Thursday, July 13, 2017

Kim's Principle of Explanatory Exclusion

Jaegwon Kim has argued that in order for there to be a workable account of mental causation, reductionism has to be true. According to his principle of explanatory exclusion: 

An event cannot have two separate and complete* explanations.
Take any human behavioral event M (A person decides to change seats, comes to understand a principle of physics,feels sorry for her little sister, etc.) For every M, there can be only one complete explanation. There cannot be two explanations which
a). individually provide a complete explanation of M, and
b). are unconnected to each other.
*An explanation is complete if the events or properties that it specifies are the only ones that need to be mentioned in order to fully explain the occurrence of that event.

Explained here. 

14 comments:

William said...

Perhaps Kim would agree that neither the objective (scientific measurable data) nor the subjective (our psychological intentional reasons) types of explanatory pictures are truly complete when it comes to human cognitive behavior? Thus, there is no over determination since each explanatory strategy by itself under-determines what we do.

In fact I'd take that a step farther and suggest that maybe that even the combination of intentional and scientific explanations under-determines many human behaviors.

Unknown said...

"An explanation is complete if the events or properties that it specifies are the only ones that need to be mentioned in order to fully explain the occurrence of that event."

Is this definition of complete somewhat circular? It says it's complete if no other facts are needed to fully explain it? But what counts as "fully" explained? Would a full explanation just be a complete explanation? If so, then this is not a very informative definition.

Maybe by full Kim means "total" in the sense that every possible relevant fact is included. So consider the explanation for the fact that it is raining . You could give one explanation that involves a complete account of all the relevant meteorological phenomena involved. Or you could give an explanation that says "given the laws of physics and the initial state of the universe, the statue of the universe would necessarily evolve in such a manner to lead to the exact configuration of atoms we have at this moment. Both explanations are correct, but are they complete? If they are complete, then Kim's principle is false because we'd have two complete true explanations. But if they are not complete then this means that no explanation could ever be complete unless it contains ALL the information down to every last detail, including the behavior of each individual atom over the history of the universe. In that case, we could NEVER offer a complete explanation for anything unless we were omniscient

StardustyPsyche said...

Ron said...
" In that case, we could NEVER offer a complete explanation for anything unless we were omniscient"
--Right, we can't. Humans use simplified models, idealizations, approximations, and probability estimates to function. That is all we can do.


July 13, 2017 2:30 PM

Victor Reppert said...

However, the combination of the theses of mechanism, causal closure, and supervenience postulate the existence of a true complete naturalistic explanation for everything, even if we cannot provide it. And the principle of explanatory exclusion implies that certain other explanations are excluded, such as mentalistic explanations. Naturalists are happy to use PEX to get rid of creation given evolution, or even Thor's hammer when an electrochemical analysis of lightning is provided. But when it comes to the mental and the physical, they are not so eager to apply it.

Unknown said...

Victor,

Are naturalists really using PEX to get rid of creation given evolution? I think theyre typically employing the weaker principle of Okham's razor. They're not (in my experience) arguing that evolution rules out creation because there can't be two true explanations at work. Rather, they are appealing to simplicity - of evolution works, don't add redundant hypotheses on top of it

Unknown said...

Victor,

How would you define "full explanation"?

How about this: explanation E is a full explanation of the event X if P(X|E)=1.0?

Victor Reppert said...

But isn't PEx just an expression of Ockham's Razor? That is just what the argument from reason is saying. If the physical is closed, and we give all the physical explanations for the event, all other explanations are redundant.

Meteorological explanations are macro-izatons of physical explanations. If you add up the physical states, the meteorological states appear to be entailed. There are mereological summatiopns of the physical, With the mental, you are not just going from micro to macro, you are going from nonintentional to intentional, from nonnormative to normative, from third-person to first-person, and from nonpurposive to purposive.

Unknown said...

I see a difference between Ockham's Razor and PEX. PEX says the additional explanation must be false. O.R. says the conjunction of the two explanation is to be discounted as less probable than the one explanation on its own. Applied to the evolution/creation example, PEX provides deductive grounds for rejecting as *impossible* the creation explanation (provided evolution is true). On the other hand, O.R. gives us a merely abductive grounds for favoring unguided evolution as more probable than evolution+creation. This is a nontrivial difference. So I don't think naturalists do (or at least don't need) to invoke PEX to argue that evolution eliminates the need for creationism

Victor Reppert said...

But the application of the two principles seems the same. And Kim realizes that there are cases of overdetermination. Thus, if scientists conclude that the asteroids caused the death of the dinosaurs, it could still be that they were dying anyway for some other reason. Nevertheless, science quits looking when it finds the explanations.

Unknown said...

How can the application be the same if the result in two very different conclusions? Application of the PEX results in the conclusion: "creationism must be false with deductive certainty." Application of ockhams razor results in the conclusion: "creationism+evolution is less probable than evolution without creationism. Nonetheless, crestionism could still be true."

Ian Thompson said...

You link does not work.
Maybe you meant to point here.
If so, then you can read a spectacular misunderstanding about dualism.

At the end of that document, we read "NOTE: Interactionist dualism (Cartesian dualism) presumes that for any mental event one can specify a complete causal explanation using only mental terms."

Looks like they terribly misunderstand the meaning of "interactionist". Interactions mean that minds can perceive nature as well as change nature. If perception is possible, then the 'complete causal explanation' for a mental event of perception MUST use physical as well as mental terms. How could Stephen O'Sullivant not see that?

William said...

Ian, thanks for the link. I agree with you that the complete causal picture for ordinary human mental events needs to include both explanations. Your link's page says:

----------------------------
"There are two possible stories.
1. One story explains M by appealing to prior mental events.
2. Another story explains M by appealing to brain (i.e. purely physical) events.

In 2, there is no need to appeal to mental events or mental categories – we simply talk about cellular activity, starting from the time M takes place and going backwards in time, listing the purely physical events that precede one another, leading up to the behavior.

Now according to the principle of “explanatory exclusion”, there cannot be two complete explanations of any event that are causally independent of one another. So #1 and #2, presumably explanations of M, cannot both be complete explanations and unrelated to each other.

-----------------

My issue is that I think both story 1 and story 2 are incomplete by themselves, and that additionally they are related. Thus (speaking from a sort of neutral monist point of view, where both the physical and the mental are features of an underlying reality which need not be either) I would deny one of Kim's assumptions about physicality.

Ian Thompson said...

Which philosophies have "both story 1 and story 2 are incomplete by themselves, and that additionally they are related."?

William suggests a neutral monist point of view gives this.
But the same conclusion follows from interactionist dualism.

Victor Reppert said...

I don't think Kim's principle requires deductive certainty. Events can be overdetermined. It could be the case that the asteroids killed the dinos, but something else (maybe excessive cigarette smoking) was killing them at the same time. But overdetermination is not normally the case.