The Diamond Fallacy
Jared Diamond's book Guns, Germs and Steel: A Short History of Everybody for the Last 13,000 Years is a fascinating and quite readable speculation on the relationship between geography and history. He has assembled a cornucopia of interesting facts and plausible insights concerning the course of events over the last 13,000 years. The result is well worth reading, despite the fact that I think the ambition of his main thesis reaches well beyond his actual achievement. That discrepancy is due, I believe, to Diamond's having little understanding of what history actually is.
The critique of Diamond's conception of history I offer here is based on the view of the historical enterprise put forward by such philosophers of history as R.G. Collingwood, Ludwig von Mises, and Michael Oakeshott. They share the view that history consists in the effort to identify the particular, past circumstances that make intelligible the subsequent occurrence of other, unique events. Any attempt to explain the human past by reference to general laws or broad patterns is, in this view, a distinctly separate way of comprehending the past from that offered by history. Furthermore, any effort to discover such "laws of history" faces inherent obstacles that prevent it from achieving the sort of success that, for instance, physics has in describing universal laws of matter and energy.
Diamond's work falls within the broad class of theories purporting to detect universal historical laws, and are therefore subject to the same criticisms that Collingwood, Mises, and Oakeshott directed against his intellectual predecessors. His attempt to discern typical patterns in humanity's past is not, in and of itself, absurd or doomed to failure. The main problem with his enterprise is that he seemingly is unaware of what sort of investigation creates the truly historical past. As a result, he proposes substituting his own "geographical past" for the genuinely historical past.
While Guns, Germs and Steel offers many interesting and plausible suggestions as to how geography may have influenced human history, his apparent ignorance of the discipline of history leads him to propose replacing true historical inquiry with a "scientific" hunt for the "ultimate causes" of historical events. Diamond's central error, besides being of interest to anyone concerned with historical methodology, also has broader political implications, which run as an implicit secondary theme through Guns, Germs and Steel, and are made more explicit in his recent book, Collapse. I will address the subject of policy in the conclusion of this article.
Diamond's central conception is that the course of history, broadly speaking, is not determined by individual actions, cultural factors, or racial differences, but by the environmental circumstances into which different groups of people accidentally wandered. More specifically, those groups that happened to wind up in places that offered a variety of plants and animals suitable for domestication, and that made acquiring domesticated species and new technologies from other societies relatively easy, wound up having a decisive advantage over groups located in environments lacking those features. As a result, when geographically advantaged societies encountered groups not so blessed, the outcome was inevitably that the former conquered or absorbed the disadvantaged culture. Thus it is geography, claims Diamond, and not greater inventiveness, a superior culture, or racial differences that is the "ultimate explanation" of why, for instance, Europeans came to rule the Americas rather than American Indians ruling Europe.
Some of Diamond's critics have accused him of excusing past atrocities, wars of aggression, genocides, and other crimes. They believe his thesis implies that the perpetrators of such acts are off the hook, since "geography made them do it." In answering that charge, Diamond quite correctly distinguishes between understanding why some event occurred and justifying the actions of the people involved. As he notes, "psychologists try to understand the minds of murderers and rapists ... social historians try to understand genocide, and ... physicians try to understand the causes of human disease" (pg. 17). Yet none of them are trying to justify murder, rape, genocide, or disease—indeed, their attempt to understand them is often motivated by the desire to prevent them.
Naturally, no conscientious scholar makes a controversial and sweeping claim, such as attributing to geography the primary causal role in history, without presenting a fair amount of supporting evidence. In fact, the bulk of Diamond's book relates historical events meant to demonstrate the soundness and explanatory scope of his claim.
In order to make his thesis plausible, Diamond must show that there were crucially important geographical differences between the homelands of those societies that wound up as conquerors and those that turned out to be the vanquished. He has exerted tremendous ingenuity in attempting to do so. I believe that he has succeeded to some extent, although it is a much more limited accomplishment than he ambitiously claims to have achieved.
The main historical outcome that Diamond seeks to explain is that the descendants of the people who 13,000 years ago occupied Eurasia1 came to rule over such a large portion of the Earth's inhabitable land. Why wasn't it American Indians or sub-Saharan Africans who colonized Europe, rather than the reverse?
