Kepler's Celestial Music Author(s): D. P. Walker Source: Journal of the Warburg and Courtauld Institutes, Vol. 30 (1967), pp.

228-250 Published by: The Warburg Institute Stable URL: http://www.jstor.org/stable/750744 Accessed: 26/05/2010 17:39
Your use of the JSTOR archive indicates your acceptance of JSTOR's Terms and Conditions of Use, available at http://www.jstor.org/page/info/about/policies/terms.jsp. JSTOR's Terms and Conditions of Use provides, in part, that unless you have obtained prior permission, you may not download an entire issue of a journal or multiple copies of articles, and you may use content in the JSTOR archive only for your personal, non-commercial use. Please contact the publisher regarding any further use of this work. Publisher contact information may be obtained at http://www.jstor.org/action/showPublisher?publisherCode=warburg. Each copy of any part of a JSTOR transmission must contain the same copyright notice that appears on the screen or printed page of such transmission. JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about JSTOR, please contact support@jstor.org.

The Warburg Institute is collaborating with JSTOR to digitize, preserve and extend access to Journal of the Warburg and Courtauld Institutes.

http://www.jstor.org

KEPLER'S CELESTIAL MUSIC By D. P. Walker

tn the long traditionofthe musicofthe sphereslKepler'scelestialharmonies whereas 1 are unique in severalrespects. First, they are real but soundless, becoming is eithermetaphorical, the Greekand medievalmusicof the spheres eventuallya literarytopos, or is audible music, which, for variousreasons, can only very exceptionalpeople, like Pythagoras,2 hear. Secondly,they are in polyphonic(that is, are harmonies the modernsenseof the word), whereas earlierones, from Plato to Zarlino,3consistonly of scales. Thirdly, they are in just intonation, that is, having consonantthirds and sixths, whereas all intonation,in which the smallestconsonance earliersystemsuse Pythagorean determined,by are is the fourth.4 Fourthly,these consonances geometrically whereasearliertheoristsderive the regularpolygonsinscribablein a circle, musical intervals arithmetically from simple numerical ratios. Finally,
in 1 There is quite a full bibliography the cf. pp. I 23-6), gives a good account of article 'Harmonie'in Die Musikin Geschichteancientopinionson the subject. 4 Just intonation is based on a scale so Musikin G. e G.), (hereafter undGegenwart v, Kassel I949-, cols. I594-99; cf. alsoJames divided that it contains the maximum which just consonances, Hutton, 'Some English Poems in Praise of numberof perfectly ed. Music', in EnglishMiscellany, M. Praz, are: octave 1, fifth 32, fourth3, majorthird 4, minor third 6, major sixth 3, minor sixth -5 ii, Rome I 95 I, pp. I-63. to andDemonic (theseratiosare of frequencies; obtainthe 2 See D. P. Walker, Spiritual p. 37, and John Hol- ratios of length of string,which Kepler and I958, Magic,London used, one has merely to of lander, rhe Untuning the Sky, Princeton his contemporaries invert the fractions). From Zarlino'stime p. 29. I96I, pt. Harmoniche, i, onwardsthe just scale is usually: 3 Zarlino,in his Istitutioni i, l'Opere, Venice I589, pp. I6-2I, c. vi (Tutte
Intervals from lowest note
maJor

tone
8

third
5

fowrth
4

fifth
3

major sixth
5 3

major seventh
15 8

octave
2 1

Intervals between notes

4
8

4
-9

R
6 11

*
98

^
190

r
9

*
16

<

majortone, minortone, semitone intonationis basedon a scale Pythagorean here arejust, except for All the consonances the fifth D-A, and the minor third D-F; to so divided that all the fifthsare just and all havingtwo the tonesequal: makethesejust wouldnecessitate D's. 2 243 27
I
9

Intervals from lowest note

81 64

4 3

3 2

16

128

vP

-|F9

_
9

S
8 256 243

*
9

@
8
9
. o _

f
8
9

#
8 25DG 243

Intervals between notes

except Here all the fifthsand fourthsarejust, but all tone is ff2. Here all the consonances, for the octave, are false. Cf.J. M. Barbour the thirdsand sixths are dissonant. East and is Equal temperament based on a scale so Tuning Temyserament, Lansing I953, in also and Musik G. e G., article'Intervall'. and dividedthatallitssemitones, therefore all its tones,are equal. The ratioof the semi228

KEPLER'S CELESTIALMUSIC

229

is mundana centred on, and perceivedfrom, the sun. These Kepler's musica are all peculiarities, of which are interconnected, the subjectof this article. First, there is the basic problem of the objective validity of Kepler's celestialharmonies. Kepler, when finding the ratios of musicalconsonances in the extremeangularspeeds of the planets as seen from the sun,5 allowed with his himselfsome marginof error. This was of coursequite in accordance metaphysics one would not expect to find an exact copy of a geometrical archetypein the naturalworld. But it is evident that, given a wide enough marginof error,one could find musicalratiosin any old set of numbers. Was playing Kepler, as AthanasiusKircher suggestedin his Musurgia(I650)6, a game with such lax rules that he was boundto win? Or did he in fact discover a pattern, a regularitywhich really does exist? I think the answeris that he was not playing too easy a game and that he had every reason to suppose that he had made a genuine discovery-for the following reason. (I 596), CSosmographicum to He had been trying, ever since the Mysterium find these ratios in the heavens, in the distances between the orbits of the 7 planets,and then in their orbitalspeeds, and he did not findthem. It is clear, therefore,that he was not willing to stretch his margin of error so that he could find them whereverhe looked for them. He found them only when he placedhimselfin the sun and lookedat the angularspeedsof the planetsfrom there. Kepler'sinsistencethat his celestial harmoniesshould be real, should be empirically confirmed by astronomicalobservation, is typical of all his thinking,and in particularof his thinkingabout music. While searchingfor, and discovering,purely metaphysicalor aestheticcausesfor things being as they are and not otherwise beautiful, simple patterns, mathematically determinedand logicallyinterconnected,he alwaysgave absolutepriorityto empiricalevidence;8if the theoreticalpattern, howeverbeautiful,did not fit the facts, it was discarded. Though Kepler was resolved that his celestial music should not be merely analogical,he by no means despisedanalogies; and, as we shall see, it is not alwaysclear whetherhis musicaland geometric analogies are only metaphoricalor whether they expressa real connexion betweenthe two termsof the analogy. The two novelties in Kepler's celestial music of polyphony and just intonation are closely connected; if thirds and sixths are not admitted as consonances,there can be no polyphony. Although in the sixteenth and seventeenthcenturiesthe questionwas debatable, Kepler believed, with the majority of competent scholars, that ancient music, though perhaps not
of 5 It is the extreme speeds that give the criticizedthe largeness Kepler'smarginof basic 'scales' of each planet; other speeds error. Mundi Libri V 7 See Kepler, Harmonices within these limits are also used to make the (hereafter Harm. Mund.), Lib. v, c. iv harmonies;cf. infra,pp. 247-8. Werke,ed. Max Gaspar, vi, Universalis, (Gesammelte Kircher,Musurgia 6 Athanasius ii, Rome I 650, p. 379: 'Ludereautemin sola MunichI 940, pp. 306-I 2), and Max Caspar's (ibid.,pp. 470ff.). nulliusingenijest, cum vix ullp Xachbericht proportione, Founda8 Cf. E. A. Burtt, TheMetaphysical numerissubiectaeres sint, quae non aliquas London I949, Science, Physical ex musicis proportionibusdenominationes tionsof Modern habeant'; he has just (ibid., pp. 377-8) pp. 50-5I.

.
230

D. P. WALKER

strictlymonodic,was not polyphonicin any way resemblingmodernmusic,9 and that this differencewas reflectedin the prevailingsystemof intonation: Pythagorean(in which the thirds and sixths are dissonant)for the ancients, and just (in which they are consonant)for the moderns.l This last opinion was also debatable. The acrimonious controversybetween Zarlino and VincenzoGalilei,ll both of whom Keplerhad read,l2was about the question whethercontemporary capella singingwas in just intonationor in somekind a of temperedscale, Zarlinoasserting formerand Galileithe latter; but both the agreedthat ancientvocal musicwas monodicand usedthe Pythagorean scale. In fact, Galileiwas almostcertainlynearerthe truth than Zarlinoand Kepler, for the followingreasons. First,all systemsof intonation,includingequaltemperament,are mathematicalideals,l3 to which actual musical practice can approximateonly very roughly. Secondly, even an approximation much is more diEcult to achieve in just intonation than in Pythagoreanor equal temperament,because, unless some of the consonancesare tempered,just intonationis hopelesslyunstableeven in the simplest diatonic music: if all the intervalsin the following example are sung justly, the singerwill end a comma (B-O) flatterthan he began:l4 . A.

ff
9 See

43

81

insteadof I.

