PHILOSOPHY OF THE SOCIAL SCIENCES / June 2000 Hattiangadi / PAC-MAN METAPHYSICS
Pac-Man Metaphysics and the Modest
Hubris of the Professional Intellectual JAGDISH HATTIANGADI York University, Toronto Randall Collinss book The Sociology of Philosophies is very impres- sive, both in its scope and range, and in its sophistication and erudi- tion. 1 Inmyopinionthe last is also its greatest weakness, at least when the reader happens not to be a sociologist. Sociology is quite replete with impressive books and writers who produce such grand, erudite tomes. Not everyone outside of sociology is quite usedto this. Abook that is so impressive is therefore not likely to be readcarefully by non- specialists. It may even be dismissed by a casual reviewer and reader on flimsy grounds, merely because it is intimidating, especially if the reader or reviewer does not like to be tested and found wanting. But such reactions are unfortunate. This is an important book for theories of the development of ideas. Since Thomas Kuhns The Structure of Sci- entific Revolutions became stale news, there has not been any account of ideas with such a synoptic view that provides a new direction to our thought. Nowthis bookmakes a beginninginthis directiononthe subject of theories and theory change. The main feature of the book is that it is sociology. It asks to be read as a sociological account of philosophies. In this respect, it is not like Kuhns contribution, which seemedto showKuhn as an independent observer (a physicist turning to history with philosophical concerns) rather than as a modest, working professional. Collinss book includes some sophisticated philosophical discussion, too, that philosophers cannot dismiss as jejune, as many other sociological excursions into philosophical issues are. But then, in many of those cases, the sociolo- gists under scrutiny seemto be more interestedin attracting attention by holding forth overstated and provocative philosophical opinions than, as in Collinss case, trying to state a clear and convincing descriptive account. But Collinss contribution, however sophisti- 284 This article is based on a paper read at the symposiumon Randall Collinss ASociology of Philosophies at York University, Toronto, 17 September 1999. Philosophy of the Social Sciences, Vol. 30 No. 2, June 2000 284-298 2000 Sage Publications, Inc. cated be it in nuances of philosophical terminology and concept, remains primarily a contribution to the sociology of knowledge. Collinss writingis verydifferent fromthat of Thomas Kuhn. Kuhn was an migr from quantum physics who, under the influence of James Bryant Conant (thenpresident of HarvardUniversity andhim- self an aspiring philosopher and social theorist of science), adopted first history of science and then philosophy as his field for scholarly study, although he was never fully accepted among either group of scholars, notwithstanding the many honors showered upon him by both. Randall Collins is, by contrast, a card-carrying and distin- guished professional sociologist, addressing some of the central con- cerns of sociology in this book. In the process of doing this, he also provides important new insights into the subject of theories and the- ory change whichcouldeasily be neglectedif the sociological account is not understood in perspective. I will suggest, however, that this book does not provide the right perspective for its own contributions, synoptic though it may be. Myrecommendationtoa philosopher whopicks upthis large book to read all 1,100 pages of it, less 2, is to begin at the very end, which is the Epilogue. I would then suggest going back to the chapter preced- ingit, whichis a verycomplete statement of his interpretationof what he thinks philosophies do. These two chapters are titled Meta- Reflections. They are useful to orient the reader concerning Collinss theoretical assumptions. Only after that would I read the preface, the introduction, and the first few chapters setting up the theoretical apparatus of the book. The subsequent middle chapters on the his- tory of philosophies, which constitute about 80%of the book by page count, I wouldleave tothe end. These are well worthremarkingupon, but I will not have the time to do so here. In this brief critique, I will leave out of account many of the best features of the book to concen- trate on his theoretical stance. In the Epilogue, Collins tells us that he subscribes to the position that reality is socially constructed. His thesis therein, which is summa- rized helpfully at the beginning of the Epilogue, is as follows: It is oftensupposedthat social constructivismundermines truth. If real- ity is socially constructed, there is no objectivity or reality. I deny the conclusion. Social constructivismis sociological realism; andsociologi- cal realism carries with it a wide range of realist consequences. 