Saul Kripke’s Locke Lectures were delivered in Oxford in 1973. This volume is a publication of the lectures in book form for the first time. Kripke mainly discusses two topics. One is the problem of empty names, typical examples of which are fictional names such as ‘Sherlock Holmes’ (‘SH’ hereafter). The other is that of semantic referent and speaker’s referent.

On the problem of empty names, Kripke deals with five connected but different issues. The first is ‘whether the existence of fictional works should be held to be conclusive against Mill’s semantics of names,’ according to which a name’s whole semantic function is its referent. It seems that the answer should be yes, because in a fictional work of SH, ‘SH’ is meaningful, but SH doesn’t exist at all. Kripke’s view is that the existence of fictional works is not a powerful argument for any semantic theory, since it is part of the pretense of fiction “that the criteria of naming, whatever they are … are satisfied” (23).

The second issue is ‘how to analyze negative existential statements’ such as ‘SH doesn’t exist.’ The Russellian analysis cannot be successful, according to which ‘SH’ means the unique person satisfying the properties stated in the story: (a) the story needn’t say anything to identify SH uniquely; (b) it is inconsistent with the author’s claim that the names used in the story are fictional and any resemblance to characters is coincidental; and (c) it leaves no room for historical fictions (25-28). It is unhelpful to appeal to the ontology of fictional characters, since as a fictional character, SH does exist (which will be shown in the fourth issue section). Metalinguistic analysis also fails (151-155). Kripke’s tentative view is that, because of our careless use of ‘false,’ we assert ‘SH doesn’t exist,’ by which we mean it is ‘false’ that there is a true proposition that SH exists, or we simply mean that “there is no such true proposition as that SH exists” (159). The assertion is true, since there is “really no such proposition at all as that Sherlock Holmes exists” (159). The case is similar for general negative existential statements, such as, ‘there are no dragons.’

The third issue is ‘whether SH might have existed.’ The Russellian answer is ‘yes,’ according to which ‘SH might have existed’ means ‘it is possible that there is a unique person having those properties stated in the story.’ Kripke argues that this is wrong, since ‘SH’ is introduced to name a particular character who would have done certain things, but not just any character who did these things (41). Kripke’s view is that there is no such possibility as that SH exists, since the proposition that SH exists itself does not exist (42). Fictional predicates are similar: “one cannot say under what circumstances there would have been bandersnatches” (52).

The fourth issue is ‘is there any sense in which SH exists?’ Kripke offers two such senses. One is the ‘according to the story’ sense. It is true in this sense, since it is a correct report on what exists in the story. The other is the ‘as a fictional character’ sense, in which the statement means ‘as a fictional character, SH exists.’ It is true, since it follows from true literary criticism statements, such as, ‘SH is a fictional character,’ and intentional statements such as ‘Kripke admires SH.’ Kripke then proposes his ontology of fictional characters. Fictional characters are conceived as abstract but created objects of their authors (73).

The fifth issue is ‘whether a theory of fiction can be helpful to solve problems of perception.’ Kripke answers ‘yes.’ He extends his account of fictional characters to deal with the problem of what we see. When a person in hallucination reports that she sees a pink rat, given the first analogy with fiction, Kripke’s suggestion is that she really sees something—a hallucinatory rat (94). As long as hallucinatory objects are introduced, given the second analogy, their predicates can be understood in two senses— the ‘out-and-out’ sense and the ‘according to a vision' sense. ‘Was caused by such and such medical problem’ is ‘out-and-out’ usage, but ‘is colored pink’ is the ‘according to a vision’ usage (95).

On the topic of semantic referent and speaker’s referent, Kripke makes three important points. First, the notion of speaker’s referent can’t be used to show that ‘SH’ is referential in a semantic sense, because speaker’s referent is not a semantic notion, but a pragmatic notion (131). Kripke thinks, however, speaker’s referent has certain significance to the semantics of pronominalization. Second, given the distinction between semantic referent and speaker’s referent, it can be shown that Donnellan’s argument against Russell’s theory of definite descriptions is unsuccessful. According to Russell’s analysis, a definite description’s referent (if any) is the unique object that satisfies its descriptive content. To Donnellan, this analysis is refuted by the linguistic phenomenon that a speaker can use a definite description to talk about an object when the object does not uniquely satisfy the descriptive content. Kripke argues that Donnellan is wrong, because we can easily imagine a hypothetical community, in which Russell’s analysis is stipulated to be correct, but in terms of the general principle of the divergence of speaker’s referent from semantic referent, the linguistic phenomenon mentioned by Donnellan will be exhibited (127). Third, a speaker’s referent of a singular term may be transmitted from one speaker to another, and might eventually become the semantic referent (135-136). Gareth Evans’s case of ‘Madagascar’ is an example.

Reviewer’s comments: (1) Kripke should give an explanation of how an author can create an abstract fictional character, without which his ontology of fictional characters is mysterious; (2) Kripke’s analogy of perception with fiction seems incorrect: if the object of sight must be physical in some sense, and if hallucinatory objects are non-physical (94), then they cannot be objects of sight; (3) ‘there is no such true proposition as that SH exists’ is apparently about the proposition that SH exists, which is denied by Kripke, so it is hard to understand that his ‘metapropositional’ analysis is an analysis; (4) the lectures are full of rich philosophical ideas, which have influenced and will continue to influence philosophers in metaphysics, philosophical logic and philosophy of language.