When it comes to details, the classical narrative cannot account for the very scientific facts that have been preserved either in the classical historical sources of the period or in the scientific texts themselves. For example, more than one historian tells us[34] that when the caliph al-Manṣūr wished to build the city of Baghdad, in 762 A.D., he assembled three astrologers and charged them with casting the horoscope for the future city. They were supposed to choose the time for the foundation so that no potentate would be killed in the city. The horoscope itself is preserved in the Chronology of Bīrūnī, and in several other sources. Most sources agree that the astrologers who were assigned that task included Nawbakht (a Persian astrologer who became the progenitor of the Nawbakht family of astrologers, which served caliphs for a whole century), Ibrāhīm al-Fazārī, and Māshā'allāh al-Fārisī. Bīrūnī states explicitly that it was Nawbakht who determined the day for the foundation of the city to coincide with the propitious 23rd of July of that year.
If the ancient Greek sciences were supposed to have been brought into Arabic by the Persian-leaning elements of the Abbāsid dynasty, even if we grant that this interest started with al-Manṣūr himself, and if we grant that they could recruit for the purpose of the horoscope the Persian astronomers Nawbakht and Māshā'allāh, then who was this Ibrāhīm al-Fazārī, obviously an Arab from the tribe of Fazāra, who was also invited to join them, and where did he acquire the kind of advanced astronomical knowledge that he would have needed for casting such a horoscope at that early time in the Abbāsid reign? Where did his usual collaborator Ya'qūb b. Ṭāriq learn his own astronomy so that he could produce, together with Fazārī, a translation of the Sanskrit Sidhanta (al-Sindhind), which was completed during the caliphate of al-Manṣūr (754-775 A.D.)?[35] Later sources always joined those two names together,[36] so it is sometimes difficult to determine who did what. For the purposes of the Baghdad horoscope, we may stipulate that Fazārī may have learned his craft in Persia. But the sources are silent on that, and we do not know much about the Persian astronomy of the time beyond the existence of the Shariyār zīj (which was quoted in later sources). Furthermore, the historical sources that connect the two assert that this very same Fazārī and/or Ibn Ṭāriq also wrote a theoretical astronomical work called Tarkīb al-aflāk, which seems to have been lost. The same Fazārī is also credited with the authorship of his own zīj, in which he used the "arab years" ('alā sinīy al-'Arab).[37] Writing a theoretical astronomical text, transferring a zīj to a different calendar with a completely different intercalating scheme, and producing astronomical instruments such as astrolabes—as we are also told about these men—could not have been done by amateur astronomers. Who educated Fazārī and Ibn Ṭāriq in all these fields of astronomy? And even if we believe that the three astrologers also used the Persian Zīj-i Shahriyār for the purposes of the horoscope, we should also ask about another Arab, 'Alī b. Ziyād al-Tamīmī, from the tribe of Tamīm, who was supposed to have translated this zīj into Arabic.[38] Who taught al-Tamīmī how to translate a zīj, and when he did so did he also transfer it into Arab years (as we are told Fazārī had done)?
All this evidence indicates that there was a class of people, who were already in place by the time the Abbāsids took over from the Umayyad dynasty, who were competent enough to use sophisticated astronomical instruments, to cast horoscopes, to translate difficult astronomical texts, and to transfer their basic calenderical parameters, as well as to compose theoretical astronomical texts such as Tarkīb al-aflāk. Such activities could not have been accomplished by people who were just learning how to translate under the earliest Abbāsids, as the classical narrative claims.
The situation gets more complicated, again on the level of details, when we look at the works that were produced about 75 years later by people like al-Ḥajjāj b. Maṭar (fl. ca. 830), who translated the two most sophisticated Greek scientific texts: Euclid's Elements and Ptolemy's Almagest. We know, for example, that al-Ḥajjāj finished his translation of the Almagest in the year 829, as is attested in the surviving copy now kept at the Library of Leiden University (Or. 680). And when we look at this translation we are immediately struck by two most startling phenomena: the language of the text is impeccably good Arabic, technical terms and all; and the Arabic translation even corrects the "mistakes" of the original Greek Almagest. Who taught al-Ḥajjāj the technical terms, and who taught him how to correct the mistakes of the Almagest? Neither of these questions is resolvable if we continue to believe the classical narrative that dates the beginning of the serious translations to the time of al-Ma'mūn (813-833). Early translations usually struggle with technical terminology, and usually do not go beyond the letter of the text and would never dare correct its mistakes, if they could understand the text in the first place.
Furthermore, we know that al-Ḥajjāj's translation of those scientific works was not the first. In fact, we are explicitly told by some sources that those two books were already translated under the patronage of Khālid al-Barmakī, the vizier of Hārūn al-Rashīd (d. 809), and maybe by al-Ḥajjāj himself, and by others that they were translated during the time of al-Manṣūr (754-775).[39] But the farther back these translations are pushed, the more complicated the story becomes, for the question of the development of technical terminology would still persist and actually becomes even more difficult to answer. In any event, the text as it is now preserved in the 829 A.D. translation reveals a maturity that could not have come from one generation of translators. And thus we must allow for a longer period of translation so that more than one generation of translators would create enough output to produce technical terminology and teach the sophisticated mathematics and linguistic skills that were required to render the Almagest, the Elements, and similar books into the kind of coherent Arabic in which they are preserved.
During the same early period — that is, during the reign of al-Ma'mūn — we also witness the creation of the new discipline of algebra by Muḥammad b. Mūsā al-Khwārizmī (fl. ca. 830),[40] already in a mature format—treating, for example, the field of second-degree equations in its most general form. This happened before the translation of the work of Diophantus and other Greek sources. This does not mean that classical Greek sources, or for that matter ancient Babylonian sources, did not include algebraic problems, but the coinage of the new term for algebra (al-jabr), and the statement of the discipline in general as different from arithmetic,[41] required a kind of maturity that could not have come with the first generation of translators if we assume that translations began with the early Abbāsid times as the classical narrative stipulates. Under such circumstances we are entitled to ask " Who taught al-Khwārizmī to do what he did?"
34
See, for example, the work of the eleventh-century polymath Abū al-Raiḥān al-Bīrūnī,
35
On the role played by those astrologers in appropriating the Indian sources under the patronage of the caliph al-Manṣūr, see Nallino, pp. 141-215.
36
On the surviving fragments from the works of those two astronomers, see David Pingree, "The Fragments of the Works of al-Fazārī",
39
For the report on the
40
See the edition and translation of this work by Frederic Rosen, The Algebra of
41
For a similar argument see Carl Boyer,