Saturday, October 22, 2011

A Long History of Something or Other (Part 1): Anecdotes and Anthropology

So scientific popularizer Steven Pinker has written a book on violence. The Better Angels of Our Nature claims that violence has declined consistently over the course of human history. and that we can isolate the various causes of that decline. At 700 pages with wide margins, small type, and many footnotes and graphs, it's the kind of book that some commentators take seriously based only on appearance.  "Look at all the evidence," they'll declare, as if the mere assembly of information in an academic format constitutes meaningful argument. In fact The Better Angels of Our Nature is, on the basis of what I've read so far, a very strange book, lacking the rigor of serious intellectual inquiry and inclined to throw selective data at the reader with insufficient attention to its context and drawbacks. As is often the case with books of this type, a traditional review can't capture the scope of the problem, can't communicate to the reader just how consistently sloppy and simplistic its thinking is. So I've decided to put together an in-depth, multi-part analysis of the book, looking at each major section of Pinker's argument and explaining how, for all the detail he puts forth, he fails to be as comprehensive and conclusive as his tone suggests. This first part will deal with the book's first two chapters.

Perhaps the most salient problem of The Better Angels of Our Nature is that it never attempts to define its central concept. What, in Pinker's view, is violence? This may sound like a stupid question; we all have a pretty good idea of what violence is. But when it comes to serious analysis, the things that we all have a pretty good idea of turn out to be the biggest problems. By failing to define violence clearly and comprehensively, Pinker lets himself have it both ways.  He can pull together a lot of data that aren't obviously connected and use them to argue that the abstraction he calls violence is on the way down, but when it comes to arguing that specific forces cause "violence" to decline, he can present correlations based on whichever measures of violence give the best result. Eventually I'll get into some specific problems caused by this methodological slipperiness.

Chapter 1

Before Pinker offers any hard evidence, though, he devotes his entire first chapter to anecdotal accounts of violence throughout history.  He justifies this as follows:
Scientists often probe their conclusions with a sanity check, a sampling of real-world phenomena to reassure themselves they haven't overlooked some flaw in their methods and wandered into a preposterous conclusion. The vignettes in this chapter are a sanity check on the data to come.
But the point of a sanity check is that it comes after the data, not before it, as a quick justification of scientifically-reached conclusions rather than an attempt to plant those conclusions in the reader's mind before any evidence has been presented.  Pinker's anecdotes, long on extreme and sadistic violence and in no way representative or statistically controlled, have an obvious prejudicial effect on the reader's mindset. Throughout the book's opening chapters, in fact, he revels in elaborate descriptions of particular cruelties, as though given societies should be thought of as more violent because they are sadistically violent. One could make an argument to that effect, but it would depend on a particular theoretical definition of violence, with which Pinker hasn't bothered. I also doubt he'll be regaling us with such elaborate descriptions of the effects of nuclear weapons, mustard gas, or anthrax.

He begins with carefully-selected descriptions of prehistoric corpses that suffered violence, then tendentiously asks “What is it about the ancients that they couldn’t leave us an interesting corpse without resorting to foul play?” Of course there are plenty of archaeological corpses that show no evidence of foul play; it's just that they aren't "interesting" because they don't tend to support Pinker's argument. In fact, one of the corpses he does discuss raises another question about definitions. Lindow Man, a two-millenia old bog body, was possibly a religious ritual sacrifice, in which case it is further possible that he went to his death willingly. Is this sort of consensual behavior violence? If so, what theory of violence can explain at once the motivations behind it and those behind coercive violence? What I'm getting at here is that, by failing to define violence, Pinker fails to come to terms with the complexity of its causes, his discussion of which, as we shall shortly see, is incredibly reductive.

But back to his morbid anecdotes. The next two sections are based not on historical records of any kind, but on Greek and Israelite myth: that is, on the the Homeric epics and the Hebrew Bible. Pinker acknowledges that these are fictions, but slips them in on the claim they reflect the cultures of their times and places. Which, up to a point, they obviously do, but it's also obvious that that process of reflection is not an easily comprehended one. There are a great many more depictions of violence in our media than there are acts of violence, after all. One could actually suggest that fictional accounts of violence resolve violent impulses in actual people, meaning violent art would imply a less violent society. I don't really think that's the case; it's a facile theory. But even considering such things would require Pinker to engage more deeply with the meaning of violence than he evidently cares to.

