Slides-1 [Dynamic Semantics] - UCLA Department of Linguistics
Slides-1 [Dynamic Semantics] - UCLA Department of Linguistics
Slides-1 [Dynamic Semantics] - UCLA Department of Linguistics
Create successful ePaper yourself
Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.
Presupposition<br />
<strong>UCLA</strong>, Fall 2007<br />
Philippe Schlenker<br />
(<strong>UCLA</strong> & Institut Jean-Nicod)
Presupposition<br />
Approximation: A presupposition <strong>of</strong> S is a condition that<br />
must be met for S to be true or false.<br />
Presuppositions<br />
a. John knows that he is incompetent.<br />
π: John is incompetent.<br />
b. Does John knows that he is incompetent?<br />
π: John is incompetent<br />
c. John doesn’t know that he is incompetent.<br />
π: John is incompetent.<br />
Entailments<br />
a. John is French. => John is European.<br />
b. Is John French? ≠> John is European.<br />
c. John isn’t French. ≠> John is European.<br />
2
Why Study Presupposition ?<br />
I. Presuppositions are ubiquitous<br />
John regrets that he is incompetent.<br />
π: John is incompetent.<br />
John has stopped smoking.<br />
π: John used to smoke.<br />
It is John who left.<br />
π: Someone left.<br />
What John drank was vodka.<br />
π: John drank something.<br />
She is clever!<br />
π: The person pointed at is female.<br />
3
John too was jailed.<br />
π: Someone other than John was jailed.<br />
John was jailed again.<br />
π: John was jailed before.<br />
Only John was jailed.<br />
π: Somebody was jailed.<br />
4
Why Study Presupposition ?<br />
II. Presuppositions and <strong>Dynamic</strong> <strong>Semantics</strong><br />
Static View <strong>of</strong> Meaning<br />
Meaning = Truth Conditions<br />
<strong>Dynamic</strong> View <strong>of</strong> Meaning (after the 1980’s)<br />
Meaning = Context Change Potential<br />
= potential to change beliefs<br />
Motivations for the dynamic view<br />
a. Pronouns, e.g. Every man who has a donkey beats it.<br />
b. Presuppositions.<br />
5
Why Study Presupposition ?<br />
III. The <strong>Semantics</strong> vs. Pragmatics Divide<br />
<strong>Semantics</strong> = study <strong>of</strong> meaning as it is encoded in words<br />
John is an American student<br />
=> John is a student<br />
John is a former student<br />
≠> John is a student<br />
Pragmatics = study <strong>of</strong> the additional information that<br />
can be obtained by reasoning on the speaker’s motives<br />
Mr. Smith is unfailingly polite and always on time<br />
=> Smith is a bad student<br />
6
<strong>Semantics</strong> vs. Pragmatics<br />
7
Entailments vs. Implicatures<br />
Difference 1: Entailments follow from what is<br />
linguistically encoded. Implicatures do not.<br />
Difference 2: Entailments satisfy the following test.<br />
Implicatures generally don't.<br />
To check whether p entails q, check whether:<br />
In every conceivable situation in which it is true that p, it<br />
is true that q.<br />
Difference 3: Implicatures can be cancelled. Entailments<br />
cannot be.<br />
8
Scalar Implicatures<br />
a. Rick is a philosopher or he is a poet<br />
(B. Schwarz)<br />
b. John will leave or Mary will leave.<br />
c. Paris is pleasant or London is pleasant.<br />
9
Hypothesis 1.<br />
Disjunction is unambiguously exclusive.<br />
[[ [i or i'] ]] = true iff exactly one <strong>of</strong> [[i]], [[i']] is true<br />
Notational variant (with 1 = true, 0 = false)<br />
[[ [i or i'] ]] = 1 iff exactly one <strong>of</strong> [[i]], [[i']] is equal to 1<br />
10
1. a is predicted to be a contradiction; it should have<br />
the same status as b.<br />
a. Rick is a philosopher or he is a poet. In fact, he is both.<br />
b. #Rick is a philosopher or he is a poet but he is not both.<br />
In fact, he is both.<br />
2. Incorrect predictions<br />
a. Every Italian who is a philosopher or a poet is a<br />
socialist.<br />
b. Whenever I invite a philosopher or a poet to a party, it<br />
ends up being a success.<br />
11
Every Italian who is a philosopher or a poet is a socialist.<br />
i 1, is a philosopher but not a poet, and he is a socialist.<br />
i 2, is a poet but not a philosopher, and he is a socialist.<br />
i 3, is both a philosopher and a poet, but he is not a socialist.<br />
12
Hypothesis 2.<br />
Disjunction is ambiguous<br />
1. Cross-linguistic morphology<br />
2. The ambiguity theory predicts that a. could be<br />
understood as true in the situation we described earlier.<br />
3. Ellipsis (Fox, crediting T. Stephenson)<br />
John read Chomsky or Montague. Mary did too. In fact,<br />
she read both<br />
13
General observation about ellipsis<br />
John went to the bank. Mary did too.<br />
bank is ambiguous:<br />
bank 1 = slope near the side <strong>of</strong> a river<br />
bank 2 = financial institution<br />
Ok John went to the bank 1. Mary went to the bank 1 too.<br />
Ok John went to the bank 2. Mary went to the bank 2 too.<br />
* John went to the bank 1. Mary went to the bank 2 too.<br />
* John went to the bank 2. Mary went to the bank 1 too.<br />
Ok John went to the bank 1. Mary did go to the bank 1 too.<br />
Ok John went to the bank 2. Mary did go to the bank 2 too.<br />
* John went to the bank 1. Mary did go to the bank 2 too.<br />
* John went to the bank 2. Mary did go to the bank 1 too.<br />
14
4. Yet another problem...<br />
a. It is certain that John will read Chomsky or Montague.<br />
b. Every student read Chomsky or Montague.<br />
15
Hypothesis 3.<br />
Scalar Implicatures<br />
Hypothesis:(i) ‘or’ is inclusive disjunction. (ii) an<br />
implicature is responsible for the not and inference.<br />
S said: F or G<br />
form a scale: F and G entails F or G.<br />
If S believed that F and G, it would have been more<br />
cooperative to say: F and G<br />
Primary Implicature: NOT S believes (F and G)<br />
If John is well informed and either believes or disbelieves<br />
(F and G), we also get:<br />
Secondary Implicature: S believes NOT(F and G)<br />
16
I. Alternatives<br />
Alt(S) = {S': S' is a sentence obtained from S by replacing<br />
simultaneously any number <strong>of</strong> occurrences <strong>of</strong> or by and<br />
and any number <strong>of</strong> occurrences <strong>of</strong> and by or}.<br />
a. S 1 = Rick is a philosopher or a poet<br />
Alt(S 1) = {Rick is a philosopher or a poet, Rick is a<br />
philosopher and a poet}<br />
b. S 2 = Rick is a philosopher and a poet<br />
Alt(S 2) = Alt(S 1) = {Rick is a philosopher or a poet, Rick is<br />
a philosopher and a poet}<br />
c. S 3 = I doubt that Rick is a philosopher and a poet<br />
Alt(S 3)={I doubt that Rick is a philosopher and a poet, I<br />
doubt that Rick is a philosopher or a poet}<br />
17
Ordering<br />
II. Ordering and Cooperation<br />
Let S be a sentence and let S' be a member <strong>of</strong> Alt(S).<br />
S' is better than S if:<br />
