State Library Victoria > La Trobe Journal

No 81 Autumn 2008


Charles Burnett
Why Study Ptolemy's Almagest? The evidence of MS Melbourne, State Library of Victoria,
Sinclair 2241

Intelligens est qui semper linguam suam refrenat nisi ad hoc ut de Deo loquatur He ys wise that settithe his tunge to speke of God
Ptolemy of Alexandria2


Ptolemy's Almagest, written in Alexandria a little after 150 AD, is rightly regarded as Classical Antiquity's greatest monument of mathematical astronomy. It remained the key work for understanding and measuring the movements of the heavenly bodies in the Western and Islamic world until the early modern period. Yet, in the first chapter of the work, Ptolemy emphasises that knowledge of astronomy is not the ultimate aim of the work. Rather, the teaching and study of the exact science of mathematical astronomy has a moral purpose. This is evident throughout the first chapter of the first book of the Almagest (the 'author's preface'), in which we find phrases such as the following:
Hence we thought it fitting to guide our actions…in such a way as never to forget, even in ordinary affairs, to strive for a noble and disciplined disposition, but to devote most of our time to intellectual matters in order to teach theories which are so many and beautiful, and especially those to which the epithet 'mathematical' is particularly applied… We were drawn to the investigation of the mathematical part of theoretical philosophy… and especially to the theory concerning divine and heavenly things. For that alone is devoted to the investigation of the eternally unchanging.…With regard to virtuous conduct in actions and character, this science (i.e. mathematics) above all things could make men see clearly; from the constancy, order, symmetry and calm which are associated with the divine, it makes its followers lovers of this divine beauty, accustoming them and reforming their natures, as it were, to a similar spiritual state.3
Knowledge of the regular movements of the heavenly bodies, which are the most lofty beings in the universe, is the highest kind of knowledge that a human being can attain, and can also give us an inkling of the nature of transcendental being. Ptolemy does not refer to a personal God in His heaven or even a panoply of Classical gods, but rather follows the Greek philosophical tradition in conceiving of the deity as the unmoved mover, and divine science as metaphysics. He departs from other Greek philosophers in regarding mathematics, rather than metaphysics, as the highest form of study for a philosopher. As Liba Taub has remarked, 'by studying and teaching mathematics, Ptolemy considered himself to be working toward the highest goal of the ancient philosophers—the achievement of the sort of immortality appropriate to man'.4
For the mathematicians and philosophers who wished to enrich the Arabic language with Greek learning the Almagest was a key text. References to several Arabic translations of the text testify to the interest in this work. Manuscripts of two of these translations survive, one made by al-Hajjaj ibn Yusuf ibn Matar in A.D. 827-28, a second made by Ishaq ibn Hunayn (ca. AD 880) and revised by Thabit ibn Qurra (d. AD 901). When Latin scholars wished to restore the science of astronomy in the West in the mid twelfth century, the Almagest, again, was their goal. Robert of Ketton, Hermann of Carinthia and Gerard of Cremona sought out Arabic versions of the text, while the translator of the Greek version, working in Sicily in about 1165 A.D., made great efforts to get hold of a copy that he had heard had been brought to Palermo from Constantinople.5 The lure of the Almagest was great indeed! And, when it was found, it was on the first chapter, in which its lofty purpose is described, that translators and interpreters lavished their greatest attention.
Manuscript RARESF 091 P95A (Sinclair 224) in the State Library of Victoria, Melbourne, gives a particularly striking example of this interest. This manuscript is a copy of the Almagest written by two contemporary hands in the thirteenth century (probably early in that century) in the north of Italy. It belonged to the library of the Convent of San Marco, Florence, a library famous for the quality of its mathematical manuscripts.6 The manuscript bears a similarity in layout, in the quality of its text, and in the appearance of its script to the best-known early manuscripts of Gerard of Cremona's mathematical translations—Paris, BNF, lat. 9335 and Vatican City, BAV, Ross. lat. 579—and it is tempting to believe that all three manuscripts are close to Gerard's own manuscripts allegedly taken back to Cremona, Gerard's birthplace, after his death in 1187.7 The manuscript originally consisted of the Almagest alone, written with great attention both to accuracy and to appearance. Each of the thirteen books of the Almagest starts with two illuminated initials—one at the beginning of the list of chapters and one at the beginning of the text itself (see Fig. 1).8
The text of the Almagest in the Melbourne manuscript belongs to Gerard's revised version (called by Paul Kunitzsch, class B).9 Gerard's translation combines the two extant Arabic translations, bks I-IX following al-Hajjaj's version, bks X-XIII, Ishaq/Thabit's version. One of the characteristics of Gerard's revision is that it includes many additions from the second Arabic version of the text. Thus when the main text is that of al-Hajjaj, translations of the equivalent passages from Ishaq's translation appear in the margin, and vice versa. These alternative translations were probably already present in the Arabic manuscript Gerard was using, since we find them in a Tunis manuscript of the Arabic Almagest written in 1085.10 In the case of the first chapter of the book, the whole chapter appears in both the al-Hajjaj and the Ishaq translation in several manuscripts of Gerard's revision. In the Melbourne manuscript, we have in addition a second version of the al-Hajjaj translation and the beginning of a lemmatised commentary on the first chapter, and a detailed commentary on the same chapter added in the margin.