D. P. Walker, 'Musical Humanism humanaeuniusaccommodationem lyram, ad in the I6th and early I7th centuries'(here- quaehodievoluptatis causapassim revocatur: after'MusicalHumanism'), TheMusicReview,sed unius simplicis vocis modulationem suavioremesse quatuorvocibus in varietate I94I-42, sectionsix-xi, and infra,p. 23I. 10Kepler does not explicitly state this identitatemtuentibus,numquamcredidero. connexion the Harmonice in Mundi, I think At nuspiam legimus cecinisse illos diversis but only because it is so obvious; he is a very vocibus in unum.') Reimarus Ursus was ellipticalwriter. It is a commonplace the Imperial Mathematicianand an enemy of in I6th century musical treatiseshe used (vide Tycho Brahe; I have not been able to find infra, note I2). In a letterof I599 to Herwart anythingaboutmusicin his published works. von Hohenburg (Kepler, Ges. Werke, xiv, 11See Walker, 'MusicalHumanism',end p. 72) Keplerdoesmakethis connexion quite of sect.iv. 12 Caspar(Kepler, Ges. Werke, p. 477) vi, clearly: after explaining how the Greeks, owingto theirPythagorean intonation,failed gives a list of musicaltreatises which Kepler to use the imperfect consonances, goes on: had read; this does not include Zarlino,but he 'Since this is the case, I am extremelysur- Keplerciteshim in the Harm. AWund. (Kepler, prised. . . that Ursusshouldthinkthe music ibid., p. I39). The list wrongly ascribesto of the ancients much nobler than ours. I J. T. Freigius a work by F. Beurhusius MusicaeLibri duo, Noribergae believethat one voice singingto the lyre had (Erotematum its own grace,and thisfor the sakeof pleasure I580), for which he wrote the preface. l3Just intonation,and to a lesser degree is being revivedeverywhere nowadays;but I have a physicalbasis:the series shall never believe that monody is more Pythagorean, delightfulthan four voices preservingunity of overtones, and combination tones; but in in variety . . .' ('Quae cum ita sint, vehe- thesewerenot yet discovered Kepler'sday. menter miror Ursum (et antea quoque Cf. infra,p. 24I. 14 Cf. J. M. Barbour,op. cit., pp. I96-9, mirabar, quam haec scirem), qui veterum der Musicam putat longe nobiliorem fuisse and WilhelmDupont, Geschichte musikalinostra. Credogratiamhabuissesuam, vocis schen Eemperatur, KasselI 935, pp. I I-I 2.

KEPLER'S CELESTIALMUSIC

23I

Nevertheless, thoughnone of thesesystemsis everexactlyput into practice, it is not a matterof indifference practicalmusicwhich of them prevailsas for a theoreticalideal, because musicianswill attempt to attain it and produce diXerentresults,which are easily distinguishable ear.l5 For music which by is monodic,or in which the interestis concentratedon melody, Pythagorean intonationis more suitable than just, since all the fifths and fourthscan be untempered,and the very narrow semitonesgive greater sharpnessto the shape of the melody.l6 For polyphonicmusic such as that of the sixteenth to the nineteenthcenturies,in which the majortriad occupiesa dominating and central position,just intonationhas the advantageof makingthis chord as sweetas possibleand in generalof makingall chords,both majorand minor, more consonant,though it has the disadvantageof much greaterinstability of pitch, of unequaltones,and of much widersemitones(1lW comparedwith as 22w3-, diffiering a comma, 84). These remarksare borne out by the history by of Western music. Music in the ancient world was monodic, or at least dominated by melody, and the standard intonation was Pythagorean; althoughPtolemy,and beforehim Didymus,gave the ratiosofjust intonation, they did not accept thirds and sixths as consonances.l7Pythagorean intonation, transmittedmainly through Boethius and Macrobius,l8was the only system known to medieval musical theorists;but with the full development of polyphonyin the later Middle Ages theoristsbegin to accept thirds and sixths as 'imperfectconsonances',though still giving the Pythagoreanratios (WiF 32272 217 2 l-82-l8) . 1 9 If medieval musicians were aiming at Pythagorean intonation,their majortriadswould be no more consonantthan their minor triads; and in fact it is not until the later sixteenth century that harmony beginsto be dominatedby the majortriad, as opposedto the minor,and that majorand minor tonality begins to replacethe modes. This bringsus to the period of Zarlino, the first widely read and influentialtheoristto advocate
15To convince himself that he can hear the diffierence betweenjust and Pythagorean thirds and sixths, the reader who owns a violinor 'cellomay makethe following simple experiment. Having tuned the instrument as accurately as possible, play E on the D-stringwith the open G-string;then, taking carenot to move yourfinger,play the E with the openA-string. If the majorsixthhasbeen madeas sweetas possible, will be foundthat it the finger has to be leaned considerably forward to produce a perfect fourth. The differencebetween the two E's is a comma (80) Then try the experiment other way the round. 16 Present-day violinistswho believe that they are playing in 'natural' or 'true' intonation,as opposedto equal temperament, make very narrowsemitonesby sharpening upward leading-notesand flatteningdownward ones. In consequence,e.g. G sharp followedby A is sharperthan A flat followed by G; and in consequence thistheirdoubleof stoppedthirdsand sixthsare very harsh. In other words they are attemptingto play in Pythagoreanintonation, as befits a melody instrument. They are of coursealso pushed towardsthis kind of intonationby the fact that their instrumentis tuned in fifths (cf. J. M. Barbour, cit.,p. 200). op. 17Ptolemy,Harm., i, c. xv, and Musik lib. in G. &)9 (hereafter G. MGG),articles'Intervall', 'Didymos'. 18 Boethius, De Institutione Musica,lib. i, c. vii and passim; Macrobius,Commentariorum in Somnium Scipionis LibriII, lib. ii, c. i (it is interestingthat, although Macrobiustransmits the Pythagorean ratioscorrectly,he did not understand that they were ratios;having stated, rightly, that the tone cannot be exactlydividedinto two halves,he givesas a reason that g cannot be divided into two equal integers the true reason being, of course,that there is no rationalsquareroot of 98). 19MGG,article'Intervall',cols. 1344-5.

232

D. P. WALKER

the use ofjust intonation,20 which the majortriadis muchsweeterthan the in minor. The great growth of instrumentalmusic and the developmentof harmonytowardsgreaterfreedomof modulationfrom the sixteenthto the eighteenthcenturiesare reflectedin the eventualtriumphof equal temperament as the ideal intonation. All instruments fixed intonationmust have of some kind of temperament, and in the sixteenthcenturyan attemptat equal temperament was used for frettedinstruments(lutes and viols), and usually meantonetemperament,2l which provides just thirdsand sixths,for keyboard instruments. For instrumentalmusic and for any music which modulates freely equal temperament had enormousadvantages:as comparedwith just intonationor meantonetemperament, its fifthsand fourthsare very nearly all true and its semitones narrower; compared as with Pythagorean, thirdsand its sixths are less dissonant;as comparedwith any other system, all keys are equallyin tune, and there is no instabilityof pitch. We may say then that Kepler was right in accepting a real connexion betweenthe growthof polyphonyand the prevalenceof just intonationas an ideal, thoughhis reasons this acceptanceare not of coursethe sameas those for I have just given. For Kepler just intonation and polyphony had finally prevailedbecausethey were natural,that is, they corresponded the archeto types in the mind of God,22on which the createdworld was modelled,and which are also in the mind of man, the image of God. Modern music and intonation are thus justified and welcomed by Kepler in two ways: first, empirically,because an unprejudicedobservergifted with a good ear can realizethat thirdsand sixthsare consonant they will please and satisfyhim because they correspondto the archetypesin his mind, and, by using a monochord,he can discoverthat their ratiosare 5: 4, 6: 5, etc.; secondly,the investigator of nature can find these consonancesin God's creation, in Kepler's case in the harmony of the spheres, and thus confirm his own empirical knowledge and the instinctivelynatural, polyphonic practice of modernmusicians. The finalstep is to showby reasonsdrawnfromgeometry, that supremeset of archetypes which is coeternalwith God, why these ratios and no othersproducemusicalconsonances. Kepleris alwaysmost emphaticin affirmingthat polyphonyis a modern inventionand thereforequite unknownto the ancients,though for historical evidencehe merelyrefersthe readerto Galilei'sDaalogo Musaca della antica et moderna58I ) ;23 and, unlike most of his contemporaries sees this as the (I he
20 Ramis de Pareia (Musica Practica, GirolamoMei (see C. V. Palisea, Girolamo BolognaI482) and Foligno (MusicaTheorica, Mei (I5I9-I594) Letters Ancient Modern on and Veniee I 529) had both giventhe ratiosofjust Music toBardi Vincenzo Galilei and Giovanni eonsonanees. n.p. (Ameriean Institute of Musieology), 21 So ealled beeause the major third is I960), waswritingagainstmodernpolyphony dividedinto two equal tones (ratio t/25); see andjust intonation,and in favourof a revival PercyC. Buck,Acousticsfor .Musicians, Oxford of ancient monody and PythagoreaninI 9 I 8, pp. 95-99. tonation. But in a letterof I 6 I 8 to Matthaus 22 For Kepler's arehetypes Pauli'seon- Waekervon Waekenfels, see Keplertells how in tributionto C. G. Jung and W. Pauli, The Oetober I6I7, when settingoSfrom Linz to Inter87retation of J%ature thePsyche, and London Regensburg,he foresawa slow journey and I 955. thereforetook with him Galilei's dialogue, 23 In the Harmonice Mundi Keplerdoes not whieh, thoughhe found the Italian diffieult, mentionthat Galilei, under the influeneeof he read with the greatestpleasure;in it he