a. S' entails S and S does not entail S'<br />
[terminology: we say that S' asymmetrically entails S]<br />
b. The speaker believes that S'<br />
Cooperation<br />
A sentence S is not uttered cooperatively if for some S' in<br />
Alt(S), S' is better than S.<br />
18
Scalar Implicatures<br />
a. Rick is a philosopher or a poet<br />
b. Alt(a)={Rick is a philosopher or a poet, Rick is a<br />
philosopher and a poet}<br />
c. __ and __ >> __ or __<br />
a. is not uttered cooperatively if the speaker believes that<br />
Rick is a philosopher and a poet.<br />
-Primary Implicature: If the speaker is cooperative, it's<br />
not the case that the speaker believes that Rick is both a<br />
philosopher and a poet.<br />
-Secondary Implicature: If the speaker has an opinion on<br />
this matter, it must be that he believes that Rick is not both<br />
a philosopher and poet.<br />
19
Scalar Implicatures<br />
a. Rick is a philosopher and a poet<br />
b. Alt(a)={Rick is a philosopher and a poet, Rick is a<br />
philosopher or a poet}<br />
c. No member <strong>of</strong> Alt(a) asymmetrically entails a, so<br />
nothing additional is inferred.<br />
20
'Scale Reversal'<br />
a. I doubt that Rick is a philosopher and a poet<br />
b. Alt(a)={I doubt that Rick is a philosopher or a poet, I<br />
doubt Rick is a philosopher and a poet}<br />
c. I doubt that __ or __ >> I doubt that __ and __<br />
a. is not uttered cooperatively if the speaker doubts that<br />
Rick is a philosopher or a poet.<br />
... hence if the speaker is cooperative, the speaker does not<br />
doubt that Rick is a philosopher or a poet (i.e. he believes<br />
that Rick is a philosopher or a poet)<br />
a philosopher and poet.<br />
21
'Scale Reversal'<br />
a. Every Italian who is a philosopher or a poet is a<br />
socialist<br />
=> no additional inference (because the version with and<br />
would be less informative)<br />
b. Every Italian who is a philosopher and a poet is a<br />
socialist.<br />
=> it’s not the case that every Italian who is a philosopher<br />
or a poet is a socialist,<br />
i.e. some Italian who is a philosopher or a poet (but not<br />
both) is not a socialist.<br />
22
'Scale Reversal'<br />
a. Whenever John is next to Mary or Ann, he behaves like<br />
an idiot<br />
=> no additional inference<br />
b. Whenever John is next to Mary and Ann, he behaves<br />
like an idiot.<br />
=> It's not the case that whenever John is next to Mary or<br />
Ann, he behaves like an idiot.<br />
23
Some, Most, Every<br />
a. Some <strong>of</strong> my friends are clever<br />
=> Not all <strong>of</strong> my friends are clever.<br />
=> A minority <strong>of</strong> my friends are clever.<br />
b. Some <strong>of</strong> my friends are clever. In fact, all <strong>of</strong> them are.<br />
a. Most <strong>of</strong> my friends are clever<br />
=> Not all <strong>of</strong> my friends are clever.<br />
b. Most <strong>of</strong> my friends are clever. In fact, all <strong>of</strong> them are.<br />
a. Whenever most <strong>of</strong> the students come to class, there is a<br />
pleasant atmosphere.<br />
b. Every student who read most <strong>of</strong> the articles on the<br />
reading list will get an A.<br />
24
<br />
Extensions<br />
<br />
<br />
<br />
<br />
<br />
25
a. John read some book.<br />
Why are Scales Necessary?<br />
b. John read exactly one book.<br />
c. (b) is more informative than (a), therefore the speaker<br />
was not in a position to assert (b)<br />
d. Therefore it is likely that John didn’t read exactly one<br />
book.<br />
☞ This is the opposite <strong>of</strong> the result we want!<br />
26
Other Implicatures<br />
John is in Paris or he is in Rome<br />
=> it is not the case that:<br />
a. the speaker believes that John is in Paris.<br />
b. the speaker believes that John is not in Paris.<br />
c . the speaker believes that John is in Rome.<br />
d. the speaker believes that John is not in Rome.<br />
If John is in Paris, he is there for business.<br />
=> the speaker takes it to be possible but not certain that<br />
John is in Paris<br />
27
Experiment - Scalar Implicatures<br />
(Crain & co-workers, U. Maryland)<br />
28
[Credits: Crain & co-workers, U. Maryland]<br />
30
[Credits: Crain & co-workers, U. Maryland]<br />
31
[Credits: Crain & co-workers, U. Maryland]<br />
32
[Credits: Crain & co-workers, U. Maryland]<br />
33
[Credits: Crain & co-workers, U. Maryland]<br />
34
Children and Scalar Implicatures<br />
Children appear not to compute Scalar Implicatures in<br />
some environments where adults do.<br />
Paradox: children appear to be 'more logical' than adults!<br />
35
Scalar Implicatures Take Time<br />
Noveck and Posada 2003<br />
38
Three Properties <strong>of</strong> Scalar Implicatures<br />
Unlike entailments, they can be cancelled.<br />
They ‘disappear’ in certain environments (and ‘appear’ in<br />
others).<br />
They are acquired relatively late by children.<br />
They take time to compute.<br />
40
Presuppositions<br />
41
Presuppositions vs. Entailments<br />
Difference 1 (dubious)<br />
If an entailment <strong>of</strong> S is false, S is false, not weird.<br />
-John is French.<br />
-No. He is South African.<br />
-John knows that he is going to be fired.<br />
-No. He doesn’t know it.<br />
- No. He is going to keep his job.<br />
42
Presuppositions vs. Entailments<br />
Difference 2 (very clear)<br />
Presuppositions ‘project’ differently from entailments.<br />
a. Is John French? ≠> John is European<br />
b. John is not French. ≠> John is European<br />
c. None <strong>of</strong> these 10 students is French<br />
≠> Each <strong>of</strong> these 10 students is European<br />
≠> Some <strong>of</strong> these 10 students is European<br />
a. Does John know that he is incompetent?<br />
=> John is incompetent<br />
b. John does not know that he is incompetent<br />
=> John is incompetent<br />
c. None <strong>of</strong> these 10 students knows that he is incompetent<br />
=> Each <strong>of</strong> these 10 students is incompetent<br />
43
Presuppositions vs. Entailments<br />
a. Does John take care <strong>of</strong> his computer?<br />
=> John has a computer<br />
b. John doesn’t take care <strong>of</strong> his computer<br />
=> John has a computer<br />
c. None <strong>of</strong> these 10 students takes care <strong>of</strong> his computer<br />
=> Each <strong>of</strong> these 10 students has a computer<br />
a. Did John stop smoking?<br />
=> John used to smoke.<br />
b. John didn’t stop smoking<br />
=> John used to smoke<br />
c. None <strong>of</strong> these 10 students stopped smoking<br />
=> Each <strong>of</strong> these 10 students used to smoke<br />
44
Presuppositions vs. Implicatures<br />
An analysis <strong>of</strong> presuppositions as implicatures<br />
Hypothesis: If pp’ is a clause described as having<br />
presupposition p and assertion p’:<br />