2 Hattiangadi / PAC-MAN METAPHYSICS 285 In just making this claim, he shows more sophistication and under- standing of philosophy thanmost sociological theorists do whenthey are writing on the social construction of reality. But the claim made, although sophisticated, is still very puzzling. Underlying it is a strange question that is left unasked, namely, What must we admit being real as sociologists? And the peculiarity of this is in those last phrases as sociologistsor as social constructivists.Let meexplain. The doctrine that reality is a social construction apparently entails that mathematical entities, for instance, considered as constructs, are not quite real. Theoretical entities postulated in physics are not quite real, either, or not all of them. In this respect, he says, numbers are like other symbols which make up human discourse. Their universality comes fromtheir universal use, not fromany charac- ter of the objects onwhichtheyare used. . . . Numberingis the operation of making items equivalent by counting them off; they become items because they are treated as such. This does not mean that numbers are illusions. They are real, as operations carried out by human beings, activities carried out in time and spatial location. 3 As sociologists, we ask, What dowe needtoadmit inorder tomake sociological sense of the activity of the mathematician? Well, we must admit that there are mathematicians and that there are those typical things that they do, peculiar though they may be. As long as we can describe everything lying within mathematics in terms of something done characteristically by these mathematicians, whose existence we have admitted as sociologists, and whose activities, and especially whose social activities, are precisely what sociologists study, we are doing well. Mathematical entities (numbers, spaces, etc.) are only real insofar as they are reducedto the social, not as entities independent of the social and all that it may entail. Physics does not fare much better. Taking a page from the book of philosophy of science, he questions the reality of theoretical entities, on grounds that I need not repeat to philosophers. His only doubt is also a classical one to be found in Duhemand in more recent writers, too. In discussing the mathematical entities postulated by modern physics, which Collins calls rapid-discovery science (which leaves it open whether there is real discovery, and it was rapid for the first time, or just the rapidity of what is socially constructed as discovery), he writes, 286 PHILOSOPHY OF THE SOCIAL SCIENCES / June 2000 The social construction of classical entities leads at least to quasi real- ism. Although not all scientific entities, as intellectual constructs, have the same claimto be regarded as real, some of themare so closely inter- twinedwith human sizedbanal reality that the borderline is difficult to draw. Although their epistemological justification is more complex than the irrefutable realities of immediate social experience, they are at least close kin. 4 Once again, it is clear that entities a physicist invokes are real only if they are akin to the sociological ones, whose reality a sociologist finds quite banal. If, however, the putative entities happen to be needed by a physicist in the pursuit of physics, but not by the sociologist, then they are considered quasi-real. This is a very curious criterion for what there is: it is a requirement that existence be definedwithinthe operationof a sociological method. We may well ask, Why can a mathematician not work out her own list of what there is? And why does the sociologist think that he gets to decide what there is for the rest of us? Consider the case of an evolutionary psychologist, or better still, of a sociobiologist. This person studies some very well established results in biology (genetics, in fact) and asks, As geneticists, what do we need admit being real? No doubt, he or she can continue, there are all sorts of experiences that are described as social. I can provide an account of them, in principle, without making any reference to social reality, except as a construct fromwithin genetic theory. They must all be genetically determined. These are indeed quasi-real. The sociologist may well reply that the sociobiologists account, however satisfying it may be to the geneticist, is quite unsatisfactory for the pursuit of sociology. Sociologyhas its owngoals, concerns, and criteria, and therefore must describe reality independently of the account of it by a genetic reductionist or a psychological reductionist. But the mathematician can respond to the sociologist likewise. The account of mathematics given by Randall Collins is not mathematics, by any means. A mathematician, doing mathematics, cannot possibly think of numbers as operations of the kind Collins invokes. Why shouldsociologyget anyprecedence over mathematics if genetics has to relinquish its precedence over sociology? Why are numbers con- structs that do not attach to reality, but the law of small numbers, which I will explain below, attaches to a social reality? This seems extremely odd. Hattiangadi / PAC-MAN METAPHYSICS 287 This brings me to the title of this article, why I call this pac-man metaphysics. The metaphysical claimthat I have described to you is not one peculiar to sociology of knowledge but one endemic to modernisms, or modern metaphysical schemes, as such. Let me describe this by a story, although one somewhat shorter and much less learned than the story that Randall Collins tells. In the 17th century, soon after the birth of rapid-discovery sci- ence, Descartes made a remarkable discovery. He discovered an algebraic geometry inwhichhe was able to describe all spatial figures whatsoever in terms of a system of equations, which were closed under the four operations of arithmetic. If we add time as another coordinate to the three spatial ones, then this algebraic method, he