November 12th, 2011
42nd Meeting of the North East Linguistics Society
From Quantification and Intensification to Slack Regulation: Adjectival ALL Heather Burnett (UCLA/ENS, Paris)1
[email protected] Lucille: How’s my son? Dr. Fishman: He’s going to be all right. Oscar: Oh thank god. Dr. Fishman: That’s a great attitude. I gotta tell you, if I was given this news, I don’t know if I would take it this well. Lucille: But. . . you said he was alright. Dr. Fishman: Yes. He’s lost his left hand, so he’s going to be all right. –Arrested Development. S2.E12. (Hand to God )
1
Introduction
This paper presents a novel empirical contribution to the open debate concerning the syntactic and semantic analysis of the lexical item all and its cognates in the Indo-European languages (henceforth ALL). • A particular focus on the Romance family (ex. French tou(te)s, Italian tutto, Romanian tot- etc.). (1)
a. b.
French: Toutes les filles ont saut´e dans le lac. ALL the girls have jumped in the lake Italian: Tutte le ragazze sono saltate nel lago. ALL the girls were jumped in the lake ‘All the girls jumped in the lake.’
• A secondary focus on some languages that have an all/whole alternation (ex. German alle/ganz ). (2)
German: Der ganze Apfel ist rot. The ALL apple is red. ‘All of the apple is red/The whole apple is red.’
´ e, Thomas Graf, Ed Keenan, Hilda Koopman, Jessica I would like to thank Melanie Bervoets, Paul Egr´ Rett, Emanuela Sanfelici, and Dominique Sportiche for comments and grammaticality judgements. This research was funded in part by the following grants from the Social Sciences and Humanities Research Council of Canada: Doctoral fellowship (#752-2007- 2382 (H. Burnett)), the MCRI Mod´eliser le changement: les voies du fran¸cais (#412-2004-1002 (F. Martineau)), and a France-USA Partner University Fund (PUF) ´ grant between the Ecole normale sup´erieure and UCLA. 1
1
November 12th, 2011
1.1
42nd Meeting of the North East Linguistics Society
The Debate: Quantifier or Slack Regulator?
Quantifier: The majority view in the literature is that all and its cognates are universal quantifiers (∀). • Recent analyses in this vein have been proposed by Winter (2001), Zweig (2008), and Champollion (2011). Slack Regulator: An alternative view proposes that the contribution of all to the meaning of the sentence is pragmatic, not quantificational. • Advocated primarily by Lasersohn (1999) (and Brisson (1998), Brisson (2003)). • Lasersohn (1999): All is a slack regulator (an item that eliminates (or significantly reduces) ‘pragmatic slack’ or vagueness/imprecision associated with its complement). – ALL forces a more precise interpretation of its complement.
2
My Proposal: Slack Regulator
A new argument in favour of the SR analysis: Adjectival ALL in Romance and Germanic. • From a typological perspective, English all has a very limited distribution, appearing primarily in the nominal domain (attaching to DPs (1)). – Moltmann (1997): English all/whole is ‘deficient’ compared to its cognates other languages. • In many other languages, ALL appears in the adjectival domain2 . (3)
Adjectival ALL a. French: La salle est toute vide. The room is ALL empty b. Italian: La stanza `e tutta vuota. The room is ALL empty c. German: Das Zimmer ist ganz leer. The room is ALL empty ‘The room is completely empty’ (lit. all empty)
I argue: 1. A universal quantifier approach cannot account for the interpretative pattern observed in the adjectival domain. • We find interpretations of adjectival ALL that are intensive, not universal. 2
In some languages, it can even even appear in the verbal domain (cf. Moltmann (1997) for German ganz and Lemieux et al. (1985) for Qu´eb´ecois tout).
