Goodman’s (1995) goal is to show that there cannot be a purely formal theory-of-confirmation. He thinks that confirmation and induction works differently than what logical-empiricism has thought.
Any argument with the following form is deductively valid, and as long as the terms we substitute pick out definite properties / classes of object, and retain the same meaning throughout the argument, the argument will be valid.
- All F’s are G’s
- X is F
- Therefore, X is G
Based on this, the deductive validity of arguments only depends on the form or pattern of the argument, not the content. logical-empiricism attempts to incorporate this feature of deductive arguments into their theory of induction and confirmation that Goodman aims to show is impossible.
The Problem
Legitimate form of inductive argument:
All the many E’s observed, in diverse circumstances, prior to 2050, have been G. Therefore most likely, all E’s are G.
Consider =
[! Argument 1]
- All the many emeralds observed, in diverse circumstances, prior to the year 2050, have been green.
- Therefore most likely all emeralds are green.
[! Argument 2]
- All the many emeralds observed, in diverse circumstances, prior to 2050, have been grue.
- Therefore most likely, all emeralds are grue.
- Where, grue = an object is grue if and only if it was first observed before 2050 and is green, or if it was not first observed before 2050 A.D. and is blue.
Note in Argument 2, any object observed before 2050 that is green will be grue.
Also both arguments 1 and 2 follow the same pattern as the inductive form given above, even though argument 2 appears to be a bad argument while argument 1 appears to be a good argument. Goodman’s point is that two arguments could have the same form, but one can be better than the other. Hence, it is not the form that makes an inductive argument good or bad, and consequently, there can be no purely formal theory of induction and confirmation.
Proposed solutions to the problem
A proposed solution to this problem is that there is something wrong with “grue”, and that a good theory of induction should propose a restriction on the terms that occur in inductive arguments.
One problem with grue is that it’s definition includes a reference to time.
- Goodman’s response = whether or not a term is defined in this way depends on
The problem with this argument =
- intuitively = the inductive argument is giving us conflicting predictions after the year 2050 for emeralds first observed after this date
- grue illegitimately introduces time (year 2050)
- Natural kinds = grue is a problem because it does not define a REAL property (colors, masses, shapes, molecular structure, etc. are real properties)
- how do we know what the natural kinds are? if we can’t identify natural kinds, we cannot distinguish good inductive arguments from bad ones
Skeptical conclusion of this argument = inductive arguments are not good arguments