thought, can also describe all motion. He came to regardspace, which his algebra could describe so completely, as the physical Universe itself. (Later Newton treated this coordinate system of the world as empty space that houses everything that is material within it.) This conception of space provided us with the first criterion for existence of which we are aware in philosophy anywhere. Nothing exists as a physical object, except as it manifests itself in the spatiotemporal coordinate systemof the world. Descartes made exceptions of minds andGod, as immaterial things falling outside of space, but soon those exceptions came to be abandoned by other philosophers. The reductionof all things to mechanics was a great andcelebrated issue. It continues to exercise philosophers, now and again. His- torically, it is the first great reductionism, but by no means the last. Idealism, when Berkeley introduced it, came with a powerful critical argument, but its main attraction was that it rivaled Descartes reductionist scheme in its reductive power. If, as Locke says, all our knowledge of the physical universe comes solely from sense percep- tions (and reflections upon them, combinations of them, and abstrac- tions fromthem), then howcan we conclude on the basis of our expe- rience that we live in a mathematical world quite unlike our perceptual experience? Berkeley proposed that we accept the empiri- cal principle at face value: all we can know are percepts, as ideas in some mind. Berkeleys idealism argued for banal or commonsense realities and questioned the theoretical constructs of what seemed to him to be a potentially Godless Newtonian science. Berkeleys idealism has a very special quality, though, which makes it animportant philosophical contribution. His idealismis spe- cial because it rivals Cartesian reductionismin one seemingly impor- tant feature: it provides an interpretative scheme for all existence. It 288 PHILOSOPHY OF THE SOCIAL SCIENCES / June 2000 gives us a newcriterion, esse est percipii. Everything that exists does so as anidea insome mind. This is the onlywaythat somethingcanexist, unless it is a mind (spirit) itself. This does not mean that trees do not exist, because, of course, they exist as perceived. Gravity, insofar as it helps us predict natural events, has that kind of quasi reality, but no more. Science canbe reinterpretedas givingus laws of nature govern- ing perceptual patterns. In fact, as Wittgenstein says much later of another interpretative scheme that he proposed, it leaves everything as it is. Berkeley provided us with a criterion of existence every bit as general as the materialist one based on space and time. For many years, the clash between these two types of grand inter- pretative schemes (of each of which there were many forms) was the stuff of heroic modern philosophical dispute. But in the 19th century, professionalism began to take hold of intellectuals employed at uni- versities. 5 Manynewmodernsubjects came upthat claimedthe status of a science. It seemed to the practitioners, however, that the status of their science reaches a pinnacle only when it could rival Descartes all-consuming space or Berkeleys all-encompassing ideas, in provid- ing an interpretation of all there is. How, we may ask, is this accom- plished by a new science? Aprofessional is, among other things, an expert, whose profession is defined, inpart, by the tools andskills that the practitioners possess as a member of the profession. The interpretative scheme of a profes- sion is nothing more than reality to the extent that the tools of the trade require. This is the modesty of the professional intellectual. I am only doing my job, he says, and these are my tools. My skills require that I admit tothese entities, or else I cannot domyhumble job. But the job entails a metaphysical or interpretative scheme. In this scheme, I get to decide, as a modest technician, on what exists and what does not. This is the hubris that follows from the careful cultivation of the modesty of the professional intellectual, which is being suggested in the title of this article. Starting modestly, the professional and thor- oughlymodernintellectual goes ontodecide for all of humanitywhat exists and what does not. It was soon found, after the rise of professionalism in academia, that interpretative schemes are many anddifferent. To cite contempo- rary instances, some logicians claim that to be is to be the value of a bound variable. This is Quines dictum, who, apparently, does not need to ask if variables exist. Why should he? He wrote logic texts, and everyone knows that logic needs variables. Putnam, an analytic philosopher of sorts, has espoused a version of realism called inter- Hattiangadi / PAC-MAN METAPHYSICS 289 nal realism, which allows that things exist to the extent that they are admitted to exist internally to a scheme or language. But he has no doubts of the existence of language. Postmodernists, similarly, find that texts, liberally interpreted, form a complete interpretative scheme, andif we take care to deconstruct the author interms of texts, and if we are sufficiently liberal in deciding what counts as a text and its interpretation, we can interpret anything and everything as text. Interpretative schemes aboundinmodern philosophy andinthe new sciences that seek philosophical prestige. Pac-Man is a well-known computer