Steven Pinker on Analogy
[ by Charles Cameron — importance of analogy as an under-developed cognitive skill ]
.
There was a interview with five prominent “science writers” in the Guardian a few days back, titled Science writing: how do you make complex issues accessible and readable? and one of the writers, Steven Pinker, makes two highly interesting observations:
**
There are a couple of things going on here that I’d like to note. One is that without intending to do so specifically, he is in essence formulating a view about a possible, central difference between scientific and religious thinking, since what he says about the humanities in general applies with great specificity to religion and the arts: in both religion and art, the expansive nature of “symbolism” is a key to the experience.
And that in turn prompts me to suggest that perhaps both the arts and religion are geared towards provoking, evoking or invoking an experience — whereas the sciences are geared towards obtaining an understanding.
I’ll have to think about that, and come to some sort of understanding of my own — perhaps expressed via symbolic means.
**
My second point of interest is that there’s an analogy to be made between Pinker’s two remarks: each of them has a form I could portray thus in terms of cause :: effect —
science : humanities :: simplicity : complexity
Nobody present — the interviewer, Pinker himself, and four other very bright science writers — picked up on the close correspondence between those two statements at the time. And I find that very interesting.
I find it very interesting because the six of them were more interested in seeing what they could say (of what they already thought) than in saying what they could see (in light of the ongoing, immediate conversation).
I think we all tend to do that — which is why David Bohm‘s approach to dialogue is so important: if brings us to speak more into the moment as it surrounds us, not quite so much from the past as it has informed us.
**
Then there’s the interesting fact that Pinker’s sense of the difference between modes of thought in the humanities and the sciences as expressed in the top quote translates so directly to the difference between uses of analogy in the second — and his fairly emphatic statement:
one could argue that we understand everything except for the physical world of falling objects by analogy.
Analogy is the central device in our mental toolkit, and yet we know far more about trains of logic than we do about analogical leaps. We know so little, in fact, that distinguishing between “literary metaphor” and “scientific analogy” (both of which are based in the recognition of resemblance) on the basis of one looking for multiple, rich connectivity and the other for a single tight connection is something noteworthy enough for Pinker to bother to point it out. It is indeed a provocative and perhaps essential insight. But it is also pretty basic — dividing a field up into significant chunks, the way anthropology got divided into “cultural”, “archaeological”, “linguistic” and “physical anthropology”…
It’s time we learned to understand and use analogic with the same rigor we’ve applied to learning and using logic — and Sembl is just the tool for this.
**
Experience wants to be rich: factual understanding wants to be clear.
December 7th, 2012 at 5:16 am
Superb post Charles
.
Science looks for the simplest expressions in part because they are seeking the discovery of laws whose expression can be quantified and stated mathematically – a legacy of Descartes and Newton. Prior to the scientific revolution, science was Aristotelian and therefore empirical but descriptive, just like the humanities are interpretive and descriptive. I wonder how much of a role descriptiveness actually plays in science today – more than I think many scientists would care to admit, not as “proof” (“Anecdotal!”) but as cognitive catalysts for moments of insight.
.
Then again, I am not a scientist. Maybe some will weigh in.
December 7th, 2012 at 5:24 am
Pinker is also raising the point regarding the fractionating, isolating, analytical reductionist nature of scientific investigation. Good scientific experimental conditions where you can rule out the influence of extraneous variables are required for validity but they are sometimes, a highly artificial and “unnatural” environment compared to messy physical reality
December 7th, 2012 at 3:11 pm
Charles, as usual, you raise far too many issues to be easily addressed. Further, I might address each one of them in multiple ways, and I think that would be a scientific approach, although it goes to complexity rather than simplicity.
.
It happens that I am engaged in a project now that has raised an issue very like one of the ones in this post. I have collected a fair amount of information and interview material that I now hope to write up for Nuclear Diner. My biggest problem is how to focus the material, which parts I want to emphasize.
.
All of us who blog or write anything know that problem: we have a finite number of words, a finite amount of time, and we must focus. Or we have a point we want to make. But that’s not the issue in common with Charles’s post.
.