Diamond asserts that it was the combination of the Eurasians' superior technology, and of the various diseases that they carried proving massively lethal to many of the other people they eventually encountered, that led to that outcome. (Thus, the "guns, germs and steel" in the title of the book.) He calls those facts the "proximate causes" of the present Eurasian dominance of the world. But, quite understandably, he is not satisfied to halt his inquiry at that point. Why, he goes on to ask, did Eurasians come to possess better technology than did the inhabitants of the other continents? And why did Europeans carry germs so deadly to American Indians that many Indian nations were wiped out in advance of any direct contact with Europeans (through disease transmission from Indian tribes that did have direct contact with colonists), rather than the Europeans falling in droves to diseases they contracted from Indians?
To answer the first of the above questions, Diamond begins by noting that the pace of technological development in a society depends heavily on its ability to create a food surplus. That enables the emergence of producers who can specialize in craft manufacture, because once a society reaches a situation in which the labor of one person can supply more food than is needed simply to keep him alive, some members of the group do not need to devote themselves to procuring sustenance. And a food surplus generally only comes about once a society learns to deliberately produce its food, rather than relying on finding it naturally produced and then hunting it down, digging it up, or picking it.
Diamond makes a convincing case that, given enough time living in one place, all modern humans (meaning homo sapiens) will tend to figure out how to domesticate any of the indigenous plants and animals that are suitable for agriculture. (For example, there are nine widely separated places on the globe—the Fertile Crescent, China, Mesoamerica, the Andes, the African Sahel, tropical West Africa, Ethiopia, and New Guinea—where food production seems likely to have arisen independently.) However, it happens that Eurasia was blessed with far more species suitable for domestication than any other continent. Of today's major food crops, more originated from there than anywhere else. Of the fourteen mammals over 100 pounds that humans have domesticated, every one of the "major five" (cattle, sheep, goats, pigs, and horses) is Eurasian in origin. Nor is it the case that Eurasians were simply more clever than the residents of other continents at learning how to domesticate the local flora and fauna—despite the fact that they eventually came to occupy every inhabitable continent, and despite all of the advances in technology and the increased understanding of breeding techniques that have taken place in recent centuries, European colonists have domesticated no new species of major agricultural importance in any of the lands they came to conquer.
Diamond also contends that the "east-west axis" of Eurasia, as opposed to the "north-south axis" of the Americas and Africa, made an important contribution to Eurasian's current global dominance. Because Eurasia extends mostly from east to west, it provided a vast area of roughly similar climatic conditions over which a multitude of societies could share agricultural innovations. The result was an enormous, integrated area of agricultural practices and common crops stretching roughly 6000 miles, from Ireland to Japan. In contrast, the crops and the agricultural economy developed in tropical West Africa could not spread south into the Mediterranean climate of South Africa or north into the Sahel. The species domesticated in the Andes never reached central Mexico, nor visa-versa, because they were useless in the intervening tropics of Central America. Although Mexican corn eventually was cultivated in eastern North America, it took millennia to spread to there, because of the two regions' different climates and the arid stretches separating them.
It is not terribly difficult to imagine that an advantage in food production can result, over time, in a technological advantage as well. But how can Diamond account for the diseases that Europeans carried to distant shores being so much deadlier to the locals than the diseases they carried were to Europeans? Quite cleverly, it turns out.
He launches his explanation by suggesting that epidemic, or "crowd," diseases, such as influenza, measles, smallpox, and bubonic plague, cannot easily sustain themselves among small bands of hunter-gatherers. They will tend to wipe out the entire population, which, unfortunately from the point of the microbe causing the disease, wipes the microbe out as well. It is only among large populations of humans, in close contact with other populous groups nearby, that epidemic diseases have a chance to persist over long time-spans, both because the likelihood of a few individuals having natural immunity to the disease is greater, and because the microbe can shift back-and-forth between neighboring populations, surviving dormant in a group that has recently built up immunity to it until it can jump to another, more susceptible neighboring population.
But where do such microbes come from? They can't conjure themselves into existence out of thin air once a sufficiently dense human population emerges. No, Diamond argues, they come from the only place they plausibly could—they are mutations of microbes that evolved to survive amidst dense populations of other mammal species, specifically, among the herd animals, a number of which humans domesticated and came to live with in close quarters. Therefore, agriculture provided the necessary conditions for the survival of epidemic diseases among humans, and animal domestication, especially of herd mammals, supplied the source of microbes able, through natural mutation, to make the relatively small adjustment from being hosted by cows or pigs to being hosted by humans. As a result, when Europeans first encountered American Indians, it was the Europeans, and not the Indians, who carried the deadly crowd diseases.