Fig. 1. State Library of Victoria, MSRARE 091 P95A (Sinclair 224), fol. 33v

The versions of the first chapter in the Melbourne manuscript and their accompanying commentaries are as follows:
  • 1)
    Fols 1v-2r. The translation of al-Hajjaj's chapter 1: (Fig. 2)
    Bonum, domine [hoc dicit lectori],11 fuit quod sapientibus non deviantibus visum est, cum partem speculationis a parte operationis diviserunt, que sunt due sapientie partes. Licet enim contingat ut operatione sit speculatio prius, inter eas tamen non parva existit differentia. Non solum quia etsi quorumdam excellentium accidentium anime bonitatem possibile sit pluribus hominum inesse absque doctrina non tamen totius scientiam absque doctrina comprehendere esse possibile…(= V <ulgata>)12
  • 2)
    Interlinear glosses written by the same hand, explaining individual words or phrases. Those for the above-quoted section are: non deviantibus] i.e. recte docentibus; partem speculationis] i.e. theoricam; parte operationis] i.e. pratica; sit speculatio prius] i.e. in quantum est scientia.
  • 3)
    Marginal glosses written in the same hand. The first gloss refers to this opening section: 'Hic notandum est in omni scientia theoricam esse et praticam et est utraque scientia, sed differunt in hoc quod theorica est illa cuius finis est solam scientiam inquirere et pratica cuius est investigare ea que in operatione existunt'.
  • 4)
    Interlinear glosses written by a later hand, but this time also giving equivalents in another source (another translation, or another manuscript, signalled by 'al.'). The glosses to the passage quoted above are as follows: domine] al. o iesure; fuit] scilicet hoc; sapientibus] scilicet antiquis phylosophis; que sunt] scilicet theorica et practica; inter eas] scilicet theoricam et practicam; differentia] scilicet alia; sapientie] i.e. scientie; partes] scilicet potest <…<; accidentium anime] id est morum al. quasdam bonas virtutes anime ut magnanimitas, liberalitas et ht. (?); pluribus] quosdam; inesse] scilicet absque; doctrina] scilicet ab alio; totius scientiam] id est theoricam; absque doctrina] scilicet ab alio, scilicet quia sunt quidam qui sine magistro intelligunt multa recte.
  • 5)
    A lemmatised commentary written by the same later hand. It begins with the lemma of the first two words of the first chapter: Bonum domine. et cet. posito prohemio translatoris hic ponitur prohemium auctoris scilicet Ptholomei… primo dat intentionem suam. Secundo dat causam sue intentionis sive utilitatem huius scientie. Tertio modum procendendi. Quarto ordinem… After indicating where each of these divisions begins, and stating that proemia sometimes precede, sometimes come after, the text, the commentator rehearses the divisions of theoretical science mentioned in this chapter. Then he states that 'huius proemii due reperiuntur translationes, scilicet Greca et Arabica… in Arabica reperitur duplex litera, una dicit 'o iesure' ut etiam dicit in principio Quadripartiti, alia dicit 'o sire'. Et utraque potest esse adloquutio vel alicuius principis vel alicuius sui discipuli ad cuius instantiam fecit hunc librum, vel potest dici, ut dicit Haly, quod dicit 'o domine' ut honoret volentes studere in hoc suo libro. Per 'sapientes non deviantes' intelligit 'philosophos naturales'. Verum alia translatio dicit

    Fig. 2. State Library of Victoria, MS RARESF 091 P95A (Sinclair 224), fol. 1v

    'sapientibus non fallacibus'. Et Avicenna vocat philosophos verificatores vel certificatores. Et etiam Averroys in suo Colliget in multis locis…'
    These excerpts (which provide only a fragment of the commentary to the section quoted above) give an idea of how detailed and scholarly the commentary is. It may have been written into the manuscript at the same time as the astronomical works added at the end. It postdates the translation of Averroes kitāb al-kulliyyāt (Colliget), made by Bonacosa in Padua in the second half of the thirteenth century, and 'Ali ibn Ridwan's Commentary on Ptolemy's Quadripartitum made by Aegidius de Thebaldis at about the same time.13

