Dedication: To Bill Clinton who taught us that much can ride on what the meaning of 'is' is.
The Opponent has a very good post in which he raises the question whether the standard analytic distinction between the 'is' of identity and the 'is' of predication is but fallout from an antecedent decision to adhere to an absolute distinction between names and predicates according to which no name is a predicate and no predicate is a name. If the distinction is absolute, as Gottlob Frege and his epigoni maintain, then names cannot occur in predicate position, and a distinction between the two uses of 'is' is the consequence. But what if no such absolute distinction is made? Could one then dispense with the standard analytic distinction between the two uses of 'is'? Or are there reasons independent of Frege's function-argument analysis of propositions for upholding the distinction between the two uses?
To illustrate the putative distinction, consider
1. George Orwell is Eric Blair
2. George Orwell is famous.
Both sentences feature a token of 'is.' Now ask yourself: is 'is' functioning in the same way in both sentences? The standard analytic line is that 'is' functions differently in the two sentences. In (1) it expresses (numerical) identity; in (2) it expresses predication. Identity, among other features, is symmetrical; predication is not. That suffices to distinguish the two uses of 'is.' 'Famous' is predicable of Orwell, but Orwell is not predicable of 'famous.' But if Blair is Orwell, then Orwell is Blair.
Now it is clear, I think, that if one begins with the absolute name-predicate distinction, then the other distinction is also required. For if 'Eric Blair' in (1) cannot be construed as a predicate, then surely the 'is' in (1) does not express predication. The question I am raising, however, is whether the distinction between the two uses of 'is' arises ONLY IF one distinguishes absolutely and categorially between names and predicates.
Fred Sommers seems to think so. The Opponent follows him in this. Referencing the example 'The morning star is Venus,' Sommers writes, "Clearly it is only after one has adopted the syntax that prohibits the predication of proper names that one is forced to read 'a is b' dyadically and to see in it a sign of identity." (The Logic of Natural Language, Oxford 1982, p. 121, emphasis added) The contemporary reader will of course wonder how else 'a is b' could be read if it is not read as expressing a dyadic relation between a and b. How the devil could the 'is' in 'a is b' be read as a sign of predication?
The question can be put like this. Can we justify a distinction between the 'is' of identity and the 'is' of predication even if we do not make an absolute distinction between names (object words) and predicates (concept words)? I think we can.
Is it not obvious that if an individual has a property, then it is not identical to that property? Tom is hypertensive. But it would be absurd to say that Tom is identical to this property. This is so whether you think of properties as universals or as particulars (tropes). Suppose the property of being hypertensive (H-ness) is a universal and that Tom's brother Sal is also hypertensive. It follows that they share this property. So if Tom = H-ness, and Sal = H-ness, then, by the transitivity and symmetry of identity, Tom = Sal, which is absurd.
If properties are tropes, we also get an absurdity. On a trope bundle theory, Tom is a bundle of tropes. But surely Tom cannot be identical to one of his tropes, his H-trope. On a trope substratum theory, tropes are like Aristotelian accidents inhering in a substance. But surely no substance is identical to one of its accidents.
So whether properties are universals or tropes, we cannot sensibly think of an individual's having a property in terms of identity with that property. If H-ness is a universal, then we would speak of Tom's instantiating H-ness, where this relation is obviously asymmetrical and for this reason and others distinct from identity.
Now 'H' is a predicate whereas 'H-ness' is a name. But nothing stops us from parsing 'Tom is hypertensive' as 'Tom instantiates hypertensiveness.' This shows that we can uphold the distinction between the 'is' of identity and the 'is' of predication with a two-name theory of predication, and thus without making Frege's absolute distinction between names and predicates. It appears that Sommers is mistaken in his claim that "Clearly it is only after one has adopted the syntax that prohibits the predication of proper names that one is forced to read 'a is b' dyadically and to see in it a sign of identity."
I am assuming of course that we cannot eke by on predicates alone: we need properties. By my lights this should not be controversial in the least. My nominalist Opponent will demur. In 'Orwell is famous' he seems to be wanting to say that 'Orwell' and 'famous' refer to the same thing. But what could that mean?
First of all, 'Orwell' and 'famous' do not have the same extension: there are many famous people, but only one Orwell. 'Orwell is famous' is true. What makes it true? Presumably the fact that 'Orwell' and 'famous' denote one and the same individual. And which individual is that? Why, it's Orwell! But Orwell might not have been famous. Since it is contingent that Orwell is famous, but noncontingent that Orwell is Orwell, the truth-maker of 'Orwell is famous' cannot be Orwell alone. It has has to be the fact of Orwell's being famous, which fact involves the property of being famous in addition to Orwell.
Nominalists insist that we ought not multiply entities beyond necessity. They are right! But there is no multiplication beyond necessity here since we need to admit properties as features of extralinguistic reality. To explain why 'Orwell is famous' is contingent, one must distinguish Orwell from his contingently possessed properties. Man does not live or think truly by predicates alone.