Steven Pinker Versus Complexity

The evolutionary psychologist, Steven Pinker has published a new book: The Better Angels of Our Nature: Why Violence has Declined. I came across a review of the book in the Dec/Jan 2012 edition (Volume 18, Issue 4) of Book Forum, which, by the way, I thoroughly enjoy. It is a great periodical. CLICK HERE TO SEE THE REVIEW

I decided to quickly blog on the book because, given that this is a blog about complexity, I think Pinker’s book presents an interesting idea that works along with my previous posting about Andrew Wilson's website on complexity and psychology.

In a very ridiculous nutshell, Pinker's basic argument is that humans have progressed to a place of less violence, thanks in large measure to cultural and social forces impinging upon the better half of our evolved nature; "the better angles of our nature," as Lincoln famously stated.

Pinker's study has two historical foci: (1) the longue durée of human history and, more specifically, (2) the last 20 years of globalization-induced history.

Such a provocative thesis often requires careful review and critique. So, it is not a matter of Pinker simply being right or wrong. (As a side note, "Pinker, Right or Wrong?" seems to be the nature of most debates I have seen on the book, with little thinking in the grey area of, "hey, just how well does the model fit?")

As a complexity scientist, I am primarily interested in his notion that the world-wide decrease in violence has a scale-free character to it. Pinker argues that, from macro-level wars to micro-level views on spanking children, the world has become a less violent place both over the longue durée of human history and, more specifically, over the last 20 years of globalization-induced history.

I think further exploration of this argument is a great dissertation or study for a complexity scholar in history, anthropology, epidemiology or applied statistics to examine.

A note on method: Pinker is clear that his focus is physical violence, not violence as a metaphor for political, cultural or economic oppression. Violence as physical violence. His method of analysis is rather simplistic: rates (ratios expressed over time), computed as a basic prevalence--people harmed by violence divided by the total population. There are lots of epidemiological issues that emerge when one thinks of this approach, but we will confine ourselves to just two:

One of the immediate issues that emerges is that, as the world gets larger, large-scale violent events such as wars, by definition, decrease in their rate of harm. If there are several billion people on the planet and a world war emerges where several million people are killed, this comes across as not as bad as the Roman empire killing people in a much smaller world. Is it true, then, that, at a macro-level scale, we live in a less violent world? What, for example, if we used a network analysis approach and looked at degrees of separation. Even in a world of several billion people, are humans less separated from violence than they were 2 thousand years ago? Also, is there, perhaps, some sort of tipping point here, where the world, past a certain population threshold, becomes too large at the macro-level for people to inflict an increasing rate of violence?

And, as another issue, what about regional scale--here I am thinking of path dependency issues in terms of different complex socio-political systems bounded by particular geographies? Should one's focus be broken down into regions? For example, if one just studied Europe, would Pinker's thesis hold--particularly in terms of his argument that smart governments, not too corrupt and reasonably democratic lead to less violence? Or, is it that, at smaller levels of scale, more democratic government leads to less daily violence in the criminal justice system, people-to-people interactions, discriminatory violence, etc. But, at larger scales, particularly country to country, violence through war has not decreased? It is true that over the last 20 years macro-level violence has decreased. But, I am just not sure what to make of that phenomenon. Anyway, these are the sorts of questions that emerged in my head as I worked through Pinker's ideas.

Bottom line: I think Pinker has an interesting thesis, but I think a lot more work needs to be done before it is embraced. In particular, I think his topic is far too complex to be analyzed in terms of simple rates. It needs to be grasped in complex systems terms and truly examined for its scale-free character and regional context. My initial response is that Pinker's findings are more scale-dependent and context-sensitive than they initially seem. But, without conducting a study, it is nothing more than a conjecture on my part.


  1. Belated thanks for this post. I'm reading the book now and am grappling with exactly these issues. For instance, can we even talk about rates of murder per 100,000 in populations limited to 1000s?

    I think Pinker makes an important point - there is less violence now than the rhetoric implies but is so ham-fisted about it, his credibility is undermined. So I would definitely welcome more exploration of the data along the lines you suggest.

  2. I agree! if you have any ideas when you finish the book, let me know.