video game (fromvideo game infancy) in which figures swimming on a screen go around swallow- ing or getting swallowed by other figures. The object of the game is to have the figure under ones control swallow other figures but not be swallowed in turn by them. Like pac-man swallowing figures, inter- pretative schemes in modern metaphysics swallow the reality of other modernist schemes and try to resist being swallowed them- selves. Swallowing an alien reality is accomplished thus: one merely shows how the intruder is a construct built out of our own favorite theoretical building blocks. Elaborate arguments are pre- sentedat one endof a thoroughlymodernsubject showingwhyit can- not and should not be reduced to another modern scheme. At the same time, at the other end of the subject, it is vigorously trying to prove that all reality is reducible to its favorite modernist scheme. Social constructivism is no different. As pac-man modern metaphys- ics, it claims that there is an irreducible social reality andthat all of the mathematical reality, for instance, is real onlyinsofar as it is social. But its own reality is, of course, not to be doubted. There is another reason why I call these schemes pac-man schemes of modern metaphysicsit is because, like Pac-Man, they are merely games (see Figure 1). In the 18th century, this was not realized, because it seemed then that the debate between materialismand ide- alism is deep and central to our understanding of the most basic fea- tures of the universe. Now that modernisms are many and diverse, one cannot help but notice this frivolous feature that they possess. You are always aware of the fact that if you lose your game, you can turn the computer game on again, starting afresh. In this way, the game is both civilized and nonviolent. Everyone understands that if postmodernists deconstruct a text and thereby make its author unreal, he or she neednot feel his or her pulse to check that he or she is alive. If the text has been decentered, he or she need not look in a mir- ror to count how many of him or her there are. This is not what the 290 PHILOSOPHY OF THE SOCIAL SCIENCES / June 2000 analysis does. Like all modernism, it claims toleave everythingas it is. It is a great irony that the point of viewthat bills itself as having tran- scended the very era of modernismis just a pac-man modernismlike any other. When a social constructivist relegates a number to quasi reality, or even to the class of outright, less-than-real, constructs, it is under- stood that this is done in such a way as to make no difference to the waythings appear or howyouact or transact withnumbers. That is its main stipulation. Like pac-man, the social construction of reality is only a game, for that reason. But as I have pointed out, the pac-man modern metaphysical game passes off as the most sophisticated game in town, at least within philosophy. It is therefore important that philosophers take note of Randall Collinss defense of social constructivism, because at last it has an exponent skillful enough to make it a player in the game. Moving backwardto chapter 15, the last chapter of the book before the Epilogue, we find out the answer to what puzzled me immensely while I was reading the book itself, which is this: what does Collins think philosophy is? He seems to know it well enough to identify it across civilizations and vast stretches of time. Yet he does not tell us, except by identifying examples. The questionis not as easy to settle as Hattiangadi / PAC-MAN METAPHYSICS 291 Figure 1: A Logo for Modern Metaphysics may seem at first. If we take another subject of study, let us say soci- ology, or geology, or comparative literature, we would be hard- pressedto give an account of each in all the different civilizations. We would have the same difficulty with understanding history as mani- fest in every civilized tradition of thought. Why this difference? Why are mathematics and philosophy so easily spotted transculturally? Collins tells us, inthe 15thchapter, that philosophyis the activityof abstraction and reflexivityabstraction being the consideration of concrete things more and more generally, with less and less regard to situation and detail, and reflexivity is the activity of examining its own presuppositions, digging its own foundations, as he puts it. 6 Collins is quite certain that in philosophy the main flowof ideas is in an abstraction-reflexivity sequence. It seems that this is how he identifies philosophy across civilizations, since no evidence is given to adduce it. I do not knowof a clear criterion for levels of abstraction and why philosophy is cast as the most abstract of all studies. Collins seems to know perfectly clearly which is more abstract and which more concrete, when he makes judgments about them in particular instances. It wouldbe a help to the reader to be toldhowto do it with- out relying entirely on Collinss word. But we can now tackle Collinss book knowing that he is a social constructivist who allows only for those things that a professional sociologist must allow to practice his or her craft, and he thinks that philosophy is an attempt at the greatest possible abstraction and reflexivity. Now we can start on his intriguing ideas on the sociology of philosophies. Collinss account of the development of ideas, starting from the Introduction andgoing into the thirdchapter, which