MIT OpenCourseWarehttp://ocw.mit.edu 24.910 Topics in Linguistic Theory: Propositional Attitudes Spring 2009 For information about citing these materials or our Terms of Use, visit: http://ocw.mit.edu/terms.Feb. 3 24.910, Spring 2009 (Stephenson) Preview of Course Topics Course Theme: Propositional attitudes (and related topics) Preliminaries:  Propositions ¾ The kind of thing that is expressed by a declarative sentence ¾ The semantic type of a sentence / clause To give some (lame) examples:  It’s snowing (in Cambridge) (on Feb. 3, 2009)  2 + 2 = 4  All books have pages  Propositional attitudes ¾ Mental states that we might have towards propositions. For example: belief, knowledge, suspicion, discovery, desire ¾ Some attitude predicates: believe, know, realize, think, discover, want … (and similar expressions in other languages)  Topics to be explored in this course: ¾ Central theme: The semantics of attitude predicates Expanding the domain: ¾ Speech predicates (say, ask, etc.) (since these have many semantic and syntactic parallels with attitude predicates) ¾ Syntax of sentence embedding (since speech and attitude predicates typically take sentential complements of various kinds) ¾ Formal Pragmatics – Theory of conversation and common ground that is, the ways that shared information is updated in the course of conversations (since this involves many of the same formal tools as the semantics of propositional attitudes) 1Feb. 3 24.910, Spring 2009 (Stephenson) Specific Topics (The list below will have substantial overlap with – but will not be identical to – the set of topics we actually talk about.) Possible World Semantics for Propositional Attitudes  Possible Worlds ¾ Possible world: complete history of the entire universe, specifying every detail about how everything happens. ¾ The actual world: the particular possible world that we are living in Lewis (1986, quoted in H&K): ¾ A proposition: can be construed as a set of possible worlds (intuitively, the set of worlds where that proposition is true) ¾ This is parallel to treating a predicate such as red or smokes as a set of individuals.  Representing knowledge and desire states (for example): ¾ If we knew everything there was to know about the world (omniscient deity), we would know which possible world was the actual one. BUT we don’t, so we can only narrow down worlds to a set that are compatible with what we know – i.e., which could be the actual one for all we know. So we can construe “what x knows” as a set of possible worlds. ¾ Similarly, if we had very specific desires about how the world should be (a control-freak deity?), then we could specify exactly which world we would like to be in. BUT again our desires aren’t this specific – for example, I might want to have pizza for dinner tonight, but not particularly care whether it has olives on it or not. 2Feb. 3 24.910, Spring 2009 (Stephenson) So we just narrow down the worlds to the set compatible with our desires (i.e., worlds which would all be perfect as far as we are concerned.) ¾ Again: construe “what x wants” as a set of possible worlds.  Note about mental states and propositions (Note that on this view a person’s entire knowledge state (belief state, desire state, etc.) is represented as the same type of thing (set of worlds) as a sentence such as It’s snowing.) To help make sense of this, keep in mind that: ¾ There is a natural mapping from sets of propositions (sets of sets of worlds) to sets of worlds. ¾ Given a set of sets of worlds, e.g.: let S = { {w1, w2, w3, w4}, {w2, w3, w4}, {w1, w2, w3} } We can take the intersection of all of the member sets to get a set of worlds: ∩ S = {w2, w3}  Attitudes as sets of worlds ¾ Let Knowx = {w: w is compatible with what x knows} (the set of possible worlds compatible with what x knows, where x is an individual) Similarly: ¾ Believex = {w: w is compatible with what x believes} ¾ Wantx = {w: w is compatible with what x wants to be the case}  Compositional semantics of attitude reports ¾ [[α]] = def the semantic value (meaning) of expression α ¾ Note, for a sentence S, [[S]] /a/ will be a proposition (set of worlds) Some general rules: ¾ [[x knows S]] = 1 [true] iff knowx ⊂ [[ S ]] i.e., x knows S is true iff every world compatible with what x knows is a world where S is true. Similarly: ¾ [[x believes S]] = 1 iff believex ⊂ [[ S ]] ¾ [[x wants S]] = 1 iff wantx ⊂ [[ S ]] 3Feb. 3 24.910, Spring 2009 (Stephenson) Presupposition & Entailments Factive predicates: (1) Sue knows that it’s snowing. (2) Sue doesn’t know that it’s snowing. (3) Sue realized that it was snowing. (4) Sue didn’t realize that it was snowing. All say something about Sue’s mental state; but they also suggest that it is, in fact, snowing. (5) # Sue knows that it’s snowing, but it isn’t. (6) # Sue realized that it was snowing, but it wasn’t. ¾ [[x realized S [at time t1 ] ]] = 1 if [[S]] = 1 AND knowx ⊄ [[S]] before t1 and knowx ⊂ [[S]] at t1 0 if [[S]] = 1 AND knowx ⊄ [[S]] before t1 and knowx ⊄ [[S]] at t1 undefined if [[S]] ≠ 0 (or if knowx ⊂ [[S]] before t1 undefined if [[S]] ≠ 0 [Some of this might come from what counts as “knowledge” in the first place] ¾ [[x knows S]] First pass: ¾ [[x knows S]] = 1 if [[S]] = 1 AND knowx ⊂ [[ S ]] 0 if [[S]] = 1 AND knowx ⊄ [[ S ]] = 1 if [[S]] = 1 AND knowx ⊂ [[ S ]] 0 if [[S]] = 1 AND knowx ⊄ [[ S ]] undefined if [[S]] ≠ 0 Tense under Embedding  Two readings for embedded past tense in English: (7) Sue said that was happy. (i) Sue said, “I’m happy” [she said at t1 that she was happy at t1] (ii) Sue said, “I was happy” [she said at t1 that she was happy at t0] Reading (i) shows “sequence of tense.” 4Feb. 3 24.910, Spring 2009 (Stephenson)  Russian embedded present Russian embedded tense (Schlenker 2003 citing Kondashov & Kondrashova, Kusumoto): Extra complication: it matters what the embedding predicate is:  “Double-access” readings (English) (Abusch, 1988, 1991; Ogihara, 1995) (8) Sam found out that Mary was pregnant. (Ogihara, 1995, no. 28) (9) Sam found out that Mary is pregnant. (Ogihara, 1995, no. 27) [Mary must still be pregnant at the time of utterance] Logophors & Shifting Indexicals  Logophors (West African type) Logophors: pronouns