Comments on Philippe Schlenker’s Be Articulate! A Pragmatic Theory of Presupposition Projection Berit Brogaard Philosophy RSSS The Australian National University [email protected] October 6, 2007 A presupposition of a sentence S is a precondition p such that if p is not taken for granted, S is infelicitous Previous Proposals Robert Stalnaker: “When a speaker says something of the form A and B, he may take it for granted that A (or at least that his audience recognizes that he accepts that A) after he has said it. The proposition that A will be added to the background of common assumptions before the speaker asserts that B. Now suppose that B expresses a proposition that would, for some reason, be inappropriate to assert except in a context where A, or something entailed by A, is presupposed. Even if A is not presupposed initially, one may still assert A and B since by the time one gets to saying that B, the context has shifted, and it is by then presupposed that A” (Stalnaker 1974)

Stalnaker’s proposal predicts that (1) presupposes that France has a king, but (2) doesn’t, because when the first conjunct in (2) is uttered, the speakers will add ‘France has a king’ to the common ground. (1) (2)

The king of France is bald France has a king, and the king of France is bald

Problems (according to Schlenker): A: The audience may not take for granted what the speaker asserts. If the speaker asserts that the earth is flat, the audience may just conclude that he is an idiot. B: The proposal does not extend to embedded conjunctions. For example (from Schlenker), “none of my students is both incompetent and aware that he is” presupposes nothing; “none of my students is aware that he is incompetent” presupposes that all of my students are incompetent. But because the conjuncts are embedded under a negative quantifier, they are not asserted. So, Stalnaker’s analysis does not extend to these cases. C: Not clear how to extend Stalnaker’s proposal to sentences with connectors other than conjunction (e.g. disjunction).

Irene Heim:

1

The lexical items are dynamic. Stalnaker’s update rule is grounded in the semantics of ‘and’. Meaning = context-change potential. Update rules for other connectors will follow from the semantics. Problem (according to Schlenker) Descriptively adequate but not explanatorily adequate. No way to explain why English contains ‘and’ rather than ‘*and’, because meaning of connectors must be stipulated up front. Schlenker’s Proposal Pragmatic account Let pp’ be a propositional meaning with truth-conditional content: p and p’, where p is the pre-condition of pp’. For example, the truth-conditional content of ‘John knows it is raining’ can be written as ‘it is raining, and John knows it’. Be Articulate! In any syntactic environment, express the meaning of an expression pp’ as (p and pp’), unless independent pragmatic principles rule out the full conjunction. In other words: if possible, say p and pp’ rather than just pp’ Incremental Presuppositional Transparency Roughly: a conjunct is transparent if it is superfluous given the context set, and transparent conjuncts should be left unsaid. More precisely: The beginning of a sentence α (p and uttered in a background of assumptions C is infelicitous if no matter what follows this string, the expression p and could be eliminated without modifying the contextual meaning of the result. Formally: Given a context set C, a predicative or propositional occurrence of pp’ is acceptable at the beginning of a sentence α pp’ If the ‘articulated’ competitor α (p and p’) is ruled out because p is transparent if for any expression γ of the same type as p and for any sentence completion β, C ⏐= α (p and γ) β ⇔αγβ It may be ok to include a conjunct that is indispensable if its indispensability can only be determined after the conjunction has been uttered. Background: none (3) Mary is pregnant, and she is expecting a son (4) #Mary is expecting a son, and she is pregnant

2

Schlenker allows for local and global accommodation. Background: it is not presupposed that A has a sister. A: My sister is pregnant. B: Oh, when is she due? Background: it is presupposed that A does not have a sister. A: My sister is pregnant. B: But you don’t have a sister. What the h … are you talking about? Furthermore, Schlenker allows for global presuppositional transparency GPT (i.e., the transparency of the second conjunct may affect acceptability). That is, one may rule out a competitor conjunction by taking into account the entire sentence. But relative strengths are assigned to IPT and GPT (IPT is stronger).

Some Questions for Schlenker 1. Anaphora: Background: it is taken for granted that Mary is pregnant (5) # Mary is pregnant, and she is expecting a son. (6) Mary is pregnant, and that’s why we need a bigger house (5) is alleged to be infelicitous because ‘Mary is pregnant and’ could have been eliminated without modifying the contextual meaning of the result, but as (6) shows, it is not clear that there are any cases where no matter what follows, the expression p and could be eliminated without modifying the contextual meaning of the result. Background: It is taken for granted that Alice accepted a job in California (7)

John: Why are you leaving New York? Alice: I accepted a job in California, and that’s one reason I am leaving. Another is …

Question: Are these cases ruled out by “γ is of the same type as p”? If not, what rules them out?

2. Aide Memoire Background: We are at Alice’s farewell dinner. Alice is giving a toast. It is taken for granted that Alice accepted a job in California, and that everyone knows it. (8)

As you all know, I accepted a job in California, and I am very excited but I will miss you all terribly.

3

Question: (8) is felicitous, but as ‘as you all know, I accepted a job in California’ is transparent, “Be Articulate!” predicts that (8) is infelicitous. Why is ‘as you all know, I accepted a job in California’ not transparent in this context?

