91, 95f,
Newton-Smith, William H. 50 98
Nolting-Hauff, Use 116 Shaftesbury, Anthony Ashley Cooper
Novalis 78 of 104
Nowvieskie, B. XI Shakespeare, William 234
Oaklander, L. Nathan 53 Shelley, Mary 70
O'Donnell, James O. 181f., 185 Simons, Peter 63, 65
Pan, Feng 224 Sklar, Lawrence 172
Petrarca, Francesco 70 Smart, John J.C. 53
Petsch, Robert 69, 103, 105, 134 Smith, Quentin 52
Pfister, Manfred 113f, 131 Staiger,Emil 69
Piaget,Jean 39 Starritt, Alexander XXI
Picon, Gaetan 105 Steegmuller, Francis 220
Pier, John 200 Stein, Edith 186
Pindar 102 Swinburne, Richard 65
Poe, Edgar Allan 206 Stanzel, Franz K. 133,147,153
Poppel, Ernst XI, 233 Stenzel, Julius 104
Pouillon, Jean 105 Sternberg, Meir 180,191
Priest, Graham 58 Sterne, Lawrence 105
Prior, Arthur N. 52, 59f, 62f Stierle, Karlheinz 110,147
Propp, Wladimir 112 Stifter, Adalbert 77, 96
Proust, Marcel 89, 122, 226 Strawson, Peter F. 13
Pustejovsky, James XII, 224 Sue, Eugene 125
Quine, Williard Van Oman 15 Thieberger, Hans 105
Radcliffe, Anne 125 Todorov, Hristo 113
Ranke, Leopold von 94 Todorov, Tzvetan 110, 112ff, 120,
Reichenbach, Hans VIII, XIVf, 144
125, 130, 132f
Reichenbach, Maria VIII
Tolstoy, Lev 69
Ricardou, Jean 109, 113f, 132, 136
Tomasevskij, Boris 143f, 147
Ricoeur,Paul IXf, 171, 175, 180
Torricelli, Evangelista 44
260 Name Index

Turner, Frederick 233 Woolf, Virginia 89,133

Vargas Llosa, Mario 125 Wunderlich, Dieter 113,144
Varzi, Achille 177f. Wundt, Wilhelm 32
Vercoutter, Jean 44 Zalta, Edward 173, 178, 193,232
Walzel, Oskar 68, 101 Zielinski, Thaddaeus 105,143,158
Weinrich, Harald 113,132 Zimmerli, Walter, Ch. XII
Weizsacker, Viktor von 92 Zwaan,RolfA. 234
Wolfram von Eschenbach 70