Starting with Frege, the semantics (and pragmatics) of quotation has received a steady flow of attention over the last one hundred years. It has not, however, been subject to the same kind of intense debate and scrutiny as, for example, both the semantics of definite descriptions and propositional attitude verbs. Many philosophers probably share Davidson's experience: ‘When I was initiated into the mysteries of logic and semantics, quotation was usually introduced as a somewhat shady device, and the introduction was accompanied (...) by a stern sermon on the sin of confusing the use and mention of expressions’ (Davidson 1979, p. 79). Those who leave it at that, however, miss out on one of the most difficult and interesting topics in the philosophy of language. (shrink)

This is a lively, provocative book and many of its arguments are convincing. In this critical study I summarize the book, then discuss some of the authors’ claims, dwelling on three issues: their objections to the view of François Recanati on “pre-semantic” effects; the relation between their theory of quotation and the Tarskian “Proper Name Theory,” which they reject; and their treatment of mixed quotation, which rests on the claim that quotation expressions are “syntactic chameleons.” I argue that the objections (...) to Recanati don’t expose any problem with his view, and that the “Proper Name Theory” has all the virtues of their own proposal. Finally I raise some queries about the technical apparatus of syntactic chameleonism. (shrink)

It has proved challenging to account for the dual role that a directly quoted part of a 'that'-clause plays in so-called mixed quotation. The Davidsonian account, elaborated by Cappelen and Lepore, handles many cases well; but it fails to accommodate a crucial feature of mixed quotation: that the part enclosed in quotation marks is used to specify not what the quoter says when she utters it, but what the quoted speaker says when she utters it. Here I show how the (...) Davidsonian can do better. The proposal rests on the idea that mixed quotation involves deferred demonstration: a mixed quotati on specifies what the subject says partly by demonstrating the quoter's utterance of the unquoted part and partly by deferred-demonstrating the subject's utterance of the quotation-marked part. (shrink)

The issues addressed in philosophical papers on quotation generally concern only a particular type of quotation, which I call ‘closed quotation’. The other main type, ‘open quotation’, is ignored, and this neglect leads to bad theorizing. Not only is a general theory of quotation out of reach: the specific phenomenon of closed quotation itself cannot be properly understood if it is not appropriately situated within the kind to which it belongs. Once the distinction between open and closed quotation has been (...) drawn and properly appreciated, it is tempting to consider that only closed quotation is relevant to semantics. Open quotation is more a matter of pragmatics: it is a matter of what people do with words, rather than a matter of content and truth-conditions. In this way one can provide the beginning of a justification for the neglect of open quotation in current semantic theorizing. There is some truth in this view, yet the phenomenon of ‘mixed quotation’, investigated at length in this paper, is interesting precisely because it shows that things are not so simple. Important issues concerning the interface between semantics and pragmatics will thus be raised. (shrink)

Most writers working on simultaneous use and mention assume a distinction between mixed quotation and scare quoting. The consensus is that MQ affects truth-conditions. Hence, many writers regard MQ as a semantic phenomenon. There is no such consensus about ScQ. On the face of it, there is a clear difference between: Alice said that life “is difficult to understand”. Several ‘groupies' followed the band on their tour. The words quoted in are attributed to Alice, and would seem false if Alice (...) had not uttered them. No such intuition is available for. This induces Cappelen & Lepore, leading writers in the field, to regard as a blend of Direct and Indirect Discourse and to propose the following semantic analysis : there is an utterance u such that Alice said u and u samesays “Life is difficult to understand” and u sametokens “is difficult to understand”. As for, they regard the impact of the quotation in it as being entirely pragmatic. I show, however, that there is no principled distinction between MQ and ScQ. The main points are: speech attribution to Alice in is defeasible; there are cases of simultaneous use and mention without “say” which are semantically relevant ; there is apparent MQ with reporting verbs that pose a problem because they cannot govern Direct Speech reports. Hence, it is illegitimate to appeal to the sametokening predicate in logical form. These examples that ‘refuse' to fall neatly under MQ or ScQ seal the fate of Cappelen & Lepore's account. To conclude, I discuss Benbaji's fully semantic theory. Benbaji adopts Cappelen & Lepore's views on MQ. But he also proposes a semantic account of ScQ. One option for him is to give up the Cappelen & Lepore story and extend his views on ScQ to all instances of simultaneous use and mention. I show, however, that this too fails. Therefore, in the end, only a pragmatic account of simultaneous use and mention remains as a viable option. (shrink)

I distinguish, and discuss the relations between, five types of context-shift involving indexicals. For 'intentional' indexicals - indexicals whose value depends upon the speaker's intention - we can shift the context more or less 'at will', by manifesting one's intention to do so. For other indexicals we can shift the context through pretense. Following a number of authors, I distinguish two types of context-shifting pretense, corresponding to two sets of linguistic phenomena. The fourth type of case is that of expressions (...) which are not really indexical, but perspectival, and for which we do not need to appeal to the notion of context-shift in order to account for their shifty behaviour. The fifth category I introduce and discuss is that of 'shiftable indexicals'. Do we need it, given that we already have the other four categories? My answer is a qualified 'yes'. (shrink)