KEPLER'S CELESTIALMUSIC

233

and unique advance that it was,24an advance that for him is extraordinary paralleledby the new astronomyand his own discoveryof the celestialpolyMundi,after he has gone through phony. In the Fifth Book of the Harmonice the 'scales'played by each planet, which are like 'simplesong or monody,the he only kind knownto the ancients',25 beginshis chapteron the chordsmade and by five and by four of them, thus:27 by all six planets,26 Now, Urania, a more majestic sound is needed, while through the harmonicladder of celestial movementsI ascend yet higher, where the true Archetypeof the world'sstructurelies hidden. Follow me, modern musicians,and expressyour opinion on this matter by means of your arts,28unknown to antiquity; Nature, always generous with her gifts, has at last, having carriedyou two thousandyears in her womb, brought of you, the firsttruelikenesses the universe; you forthin theselast centuries, by your symphonies of various voices, and whispering through your ears, she has revealed her very self, as she exists in her deepest recesses, to the Mind of man, the most beloved daughterof God the Creator. Though modernmusic revealsthe archetypicalstructuresof the heavens, it is not an imitation of the celestial music,nor derived from it; but both are
abouttheancients, foundvaluableinformation and, although he often disagreedwith the author'sopinions,he enjoyed the virtuosity with which Galileiexpoundedviewsopposite to his own, by extolling ancient music and xvii, modern(Kepler,Ges.Werke, denigrating p. 254: 'inveni enim thesaurumantiquitatis egregium, et quamvis in re ipsa crebro ab ipso dissentiam,delectatustamen sum artificio disputantis in contrarium, et in re Mathematica oratorem agentis, praesertim ubi veterem Musicam extollit, novam deprimit'). Kepler cites Galilei much more frequentlythan any other modernwriteron
music. letterquotedabove,p. 230, 24 Cf. Kepler's pp. n. I 0, and infra, 234-5. vi, 25 Kepler,Ges.Werke; p. 3I6: '. . . quae

est proportio Cantussimplicisseu Monodiae, quam ChoralemMusicamdicimus, et quae sola Veteribus fuit cognita, ad cantum plurium vocum, Figuratum dictum, ineadem est saeculorum: ventum proximorum proportio Harmoniarum,quas singuli dejunctorum.' signantPlanetae,ad Harmonias 26 The moon is excluded because 'Luna seorsimsuam Monodiamcantillat,Terrisut cunisassidens'(Kepler,ibid.,p. 323). In the translationof the Fifth Book of the Harm. Mund.by CharlesGlenn Wallis (in Ptolemy, . TheAlmagest. ., (GreatBooksof the Western World,no. I6), I952, p. I040), the lastpartof this sentenceis rendered:'like a dog sitting
I6

being read on the earth', 'canis'presumably for 'cunis' not a happy emendation; the translationas a whole is very poor. The Germantranslationby Casparof the whole of the Harm. Mund. (Kepler, Welt-Harmonik, I Munich-Berlin 939) is of courseexcellent. 27 Kepler, ibid., p. 323: 'Nunc opus, Uranie, sonitu majore: dum per scalam coelestiummotuum,ad altiora Harmonicam fabricae Archetypus qua conscendo; genuinus asservatur.Sequimini Mundanaereconditus Musicimoderni,remquevestrisartibus,antiquitati non cognitis,censete:vos his saeculis exemplagenuina, ultimis,primauniversitatis bis milliumannorumincubatu,tandemproduxit sui nunquam non prodiga Natura: vestris illa vocum variarum concentibus, perque vestras aures, sese ipsam, qualis sinu, Menti humanae,Dei existatpenitissimo insusurravit.' Creatorisfiliae dilectissimae 28 I.e. 'prove me right by using justly intoned polyphony'. In a side-noteKepler shouldwrite that moderncomposers suggests six-partmotetson one of the Psalms,or some text, in returnfor this eulogy otherscriptural he has given them. Keplerwill see that they are published,and says that: 'He who most the nearlyexpresses celestialmusicdescribed a in this book; to him Clio promises garland, Urania promises Venus as a wife' ('Qui propius Musicam coelestem exprimet hoc opere descriptam;huic Clio sertum,Urania Veneremsponsamspondent.')

234

D. P. WALKER

of likenesses the same archetypes,the geometricbeauties coeternalwith the Creator; and modern music, as we shall see, thereby even allows us to in experiencesomethingof God'ssatisfaction His own handiwork. Kepler's whole-heartedand joyous acceptance of polyphony as a step forward,comparablein importancewith the Copernicanrevolution,is in even of those who marked contrast to the attitude of his contemporaries, such alsobelievedthat ancientmusicwas monodic. Zarlinoand his followers, of or a mesuree l'antique the FlorentineCamerata as the composersof musique Bardi, concede that modern music has acquired additional sweetnessand varietythroughthe use of polyphony,but they also believe that, with regard to rhythm and the treatmentof text, we still have much to learn from the thereis no feelingthat musichas acquiredanotherdimension,but ancients;29 merelythat one aspectof the art has been elaborated,while anotherequally, or even more important aspect has degenerated.30Another, more subtle contrast with Kepler is provided by Sethus Calvisius,with whom Kepler and on had a long correspondence musicand on chronology,31 whosemusical 32 Mundi. treatises he recommends, rather lukewarmly, in the Harmonice eo rebus spectantibus aliisque Musices, Calvisius,in his essayDe Initioet Progressu (I600),33gives a competent,if brief historyof musical theory and practice from the Flood to the presentday, and from it Kepler could have gathered all the elementsnecessaryto producea realizationof musicalprogress. The ancientsrejectedthirdsand sixths; their musicwas monodicor nearlyso; at some time in the Middle Ages polyphonywas invented, and soon became 'the at decadentlyover-complicated;34 the time ofthe Reformation, repurging of celestialdoctrine,togetherwith othergood artsand languages',an improvement in musical style began, especiallyin the treatmentof text, which has reached its culmination with Orlando di Lasso and other more recent Music has now attained such heights that no furtherprogress composers.35 seems possible;all we can do now is to use it to thank God that in this last age of the world He has advancedthis art, 'among the other liberal arts, to as its highestperfection', a preludeto the musicof the ChurchTriumphantin however heaven,soonto beginand neverto cease.36UnlikeKepler,Galvisius thoughhe knowsand statesthat the ancientshad no polyphony,neversingles
&? recte vocant, cognoscendis, dijudicandis. See Walker,'MusicalHumanism',sect. Tonos de Posterior, Initio. . ., Lipsiae I 600. 9I-94, I24-8. 34 Calvisius, op. cit., pp. 30 G. M. Artusi, another author whom Kepler cites (ibid., pp. I8I, I82, I85), iS CalVlSlUS 1S referring mainly to the comalmost as severe on modern polyphony as plexities of medieval rhythmical notation. delle 35 Calvisius, ibid. pp. I33-5, 'usque ad overo Galilei and Mei. In his L'Artusi coelestis doctrinae, una cum bonis artibus & Ragionamenti Musica della Imperfettioni Moderna as linguis, repurgationem . . .'. is Venice I600, polyphony condemned due, (last page), 36 Calvisius, ibid., p. I38 because,by its mixture positivelypernicious, modesand genera, '. . . quod hoc ultimo mundi articulo, inter of and confusion rhythms, alias liberales artes, hanc etiam ad summam of the it prevents production the 'eXects'. 31 The letters concerning music are: perfectionem deducere, & quasi wpoauxLov xv, Kepler, Ges. Werke, pp. 469ff.; xvi, pp. praelusionem fieri voluit [sc. Deus], perfectissimae illius Musicae in vita coelesti, ab 47i, 55ff8., 2 I 6ff.; xviii, pp. 455i universo triumphantis Ecclesiae & beatorum vi, p. I85. 32 Kepler,ibid., MusicaeDuae. Angelorum choro, propediem inchoandae, 33 Calvisius, Exercitationes quosvulgo & per omnem aeternitatem continuandae'. Priorest, de Modismusicis, Quarum
29 1X.

