« My pen: R.I.P. | Main | The cold cometh »

October 21, 2005

In defense of reification

All right, y'all, I'm newly full of book-learnin' on intelligence and home sick with a sinus infection, so we're going deep.

Reification is the process of regarding something abstract as something real, concrete, and/or material. We are reifying when we decide that something we are looking at, talking about, or thinking of is real -- it's not just a useful model of how things work or a clever trick of the mind. According to some schools of thought, notably epistemological realism, to talk of anything that is not indisputably concrete as if it is commits a fallacy. For example, someone who says "we are giving some of our democracy to the Iraqi people" is wrongly implying that there is such a thing as democracy, that it's something that can be possessed, that we possess it, and that it can be doled out in part to other people. There are further fallacies that could be asserted. I once held the (rather dogmatic) view that there was no such thing as "society". One could similarly argue that there is really no such thing as "the Iraqi people" in any coherent sense, only groups of disparate individuals. Or consider the Xinhua example "the Chinese people believe that Taiwan is part of China". If we are to reify "the Chinese people" as the official Chinese government press would intend, then the statement makes no sense. Granting Taiwan as part of China makes the Taiwanese people part of "the Chinese people"; thus, there should be no dispute between Taiwan and the mainland Chinese government, but there obviously is.

The anti-reificationist position (I'll call it "AR" for short) has considerable value, especially with respect to political discussions. Clearly, terms like "soccer mom", "blue stater", and "welfare queen" poorly reflect the essential attributes of the people to which they allegedly apply; they are stereotypes, usually coined to serve a political purpose that requires a group of people to be viewed in a narrow and (usually) misleading way. Pointing out this distortion of the truth is a valuable contribution to rational discourse. Nevertheless, despite its utility I have problems with AR.

My first problem with AR is that its use is not so much focused on the form of the argument as the object. A fallacy is generally a statement that is not true on its face, regardless of its subject, because of its form. For example, if I say "Some cats have three legs, therefore my cat has three legs," there's a problem with the proposition even if it is true that some cats have three legs and that group includes my cat. The problem is the "therefore"; it is simply not the case that my cat has three legs because cats have been known to be in a three-leggish way from time to time. AR does not work like this. If I say "cats are spunkier than dogs", you might say via AR that I'm wrongly arguing there's such a thing as spunk, that it can be had, and that it can be quantified and measured. And you'd be right. If I'd said, on the other hand, "monkeys have more hair than people do", then I am not reifying and it's fine. The only thing that distinguishes statements that violate the AR principle is whether there's any question about the reality of the subject. Unfortunately, there is often quite a bit of controversy about whether something is real: consciousness, design in nature, institutional racism... you name the topic, there's some kind of existential argument going on that supercedes AR.

For example, if I say "people have more intelligence than cows", then those on the watch for "speciesism" will take umbrage at my claim of human superiority. But they'll do so, not by saying they disagree with my factual claim (though they obviously do), but by asserting that I'm not actually making a factual claim; I'm just reifying intelligence and my statement is wrong, ipso facto. What's the difference? Well, if you told me you thought I was wrong about this, I could try to convince you that I'm not. I could show you studies, say, that demonstrate humans are unparalleled in their use of tools and symbolic language, and that cows seem to have little to no ability in these things. And I could argue that, whatever intelligence is, those properties are part of it and they can be compared qualitatively at the very least. But if you tell me it is not possible that I am right, then the conversation has to stop there. It's as if you have said intelligence does not exist, and I am either a fool or a blackguard for suggesting that it does. Maybe you really wanted to say "who are you to say what's intelligent? You're just an arrogant bigot justifying his meat-eating lifestyle!" but that would have been ad hominem. Asserting AR is far more refined -- and rhetorically effective.

My second problem with AR is that, by categorically labeling all such arguments as a priori fallacious, it may well do violence to the very notion of what it means to be human. This is where Jeff Hawkins' book On Intelligence enters my rambling prose. He is an engineer by trade, lead designer for such ubiquitous devices as the PalmPilot and cellular phones, and he has always been fascinated by the brain. Hawkins is among those neuroscience theorists who argue for a "connectionist" model of the brain, but he goes further still. He notes that a study has demonstrated that there are neurons in 21st-century American brains that fire only when viewing a picture of Bill Clinton. We can't even get computers to recognize general faces consistently, yet the brain can unambiguously specify Bill Clinton's face. Computers are, in principle, much faster but they haven't got what the brain's got, and Hawkins thinks he knows why:

There are as many if not more feedback connections in visual cortex as there are feedforward connections. For many years most scientists ignored these feedback connections. If your understanding of the brain focused on how the cortex took input, processed it, and then acted on it, you didn't need feedback. All you needed were feedforward connections leading from sensory to motor sections of the cortex. But when you begin to realize that the cortex's core function is to make predictions, then you have to put feedback into the model; the brain has to send information flowing back toward the region that first receives the inputs. Prediction requires a comparison between what is happening and what you expect to happen...

By utilizing a hierarchical structure of prediction-making machines, suggests Hawkins, the brain is constantly forming sets of complex predictions that become what he calls "invariant memories". No one knows what the world is like directly, but only indirectly, through one's senses. Intelligence is, in effect, a nested prediction algorithm making and adjusting predictions of everything: whether something is preceived as a chair, a dog, or Bill Clinton's face depends on what we have predicted in the past about the properties these things should have. If Hawkins is right, then he's not reifying anything by speaking of intelligence: viewed from an engineering standpoint, saying a person is equipped with intelligence is no different than saying my car is equipped with an automatic transmission.

But even if he's not reifying anything, our brains are. It is, in fact, all they do:

People are real, trees are real, my cat is real, the social situations you find yourself in are real. But your understanding of the world and your responses to it are based on predictions coming from your internal model... Throughout this book, you could substitute the word stereotype for invariant memory... without substantially altering the meaning. Prediction by analogy is pretty much the same as judgment by stereotype... If my theory of intelligence is right, we cannot rid people of their propensity to think in stereotypes, because stereotypes are how the cortex works. Stereotyping is an inherent feature of the brain. The way to eliminate the harm caused by stereotypes is to teach our children to recognize false stereotypes...

In summary, reification, far from being a fallacious process, seems to be the only way we can know anything. We should focus on improving our understanding of abstract things rather than denying their reality.

Posted by The Greatness at October 21, 2005 10:30 AM