(i) pp’ has as its meaning the conjunction <strong>of</strong> p and p’<br />
(ii) but forms a scale<br />
Examples<br />
a. <br />
b. <br />
c. <br />
45
Predictions I<br />
pp’ entails p <br />
a. John knows that he is incompetent<br />
=> John is incompetent<br />
b. I’ll invite John and Mary<br />
=> I’ll invite John or Mary<br />
not pp’ implicates p <br />
because (not p) is more informative than (not pp’) !<br />
a. John doesn’t know that he is incompetent<br />
implicates: John is incompetent<br />
b. I won’t invite (both) John and Mary<br />
=> I’ll invite John or Mary<br />
46
Predictions II<br />
No student PP’ implicates Some student P<br />
because No student P<br />
is more informative than No student PP’<br />
hence the inference that not No student P<br />
i.e. Some student P<br />
Presuppositions vs. Entailments:<br />
An Experiment (French, Chemla 2007)<br />
48
Experimental Conditions<br />
Triggers<br />
• Presuppositions<br />
attitude verbs: know, be unaware<br />
change <strong>of</strong> state: start, stop<br />
definite descriptions: his computer<br />
• Implicatures: , , <br />
Environments<br />
-Inferences: universal-like and implicature-like<br />
-Operators: John ___, I doubt that John ___, More than 3<br />
<strong>of</strong> these 10 students ___ , Each <strong>of</strong> the 10 students ___,<br />
None <strong>of</strong> these 10 students ___, Exactly 3 <strong>of</strong> these 10<br />
students ____.<br />
49
Examples<br />
Less than 3 <strong>of</strong> these 10 students know that their father is<br />
about to receive a congratulation letter.<br />
=>? The father <strong>of</strong> each <strong>of</strong> these students is about to receive<br />
a congratulation letter.<br />
=>? The father <strong>of</strong> at least 3 students is about to receive a<br />
congratulation letter.<br />
None <strong>of</strong> these 10 students read the handout and did an<br />
exercise.<br />
=>? Each <strong>of</strong> these 10 students did (at least) one or the other<br />
=>? At least 1 <strong>of</strong> these 10 students did (at least) one or the<br />
other<br />
50
Main Results (Chemla 2007)<br />
Presuppositions display a different a behavior from<br />
scalar implicatures under no:<br />
-Non-universal inferences for implicatures<br />
-Universal implicatures for presuppositions<br />
Not all quantifiers behave on a par:<br />
at least 3, more than 3, exactly 3 display an intermediate<br />
behavior (universal inferences half the time).<br />
Not computing a presupposition takes time.<br />
51
NO and Universal Inferences<br />
Left, from left to right<br />
1. Every student stopped<br />
smoking => every student<br />
smoked<br />
2. No student stopped<br />
smoking => at least one<br />
student smoked<br />
3. No student stopped<br />
smoking => every student<br />
smoked<br />
Right, from left to right<br />
1. Every student did A and<br />
B => every student did (at<br />
least) one<br />
2. No student student did A<br />
and B => at least one<br />
student did (at least) one<br />
3. No student did A and B<br />
=> every student did (at<br />
least) one<br />
52
NO and Universal Inferences<br />
53
Less than three and Universal Inferences<br />
54
Universal Inferences for Various Quantifiers<br />
55
Reaction Times: Universal Inferences<br />
56
Triggering Problem<br />
Questions<br />
Why do some elementary clauses have presuppositions?<br />
a. John knows that it is raining<br />
π: It is raining.<br />
b. John rightly believes that it is raining<br />
π: none, or possibly: John believes that it is raining.<br />
57
Projection Problem<br />
Questions<br />
How do the presuppositions <strong>of</strong> elementary clauses get<br />
transmitted to complex clauses ?<br />
a. If John is realistic, he knows that he is incompetent.<br />
π: John is incompetent<br />
b. If John is an idiot, he knows that he incompetent<br />
π: none, or possibly: if John is an idiot, he is incompetent<br />
58
Questions<br />
Architectural Question<br />
Where do presuppositions belong in the architecture or<br />
language?<br />
Are they a semantic or a pragmatic phenomenon?<br />
59
The Projection Problem<br />
60
Conjunction<br />
a. John knows that he is incompetent<br />
b. Is it true that John knows that he is incompetent?<br />
π: John is incompetent<br />
c. I doubt that John knows that he is incompetent<br />
π: John is incompetent<br />
d. None <strong>of</strong> these 10 students knows that he is incompetent.<br />
π: Each <strong>of</strong> these 10 students is incompetent.<br />
a. John is incompetent and knows that he is.<br />
b. Is it true that John is incompetent and knows that he is?<br />
π: none<br />
c. I doubt that John is incompetent and knows that he is.<br />
π: none<br />
d. None <strong>of</strong> these 10 students is incompetent and knows it.<br />
π: none<br />
61
Conjunction<br />
a. John is depressed and his boss knows that he is<br />
incompetent<br />
b. Is it true that John is depressed and that his boss knows<br />
that he is incompetent?<br />
π: John is incompetent<br />
c. I doubt that John is depressed and that his boss knows<br />
that he is incompetent.<br />
a. John is an idiot and his boss knows that he is<br />
incompetent.<br />
b. Is it true that John is an idiot and that his boss knows that<br />
he incompetent?<br />
π: if John is an idiot, he is incompetent (?)<br />
c. I doubt that John is an idiot and that his boss knows that<br />
he is incompetent.<br />
62
Conjunction<br />
p and qq’ presupposes p ⇒ q<br />
(... to be refined)<br />
John is incompetent and he knows it / that he is<br />
π: none<br />
John is an idiot and he knows that he is incompetent<br />
π: if John is an idiot, he is incompetent<br />
John is depressed and his boss knows that he is<br />
incompetent<br />
Predicted π: If John is depressed, he is incompetent<br />
Actual π: John is incompetent<br />
Maybe because: the most plausible way to make the<br />
conditional true is to assume that its consequent is!<br />
63
Conditionals<br />
a. If John is incompetent, he knows that he is.<br />
b. Is it true that if John is incompetent, he knows that he is?<br />
c. I doubt that if John is incompetent, he knows that he is.<br />
a. If John is realistic, he knows that he is incompetent.<br />
b. Is it true that if John is realistic, he knows that he is<br />
incompetent?<br />
c. I doubt that if John is realistic, he knows that he is<br />
incompetent.<br />
a. If John is over 65, he knows he can’t apply.<br />
b. Is it true that if John is over 65, he knows he can’t apply?<br />
c. I doubt that if John is over 65, he knows he can’t apply.<br />
64
Conditionals<br />
a. If John knows that he is overqualified, he won’t apply.<br />
b. Is it true that if John knows that he is overqualified, he<br />
won’t apply?<br />
c. I doubt that if John knows that he is overqualified, he<br />
