Non-transferable Knowledge: Arabic and Hebrew Onomancy into Latin

Summary

As a divinatory device based on the numerical values of names, onomancy requires a system of letter-number equivalents. In Greek and the Semitic languages, a unique system is used, which consists of ascribing the first nine letters of the alphabet to the units (1–9), the following nine letters to the tens (10–90), and the remaining letters to the hundreds (100-). Given the structural similarities between those languages, the transfer of onomancy between Greek and Semitic cultures does not pose particular problems. Difficulties arise when onomancy is translated into Latin, where there is no such system of letter-number equivalents. This article examines the various solutions put forward by the translators and authors of the earliest Latin onomantic texts of Arabic origin, the so-called Alchandreana.

Acknowledgements

I am grateful to Charles Burnett and Benno van Dalen for their comments on an earlier version of this article.

Notes

¹There is no general study of onomancy. For a first approach, albeit somewhat outdated, see Franz Dornseiff, Das Alphabet in Mystik und Magie, 2^nd edition (Leipzig-Berlin, 1925), 91–118.

² Auguste Bouché-Leclercq, Histoire de la divination dans l'Antiquité, I (Paris, 1879), 261–4; Id., L'astrologie grecque (Paris, 1899), 537–41; Paul Tannery, ‘Notice sur des fragments d'onomatomancie arithmétique’, in Notices et extraits des manuscrits de la Bibliothèque Nationale et autres bibliothèques, XXXI.2 (Paris, 1886), 231–60; Marcellin Berthelot and Charles-Emile Ruelle, Collection des anciens alchimistes grecs, I (Paris, 1887), 86–92; André-Jean Festugière, La révélation d'Hermès Trismégiste, I: L'astrologie et les sciences occultes (Paris, 1949), 125, 327 and 336–9; Otto Neugebauer and George Saliba, ‘On Greek Numerology’, Centaurus, 31 (1988), 189–206. The relevant texts are edited in the Catalogus Codicum Astrologorum Graecorum, eds. Franz Cumont, Franz Boll, a.o., 20 vols (Bruxelles, 1898–1953), I, 128; IV, 120–1; VII, 161, 191; IX.2, 138–9; X, 57, 98–100, 243; XI.2, 139–44, 148–50, 152–4, and 163–4; in Armand Delatte, Anecdota Atheniensia, I (Liège–Paris, 1927), 68 and 133-5; and in Karl Preisendanz, Papyri Graecae Magicae, II, 2^nd edition (Stuttgart, 1974), 81 (trans. Hans D. Betz, The Greek Magical Papyri in Translation, Including the Demotic Spells, I, 2^nd edition (Chicago–London, 1992), 165–6).

³ The main available source is an astrological compilation edited and translated from a twelfth-century manuscript (but whose material is much older) in E. A. Wallis Budge, The Syriac Book of Medicines. Syrian Anatomy, Pathology and Therapeutics in the Early Middle Ages, 2 vols (London, 1913), see translation, II, 524–6, 531–5, 538–41 and 620–7.

⁴ The Arabic tradition remains largely unexplored. See Edmond Doutté, Magie et religion dans l'Afrique du Nord (Alger, 1909), 380–1; J.-M. Faddegon, ‘Notice sur un petit traité d'astrologie attribué à Albumasar (Abu-Ma‘Šar)’, Journal Asiatique, 213 (1928), 150–8: 152–3; Franz Rosenthal, Ibn Khaldûn: The Muqaddimah. An Introduction to History, I (New York, 1958), 234–8; Toufic Fahd, La divination arabe. Etudes religieuses, sociologiques et folkloriques sur le milieu natif de l'Islam (Leiden, 1966), 217 n. 1 and passim between 219–45; Mario Grignaschi, ‘L'origine et les métamorphoses du Sirr-al-Asrār’, Archives d'Histoire Doctrinale et Littéraire du Moyen Age, 43 (1976), 7–112: 10 and 93–7; Emilie Savage-Smith, ‘Introduction’, in Magic and Divination in Early Islam, edited by E. Savage-Smith (Aldershot, 2004), xxxv–xxxvi.

⁵ Reimund Leicht, Astrologumena Judaica. Untersuchungen zur Geschichte der astrologischen Literatur der Juden (Tübingen, 2006), 142, 151, 156, 157, 159, 160 and 165. See also Moses Gaster, ‘The Hebrew Version of the Secretum Secretorum’, Journal of the Royal Asiatic Society (1907–8) (reprinted in Id., Studies and Texts in Floklore, Magic, Mediaeval Romance, Hebrew Apocrypha and Samaritan Archaeology, II (London, 1925–1928), 742–813: 797–9).