Diamond illustrates the patterns he has detected with a number of historical examples, including less familiar ones such as the Austronesian expansion, from Taiwan to the Philippines, then to Indonesia and Malaysia, and on westwards to Madagascar and east across the Pacific, eventually reaching Hawaii and Easter Island. Given what we have examined so far, Diamond's work suggests the following, quite sensible ideas:
- When two societies first encounter each other, the one that is more technologically advanced will frequently conquer or absorb the less advanced;
Advances in technology depend heavily on a food surplus, and, therefore, on agriculture;
The degree to which agriculture could be practiced in any location, before the advent of world-wide commerce, depended heavily on what species were locally available for domestication or could be acquired from neighboring cultures sharing a similar climate;
Agriculture and the domestication of herd animals are also prerequisites for the emergence of epidemic diseases among humans; and
Therefore, agricultural, herding societies will carry deadlier germs than will hunter-gatherers or people that farm only plants.
But Diamond is not satisfied with merely having discovered certain factors that frequently have been influential in humanity's past. Instead, he aims to transform the entire discipline of history into a natural science that discovers deductive-nomological2 explanations, one that determines the "ultimate causes" of historical events, rather than mere "proximate causes," such as the actions of people or the ideas that they held. In adopting that grandiose project, Diamond turns what would have been an enlightening and sound exploration of some common historical patterns into a deeply flawed attempt to reform a subject he does not really understand.
For example, in his effort to squeeze the course of real events into his conceptual scheme and thereby demonstrate his "laws," Diamond often has to put a good deal of spin on historical episodes. In attempting to explain why the Vikings did not successfully colonize the New World, while the Spaniards and the Europeans who followed in their wake did, he writes, "Spain, unlike Norway, was rich and populous enough to support exploration and subsidize colonies" (pg. 373). But this declaration simply brushes over the fact that Norway did successfully explore the North Atlantic, and did successfully colonize the Faeroe Islands and Iceland. If Diamond were true to his project of turning history into a deductive-nomological science, he ought to proceed to formulate a quantitative law governing just how far from the mother country a colony can survive, given any particular amount of wealth and any number of residents in the colonizer. However, simply to state that requirement is to expose the attempt to stuff human history into a deductivist framework as the absurdity that it is.
Another instance of forcing the facts to fit the theory is Diamond's "law of history" asserting that agricultural societies will inevitably come to dominate their non-agricultural neighbors. He ignores the multitude of instances where settled farmers were conquered by nomadic horsemen: the Hittite conquest of the ancient Middle East, (possibly) the invasion of Greece by the Dorians, the successive movements of the Celtic and Germanic people across Europe, the Aryan migration into India, the Turkish conquest of much of the Moslem world that began in the 11th century, and the vast Mongolian conquests of the 13th and 14th centuries.
In fact, such examples led both the political theorist Albert Jay Nock and the economist Murray Rothbard to suggest a typical pattern in history nearly the opposite of Diamond's. They hypothesized that states arise when some nomadic people, who have been repeatedly raiding a nearby society of relatively peaceful farmers over an extended period, come to realize that it is more profitable to settle right in the farming community as rulers, enabling them to continually raid the productive population in the form of taxes. (See Nock, 1935, and Rothbard, 1978.)
I don't wish to enter here into disputes as to how the state came to be, or as to whether the pattern noted by Diamond is more or less common than that detected by Nock and Rothbard. I don't contend that such counter-examples make nonsense of Diamond's observations, much less that they demonstrate a "law of history" such as "nomadic horsemen always will defeat settled farmers." But they do show that the complexity of history defies attempts to deduce universal laws from its complex patterns. It is only by "cherry picking" his examples that Diamond can defend his claim that he has found "ultimate causes" in history.