    Fig. 3 State Library of Victoria, MS RARESF 091 P95A (Sinclair 224), fol. 172v

  • 6)
    Fol. 172va-173ra. A version of al-Hajjaj's translation which is, apparently, unique to the Melbourne manuscript: Bonum, O Sure, fuit quod sapientibus non fallacibus visum est, cum partem theoricam a parte practica diviserunt, que sunt due sapientie partes. Nam etsi accidat ut sit practica theorica prius, non est tamen inter eas differentia parva. Non tantum quantum quasdam bonas virtutes anime possibile est esse in pluribus hominum sine doctrina, et non est possibilis comprehensio totius sine doctrina… (= M<elbourne>) (Fig. 3).
  • 7)
    Fol. 173ra. The beginning of a commentary written in the text space, in the same hand as no. 6 above, to this much of the translation:
    bonum sure. Hoc lectori dicit. Et est sciendum quod hoc nomen 'sure' in Greco est scriptum in principio sui per litteram sextam que est 'che' et significat in Latino 'domine' ('He addresses this to the reader. One should know that this name "Sure" in Greek is written with the sixth letter, which is che and means in Latin "Lord"'14).
    partem theoricam. Notandum est in omni scientia theoricam esse et practicam. Et est theorica pars illa in qua auctoris intentio non est nisi ad sciendum; practica vero in qua est auctoris intentio perveniendi ad operandum ('One should note that in every science there is theory and practice. The theoretical part is that in which the author's intention is nothing other than to know; the practical, in which the author's intention is arriving at putting into action.')
    Nam etsi contingat ut sit. Hoc sic est intelligendum: practica est theorica prius, id est illud ad quod intendit practica prius intelligitur secundum theoricam ('This should be understood in this way: practice is theory first, i.e., the aim of practice is first understood according to theory'15)
    Non solum quia. Prima differentia quam ponit inter theoricam et practicam non patet, secunda satis manifesta est: quia theorica acquiritur studendo in scientia et practica studendo in operatione ('The first difference which he makes between theory and practice is not clear, the second is clear enough: that theory is acquired by concentrating on the science and practice by concentrating on action') = C<ommentary>
    It is noticeable that the first two lemmata of this commentary follow the text of the version immediately preceding it (M). The third lemma is a mixture between this text and the version of al-Hajjaj's translation that comes at the beginning of the manuscript (V), the fourth lemma is entirely that of this latter version. Moreover, parts of the first two comments are included as glosses to V ('hoc lectori dicit' and 'Notandum est in omni scientia theoricam esse et practicam'). This contrasts with the later commentary (no. 5 above) whose lemmata conform entirely to the V(ulgate) text that it accompanies.
  • 8)
    Fol. 173r-v. The translation of Ishaq's chapter 1: Capitulum primum in prologo huius libri. Bonum quod fecerunt in eo quod video illi qui perscrutati sunt scientiam philosophie, Ie Kirie, in hoc quod partiti sunt partem philosophie speculativam ab activa. Sed, quoniam activa, quamvis antequam sit activa est speculativa, tamen quod inter eas de diversitate reperitur est magnum. Non propterea quod quasdam bonas virtutes animales possibile est quandoque ut sint in multis hominum sine doctrina, sed ad scientiam omnium rerum speculativarum non est possibile aliquem pervenire absque doctrina…(= I<shaq> )
Thus we can see in the Melbourne manuscript intense interest in the first chapter of the Almagest: three versions of two Arabic translations of the text; interlinear and marginal glosses which may well be contemporary with the translations of these versions; and a scholarly commentary added in the fourteenth or fifteenth century.
The most economic way to explain the origin of the multiple versions of the first chapter is to propose that the version unique to the Melbourne manuscript (M) is a first attempt at a translation of the al-Hajjaj translation, while the version that is included at the beginning of the Almagest (and became the vulgate; V) is a revision,16 which was made perhaps concurrently with an explanatory gloss (C). Possible indications of this are:
  • 1)
    M is closer to the Arabic text (A);17 V substitutes words and phrases which are better Latin and further away from the Arabic:
    • a) Yā Sūrī A, O Sure M, domine V18
    • b) al-sabab al-awwal alladhī li-l-haraka al-ūlā A; causa prima que est motui primo M; causa primi motus V
    • c) alladhī bi-hi yubhath 'an asnāf al- 'unsuriyya A; qua fit perscrutatio de speciebus materialibus M; qua species materiales investigantur V
    • d) wa-nakhuṣṣuhu bi-smi al- 'ilm A; appropriabo eam nomine scientie M; vocabo eam proprie nomine scientie V
    • e) al-jinsayn al-ākharayn alladhayn humā min qismat al-naẓar A; duo alia genera que sunt de divisione theorice M; duo reliqua genera divisionis theorice V
    • f) bi-qadr quwwatinā 'alayhi A; cum quantitate nostre potentie super eam M; quantum possumus V
    • g) wa fī dhālika A; et propter hoc M; quapropter V
    • h) shiddat bahthin A; vehementia inquisitionis M; vehementi investigatione V.19
    • i) In the passage quoted below (p. 134) M retains the noun ('adiutorium') in Arabic ('Et non est parvi adiutorii'), while V's 'Hoc quoque non parum valet' substitutes a verb ('valet').
    • j) M retains the Arabic word order where V changes it in 'nahnu sanatakallaf an nazīd ft 'ishq A: et nos quidem laborabimus ut addamus in amorem M; Nos autem laborabimus ut in amorem… addamus V.
  • 2)
    A gloss in M, paraphrasing the text,20 has become incorporated into the text in V: muntaqālāt A; motus motorum M, glossed as 'scilicet localiter'; eorum que localiter moventur V. M's gloss brings out the difference between the root 'n-q-l' ('to move from place to place') and 'h-r-k' ('to move' in general).
  • 3)
    A word in M has been written in V and then replaced: akthar farāghinā A; plurimum nostri otii M; plurimum nostri otii studii V
Both M and V appear to have used an Arabic text directly, as can be seen, for example, by V's mistake in reading wuḍiḥ (translated correctly by M as 'explanata') as wuḍi Vponitur', and by M and V taking up two meanings of bāliya:21 'worn out' = M's corruptibilibus, and 'old' = V's 'antiquis (vel vestustis)'. Both M and V, however, show knowledge of the Ishaq translation, phrases of which are given in exactly the same version as appears alongside the
vulgate Latin version of al-Hajjaj. For example, the phrase 'al-umūr allatī khuṣṣat an summiyat ta 'alīmiyya' (Ishaq) appears in the words of I as the main text of M, while the translation of al-Hajjaj appears as a gloss ('in alio'): 'proprie (in alio: appropriata nomine scientie) in rebus quibus apropriatum est ut nominentur doctrinales'. V, on the other hand, seems to give a combination of the two: 'precipue que proprie nominatur scientia'. In a more conspicuous case, M presents the Ishaq version in the text immediately after the Hajjaj version, while V adds the Hajjaj version as a marginal gloss, signalled by 'al.':
Translation of Greek Ishaq version (I) al-Hajjaj (Melbourne version: M) al-Hajjaj (Vulgate version: V)
secondly, <the subject matter of mathematics> is an attribute of all existing things without exception, both mortal and immortal: for those things which are perpetually changing in their inseparable form, it changes with them, while for eternal things which have an aethereal nature, it keeps their unchanging form unchanged. Est communis iterum omnibus corporibus que corrumpuntur ex eis et que non corrumpuntur. Ipsa igitur est in eo quod corrumpitur ita quod alteratur cum eo in forma que seperatur a materia, et in eo quod non corrumpitur, scilicet in natura celesti remanet in forma, absque alteratione. (a) in omnibus que de non esse venerunt ad esse in eis que moriuntur et in eis que non moriuntur, alterata cum alteratis in forma numquam separata, adherens formis rerum sempiternis perpetuis que sunt nature etheree absque alteratione (b) ipsa est communis in omnibus corporibus, que de eis corrumpuntur et que non corrumpuntur, et in eis quidem que corrumpuntur ipsa alteratur cum eis in forma que non separatur a materia, et in eis que non corrumpuntur, scilicet in natura celesti, remanet in forma eorum (vel sua) sine alteratione. (b) ipsa communiter consistit in omnibus corporibus que corrumpuntur et que non corrumpuntur. In eis autem que corrumpuntur existit cum alteratione forme que non separatur a materia, set in eis que non corrumpuntur, scilicet in natura celesti, remanet in forma sua absque alteratione. (a) al. existit in omnibus essentiis mortalibus et inmortalibus alterata cum alteratis in forma numquam separata, comitans formas rerum semper permanentium que sunt nature etheree absque alteratione.
The free transference of phrases from one translation to another would seem to corroborate Kunitzsch's contention that Gerard of Cremona was the author of the Latin versions of both translations. M could then be regarded as a 'first draft' of Gerard's translation, which was later polished up by the author. It is quite conceivable, however, that several hands were at work almost simultaneously in the entourage of Gerard. We know of a certain Ghalib ('Galippus') the Mozarab who helped Gerard.22 He may indeed have made a first version, whose Latin needed polishing. The variation in the degree of literalness between the versions may also indicate the involvement of more than one translator: both al-Hajjaj versions follow the Arabic less literally than does the Ishaq translation. Finally, one may point to a terminological difference which has been alleged to differentiate the translations of Gerard from his Toledan contemporary, Dominicus Gundissalinus, namely that between translating the Arabic 'aql as 'intellectus' (Gundissalinus) and as 'ratio' (Gerard).23 In M we find the Arabic ma'qūl ('understood') translated as 'rationatum' whereas in V we have the verbal equivalent 'intelligitur'.24


