Brains, Persons, and Society *** ABSTRACTS
   Cervelli, Persone e Società ***ABSTRACTS





Serena Corrao
Dipartimento di Filosofia e SS. Sociali, Università di Lecce

Moral realism and truth-conditional semantics: a vain marriage

In this paper I intend to draw attention to the fundamental distinction -that in ordinary language mostly collapses- between genuine normative (or evaluative) sentences and sentences descriptive of norms (or values). The former express norms (or values); the latter state that a certain norm (or value) exists or is valid in a given normative (or axiological) system. These two kinds of sentences are almost systematically homophonic (von Wright 1963), that is, the same string of words like ‘it is obligatory to keep promises’ or ‘lying is wrong’ can be used to express a norm/value or to state the existence of a norm/value (Ayer 1936; Bulygin 1982). Such ambiguity of deontic sentences in ordinary language has been largely acknowledged and it has represented a much-debated issue in contemporary legal philosophy and deontic logic, involving many distinguished philosophers like Kelsen, Hart, von Wright, Ross, Alchourron, Bulygin, Weinberger and Kalinowski. A fairly established conclusion of such debate affirms the importance of disambiguating the meaning of deontic expressions, since “one and the same expression cannot share both properties, i.e., be prescriptive and descriptive at the same time”: norms lack truth-values, while sentences descriptive of norms do have truth-values and they are mutually exclusive (Bulygin idem). Despite the great prominence enjoyed by this issue within the above mentioned branches of philosophy, it seems it has had no echo in ethics. In fact, the defence of non-naturalist cognitivism (in particular, moral realism), notoriously starts from the appeal to the descriptive phenomenology of moral language. However, moral language -I will argue- so appears, just because of the systematic homophony of sentences expressing norms and values with the corresponding sentences describing them, which causes the former to be understood as truth-apt. As Bulygin remarks (1982), the idea that normative sentences could be true or false is commonly based on an analogy with the Tarskian theory of truth for descriptive sentences. In particular, it is held that it is possible to extend to normative and evaluative sentences the T-conditions. Nevertheless, Bulygin points out that merely suggesting an analogy between the truth of descriptive sentences and the truth of normative sentences is not enough: it is necessary to spell out what it means for a normative sentence to be true in terms of the Tarskian theory, as well as what kind of reality would correspond to a true normative sentence.


The development of an extensional theory of meaning à la Davidson has been welcomed by moral realists as a response to this problem (Platts 1980; McDowell 1980). A distinguished representative of contemporary moral realism has recently nicely reaffirmed that “realists were in troubles until Davidson came along” (Dancy, Helsinki, Dec.2005). As a matter of facts, by extending Davidson’s truth-conditional semantics to moral sentences one typically obtains meaning-giving theorems, patterned after Tarski’s ‘T-sentences’, that articulate the meaning of sentences by stating the conditions under which they are true.


However, such a solution incurs a serious objection: it entirely disregards the distinction between normative or evaluative sentences and sentences descriptive of norms and values. I will show that Tarskian semantics can be appropriately applied only to the latter and that its application to the former cannot but conflate the two kinds of sentences.


I will also show that the same outcome results from applying the possible worlds semantics.


All this supports the conclusion that moral sentences cannot be cognitively understood as possessing truth-values without them being systematically conflated with their corresponding descriptive sentences; that is, no truth-conditional semantics can be applied to moral language on pain of dissolving the philosophically fundamental distinction between expression and description of norms/values.