The Axial Age & world population

By Razib Khan | December 30, 2010 11:26 pm

A few days ago Robin Hanson brought this chart of world population to my attention:

On the x-axis you have time, 12,000 years ago to the present. On the y-axis an estimate of the total world population log-transformed. The data is derived from the US Census low estimate. Granting the data’s accuracy for the purposes of reflection, Robin’s question was what could have occurred between 1000 and 500 BC to produce such a rapid population rise?

My immediate response to Robin was that perhaps the transition from widespread utilization of bronze to iron democratized tool use so that more land was brought under cultivation. Bronze tools and weapons were the privileges of the elite because of the high capital investments for the production of the alloy. Stone or copper remained the norm for peasants. With the switch to iron per unit cost of production for metal tools went down. There is a hypothesis for example that only extensive use of iron tools allowed for the clearing of the eastern Gangetic plain and the expansion of Indo-Aryan civilization to the Bay of Bengal.

A second complementary suggestion I made is that biological changes in the horse allowed for the emergence of full-fledged nomadic lifestyles with the development of mounted cavalry. In Empires of the Silk Road the author makes the argument that Inner Asian nomadic groups were much more important in being facilitators for diffusion of ideas, and even the originators of ideas, than we give them credit for. A tentative assertion is made that the Axial Age itself was the work of horse riding nomads! Whatever the reality of that specific claim, one could outline a model where the free flow of ideas accelerated during this period because of the rise of mobile populations such as the Scythians, whose cultural domain spanned the whole Ecumene.

And then there are even stranger ideas, such as Julian Jaynes’ ‘bicameral mind.’


Image Credit: Waldir, Wikimedia Commons

CATEGORIZED UNDER: Culture, Economics, History

Comments (18)


    The above paper may be relevant. The integration of different regional populations in the Old World into a world system may have started a feedback loop, if I understand the authors correctly.

  2. My first thought looking at this diagram (leaving aside that the jump from 500BC to 400BC is caused by using a single data point for 400BC from an alternative source) is not so much why the acceleration occurs from 1000BC to 500BC but rather, why does it stop? Assuming a scale effect in technology (more minds, more ideas) and that population is linked to the level of technology in this Malthusian world, you would expect greater than exponential growth in population. The graph should get steeper even with the logarithmic scale. It does that, until 400BC.

    Is the sudden decrease in population growth due to the decline of empires or higher death rates in cities? That, to me, is the puzzle.

    My other thought is that we need to look to Asia for our answer (as you have done). McEvedy and Jones, on whom all but one of the data points between 5000BC and 700AD are based, put around half of the world’s population at that time in India and China, with another quarter in the Persian Empire. A lot of the discussion on Robin Hanson’s blog was around the Roman Empire (which given Robin’s point, might be expected). However, the explanation needs to be far broader than that.

  3. bioIgnoramus

    Surely the clever chaps who can calculate what the population was can also tell us where it was? Won’t that point us to an answer?

  4. twl

    My immediate thought was NH climate.

    There was significant warming between 1000 and 500BC then a cooling before year 0. This correlates with observed population expansion and then the population drop off. In each case the weather might not have been always directly responsible i.e. crop failures, due to floods and droughts and shorter/longer growing seasons, there might have been political instability which directly caused the population changes. Although for nomadic groups in Central Asia you’d thought that length of growing season would be significant if they’re living on the Malthusian edge. There were a lot of large empires which expanded around 300-100 BC compared to say 800-500BC and there is a theory all these changes are related.

    Here is one temperature reconstruction that covers the period (tree rings). Here’s another (Greenland ice core).

    On both you can see warming then cooling.

    Direct link to image 1

    Direct link to image 2

  5. bob sykes

    I would really like to see error bars on those population numbers. I find it very hard to believe that the estimates are good enough to identify an Axial Age population burst. Even the current numbers are suspect because most countries either do not ever conduct censuses or perform them rarely for political reasons. The US is the only country with a long record of population counts.

    Also, bioIgnoramus is right. If these numbers have any validity at all, the estimators can tell us where the people lived. That would provide useful information

    Finally, let’s not bring Julian Jaynes into this. His theory would require a substantial brain reorganization for which there is no evidence. Next we’ll be citing Deepak Chopra.

  6. John Roth

    @Bob Sykes

    If you click on the link, you’ll find where the numbers are from. That gives the sources, which can then be researched if you’re interested.

    I was also kind of surprised to see Razib citing Jaynes. However, I don’t think that it would require a brain reorganization, only a reorganization in the way we use the brain – that is, function, not wiring. I agree that there’s no evidence that it happened, but there is a lot of recent evidence piling up that there’s more than one way to organize the way one thinks.

    As far as the main question goes, I think we’re asking the wrong question.

    If I understand the sequence, this is about the time when literacy expanded from priests and scribes to the elites in general, which makes me think that these people just happened to be the ones in on the ground floor soon after the “beginning of history.” In other words, there’s nothing particularly novel about their ideas other than that earlier people didn’t have the advantage of having them written for posterity.

  7. dave chamberlin

    What I find interesting is that world population growth on this log tranformed graph was virtually a staight line from approximately 5000 BC to 1600 AD, the bump caused by the the Axial Age being the only exception. Off topic as to what caused the additional growth between 1000 and 500 BC, but I find it suprising that world population growth was this predictable for such a long period of time in what I would have thought was our very chaotic past.