The Arabic Hajjaj and Ishaq/Thabit translations of Ptolemy's Almagest are totally independent of each other, and it is not surprising that Latin scholars, wishing to have as much help as possible in understanding Ptolemy's words, should turn to both translations. The revised version of Gerard of Cremona's translation, as we have seen, includes many passages in both the Hajjaj and the Ishaq/Thabit version. The appearance of the second translation in the margins of the text of the first, in several manuscripts, gives the impression that one is a commentary on the other. But, when one looks at the Latin versions of the first chapter of the Almagest one notices something more. The Latin al-Hajjaj, in both its versions, diverges from the Arabic al-Hajjaj (at least, in the two manuscripts available to me), in a significant way. This can be demonstrated by taking the passage in which Ptolemy explains how useful mathematics is for theology.
The Greek text reads:
Furthermore it can work in the domains of the other <two divisions of theoretical philosophy> no less than they do. This is the best science to help theology along its way, since it is the only one that can make a good guess at <the nature of> that activity which is unmoved and separated, because of its closeness to those attributes of beings which are on the one hand perceptible, moving and being moved, but on the other hand eternal and unchanging, <I mean the attributes> having to do with motions and the arrangements of motions.25
The Ishaq version follows the Greek more closely:
wa-yumakkin ay?an an (wa-hādha al- ?ilm ay?an L) yu ?īn ?alā ?ilm al-jinsayn al-ākharayn li-annahu yu?riq ilā al- ?ilm al-ilāhī akthar min ta?rīq sā'ir (jamī? TP) al- ?ulūm wa-dhālika li-annā bi-hādhā al- ?ilm faqa? naqdir ?alā al-wu?ūl ilā ta?awwur dhālika al-fi ?l (al- ?aql TP) alladhī yakūn bi-lā ?arak wa-lā yashūb al-mādda min qibali
qurbihi mimmā yalzam al-jawāhir al-mahsūsa (L adds 'minhu') al-mu?arrika wa-l-mu?arraka allatī hiya dā'ima ghayr (lā TP) mutaghayyira min ( an TP) al-adwār wa-nizām al-harakāt.
Which may be translated:
And it is possible that this science/knowledge ( ilm) also helps the knowledge of the other two kinds, because it provides the path to the divine science more than the other (all the TP) sciences do. This is because by this science alone we are able to arrive at conceiving that action (understanding TP) which comes about without movement and does not mix with matter, because of its closeness to those revolutions and the order of the movements that accompany sensible substances that move and are moved, which are everlasting and without change.
This is translated as:
Et possibile est ut adiuvet super scientiam duorum generum aliorum, quoniam perducit ad scientiam divinam plus quam perducant omnes scientie. Et illud est quoniam nos per hanc scientiam tantum possumus pervenire ad hoc ut esse habeat<ur> illa intelligentia26 cuius esse est sine motu et non cum cursu27 materie, propter propinquitatem suam ad illud quod sequitur substantias sensatas, motas et moventes, que assidue non alterantur, a revolutionibus et ordinibus motuum.
This may be compared with the Hajjaj version:
wa-ammā ft dark al-qismayn al-ākharayn fa-laysa ?awnuhu fīhimā bi-dūnin ammā fī al-jins al-ilāhī fa-huwa al-mu?riq al-sā'iq ilayhi li-annahu wa?dahu faqa? min jins (faqa? yumkin min ?usn qiyās Le) mā lā yataghayyar wa-?azruhu bi-lā ?amal yuqarrib (taqrīb L) al-a ?rā? allatī fī al-adwār wa-marātib al-?arakāt allawātī li-l-jawāhir al-?issiyya al-muharrakāt wa-l-muta?arrikāt al-abadiyya allatī laysa fīhā khilāfun (ikhtilāfun Le).
which may be translated:
In the comprehension of the other two divisions its help is not small. As for the divine kind, it is providing the path and driving one to it, because it alone is of the kind28 (alone makes it possible to judge that Le) which does not change, and appraising it without action brings close (i.e. makes understandable) those attributes in the revolutions and the ranks of the movements which belong to sensory substances, moving and being moved, eternal, in which there is no difference.
The Melbourne version (M) gives:
Et non est parvi adiutorii ad comprehendendam scientiam aliorum duorum modorum, et precipue scientiam Dei gloriosi, quoniam ipsa est semita ad cognoscendum Deum gloriosum propter rationem cum perscrutatione (vel perexperientia) et cogitatione. Et hec et ei similia significant significatione vera et manifesta Deum, qui non alteratur et non separatur et non est novus neque factus, quoniam ipsa est propinquior scientie nostre quam Ipse gloriosus et quia ipsa est viator perducens ad Eum. Quoniam ipsa sola tantum cum perseverantia (vel 'per perseverantiam') inquisitionis de sempiternis fixis et de genere considerationis eius quod non alteratur, et cuius estimatio est absque opere, approximat accidentibus que sunt in revolutionibus et ordinibus motuum qui sunt sub iis sensatis moventibus et motis, sempiternis in quibus non est diversitas.
The vulgate Latin version (V) gives:
Hoc quoque non parum valet ad reliquorum duorum modorum scientie29 comprehensionem, et precipue scientie Dei excelsi. Ipsa namque est via30 ad sciendum Deum altissimum propter rationem cum perscrutatione et intellectu. Que et eius similia vere et manifeste significant Deum, qui non alteratur et non movetur et non est novus31 neque est factus, quoniam ipsa nostre scientie quam Ipse altissimus vicinior existit32 et quia ipsa est semita ducens ad Eum. Ipsa namque sola tantum de rebus semper manentibus cum perseveranti inquisitione33 et de rebus que sunt et34 ex genere considerationis eius quod non alteratur, et extimatione ipsius35 absque opere vicinatur accidentibus que sunt in revolutionibus et ordinibus motuum qui sunt in substantiis sensatis, 36 moventibus et motis, sempiternis, in quibus non existit diversitas.
