In “The Naturalistic Fallacy” Frankena makes the claim that the real issue behind the so-called naturalistic fallacy is not “a logical or quasi-logical fallacy” (344). He quotes from Moore’s article in his attempt to characterize the naturalistic fallacy. Basically, according to Frankena (and here I think that he gets Moore right) the naturalistic fallacy is when one equates a natural object (like pleasure) with a non-natural object (good) (347). But what is at the heart of the fallacy is not the fact that one is a natural object and one is a non-natural object as Moore asserts that even if good were a natural object the fallacy would not be diffused (320, 347). The only difference would be that it would no longer be appropriate to call it the naturalistic fallacy. But the name, for Moore, is unimportant. If we try to equate two natural objects, orange (as in the fruit) and yellow, we err in the same way. Frankena clarifies this point by saying that the naturalistic fallacy is in fact a species of a moral general fallacy, the definist fallacy (347). The definist fallacy is, according to Frankena, “the process of confusing or identifying two properties, of defining one property by another, or of substituting one property for another” (347). This seems to be an accurate characterization of what is driving Moore’s orange/yellow example. The problem, for Moore, is that when one says that good is pleasure it is tantamount to one saying an orange is yellow, meaning that orange means nothing other than yellow (320).

Frankena goes on to say that “it is an error to construe a… synthetic proposition as a definition”, implying that a definition must be analytic (348). Furthermore, he asserts that to claim that those who say that good is pleasure are guilty of construing a synthetic proposition as a definition is to beg the question. Those who say that good is pleasure take this statement to be analytic. From these assertions two questions arise: First, must a definition be analytic? Second, is it really an error to attempt to identify good with pleasure, for example?

It seems to me that a definition need not be analytic. Consider the following statement: “Pegasus is a winged horse”. This statement appears to be a definition in the sense that a winged horse is precisely what we mean when we say Pegasus. Likewise, being a horse and being winged are necessary and sufficient conditions for being Pegasus. (For the sake of the example I am assuming that “Pegasus” does mean simply a winged horse, not something more specific like the winged horse of myth that did such and such.) But if “Pegasus” simply means “winged horse” then the statement “Pegasus is a winged horse is no more meaningful than “Pegasus is Pegasus”. For something to be a meaningful definition (have explanatory power), it has to add something. If a definition “adds something” in this way it seems that it may no longer count as an analytic statement. (I am not really all that sure about this but I think that the idea behind my explanation is at least plausible. Maybe my intuitions about the analytic/synthetic distinction are screwed up.)

Frankena believes that the only way that a proposition can be a definition is if it is analytic but he thinks that people like Moore are too quick in ruling out statements like “good is pleasure” as definitions because these propositions may in fact be analytic. Therefore, Frankena suggests that there may be nothing wrong with saying that good is definable as pleasure. I agree that there may be nothing wrong with saying “good is pleasure”; however, I suspect that this is true for a different reason. I think that “good is pleasure” is a synthetic statement. It is not meaningless like “good is good” but it explicates the nature of the good. If I am correct that definitions need not be analytic, then the fact that “good is pleasure” may be a synthetic statement does not rule it out as a definition.

-- AnnieS - 05 Dec 2005