introduces some ideas by applying themto ancient Greek thought, is a much finer the- ory about theories and theory change than anything to be found in either the pac-man sociological method or the tired old conception of philosophy as digging deeper into abstractions and reflexivity. Indeed, I will suggest that he has a fine sociology of knowledge, whose main limitation is that he is determined to limit his account to pac-mansociologism. This is perhaps its only limitation, too. Another apparent limitationmayseemtobe that youcannot readthe historical account of the philosophies unless youhave a previous knowledge of their main doctrines and arguments. But this is not a real limitation, since the book does not pretend to be an introductory history of phi- losophy. It is, rather, a sociology of philosophies. 292 PHILOSOPHY OF THE SOCIAL SCIENCES / June 2000 Collins begins his Introduction well, as follows: Intellectual life is first of all conflict anddisagreement. 7 We maycall this his first princi- ple of the theory of development. It is one that I, too, have endorsed, followingPlatos Socrates, andmanyothers since. Collins develops it, however, in a very interesting way. The most important theoretical insight into the way in which ideas are produced is what he calls interaction rituals, as these were described stirringly by Erving Goffman in 1967. Interaction rituals happeninlocal situations inwhichindividuals bindwithother indi- viduals into a moral community, and which create symbols which act as lenses through which members view the world. 8 The conditions for their occurrence are very general: there must be more than one person present, they must all focus their attention on the same object or action, be aware of this commonness of focus, andshare a common mood or emotion. The result is a certain commonness of rhythm in their synchronized actions, a kind of impromptu ritual. The ritual aspect of it is very much as Durkheim described it. Participants in a ritual interaction feel that they are members of a group, with moral obligations to one another, and their relationship becomes symbol- ized by whatever they focused on during their ritual interaction. Peo- ple who take part in interaction rituals are filled with emotional energy, in proportion to the intensity of the interaction. It charges them up like an electric battery, Collins says, giving them enthusi- asmin the absence of the group. Collins develops that into a theory of life trajectories, as interaction ritual chains. This is a very intriguing account of howwe move fromone ritual to another and howwe keep track of our symbols, feelings, and group memberships of networks throughout our lifetimes. We are, we may say, our trajectories. Collins argues that everyconversationandeveryuse of language is a ritual use, which can be understood as an interaction ritual. If we interpret thought as internalized speech, then it, too, is an interaction ritual. There is therefore a sociology of thinking, in which all thought is fully determined by its sociology. Rituals are, of course, repetitious. So, if we could understand all our actions as rituals, then it seems that they are quite predictable. All our thought would be predictable, too. Even creative thoughts, says Collins, are fully predictable: There is a social causation of creativity, even at its intimate corethe contents of newideas that flashinthe minds of intellectuals intheir cre- ative moments. The flux of interaction ritual chains determines not merelywhowill be creative andwhen, but what their creations will be. 9 Hattiangadi / PAC-MAN METAPHYSICS 293 Here is the pac-man again. The description encountered every day interms of interactionrituals andof our lives as trajectories of interac- tion ritual chains is a very pretty account of a great many features of our existence. But while rituals are repetitious, they need not be pre- dictable. I knew a philosopher once whose greetings were of one of two kinds: he wouldeither greet one witha smile anda goodday or scowl and look away. In each case his performance was quite ritual, but I could not tell on many mornings which ritual he would follow. I wouldbe verysurprisedif the keytothe predictionof his choice of rit- ual lay insociology, however muchindepth. There is also the fact that the person could also study sociology and decide not to do what he predicts, of course, a well-knownfeature of the reflexivityof sociolog- ical knowledge. We expect, following Collins, that if only we knew enough about the interaction ritual chains of individuals, we could predict even what they think. This expectation makes an illicit assumption, for which there is no evidence whatsoever. In fact, there is goodreason to the contrary. It illicitly concludes, from the statement that interaction ritual chains can help shed light upon so much of our intellectual life, which is true, to the extravagant claim that nothing is relevant to our intellectual life except interaction ritual chains. (That is what would allow us to predict thoughts on the basis of our knowledge of them.) This is an extraordinary assumption, and there is no reason to sup- pose it. Let us begin by asking what sort of thing is excluded by Col- linss scheme. Onhis scheme, the determinationof mythoughts has nothingtodo with my noticing a logical consequence of an idea that I have previ- ously been considering. Collins gives