It is immediately obvious that M and V are not translating exactly what is in the two Arabic manuscripts of the al-Hajjaj version. An English translation, based on these two Latin versions, would run as follows:
Its help is not small in the comprehension of the other two kinds, especially the knowledge of God, the Glorious. For it is the path to knowing God, the Glorious, because of <its> rationality, together with careful investigation and thought. This <science> and those similar to it truly and clearly indicate God, who does not change and is not divided into parts (does not die)37 and is not innovated or made, since this <science> is closer to our knowledge than is the Glorious One,38 and because it is the path (traveller) leading to Him. Since it alone <is> accompanied by (or 'occurs through') persistence in <things> investigation concerning eternal, fixed <things> and is of the kind of consideration of that which does not alter, and whose appraisal <is> without practice (i.e. only theoretical); it comes close to those attributes in the revolutions and the ranks of the movements which are in sensory substances, moving and being moved, eternal, in which there is no difference.
One can isolate in the Latin versions phrases that appear to be two attempts at translating single phrases in Arabic:
quoniam ipsa est semita ad cognoscendum Deum M/ Ipsa namque est via (semita B) ad sciendum Deum V
quia ipsa est viator perducens ad Eum M / quia ipsa est semita ducens ad Eum V
Both of these are equivalent to the Arabic 'it is providing the path and driving one to it'
  • 1) cum perseverantia inquisitionis de sempiternis fixis M / de rebus semper manentibus cum perseveranti inquisitione (cum perseveranter inquirit B) V
  • 2) de genere considerationis eius quod non alteratur M / de rebus que sunt (et est B) ex genere considerationis eius quod non alteratur V
Both of these are equivalent to the Arabic 'of the kind which does not change' (perhaps originally: 'of the kind of judgement that does not change').
These repetitions suggest that we are dealing with glosses which have been incorporated into the text. In the first case 'because it is the way (guide) leading to Him' is glossed 'since it is the path to knowing God… since this (science) is closer to our knowledge than the Glorious One'. In the second case 'of the kind which does not change' is glossed
'persistence in investigation concerning eternal fixed (things)'.
I am inclined to think that these glosses would have been in the Arabic text, rather than in an earlier Latin manuscript. The adjective 'altissimus' sounds as if it is a translation of Arabic taālā. The variation in terminology between the two translations suggests that the extra phrases (the 'glosses') were also in the Arabic. The variant between 'intellectus' (V) and 'cogitatio' (M) is found in two other passages, where these are, respectively, V's and M's translations for the Arabic fikr. The repetition of the root 'significant significatione vera' in M (simplified in V to 'vere…significant') is a characteristic Arabism.
The general literalness of the translations of M and V would also make it unlikely that the translators would have added anything substantial that was not already in the Arabic. All these considerations lead me to believe that we are dealing with Latin versions of an Arabic text which has not been identified, which I shall call 'the enhanced Hajjaj version'.
What is notable about this version is that the 'divine science' (which in Ptolemy, and in its strict Arabic translation ?ilm ilāhī, is tantamount to 'metaphysics'), has become the knowledge of God. Instead of putting natural science and divine science on the same level, as two sciences for which mathematics prepares the way, the translation elevates the 'knowledge of God' to the position of the major ('precipue') science for which astronomy is useful. The strictly philosophical language of Ptolemy, reflected by his Arabic translator, has been enhanced by a note of piety.
In another context God has been introduced where he doesn't occur at all in Ptolemy's text, or the Arabic translations. Where the Greek reads 'It is this love of the contemplation of the eternal and unchanging which we constantly strive to increase'39 the Latin versions of al-Hajjaj's translation have added after 'eternal and unchanging' 'up to the time that their Creator has imposed on them' ('Et nos quidem laborabimus ut addamus in amorem sempiternorum fixorum usque ad terminum quem posuit eis ipsorum Creator, in eo quod sequitur de hoc nostro libro' M; 'Nos autem laborabimus ut in amorem scientie sempiternorum, manentium usque ad terminum quem eorum Conditor eis imposuit, in sequentibus huius nostri libri addamus' V)
This added phrase does not appear in Arabic in either the Ishaq or the Hajjaj translation. It does, however, occur in Abu Ma ?shar's Great Introduction to Astrology in the Arabic form ?ilā al-waqt alladhī yashā' Allāh' in a passage reminiscent of Ptolemy's argument: 'Something else from which one may infer the nobility of the profession of astrology is that it is a lofty profession and its subject is the stars which do not alter and are not subject to coming-to-be and passing-away, for as long as God wills'.40 Once again, then, we have in the Latin versions of al-Hajjaj the addition of a pious phrase, which would seem to have come from the Arabic original.
So, in the Latin versions of al-Hajjaj's translation we find not only that studying astronomy makes one a better character, but also that it leads to the aim of every good Muslim (and, indeed, every good Christian and Jew): to know God more fully, and especially
to acknowledge that God's power is boundless. A tone of piety that is completely lacking in the philosophical language of Ptolemy's original text has been added. As a consequence of this, Ptolemy himself, in the Latin tradition, becomes both a sage and a pious man.