2
November 12th, 2011
42nd Meeting of the North East Linguistics Society
• These interpretations are systematic and predictable based on the scale structure class that the adjective belongs to. 2. The observed distribution and interpretations of ALL with both adjectives and DPs is straightforwardly expected under an analysis of this lexical item as a slack regulator that eliminates the borderline cases of its complement. 3. Appendix: I give a formal version of such an analysis in Cobreros et al. (2010)’s logical framework.
2.1
Plan
1. The Quantification vs Slack Regulation Debate. • Lasersohn’s Pragmatic Slack. 2. The Data: Adjectival ALL. 3. A homophony analysis? 4. A new slack regulation analysis. 5. Appendix 1: Slack Regulation in Tolerant, Classical, Strict. 6. Appendix 2: Comparison with Lasersohn (1999).
3
ALL: Quantifier or SR?
3.1
The Traditional View
V All is a universal quantifier (∀/ ). (4) (5)
Kalish et al. (1980) (p. 130): If all persons are mortal, then either there are exceptions or Socrates is mortal. V ( x(F x → Gx) → (P ∨ GA))
But all cannot have exactly the same meaning as ∀ in FOL: • All is possible with collective predicates like gather and meet. (6)
All the girls gathered. 6↔ ∀x(girl(x) → gather(x)) Furthermore, the distribution of all in the nominal domain is somewhat complicated: • Dowty (1987): Possible with some collective predicates (6), but not with others (7).
(7)
*All the girls are numerous.
Much of the work in this area has involved proposing more subtle quantificational analyses that can account for patterns like (6) vs (7) (cf. Winter (2001), Champollion (2011) among others). 3
November 12th, 2011
3.2
42nd Meeting of the North East Linguistics Society
The Pragmatic View
Almost since the beginning of the logical approach to natural language semantics, researchers have been skeptical of the ‘universal quantifier’ analysis for all (cf. Link (1983)). One argument: The puzzling distribution of all : • Brisson (1998) among others: (In English) all only attaches to constituents that are already quantified. Example. In most theories of the semantics of sentences with definite plurals and distributive predicates, the subject DP is universally quantified by a distributivity operator. • Denying the universality of (9) results in a contradiction (10) (Kroch, 1974), (Lasersohn, 1999). (8)
The townspeople are asleep.
(9)
# Although the townspeople are asleep, some of them are awake.
(10)
All the townspeople are asleep.
If not quantification, what is ALL’s contribution to the meaning of the utterance that it occurs in? 3.2.1
Pragmatic Slack
Depending on context, speakers might very naturally utter (9) if some small percentage of the townspeople is awake. • Sentences with definite plurals allow a certain amount of pragmatic slack (Lasersohn, 1999)/non-maximality effects (Brisson, 2003)/vagueness ((Malamud, 2006), (Burnett, 2011)) with respect to how many parts of the subject DP need to be affected in order for the utterance to be felicitous. Prefixing all to the subject eliminates (or significantly reduces) the pragmatic slack: every single townsperson must be asleep. (11)
All the townspeople are asleep.
All as a ‘slack regulator’: “This makes [(9)] truth-conditionally equivalent to [(12)]. The difference in meaning, I suggest, is not truth-conditional, but simply in how much pragmatic slack they allow, in how much deviation from the truth they permit the pragmatic situation to license.” Lasersohn (1999) (p.523) Summary: • All is licensed in constructions in which the universal quantification seems to be coming from some other source. • The difference between an expression [φ] and [All [φ]] is a matter of pragmatic slack, not quantification. 4
November 12th, 2011
4
42nd Meeting of the North East Linguistics Society
Adjectival ALL: The Data • A new argument in favour of the SR analysis: the distribution and interpretation of ALL in other domains.
4.1
The Basic Observation
In the adjectival domain, ALL can have two different interpretations. 4.1.1
Universal/Completive Interpretation
• Sometimes, ALL seems to be synonymous with completely. (12)
a. b. c.
(13)
a. b. c.
