1) One thing called “reduction” involves a more general, more encompassing, theory covering what had been seen as disparate phenomena. Rosenberg talks about this. Example: p. 78-81, Kepler, Galileo, Newton. Key features - see p. 81, claims about derivation of laws; moving toward more general, and more fundamental, theories.

            Questions about “the unity of science” and “scientific progress” arise in connection with this.

 

2) Another idea gets less emphasis in Rosenberg, but is mentioned. See top of p. 82. We think we make progress by explaining bigger things in terms of littler things. We explains wholes in terms of parts and their relations. Related to this is the thing Rosenberg mentions a hierarchy - psychology, biology, chemistry, physics. He drops social sciences from the picture - not everyone would agree about that. If we were to include that we’d have this picture:

 

a) Social theories - principles about group behavior - “Countries go to war when ...”

reduce to

b) Psychological theories about individuals and their relations

reduce to

c) Biological theories about cells and their relations

d) Chemical theories about molecules

e) Physical theories about atoms and their parts

 

3) In thinking this way, one seems to be saying that each thing is “nothing but” an organized collection of things at the next lower level. Societies are nothing but organizations of people; people are nothing but organizations of cells; cells are nothing but organizations of molecules; molecules are nothing but organizations of atoms; atoms are nothing but organizations of still littler things.

 

4) We have just mentioned two categories of things that “reduced” to other things: a) individuals (and categories of individuals), e.g., societies are just collections of individuals, and so societies are “reduced to” individuals; b) theories, e.g., psychology “reduces to” biology. It may be useful to note a third category: properties (features, characteristics). Consider a property at the psychological level - being angry, e.g. You might think that being angry reduces to some particular sort of state or process in the brain.

 

5) What is it for one property to reduce to another? a) Not definitional equivalence. The word “anger” is not defined in terms of brain processes. We learn that the reduction holds by empirical discovery, not be analyzing the concept. (Contrast what we did when we talked about “lying”.) Also, “anger” gets its meaning from its role in psychology, not from the underlying biology.

b) Not mere extensional equivalence (applying to the same things). Consider the complex property of being a professor in this room. And consider the complex property of being a tie wearer in this room. They apply to exactly the same people. But there’s no reduction. Also, you can that neither (a) nor (b) is right because those relations are symmetrical and reduction is not.

c) Another possibility brings in the idea of supervenience. This is a kind of dependence. Your psychological properties depend upon your biological properties. Consider a different example from a completely different domain: the quality of an essay depends upon grammar, spelling, organization, etc. The evaluative properties (being a good essay, being a poor essay) depend upon these other properties. If two essays are alike with respect to those properties, then their evaluative properties are alike. Similarly, if two people are alike in the relevant ways in their brains, then they are both angry (if reductionism is correct). So,

 

Properties from theory T2 supervene on properties from theory T1 iff necessarily, if 2 things are alike with respect to their T1 properties, then they are alike with respect to their T2 properties.

 

Note: i) Supervenience is not causation; ii) it does not imply symmetry - there can be different lower level ways to get the same higher level property. Think about the evaluative properties of essays: two essays with different lower level properties might be equally good.

 

6) Reduction of theories - an overly simple view. A theory includes a bunch of laws. Pretend there is a law in a higher level theory, e.g., psych. that says that anyone who has psych. property 1 also has psych. property 2. Eg., anyone who is angry is snarling (or growling, or gnashing teeth, ...) [Don’t worry that this isn’t really true.] So there is a higher level (HL) law:

            Anything that has HL property 1 has HL property 2

The overly simplistic reductive idea is that HL1 reduces to LL1 and HL2 reduces to LL2. (Go through example.) So there is a lower level (LL) law

            Anything that has LL property 1 has LL property 2

 

The simplistic picture says that the properties are equivalent - (x)(HL1x iff LL1x) . These interlevel laws are called bridge laws since they bridge the levels.

 

7) Remember the point about different essays that are equally good. Maybe there is more than one way to be angry also.“Martian” anger. Maybe Dick Cheney’s brain works differently. This illustrates multiple realization. This messes up the simple view about reduction of theories. No equivalence. And here’s the important result: maybe there is no simple law at the lower level. This is not to give up the idea that in some sense everything depends upon the lower level. Maybe there can be an enormously complicated laws - well beyond our capacity to ever formulate - that captures the higher level law.

 

8) Further, the higher level laws may be very useful ways to organize things. Computer example: here’s a higher level law: When you press the “K” key, a “K” appears on the screen. The exact pattern of the “K” on the screen can vary, and the underlying physical process - the electrons running around inside the computer - can vary. And all of this does not matter for the purpose of “theories” about using your computer. Similarly, if you are interested in theories involving angry people, the underlying stuff may not matter so much. (Of course, if you care about drug treatments you may want to know.) So the idea is that the higher level theory may contain useful information that would not be usefully contained in the lower level theory, even though it remains true that all the higher level stuff depends on the lower level. We have not abandoned the metaphysical supervenience picture. This leaves open the possibility that there are reductions of more narrowly circumscribed higher level properties - pressing “K” on Dell 6000, anger in people, Martian anger, or anger in Dick Cheney.

 

9) But there’s a further complication about the status of higher level laws. The picture is that there are higher level regularities that in effect collect together enormously complicated lower level regularities. The further complication is that things might not be quite so smooth. There will always be potential glitches and complications at the lower level to falsify any universal generalizations at the higher level. The computer example makes this easy to see.

 

10) These last points cloud the issues about reduction of theories. If supervenience is enough for reduction, then we don’t have to abandon the reductionist picture. But if reduction requires more - laws at the lower levels that we can formulate corresponding to the higher level laws - then maybe it isn’t the right picture. Two reasons for this: a) the multiple realization possibility suggests that the lower levels might be enormously complex, with lots (maybe infinite number) of disjuncts (explain); b) Glitches and complications at the lower level will be reflected in the lower level laws, but not in the higher level laws. This suggests that there may not be an exact correspondence between the higher level laws and the lower level laws.

 

11) Some people have resisted reductionism for different reasons. They think something that perhaps is captured by the saying, “The whole is more than the sum of its parts.” They say that some properties are “emergent”. These properties of the wholes which are not determined by the properties of the parts. Two proposed such properties: life, consciousness. This is a very obscure doctrine. I don’t think that it is denying the supervenience claim.

 

12) There seems to be a difference between reduction and replacement. That’s Rosenberg’s point on p. 83. But this can get very confusing. Suppose we find lots and lots of different underlying states associated with anger. We might find ourselves saying that there is no such thing, there are just the different underlying states. This idea will come again later on.

 

13) Back to a point about the reduction of entities. To the extent that reduction succeeds, we end up in a somewhat puzzling situation. See last ¶ of section II., p. 84. This suggests that it would be good to think some more about observations, theoretical entities, and the questions about the support for theoretical claims.