KEPLER'S CELESTIALMUSIC

235

out this fact as an exampleof progress,and he sees the presentgood state of music on a par with that of the other liberal arts, which have been revived afterthe long medievaldarkness. The main reason, accordingto Kepler, why this musical revelationand revolutionwas so long delayedwas that the ancientsdid not stay close enough to empiricallyestablishedfacts, to the judgementsof the ear. In the preface Mundi,which deals with practicalmusic, to the Third Book of the Harmonice discoveredby Kepler gives a brief historyof intonation. The Pythagoreans ear the perfect consonancesof the octave, fifth and fourth, and their ratios (2: I, 3: 2, 4:3); but then turned away too soon from the evidence of their ears and towards speculation in numbers a double error, first in that musical theory must not only start from observationbut also be constantly checkedby it, and secondly,in that the groundsof consonancemustbe sought not in numbers,but in geometry:37 The Pythagoreanswere so addicted to this kind of philosophizingin numbers,that they failed to keep to the judgment of their ears, though it was by means of this that they had initially been broughtto this philosophy; they defined solely by their numberswhat is a melodic interval and what is not, what is consonant and what dissonant, thus doing judgmentof the ear. violence to the naturalinstinctive Thus misled, the Pythagoreans, and following them Plato,38 restricted and therefore consonancesto ratios made out of their tetractys (I.2.3.4.), failed to include thirdsand sixths,withoutwhich there can be no polyphony. They wrongly accepted as a melodic interval the Pythagoreansemitone,or Platonic limma, 2243 (the differencebetween two major tones and a fourth: between ' X *. (8 X 8) ), and wronglyexcludedthe minortone, 1f (the difference cona just majorthird and a majortone: -w . 98) This 'harmonictyrannwr' tinued until the time of Ptolemy,who, maintainingthe judgementof the ear philosophy,admittedas melodicintervalsthe minortone againstPythagorean and (1-90) just semitone (1l-), and gave the ratios of just thirds and sixths But Ptolemy, though he had thus emended the Pythagoreans' (4, 5, 3, 5). system and rightly trusted his ear, was still misled by their preoccupation with 'abstractnumbers',and in consequenceboth wrongly excluded thirds and sixths from the consonances,which 'all well-eared musiciansof today' accept, and wronglyincluded among the melodic intervalsa divisionof the fourthinto 6 and 87, which is 'most abhorrentto the ears of all men'.39 Kepler has several reasons for insisting that the causes of consonance must be sought not in numbersbut in geometricalfigures. First, one cannot reasonwhy God shouldhave chosenthe numbersI.2.3.4.5.6. find any sufficient as those out of which consonancesshould be generated,and have excluded
vi, 37 Kepler, Ges.Werke, p. 99: 'Huic enim philosophandi formae per Numeros, tantopere fuerunt dediti Pythagoraei; ut jam ne aurium quidem judicio starent, quarum tamen indicijs ad Philosophiam hanc initio perventum erat: sed quid concinnum esset, quid inconcinnum; quid consonum, quid dissonum, ex solis suis Numeris definirent, vim
facientes instinctui naturali auditus.' vi, 38 Kepler, Ges. Werke, pp. 94-95, IOO; cf. his earlier criticism of Plato in a letter of

p. 99; Ptolemy, Harm.,lib. i, c. xv. ihe appendix to the contains a critique of Ptolemy's Mund. Harm. musical analogies (Kepler, ibid.,pp. 369ff.).

I599, xiv, pp. 7I-72. vi, 39 Kepler, Ges. Werke,

236

D. P. WALKER

7.II.I3. etc.40 The reasongiven in the fimaeus, namely the two familiesof squares and cubes generated by the triad I.2.3, itself the principle of all things: I

2
4 8

3
9

27

is no good, becauseit excludesthe number5, 'whichwill not allow itselfto be robbedof its right of citizenshipamongthe sourcesof consonances',4l that is, of the thirds and sixths, all of which in their ratios have the number 5. Secondly,numbersare not suitableas causesof musicalintervals,becausethe terms of musicalratios are continuous,not discretequantities,and therefore these causesmust be sought in geometricalfigures.42By the terms (termini) of the intervalsKepler must mean musical sounds of differentpitches, and presumablybelieves that these have no natural units by which they can be counted. This is odd, since, when explaining (correctly)sympatheticvibration,43Kepler is evidentlyconsidering musicalsound as made up of a series of pulses (ictus) caused by a vibrating string, and these would provide a naturalunit for counting,as we now countfrequencies.But Keplermusthave been in a bit of a muddle about the nature of musical sound, since he apparentlybelieved that the pitch of a string falls as the amplitude of its vibration decreases.44Finally, numbers are metaphysicallyand epistemologically inferior to geometricalfigures and proportions. Numbers do not exist in physicalthings, but only 'dispersed units' so exist; numbersare thus abstract,in the sensethat an Aristotelian tabula mind could develop them rasa by abstraction from the repetitivesense-experience any kind of unit-they of are 'of second, even of third or fourth intention'.45But this is not true of geometricalfiguresand proportions;these do exist, as imperfectcopies, in physicalthings; and the mind or soul recognizesand classesthem by comparing them with the God-implanted archetypes within itself.46In the Harmonice Mundi Kepler quotes a long passagefrom Proclus'scommentaryon Euclid, which is a defenceof the Platonicdoctrine,that all mathematicalideas exist
40Kepler,ibid.,p. IOO; cf. infra,p. 24I on Kepler'srejectionof harmonicproportion. 41 Kepler, ibid.: 'Nam causa illa de Ternario principiorum, et familia quadratorumet cuboruminde deducta,causa est nulla; cum quinariusab illa exulet, qui sibi inter MusicorumintervallorumOrtum jus civitatiseripi non patitur'; cf. Kepler, ibid., pp. 94-95; Plato, Timaeus, B. 35 42 Kepler, ibid., p. IOO: 'Cum enim intervallorum Consonorumtermini, sint quantitates continuae: causas quoque quae illa segreganta Dissonis,oportetex familia peti continuarumquantitatum,non ex Numeris abstractis, quantitatediscreta. . .' ut 43 Kepler,ibid.,pp. I05-6. 44 Kepler,Ges.Werke, p. I44. vi,
45 Kepler, ibid.,p. 43I (Apologia against Fludd): 'Omnis numerus, ut sit numerus, menti inesse debet, ut docet Aristoteles;in sensibusinque materianumerusnon est, sed unitates dispersae';p. 2 I 2, 'Numerus definitur esse multitudo ex unitatibus conflata . . .'; p. 222, 'suntenim illi [sc. numeri] secundae quodammodointentionis,imo et tertiae, et quartae, et cujus non est dicere terminum:nec habentin se quicquam,quod non vel a quantitatibus, ab alijs veris et vel realibus entibus, vel etiam a varijs Mentis intentionibus acceperint'. 46 Kepler, ibid., pp. 2I5-I6; the mind recognizes these proportionsintellectually, the soul instinctively.