First, let me say that Pinker is overgeneralizing. He probably knows he is, and it’s necessary to make the point he’s trying to make. I say this because I am going to say something that will very likely be termed overgeneralizing. I probably am being too defensive, but it seems to me that I have been getting (and seeing) an awful lot of response on the internets that has more to do with Charles’s observation than any positive contribution to understanding.
.
I find it very interesting because the six of them were more interested in seeing what they could say (of what they already thought) than in saying what they could see (in light of the ongoing, immediate conversation).
.
The issue: simplification versus complexity. In my project, my interviewee feels that science has not been done well in a particular case because there was a jump to simplification too early in the process. Here, too, there are multiple pathways I can go down, and I continue to face that problem of finite time. Let me try this one: It is often necessary to collect data/observations for quite some time before regularities emerge. Then there are many ways to test those regularities. Sometimes the regularities cannot be isolated from a complex system, so tests (experiments) must be devised to hold some or most of the variables constant so that their effects may be separated.
.
That’s very general – partly, I don’t want to say much about the project because I haven’t found a way to line up its complexity in a way that allows posts or articles to be written. But partly, I’m trying to simplify down to general principles. Now let me be more specific.
.
Historically, science has been done by observation followed by discovering and testing regularities. Even physics started out that way; in fact, the sciences were not separated from each other early on as they are now. As those regularities were mathematized into laws, they became tools for finding more regularities. The subject matter of physics was the easiest to mathematize. Skipping ‘way forward in time, physics showed a big success of those tools by developing highly destructive weapons that ended a dreadful war. So physics and physicists were allowed, by a mostly grateful world, to revel in their claim as the most developed and highest practice of science.
.
But that leaves out half or more of science, the observational part, the part that actually pushes matter around into the configurations the physicists need to make those weapons work. Had the Manhattan Project been made up entirely of physicists, we might be speaking Japanese now.
.
Chemistry and biology have long had to separate variables in complex systems without tearing those systems apart. So chemists and biologists are better at this sort of thing than physicists, who all too often want to start from their mathematics. When all the tools you have are hammers… Psychologists have so many variables to deal with in systems that will never be torn apart that they argue really a lot over experimental design.
.
There’s another way I have of looking at this: top-down and bottom-up. Thermodynamics is top-down. Kinetics is bottom-up. Most big models, like climate models, must work with both and must make them mesh at the in-between. This can be really, really hard. And the models don’t even have to be that big: I’ve dealt with this for a lab-scale chemical reactor. Modelers have figured out ways to do this. Physicists believe that physics is the queen of the sciences because, in theory, it has all the tools to make a model of the universe from the bottom up. But when you want real results that approximate a real world in finite time, you have to work top-down and bottom-up.
.
Since the physicists’ great triumph, there has been a great desire by many other scientists to emulate the physicists’ methods. Some of the results of this have been good. But we don’t know enough, or have enough computational power, to deal with biology from the first principles of physics, let alone psychology. The original methods of science need to be recalled and respected.
December 8th, 2012 at 12:06 am
Cheryl nails it — Pinker’s over-generalizing (and cramming us into the “Reductionist” bucket that Zen referenced in his comment).
A *good* scientist will embrace his humility and not ever suggest he has found “The” answer. Instead, we will see to express theories in terms of — wait for it — analogies.
Sure, we can assign variables to define relationships. F = ma simply means that our definition for force is a function of mass and acceleration, the latter of which is another derived dimension (distance per unit time per unit time).
If I can express a complex scientific phenomenon in a way that relates to multiple domains, I think that’s a “Win”.
Too many scientists become self-impressed with their own command of their domain, no matter how narrow. We should all strive to be understood by as wide a variety of our peers as possible.
December 10th, 2012 at 7:40 pm
Zen, Cheryl, Deichmans:
.
Just a quick note to say I’m not ignoring you guys, I very much appreciate your comments. I do have a response in the works, but have been under the weather and not able to do much thinking, let alone writing. Hope to post here in the next day or three…
December 11th, 2012 at 2:21 am
No worries Charles! Get healthy, the blog is always here 😉
December 27th, 2012 at 5:16 pm
[…] zenpundit.com var featureBoxVar = […]