« Now Yezidi, now you don't?* | Gene Expression Front Page | Dawkins on race »
September 25, 2004

Up from ignorance II

My post Up from ignorance elicited a lot of good responses.

First, I would like to acknowledge that my reaction (and formulation of a "solution") was framed by the commentor on Winds of Change, and my perception of his character or intellectual life, or, more properly, the hint of a character that I perceive to be endemic among thoughtful Americans. In Innumeracy John Allen Paulos notes that while many intelligent people can imagine themselves 'curling up' to a biography of Samuel Johnson, a similar picture where one is deeply pondering a multivariable calculus text before a night's rest seems implausible. There are reasons for this, math is hard. In The Number Sense Stanislas Dehaene shows that though humans have an intuitive sense of 'numerosity,' it is an analog facility. The implication is clear that abstract higher level mathematics requires complex cooperation between various cognitive domains in a somewhat unnatural fashion. Robin Dunbar makes a similar argument in The Trouble With Science about the more inductive but still abstraction filled (and counterintuitive) disciplines of natural science.

While there seems to be a niche for 'science popularizers,' who serve as the intermediaries between the scientists and the public (think Richard Dawkins channeling W.D. Hamilton), I don't get the impression that history has the same class of individuals, as professional historians can distill and reformulate their prose and substance to satisfy a general audience with relative ease. After all, it is called history! Humans love a good yarn, and they don't demand fidelity to the methods elaborated by Popper (or Carnap or Feyerabend or Kuhn if you prefer). That is the great temptation, to tell a good story, a tale that titillates the reader's need for entertainment. Of course, popularizers of science can entertain, but they are constrained by the narrow lands which science traverses, and so must make generous use of metaphor and scientific biography so that the nutritious sliver of technique and result can be interspersed into the social and stylistic narrative that acts as sugar and savor.

I think the relative ease with which history passes through our intellectual system (this is a generalization) makes it easy for a certain class of Americans to believe they are "well read" or "learned" in the human sciences even if they neglect more explicitly natural scientific formulations of the human condition. Because history is filled with n 'interpretations' they can simply choose ni that matches their own normative preferences. For some, Bernard Lewis is a salve for their own preconceptions about a "Clash of Civilizations," the mountains of erudition and deftly phrased general analysis cocooning their self-confidence in the "correctness" of their viewpoint.

But, all this said, I obviously was wrongheaded in the way I phrased my entry if I gave the impression that people should always start from basic truths and work their way up. The commenators were correct in that people should shift back and forth, shedding light throughout the spectrum of organization and complexity. What I truly find disconcerting is that many people simply remain at the extreme end of data richness and complexity and never veer back toward technical rigor and modular simplicity. Clearly, often knowledge progresses by:
Data -> Induction -> Hypothesis -> Test/Observation.

The elemental portion is preceded by a period of data collection, which reading narratives of history can fit easily under. It is in the stages of induction & hypothesis formulation that I think the problems crop up. You see a pattern, how do you explain the pattern? I believe too often thinkers not impacted by affinal areas of science are unconstrained in their imagination. There are real physiological and psychological limits and boundaries that are imposed on the human condition. Logic and analysis offer myriad possibilities, almost infinite, so how do you choose between the various plausible options? For example, what about the view that before the coming of Kurgan culture Europe was dominated by matriarchies? I suspect that this is a garbling of Marija Gimbutas' already speculative and tenditious theories, but, it is a view I have noted as implicit in the works of some feminist authors. I think one can look to anthropology, psychology and the biologically oriented subdisciplines within each as to why this "historical question" is a nonquestion. This option is already highly improbable and should be left off the table for consideration and the data should be interpreted in a different manner.

But I think that the possibilities offered by the human sciences are even greater, not only negative, but positive. Recently I have posted on the importance of 150 individuals as the upper bound of human social intelligence. A Roman military historian might find a fair amount of archival evidence where generals note the importance of centurions in winning the loyalities of the troops. What does one make of this? If the historian was aware of evolutionary psychology the importance of the centurions is clear in that they personally know both the generals and the vast numbers of rankers. The data and the model point in the same direction! But without the Rule-of-150 in mind who knows what direction a scholar might go. Additionally, the Rule-of-150 offers the hope of predictive models for what type of bureaucracy is less susceptible to the inevitable process of institutional decay and collapse.

Human intellect can form innumerable models. There have been many hypotheses of how the solar system (or what we now call the solar system) was arranged. Most of them displayed an internal coherence and even fit by rough aproximation the data on hand, but only one of the models was validated by the findings of science over the long haul. Reading Spengler or Toynbee exposes one to a wealth of erudition and an enormous capacity for analysis, but in the end the richness is more entertaining than insightful (Spengler's "organismic" conception of civilizations seems to take group selection too far). Toynbee even crosses over into the world of "alternative history" or counterfactual.

In the end what I would like to see is a genuine algebra of history. I don't know if this is possible, but right now it is very difficult to precisely comprehend what someone means when they say "significant" or "minor influence." Adding some numbers like the Rule-of-150 might be a first step, at least we know the precise number (within a narrow range) of this atomic unit. The size difference between the mean size of a male and the mean size of a female is "significant," but we also know the absolute and relative values. Every scholar has their own internal definition for these sorts of proportional statements, exactitude here and there will allow the audience to calibrate the terms effectively.

Calibration. Precision. Parsimony. Concurrent validation. Building blocks for theoretical models. Those were my real points!

Posted by razib at 03:19 AM