Wolfgang Schwarz

Semantic Properties and Bearerless Names

Linguistic expressions have all kinds of properties. In other words, they can be alike in all kinds of ways. For example, two sentences (of a particular language) can be alike in that

1. they have the same truth value
2. they attribute the same property to the same object
3. they are necessarily equivalent
4. they are a priori equivalent
5. they are such that noone who understands them could regard one as false and the other as true
6. they are cognitively processed in the same way in all speakers of the language
7. they invoke the same mental images in all speakers
8. they invoke the same mental images in some particular speaker
9. they have the same use in the community
10. they are verified by the same observations
11. they are constructed in the same way out of constituents that are alike in one way or another

and so on. All these properties are, I believe, worth investigating into, and all of them might be called "semantic".

In many cases, the semantic properties of sentences are somehow determined by or influenced by corresponding properties of the sentences' constituents. Or, at any rate, it would often make for a nice theory if this were true, so that we might want to postulate properties of constituents to satisfy this role.

Frege, for instance, postulates Sinne and Bedeutungen for all syntactic constituents of sentences. The Sinne are meant to determine something (though not quite) like property 5 in the above list, the Bedeutungen property 1.

It is, however, not entirely obvious that we need different semantic properties of constituents to generate all the different properties of sentences. (The same holds for other complex expressions; I use sentences only as an example. I also don't want to suggest that only semantic properties that are relevant to determine the properties of sentences are worth investigating into. For example, I believe that the referent is an important semantic property of proper names, even though I'm just about to say that it's irrelevant for compositional semantics of sentences.)

For example, instead of specifying a sentence's truth value by means of its constituents' Bedeutungen, we might just as well specify it by means of their Sinne. For the Sinne of the constituents determine the Sinn of the sentence, and the Sinn of the sentence determines its truth value. Bedeutungen for constituents are therefore redundant in Frege's semantics.

Conversely, some have argued that we should try doing without Sinne for at least certain constituents, namely proper names. On this view, the only semantic property of a proper name we need is its referent, i.e. its Bedeutung. The view is disturbingly peculiar, because many properties in the above list certainly can't be accounted for by ignoring all other properties of names. So the view must presumably be combined with the further claim that all the other properties, those that can't be accounted for, are not really semantic.

The problem is that this seems to exclude even property 1, truth value, and thereby make the position look very silly. For in many cases, substituting one name by another can alter the truth value of a sentence even if both names have the same referent. Moreover, some names don't have a referent but can nevertheless be used to build sentences that are clearly true or false.

So we have another good reason to abandon Frege's two-tiered semantics: If we specify the truth value of a sentence via its sense, we don't need to bother about what happens if one of its constituents is an empty name.

More generally, this illustrates that the projects of accounting for the properties in the above list are often closely connected: The best way to determine the truth values of sentences might be one that also determines many of the more "intensional" properties like 4 and 5.