237 CELESTIAL MUSIC KEPLER'S innate in the soul, against the Aristotelian epistemology of their being universalsabstractedfrom multiple sense-experience.He then concedesthat Aristotlewas right as far as numbersare concerned,and was right to refute and Pythagorean number-philosophy, Kepler himselfrejectsPlato'snumerology in the Republic; 'with regardto continuousquantitiesI am entirely but in agreementwith Proclus'.47 This inferiority numbersto geometricfiguresand ratiosis importantfor of Kepler'sattitudeto analogiesor symbols. Analogiesbasedpurelyon numbers correspondto no archetypein the soul of man or mind of God, whereas more and, in many cases,are therefore geometricanalogiesdo so correspond, than analogies:they displaythe reasonswhy God createdthings as they are and not otherwise, or why we are pleased or displeasedwith certain exagainst Fludd, but elsewhere in the periences. Not only in the Apologia Harmonice Mundi,Kepler takes care explicitly to reject any number-symbols which might suggestthemselvesto the reader.48 Harmony,musicalor of any other kind, consistsin the mind'srecognizing between two or more continuousquantities and classingcertain proportions by meansof comparingthem with archetypicalgeometricalfigures. Now we know by experiencethat there are seven musical consonances,which have these ratios: 21, 32, 43, 4, 65, , 58, and which can be multiplied indefinitelyby doublingtheir ratios,i.e. by insertingoctaves (32, e.g., a fifth, when doubled becomes31, a twelfth). What classof geometricfigureswill yield these ratios, and no others? As early as I599 Keplerwas lookingfor the answerin the arcs polygons.49 constructable, inscribed of a circlecut offby regular,geometrically There are two main reasonswhy he should have looked here. First, he had already had at least partial successin using the five regular Platonic and solidsto accountfor the numberof planetsand the size of their orbits;50 he had, as he later wrote,5l in the Mysterium Cosmographicum wronglyattemptedto deduce the numberand ratios [of the consonances] from the five regularsolid bodies, whereasthe truth is rather that both the five regularsolid figuresand the musical harmoniesand divisionsof the monochordhave a commonoriginin the regularplane figures, which by that is, the numberof the regularsolidsis determined their surfaces,
47 Kepler, ibid., pp. 2I8-22, 'De numeris Paris I 96I, pp. I 43ff. 51 Kepler, Ges. Werke, vi, p. I I9, 'Legat quidem haud contenderim;quin Aristoteles recte refutaverit Pythagoricos.... At quod curiosuslector, quae de his sectionibusante Cosmographico, attinet quantitates continuas, omnino ad- annos 22 scripsi in Mysterio Capite XII et perpendat,quomodo fuerim sentiorProclo.' super causissectionum 48 E.g. Kepler, ibid., p. I23. (There are illo loco hallucinatus six possibleconsonanttriads; this fact is not et Harmoniarum; perperam nisus earum to be explainedby the six daysof creationand numerum et rationes deducere ex numero the Trinity.) quinque corporum llegularium solidorum: 49Kepler, Ges. Werke, xiii, pp. 349-50, cum verum sit hoc potius, tam quinque letter to Herwart von Hohenburg, dated figurassolidas,quam HarmoniasMusicaset 30 May I599 (consonances from arcs of a chordae sectiones, communem habere eirele); xiv, pp. 29-37, letter to the same, originem ex figuris Regularibus planis.' 6 August I 599 (consonancesfrom regular Myst. Cosm.(I 596) on consonancesin Ges. inscribedfigures). Werke, pp. 40-43. i, 50 See A. Koyre,La Revolution astronomique,

238

D. P. WALKER

must be regular polygons, and the three basic regular polygons, triangle, square and pentagon, can generate only five solids. Secondly, Kepler had archetypicalreasonsfor using divisionsof a circle rather than of any other figure. One of his favouriteanalogies,which is certainlymore than a metaphor, is that of the sphererepresenting Trinity; the centre is the Father, the the surface the Son, and the interveningspace the Holy Gllost.52 In the Harmonice Mundi,when explaininghis use of the circle as the cause of consonances,he recallsthis 'symbolisatio' extendsit. A sectionthroughthe and centre of the sphere producesthe plane figure of a circle, which represents the soul of man; this sectionis made by rotatinga straightline, representing corporeal form,which extendsfromthe centreof the sphereto any point on its surface; thus the soul is to the body as a curve to a straightline, that is, 'incommunicable and incommensurable'; the soul is to God as a circle and to a sphere, that is, partakingof the divine three-dimensional sphericity, but joined to, and shaping, the plane generatedby the bodily line. 'Which cause', continuesKepler, 'established Circle as the subjectand sourceof the termsforharmonic proportions'. 53 Using only a rule and compasses,one can divide the circumference a of circleinto equal partsin only four basicways (with one exception,the pentecaidecagon,which will be dealt with later), namely, by inscribingin it its diameter, an equilateral triangle, a square, and a pentagon; by con54 tinuouslydoublingthe numberof sides of these figuresan infinitenumberof furtherdivisionsis possible. Figures,sucll as the heptagon,which cannot be so constructed,are not demonstrable,and are thus 'unknowable',even to God;55they are thereforeexcluded from the archetypes. The arcs cut off by thesebasic demorlstrable figuresprovidethe followingratiosby comparing the arc subtendedby one side with the whole circumference, and the arc subtendedby the remainingsideswith the whole: Onesideto whole Residue whole to diameter 1: I. Octave 2: I triangle I :3. Twelfth 2:3. Fifth square I :4. Double octave 3: 4 Fourth pentagon I :5. Double octave plus 4:5. Majorthird. majorthird. This gives us all but three of the seven basic consonances:minor third and sixth, and major sixth. The last can be obtained by dividing the pentagon into 2 and 3; which yields 2:5, a tenth, and 3:5, a majorsixth. Just as an infinite number of consonancescan be generatedby doubling the ratios, so there are an infinitenumberof regularpolygonsobtainableby doublingthe number of sides. By using two of these polygons,hexagon and octagon, we can get the missingconsonances: 5:6, minorthird, and 5:8, minorsixth.56 The salientfeatureof this method of explainirlg ratiosof consorlances the is that it does not work very well, and it does not work well becauseof the 52 Kepler,ibid., pp. 9, 23-24i, 55 Kepler, ibid., vi, pp.47ff.
Kepler,ibid., p. 224. vi, The square is really a doubling of the diameter.
53 54 56 Kepler, Ges. Werke, pp. IOI-I8. vi, minorthird, cf. infra, 243, n. 75. p.

For

KEPLER'S CELESTIALMUSIC

239

thirdsand sixths. This is the main point I want to make here: if Kepler had accepted the still current,very ancient Pythagoreanand Platonic systemof intonation,involvingonly the consonancesI: 2, 2: 3, 3: 4, he would have had no difficultiesat all; but he did not accept it, and that on purely empirical grounds because,beforehe set out on his investigationinto causes,he had already establishedby ear that just thirds and sixths are consonant. The Pythagoreansystem would have fitted Kepler's geometricalexplanationso well because he could have defined the admissiblepolygonsas those whose sides are either directlycommensurable with the diameterof the circle (the diameteritself) or commensurable square (the triangleand the square),57 in and he rleed have used no other figures. We may, I think, take Kepler's wordforit that he did originallyadoptjust intonationsolelyon thejudgement of his ear. He emphaticallystates this in the Harmonice Mundi, gives as and evidence the fact that he already used this system of consonancesin the Mysterium Cosmographicum, at a time when he was still far fromfinding that is, any satisfactory theoretical justificationof it:58 The evidence of my book the Mysterium Cosmographicum will be alone enough to protect the sense of hearing against the objectionsof sophists who will dare to derlythat the ear can be trustedin such minutedivisions [of the monochord] and such very subtle distinctionsof consonances. For the readerwill see that there I relied on the judgment of the ear in establishingthe numberof divisions[i.e. consonances],at a time when I was still strugglingto find causes, and that I did not then do what the Ancients did. They, having advanced a little way by the judgment of the ear, soon despisedtheir guide and finishedthe rest of their journey followingmistakenReason,having, as it wereforciblyled their earsastray and orderedthem to be deaf. Moreover,it is clearfromhis correspondence he was in the habit of using that a monochord,and he gives advice on how to achieve more accurateresults by checkingthe consonanceone is investigatingwith its residue;for a major third, e.g., check 45 with 1, i.e. 4, a double octave.59 He also gives, in the Harmonice Mundi, ingeniousmethodof makingaudiblethe slight errorin a an rough and ready kind of equal temperamentused on lutes and describedby
Galilei.60

In Kepler'scorrespondence the year I599, when he began his harmonic of investigationswith the regular polygons, it is always the thirds and sixths that give trouble. The pentagon, necessaryfor the majorthird and sixth, is
57 (Side of triangle)2 = 3(radius)2;(side Edem aurium illo temporesecutumesse, in of square)2 = 2(radius)2. The hexagon constituendo sectionumnumero,cum adhuc might have given trouble; but I am sure de causis laborarem;nec idem hlc fecisse, Keplerwould have found a way roundit. quod fecere Veteres; qui aurium judicio 58Kepler,ibid.,pp.IIg-20,'Igiturvelsolo progressialiquatenus,mox contemptisduciallegato mei Mysterij Cosmographicites- bus, reliquum itineris, Rationem erroneam timonio, satis est munitus auditus, contra secuti, perfecerunt;auribus vi quasi perSophistarumobtrectationes,fidem auribus tractis,et plane obsurdescere jussis'. derogare ausuros circa divisiones adeo 59 Kepler, Ges. Werke, xvi, p. I59; ef. xv, minutas,et dijudicationem concordantiarum p. 450. subtilissimam: quippe cum videat lector me 60 Kepler, ibid.,vi, pp. I43-5

gf t

12 2 52@@@ 45 6 **t .............. w21.... i f S 3 .................. 4 w< .......t ............. 4 ....... w-w-ww..............

