Conceptual Holism Concept sharing in conceptualrole semantics Conceptualrole

Conceptual Holism Concept sharing in conceptual-role semantics

Conceptual-role Semantics • The content of a concept is (partially) determined by the role the concept plays in inference. • Alternatively: The content of a concept is (partially) determined by its inferential connections to other concepts.

PLATYPUS Your concept PLATYPUS might have an inferential connection to BILL, MAMAL, LAYER-OF-EGGS etc. Each inferential connection is (loosely) a disposition to draw an inference from the application of one concept to the application of another concept. (e. g. From PLATYPUS(fred) infer MAMAL(fred)) The conceptual-role of PLATYPUS consists of all of these inferential connections taken together.

Holism If CRS is right then concepts are holistic. A property P is holistic if 1. For P to be instantiated once it must be instantiated many times. 2. How P is instantiated in one case is dependent upon how P is instantiated in other cases. Example: Being a natural number. 1. If anything is a number then a whole bunch of things are numbers. 2. Which number x is depends, in part, on which numbers all the other objects are.

Holism According to conceptual role semantics an isolated concept cannot have content. Since a concept C's content is determined by the inferential connections it has with other concepts, there must be other concepts for C to have a content. For C's conceptual role to be interesting many concepts and many connections are necessary. If these other concepts are themselves to have interesting conceptual roles even more concepts and connections are necessary. The upshot is that if an agent has any interesting concepts, she likely has a whole bunch of them. Furthermore, exactly which conceptual role a concept has, depends upon the conceptual roles of the other concepts. Conceptual role is a holistic property.

PLATYPUS • The content of PLATYPUS is partially determined by its connection to BILL. • If you remove that connection (or connect it to a different concept) then the content of PLATYPUS changes. • Changing BILL changes PLATYPUS. If our concepts are massively connected, as they presumably are if we have interesting concepts, then changing the content of one concept can create a ripple effect which changes the content of all of our concepts.

Concept Sharing Intuitively it seems obvious that people sometimes share concepts. We all have a concept PENCIL and it is the same concept. It is also widely accepted that it is vital for psychology that people sometimes share concepts. (This is a point over which Fodor does much foaming at the mouth. ) A good theory of concepts then, should allow that people can sometimes share concepts. Question: Is this possible if concepts are holistic.

Concept Sharing in CRS Sally and Tim have two concepts S-DUCK and T-DUCK with slightly different conceptual roles (Tim thinks ducks can’t fly). S-DUCK is connected to S-BILL which is connected to S-PLATYPUS. T-DUCK is connected to T-BILL which is connected to T-PLATYPUS. Since Sally and Tim don’t have the same duck concept, they also don’t have the same platypus concept. That is, S-PLATYPUS and TPLATYPUS do not have the same conceptual roles. More generally, if conceptual content is massively holistic, then if two agents differ in even one concept, they likely differ in all their concepts.

Concept Sharing in CRS The problem worsens: After talking to Sally, Tim learns that Ducks fly and he modifies the conceptual role of T-DUCK to incorporate this information. T-DUCK is connected to T-BILL which is connected to T-PLATYPUS. But T-DUCK isn’t really T-DUCK anymore. So T-BILL is connected to a new concept which means T-PLATYPUS is connected to a new concept. The result: Any time an agent changes one concept all of her concepts are changed.

To sum up, it appears we have the following argument. 1. Intuition and psychology require concept sharing. 2. CRS implies holism of conceptual content. 3. Holism of conceptual content implies concepts are rarely shared across subjects or through time within subjects. CRS is false. Corollary: Josh’s dissertation is a waste of time.

In order to satisfy our intuitions and to make use of psychology, we need to be able to say sometimes that C is the same concept as C*. The conclusions that is generally drawn at this point is that there should be a simple precise set of conditions that tell us when two concepts are the same. Those conditions should satisfy our intuitions and aid in the development of psychological generalizations. But why think that there is one simple precise set of conditions that tell us when two concepts are the same? This doesn’t seem to be the case for anything else in life.

Car Sharing Consider two situations in which we want to ask whether one car is the same as another. Situation 1: You are looking for parts at a junk yard. Situation 2: Your car was stolen and you are looking for you car at a junk yard. The notion of sameness you will use in situation 1 is very different from the notion of sameness you will use in situation 2. Concepts are like cars. Whether two concepts are the same depends on what we are interested in.

Two-Factor Theories of content Block, Rey, and others have suggested that there are two determinants of conceptual content. Narrow content is internal and is supposed to say what it is that I have in common with my Twin-Earth twin. Wide content is external and is supposed to explain how my concepts hook up to the world. It also would say what is different between my concepts and those of my twin. A complete theory of conceptual content would explain both narrow and wide content.

If we adopt a two-factor theory of content, then conceptual-role semantics is a theory of narrow content. Wide content would be handled by a causal theory or a telosemantic theory. . . If we accept a distinction between narrow and wide content then there are (at least) two ways that one concept can be the same as another: 1. They have the same narrow content. 2. They have the same wide content.

The Solution: Part 1 Although narrow content is holistic, wide content is not. Concept sharing begins by sharing wide content. So, Sally and Tim have the same concept PLATYPUS because the wide content of S-PLATYPUS is the same as the wide content TPLATYPUS.

The Disappointment I have said that Sally and Tim share PLATYPUS because the wide content of their concepts is the same. It seems right to say that their concepts are hooked up to the world in the same way, but it also seems like they use the concept the same way. In other words, Sally and Tim seem to have the same narrow content too. CRS is still leaving something out if it can’t explain this.

The Solution: Part 2 Local Conceptual Role Two concepts A and B have the same local conceptual role just in case 1. ) A and B have the same wide content and 2. ) for every inferential connection A C_a, there is an inferential connection B C_b such that C_a and C_b have the same wide content.

PLATYPUS Sally and Tim don’t have the same global conceptual role for their platypus concepts, but they do have the same local conceptual role. S-PLATYPUS and T-PLATYPUS have the same wide content. S-PLATYPUS and T-PLATYPUS are connected to S-BILL and TBILL S-BILL and T-BILL both have the same wide content. The result is a much more robust notion of concept sharing than merely wide content.