won’t apply.<br />
a. If John knows that he is overqualified, he is depressed<br />
b. Is it true that if John knows that he is overqualified, he is<br />
depressed?<br />
c. I doubt that if John knows that he is overqualified, he is<br />
depressed.<br />
a. if p, qq’ presupposes p ⇒ q<br />
b. if pp’, q presupposes p<br />
65
Disjunctions<br />
a. If John is incompetent, he knows that he is.<br />
b. Either John is not incompetent, or he knows that he is.<br />
a. If John is realistic, he knows that he is incompetent.<br />
b.Either John is not realistic,or he knows he is incompetent.<br />
a. If John is over 65, he knows he can’t apply.<br />
b. Either John isn’t over 65, or he knows he can’t apply<br />
a. If John knows that he is overqualified, he won’t apply.<br />
b. Either John doesn’t know that he is over qualified, or he<br />
won’t apply.<br />
a. p or qq’ presupposes (not p) ⇒ q<br />
b. pp’ or q presupposes p<br />
66
Stalnaker’s Pragmatic Analysis<br />
67
A Pragmatic Analysis<br />
p and qq’ presupposes p ⇒ q<br />
‘... when a speaker says something <strong>of</strong> the form A and B, he<br />
may take it for granted that A (or at least that his audience<br />
recognizes that he accepts that A) after he has said it. The<br />
proposition that A will be added to the background <strong>of</strong> common<br />
assumptions before the speaker asserts that B.<br />
Now suppose that B expresses a proposition that would, for<br />
some reason, be inappropriate to assert except in a context<br />
where A, or something entailed by A, is presupposed. Even if A<br />
is not presupposed initially, one may still assert A and B<br />
since by the time one gets to saying that B, the context has<br />
shifted, and it is by then presupposed that A.’<br />
Stalnaker, ‘Pragmatic Presuppositions’, 1974<br />
68
Assumptions<br />
Assumption 1: Sentences may be true, false or #<br />
Assumption 2: A sentence S is a presupposition failure if it<br />
has the value # with respect to at least one <strong>of</strong> the states <strong>of</strong><br />
affairs compatible with what the speech act participants<br />
take for granted.<br />
Definition 1: Common Ground = what the speech act<br />
participants take for granted.<br />
Definition 2: Context Set = set <strong>of</strong> worlds compatible with<br />
what the speech act participants take for granted.<br />
Assumption 3: The Context Set is updated incrementally<br />
in discourse and in conjunctions.<br />
69
Possible Worlds<br />
A possible world w = a complete specification <strong>of</strong> what is<br />
going on. It determines for every sentence S whether<br />
[[ S ]] w = true, [[ S ]] w = false, or [[ S ]] w = #.<br />
Different clauses give rise to different functions, e.g.:<br />
The President <strong>of</strong><br />
France is Chirac<br />
w 1 → false<br />
w 2 → true<br />
w 3 → #<br />
w 4 → #<br />
...<br />
The US<br />
President is Bush<br />
w 1 → true<br />
w 2 → false<br />
w 3 → true<br />
w 4 → #<br />
...<br />
Two plus two<br />
is four<br />
w 1 → true<br />
w 2 → true<br />
w 3 → true<br />
w 4 → true<br />
70
Non-Contradiction<br />
Further Conditions<br />
A sentence S uttered in a Context Set C is deviant if S is<br />
true in no world <strong>of</strong> C.<br />
Non-Triviality<br />
A sentence S uttered in a Context Set C is deviant if S is<br />
true in every world <strong>of</strong> C.<br />
71
Stalnaker’s Analysis<br />
John knows that he is incompetent is:<br />
-true in w if John is incompetent and believes that he is<br />
-false in w if John is incompetent and doesn’t believe he is<br />
-# in w if John is not incompetent.<br />
Suppose that the speech act participants do not know<br />
whether John is or isn’t incompetent. Suppose further that<br />
the Context Set C is C = {w 1, w 2, w 3, w 4}<br />
w 1 : John is incompetent and believes that he is<br />
w 2: John is incompetent and believes he isn’t<br />
w 3: John is not incompetent but believes he is<br />
w 4: John is not incompetent and believes he isn’t<br />
72
Stalnaker’s Analysis<br />
T = John knows that he is incompetent uttered in C is a<br />
presupposition failure because this sentence is # in w 3 and<br />
w 4, which both belong to C<br />
Suppose that the speech act participants do not know<br />
whether John is or isn’t incompetent. Suppose further that<br />
the Context Set C is C = {w 1, w 2, w 3, w 4}<br />
w 1 : John is incompetent and believes that he is<br />
w 2: John is incompetent and believes he isn’t<br />
w 3: John is not incompetent but believes he is<br />
w 4: John is not incompetent and believes he isn’t<br />
73
Stalnaker’s Analysis<br />
S = John is incompetent is:<br />
-true in w if John is incompetent in w.<br />
-false in w in all other cases<br />
(i.e. the sentence does not have a presupposition)<br />
a. Acceptability<br />
Clearly, John is incompetent uttered in C is not a<br />
presupposition failure.<br />
b. Update<br />
-Initially, the Context Set was C = {w 1, w 2, w 3, w 4}<br />
-After S is uttered,<br />
the new Context Set is: C’ = {w 1, w 2}<br />
(i.e. only the worlds compatible with S are retained)<br />
74
Stalnaker’s Analysis<br />
John is incompetent. He knows it.<br />
= S. T.<br />
Step 1.<br />
-The initial Context Set is C = {w 1, w 2, w 3, w 4}<br />
-After the first sentence is uttered,<br />
the new Context Set is C’ = {w 1, w 2}<br />
Step 2.<br />
-The second sentence is evaluated with respect to C’<br />
-By construction, in each world in C’, T has a value<br />
different from #. So T is not a presupposition failure in C’.<br />
Step 3.<br />
C’ is updated to C” = {w 1} .<br />
75
Conjunction<br />
Stalnaker’s Analysis<br />
a. Treat S and T in the same way as the discourse S. T: the<br />
assertion <strong>of</strong> a conjunction is a succession <strong>of</strong> two assertions.<br />
b. Beautiful analysis <strong>of</strong> presupposition projection:<br />
every world in C that satisfies S must satisfy T.<br />
In other words: C |= S ⇒ T<br />
Limitations<br />
a. How does the analysis extend to other operators?<br />
b. How does the analysis extend to embedded<br />
conjunctions?<br />
e.g. None <strong>of</strong> my students is rich and proud <strong>of</strong> it.<br />
76
Heim’s Semantic Analysis<br />
(following in part Karttunen 1974)<br />
77
Karttunen I: The Limits <strong>of</strong> Brute Force<br />
78
Karttunen II: Admittance Conditions<br />
a. Brute Force Method<br />
define recursively the (complex!) rules by which the<br />
presuppositions <strong>of</strong> complex sentences are computed on the<br />
basis <strong>of</strong> the presuppositions <strong>of</strong> their parts.<br />
b. Admittance Conditions<br />
(i) take as primitive the notion <strong>of</strong> a context satisfying the<br />
presuppositions <strong>of</strong> an elementary clause.<br />