I I 97 ,. 7I I

24o

D. P. WALKER

indeed constructable,but its sides are incommensurable with the diameter even in square.6l The octagorl,necessary the minor third and sixth, also for has irrationalsides even in square;but in any case by what rule do we allow it to divide the circleinto 3 and 5 parts,but excludethe divisioninto I and 7, which would producedissonantintervals?With regardto the pentagon,the answerwas to be that its irrationality involvesthe 'divine'proportionof the golden section, to which we shall return. For the octagon Kepler tried out varioussolutions,usingregularsolids,stars,comparingthe arcsnot only with the circle but also with the semicircle,etc.;62and finally arrivedat the rule he uses in the Harmonice Mundi, namely, that the harmonicsectionof a circle must be such that the two partscomparedboth with the whole and also with each other produce ratios that do not involve numberssuch as 7.9.I I.I 3., which are the number of sides of undemonstrable figures (heptagonetc.).63 Thus the octagonmay dividethe circleinto 3 and 5 parts,since-, 8, 8- involve no 'ungeometric' numbers,but not into I and 7 parts. By meansof this rule, which comes dangerously close to being an arithmeticratherthan geometric explanationof consonance,Kepler already in I 5 9 9 gives the neat table of consonantratiosthat appearsin the Harmonice Mundi:64
f S 15<

6fffffffff

rlv

8 I

These fractions are generated by adding numerator and denominatorto form a new denominator,which has as numeratorsboth numbers of the previous fraction, the generation being blocked by the appearanceof an 'ungeometric' number. These sectionsof a circle or a string are not of course harmonicin the usual mathematical and musical sense of the term. Kepler does define ordinaryharmonicproportionand give the formulafor findingan harmonic mean.65Three numbersare in harmonicproportion, the greatestis to the if least as the difference betweenthe greatestand the middleis to the difference betweenthe middle and the least (ac= (b_b); the formulafor the harmonic mean is therefore = (a2+6a)). b Thus in termsof string lengths,the harmonic division of an octave, I:2v gives us I, 3, 2, i.e. a fifth (I: 3) and a fourth
(Side of pentagon)2 = (i radius)2 Kepler, in the First Book of the Harm.Mund.,elaborates a detailed system, based on Euclid Book X, for grading the irrationality of the sides of polygons. For trouble about the pentagon in the letters of
61

(Io-2A/5).

I599, see Kepler, ibid.,xiv, pp.30-32,46-48, 65-66. 62 Kepler, ibid.,xvi, pp.3I-38,46-48. 63Ibid.,pp. 48) 66 64Kepler,ibid.,vi,p II865 Kepler, Ges.Werke, pp. IXo-I. vi,

Frequencies:

KEPLER'S CELESTIALMUSIC

24I

(3 2 I :-). Similarly,the harmonicdivisionof a fifth yields a major and a minor third; and that of a major third yields a major and minor tone. Kepler rejectsthis methodof generatingmusicalconsonances the grounds on that there is an indefinite number of harmonic proportionswhich yield consonantratiosbetweenthe extremetermsbut not betweenthe middle and the extremes,e.g. 5, 27o, 2. It was very unfortunatethat Kepler'sdislike of numbersshould have led him to reject harmonicproportion,the standard explanation,from Zarlino onwards,of the consonances. Harmonicproportion and the harmonicseries I, 2, 3, 4 . . ., togetherwith the phenomenon of sympatheticvibration,point directly to the physicalbasis of consonance, namely,that a vibratingstringdoesin fact divideitselfup into partsa 1, 13, etc. Ofthe total length, which also vibrateat frequences 3, etc. times that of the 2, whole string,thus producingthe overtones:

String lengths: I

13

1 . . .

{&
Jr , . .

_ r

^ *)

1> _/t
A :S
Fundamental.Overtones

. ,

Nor is it anachronisticto make this comment. Descartes,in his Compendium Musicae, written at the same time as the Harmonice Mundi, on the was brink of discoveringovertones,66 and this, a little later, Mersenne in his achieved.67Moreover,this is the kind of physical explanationthat would have delightedKepler;but it would also have worried him. The harmonic divisions of a string continue indefinitely, and thus producethe 'ungeometric' divisioninto 1 and an interval (a very flat minor seventh) not accepted as consonant.68He would thereforehave had great difficultyin combininghis regularpolygonswith the seriesof overtones. Another difficultywith the regularpolygons,which Kepler had cleared up, to his own satisfaction, I607, was that raisedby the pentecaidecagon. by This figureis constructable and cannot be excluded on the groundsthat its sides are irrationaleven in square,since this would entail excludingalso the pentagon and octagon. But he was determinedto exclude somehow, and did so on the groundsthat it does not have its own independentconstruction, but can be constructedonly by combining a triangle with a pentagon.69

Harmonie Universelle (I636)

66 Descartes, Oeuvres, Adam and Taned. 87, 89) suggests that long custom might lead nery, xi, Paris I908, pp. 97, 99, I03. us to accept 7-ratios as consonant. The series 67 See Hellmut Ludwig, Marin Mersenne of overtones also contains, of course, an undseine Musiklehre, Berlin I935, pp. 40ff. indefinite number of other dissonant ratios. 68 Mersenne (Harmonie Universelle, Paris 69 Kepler, Ges. Werke, vi, pp. 46-47; cf. 636, Livre Premier des Consonances, pp. xv, p. 39I.

242

D. P. WALKER

At first sight it is not clear why Kepler should be so anxious to reject this polygon,since a circlecan, on his own principles, dividedby it into I 2 and be 3 parts and thus produceconsonantratios. But in the letter to Hewart von Hohenburgof January I607, which gives a summaryof the projectedHarmonice Mundi, find the reason,expressed a typicallyenigmaticway.70 we in And so this fifteen-angledfigure is sent back among the five foolish virgins. For it comes too late after all the doors have been shut by the numbers7.9.I I . I 3. That is to say: to include the pentecaidecagonas a consonance-generating polygonwouldspoilthe neatnessandeleganceof thetableof ratiosgivenabove. The acceptanceof the pentagon,althoughits irrationality greaterthan is that of the triangle or square, is justified, as I have mentioned, by that irrationalityinvolving the golden section.7l This is the proportionbetween three quantitieswhich fulfil the two followingconditions: (I) That they are in geometric proportion, that the greatesttermis to i.e. the middleterm as the middleto the least (ab b or b2 = ac). =
(2)

That the greatestterm is the sum ofthe two lesser(a =b+c;sothat

the formulais: b- (a b b) or b2 = a(a-b); thereforeb = a(v5-I)). = In other words, a line is divided into two parts in this proportion,if the whole is to the greaterpart as the greaterpart to the less. The side of an inscribeddecagonis to the radiusof the circleas the greaterpart to the whole in the golden section (decagon side - 12radius(f'5 I) ). The square on the side of a pentagon is equal to the square on the side of the decagon inscribed the samecircleplus the squareon the radius( (pentagonside)2 in r2 +4 (6-2 a/5); thereforepentagonside = 2r| I o-2 /5 ) . Also the side of a pentagonis to the line joining two of its verticesas the greaterpart to tTle whole in the golden section. Finally, and most importantlyfor Kepler, this proportionhas the propertyof generatingitself indefinitely:by adding the greaterpart to the whole one obtainsa new whole, and the old whole becomes the new greaterpart ((a +b)-b). Since the pentagon contains the divine proportionwithin itself, as that betweena side and the linejoining two vertices,andnot only,likethe decagon, in relationto the radiusof the circlein which it is inscribed,72 Kepleris able to regardthe pentagonas the archetypical figureof this proportion and hence of generationin general. In the Harmonice Mundihe reinforcesthe belief that the golden section is the archetypeof generationby the followingcorlsideration.73An approximation this proportioncan be obtained by tlzis to sequence(Fibonacci numbers):
70 Kepler, ibid.,xv, pp. 395-6, 'Itaque haec figura quindecangulum referturinter quinque fatuas virgines. Venit enim sero postquam jam januae omnes per numeros 7.9.II.I3

occlusae sunt.' 71 Kepler, ibid., vi, pp. 42-45, 63-64 I 75 72Kepler, Ges. Werke, pp. 63-64. vi, 73Ibid., p. I75.

