There are idiots in this world who babble about the rise and fall of the Islamic Golden Age in terms of rising and falling mean intelligence!!! Amazing. Here is a simple illustration of “more people, more ideas”.
Fans of Charles Murray’s Human Accomplishment have come away with the idea that human accomplishments are concentrated at the top of the ability distribution and have also occurred disproportionately in Northwestern Europe. But one aspect of the book which neither critic nor fan seems to have internalised, is that the rate of innovations accelerates over time as we get closer to the present. Conversely, before the modern era, innovations proceeded with all the grande vitesse of a snail after a Frenchman had already cooked it in butter and garlic. From Chapter 14 of Murray’s book :
The average annual growth rate of creative output per capita — can I say that ? — was near zero throughout most of human history. What does the chart remind you of ?
In the pre-modern world, economic growth was glacially slow because technological progress was glacially slow. In our world, where scientific knowledge and technological progress measurably increase every year, we naturally think in terms of creativity as a function of geographical space — where it is located and concentrated rather than when or how often it occurs. Look up the history of technology and over the course of thousands of years there are numerous and significant inventions whose cumulative impact may be quite large. But on an annual or decennial or even centennial basis they amount to basically nothing in terms of impact on the material standard of living.
Joel Mokyr dwelt at some length in The Lever of Riches on the innovations of the “Dark Ages” which were lacking in the Roman world :
At the beginning of the twentieth century retired French cavalry officer, Richard Lefebvre des Noëttes, compared the use of horses in antiquity and the Middle Ages. He discovered that the Greeks and the Romans used a throat-and-girth harness in which two straps were wound around the belly and neck of’ the horse. The neck strap simultaneously pressed on the animal’s jugular vein and cut off the windpipe as soon as it began to exert pressure. Lefebvre des Noëttes found by experiments that a horse strapped this way lost about 80 percent of its efficiency. In early medieval times, by way of contrast, such an easily corrected waste of valuable energy was not tolerated. The solution to this problem emerged when the breast strap, which rested against the horse’s chest, and the collar harness, which rested against the shoulders, were invented. Both eliminated the yoke, and thus avoided the main shortcoming of the Roman harness. The breast strap appeared somewhat earlier than the collar, but both were more or less in place in the ninth century. Consequently, horses gradually acquired major economic importance in agriculture and in the hauling of wagons. The horse harness was supplemented by other advances in horse technology”.
So the average annual increase in agricultural productivity due to the change in this particular use of animate energy was near zero.
In the absence of standardised test scores people often infer intelligence and creativity from achievements. Well, what might one say about the pre-modern world ? In any given generation you just didn’t need a very large “smart fraction” to produce the drastic paucity of scientific, technological, and artistic innovations.
Besides, in a world which was composed of something like 90% subsistence farmers riddled with nutritional deficits, 1-5% artisans, and 1-2% merchants, how smart did you really have to be to outperform the average ? I suppose there must be some small constant of skills required to be a tax collector whether in 100BCE or 2014, but the cognitive demands are obviously much bigger now. And one can infer from age-heaping evidence that a large percentage of people in the Roman period most likely did not know their own age precisely. Here are some “rounding index” data from Roman Italy (higher the index, the worse) :
Aurelius Isidorus was a wealthy Roman landowner in Egypt who made numerous age declarations in legal documents which survived to our day as papyri fragments. But Duncan-Jones added the following note: “Isidorus, who was called agrammatos [=“unschooled”, PE] and had to declare his property through an intermediary, was evidently illiterate… The Egyptian evidence in general shows a substantial proportion of the propertied class as signing documents through intermediaries, apparently because they could not do otherwise…” I cite the above, not to argue that Isidorus & co. were dim-witted, but that in a world of peasants even an innumerate-illiterate could be king.
So how high were the cognitive requirements in the long preindustrial age ? How big a “smart fraction” did you need, in order to account for the actual pace of creative output that we did see ? What is the required “smart fraction” to account for preindustrial civilisational accomplishments ? It is in light of that question that I now turn to the Golden Age of Islam.