(ii) extend recursively the notion <strong>of</strong> satisfaction.<br />
79
Heim’s Synthesis<br />
Karttunen<br />
Separate specification <strong>of</strong>:<br />
(i) admittance conditions<br />
(ii) truth-conditional (assertive) content.<br />
Gazdar’s critique (<strong>of</strong> Karttunen & Peters): this is not<br />
explanatory!<br />
Heim<br />
a. The ‘context change potential’ <strong>of</strong> an expression cannot<br />
be derived from its assertive content.<br />
b. But its assertive content cannot be derived from its<br />
context-change potential.<br />
(... once one has the right context change potential!!!)<br />
80
Heim’s Synthesis<br />
Heim vs. Stalnaker<br />
a. Keep from Stalnaker’s analysis<br />
-the idea <strong>of</strong> an update<br />
-the analysis <strong>of</strong> presupposition projection in conjunctions<br />
b. Drop the pragmatic derivation <strong>of</strong> Stalnaker’s analysis.<br />
Heim vs. Karttunen<br />
-In Karttunen’s system, admittance conditions are specified<br />
separately from the assertive content <strong>of</strong> expressions.<br />
-For Heim, Context Change Potentials do double duty.<br />
The <strong>Dynamic</strong> Conception <strong>of</strong> Meaning<br />
-Old conception: meanings as truth conditions<br />
-New conception: meanings as Context Change Potentials,<br />
i.e. as functions from Context Sets to Context Sets.<br />
81
Heim’s Synthesis<br />
Notation: C[F] = update <strong>of</strong> the Context Set C with F<br />
Elementary Clauses<br />
a. C[John is incompetent]<br />
= # iff C = #<br />
= {w∈C: John is incompetent in w} otherwise<br />
b. C[John knows that he is incompetent]<br />
= # iff C=# or for some w∈C, John is not incompetent in w<br />
= {w∈C: John believes he is incompetent in w}, otherwise<br />
Truth<br />
If C[S] ≠ # and w∈C, then: S is true at w iff w ∈ C[S]<br />
Conjunction<br />
C[F and G] = C[F][G]<br />
82
Heim’s Synthesis<br />
Negation<br />
C[not F] = # iff C[F] = #<br />
= C - C[F] otherwise<br />
a. not F = John doesn’t know that he is incompetent.<br />
b. C[not F] = # iff C[F] = #, iff for some w∈C, John is not<br />
incompetent in w<br />
= C - C[F] otherwise,<br />
i.e. = C - {w∈C: John believes he is incompetent in w}<br />
F<br />
83
Heim’s Synthesis<br />
Negation<br />
C[not F] = # iff C[F] = #<br />
= C - C[F] otherwise<br />
This means that<br />
not(pp’)<br />
presupposes that p<br />
a. not F = John doesn’t know that he is incompetent.<br />
b. C[not F] = # iff C[F] = #, iff for some w∈C, John is not<br />
incompetent in w<br />
= C - C[F] otherwise,<br />
i.e. = C - {w∈C: John believes he is incompetent in w}<br />
F<br />
84
Heim’s Synthesis<br />
Negation<br />
C[not F] = # iff C[F] = #<br />
= C - C[F] otherwise<br />
a. not F = John doesn’t know that he is incompetent.<br />
b. C[not F] = # iff C[F] = #, iff for some w∈C, John is not<br />
incompetent in w<br />
= C - C[F] otherwise,<br />
i.e. = C - {w∈C: John believes he is incompetent in w}<br />
F<br />
85
Heim’s Synthesis<br />
Conditionals (analyzed as material implications)<br />
F<br />
C[if F, G] = # iff C[F] = # or C[F][not G] = #<br />
= C - C[F][not G], otherwise<br />
Worlds that<br />
refute<br />
if F, G<br />
G<br />
86
Heim’s Synthesis<br />
Conditionals (analyzed as material implications)<br />
F<br />
C[if F, G] = # iff C[F] = # or C[F][not G] = #<br />
= C - C[F][not G], otherwise<br />
This means that if pp’, q presupposes that p,<br />
and that if p, qq’, presupposes if p, q<br />
Worlds that<br />
refute<br />
if F, G<br />
G<br />
87
Heim’s Synthesis<br />
Conditionals (analyzed as material implications)<br />
F<br />
C[if F, G] = # iff C[F] = # or C[F][not G] = #<br />
= C - C[F][not G], otherwise<br />
Worlds that<br />
refute<br />
if F, G<br />
G<br />
88
Heim’s Synthesis<br />
if F, G = If John is incompetent, he knows it<br />
C[if F, G] = # iff C[F] = # or C[F][not G] = #<br />
But C[F] ≠ # and furthermore<br />
C[F] = {w∈C: John is incompetent in w}<br />
C[F][not G] = # iff C[F][G] = #, which is not the case (by<br />
construction). Furthermore,<br />
C[F][not G] = {w∈C: John is incompetent in w}[not G]<br />
= {w∈C: John is incompetent in w}<br />
- {w∈C: John is incompetent in w and John believes he is<br />
incompetent in w}<br />
= {w∈C: John is incompetent but doesn’t believe it in w}<br />
C[if F, G] = C - {w∈C: John is incompetent but doesn’t<br />
believe it in w}<br />
89
Summary<br />
Meaning <strong>of</strong> an elementary clause = a CCP<br />
Conjunction<br />
C[F and G] = C[F][G]<br />
Negation<br />
C[not F] = # iff C[F] = #; = C - C[F] otherwise<br />
Conditionals<br />
C[if F, G] = # iff C[F] = # or C[F][not G] = #<br />
= C - C[F][not G], otherwise<br />
Disjunction<br />
C[F or G] = # iff C[F] = # or C[not F][G] = #<br />
= C[F] ∪ C[not F][G], otherwise<br />
90
Disjunctions<br />
a. If John is incompetent, he knows that he is.<br />
b. Either John is not incompetent, or he knows that he is.<br />
a. If John is realistic, he knows that he is incompetent.<br />
b.Either John is not realistic,or he knows he is incompetent.<br />
a. If John is over 65, he knows he can’t apply.<br />
b. Either John isn’t over 65, or he knows he can’t apply<br />
a. If John knows that he is overqualified, he won’t apply.<br />
b. Either John doesn’t know that he is over qualified, or he<br />
won’t apply.<br />
a. p or qq’ presupposes (not p) ⇒ q<br />
b. pp’ or q presupposes p<br />
91
Heim’s Analysis<br />
Disjunction<br />
C[F or G] = # iff C[F] = # or C[not F][G] = #<br />
= C[F] ∪ C[not F][G] otherwise.<br />
F<br />
a. John is not incompetent, or he knows that he is.<br />
b. C[not I or K] = # iff C[not I] = # or C[not not I][K] = #,<br />
i.e. iff C[I] = # or C[I][K] = #, which is never the case.<br />
Thus C[not I or K] = C[not I] ∪ C[I][K]<br />
G<br />
92
Heim’s Analysis<br />
Disjunction<br />
C[F or G] = # iff C[F] = # or C[not F][G] = #<br />
F<br />
= C[F] ∪ C[not F][G] otherwise.<br />
a. John is not incompetent, or he knows that he is.<br />
b. C[not I or K] = # iff C[not I] = # or C[not not I][K] = #,<br />
i.e. iff C[I] = # or C[I][K] = #, which is never the case.<br />
Thus C[not I or K] = C[not I] ∪ C[I][K]<br />
G<br />
This means that<br />
pp’ or q presupposes<br />
that p,<br />
and that p or<br />
qq’presupposes<br />
if (not p), q<br />
93
Heim’s Analysis<br />
Disjunction<br />
C[F or G] = # iff C[F] = # or C[not F][G] = #<br />
= C[F] ∪ C[not F][G] otherwise.<br />
F<br />
a. John is not incompetent, or he knows that he is.<br />
b. C[not I or K] = # iff C[not I] = # or C[not not I][K] = #,<br />
i.e. iff C[I] = # or C[I][K] = #, which is never the case.<br />
Thus C[not I or K] = C[not I] ∪ C[I][K]<br />
G<br />
94
Heim’s Analysis<br />
Definition <strong>of</strong> Truth<br />
If w∈C,<br />
a. F is # in w relative to C iff C[F] = #<br />
b. If ≠ #, F is true in w relative to C iff w∈C[F]<br />