Part of the reason The Better Angels of Our Nature is so long is not that Pinker provides a great deal of substance, but that he loves a good digression. He spends 4 pages lovingly laying out the depravities of the Hebrew Bible. Among the highlights of this account are a calculation of the homicide rate at the time of Cain and Abel, a garbled version of the Samson story (it omits the actual reason Samson kills the men at his wedding feast, not that that's much better than Pinker's misreading), and the treatment of odd, unexplained stories like the death of Aaron's sons and David's census as if they proved something about human violence rather than about obscure religious tradition. Precisely because it's not a serious argument, none of this exactly matters; I dwell on it because I think it reflects Pinker's cavalier treatment of facts, which doesn't stop when he approaches substantive matters.

He goes on to reveal that Christian martyrdom, medieval knighthood, and the English monarchy were violent, shocking those readers who have never taken an introductory history course. Pinker really seems to believe that most Americans don't think of the distant past as having been at all violent. He fails to grasp that nostalgia about a non-violent past is almost always personal, reaching back to the innocence of childhood, not to the days of the Visigoths. He does get around to discussing violence in those days, the "wholesome" 1950s, but here, with an overview of advertisements and TV shows, we return to the notion that representations of violence are equivalent to actual violence. Pinker treats anecdotes about war, religious strife, violent crime, and media representation as equivalent, as all proving something about a monolithic entity called "violence." The notion that these forms of violence might conflict as much as they cohere, might fulfill different human needs, continues to go unmentioned.

Another small point with larger implications: at one point Pinker writes, without a hint of irony, that where our society once glorified soldiers, nowadays "military men are inconspicuous in public life, with drab uniforms and little prestige among the hoi polloi." I can only assume he's never attended a Fourth of July or Veterans' Day parade, or even seen a soldier in a dress uniform. (They look hot.)  What's happening here, and elsewhere in the book, is that Pinker is projecting his own upper-middle-class liberal values onto the rest of American society. This cuts right to the heart of Pinker's explanation of the decline of violence: those values, he claims, are responsible, and in places where violence persists, it's only because those values are insufficiently appreciated. That it's the presence of violence in those places that enables its absence here, that the lack of recent war among "major" and "developed" nations (which he sickeningly praises at the end of this chapter) is exactly because it's now occurring elsewhere in the world, is another thought of which Pinker is apparently incapable.

Chapter 2

The second chapter begins to lay out Pinker's theory of the decline of violence. This theory, taking in "Six Trends," "Five Inner Demons," "Four Better Angels," and "Five Historical Forces," is so loaded with simplistic numbered lists that one might imagine one had accidentally picked up a self-help title, or worse, a book by Newt Gingrich. Potshots aside, the first trend, something Pinker calls the pacification process, involves the transition from stateless prehistoric societies to early states, which were, he suggests, much less violent, because the state, as in Hobbes' theory of Leviathan, naturally curbs violence.

Before he begins to lay out such evidence as he has for this claim, Pinker gives an account of the origins of and motivations for violence. Devoted as he is to gene-driven evolutionary psychology theories --not that he mentions that these topics remain contested and there are other views on them-- Pinker treats violence as a basically rational process, something people do (or ancient societies did-- again, it's not quite clear whether he's discussing specific aspects of violence or the whole shebang) for coherent reasons related to survival, not because of irrational individual or communal impulses.  Revenge is purely undertaken for deterrent effect, not from emotional causes. This may or may not be a useful model for explaining why violence was common rather than rare in the earliest, least psychologically complex humans, but like all models it needs to be understood as a construct: useful for isolating and discussing variables but not something that exists in that precise form in nature. Having become common for sensible evolutionary reasons, violence has manifestly developed in senseless culturally-specific ways. (It's because of this that one of the sidelines here, a discussion of violence in chimpanzees, is thoroughly irrelevant, though as anecdote it probably primes certain readers' expectations.) In this chapter, however, Pinker is willing to ignore all that to focus on the variable he has decided matters: the presence or absence of a state structure.

