To motivate more Americans to study science, reduce the importance of English as a global language. This rather skewed line of analysis was motivated by this excellent article from Scientific American which discusses some reasons top students in the US tend to choose careers such as law or business over engineering and hard sciences; though the article focuses on structural problems in grant funding and academic job markets that also discourage students from going into such careers. Even if these structural issues were solved, however, there’s strong reason to suspect that top students, would, at the margin, prefer these ’soft’ careers, for much more pervasive structural reasons: a global language and culture.

The benefits of a global language monopoly (resulting primarily from numerous heavily armed contingents of fellows in tweed hats that enjoyed sailing and oppressing locals, cheerio) make the per-input-unit (years of education) returns from social-network-intense careers (law, business, finance, marketing) much more attractive for native born Americans than careers where “portable skills” are required (science, engineering and medicine, to some extent). This is one of the major reasons people like myself can command considerable wages purely by virtue of having been born in a certain geography, without many marketable skills.

For your average Chinese (or Indian or Japanese or Vietnamese…) overachiever, the choice between spending 7 years learning finance + 7 years learning English language and Western cultural taboos. 10 years learning programming are probably a better choice. I don’t mean to suggest that cultural familiarity is useless for scientific pursuits but merely to suggest that social skills facilitated by familiarity with taboos are more necessary in social-network-intensive careers.

Even if grant funding was reformed (an eminently reasonable suggestion, from the arguments presented in the article), my suspicion is that per capita, Americans would be much less likely to choose science careers from a purely rational cost-benefit perspective. English as a lingua franca is viewed by some economists as a coordinating equilibrium; even though English fluency may be very difficult for northeast Asians, for example, it is much easier than learning Hindi, German, Russian and Portuguese simultaneously. If it were not English it would probably be something else, say French (15th-17th century) or Mandarin (22nd century on?)

Much of the modern world’s coordinating activities are, as a result, conducted in English. This includes everything from the colloquial (most introductory emails to anyone overseas) to the refined and dignified (international contracts and treaties). Even for English speakers who have no direct contact with foreigners, the spillover benefits from lingua Anglica are considerable: the people they work with might be involved in overseas business, and derive significant advantages from being intimately familiar with language and cultural practices.

Furthermore, specialization would tend to occur more as globalization (in trade in goods and services) spreads and allows for greater specialization, making the above network effects even more important. With a greater possible pool of people to interact with, the rewards to learning the lingua Franca increase.

One intuitive way to examine this phenomena might be to examine intensity of science and engineering graduates, vs the Anglosphere:

And number of science/engineering graduates (green/yellow = lots, black/orange = fewer).

I suspect you’d find an even stronger relationship if you defined language families with etymological similarity to English. Could also expand the sample: China produces about 3 times as many science and engineering grads as India, the latter of which has much stronger native familiarity with English language and Anglosphere culture. The only outlier seems to be Norway (only 4 million people).

Another way to look at it would be to examine the immigrant intensity in science and engineering fields. Again, English speaking countries have a much higher percentage of immigrants. Unsurprisingly, the same pattern holds (though in this case, the US is a fairly weighty data point, given other reasons why lots of people probably might want to move there, and study science and engineering because it’s the only reasonable path to actually get in, legally).

A-priori we would expect such a world to feature greater specialization in portable skills (where language and cultural norms aren’t so important) in areas with less innate familiarity with the language and cultural norms. In regions with this familiarity (or degrees of it), we’d expect specialization in activities that become more remunerative when used with coordinating activities; as well as education (since we force everyone to speak our smooth, grammatically impenetrable mother tongue).

So, as long as English remains important as a global language, expect to see a disproportionate amount of top-talent go into finance, business school and law – evil bottom feeders that leech of the productive capacity of selfless doctors and engineers working to cure cancer and invent harmless nuclear energy. [Being one of those bottom-feeders, I would urge readers to understand the importance of coordinating activities (legal contracts, temporal shifting of obligations) in running a specialized economy that makes everyone wealthier globally.]

Broadly, these trends are probably a net gain (especially for people born in the US) and should be of concern only if: 1) you feel that science is really special and should be pursued past the point of rational costs, 2) you’re worried about national security and feel the US should have more really smart computer scientists and bio-weapon engineers. If the latter instance, there is then a very defensible argument for subsidization to overcome the innate lingua Anglica advantage, and motivate people to go into science.

As a consequence, observing career trends for top American/foreign students in the coming decades may be a leading indicator for Anglosphere-hegemony collapse. Individuals figuring out what to do with their lives are probably much more sensitive to minute changes in incentive structures, and examining their preferences provide a clearer early indicator than it would be possible to glean from numerous types of aggregate macro data. Disclaimer: personally, I tend to be much more confident in near-term U.S. prospects than a many other articles that have come out lately. Within a century, though, many things are possible.