3. Questions as parts of the Common Ground Context: It is assumed that John is looking for his birth mother. There is no question as to what else John might be looking for. (9)

#What John was looking for was a new room mate

Presupposition: John was looking for something. This is part of the common ground. Yet (9) is infelicitous because it wasn’t part of the common ground that we wanted to know what he was looking for. Context: It is known that John once competed in a local math competition. No one is wondering where John was or wasn’t competing. (10)

#It was in a local swim competition that John competed

Presupposition: John competed somewhere. This is part of the common ground. Yet (10) is infelicitous because it wasn’t part of the common ground that we wanted to know where he was competing. Question: How does the “Be articulate!” proposal account for questions as presuppositions? Questions cannot be stated as pre-conditional conjuncts, witness the infelicity of ‘Where did John compete? and it was in the local swim competition that John competed’.

4. Questions as Triggers (11)

Is Sydney in Australia or America?

Presupposition: Sydney is either in America or in Australia. Yet ‘Sydney is either in America or in Australia, and is Sydney in Australia or America?’ is odd. Question: How does the “Be Articulate!” proposal generalize to questions?

5. Philosophical Contexts Context: Nothing is assumed about Mary

4

(12) (13) (14)

#If Mary is expecting a son, she is pregnant and her parents know it. Context: Philosophy room. 1 If presentism is true, then there are no cross-time relations, and that is plainly absurd/and presentists know it. If genuine modal realism is true, then there is a multiplicity of concrete spatio-temporally disconnected worlds, and that is plainly absurd/and modal realists know it.

In (12) ‘she is pregnant’ is transparent, and this is supposedly what causes the infelicity of (12). ‘She is pregnant’ is transparent because ‘Mary is expecting a son’ stands in some relevant consequence relation to ‘Mary is pregnant’. However, (13) is not infelicitous, despite the fact that ‘presentism is true’ and ‘there are no cross-time relations’ stand in some consequence relation (perhaps of the very same sort). And likewise for (14). Question: What explains the difference between (12) and (13)/(14)?

6. On Stalnaker Ad A) Given Stalnaker 2002, why is problem (A) a problem at all? In Stalnaker (2002) Stalnaker says that one only needs to assume, or present oneself as presupposing, p, for the common ground to include p. This is quite common in philosophical discourse. Consider: -Let’s grant, at least for argument’s sake, that … -Suppose A, B, C … -Eternalist: if presentism is true, then there are no cross-time relations. But that can’t be right. Question: Why can’t we substitute ‘take for granted’ in the pretend-sense for ‘take for granted’ in the belief-sense? Ad B) ‘None of my students is aware that he is incompetent’ presupposes that all of my students are incompetent. Given Jackendoff’s X Syntax, however, the sentence cashes out to: [The X: students of mine X]([∀x: Xx](x is not aware that x is incompetent)). This asserts of my students that all of them are such that they are unaware that they are incompetent. Since ‘unaware’ is factive, this presupposes that all of my students are incompetent. 1

The indicative conditional, of course, cannot be interpreted as a material conditional.

5

Question: Is there any way to extend this proposal to non-partitive constructions? Could they perhaps be treated as implicit partitives (e.g. ‘nobody who applied is aware that he is incompetent’ Æ ‘the ones who applied are such that none of them are aware that they are incompetent’. ‘at most three of those who applied are aware that they are incompetent’ Æ ‘the ones who applied are such that at most three of them are aware that they are incompetent’). If so, then (B) is not a good reason to discard Stalnaker’s original proposal.

7. Entailment 1 Schlenker: A clause pp’ with presupposition p and assertion p’ is felicitous in a context C just in case C entails p. This is too strong. Background: For years and years a mathematician, Ruth, has been trying to figure out whether Goldbach’s conjecture is true, and she has finally been able to prove it. One day she says the following to her colleague, who has no idea that she has been working on the problem, let alone been able to prove that the conjecture is true: (15)

#I am so happy that Goldbach’s conjecture is true

This presupposes that Goldbach’s conjuncture is true, and since every necessary truth is entailed by anything, C entails that GC is true. Yet (15) is infelicitous. Better: A clause pp’ with presupposition p and assertion p’ is felicitous in a context C just in case p is an obviously relevant and necessary consequence of C. Relevance (proposal 1): q is a relevant consequence of p iff q does not introduce any new non-logical constants. Too strict. Relevance (proposal 2): q is a relevant consequence of p iff, if q introduces a new nonlogical constant C2, then for some minimal constant C1 in p, necessarily, all C1 is C2. Question: Should we perhaps substitute ‘obviously relevant and necessary consequence’ for ‘consequence’, as just illustrated?

8. Entailment 2 Schlenker: In fact, it can be checked independently that conjunctions are entirely infelicitous when the second conjunct is entailed by the first one. #John has cancer and is sick Not quite:

6

(16)

Mathematician: I have got some good and some bad news for you, colleagues: I didn’t get the grant I applied for but Goldbach’s conjecture is true [second conjunct entailed by the first]

(17)

Mathematician: I have got some great news for you, colleagues: Fermat’s last theorem is false, and I got the grant I applied for. [second conjunct entailed by the first]

Question: Perhaps obviously relevant and necessary consequence will do?

9. Attitude Ascriptions Contrary to what is claimed, attitude reports seem unproblematic, as ‘desire’ is not closed under any interesting closure principle. For example, ‘I want to kill your cat’ does not entail ‘I want your cat to exist’. wh-complement clauses sometimes have presuppositions, for example, ‘Lois Lane knows what Superman did at 3 p.m.’ presupposes that there is something Superman did at 3 p.m. Question: Do attitude ascriptions pose a problem for the “Be Articulate!” proposal, or does the problem lie elsewhere (e.g. in the closure principles standardly assumed for attitudes)?

7