§ § § §
Conventional history dates the Golden Age of Islam from the foundation of the Abbasid Caliphate in ~750 to the destruction of Baghdad by the Mongols in 1258. I personally would push the start back to the foundation of the Umayyad Caliphate in ~650 and date the end farther in time to the decline of the Ottomans starting in the 17th century. All the same, the earlier part of that millennium is famous for both original output in science, philosophy, mathematics, and the arts, as well as the (partial) preservation of ideas from classical Greece and Rome. Less well known is that Islam served as the conduit for the westward transmission of ideas from Persia, China, and India.
The Islamic world — Dar al Islam — that existed as of 1250 was politically fragmented but constituted a unified cultural space, a world system. Whereas the inhabitants of the Islamic and Christian worlds could not freely cross back and forth across their civilisational frontier, five of the major ancient religions — Jews, Zoroastrians, Christians, Muslims, and Hindus — inhabited the Islamic world and they were connected in a global network that stretched from the Atlantic Ocean to the frontiers of China and Southeast Asia.
It was an unprecedented unification of disparate cultural spaces. Neither the Persian, nor the Hellenistic, nor the Roman, nor the Byzantine empires had quite brought together so many different areas of the world that were so civilisationally dense. Even in 750 the Islamic world included all of Roman North Africa, Levant, and Spain; as well as what had been Persian Mesopotamia and Iran; plus the western part of the Hindu-Buddhist world; plus the southern portion of Central Asia wholly encompassing the Silk Road routes. This was as of 750.
Direct translations from Greek into Arabic took place in Egypt. But less well known is that a large body of Greek thought had already been translated into other ancient languages, such as Syriac, and many of the Arabic versions of Greek texts were actually translations of these earlier translations. The Nestorian Christians — refugees from Byzantine anathematisation — played a commanding role in rendering these Syriac texts into Arabic. According to Dimitri Gutas in Greek Thought, Arabic Culture, Middle Persian translations of Greek works were also retranslated into Arabic.
But via Persian, the Arabs inherited a more eastern source of knowledge :
Indian scientific material in astronomy, astrology, mathematics, and medicine passed into Arabic mainly through Persian (Pahlavi) intermediaries during the ‘Abbasid period, and as such it is to be seen in the context of the translation movement…. Beyond, that, however, the pre-‘Abbasid translation of some of these astronomical texts into Arabic is significant for establishing the existence of a sufficient number of international scholars, as I called them above (chapter 1.1), whose talents could be drawn upon to serve the translation movement set into motion by the early ‘Abbasids.23
The importance of the Iranian element clearly shows up in a simple enumeration of Golden Age figures which mattered most to European mediaeval scholastics of the 13th & 14th centuries :
- Alberonius (al-Biruni)
- Algazel (al-Ghazali)
- Alfarabi (al-Farabi)
- Algoritmi (al-Khwarizmi)
- Avicenna (ibn Sina)
- Averroes (ibn Rushd)
- Aljazari (al-Jazari)
- Maimonides (Moshe ben Maimon)
These figures would loom pretty large in any “citations index” of European mediaeval Scholasticism. For example, in Summa Theologica, Thomas Aquinas was sometimes baffled by Aristotle’s metaphysics and turned again and again for guidance to the ‘Commentator’ — Averroes. (The ‘Rabbi’ of the text is Maimonides and Avicenna was cited by Aquinas directly.) With the exception of Maimonides, a Jew from Islamic Spain, and Averroes, an Andalucian, everybody else in the above list is some kind of Iranian. And by ‘Iranian’ I mean it in the sense of “Greater Iran“: large parts of Turcophone Central Asia were once populated by speakers of Iranian languages, and today at least 10 countries lay claim to some of these ‘Iranian’ Golden Age figures. Amongst them was even the Soviet Union, by virtue of claiming to be “one of the largest Muslim countries” :So the Golden Age of Islam involved a population of unconverted and converted peoples, with the latter primarily from regions of the world that had already been civilisationally dense half a millennium prior to the Islamic conquest. This cultural “free trade zone” encouraged a mobile, transnational intelligentsia seeking patronage in what ever part of that world was richest and most powerful at any given time. Thus, the Andalucian Jewish philosopher-rabbi Maimonides spent much of his career in Egypt.