John is incompetent. He knows it.<br />
= S. T.<br />
C = {w 1, w 2, w 3, w 4}<br />
C[S] = C’ = {w 1, w 2}<br />
C[S][T] = C” = {w 1} .<br />
a. Relative to w1, C, the discourse is true, since w1∈C[S][T] b. Relative to w2, C, the discourse is false , since<br />
w2∉C[S][T] 95
Heim’s Explanatory Problem<br />
Problem: is the account explanatory? (Soames 1989)<br />
C[F and G] = (C[F])[G]<br />
C[F and* G] = (C[G])[F]<br />
When F and G are not presuppositional,<br />
C[F and G]=C[F and* G]={w∈C: F is true in w and G is<br />
true in w}<br />
96
Heim’s Explanatory Problem<br />
There are many ways to define the CCP <strong>of</strong> or...<br />
F<br />
C[F or 1 G] = C[F] ∪ C[G], unless one <strong>of</strong> those is #<br />
C[F or 2 G] = C[F] ∪ C[not F][G], unless one <strong>of</strong> those is #<br />
C[F or 3 G] = C[not G][F] ∪ C[G], unless one <strong>of</strong> those is #<br />
G<br />
97
Gazdar’s Account<br />
98
Non-Contradiction<br />
Reminder 1: Non-Triviality<br />
A sentence S uttered in a Context Set C is deviant if S is<br />
true in no world <strong>of</strong> C.<br />
Non-Triviality<br />
A sentence S uttered in a Context Set C is deviant if S is<br />
true in every world <strong>of</strong> C.<br />
99
Reminder 2: Other Implicatures<br />
John is in Paris or he is in Rome<br />
=> it is not the case that:<br />
a. the speaker believes that John is in Paris.<br />
b. the speaker believes that John is not in Paris.<br />
c . the speaker believes that John is in Rome.<br />
d. the speaker believes that John is not in Rome.<br />
If John is in Paris, he is there for business.<br />
=> the speaker takes it to be possible but not certain that<br />
John is in Paris<br />
100
An Explanatory Account ?<br />
Step 1: Compute the various implicatures <strong>of</strong> a sentence<br />
Step 2: Keep only those presuppositions that are<br />
consistent with all implicatures.<br />
John is incompetent and he knows that he is.<br />
a. Implicature: If John is incompetent is uttered, it<br />
cannot be trivial that John is incompetent,<br />
i.e. C |≠ John is incompetent<br />
b. Potential Presupposition: the second conjunct triggers<br />
the potential presupposition that John is incompetent.<br />
c. Filtering: The presupposition is filtered out because it<br />
is inconsistent with the implicature.<br />
101
An Explanatory Account ?<br />
John is depressed and he knows that he is incompetent.<br />
a. Implicature: If John is depressed is uttered, it cannot<br />
be trivial that John is depressed,<br />
i.e. C |≠ John is depressed<br />
b. Potential Presupposition: the second conjunct triggers<br />
the potential presupposition that John is incompetent.<br />
c. Filtering: The presupposition is not filtered out because<br />
it is consistent with the implicature.<br />
Note: Gazdar thus predicts that the entire sentence<br />
presupposes that John is depressed. Stalnaker and Heim<br />
predict: if John is depressed, he is incompetent.<br />
Most examples go in Gazdar’s direction.<br />
102
An Explanatory Account ?<br />
If John is incompetent, he knows it.<br />
a. Implicature: The speaker cannot utter If F, G<br />
felicitously if he knows that F is true. If we represent as S<br />
the set <strong>of</strong> worlds compatible with what the speaker believes<br />
S |≠ John is incompetent<br />
from which it follows that<br />
C |≠ John is incompetent.<br />
b. Potential Presupposition: the main clause triggers the<br />
potential presupposition that John is incompetent.<br />
c. Filtering: The presupposition is filtered out because it<br />
is inconsistent with the implicature.<br />
103
Reminder: Other Implicatures<br />
John is in Paris or he is in Rome<br />
=> it is not the case that:<br />
a. the speaker believes that John is in Paris.<br />
b. the speaker believes that John is not in Paris.<br />
c . the speaker believes that John is in Rome.<br />
d. the speaker believes that John is not in Rome.<br />
If John is in Paris, he is there for business.<br />
=> the speaker takes it to be possible but not certain that<br />
John is in Paris<br />
104
An Explanatory Account ?<br />
Either John is not incompetent, or he knows that he is.<br />
a. Implicature: The speaker cannot utter F or G<br />
felicitously if he believes that F is false<br />
S |≠ John is incompetent<br />
from which it follows that<br />
C |≠ John is incompetent.<br />
b. Potential Presupposition: the second clause triggers the<br />
potential presupposition that John is incompetent.<br />
c. Filtering: The presupposition is filtered out because it<br />
is inconsistent with the implicature.<br />
Note: Gazdar thus predicts that the entire sentence<br />
presupposes that John is depressed. Stalnaker and Heim<br />
predict: if John is depressed, he is incompetent.<br />
105
An Explanatory Account ?<br />
If John is depressed, he knows that he is incompetent.<br />
a. Implicature: S |≠ John is depressed<br />
from which it follows that<br />
C |≠ John is depressed<br />
b. Potential Presupposition: the main clause triggers the<br />
potential presupposition that John is incompetent.<br />
c. Filtering: The presupposition is not filtered out because<br />
it is consistent with the implicature.<br />
Note: Gazdar thus predicts that the entire sentence<br />
presupposes that John is depressed. Stalnaker and Heim<br />
predict: if John is depressed, he is incompetent.<br />
Most examples go in Gazdar’s direction - but not all do!<br />
106
Problem for Gazdar’s Account<br />
If John is French, he must know that he can travel within<br />
the European Union without a passport.<br />
a. Gazdar’s prediction: π = John can travel within the<br />
European Union without a passport.<br />
b. Actual presupposition: probably none.<br />
a. Implicature: S |≠ John is French<br />
from which it follows that C |≠ John is French<br />
b. Potential Presupposition: the main clause triggers the<br />
potential presupposition that John can travel within the<br />
European Union without a passport.<br />
c. Filtering: The presupposition is not filtered out because<br />
it is consistent with the implicature!<br />
If John has twins, then Mary will not like his children.<br />
107
A Very Partial History<br />
1973-1974<br />
-Stalnaker’s analysis: pragmatics + local contexts.<br />
-Karttunen’s analysis: recursive admittance conditions +<br />
local contexts.<br />
1970’s<br />
-Karttunen & Peters<br />
-Gazdar’s recursive pragmatics<br />
1980’s<br />
-Heim’s theory <strong>of</strong> presupposition projection<br />
-Overgeneration problem (Soames, Rooth).<br />
1990’s<br />
van der Sandt & Geurts’s critique <strong>of</strong> Heim. DRT analysis<br />
108
Back to Heim’s Account!<br />
Accommodation<br />
109
My sister is pregnant.<br />
Global Accommodation<br />
'... it's not as easy as you might think to say something that<br />