Diamond also glosses over the divergence between his hypothesis that a lead in food production and subsequently other technologies is the "ultimate cause" of one civilization's dominance over another, and the inconvenient fact that the first region to develop agriculture, animal husbandry, and writing was the Fertile Crescent, roughly located in what today is Iraq, Syria, and Turkey. None of those countries are dominant powers in today's geo-political scene. He attempts to explain this anomaly by noting the environmental degradation of the relatively fragile ecology of the Near East due to extensive human exploitation of the area's natural resources, such as the almost complete deforestation of the region that occurred as its residents cut down trees for timber and to clear land for farming. He declares that, as a consequence, "with the Greek conquest of all advanced societies from Greece east to India under Alexander the Great in the late fourth century B.C., power finally made its first shift irrevocably westward" (pg. 410). Diamond fails to explain exactly how, if this shift of power was "irrevocable" and was an inevitable result of human damage to the Near East's ecology, power, as well as the cutting edge of scholarship, shifted back to the Near East during the first six or seven centuries after the fall of Rome.
After all, if Alexander's triumph was merely a "proximate cause" of the waning dominance of southwestern Asian culture, while the "ultimate cause" was environmental, then it should have been impossible for the region to ever regain its former glory. Nevertheless, many centuries after "power" had scurried "irrevocably westward," the territory ruled by the Muslim caliphate exceeded that of the grandest empires of the ancient Near East by perhaps an order of magnitude. Nor is it obvious that Alexander's triumph over the Persian Empire had anything to do with the ecological state of affairs in the Near East – it seems, by truly historical accounts, to have been primarily due to Alexander's brilliance and tenacity as a general. (See Green, 1992, for more on this point.) And it would seem ludicrous to contend that the "ultimate cause" of Alexander's conquests was some environmental advantage held by Macedon, a late-to-develop and resource-poor backwater of the Greek world.
Diamond Does Not Comprehend the True Character of History
I believe that Diamond's desire to transform the practice of history stems chiefly from the fact that he understands neither the nature of the material from which the historian launches his inquiries, nor what the historian's task is in relation to that material. Diamond has reverted to the view of history held by 19th-century positivists, who believed that the historian is presented with a collection of "historical facts," and that his job is to discover the "laws" or "historical forces" that explain those facts.
For example, Diamond declares that, since the "whole modern world has been shaped by lopsided outcomes [in clashes of different cultures] ... they must have inexorable explanations, ones more basic than mere details concerning who happened to win some battle or develop some invention on one occasion a few thousand years ago" (pg. 25). Yet he neither refutes the idea that historical contingency can offer adequate explanations in this regard, nor does he defend his insistence upon "inexorable explanations" of the human past.
Now, despite the recent emphasis in the philosophy of science on how all facts are "theory laden," there is a sense in which it is true that the natural scientist does have the facts to be explained presented to him as a given starting point for his investigations. A certain star just does produce a certain spectral pattern. There may be disagreement as to what the pattern means, or even as to whether it is significant, but there it is. If some astronomer doubts it is so, he can re-create the pattern for himself. Compound A and compound B just do produce a certain amount of heat when combined. The chemist skeptical of the fact as reported can combine them herself and make her own measurement.
But no similar facts are given to the historian. Instead, he is faced with certain artifacts that have survived into the present, and which he takes to be signs of past events that are not present before him, events that it will never be possible to re-create. Nor can the surviving pieces of evidence of past happenings be taken at face value. A text purporting to describe a battle may have been composed to glorify the victor or excuse the loser. A politician's memoirs may have been written with an eye to making him look good to future generations. The inscription on a statue may have been re-inscribed at the behest of a ruler jealous of his illustrious predecessor's accomplishments. The historian is always presented with a collection of initially ambiguous and often, on their face, mutually contradictory pieces of evidence, on the basis of which he attempts to determine what the facts really were. The "facts of history" are not the starting point of his inquiry, but are instead its end product. As Collingwood notes, "The fact that in the second century the legions began to be recruited wholly outside Italy is not immediately given. It is arrived at inferentially by a process of interpreting data according to a complicated system of rules and assumptions" (1946, pg. 133).
To denigrate historical inquiry because it does not mimic the natural sciences in attempting to discover universal laws is to declare that there is no value in simply determining what really happened in humanity's past. Setting aside, for the moment, the question of whether it is even feasible to formulate "laws of history," a question that we will address below, I contend that the effort to discover the historical past is worthwhile in its own right, even if there is another discipline that could discover historical laws. To learn what really occurred in the past is to understand how we came to be where we are today. The knowledge gained through historical inquiry enables us to see how the myriad decisions and actions of our predecessors, the ideas they held, the ideals to which they aspired, the gods they worshipped, and the demons they feared, all combined to create the world in which we find ourselves today.
