Friday, 17 September 2010
::: What's a SOU worth?
The philosophical analysis of scientific explanation has been a Pittsburgh mainstay. Decades ago, when I came to Pitt, the great heros the age of explanation, Carl "Peter" Hempel and Wes Salmon were walking the halls and standing in the elevators. Recent hiring in both Departments of Philosophy and of HPS continue the tradition.
One of the foundational tenets of this literature is that explanation is not something merely psychological. While it may give us a warm feeling of understanding, that feeling is an incidental by-product of something more substantial. Explanations are deductive relations among sentences in theories; or perhaps probabilistic relations among them; or perhaps something else along these lines. With an eager but novice student, an early order of business for Wes Salmon was to disabuse them of the illusion that explanation is merely psychological. It was a thought that became unspeakable and then unthinkable.
J. D. Trout is speaking today. It is early in a term over which we expect to see a lot of him. Last year he declared his interest in visiting routinely through the term, making trips from his home in Chicago. It was a welcome declaration and we were eager to make it official by appointing him as a Visiting Scholar.
His talk today is on explanation and it is taking us right back to the unspeakable, unthinkable foundational tenet. "Why Fluent Explanations Feel Truth-y" is his announced title. After we are settled in our seats, J. D. explains his project. It will be largely descriptive: how do we use explain? He will be centrally interested in how the sense of understanding (SOU) arises that makes us feel explanations are delivering us truths.
He will be urging that the comfortable fluency of a good explanation is not "diagnostic of the truth," a little piece of unfamiliar jargon that flagged a mind steeped in the medical-psychlogical literature. The bigger project, he announced, is to understand how explanation can have an epistemic role in scientific progress if its truthiness is misleading.
Yes, indeed, I thought as he said this--I'm very interested to hear how this rather delicate matter will be handled.
The talk proceeded with the developing of two interwoven lines of analysis. The first was clearly informed by J. D.'s background in philosophy and cognitive science. His PhD is in both combined and he is now Professor of Philosophy and Psychology at Loyola University in Chicago. He began to analyze just what it is in our psychology that makes certain explanations feel "truth-y."
The narrative tied together ideas that have been subject to psychological investigation: the notion of fluency, how it relates to familiarity and the sense of understanding. All this was peppered with the inflated jargon of the psychologists. "Positively hedonically marked," J. D. pointed out with a mischievous twinkle, just means "feels good."
The results of the psychological investigations were both fascinating and alarming. Tell a group of subjects this: the idea that Obama is a muslim is just plain false. They'll remember it, for a while. But eventually a sizable portion of them will remember only the first part--that Obama is a muslim. That idea will then become familiar and start to feel truthy. Our efforts to debunk a myth have furthered the myth. Oh, how elusive truth is in all our thinking!
There was much more. Practice makes pleasant, it has been found. The more familiar is the more fluent; and mere familiarity makes us judge ideas more beautiful.
The second line of his narrative was to recapitulate how our past use of explanation in science has been deceptive. Old explanations that had all the right characteristic of fluency and SOU proved wrong. He began a litany of examples that were either funny or awkward, depending on just how much sympathy you have for older theoretical systems.
Copernicus, in introducing the heliocentric system, exulted on the appropriateness of an illuminating sun in the center of our system. Aristotle believed the gender of offspring depended on which way the wind was blowing; and that twinning derived from the richness of pastures. Descartes thought birthmarks arose when a pregnant mother looked at the sun. This is an HPS'ish audience and I could sense the discomfort.
We were coming perilously close to the end of the talk and the moment I'd waited for had still not come. How are we to explain the epistemic role of explanation in scientific progress given this account of the widespread failure of explanation to connect to the truth?
In the closing minutes, the answer came. In the seventeenth century, the errors of the earlier theories were largely corrected. We were now in possession of theories that could provide correct explanations. So when we employed explanations in those theories and a SOU developed, it would correspond to the truth of the explanation.
The talk was over and I could see that J. D. had pushed many buttons. Are our modern explanations so bad, the philosophically inclined were wondering. And were they so bad then, the historically minded added to this unspoken mood. Ioannis was eager to ask the first question, so eager that he wanted to neglect our five minute donut and bagel break. When we reconvened, the questions came quickly and urgently.
Sandy was concerned that there might be circumstances in which the fluency of an explanation might well be a good indicator of truth. She developed the example of cultural learning. Peter Machamer was worried by what was worrying me: why concentrate so much on feelings, when the issue is truth?
At the end, taking the next but last question, I tried to wrap it up with this thought. Is J. D. allowing the psychology of explanation to replace the philosophy of explanation? I expected that he would deny it, but the thought seemed not to bother him at all. Was I presuming that psychological analysis cannot be normative? It can be, he reassured me, drawing on examples of the psychological investigation of other forms of reasoning.
I'll admit I was not satisfied. I think I'd like to see the psychological investigation supplement the philosophical and not replace it. I can see that reading in the literature on psychology can be tremendously illuminating for philosophers. I've been amazed at how judgments of beauty can invert. Einstein's general theory of relativity was judged hideously complicated in the 1910s. By the 1970s, the definitive textbook declared Newton's theory monstrous in comparison to the beauty of Einstein's theory. Beauty comes with familiarity. It might just be that simple psychological mechanism at play. It is hard to see anything but psychology explaining this inversion.
However there are real dangers if we allow the psychologists' analysis to replace the philosophers'. For I've found their judgment can be lacking when they decide just what is good and bad thinking. Perhaps the clearest example of the dangers was when J. D. paraded his examples of ridiculous failures of explanation from the past theories. The trouble is that the examples are only ridiculous if you have no historical perspective and demand that a good explanation has to be a true explanation. It is a simple minded standard likely to lead us away from what matters. Many of us could see that the examples being ridiculed might well be perfectly cogent explanations by other standards, but merely unfamiliar to us since they were drawn from an unfamiliar theoretical context. I think it was Benny Goldberg who had interjected amid the recounting of one: "that's just the law of similarity at play." The psychological literature seems hardly poised to tell us what those other standards might be.
I'm guessing that outside of the speedy exchanges at the end of question time, J.D. will have more to say. There's a whole term to hear it. I look forward to it.