no argument here. But we see how an argument might go. Collins could say, yes, but I can give an interpretation of logical consequence in terms of sacred objects and symbolic relations, and interaction rituals, so the case that you have given is fully covered in the account of Interaction Rituals more fully understood. But the trouble is that this is pac-man sociologism. It assumes that because there is an interpretation of the notion of logi- cal consequence in a certain style of sociological theory, and this interpretation is satisfactory for sociological purposes, there must be no such thing as a logical consequence, except as so constructed. But, of course, this does not follow. Tosaythis is all the sociological reality there is is quite different fromsayingthis is all there is, whensocio- logical reality is whatever a social constructivist must concede exists. 294 PHILOSOPHY OF THE SOCIAL SCIENCES / June 2000 Things that exist exist, withor without the permissionof the sociolog- ical profession. 10 I have also said that there is considerable evidence to show that Collinss hypothesis is false. The evidence is not merely the fact that sociology cannot produce even one example of predicting a creative result in logic or mathematics (but we couldtake any other subject for that matter). It is that the very idea does not make sense. Let us imag- ine an outstanding problem; let us say, in biology, concerning, for example, gene sequencing. Imagine that it is to be discoveredone day that all the developmental genes of the vertebrate form evolved as variations on a single gene. Presumably the biologist does 3 years of concentrated research to find this out, but, interestingly enough, the sociologist studying interaction rituals and their chains cannot only tell who is going to discover what but also that that thesis is the best account of that evidence. If the sociologists thesis, whichis the identi- cal biological thesis about the familyof genes, is true, it will not satisfy the biologist, because the sociological prediction is based on interac- tions amongsociologists, which, for the biologist, is the wrongkindof evidence. In Collinss world, this evidence would not evoke the right feelings of solidarity and symbolic reverence among the biologists with whomthe thesis has to be accepted. In a more banal or common- sense way, we cansee that a biological thesis basedonbiological judg- ment cannot possibly be the same judgment as one that looks merely at interaction rituals among the researchers, before they have even done the research. The same argument would be true for a logician or a mathemati- cian. Nologicianor mathematicianwouldaccept as a proof of the con- tinuum hypothesis any argument based solely on the analysis of interaction rituals. Thus, Collinss idea of a universal sociology is impossible on its own grounds, since the mathematician and the logi- cian have local rituals that are different from those of the sociologist. Pac-man sociologism is not compatible with sociology, because the former is just not local enough. And if they were indistinguishable, then Collins could not distinguish between mathematics and soci- ology, andbetweenbiologyandsociology, whichwouldmake his the- ory even less satisfactory. If we rid Collinss account of the pac-man, it will remain an incom- plete account of Everything, but a very beautiful one of intellectual life and of theory change. I would recommend it to anyone interested inunderstanding a neglectedaspect of what it is like to be a contribut- Hattiangadi / PAC-MAN METAPHYSICS 295 ing intellectual. It is in this spirit that I would urge the philosopher to read this remarkable book. In interaction between intellectuals two things are possible: One could disagree with another, or one could agree with another. (We assume that they are focused on the same object, since they are involved in one interaction ritual. If not, there are other possibilities.) Collins suggests that disagreement drives intellectual creativity. Col- lins is a conflict sociologist. 11 When we disagree with another, he thinks, we doit toseekattentionfor ones ideas as newandimportant. But whenwe disagree anddrawa crowdtoour dispute, the result is to take away from the attention space of another disagreement or from the attention space of other possible disagreements. If there are too many disagreements, then attention will focus on a few of them, numbering from three to six. Attention space, he suggests, is very inflexible in this regard. He calls this limitation of the number of disagreeing parties that can share the limelight at any time by the appellation the law of small numbers. Curiouslyenough, he treats attentionspace as if it is real, andnot as a construct, and the law of small numbers as if it describes a genuine law, and not just a summary description of an odd fact about the his- tory of philosophies. He has to choose here between the law of small numbers and pac-man, who has eaten the numbers up. This point would need to be argued more completely, but I believe it is a serious one. Personally, I prefer his law of small numbers. Collins comes armed with three tools, or doctrines, to the study of philosophies: (1) intellectual life is about disagreements, (2) the lawof small numbers, and (3) the doctrine of interaction rituals. In terms of these doctrines, Collins seeks to interpret the entire history of