[From Bill Powers (2008.10.09.1041 MDT)]
Martin Taylor 2008.10.09.10.21 --
I agree with both you and Jeff Vancouver. Always try to improve the old constructs, but keep a strainer over the drain.
The problem with prescientific constructs is that they are primarily subjective and descriptive: "depression" feels like being pressed down, but of course nothing is pressing down on the person except metaphorically. We have to examine all these traditional constructs with one question in mind: what is wrong with a person who presents these appearances and reports these experiences? If you say a person "has" "schizophrenia," what, exactly, is it that this person does and experiences differently from other people -- and what is wrong with doing and experiencing those things? If a person is "anxious," don't ask what the persion is anxious about, because that assumes that there is a definition of anxiety, which there isn't: there are synonyms, but no definitions. Ask what it is that this person feels and thinks and tries to do unsuccessfully.
In addition to this, remember that a construct of the traditional kind is a dead end as far as explanations are concerned. If you say that a person's problem is that of having low self-esteem, you're saying that low self-esteem is the explanation of the problem. But what is causing the low self-esteem? Is this just a condition you catch, like a cold, or an inherited problem like sickle-cell anemia? Or. much more likely, is it a manifestation of a process that can be traced deeper and farther back in time? The same goes for the current fads in finding neurotransmitters that, because of excesses or deficiencies, cause psychological or physiologial problems. A real explanation has to be traced back to the actual independent variable; you have to ask what is making the level of neurotransmitter too high or too low. And I suspect that if you trace far enough, you will quite often find, instead of an independent variable, one of the variables whose state you were trying to explain. You'll find that the causal chain is a closed loop.
There seems to be a new movement afoot called "transdiagnostic processes", which makes the radical proposal that what goes wrong with people might not be confined to the traditional boundaries of "disorders." In effect, this is a recognition that traditional categories are (in the absence of any fundamental theories) essentially arbitrary and ideosyncratic to the diagnoser. If we were to find that a person who is depressed has some of the same disabilities as a person with schizophrenia and a person with anxiety, that would raise the suspicion that these three categories actually slice through several of the dimensions in which the real problems exist. This would make traditional assessments look similar to trying to explain why some cars ran fast and other ran slowly in a race by analyzing their colors. Yes, blue cars run demonstrably slower than green ones, or perhaps the other way around, p < 0.05, but that is completely irrelevant as an explanation. See Martin's example of explaining why cars bunch up on a road, if you didn't notice the stop light.
So while I can agree with Jeff and Martin that we must be careful not to discard useful aspects of tranditional constructs, we should not get lazy or defensive about them just because they seem important. Other ways of looking at problems may turn out to be much more important.
Best,
Bill P.