Lacking an understanding of what real historical research consists of, Diamond winds up doing "scissors and paste" history. His approach fails him in at least the one instance he discusses with which I have the most familiarity: the story of the QWERTY keyboard. He declares "trials conducted in 1932 with an efficiently laid-out keyboard showed that it would let us double our typing speed and reduce our typing effort by 95 percent" (pg. 248). If that were really true, then the fact that no company employing large numbers of typists, and wishing to double their productivity while at the same time making their jobs much easier—surely a profitable move!—chose to break with convention and switch to this efficient keyboard layout is astonishing.
But we can contain our astonishment. It turns out that the study Diamond cites was severely flawed, showing no evidence of using a genuine control group or random sampling to choose participants. Furthermore, it was conducted by none other than August Dvorak, the inventor of the purportedly more efficient keyboard, who, holding the patent to his design, had a large financial stake in proving the superiority of his model. Later, independent studies did not confirm Dvorak's outlandish claims. (See Liebowitz and Margolis, 1996, or my summary of their findings.)
Diamond also periodically employs the long discredited idea that there is a significant division between "human history" and an earlier time, before the invention of writing, called "pre-history." To the contrary, as Collingwood puts it:
"A consequence of the error which regards history as contained ready-made in its sources is the distinction between history and prehistory. From the point of view of this distinction, history is coterminous with written sources, and prehistory with the lack of such sources. It is thought that a reasonably complete and accurate narrative can only be constructed where we possess written documents out of which to construct it, and that where we have none we can only put together a loosely assemblage of vague and ill-founded guesses. This is wholly untrue: written sources have no such monopoly of trustworthiness or informativeness as is here implied, and there are very few types of problems which cannot be solved on the strength of unwritten evidence" (1946, pg. 372).
Diamond opens his book with a question asked of him by Yali, a New Guinean whom the author met while undertaking biological research on the island: Why is it that Europeans have so much more "stuff" than New Guineans? He laments that most professional historians "are no longer even asking the question" (pg. 15). It doesn't seem to occur to him that the reason for that might be that it is not an historical question. If history consists in showing how the occurrence of some unique event in the past is made intelligible by the particular circumstances that led up to it, then it is categorically unable to address such questions as "Why are Europeans generally wealthier than New Guineans?" As Oakeshott says:
"[The] alleged task is to discern [an historical event's] 'true' character by coming to understand it as an example of the operation of a 'law of history' or a 'law of historical change.' In order to perform this task [the historian] must equip himself with such a 'law' or 'laws.' And he is said to do this in a procedure of examining (and perhaps comparing) a number of such occurrences and situations and coming to perceive them as structures composed of regularities. But this, also, is clearly a mistake: no such conclusion could issue from such a procedure. What this 'historian' needs and what he must devise for himself is a collection of systematically related abstract concepts ... in terms of which to formulate 'laws.' How he may set about this enterprise we need not enquire ... But what is certain is that they cannot be laws of 'history' or 'historical change' because they do not and cannot relate to the circumstantially reported situations he designs to explain, but only to model-situations abstracted from them in terms of these 'laws.' In short, the distinction between such a model-situation (explicated in terms of regularities) and a circumstantially reported situation is not a difference of truth and error; it is an unresolvable categorical distinction" (1983, pp. 81–82).
In the same vein, Mises notes, "The notion of a law of historical change is self-contradictory. History is a sequence of phenomena that are characterized by their singularity. Those features which an event has in common with other events are not historical" (1957, pg. 212).
Diamond does not comprehend the nature of historical inquiry, rendering his attempt to replace what he has failed to understand with his own brand of "scientific history" badly misguided. Nevertheless, I believe that he quite usefully has described a number of common patterns in human affairs. The economist Tony Lawson calls such patterns "demi-regs," by which he means "a partial event regularity which prima facie indicates the occasional, but less than universal, actualization of a mechanism or tendency, over a definite region of time-space" (1997, pg. 204).