This tone of piety is already apparent from the manuscripts of Gerard of Cremona's translation, which include a preface before the beginning of Ptolemy's text. The last part of this preface gives, as one might expect, an account of the circumstances of the translation of the Greek text from Arabic by al-Hajjaj and Sergius and the title and author of the work. But before this is added something which is found in no other translation from Arabic into Latin made by Gerard of Cremona: a description of Ptolemy's character, a set of moral maxims attributed to him, and a physiognomy, all of which have been lifted from Abu-l-Wafa' al-Mubashshir ibn Fatik's Mukhtar al-Hikam. This is a popular book of sayings attributed to the ancient sages which was written in 1053 AD and translated into Castilian under the title Los Bocados de Oro in the mid-thirteenth century. This version was the source of translations into several other European languages, until it became the first English text to leave Caxton's press in the late fifteenth century.41 It is likely, but not certain, that Gerard himself added this material from Abu-l-Wafa', whose translation here is different from all the other extant translations.42 In this preface one can read not only how Ptolemy was a wise man and a supreme mathematician, but how he taught that an intelligent man must feel ashamed before God when his thoughts are not obedient to His will, that he must shackle43 his tongue, except to allow it to praise God the exalted; the ignorant man is the one who does not know his own destiny (the first three of his maxims).44
The effect of this religiously enhanced version of al-Hajjaj's translation can be seen in the reception of Ptolemy in the Latin West. In 1271 Robertus Anglicus completed for his students a commentary on the basic elementary textbook on cosmology used throughout the Middle Ages, the Sphere of John of Sacrobosco. As the culmination of the praise of astronomy which constitutes the first paragraph, Robertus writes 'Et ideo dicit Ptholomeus in principio Almagesti quod ista scientia est quasi semita ducens ad Deum.'45 This is a direct quotation from the Vulgate version of the Hajjaj chapter (with 'Deum' making the referent of 'Eum' explicit).
A fifteenth-century manuscript includes an anonymous 'Recommendatio astronomiae' or 'Defense of Astrology', which begins with, precisely, the statement that astronomy is the science that leads to the knowledge of God:
Astronomia "est sciencia ad sciendum Deum altissimum". Hanc propositionem scribit Ptholomeus vir sapiens, princeps astronomorum, libro suo magno capitulo primo "in quo huius sciencie ad alias excellenciam" ponit.46
Later in the same Recommendatio the author cites two other passages from the texts discussed in this article: 'Hec quoque non parum valet ad reliquorum duorum modorum comprehensionem, et precipue sciencie Dei excelsi' and 'Laborabimus in amore sciencie et
sempiternorum et permanencium usque ad terminum quem Conditor eis imposuit'.47 These are almost the only quotations from the Almagest in the text,48 and yet, as we have seen, they hardly represent the words of Ptolemy himself. Rather, they reflect the adaptation of the purpose of the Almagest to the monotheistic religion to which both an anonymous Arabic editor of the al-Hajjaj translation and the scribe of the Melbourne subscribed. Studying Ptolemy's Almagest did not just make you a virtuous philosopher, it also made you a good Muslim or Christian.