The difficulty in comparing violence in state and nonstate societies is that all of our reliable statistics come from the former. To get around this, Pinker pulls in whatever fragmentary data are available on violence in nonstate societies, even if it's not remotely parallel to what we know about states. His charts lay out all these numbers neatly as though they're comparable, but in the text we get some hint of how rough these numbers are: "even if we tripled or quadrupled the estimate to include"... "What if we added"... "Even if we throw in"... "even if we generously multiplied it by twenty." If Pinker were dealing with solid data, these vague attempts at adjustment would be horrifying, but because his numbers measure different things, it barely matters that he's willing to maul them.

His first set of data on violence in nonstate societies comes from archaeological efforts to identify cause of death in corpses from ancient cemeteries. Quite aside from the ambiguities of this process (depending on the methodologies involved, it may or may not take into account non-fatal injury, accident, and animal attack), we are forced to rely on data from the cemeteries that happen to have survived, which are not necessarily representative of all deaths in given societies. It's all important to recognize that Pinker, again crushing difference in pursuit of the state/nonstate dichotomy, treats as equivalent cemeteries spread across the entire world and over at least 12,000 years of history, as if none of those locations might have atypical rates of violence unrelated to the presence or absence of a state. He does average these numbers, but an average of non-representative figures does not somehow become representative itself.

The second nonstate data comes from reasonably contemporary societies where tribal structures happen to have endured. Pinker has been using this data for a while, and his approach to it has come in for some criticism. The gist of it is that at least some of the societies he mentions in The Better Angels of Our Nature (there are a few more mentioned in the book than in the presentation discussed at that link), while they may be nonstate, are not representative of the ancient nonstate hunter-gatherer societies, and therefore don't prove anything about the distant past. In the book Pinker seems vaguely aware of this critique-- he moves some societies into a separate category of "hunter-horticulturists and other tribal groups"-- but, as I've been emphasizing, there's still a basic failure to acknowledge that nonstate societies are not interchangeable, and that high rates of violence in the specific communities for which we have data may not represent a larger trend of all nonstate societies.

Pinker is, of course, not an anthropologist, and a lot of his data and ideas seem to derive from a particular group of anthropologists, whom he treats as the only "scholars with no political ax to grind," as opposed to the "anthropologists of peace," or the "Peace and Harmony Mafia," who believe in less-violent ancient societies for what Pinker chooses to present as purely modern political motives. I have no idea what the facts on ancient violence are, but believing that only one side in a given debate has non-substantive reasons for its interpretations is naive, and the very ferocity with which Pinker dismisses his opponents' political motives becomes, whether he realizes it or not, a political motive of its own. But I don't see much value in open speculation on sources of bad faith, either in Pinker or in his opponents. Instead, I think it's important to return to the observation above that theoretical models can mislead us into projecting them too confidently onto reality. This is especially true in academic communities, where discussion of the nuances of these models can have the unintended effect of making one take their basic assumptions as fact.  Pinker is clearly convinced that his numbers reveal a wide-ranging historical truth, but looked at unto themselves, without the ancient historian's or anthropologist's habit of speculating and generalizing, as one must to produce substantive arguments, they're just a set of discrete facts that don't intersect. They may be, in a loose sense, suggestive, but they can't prove anything.

It might be objected that I, having anarchist political leanings, have special reason to find fault with Pinker's theory of the rise of the state as a reducer of violence, and it would be disingenuous not to admit that I'm reading this book in the first place because I came across a review of it and found his notions about the state, as the review presented them, risible. Having read this part of the book, though, I don't think that even were I to grant Pinker's argument about ancient nonstate violence, I would thereby do damage to contemporary anarchism. Despite Pinker's uncritical treatment of it, as though it expressed a universal truth, Hobbes' Leviathan is the product of a particular intellectual, political, and social context. Precisely because all nonstate societies are not identical across time, the specific pacification process Pinker theorizes in prehistory doesn't require the belief that eventual nonstate (or rather poststate) societies would see an increase in violence. An anarchist future need not be a reprise of the tribal past, although if it's Pinker opponents who are in the right about that past, certain aspects of it might be worth a second look.

No comments:

Post a Comment