French: La salle est toute vide. The room is ALL empty Italian: La stanza `e tutta vuota. The room is ALL empty German: Das Zimmer ist ganz leer. The room is ALL empty ‘The room is completely empty’ (lit. all empty) French: Jean est tout chauve. John is is ALL bald Italian: Gianni `e tutto pelato. John is ALL bald (Austrian) German: Hans ist ganz glazert. John is ALL bald ‘John is completely bald’ (lit. all bald)
Call the ‘completely’ interpretation, the completive/universal interpretation. • At first glance, it looks promising to analyze ALL in these examples as some kind of universal quantifier. • The sentences in (13) and (14) have universal truth conditions: tout chauve requires its subject to have the highest degree of baldness (cf. Kamp and Rossdeutscher (1994), Yoon (1996) and below). – The completive interpretation serves as an argument in favour of developing a unified analysis of adjectival ALL (tout vide) and nominal ALL (toutes les filles). • Junker (1995) (for French): tout is a universal (event) quantifier. Note: English does has adjectival ALL with a completive interpretation, but productive adjectival all seems to be limited to only certain dialects. (14)
Bolinger (1972) (p.47) a. The can is all empty. 5
November 12th, 2011
b. c. d. 4.1.2
42nd Meeting of the North East Linguistics Society
Some of the buds are all shut, but others are all open. Is he all well again? Everything all set?
Intensive Interpretation
• Other times, ALL seems synonymous with very or really. (15)
content/contento/gl¨ ucklich ‘happy’ a. French: Jean est tout content. John is is ALL happy b. Italian: Gianni `e tutto contento. John is ALL happy c. German: Hans ist ganz gl¨ ucklich. John is ALL happy ‘John is really happy’ (lit. all happy)
(16)
triste ‘sad’ a. French: Jean est tout triste. John is is ALL sad b. Italian: Gianni `e tutto triste. John is ALL sad ‘John is really sad’ (lit. all sad)
(17)
petit/klein ‘small’ a. French: Maria est toute petite. Maria is ALL small ‘Maria is really small.’ b. German: Maria ist ja noch ganz klein. Maria is still PRT ALL small ‘Maria is still really small.’
(18)
French: Synonymy with tr`es ‘very’: a. Jean est tout content ≈ Jean est tr`es content.
Again, English has this use (possibly in a more restricted version3 ). (19)
Bolinger (1972) (p.47)
3
English all also seems to have a more productive exclamative adjectival all, which frequently requires the presence of and stuff/shit. This use of all differs from the uses discussed in this talk in its interpretation and its distribution. For example, exclamative all is possible with non-scalar adjectives. (i)
And then his ex-girlfriend comes in, all pregnant and stuff !
(ii)
You’re a disrespectful fucktard that’s causing a ruckus when a fine upstanding citizen is getting all commemorated! –Jason Stackhouse, ”True Blood” Season 3, Episode 8.
I leave establishing the relationship between exclamative all and slack regulator all to future research.
6
November 12th, 2011
a. b. c. d.
42nd Meeting of the North East Linguistics Society
He was all foolish in his embarrassment. This is all nasty and old; throw it away. I’m just all heartsick with the news. It’s all neat and tight.
Observation: • The sentences in (16) and (17) have existential, not universal, truth conditions. • tout content requires its subject to have some degree of happiness significantly higher than the contextual ‘happiness’ standard. (20)
Adj. ALL in Indo-European: a. Productive Completive/Intensive Alternation: French tout, Italian tutto, Romanian tot-, German ganz, Breton holl . . . b. Restrictive Completive/Intensive Alternation: English all, Dutch helemaal . . .
Note. The availability of intensive ALL is subject to a fair amount of speaker, register, and even constructional variation. • For most speakers of French, tout is awkward in predicative contexts with grand ‘tall/big’. – Intensive tout exists in many expressions (23). (21)
(22)
?Jean est tout grand. Jean is ALL tall ‘John is really tall’ a.
b.
un tout grand merci a ALL big thankyou ‘a very big thank you’ ouvrir tout grand open ALL big ‘to open very wide’
Open Question: Are gaps in the productivity of adjectival ALL are principled or accidental?