  8. Matt

    If these numbers have any validity at all, the estimators can tell us where the people lived. That would provide useful information – links to a chart from, in which context this chart is purported to show a regional, albeit purely Eurasian (East Asia, South Asia, Europe, West Asia), break down of population growth from the same source (McEvedy and Jones – Atlas of World Population History). This breakdown is from an earlier version of the source than Razib’s chart (1975 rather than 1978).

  9. I have some familiarity with the way these numbers are compiled. They all start from the assumption of one-parameter (exponential) growth. Sometimes they draw curves between a series of point estimates (resulting in “kinks”). Point estimates are generally based on area times density. Density is a rough estimate based on This chart attempts to summarize world population, hence it is a combination of estimates from different sources. That tends to smooth things out, as does the assumption of exponential growth everywhere. The squiggles around 1000-1500 AD are the attempts to quantify plague losses.

    If you ask “why faster growth from 1000 to 500 BC”, the most probable answer is the transition from purely archaeological evidence to historical evidence. When they have to start grappling with numbers for army sizes in the eastern Mediterranean, there is a totally different method of accounting. You can’t take these historical numbers at face value, but the people who work with them are not the same people as those who work with site densities and other kinds of archaeological indicators. So you get an incongruity.

  10. Doug1

    I found this to be the most interesting bit in the short Wikipedia article on the “Axial Age”:

    Jaspers [originator of the term and meme] argues that these characteristics appeared under similar political circumstances: China, India and the Occident each comprised multiple small states engaged in internal and external struggles.

    Your theory that the switch to iron for tools and weapons might be a significant factor in the apparent doubling in world population (no doubt largely in EurAsia) is also intriguing and plausible.

    Might they be related? Highly productive but also disruptive technological innovations centered around iron lead to both greater prosperity (and hence soon population), rebellious competing small states, and also a rich marketplace of idea.

    One thing about these small states of the Axial age seems to be that many of them had language and more or less culture in common — e.g. Greece, parts of India, and China. This would facilitate the spread of ideas, while because political power was fractured, also retard the imposition of orthodoxy.

    One theory for why the European West rose so highly and developed both experimental science and secular liberalism for the first time on earth (more or less) in the last 500 years, is that it consisted of increasingly per capita prosperous smallish states which share much culturally (and language among the intellectual elites, first Latin, then French, then English), but was also divided politically, such that inter state competition was intense, and increasingly orthodoxy became harder to impose — there was generally another Euro country that would give asylum to the intellectual luminary persecuted at home.

    Was something similar going on in the high spots of Eurasia during the Axial Age?

  11. Wil

    It is both convenient and tempting to graph these population numbers on a semi-logrithmic graph, as has been done above. Additional insight can be had by graphing the same data non-logarithmically as well.

    Using a non-logarithmic graph, if the “y” axis tops out at 6 billion, then the world’s population will appear as approximately “zero” for the first 11,000 years, and then rise almost vertically to 6 billion during the last 1000 years.

    Since that does not yield much light, modify the non-logarithmic graph by chopping off the last 1000 years, and then rescale the “y” axis to top out at 250 million. Very interesting.

    Razib, would you mind posting the graph described above?

  12. Wil

    Historical human population was (and is) a function of many different things. The fertility rates of various societies were affected by existing disease rates, cultural norms at the time, and many other factors.

    Historical death rates were a function of climate, wars, the locations of each society, the level of medical care, famines due to drought or insects, the level of specialization in the society, the ratio of urban to rural population, each society’s political stability, their average wealth, their average literacy level, their general level of industrialization, and their sophistication regarding farming and husbandry.

    Modeling human population levels is not as easy as modeling animal or microbial populations. A few of the many differences between human and animal reproduction:

    1. Animals and microbes do not engage in periodic wars
    2. Animals are restricted to their specific habitats, and can not move literally anywhere at any time (like people do)
    3. Humans do not become fertile instantly upon birth (like microbes are) or within one to three years after birth (like many animals are)
    4. Animals do not farm or practice husbandry
    5. Animals do not practice medicine
    6. Animals do not pass complex knowledge from generation to generation, such as knowledge of pesticides, fertilizers, safer places to live, how to cure diseases, etc.
    7. Humans do not uniformly reproduce every year until they become unfertile or die
    8. Unlike animals, humans specialize in the mass production, storage, and transporation of food, particularly in more industrialized societies

    Any valid explanation of historical human populations would need to take all of the above into account, and more.

  13. Matt B.

    Was it possible to chop down trees before iron axeheads were available? I can’t imagine a bronze axe being all that good at it, and I’m not sure about stone axes, since the edges aren’t straight. Being able to clear forest and build houses out of wood instead of mud brick would be a boon to population growth.

  14. Clark

    My inclination is to question the development and maturity of trade routes. That would help diffuse technology and practices not to mention seeds.


Discover's Newsletter

Sign up to get the latest science news delivered weekly right to your inbox!

Gene Expression

This blog is about evolution, genetics, genomics and their interstices. Please beware that comments are aggressively moderated. Uncivil or churlish comments will likely get you banned immediately, so make any contribution count!

About Razib Khan

I have degrees in biology and biochemistry, a passion for genetics, history, and philosophy, and shrimp is my favorite food. In relation to nationality I'm a American Northwesterner, in politics I'm a reactionary, and as for religion I have none (I'm an atheist). If you want to know more, see the links at


See More


Edifying books

Collapse bottom bar