⁶ Marcel Griaule, ‘Notes sur l'arithmomancie éthiopienne’, Journal de la Société des Africanistes, 4 (1934), 25–31; Otto Neugebauer, Ethiopic Astronomy and Computus (Wien, 1979), 162–3.

⁷ Ernest Wickersheimer, ‘Figures médico-astrologiques des 9^e, 10^e et 11^e siècles’, Janus, 19 (1914), 157–77: 164–74; Henry E. Sigerist, ‘The Sphere of Life and Death in Early Mediaeval Manuscripts’, Bulletin of the History of Medicine, 11 (1942), 292–303; Thomas G. Tolles, ‘The Latin Tradition of the Epistola Petosiridis’, Manuscripta, 26 (1982), 50–60; Linda E. Voigts, ‘The Latin Verse and Middle English Prose Texts on the Sphere of Life and Death in Harley 3719’, Chaucer Review, 21 (1986), 291–304; Charles Burnett, ‘The Eadwine Psalter and the Western Tradition of the Onomancy in Pseudo-Aristotle's Secret of Secrets’, Archives d'Histoire Doctrinale et Littéraire du Moyen Age, 55 (1988), 143–67 (reprinted in Id., Magic and Divination in the Middle Ages. Texts and Techniques in the Islamic and Christian Worlds (Aldershot, 1996), XI); John S. Lucas, Astrology and Numerology in Medieval and Early Modern Catalonia. The Tractat de prenostication de la vida natural dels hòmens (Leiden-Boston, 2003), esp. 52–3 and 96–8; David Juste, ‘Comput et divination chez Abbon de Fleury’, in Abbon de Fleury: philosophie, science et comput autour de l'an mil, ed. B. Obrist (Paris-Villejuif, 2004, reprinted 2006), 95–127; Roy M. Liuzza, ‘The Sphere of Life and Death: Time, Medicine, and the Visual Imagination’, in Latin Learning and English Lore. Studies in Anglo-Saxon Literature for Michael Lapidge, edited by K. O'Brien O'Keeffe, A. Orchard, II (Toronto-Buffalo-London, 2005), 28–52; David Juste, Les Alchandreana primitifs. Etude sur les plus anciens traités astrologiques latins d'origine arabe (Xe siècle) (Leiden–Boston, 2007), esp. 141–55 and 206–17; László S. Chardonnens, Anglo-Saxon Prognostics, 900–1100. Study and Texts (Leiden–Boston, 2007), 181–222.

⁸ The term ‘abjad’ derives from Arabic and properly applies to the Arabic language only. For convenience, I use it here for all languages.

⁹ O. Neugebauer, (note 6), 162 n. 3, notes however that the Ethiopic onomantic texts he surveyed use a numerical alphabet adapted from Greek.

¹⁰ All the texts are edited in D. Juste, Alchandreana (note 7).

¹¹ All the texts are edited in D. Juste, Alchandreana (note 7), 220–221.

¹²‘a sapientissimo Caldaeorum magistro me imbuente in Latinum ex Caldaeo eloquio translatum’ (Proportiones, 48.1). ‘Caldaeus’ is often used for ‘Arabicus’ in the Alchandreana.

¹³ ‘Ipsorum igitur nomina noscas more Aluaten Sarraceni magistri ueraciter descripta’ (Proportiones, 58.2).

¹⁴ D. Juste, Alchandreana, (note 7), 226–7. Liber Alchandrei, 27 is a literal translation of Baraitha di-Mazalot, 10 or Baraitha di-Shmu'el, 6 (ibid., 41 and 111–2; see also R. Leicht, (note 5), 87, and especially Gabriele Mancuso, ‘La scienza astrologica nel mondo ebraico. Dalla Baraita di Shemu'el agli Alchandreana’, Materia Giudaica, 13 (2008), 135–41, where the Hebrew and Latin texts are compared).

¹⁵ In Greek, see Catalogus codicum astrologorum Graecorum, (note 2), VII, 191; IX.2, 138–9; X, 57, 98–9, and A. Delatte, (note 2), I, 68; in Syriac, E. A. W. Budge, (note 3), II, 620–1; in Arabic, J.-M. Faddegon, (note 4), 151–2, and note 21 below; in Hebrew, R. Leicht, (note 5), 142, 159, 160 and 165, and M. Steinschneider, (note 5), 138 n. 32 and 144; in Ethiopic, O. Neugebauer, (note 6), 162–3.

¹⁶ D. Juste, Alchandreana, (note 7), 141–7. Other onomantic practices deriving from the spheres of life and death and the victorious and the vanquished are also found in the Alchandreana (ibid., 206–17); these will not be discussed here.

¹⁷ The lunar mansions are the result of a division of the ecliptic into 28 equal sectors of about 13° each, corresponding to the daily mean movement of the Moon through the zodiac.