4.2
Summary
1. There are uses of ALL in the adjectival domain that are universal. • These cross-domain parallels suggest that we should develop a unified analysis of nominal and adjectival ALL. 2. There are uses of ALL in the adjectival domain that are intensive; i.e. not universal. Problematic Question for the ∀ approach: If ALL is a universal quantifier, why does it lose its universal force in examples like (16) and (17)? 7
November 12th, 2011
5
42nd Meeting of the North East Linguistics Society
A Homophony Analysis?
A way of saving the ∀ analysis: • A suggestion by Junker (1995): There are two homophonous ALLs: ALL1 (a universal quantifier) and ALL2 (an intensifier). – In a degree semantics framework, we could treat ALL1 as a universal degree quantifier and ALL2 is an degree modifier (`a la Cresswell (1977) and Kennedy (1997)). (23)
5.1
Possible analysis for French tout: a. Quantifier: Jtout1 K = λP λx (P (x, d) & ∀d0 (d0 6= d → d > d0 )) b. Intensifier: Jtout2 K = λP λx (∃d(P (x, d) & d is much greater than ds )).
Argument Against a Homophony Analysis
A principled generalization governs the distribution of the completive/intensive interpretation: (24)
Empirical Generalization: The completive and intensive interpretations of ALL are in complementary distribution. a. If P is a total absolute adjective, then ALL P has a completive interpretation. b. Otherwise, ALL P has an intensive interpretation4 .
Basic Scale-Structure Distinctions: • See Cruse (1986), Bierwisch (1989), Kamp and Rossdeutscher (1994), Yoon (1996), Rotstein and Winter (2004), Kennedy and McNally (2005), Kennedy (2007) and many others for tests that distinguish these classes. • Total absolute adjectives are associated with scales that have top endpoints. (25)
Total Absolute Adjectives: bald, empty, clean, smooth, dry, straight, flat, closed. . .
(26)
Completive Interpretation (French): tout chauve, tout vide, tout propre, tout lisse, tout sec, tout droit, tout plat, tout ferm´e. . .
• Relative adjectives are associated with scales that lack a top endpoint. (27)
Relative Adjectives: tall, short, expensive, cheap, nice, friendly, intelligent, stupid, narrow, wide. . .
4
Note that this generalization describes only scalar adjectives; adjectival ALL is generally ungrammatical with non-scalar adjectives.
8
November 12th, 2011
(28)
42nd Meeting of the North East Linguistics Society
Intensive Interpretation: tout grand, tout petit, tout cher, tout gentil, tout nul, tout ´etroit. . .
• Partial absolute adjectives are associated with scales with no top endpoint. (29)
Partial Absolute Adjectives: dirty, bent, wet, curved, crooked, dangerous, awake. . .
(30)
Intensive Interpretation: tout sale, tout tordu, tout mouill´e, tout courb´e, tout croche, tout r´eveill´e. . .
Argument: • If tout was ambiguous between a quantifier and an intensifier, we should expect sentences with total absolute predicates to be ambiguous between a completive and intensive interpretation. – Recall that the tout that appears with RAs and partial AAs is synonymous with the intensifier tr`es (19). (31)
Total AAs are compatible with intensification a. Jean est tr` es chauve. Jean is very bald ‘Jean is very bald.’ b. La salle est tr` es vide. The room is very empty ‘The room is very empty.’
Wrong Prediction: Sentences with total AAs and tout are not ambiguous. • Conjoining compl`etement vide with pas toute vide is a contradiction: (32)
a. #La salle est toute vide, mais elle n’est pas compl` etement vide. The room is ALL empty, but it NEG-is not completely empty Lit. ‘The room is completely empty but not ALL empty.’ (contradiction) b. La salle est tr` es vide, mais elle n’est pas compl` etement vide. The room is very empty, but it NEG-is not completely empty. ‘The room is very empty, but it’s not completely empty.’