all phi- losophies in world history. If this seems extravagant, then we must remember that to be a social scientist, Collins cannot assume the truth of those views that are locally upheld, merely because he has not reflected on them. If he takes reflexivity and abstraction seriously, as a sociologist, then his sociological insights must be free from ethnocentrism. Thus, the immense scope of his writingbecomes a nat- ural consequence of his sociological method. Of course, he cannot escape the fact that sociology is a science and so adopts some current vices, willy-nilly. He could inquire into what these vices may be. But Collins does not go there. This is a level of reflexivity that Collins does not scale. Infact, inhis historyof the late 19thandearly20thcenturies, he neglects to recount the story of the emergence of sociological phi- losophies, andthensociology itself as a science, or at any rate as a pro- 296 PHILOSOPHY OF THE SOCIAL SCIENCES / June 2000 fessional discipline. This is particularly surprising since he has pub- lished a very popular book on the subject before this, and it would have been a good way to integrate his thought. 12 What is undoubtedly true, however much we may doubt the truth of his pac-man sociologistic account of it, is that when intellectual effort is recordedonthe grandscale that he does, conflicts standout in bold relief, and there is no doubt that at any time a lawof small num- bers applies. Neither of these facts can be understood on currently fashionable theories of theory and theory change. Philosophers need to take note. But the adherence to a strictly sociological method precludes the consideration of other ways in which the lawof small numbers could be explained, for instance, by considering the logic of problems as dif- ficulties, which I have worked out previously, although it is inappro- priate to go into it today. Moreover, because Collins interprets conflict as conflict occurring withinlocal interactionrituals, he misses the significance of the differ- ence between the oral and written traditions. Galileos dialogues concerning the two chief word systems can be read as an extended dialogue with Aristotle, which would have been impossible in a cul- ture that was strictly oral. Whether or not Plato or Socrates met Parmenides, the latter had a dialogue with Socrates in written form. Writing changed the dynamics of philosophy in a remarkable way, and the invention of the printing press, and later of the printed jour- nal, changed the way in which philosophy and science is done. These differences are of a kind not envisaged in Collinss theory, which is all face-to-face, andsohis account remains incomplete evenas sociology. But we should not demand cosmic completeness of account as a desideratum for good science. Pac-man modernism is a harmless modern vice, if we do not take it too seriously. Collinss monumental book is far better than his pac-man sociologism allows it to be. It makes a new and interesting start toward a better understanding of the development of ideas. Moreover, there are many details of his studies of various historical philosophies that are most interesting, but I will havetoawait another opportunityfor commentinguponthem. NOTES 1. Randall Collins, The Sociology of Philosophies: AGlobal Theory of Intellectual Change (Cambridge, MA: Harvard University Press, 1998). Hattiangadi / PAC-MAN METAPHYSICS 297 2. Ibid., 858. 3. Ibid., 867. 4. Ibid., 872. 5. One of the best things in the book is the account Collins provides of the restruc- turingof the universityinGermanyinthe earlypart of the 19thcenturyunder the influ- ence of Fichte (see chap. 12, Ibid., titled Intellectuals Take Control of Their Base; The German University Revolution). The reorganization of the university is the forerun- ner of intellectual professionalism, whichtook holdof almost all the subjects thereafter. Collins notes the importance of philosophy for the social structures that are formedbut does not pursue the importance of that particular philosophy to the professionalism today. Of course, this includes the professionalism of sociology. It seems that our very professionalism gets in the way of seeing its intellectual underpinnings. 6. Ibid., 787-91, contains the general thesis, but it cannot be fully understood with- out the accompanying descriptions, analyses, and examples that follow. 7. Ibid., 1. 8. Ibid., chap. 1, Coalitions of the Mind, 21-22. 9. Ibid., 53. In a certain sense, the sociology of thought itself is also the most impor- tant element of his theoretical system, but I found it very hard to understand why he makes this claim, which is extravagant as I try to show in the text following, until the Meta-Reflections at the end of the book are read and kept in mind. 10. Philosophers who concur heartily with this sentiment about the social construc- tion of reality must recognize that the same is true for all of modern philosophy. 11. Randall Collins, Conflict Sociology: Toward anExplanatory Science (NewYork: Aca- demic Press, 1975). 12. Four Sociological Traditions (NewYork: OxfordUniversityPress, 1994) is a revised edition of the very widely read Three Sociological Traditions (1985). No matter how widely read, one supposes that, in the interest of logical consistency, Collins will be asked to produce no more than two more revised editions. 298 PHILOSOPHY OF THE SOCIAL SCIENCES / June 2000