The centerpiece of what should eventually become my thesis is the Analytic Imperative. I want to argue that there in an imperative operating in science – an implicit ought which structures scientific methodology. That imperative is (you guessed it) to explain things analytically. By that I mean the imperative is to explain things in isolation from their context, and as a member of some category. That is as opposed to being holistic, which is just the opposite – treating a thing to be explained as a unique individual, situated in some particular context.
This analytic/holistic thing has been with me since my early undergraduate work, and I’ve had a variety of opinions on the topic. What I think the distinction means changes every time I write on it, and it never fails to get me excited. I see it practically everywhere, structuring methodologies, and shaping investigations. For instance, I’m doing some reading on the Modern Evolutionary Synthesis that came together in the 1940s now, and I can’t help but see the Analytic Imperative shaping the possible routes of investigation available to American scientists like T.H. Morgan. They had to assume, in order to be analytic, that the units of inheritance were particulate, having regular phenotypic effects, and that their interactions could be ignored. That is, they had to assume that genes had some kind of context-insensitive properties.
On the other side of the pond, Russian scientists were assuming the exact opposite. Because they were all officially dialectical materialists, which at the time apparently meant
holism in all things, the Mendelian theory of particulate inheritance was regarded as a bourgeois metaphysical fiction. Lysenkoism was preferred, which is the Lamarckian view that evolution mostly functions by the inheritance of acquired characteristics. This is a more holistic way of thinking, in the sense that any individual is entirely the product of their environment, their context. To explain the properties of any organism, you would look to their relations with their surroundings. But on the American, Mendelian view of heredity, to explain the properties of an organism you only need cite the intrinsic causal powers of their genes.
As I said, I see this analytic/holistic stuff operating bloody everywhere. The modern synthesis is just today’s example. A few other favourites include Merleau-Ponty’s brilliant critique of Behaviourism in The Structure of Behavior. His argument is more or less that science is stuck in an analytic mode, but that such a mode is inappropriate for dealing with the nervous system, at least in terms of stimulus and response. He writes:
To explain nerve functioning [on the classic scientific view] can only be to reduce the complex to the simple, to discover the constant elements of which behavior is constituted. Thus one would decompose the stimulus as well as the reaction until one encountered the “elementary processes” composed of a stimulus and a response which were always associated in experience. […] In principle, to each part of the stimulus there should correspond a part of the reaction.”(p.11)
However, he argues that very frequently, this cannot be achieved. Parts of stimuli modify and modulate each other in complex and unpredictable ways. I think no one would doubt this today. What that means it that there can be no science relating stimuli to responses – we cannot be analytic about that.
When we’re unable to be analytic, a couple of options present themselves. If you want to remain in a scientific mode, your best bet is to find a new set of parts to think about. Instead of studying stimuli and responses, psychologists now talk about personality types, neurochemistry and cognitive biases. These seem to be working out better for them as categories. Or, if no good categories present themselves, you could go holistic. I briefly touched on this strategy here, when I talked about what to do with messy complexity. You can deal with particular individuals in particular contexts, tracing out their relationships with their environments, seeing how the parts and embedded in the whole.
The nice thing about the holistic strategy is that it’s always available. There’s no problem so complex and ugly that you can’t be holistic about it, whereas there are lots of situations in which the Analytic Imperative can be frustrated. The problem with it is that you loose the ability to make projectable generalizations. The nice thing, the really nice thing about scientific knowledge is that it helps you make predictions about cases you haven’t even seen yet. If you know everything there is to know about one electron, you’re in a pretty good position to make generalizations about the rest of them. But if you’re being holistic, you have to treat every individual as unique. They all have unique contexts, and so each is different. Think of a historian who treated as wars as basically the same, or all revolutions – it’s just not reasonable to expect generalizations to hold.
The attitude I’ve come to hold is that there is, in fact, something special about the analytic strategy. It’s what modern Western culture has contributed to the world – though we aren’t the only ones to use it, we certainly have brought it to it’s full flower over the last few centuries. Of course it isn’t the only cognitive strategy available to us, and any person who tried to get by without being holistic would be more or less buggered. Noam Chomsky once told Richard Lewontin that he hated the word “dialectical”, because when you got right down to it it meant something like ‘reasoning properly’. I think that captures it perfectly. Being reasonable means knowing which strategy is appropriate to a given situation. But at the same time, we shouldn’t be shy about privileging the analytic strategy in a scientific context. It’s fundamental to the scientific method, I believe, and the success of that method shows that it’s a good strategy.