On Reporting Attitudes: an Analysis of Desire Reports and their Reading-Establishing Scenarios∗

I argue that conventional wisdom about attitude reports lumps together theorizing about the attitudes themselves with theorizing about the meaning of attitude reports. Instead of conflating these, I propose to treat them as distinct (although obviously related) topics. A central reason for this separation is the distinction between anaphoric constructions that are internally about the same object and anaphoric constructions that are externally about the same object. Anaphoric constructions that are internally about the same object reveal the structure of the attitudes themselves – e.g. that desires are parasitic on beliefs – whereas anaphoric constructions which are externally about the same object make explicit the structure of attitude reports. I correlate the distinction between the attitudes themselves and attitude reports with the asymmetry between the firstand third-person perspective on attitudes. On the basis of this asymmetry, I propose a semantics of desire reports according to which a desire report like Adrian wants to buy a jacket like Malte’s is not ambiguous between different readings. Conventional wisdom readings are reconstructed as a third person’s pragmatic strategies for justifying the truth of an attitude ascription. I conclude with an outlook on the transfer of the analysis proposed to reports involving other non-doxastic predicates like wish or hope and reports involving doxastic predicates such as believe. 1 Three concerns about the conventional wisdom analysis of attitude reports Conventional wisdom has it that an attitude report like (1) is ambiguous between a range of so-called readings of (1) that appear when (1) is interpreted against a certain set of background assumptions. (1) (A reporter says:) Adrian wants to buy a jacket like Malte’s. Background assumptions – often called scenarios – for the identification of readings usually comprise (a) a doxastic specification of the descriptive content of the report, e.g. whether or not Adran knows Malte or his jacket and (b) a specification of the reality in which Adrian finds himself, e.g. whether or not Adrian stands in a causal relation of acquaintance to a specific jacket like Malte’s. For example, a scenario in which Adrian has decided to buy a certain jacket but has no idea that the jacket he wants to buy is like Malte’s gives rise to the so-called de re reading of (1). According to this reading, there is a jacket like Malte’s which Adrian wants to buy. Or, in a scenario in which Adrian has not decided which jacket he wants to buy but he wants it to be like Malte’s, (1) receives a de dicto interpretation, according to which Adrian wants to buy something that is a jacket like Malte’s. Under the terms of conventional wisdom, the existence of readings is established by the difference that scenarios make to the interpretation of an attitude report. As (von Fintel and Heim, 2011, p. 85) put it: “[. . . ] construct ∗I would like to thank Hans Kamp and Emar Maier for their comments and helpful discussion. Only I am responsible for any errors that remain. This work was supported by a DFG grant to the project B4 ’Lexikalische Information und ihre Entfaltung im Kontext von Wortbildung, Satz und Diskurs’, as part of the Collaborative Research Center 732 ’Incremental Specification in scenario’ at the University of Stuttgart.