• Sentences with tr`es strongly imply (or even entail) the negation of their counterparts with tout. (33)
La salle est tr` es vide La salle n’est pas toute vide.
Conclusion: A homophony analysis over-generates intensive interpretations with total AAs. • The observed interpretative variation should be derived from the linguistic context in which ALL appears from a single underlying lexical meaning. 9
November 12th, 2011
5.2
42nd Meeting of the North East Linguistics Society
Summary
1. Adjectival ALL has a completive/universal meaning when it appears with adjectives whose associated scales have a top endpoint (total AAs). • We would like to unify the universal use of adjectival ALL with the universal use of nominal ALL. 2. Adjectival ALL has an intensive meaning when it appears with adjectives whose associated scales have no top endpoint. • Completive ALL and intensive ALL are in complementary distribution; therefore, we would like a unified analysis of these two interpretations. (34)
ALLN = ALLcomp = ALLintens
(35)
a.
b.
c.
6
Toutes les filles ont saut´e dans le lac. ALL the girls have jumped in the lake. ‘All the girls jumped in the lake.’ La salle est toute vide. The room is ALL empty. ‘The room is completely empty.’ Jean est tout petit. Jean is ALL small ‘Jean is really small.’
A New Slack Regulation Analysis
Proposal: • What definite plurals ( ‘the girls’) and scalar adjectives ( ‘small’, ‘empty’ etc.) have in common is that these constituents are vague. • ALL applies to a constituent X and eliminates X’s borderline cases (a notion defined w.r.t. pragmatic slack/vagueness). For a constituent X, let BC(JXK) be the set of borderline cases of X. (36)
Informal Proposal: JALL XK = JXK − BC(JXK)
• The proposal is formalized within (extensions) of Cobreros et al. (2010)’s Tolerant, Classical, Strict logical framework for modelling the properties of vague language in Burnett (2012).
10
November 12th, 2011
6.1
42nd Meeting of the North East Linguistics Society
Borderline Cases
Three characterizations of vagueness: (Keefe (2000), Fara (2000), Smith (2008)) 1. ⇒ Borderline Cases: Objects for which it is difficult or impossible to tell whether or not they satisfy the predicate. 2. Fuzzy boundaries: There are (or appear to be) no sharp boundaries between a predicate and its negation. 3. Sorites Paradox: A paradox for systems based on first order logic that follows from the fuzzy boundaries property. Diagnosing Borderline Cases 1. Epistemic (Peirce, 1901): Borderline cases give rise to uncertainty w.r.t. bivalence. 2. Syntactic/Semantic (Alxatib and Pelletier (2010), Ripley (2011) (a.o.)): Borderline cases tolerate overt contradictions.
6.2
Borderline Case Elimination
For each of the four classes of constituents (RAs, Partial AAs, Total AAs, definite plurals), I show: 1. What its borderline cases are. 2. How the analysis of ALL as a ‘borderline case’ eliminator creates the desired interpretation. 6.2.1
Relative Adjectives: (small etc.)
• Standard view: they are evaluated with respect to a comparison class, and their extension is determined by a contextually given standard (ds ). • An individual is in the extension just in case they satisfy the predicate to a degree higher than ds . Context: Suppose we are evaluating the heights of children in grade three. • Anyone taller than 4ft1” (the average) is going to be clearly not small, and anyone under 3ft2” is going to be clearly small. Borderline cases: 1. Uncertainty: What about kids that measure 4ft?, 3ft11”? Are they small or not small ? Both? Neither?
11
November 12th, 2011
42nd Meeting of the North East Linguistics Society
2. Overt contradictions: Recent experiments (Alxatib and Pelletier (2010), Ripley (2011)) have shown that participants accept overt contradictions of the form in (38) with individuals meeting the uncertainty condition5 . (37)
a. b.
Mary is both small and not small. Mary is neither small nor not small.