will be unacceptable for lack <strong>of</strong> required presuppositions.<br />
Say something that requires a missing presupposition, and<br />
straightway that presupposition springs into existence,<br />
making what you said acceptable after all.' I said that<br />
presupposition evolves in a more or less rule-governed way<br />
during a conversation. Now we can formulate one<br />
important governing rule: call it the<br />
Rule <strong>of</strong> accommodation for presupposition<br />
If at time t something is said that requires presupposition P<br />
to be acceptable, and if P is not presupposed just before t,<br />
then - ceteris paribus and within certain limits -<br />
presupposition P comes into existence at t."<br />
110
Local Accommodation<br />
a. The king <strong>of</strong> France is not wise because there is no king <strong>of</strong><br />
France.<br />
b. None <strong>of</strong> my students takes good care <strong>of</strong> his car because<br />
none <strong>of</strong> my students has a car!<br />
c. John doesn't know that he is incompetent because he just<br />
isn't incompetent!<br />
a. It's not the case there is a king <strong>of</strong> France and he is wise<br />
because ...<br />
b. None <strong>of</strong> my students has a car and takes good care <strong>of</strong> it<br />
because...<br />
c. It's not the case that John is incompetent and knows it ...<br />
Question: can we do without Local Accommodation by<br />
appealing to meta-linguistic uses <strong>of</strong> various operators?<br />
111
Global vs. Local Accommodation<br />
C[not F]= # iff C[F]=#<br />
= C - C[F], otherwise.<br />
Global Accommodation:<br />
C' = {c∈C: France is a monarchy at the time and in the<br />
world <strong>of</strong> c}.<br />
We then compute C'[the king <strong>of</strong> France is not powerful].<br />
Local Accommodation: Instead <strong>of</strong> computing<br />
C - C[F] (which wouldn't even be defined, since C[F]=#),<br />
we compute:<br />
C - C'[F], where C'={c∈C: France is a monarchy at the time<br />
and in the world <strong>of</strong> c} (as in A.)<br />
112
Directions<br />
Allow for local accommodation whenever global<br />
accommodation would contradict<br />
a. the literal meaning <strong>of</strong> a sentence<br />
b. or an implicature <strong>of</strong> a sentence [or possibly: certain types<br />
<strong>of</strong> implicatures, e.g. primary implicatures]<br />
In effect, this allows us to capture the good properties <strong>of</strong><br />
Gazdar’s system within Heim’s dynamic semantics.<br />
113
Summary<br />
Presuppositions cannot be analyzed as implicatures.<br />
The dilemma <strong>of</strong> dynamic semantics<br />
a. Stalnaker’s approach is explanatory but not general<br />
Update the context set in time as you process a sentence.<br />
b. Heim’s approach is general but not explanatory<br />
The meaning <strong>of</strong> words is dynamic from the start, i.e. their<br />
lexical entries specify how they change the context set.<br />
Gazdar’s account was explanatory and general but<br />
incorrect<br />
(i) Compute the implicatures <strong>of</strong> a sentence.<br />
(ii) Project those potential presuppositions that don’t<br />
contradict the entire sentence or one <strong>of</strong> its implicatures.<br />
114
The Proviso Problem<br />
a. If the problem was easy, it is not John who solved it.<br />
b. John knows that if the problem was easy, someone<br />
solved it (Geurts 1999)<br />
Predicted presupposition <strong>of</strong> (a) and (b):<br />
If the problem was easy, someone solved it<br />
Actual presupposition <strong>of</strong> (a)<br />
Someone solved the problem<br />
Actual presupposition <strong>of</strong> (b)<br />
If the problem was easy, someone solved it<br />
115
The Proviso Problem<br />
John is an idiot and he knows that he is incompetent<br />
π: if John is an idiot, he is incompetent<br />
John is depressed and he knows that he is incompetent<br />
Predicted π: If John is depressed, he is incompetent<br />
Actual π: John is incompetent<br />
Maybe this is because the most plausible way to make<br />
the conditional true is to assume that its consequent is!<br />
... but this kind <strong>of</strong> reasoning fails to address the minimal<br />
difference between:<br />
-If the problem was easy, it is not John who solved it<br />
-John knows that if the problem was easy, someone solved<br />
it (Geurts 1999).<br />
116
The Proviso Problem<br />
Direction 1 (van der Sandt 1992, Geurts 1999)<br />
-This problem refutes the standard dynamic approaches -<br />
as well as all approaches that make similar predictions.<br />
-A different analysis must be proposed, in which<br />
presuppositions are treated in a more syntactic fashion<br />
(‘Discourse Representation Theory’)<br />
This is a major contender among current theories.<br />
Direction 2 (still promissory)<br />
With enough pragmatic reasoning, we can stick to Heim’s<br />
predictions - which in any event seem to be correct in other<br />
cases, e.g.<br />
If John is over 65, he must know that he is too old to apply<br />
117
Back to Heim’s Account!<br />
Quantification<br />
118
Replacing Worlds with Contexts<br />
Example1. An amnesiac gets lost...<br />
“An amnesiac, Rudolf Lingens, is lost in the Stanford library. He<br />
reads a number <strong>of</strong> things in the library, including a biography <strong>of</strong><br />
himself, and a detailed account <strong>of</strong> the library in which he is lost...<br />
He still won’t know who he is, and where he is, no matter how<br />
much knowledge he piles up, until that moment when he is ready to<br />
say, “This place is aisle five, floor six, <strong>of</strong> Main Library, Stanford. I<br />
am Rudolf Lingens.” [Perry 1977]<br />
“It seems that the Stanford library has plenty <strong>of</strong> books, but no<br />
helpful little maps with a dot marked “location <strong>of</strong> this map.” Book<br />
learning will help Lingens locate himself in logical space. (...) But<br />
none <strong>of</strong> this, by itself, can guarantee that he knows where in the<br />
world he is. He needs to locate himself not only in logical space but<br />
also in ordinary space”. [Lewis 1979 p. 138]<br />
119
Standford Harvard<br />
120
Standford Harvard<br />
121
Example 2. 'My pants are on fire'<br />
‘If I see, reflected in a window, the image <strong>of</strong> a man whose<br />
pants appear to be on fire, my behavior is sensitive to<br />
whether I think, ‘His pants are on fire’, or ‘My pants are on<br />
fire’, though the object <strong>of</strong> thought may be the same'<br />
(Kaplan)<br />
122
123
Referential Uncertainty<br />
Situation: Lingens, who is lost in the Stanford library, knows<br />
everything there is to know about the world.<br />
I wear a coat. My coat is black.<br />
[Lingens, a well-read amnesiac, knows everything there is to<br />