¹⁸ There are variations in the texts concerning the value of the divisor (9 or 7) and the planets allocated to remainders 8 and 9 (see D. Juste, Alchandreana, (note 7), 146–7), but these do not affect the present discussion.

¹⁹ Chapters of onomantic astrology, together with their corresponding alphabets, occur as follows: In principio, 9 (instructions for mansion/sign, with alphabet S1) and 10 (planet, P1); Benedictum, 1 (planet, P1) and 13 (mansion/sign, S1 together with an Arabic alphabet); Quicumque, 15 (planet, P2 and P4) and 16 (mansion/sign, S2); Proportiones, 9, 34.4–7, 57 (mansion/sign, no alphabet) and 34.1–3 (planet, no alphabet); Liber Alchandrei, 21 (mansion/sign, with a Hebrew abjad) and 29 (planet, no alphabet). The Proportiones, with the exception of their first version (see below), contain no numerical alphabets in conjunction with the onomantic procedures, but alphabet S1 occurs separately in chapter 58^bis – with no explanation as to its purpose. The Liber Alchandrei also contains a Greek abjad (c. 3), again with no explanation as to its purpose. In addition to the five main texts, onomantic astrology appears in the isolated chapters Fragmenta Alchandreana, 11 (mansion/sign, S3) and 12 (planet, P3); and 22.2–4 (mansion/sign, S1) and 22.5–8 (planet, P1). For a detailed account of the numerical alphabets in the Alchandreana, see D. Juste, Alchandreana, (note 7) 147–5 and 675–2.

²⁰ Richard Lemay, ‘Arabic Numerals’, in Dictionary of the Middle Ages, edited by I. Strayer (New York, 1982), 382–98 (386) (see also Id., ‘The Hispanic Origin of Our Present Numeral System’, Viator, 8 (1977), 435–62: 441–2).

²¹E.g. Paris, BNF, ar. 2582, f. 1v-2r and 76v; Paris, BNF, ar. 2584, f. 2r and 4v-5r; Paris, BNF, ar. 2587, f. 1r-2v. These manuscripts, which I consulted with the help of Sophia Vasalou, preserve various versions of a text on nativities attributed to Abu Ma‘shar, on which see also J.-M. Faddegon, (note 4); and Fuat Sezgin, Geschichte des arabischen Schrifttums, VII (Leiden, 1979), 145, item 8.

²²For example, kāf=20/12, remainder 8; lām=30/12, remainder 6 etc. All indicated values are correct and the values of the first ten letters, being inferior to 12, logically retain their abjad values.

²³ Small numerations (reductions modulo 9 and 7) are also found in Greek (Hippolytus of Rome alludes to them in his Refutatio omnium haeresium written in about 230 A.D., see P. Tannery, (note 2), 232–3; O. Neugebauer and G. Saliba, (note 2), 196) and Syriac (E.A.W. Budge, (note 3), II, 533). In Arabic, besides al-jummal al-ṣaghīr, reduction modulo 9 of the western abjad is found in the form of the mnemonic formula 'yqsh [1], bkr [2], jls [3], dmt [4], hnth [5], wṣkh [6], z'dh [7], hfẓ [8], ṭḍgh [9], where 'yqsh stands for alif-yā’-qāf-shīn, bkr for bā’-kāf-rā’ etc., according to Ibn Khaldun (1332–1406), who attributes the formula to the ‘onomancers’ (see F. Rosenthal, (note 4), I, 237). Ibn Khaldun also gives a second mnemonic formula, whose structure escapes me: ’rb [1], ysqk [2], jẓt [3], mdwṣ [4], hf [5], tkhdhn [6], ghsh [7], ḥ‘ [8], tḍẓ [9] (ibid., 238; M. Grignaschi, (note 4), 95).

²⁴ All values are correct unless otherwise indicated. Modern Arabic letters are given in square brackets.

²⁵ Readers who are skeptical about this claim should try with the name ‘Stephanus’.

²⁶ The association p = bā’ (instead of fā’) seems to indicate that the translator was not Jewish.

²⁷ As shown by the double occurrence of letters m and n in S2 and P4 (see Table 3), reflecting the values of those letters according to their initial/medial or final position in Hebrew words. Likewise, the value f = 16 in most manuscripts of S2 can only correspond to that of the final pe in Hebrew (see also note 38).

²⁸ The presence of two alphabets (P2 and P4) for the calculation of the planet is also puzzling, and alphabet P4 (reduction modulo 7) seems irrelevant since the text refers to a division by 9 only. Chapter 15 is extant in two manuscripts, Vatican, BAV, Reg. lat. 1324 (the base manuscript of the edition), which gives alphabet P4, and Florence, BML, Plut. 30.29, which gives both P2 and P4, so that the issue cannot be sorted out. This problem, however, does not affect the present discussion.