Summary: For RAs like small, there exist individuals to whom we seem to be able to meaningfully apply both a predicate and its negation. Proposal: Applying ALL excludes these individuals, leaving only the individuals that clearly satisfy the predicate. • Prediction: Intensive interpretation. X Example: petit ‘small’ vs tout petit ‘ALL small’ (38)
If Marie is toute petite, a. True: Marie est petite. b. False: Marie n’est petite.
(39)
If Marie is 4ft: a. OK/?: Marie est petite et pas petite. b. #: Marie est toute petite et pas toute petite.
6.2.2
Partial Absolute Adjectives (dirty etc.)
• Kamp and Rossdeutscher (1994), Yoon (1996), Rotstein and Winter (2004) (a.o.): Partial AAs have existential truth conditions: • An individual is in the extension just in case they satisfy the predicate to some non-zero degree. 1. Uncertainty: Is a pair of pants with a couple of specks of dirt in it dirty or not dirty? 2. Overt contradictions: These pants are both dirty and not dirty. (They have some dirt on them, but I don’t need to wash them yet.) Prediction: Intensive interpretation X 6.2.3
Total Absolute Adjectives (bald etc.)
• Yoon (1996), Rotstein and Winter (2004), Kennedy and McNally (2005) (a.o.): Total AAs have universal truth conditions. 5
Alxatib and Pelletier (2010) studied the predicate tall, and Ripley (2011) studied the relative predicate near.
12
November 12th, 2011
42nd Meeting of the North East Linguistics Society
• An individual is in the extension just in case they satisfy the predicate to the highest degree (the endpoint of the scale). • Lasersohn (1999); Kennedy (2007): Speaking loosely, we can apply these predicates to individuals that are not at the endpoint. Context: Suppose we are evaluating the baldness of men on the street. • Men with zero hair are clearly bald, and men with full heads of hair are clearly not bald. • Borderline cases with total AAs appear as soon as we deviate from the endpoint of the scale. 1. Uncertainty: What about men with a few hairs on their head? A quarter head of hair? 2. Overt contradictions: Homer Simpson is both bald and not bald. Prediction: Completive/Universal interpretation X Example: chauve ‘bald’ vs tout chauve. (40)
a. Ok: Homer est chauve et pas chauve. b. #Homer est tout chauve et pas tout chauve.
6.2.4
Definite Plurals: (the girls etc.)
• See above: A plural distributive predicate is true of a group the girls just in case it affects ever member of the group. • See above: Speaking loosely, we can apply the distributive predicate even if some irrelevant parts of the subject are not affected. • Like with AAs, borderline cases appear as soon as we deviate from the top element of the (mereological) scale. 1. Uncertainty: Is The girls jumped in the lake true if one or two of the girls stay dry? 2. Overt Contradictions: The campers jumped in the lake, but Mary, a camper, decided to stay dry6 . Prediction: Universal interpretation X Example: les filles ‘the girls’ vs toutes les filles ‘all the girls’ 6
Negation works differently with DPs such that contradictions involving negation are very different from the same contradictions with scalar adjectives (cf. the literature on the homogeneity effect Fodor (1970); L¨ obner (2000)).
13
November 12th, 2011
(41)
7
42nd Meeting of the North East Linguistics Society
a.
Les filles sont en retard, mais Marie est a` l’heure. The girls are late, but Marie is on time ‘The girls are late, but Marie is on time.’ b. #Toutes les filles sont en retard, mais Marie est `a l’heure.
Conclusion
I presented new data on the distribution of ALL in the adjectival domain. • It is not obvious how to extend a universal quantifier analysis to the intensive use of adjectival ALL. • A slack regulator analysis in which ALL applies to a vague constituent and eliminates its borderline cases captures both the contribution of nominal ALL to the meaning of the utterance it appears in and the interpretative variation in the adjectival domain. Conclusion: The distribution and interpretation of adjectival ALL constitutes a strong argument in favour of the Slack Regulator approach, against the ‘universal quantifier’ approach.