I am very grateful to John Crossley for introducing me to State Library of Victoria, MS 224, photographing folios of it, and accommodating me in Melbourne, to Paul Kunitzsch for sending me copies of texts, helping me to interpret the Arabic, and correcting several mistakes and to Anna Akasoy for advice on Arabic.


From the portion of Abu-l-Wafa' al-Mubashshir ibn Fatik's Mukhtar al-Hikam accompanying Gerard of Cremona's translation of Ptolemy's Almagest (British Library, Burney 275, fol. 390v), with the translation of Stephen Scrope, The Dicts and Sayings of the Philosophers, ed. C. F. Bühler, London: Oxford University Press, 1941, p. 224.


Ptolemy's Almagest, translated and annotated by G. J. Toomer, London, Duckworth, 1984, pp. 35-7.


Ptolemy's attitude is described and analysed in detail in Liba C. Taub, Ptolemy's Universe, Chicago and LaSalle, Illinois: Open Court, 1993, pp. 19-37.


For Hermann of Carinthia and Robert of Ketton see Charles H. Haskins, Studies in the History of Mediaeval Science, 2nd edition, Cambridge, Mass.: Harvard University Press, 1927, p. 121; for Gerard of Cremona, C. Burnett, 'The Coherence of the Arabic-Latin Translation Program in Toledo in the Twelfth Century', Science in Context, 14 (2001), pp. 249-88 (p. 275); for the anonymous Sicilian translator's eagerness to get hold of a copy of the Almagest, see Haskins, Studies, pp. 191-3.


Axel A. Björnbo, Die mathematischen S. Marcohandschriften in Florenz, new edition, Domus Galilaeana, Quaderni di storia e critica della scienza, nuova serie, 8, Pisa, 1976, describes two other manuscripts of the Almagest in the monastery's collection (nos 177 and 182) but does not mention this manuscript (no. 179 in the monastery's collection). For a detailed description see Keith V. Sinclair, Descriptive Catalogue of Medieval and Renaissance Western Manuscripts in Australia, Sydney: Sydney University Press, 1969, pp. 382-386, which largely coincides with idem, 'An Unnoticed Astronomical and Astrological Manuscript', Isis, 54, 1963, pp. 396-9.


For the accounts of the return of Gerard's books (and even his body) to Cremona see Marika Leino and Charles Burnett, 'Myth and Astronomy in the Frescoes at Sant'Abbondio in Cremona', Journal of the Warburg and Courtauld Institutes, 66 (2003), pp. 273-88. It is significant that of these ten frescoes, painted in 1513, one depicts Ptolemy, another Thabit ibn Qurra, the reviser of the Arabic text of the Almagest (though this may not have been known in the Latin West), and two depict authorities mentioned in the Almagest (Hipparchus and Timocharis).


In the fifteenth century further astronomical and astrological texts were added, but these do not concern us here.


Paul Kunitzsch, Der Sternkatalog des Almagest: Die arabisch-mittelalterliche Tradition, 3 vols, Wiesbaden: Harrassowitz, 1986-91, vol. II, p. 17.


Paul Kunitzsch, 'The Role of Al-Andalus in the Transmission of Ptolemy's Planisphaerium and Almagest', Zeitschrift für Geschichte der arabisch-islamischen Wissenschaften, 10, 1995-6, pp. 147-55 (see pp. 149-50).


These three words have been signalled as being redundant by being marked with the word 'vacat'.


For comparison one may quote Toomer's translation of the opening of the first chapter (p. 35 = ed. Heiberg, p. 4, lines 7-13): 'The true philosophers, Syrus, were, I think, quite right to distinguish the theoretical part of philosophy from the practical. For even if practical philosophy, before it is practical, turns out to be theoretical, nevertheless one can see that there is a great difference between the two: in the first place, it is possible for many people to possess some of the moral virtues even without being taught, whereas it is impossible to achieve theoretical understanding of the universe without instruction…'


In addition to the Commentary on the Quadripartitum the commentator mentions a commentary on 'this book' (i.e the Almagest) by 'Haly' (i.e. 'Ali ibn Ridwan), which is not otherwise known in Latin (and doesn't seem to be known in Arabic): 'Et in Haly in commento huius libri ultimo exponit hoc prohemium'.


This note, which can also be found in MSS Florence, Laur. 89, 45 and Vienna, ÖNB, cod. 4799 (Paul Kunitzsch, Der Almagest: Die Syntaxis Mathematica des Claudius Ptolemäus in arabisch-lateinischer Überlieferung, Wiesbaden: Harrassowitz, 1974, p. 133) appears to be confused, but interprets the Arabic 'sure' (a close transliteration of the Greek 'syre') as 'kurie' (= Lord), which, unaccountably, the annotator considers to start with the 'sixth letter (of the Greek alphabet?)'.


This is a good interpretation of an obscure passage in the original, and agrees with that which Toomer arrives at (see n. 12 above), by reference to Theon's commentary.


The two versions do not coincide with the Gerard's original and revised versions of the Almagest, classified by Kunitzsch as Class A and B respectively (Der Sternkatalog, n. 9 above, pp. 1-7). The less significant variants in version V, between Class A and B, are always noted in the quotations below.


For the Arabic text of al-Hajjaj's translation, British Library, Oriental MSS, add. 7474 (= L) has been used, checked against Leiden, cod.or. 680 (= Le); for Ishaq/Thabit's translation, the marginalia of the British Library manuscript (= L), and MSS Tunis, National Library, 07116 (= T) and Paris, BNF, ar. 2482 (= P); for the Latin of all the versions, Melbourne, Victoria State Library, MS 224 (= M); for that of the Ishaq version, Wolfenbüttel, Gud. lat. 147 (= W), for that of the Hajjaj version V, British Library, Burney 275 (= B). The Melbourne and Wolfenbüttel manuscripts belong to Kunitzsch's Class B, the Burney manuscript to Class A.


MS B gives 'fuit scire' for 'domine fuit'.