KEPLER'S CELESTIALMUSIC

243
a

c or a-b
2 3 5 8

b
2 3
2

5 8
I3
2

5 8
I3 I

etc.

satisfythe secondof the abovetwo conditions(c a b); Thesesetsof numbers they fail to satisfythe firstcondition(ac-b2) in such a way that acalternately exceedsor falls short of b2 by unity, so that as the sequenceis carriedon b2 indefinitelynearerin value to acor a(a b): approaches
masc. fem. m. f. m.
r

a(a - b) 2 3
IO

4
65 I68

4 9

Where a(a b) exceeds b2, the number is, Kepler says, masculine,where it falls shortfeminine. He then continues:74 Since such is the nature of this [golden] section, which is used for the demonstrationof the pentagon, and since God the Creator has fitted the laws of generation to that [proportion] to the genuine and by itself perfect proportionof ineffable terms [has fitted] the propagation of plants which each have their seed within themselves; and [to] the of paired proportions numbers(of which the one falling short by unity is by compensated the other exceeding[by unity]) [has fitted] the conjunction of male and female what wonderthen, if the progenyof the pentagon, the major third or 4:5 and minor third, 5:6,75 move our souls, images of God, to emotionscomparableto the businessof generation. Kepler is so fond of sexual, male-female,analogies76that we become inclined to accept them even when, as in this case, it is not obviouswhy he
74Ibid.,pp. I75-6, 'Haec cum sit natura hujus sectionis, quae ad quinquangulidemonstrationemconcurrit; cumque Creator leges generaDeus ad illam conformaverit tionis; ad genuinam quidem et seipsa sola perfectam proportionem ineSabilium terminorum, rationes plantarum seminarias, haberejussae quae semensuumin semetipsis vero binas Numerosunt singulae:adjunctas rum proportiones(quarum unius deficiens unitas alteriusexcedentecompensetur)conjunctionemmaris et foeminae:quid mirum igitur, si etiam sobolesquinquanguliTertia dura seu 4.5. et mollis 5.6. moveat animos, Dei imagines, ad affiectus, generationis negocio comparandos?' thispassageby arguing 75 Keplereontinues that the minor third (5) derives primarily, not fromthe hexagon,but fromthe decagon, and thereforeis of 'the class of five-angled figures'. He refersthe readerbaek to ch. iii of this Book(Third). This mustbe a mistake for eh. ii, where,on pp. I I5-I6, one findsthe XIII. I do not find relevant Propositio Kepler'sargumentconvineing;but I am not it sure that I have understood fully. Where it suits him, Kepler derivesthe minor third fromthe hexagon (see next note). pp. I35 (majorthirds male, 76 E.g., ibzd., minor feminine, beeause former from the irrationalpentagon,latter from the rational hexagon), 292 (cube and dodeeahedron

25 64 I 69

244

D. P. WALKER

has chosenone term for male and the other for female. In a letter of I608 to Joachim Tanckius we find a fuller treatrnentof the golden section series, though here he does not connectit with majorand minor thirds. As in the Harmonice Mundi,Kepler represents numbersthus: the
-

m.

;n
r - F
r

3
IO

fr
4

4
7 11

m.

i l11

L3
r

I.

24

25

but herehe addsthe explanation: Non putome posseclarius palp et abilius explicare, rem quamsi dicamte videreimagines mentulae, vulvae. illic hic andmoreover numbers shownin compromising the are positions:
. , _

L__ _

| 8

rJ

}tn

[E3

masculine, octohedron and icosahedron feminine, becauselatterinscribable former; in tetrahedronandrogynous,because inscribable in itself), 326-7 (Earth male, Venus female) On p. I 76 is a truly remarkable . analogy, which shows what a sensitive musician Kepler was. At a time when no other writer on music even remotely approachesthe concept of a leading note, he clearly expressesthe feeling of expectation produced by semitones. The major third, says Kepler, is masculine because, when singinge.g. C D E, one feelsan urge to overcome the semitoneand reach the fourth,F; this third, then, is 'active and full of effiorts' ('actuosaet conatuumplena'), its generative

power ('vis yowuos')is striving to reach the fourth, and when the semitone is sung and the fourth reached this is like an ejaculation ('quaerens Enem suum, scilicet Diatessaron, cujus semitonium est ei [sc. tertiae durae] quasi onXt)sCs, conatu quaesita'). The toto minor third is feminine because, when singing e.g. D E F, one feels a tendency to sink back a semitone on to E; this third therefore is passive, and is always sinking to the ground, like a hen ready to be mounted by a cock ('semper se, veluti gallina, sternit humi, promptam insessori gallo'). This is not of course identical with the modern concept of a leading-note, since Kepler is thinking here in purely melodic terms.

KEPLER'S CELESTIALMUSIC

245

This letter,77as Casparpoints out, is importantfor an understanding of Kepler'sattitude to analogiesor symbols. He had been sent by Tanckiusa workon the monochord AndreasReinhard(Monochordum, by Leipzig I 604),78 which contained some sexual analogies. Having commented playfully on these, Keplergoes on to say that 'by this titillation'Reinhardhas excitedhim to vie with him in finding symbolsof male and female; and then, after the long passageon the golden section, 'which the lecherousfeelingsroused by Reinhard'sspeculations had forcedout of him',79he states:80 I too play with symbols, and have planned a little work, Geometric Cabala, which is about the Ideas of natural things in geometry; but I play in such a way that I do not forgetthat I am playing. For nothingis proved by symbols, nothing hidden is discoveredin natural philosophy through geometricsymbols; things already known are merely fitted [to them]; unless by sure reasonsit can be demonstrated that they are not merelysymbolicbut are descriptions the ways in which the two things of [i.e. the two termsof the analogy] are connectedand of the causesof this
connexion.

He gives as an exampleof a geometricanalogywhich is also a causalexplanation his theory of the weather. Bad weather accompaniescertain planetary aspects because there is a Soul of the Earth or Archeus Subterraneus which is capableof perceiving geometricrelationships, thiscase,the anglesformedby in planetaryrays meeting on the Earth, and is thereby excited to expel 'subterraneanhumours'. Thus 'the geometryof the aspectsbecomesan objective cause',8lwhereasit would be uselessto rely on such 'symbolisations' that as Saturnbringssnow, Mars thunder,Jupiter rain, etc. When in this letter Kepler is expoundingthe analogybetweenthe golden sectionand generation,he may be only 'playingand not forgettingthat he is playing'. But I think there is little doubt that by the time of the Harmonice Mundihe is convincedthat this analogy shows also a real causal connexion. Such a deeply pious man would not write in jest that 'creatorDeus' has fitted the modes of vegetable and animal generationto the archetypicalfigure of the pentagon. Polyphonicmusic,with its thirdsand sixths,excitesand moves us deeply as does sexual intercoursebecause God has modelled both on the same geometric archetype. There are also, as we have seen and shall see, other archetypicalcauses of the emotive power of music: the connexions betweenour music and the celestialharmonies. I shall not here give a general descriptionof Kepler's celestial music,
Kepler, Ges.Werke, pp. I54ff. xvi, I have not been able to see this work,as the BritishMuseumcopy has been destroyed. 79 Kepler,ibid., p. I58, 'Atquehic excursus esto, quem mihi extorsitprurigoa Reinhardi speculationibus concitata'. 80 Ibid., 'Ludo quippe et ego Symbolis,et opusculuminstitui, Cabalam Geometricam, quae est de Ideis rerumNaturaliumin Geometria: sed ita ludo, ut me ludere non
77 78

obliviscar. Nihil enim probatur Symbolis, nihil abstrusi eruiturin Naturaliphilosophia, per Symbolasgeometricas, tantumante nota accommodantur: certis rationibusevinnisi catur, non tantum esse Symbolicased esse descriptosconnexionisrei utriusquemodos et causa.' 8l Ibid., 'geometria aspectuum fit causa objectiva'.

oc

Was

246

D. P. WALKER

since this has alreadybeen done by severalmodernscholars.82I wish merely to discussa few problemsraisedby it and aspectsof it which have not, I think, been dealt with before. There are some difficulties Kepler'splanetarychordsthat are due, at in least in part, to his use of the musical terms durus mollis.In Book III and he describes two genera,molle durum, such a way that they seem to the and in be the same as our minor and major modes or tonalities. We are therefore disconcertedwhen we come to the planetaryharmoniesin Book V to find that the two chordsof all six planets'generisduri'consistof an E minor3 and a C major-4, and those 'generismollis'of an E flat major3- and a C minorw; the two chords, durum molle, five planets (Venus omitted), on the and for other hand, are what one would expect: a G major 3 and a G minor 3 83 Harmoniae Planetarum Quinque Harmoniae Planetarum Omnium Generis Generis Generis Mollis Duri Mollis Generis Duri

I think that here Kepler is using the terms in their originalsense simply to mean respectively scale or chordwhich containsB sharp (durum) which any or containsB flat (molle); is not surprising, this since the theoreticaldistinction of our majorand minormodeswas onlyjust beginningto emergeat the time he was writing. The inclusion of a W chord is odder, since in practice this
82W. Harburger, ohannes Seplers Kos- to the Harm.Mund.(Kepler, Ges. Werke, vi, mische Harmonie, Leipzig I925; A. Koyre,La pp. 46Ifffl. Revolution astronomique (hereafterRev. astr.), 83 Kepler, Ges. Werke, pp. 325-7. vi, Paris I 96I, pp. 328-45; Caspar's Jfachbericht