8
Appendix 1: Slack Regulation in TCS
Cobreros et al. (2010)’s non-classical logical framework for modelling the properties of vague language. Language • The language of first order logic + a distinguished ‘slack regulator’ symbol: ALL. (42)
If P is a predicate, then ALL P is a predicate.
Semantics • Three notions of satisfaction: classical truth, tolerant truth, and its dual, strict truth. • Formulas are tolerantly/strictly satisfied based on classical truth and predicate-relative, possibly non-transitive indifference relations. – ∼P relates those individuals that are indistinguishable with respect to P . – ∼tall = {hx, yi : x looks to have about the same height as y} Definition 1. C(lassical) Model. A c-model is a tuple hD, mi where D is a non-empty domain of individuals and m is an interpretation function for the non-logical vocabulary: for a constant a1 , m(a1 ) ∈ D; for a predicate P , m(P ) ∈ P(D). 14
November 12th, 2011
42nd Meeting of the North East Linguistics Society
Definition 2. T(olerant) Model. A t-model is a tuple hD, m, ∼i, where hD, mi is a c-model and ∼ is a function that takes any predicate P to a binary relation ∼P on D. For any P , ∼P is reflexive and symmetric (but possibly not transitive). • C(lassical)-truth is defined as classical truth in either a c-model or a t-model. Definition 3. t-truth and s-truth. Let M be a t-model. Partial definition (cf. Cobreros et al. (2010) for quantifiers etc.): 1. M t P(a1 ) iff ∃a2 ∼P a1 : M c P(a2 ) 2. M s P(a1 ) iff ∀a2 ∼P a1 : M c P(a2 ) John is tall is tolerantly true ↔ John has a very similar height to someone who is classically tall. Definition 4. Negation. 1. M t ¬P(a1 ) iff M 6s P(a1 ) 2. M s ¬P(a1 ) iff M 6t P(a1 ) Definition 5. Borderline Case. a1 is a borderline case for P in a model M iff M t P (a1 ) and M t ¬P (a1 ). Proposal: ALL forces a strict interpretation of its complement. 1. M c ALL P (a1 ) iff M c P (a1 ) 2. M t ALL P(a1 ) iff M s P(a1 ) 3. M s ALL P (a1 ) iff M s ALL P (a1 )
9
Appendix 2: Comparison with Lasersohn (1999)
In our notation (where Lasersohn’s close-enough to truth = t-truth): (43)
Comparison: a. Lasersohn (1999): M t ALL P (a1 ) iff M c P(a1 ) b. This paper: M t ALL P (a1 ) iff M s P(a1 )
Strict truth is necessary for the intensive interpretation with relative adjectives and partial AAs. • Lasersohn predicts that toute petite ≈ petite.