But Diamond fails to realize the contingent nature of all such regularities in the social world. As Lawson notes,
"in the social realm, indeed, there will usually be a potentially very large number of countervailing factors [to any particular cause] acting at any one time and/or sporadically over time, and possibly each with varying strength.... [And] the mechanisms or processes which are being identified are themselves likely to be unstable to a degree over time and space.... Indeed, given the fact of the dependence of social mechanisms upon inherently transformative human agency, where human beings choose their courses of action (and so could always have acted otherwise), strict constancy seems a quite unlikely eventuality" (1997, pp. 218–19).
One of Diamond's chief motivations in writing the book under review seems to have been to discredit racial explanations of the course of history. However, if he had comprehended the true character of historical explanation, he would have seen that he was battling a chimera. Race can no more substitute for genuine historical understanding than can geography. How could it possibly explain the concrete particularities of history, when the past presents us with Germans as different as Johann Goethe and Adolf Hitler, Jews as dissimilar as Karl Marx and Ludwig von Mises, Irishmen as far apart as James Joyce and Gerry Adams, Chinese as divergent as Lao Tsu and Mao Tse Tung, blacks like George Washington Carver and Idi Amin, and so on.
Mises categorized the type of history Diamond proposes as "environmentalism." He said of it, "The truth contained in environmentalism is the cognition that every individual lives at a definite epoch in a definite geographical space and acts under the conditions determined by this environment." But, he goes on to note the flaw inherent in all attempts to regard the environment as the "ultimate cause" of historical events: "The environment determines the situation but not the response. To the same situation different modes of reacting are thinkable and feasible. Which one the actors choose depends on their individuality" (1957, pg. 326).
Diamond, I believe, has discovered some very interesting "demi-regularities" in the human past. But he has not realized that, quite apart from the search for such demi-regs, there is a different and quite legitimate discipline called history that concerns itself with discovering the particular antecedents of some unique going-on that explain its occurrence, based on critically analyzing artifacts from the past that have survived into the historian's present.
As I mentioned in the introduction, Diamond's mistake is not merely of concern to scholars. The view that "vast, impersonal forces" largely determine the course of history, whether those forces are taken to be "the material conditions of production," as in Marxism, or geographical circumstances, as in Diamond, naturally suggests that individuals can do little to affect their own future. As a logical consequence, in order to improve the lives of those who have been dealt a poor hand by those forces, it seems necessary to counteract them with another vast, impersonal force, namely, the State. Huge international programs intended to redress the arbitrary outcomes brought about by historical forces are recommended. The cases of countries with few geographic advantages but relatively free economies, such as Japan, prospering, and those of nations blessed with natural resources but ruled by highly interventionist governments, for example, Brazil or Nigeria, lagging behind, are easily dismissed as anomalies by those who are convinced that human action plays an insignificant part in history.
While Diamond's book is filled with valuable insights, it is not, as he would like to believe, the first step in the reformation of history along more "scientific" lines, but only another interesting vantage point from which to contemplate humanity's past. Furthermore, the policy implications of his overreach are a danger to both human welfare and freedom.
- 1. Diamond prefers to regard Europe and Asia, with which he includes North Africa, as a single large region, a choice that appears quite reasonable when one considers a map of the world as well as the long and significant cultural connections between European, Asian, and North African cultures.
- 2. The deductive-nomological model of science claims that a genuinely scientific explanation of an event consists in deducing the occurrence of the event from a set of empirical laws and initial conditions. It is also called the covering law model.
Collingwood, R.G. (1946) The Idea of History, Oxford, England: Oxford University Press.
Diamond, J. (1998) Guns, Germs and Steel: A Short History of Everybody for the Last 13,000 Years, London: Vintage.
Green, P. (1992) Alexander of Macedon 356-323 B.C.: A Historical Biography, Berkely, Los Angeles, Oxford: University of California Press.
Lawson, T. (1997) Economics and Reality, London and New York: Routledge.
Liebowitz, S. and S.E. Margolis (1996) "Typing Errors?" Reason Magazine, June issue.
Mises, L. von (1957) Theory and History, Auburn, Alabama: Ludwig von Mises Institute.
Nock, A.J. (1935) Our Enemy the State.
Oakeshott, M. ( 1985) Experience and Its Modes, Cambridge, England: Cambridge University Press.
— (1983) On History, Oxford, England: Basil Blackwell Publisher Limited.
Rothbard, M. (1978) For a New Liberty: The Libertarian Manifesto.