KEPLER'S CELESTIALMUSIC

247

chord was treated as a dissonanceand was banned by most theorists-but not by all. Zarlino,thoughof courseawareof currentpractice,givesa defence ofthe fourthas a consonance,84 Keplerhad read Zarlino.85 In his earlier and attemptsto find harmoniesin the orbital speedsof planets Kepler also gives w chords. In one letter he notes that modern musiciansmay object to the fourthinsteadof the fifth being at the bottomof the chord, and says that he has answersto this objection; but he does not unfortunatelygive them.86 Since he relied so much on the judgement of his ear, he may well just have observedthat- chordsare nearlyas sweetas 3 ones, and considerably sweeter than 3 ones. It is indeed a still not fully explainedmysteryof musicalhistory how the 4 came to be treatedas a dissonance. In any case, we are not justifiedin expectingthe celestialharmoniesto be exactly the same as our music. As I have alreadymentioned,Kepler makes it clear that our musicis not an imitationof celestialmusic; their relationship is that of two independentproductsof the same geometricarchetypes. Since Kepler does not stressthe differences betweenthe two-quite naturally,since he is interested theirsimilarities it may maketheirrelationship in clearer,if I point out two of the most importantof these differences. Each planet has its own scale, determined by its extreme speeds (at aphelion and perihelion).87 Saturn and Mercury, for example, have respectively :88 4 octaves higher
W

ol

1(;

,, oDo,,o<?

>u0OS

But, as Kepler points out, their passagefrom the lowest to the highest note and back again is not really articulatedinto stepsof tone and semitone,as in a musicalscale, but represents continuousaccelerationand decelerationof a the planet's speed, so that, if they actually emitted sounds (which they do
author'sviews, the first of which must be unique in Western music in that its final 302-4). Zarlino,in recommending use of chordis a major-4. the major 6 chords,relies chiefly on the judge85 Vide supra, 230 n. I2. p. ment of the ear. He cites an earlierwork in 86 Kepler, Ges. Werke, xiv, p. 52; cf. ibid., defenceof the fourth: AndreasPapinus,De p. 27. Consonantiis, pro Diatessaron seu Libri Duo, 87 Kepler,Ges.Werke, p. 322. vi, Antwerp I58I. This is a long book; it ends 88 The originalhas for Mercury's scale: with two musical examplesto illustratethe
84 Zarlino, Istitutioni Harmoniche, iii, pt. C. V, lx (Tutte l'Opere, Venice I589, pp. I86-8,

Q a

<

o l

I am assumingthat the C clef shouldbe on the bottomline, as the Frischeditiongivesit, and as Koyretranscribes (Rev.astr.,p. 339). it This emendationis necessaryfor Mercury's scale to fit into the planetarychords,and it fits Kepler'splanetaryspeedsbetter, according to which the highest note of Mercury

should be a major sixth and seven octaves above Saturn'slowestnote (see Koyre,ibid.). On the other hand, Kepler states that Mercury'sscale begins on A (Kepler, Ges. Werke, pp. 3XI-2), and gives its compass vi, as an octaveand a minorthird (ibid.,p. 3I2).

1 I

-L

248

D. P. WALKER

not), their 'scales'would sound like a siren giving an air-raidwarning. scale Mercury's thenwouldbe like

A Pt-

played with one fingeron a violin. The chordswhich two or more planets can make are determinedby the intervalscomprisedby their scales and the intervalsbetween these scales.89 SaturnandJupiter, for example,have respectively:

o ---n
b

A
O

X
, ,,

and can producemajortenthsbetween

IJminor tenthsbetween

o
>
,_<p *_

-i

J
eleventhsbetween

- ,17

Ldo

--n

and one twelfth

,

,
more But with chords of more notes the possibilitiesbecome progressively limited. Since the Earth and \;enus have very narrowranges:
89

ibid.,pp. Kepler,

3I4-I6.

:.

KEPLER'S CELESTIALMUSIC
0 s

249
ls -

and

{vv
,

ie Z

7s
]

;) 7r

iE

,.

- j

*E

respectively,they can combine with other planets in only a very restricted number of ways. It is these two planets, then, which determine the four possiblechordsof all six planets,which must all contain the majoror minor sixths:90

:;

II

These six-planet chords, already given above, can evidently occur only at very long intervalsof time. Kepler doubts whether any of them can yet have happenedtwice, and conjectures that perhapsit was only at the moment of creation that such a perfectlyharmoniouscombinationof all six planets occurred.9l One may also, I think, supposethat Kepler believed that at the Last Day the heavenswould, before their music ceased for ever, once again sound a perfectchord. In any case, here is anotherbasic differencebetween earthly and heavenly music. In the latter, the unique piece of divine music begins with a concord, passes through an immense series of dissonances, which are only finally resolved (perhaps) on the ISnalchord; whereas (at least in Kepler's time) human polyphony consistslargely of concords,and dissonances rapidlyresolved. Kinds of human musicwhich come near to are Kepler'sheavenlymusic would be: the cadenza a classicalconcerto,which of is a very long interpolationbetweena -4- chordand its 3 resolution,or a piece written entirelyon a pedal-note (e.g. one of Bach'sMusettes). When, however, Kepler himself describesthe likeness between earthly and heavenlymusic, he evidentlyassumesthat there will be more than only one or two perfectly consonant chords in the whole course of the world's history: 92
90I do not know why Kepler did not use the possibility:
,-

A [e

bs, pu.e 1-.:

imitanturistasdissonantiasnaturales) tendens in certas et praescriptasclausulas,singulas sex terminorum(velutiVocum) ijsqueNotis immensitatemTemporis insigniens et distinguens;ut mirumampliusnon sit, tandem inventam esse ab Homine, Creatoris sui

91Kepler, Ges. Werke, p. 324. vi, Kepler, Ges. Werke, p. 3{?8,'Nihil vi, igitur sunt motus coelorum,quam perennis quidamconcentus (rationalis nonvocalis)per dissonantes tensiones, veluti quasdamSyncopationes vel Cadentias (quibus homines
92
I7

250

D. P. WALKER

The motions of the heavens, therefore,are nothing else but a perennial concert (rationalnot vocal) tending, through dissonances,through as it were certain suspensions cadentialformulae93 which men imitate or (by those natural [i.e. celestial] dissonances), towards definite and prescribedcadences93, each chord being of six terms (as of six voices), and by these marks [sc. the cadences] distinguishingand articulating the immensityof time; so that it is no longera marvelthat at last this way of singing in several parts, unknown to the ancients, should have been inventedby Man, the Ape of his Creator;that, namely, he should,by the artificialsymphonyof several voices, play out, in a brief portion of an hour, the perpetuityof the whole durationof the world, and should to some degree taste of God the Creator'ssatisfactionin His own works, with a most intenselysweet pleasuregained from this Music that imitates God. Once again, I wish to emphasize that this comparisonis not only a metaphor; through the geometricalarchetypesthere is a real causal conrlexionbetween the two polyphonies,a connexion which accountsfor their likeness. By this causal analogybetween human music and planetarymovements,and betweenmusicand sexualdesire,Keplergives to musica meaning and value that had not previouslybeen attributedto it, a meaning which only polyphonicmusic, unknownto the ancients,could possiblyhave. The marvellouseffects of music, emotional, moral and religious, are of course familiar enough; but in that tradition it was always music together with wordsthat producedthe Ceffects', it was alwaysthe wordsthat bore the and specific meaning, that determinedthe particulareffect. That music alone could have a precise and profoundmeaning,was, I think, in Kepler's time an entirelynovel idea. It is an idea that we have all come to accept, and, although we may find Kepler's explanationof it unconvincing,we cannot claim to have found a better one.

Simia, rationem canendi per concentum, ignotam veteribus; ut scilicet totius Temporis mundani perpetuitatem in brevi aliqua Horae parte, per artificiosam plurium vocum symphoniam luderet, Deique Opificis complacentiam in operibus suis, suavissimo sensu voluptatis, ex hac Dei imitatrice Musica perceptae, quadamtenus degustaret.' 93 Kepler, like Calvisius, uses the term

'clausula' for cadence. The only time he uses the term 'cadentia' (ibid., p. I82) iS when discussing suspensions; he suggests that the word 'cadentia' derives from the fact that the dissonant suspended note falls to its resolution. In this passage, therefore, I believe that, in coupling 'cadentiae' with suspensions, Kepler was thinking of the regular r35 SUSpensions at perfect cadences.