15
November 12th, 2011
42nd Meeting of the North East Linguistics Society
References Alxatib, S. and Pelletier, J. (2010). The psychology of vagueness: borderline cases and contradictions. Mind & Language, (forthcoming). Bierwisch, M. (1989). The semantics of gradation. In Bierwisch, M. and Lang, E., editors, Dimensional Adjectives, pages 71–261. Springer, Berlin. Bolinger, D. (1972). Degree words. Mouton de Gruyter, The Hague. Brisson, C. (1998). Distributivity, Maximality, and Floating Quantifiers. PhD thesis, Rutgers University. Brisson, C. (2003). Plurals, ALL, and the nonuniformity of collective predication. Linguistics and Philosophy, 26:129–184. Burnett, H. (2011). Vagueness and plural predication. In Lassiter, D., editor, Proceedings of the 2011 ESSLLI Student Session, pages 34–45. Burnett, H. (2012). The Grammar of Tolerance: Vagueness and Quantification in Natural Language. PhD thesis, University of California, Los Angeles. Champollion, L. (2011). Parts of a whole: Distributivity as a bridge between aspect and measurement. PhD thesis, University of Pennsylvania. ´ e, P., Ripley, D., and van Rooij, R. (2010). Tolerant, classical, strict. Journal Cobreros, P., Egr´ of Philosophical Logic, (forthcoming). Cresswell, M. (1977). The semantics of degree. In Partee, B., editor, Montague Grammar, pages 261–292. Academic Press, New York. Cruse, D. (1986). Lexical Semantics. Cambridge University Press, Cambridge, UK. Dowty, D. (1987). Collective predicates, distributive predicates, and all. In Marshall, F., editor, Proceedings of the 3rd ESCOL. Ohio State University. Fara, D. (2000). Shifting sands: An interest-relative theory of vagueness. Philosophical Topics, 28. Fodor, J. (1970). The Linguistic Description of Opaque Contexts. PhD thesis, Massachusetts Institute of Technology. Junker, M. (1995). Syntaxe et s´emantique des quantificateurs flottants tous et chacun: distributivit´e en s´emantique conceptuelle. Droz, Geneva. Kalish, D., Montague, R., and Mar, G. (1980). Logic: Techniques of formal reasoning. Oxford University Press, Oxford. Kamp, H. and Rossdeutscher, A. (1994). DRS-construction and lexically driven inferences. Theoretical Linguistics, 20:165–235. 16
November 12th, 2011
42nd Meeting of the North East Linguistics Society
Keefe, R. (2000). Theories of vagueness. Cambridge University Press, Cambridge. Kennedy, C. (1997). Projecting the Adjective. PhD thesis, University of California, Santa Cruz. Kennedy, C. (2007). Vagueness and grammar: The study of relative and absolute gradable predicates. Linguistics and Philosophy, 30:1–45. Kennedy, C. and McNally, L. (2005). Scale structure and the semantic typology of gradable predicates. Language, 81:345–381. Kroch, A. (1974). The Semantics of Scope in English. PhD thesis, MIT. Lasersohn, P. (1999). Pragmatic halos. Linguistics and Philosophy, 75:522–571. Lemieux, M., St-Amour, M., and Sankoff, D. (1985). /TUT/ en fran¸cais de Montr´eal: un cas de neutralisation morphologique. In Langue et soci´et´e, pages 7–89. Government of Qu´ebec. Link, G. (1983). The logical analysis of plurals and mass nouns: A lattice-theoretic approach. In Bauerle, R., Schwartze, C., and von Stechow, A., editors, Meaning, Use and the interpretation of language, pages 302–322. Mouton de Gruyter, The Hague. L¨obner, S. (2000). Polarity in natural language: Predication, quantification and negation in particular and characterizing sentences. Linguistics and Philosophy, 23:213–308. Malamud, S. (2006). Non-maximality and distributivity: a decision theory approach. In Proceedings of Semantics and Linguistic Theory 16, Amherst, Massachusetts. GSLA Publications. Moltmann, F. (1997). Parts and Wholes in Semantics. Oxford University Press, Oxford. Peirce, C. (1901). Vague. In Baldwin, J., editor, Dictionary of Philosophy and Psychology, page 748. Macmillan, New York. Ripley, D. (2011). Contradictions at the borders. In Nouwen, R., , van Rooij, R., Sauerland, U., and Schmitz, H., editors, Vagueness in Communication, page forthcoming. Springer. Rotstein, C. and Winter, Y. (2004). Total vs partial adjectives: Scale structure and higherorder modifiers. Natural Language Semantics, 12:259–288. Smith, N. (2008). Vagueness and degrees of truth. Oxford University Press, Oxford. Winter, Y. (2001). Flexibility Principles in Boolean Semantics. MIT Press, Cambridge. Yoon, Y. (1996). Total and partial predicates and the weak and strong interpretations. Natural Language Semantics, 4:217–236. Zweig, E. (2008). Dependent Plurals and Plural Meaning. PhD thesis, New York University.
17
