Phys Rev Letters meets Homo ideologicus

According to Wikipedia, Physical Review Letters’ “focus is rapid dissemination of significant, or notable, results of fundamental research on all topics related to all fields of physics. This is accomplished by rapid publication of short reports, called ‘Letters'”. That’s what I always thought, so I was somewhat taken aback to find an article in last week’s issue entitled, “Encouraging Moderation: Clues from a Simple Model of Ideological Conflict.” Unfortunately, you can’t see the whole thing without a subscription, but here’s the abstract:

Some of the most pivotal moments in intellectual history occur when a new ideology sweeps through a society, supplanting an established system of beliefs in a rapid revolution of thought. Yet in many cases the new ideology is as extreme as the old. Why is it then that moderate positions so rarely prevail? Here, in the context of a simple model of opinion spreading, we test seven plausible strategies for deradicalizing a society and find that only one of them significantly expands the moderate subpopulation without risking its extinction in the process.

That’s physics?! Not according to any of the definitions in my ancient copy of Webster’s Dictionary.  Evidently some new ones have cropped up since it was published, and nobody bothered to inform me.  In any case, tossing in this kind of stuff doesn’t exactly enhance the integrity of the field.  If you don’t have access to the paper, I would not encourage you to visit your local university campus to have a look.  I doubt the effort would be worth it.

Where should I start?  In the first place, the authors simply assume that “moderate” is to be conflated with “good”, without bothering to offer a coherent definition of “moderate.”  In the context of U.S. politics, for example, the term is practically useless.  People with an ideological ax to grind tend to consider themselves “moderate,” and their opponents “extreme.”  Conservatives refer to the mildest of their opponents as “extreme left wing,” and liberals refer to the most milque-toast of their opponents as “ultra right wing.”  Consider, for example, a post about the Muhammad film flap that just appeared on a website with the moniker, “The Moderate Voice.”  I don’t doubt that it might be termed “moderate” in the academic milieu from which papers such as the one we are discussing usually emanate, but it wouldn’t pass the smell test as such among mainstream conservatives, and has already been dismissed in those quarters as the fumings of the raving extremist hacks of the left.  Back in the 30’s, it was a commonplace and decidedly “moderate” opinion among the authors who contributed articles to The New Republic, the American Mercury, the Atlantic, and the other prestigious intellectual journals of the day was that capitalism was breathing its last, and should be replaced with a socialist system of one stripe or another as soon as possible.  Obviously, what passes as “moderate” isn’t constant, even over relatively short times.  Is the Tea Party Movement moderate?  Certainly not as far as most university professors are concerned, but decidedly so among mainstream conservatives.

According to the authors, the types of ideological swings they refer to occur in science as well as politics.  One wonders what “moderation” would look like in such cases.  Perhaps the textbooks would inform us that only half the species on earth evolved, and God created the rest, or that, while oxygen is necessary to keep a fire burning, phlogiston is necessary to start one, or that only the most visible stars are imbedded in a crystal ball surrounding the earth known as the “firmament,” while the other half are actually many light years away.

Undeterred by such considerations, the authors created a simple mathematical model that is supposed to reflect the dynamics of ideological change.  Just as the economic models are all infallible for predicting the behavior of Homo economicus, it is similarly effective at predicting the behavior of what one might call Homo ideologicus.  As for Homo sapiens, not so much.  There is no attempt whatsoever to incorporate even the most elementary aspects of human nature in the model.  It is inhabited by “speakers” and “listeners,” who are identified as either AB, the inhabitants of the moderate middle ground, or A and B, the extemists on either side of it.  For good measure, there is also an Ac, inhabited by “committed” and intransigent followers of A.  The subpopulations in these groups are, in turn, labeled nA, nB, nAB, and p.  Only moderate listeners can be converted to one of the extremes, and vice versa, although we are reliably informed that, for example, the Nazis found some of their most fertile recruiting grounds among the Communists at the opposite extreme, and certainly not just among German moderates.  With the assumptions noted above, and setting aside trivialities such as units of measure, the authors come up with “dynamic equations” such as,

nA = (p + nA)nAB – nAnB
nB = nBnAB – (p + nA)n

There are variations, complete with parameters to account for “stubbornness” and “evangelism.”  There are any number of counterintuitive assumptions implicit in the models, such as that all speakers are equally effective at convincing others to change sides, opinions about given issues are held independently of opinions about other issues, although this is almost never the case among people who care about the issues one way or the other, that a metric for deciding what is the moderate “good” and what the extreme “evil” will always be available to the philosopher kings who apply the models, etc.  The models were tested on “real social networks,” and (surprise, surprise) the curves derived from a judicious choice of nA, nB, etc., were in nice agreement with predictions.

According to the authors,

Since we present no formal evidence that the dynamics of (the equations noted above) do actually occur in practice, our work could alternatively be viewed as posing this model and its subsequent generalizations as interesting in their own right.

While I heartily concur with the first part of the sentence, I suggest that the model and its subsequent generalizations might be of more enduring interest to sociologists than physicists.  Perhaps the editors of Phys Rev Letters and their reviewers will consider that possibility the next time a similar paper is submitted, and kindly direct the authors to a more appropriate journal.