Warriors of the Cloisters: The Central Asian Origins of Science in the Medieval World

In the history of science writ large, the scientific revolution has deeper roots than we once thought. Although it would have been difficult to imagine even two generations ago, scholarly consensus now acknowledges the medieval foundations of early modern science. It appreciates not only the central role of the universities in anchoring science in European culture after 1200, but also their fundamental debt to advanced Islamic science. Against this backdrop, the adjective “Central Asian” in Christopher Beckwith’s title is a bombshell that believers in method will love. But first the nutshell.

Beckwith anoints two features of late medieval Europe as key to its “full scientific culture”: the college and the “recursive argument method” (his name for the scholastic method), typically considered indigenous. Beckwith, however, argues that they developed in 1st-2nd century AD Buddhist Bactria/Gandhara, passed into 11th-century Islamic civilisation (where they remained scientifically marginal) and then thrived in the Latin world, setting European science on the path it still treads. Impressive for its scope and detailed analyses, Beckwith’s argument is necessarily a long chain, which he constructs from formal resemblances (less transitive than one might wish) linked by transmission. After forging the initial Central Asian links himself, Beckwith joins them to Arabic studies scholar George Makdisi’s prior linkage of the Islamic madrasa/college and disputation with Europe, and shakily extends that controversial argument to science.

But as a key to the developing medieval Latin scientific culture, the college thesis is unconvincing. Whether rooted or not in the Buddhist vihara or the madrasa, and whatever their later role, privately endowed colleges between 1200 and 1500 housed a small minority of students. It was the universities’ faculties of arts that did the statistically significant work of introducing Graeco-Arabic science to hundreds of thousands of students and fighting to give it an institutional home. Although he is aware of the distinction, Beckwith sometimes blurs it with Makdisi’s “college-university” (a hyphen that looks set to snap).

What about that other juxtaposition of nouns with distinct meanings, the “recursive argument method” (hereafter RAM)? For Beckwith, it names a genre - neither dialogue nor treatise - that is characterised by the recursive pattern of argument, sub-argument, sub-sub-argument, and so on. Instantiated in Buddhist legal disputations, the Central Asian polymath Avicenna’s natural philosophy and the great Latin summae, it becomes the “medieval scientific method”. Indeed, the book’s conclusion finds that structure even in modern scientific articles.

Behind these formal resemblances, Beckwith sees transmissions. After encountering remnants of pre-Islamic legal disputation, Avicenna (who died in 1037) emulated the RAM, which first entered the Latin world cloaked in the 12th-century translation of his commentary on Aristotle’s De Anima.

Here some asymmetries emerge. At the Western end of his argument, Beckwith’s stringent criteria for the RAM exclude all antecedents before the Latin Avicenna; ergo it was imported. Yet his criteria relax noticeably when linking the Central Asian disputation with Avicenna. Now the structure becomes: question; responses to the question; objections to the responses. But do not many earlier texts exhibit and repeat this pattern, from book 1 of Aristotle’s Physics to the Bible’s Book of Job?

The pressures of abstraction necessary for separating form and method from content sometimes squeeze the lifeblood from the subject, and so can excessive generalisation. Although it is widespread in natural philosophy, the RAM is rare in the mathematical sciences. Making it “the medieval ‘scientific method’” bypasses such highly successful theoretical and quantitative efforts as optics and astronomy, Arabic and Latin. More distressing, by judging the RAM and the college to be essential to medieval Latin science but “not integrated into Islamic culture”, Beckwith’s thesis restores to the latter its long discredited role of marginal and passive receiver. Since Islamic science was vastly richer than this, Beckwith’s content-free symptoms of success have clearly missed something awfully important.