§ § § §
So what is the “smart fraction” required to produce the flowering of the Islamic Golden Age ? What mean intelligence might be required ? Here is how I go about the guess.
We can infer the size of the intelligentsia in the European Middle Ages from the clerical population. According to the poll tax records, the clergy of England circa 1377 were approximately 2% of the population. This would include everyone from the most illustrious canon of the church down to the village friar. Without double-counting the rich beneficed clergy in the poll of landowners, I add in non-clerical landowners for a total of 3% of the population. And 3% is the figure I will apply to the Islamic world.
Based on data in Maddison I’ve come up with 40-50 million people living in Dar al-Islam of the year 1000. This would seem to be an underestimate compared with the >50 million from this source. So in the year 1000 the population of the Muslim-ruled parts of the world must have been 15-20% of the global population. This is roughly double the share of Western Europe, comparable with China (at the upper end of the Muslim range), and smaller than India (whether or not including the Muslim-ruled areas).
As reference I offer the following computations of “smart fraction” percentages for several mean IQs and two different standard deviations :
This would be per generation. It seems to me, even at IQ=70, the sheer size of Islamic civilisation alone could have accounted for the ultra-elite intellectual eminences of the Golden Age. If one were to enumerate, à la Murray, all the scientists, mathematicians, physicians, engineers, writers, artists, and historians, plus discrete anonymous accomplishments, throughout the entire period of 650-1350, the roster may not even reach 1200 individuals (the number for a single generation with IQ>130 for a population of 40 million with µ=70 and σ=15).
The number of eminences of the Golden Age was barely distinguishable from noise. So if much of its achievements were due to the scale effects of a vast multinational civilisation, and perhaps also to the increasing returns to scale from assembling the smartest fractions of the ancient world, then you can easily imagine how fragile a phenomenon it probably was. Thus, the slow disintegration of the Andalusian caliphate by civil war and by Christian reconquest ; the pressure created by the arrival of the wild, barbarian Turkic peoples ; the utter destruction of the Abbasid caliphate and its capital Baghdad by the Mongols ; and the further devastation wrought by Tamurlane who basically repeated the Mongol destruction, etc. — all these must have all taken a toll. And by “taking a toll”, I mean raze the cities and kill the inhabitants. And since the intelligentsia were hyperconcentrated in cities in mediaeval times, almost certainly these depredations amounted to an intellectual holocaust.
But of course I would argue that the Golden Age did not end with the sack of Baghdad. It simply shifted west to the Ottomans, who maintained their dynamism into the 17th century, and to the Mamluks in Cairo, not to mention more easterly precincts, with the Timurids and the Mughals.
Many (most?) pre-modern cultural efflourescences might also be similarly explicable by population scaling effects, including China, the Roman Empire, the Hellenistic age (the successor states of Alexander the Great), and perhaps (perhaps) even classical Greece. I’m reasonably sure if one were to apportion the “eminent Greek accomplishments” out by generation and by population the apparent density of achievement in Athens during a single century would not appear nearly as magical.
Edit: Some people were confused. What I wanted to say was :
- in most of the premodern period, the density of observed achievement (relative to population, time, space) was so small that you don’t need very many intelligent people to explain it;
- I really don’t know what the intelligence of premodern peoples was, but we probably shouldn’t infer the population mean from premodern achievements;
- there’s no need to invoke dysgenic or eugenic reasons for the fluctuations in the fortunes of civilisations, as so many cranks are wont to do.