²⁹ ‘Nota autem quod, secundum Hebraeos, si m uel n in capite nominis fuerit, uel in medio, numerum quem habent accipies, scilicet pro m V, pro n I, in fine uero pro m V, pro n VII; pro a V¹ in feminino nomine, i raro tollitur, u numquam nec o; quoniam t pro c ponitur, habetur pro VIIII; item, m, n uel c in aliquo nomine bis inuenitur, ultimum enim m pro DC tene, n pro DCC, c pro DCCCC’ (Quicumque, 15.2). The values given for m and n in the opening are those of P4.

³⁰ On this version, see D. Juste, Alchandreana, (note 7), 87–94. I have not edited the text, with the exception of its ‘original’ chapters (= Fragmenta Alchandreana, 11–21). The text is extant in MS Burgo de Osma, Archivo de la Catedral, 7, f. 92v-111r, a manuscript copied in the eleventh or early twelfth century, but of which there are good reasons to believe that it is a faithful copy of the original.

³¹ Fragmenta, 11–12 (edited by D. Juste, Alchandreana, (note 7), 646).

³² The Latin abjad was therefore not invented by the twelfth-century translator Hermann of Carinthia, as Richard Lemay suggested in his ‘Nouveautés fugaces dans des textes mathématiques du XIIe siècle. Un essai d’abjad latin avorté’, in Sic itur ad astra. Studien zur Geschichte der Mathematik und Naturwissenschaften. Festchrift für den Arabisten Paul Kunitzsch zum 70. Geburtstag, edited by M. Folkerts and R. Lorch (Wiesbaden, 2000), 376–91: 377–80.

³³MS Burgo de Osma, Archivo de la Catedral, 7, f. 104v (some letters are partly or entirely illegible).

³⁴ Only letters a, b, d, and possibly h, have the same value in the Latin and Hebrew/Arabic abjads (see Table 1).

³⁵See note 19 above.

³⁶ ‘Accipiat nomen suum maternumque naturale et fatale, non illatum aliunde nec deprauatum uulgaris appellationis simplicitate, consideretque apices eorundem nominum iuxta Hebraicarum figurarum supputationes, quas michi uidetur rationem formasse et naturam, non cuiusque sui competentem uoluntatem’ (Liber Alchandrei, 21.3).

³⁷ ‘Litterae autem Hebraeae… hic cum numeris et expressionibus suis subscribantur, ut de hac re disputanti iugiter in aperto habeantur’ (Liber Alchandrei, 21.7).

³⁸ These associations are basically identical to those of the Quicumque, including f= final pe (see note 27).

³⁹ ‘Cognitis his litteris cum computatione earum, illud diligenter adtende quod ubicumque huiusmodi nomen aliquod iuxta Hebraeos apices computes. Si praecedat he, alef pro ea subroges; et in omnium e contra muliebrium nominum fine, pro alef he supputes; in masculorum autem non ita. Memento etiam quod in initio uocabulorum omnium si uau, quae modo o modo u litterarum uim obtinet, uel iod simplex et loco uocalis praecesserit, alef semper praemitte, et post eandem uocalem quae praecessit, cum numero suo debes supputare. Illud etiam prae omnibus recordare, he et alef ubique in mediis dictionibus uim suam perdere si consonans praecesserit quae eandem uocalem in se naturaliter habuerit; alioquin pro e pura, alef scias semper supputari’ (Liber Alchandrei, 21.9–12).

⁴⁰ Almost all numerical alphabets copied in manuscripts of the Alchandreana contain at least one mistake and, in some cases, alphabets S were mistaken for alphabets P, and vice versa (see D. Juste, Alchandreana, (note 7), 155 and 676–81). Alphabets S1-S2-S3 and P1-P2-P3-P4 subsequently spread, in more or less corrupt forms, in other medieval onomantic texts, e.g. alphabet S2 is used in the Catalan Tractat de prenostication de la vida natural dels hòmens (see J.S. Lucas, (note 7), 96). In his Arithmologia published in 1665, Athanasius Kircher cast aspersions on alphabet S2, whose numerical values, he says, ‘were undoubtedly invented by men or demons devoid of common sense’ (‘In hoc alphabeto singularum literarum valorem habes numericum, quem sine ratione vel homines vel daemones invenisse, certum est’, Arithmologia sive de abditis numerorum mysteriis (Rome, 1665), 227). Likewise, those alphabets have puzzled modern scholars, see, besides Lemay (as above), Hans J. Vermeer, ‘Quellen der Namenmantik und Christliche Piraterie’, Leuvense Bijdragen, 54 (1965), 111–7: 112–3; and J. S. Lucas, (note 7), 52 and 55 n. 24.