On the other hand, in the phrase 'alladhī yub?athu bi-hi ?an ?alab al- ?ilm bi-hi' Vs translation ('qua inquiritur indagatio sciendi de Deo') is more literal than M's 'qua Ipse scitur'.


This gloss is contemporary with the text, as are two further glosses: 'perscrutatione] vel perexperientia' (sic) and cum percrutatione] per perscrutationem (see p. 136 below). Two interlinear glosses at the beginning of the text of M (differentia] scilicet alia; totius] i.e. theorica) simply reproduce glosses to the first al-Hajjaj version (no. 4 above) and appear to be in the same, later, hand.


The London MS has 'kā'ina' in place of the Leiden MS's 'bāliya'.


'Galippus mixtarabs' is referred to by Daniel of Morley in his Philosophia, ed. Gregor Maurach, Mittellateinisches Jahrbuch, 14, 1979, pp. 204-55 (at pp. 244-5).


This terminological difference has been pointed out by Manuel Alonso Alonso in Gundissalinus' and Gerard's respective translations of al-Farabi's On the Division of the Sciences, but is apparent also in the two different versions of al-Kindi's De ratione/De intellectu: Manuel Alonso Alonso, 'Traducciones del Arcediano Domingo Gundisalvo', al-Andalus, 12, 1947, pp. 295-338. See the critical appraisal of the situation by Anna Akasoy in 'Die lateinischen Übersetzungen der Risāla fī' l-?aql al-Kindīs', in Intellect et imagination dans la Philosophie Médiévale, Actes du XIe Congrès International de Philosophie Médiévale de la Société Internationale pour l'Étude de la Philosophie Médiévale, Porto, du 26 au 31 août 2002, Turnhout, Brepols, 2006, pp. 689-701.


It may be significant that V regularly writes 'existere' where M gives 'esse' (10 occasions), V writes 'intellectus' where M gives 'cogitatio' (3 occasions) and V writes 'indicantem' where M gives 'significantem' (one occasion).


Trans. Toomer (slightly adapted), p. 36 = ed. J. L. Heiberg, p. 7, lines 4-10.


This passage is unclear in the Latin. The translator followed the reading of TP: ?aql ('intelligentia') rather than that of L: fi?l ('actio').


The Wolfenbüttel manuscript gives 'non incursu' which is a more plausible reading.


The Latin versions imply a combination between the two versions: 'min jins qiyās' ('ex generatione considerationis').


'scientie' is omitted in MS Burney 275 (= B), and is absent in the Arabic.


semita B = the text of the Melbourne version (M).


et non movetur et non est novus] et non morietur et non est accidens B


quam Ipse altissimus vicinior existit] quam de Deo habemus altissimo vicina existit B


cum perseveranti inquisitione] cum perseveranter inquirit B


et de rebus que sunt et] et est B (closer to the Arabic)


eius existimatio B


sensilibus B


'movetur/morietur' in V may be a banalization of 'separatur', which could have been more difficult to understand in a non-Islamic context.


This would seem a more accurate version than that of the Burney manuscript which would translate 'since it is close to the knowledge which we have concerning God, the most exalted'. The gloss in V makes the meaning clear: 'he says this because he cannot comprehend these things truly'.


Ed. Heiberg, p. 7, line 26-p. 8, line 1, trans. Toomer, p. 37.


Abu Ma 'shar's Great Introduction to Astrology, Book 1, chapter 2, ed. R. Lemay, 9 vols, Naples, 1995-6, Istituto universitario orientale, vol. 2, p. 14, lines 299-301. Compare the Latin translation by John of Seville, ibid., vol. 5, p. 16, lines 532-6: 'Ex hoc intelligitur dignitas magisterii astrorum, quia magisterium astrorum altius est et loca eius planete que non corrumpuntur nec recipiunt augmentum neque diminutionem, effectum neque detrimentum usque ad tempus quod Deus voluerit'; cf. also Book 1, chapter 5, vol. 2, p. 30, lines 685-6, Latin translation, vol. 5, p. 34, lines 1187-90.


The Dictes and Sayings of the Philosophers printed by William Caxton in 1477.


Kunitzsch, Der Almagest (n. 14 above), pp. 98-9. The Vita of Ptolemy (without his dicta) from the Mukhtar al-Hikam was added to the Leiden manuscript of the al-Hajjaj version by its owner, Amir Katib ibn Amir 'Umar al-Amid (1286-1357).


A play on the root c-q-l which means both 'intelligence' and 'to hobble, confine'.


These are translations from the Arabic text edited by A. Badawi, Los Bocados de Oro (Mujtār al-Hikam) Abū-l-Wafa al-Mubaššir ibn Fātik, Madrid, Instituto Egipcio de Estudios Islámicos, 1958, p. 251. For the Latin translation of the second of them see the motto at the beginning of this article.


Lynn Thorndike, The Sphere of Sacrobosco and its Commentators, Chicago, University of Chicago Press, 1949, p. 143. I owe this reference to David Juste of Sydney University who, in turn, was informed of this text by his student, Richard Jones.


This text is edited and discussed in Paolo Lucentini and Antonella Sannino, 'Recommendatio astronomie: un anonimo trattato del secolo XV in difesa dell'astrologia e della magia', in Magic and the Classical Tradition, eds C. Burnett and W. F. Ryan, London and Turin: The Warburg Institute and Nino Aragno Editore, 2006, pp. 177-93 (see p. 187, lines 1-4).


Recommendatio astronomie, p. 187, lines 30-32 and p. 188, lines 62-3.


A further quotation cites the opening phrase of the first chapter: Recommendatio astronomie, p. 190, lines 157-8.