know about the world; but he does not know whether he is<br />
Alfred, who is having a conversation with Berenice, or<br />
Charles, who is having a conversation with Denise.<br />
Berenice used to smoke but Denise never did]<br />
Compare:<br />
Did you stop smoking?<br />
You are Berenice. Did you stop smoking?<br />
124
Referential Uncertainty<br />
Situation: Lingens, who is lost in the Stanford library, knows<br />
everything there is to know about the world.<br />
I wear a coat. My coat is black.<br />
[Lingens, a well-read amnesiac, knows everything there is to<br />
know about the world; but he does not know whether he is<br />
Alfred, who is pointing towards Berenice, or<br />
Charles, who is pointing towards Denise.<br />
Berenice used to smoke but Denise never did]<br />
Compare:<br />
Did she stop smoking?<br />
She is Berenice. Did she stop smoking?<br />
125
Static Account with worlds<br />
[[ it is raining ]] w = false<br />
[[ PS is in Los Angeles ]] w = true,<br />
[[ the British President is happy]] w = #<br />
Rule<br />
[[ Pro VP ]] w = true if and only if [[ Pro ]] w + w ∈ [[VP ]]<br />
where [[VP ]] + w is the set <strong>of</strong> things <strong>of</strong> which VP is true in w<br />
[[ Pro VP ]] w = false if and only if [[ Pro ]] w - w ∈ [[VP ]]<br />
where [[VP ]] - w is the set <strong>of</strong> things <strong>of</strong> which VP is false in w<br />
[[ Pro VP ]] w = # in all other cases!<br />
126
Static Account with contexts<br />
A context = <br />
<br />
[[ I smoke ]]<br />
= true if and only if PS smokes in w<br />
= false if and only PS does not smoke in w<br />
[[ She 2 smokes]]<br />
<br />
= true if and only if Mary smokes in w<br />
= false if and only Mary does not smoke in w<br />
[[ She 2 stopped smoking]]<br />
<br />
= true if and only if Mary used to smoke but doesn’t now in w<br />
= false if and only Mary used to smoke and still does in w<br />
= # if and only if Mary didn’t use to smoke.<br />
127
<strong>Dynamic</strong> Account with contexts<br />
The rules are exactly the same as before, replacing<br />
worlds with... contexts!<br />
Elementary Clauses [now C is a set <strong>of</strong> contexts]<br />
We write as c w the world <strong>of</strong> c, as c(1) the denotation <strong>of</strong> pro 1<br />
a. C[John is incompetent]<br />
= # iff C = #<br />
= {c∈C: John is incompetent in c w} otherwise<br />
b. C[John knows that he is incompetent]<br />
=# iff C = # or for some c∈C, John is not incompetent in c w<br />
= {c∈C: John believes he is incompetent in c w}, otherwise.<br />
c. C[she 2 stopped smoking]<br />
= # iff C = # or for some c∈C, c(2) didn’t smoke in c w<br />
= {c∈C: c(2) doesn’t smoke in c w}, otherwise.<br />
128
Quantification in a Static Setting<br />
[[ [no x 1: x 1 student] x 1 smokes]]<br />
≠ # iff for every d which is a student in w,<br />
[[x 1 smokes ]] ≠ #.<br />
If ≠ #,<br />
= true iff for no d which is a student in w,<br />
<br />
[[x 1 smokes]] = true.<br />
= false iff for some d which is a student in w,<br />
[[x 1 smokes]] = true.<br />
129
Quantification in a Static Setting<br />
<br />
[[[no x1: x1 student] x1 stopped smoking]]<br />
≠ # iff for every d which is a student in w,<br />
[[x 1 stopped smoking]] ≠ #.<br />
If ≠ #,<br />
= true iff for no d which is a student in w,<br />
[[x 1 stopped smoking ]] = true.<br />
= false iff for some d which is a student in w,<br />
[[x 1 stopped smoking ]] = true.<br />
130
Quantification in a <strong>Dynamic</strong> Setting<br />
Notations<br />
c[i → d] = that context which is exactly like c except that<br />
pro i denotes d<br />
C[i → d] = {c[i → d]: c∈C}<br />
C[[no x i: x i NP] x iVP] = # iff<br />
C = # or {c[i→d]: c∈C and d is an object}[x i NP] = #<br />
or {c[i→d]: c∈C and c[i→d] ∈ C[i→d][x i NP]} [x i VP] = #.<br />
If ≠ #, C[[no x i: x i NP] x i VP] = {c: c∈C and for no object<br />
d, c[i→d] ∈ C[i→d][x i NP] and c[i→d] ∈ C[i→d][x i<br />
NP][x i VP]}<br />
C[[every xi: xi NP] xiVP]: same thing as for no ..., replacing<br />
no with every.<br />
131
Quantification in a <strong>Dynamic</strong> Setting<br />
[no x 1: x 1 student] x 1 smokes<br />
Let us assume that C ≠ #. Then:<br />
C[[no x 1: x 1 student] x 1 smokes] ≠ # because<br />
C = #, {c[1→d]: c∈C and d is an object}[x 1 student] ≠ #,<br />
and {c[1→d]: c∈C and c[1→d] ∈ C[1→d][x 1 student]} [x 1<br />
smokes] ≠ #. Furthermore,<br />
C[[no x 1: x 1 student] x 1 smokes]<br />
= {c: c∈C and for no object d, c[1→d] ∈ C[1→d][x 1<br />
student] and c[1→d] ∈ C[1→d][x 1 student][x 1 smokes]}<br />
= {c: c∈C and for no object d, d is a student in c w and d<br />
smokes in c w}<br />
132
Quantification in a <strong>Dynamic</strong> Setting<br />
[no x1: x1 student] x1 smokes<br />
Let us assume that C = {c1, c2, c3, c4} and for each i, ci = , with:<br />
w1 : All students used to smoke. All students still smoke.<br />
w2: All students used to smoke. One doesn’t any more.<br />
w3: One student didn’t use to smoke. No student smokes.<br />
w4: One student didn’t use to smoke. One student smokes.<br />
C[[no x1: x1 student] x1 smokes]<br />
= {ci: i ∈ {1, 2, 3, 4} and for no object d, ∈ {: i ∈ {1, 2, 3, 4}}[x1 student] and<br />
∈ {: i ∈ {1, 2, 3,<br />
4}}[x1 student][x1 smokes]}<br />
= {ci: i ∈ {1, 2, 3, 4} and for no object d, d is a student in<br />
wi and d smokes in wi} = {c3} 133
Quantification in a <strong>Dynamic</strong> Setting<br />
[no x 1: x 1 student] x 1 stopped smoking<br />
Let us assume that C ≠ #. Then:<br />
C[[no x1: x1 student] x1stopped smoking] = #<br />
iff {c[1→d]: c∈C and d is an object}[x1 student] = #,<br />
or {c[1→d]: c∈C and c[1→d] ∈ C[1→d][x1 student]} [x1 stopped smoking] = #,<br />
iff {c[1→d]: c∈C and c[1→d] ∈ C[1→d][x1 student]} [x1 stopped smoking] = #<br />
iff for some c∈C, for some d, d is a student in cw and d<br />
didn’t use to smoke in cw. If ≠ #,<br />
= {c: c∈C and for no object d, d is a student in cw and d<br />
stopped smoking in cw} 134
Quantification in a <strong>Dynamic</strong> Setting<br />
[no x i: x i student] x i stopped smoking<br />
Let us assume that C = {c 1, c 2, c 3, c 4}<br />
(with c 1, c 2, c 3, c 4 defined as before)<br />
C[[no x 1: x 1 student] x 1 stopped smoking] = # because<br />
{c[1→d]: c∈C and c[1→d] ∈ C[1→d][x 1 student]} [x 1<br />
stopped smoking]<br />
= {: i ∈ {1, 2, 3, 4} and d is a student in<br />
w i} [x 1 stopped smoking]<br />
= # because in w 3 and w 4 there are students who didn’t use<br />
to smoke.<br />
135
Quantification in a <strong>Dynamic</strong> Setting<br />
[no x i: x i student] x i stopped smoking<br />
Let us now assume that C’ = {c 1, c 2}<br />
(with c 1 and c 2 defined as before)<br />
It can be shown C’[[no x 1: x 1 student] x 1 smokes] ≠ #.<br />
Furthermore, C’[[no x 1: x 1 student] x 1 stopped smoking]<br />
= {c: c∈C’ and for no object d, d is a student in c w and d<br />
stopped smoking in c w}<br />
= {c 1}<br />
136