Behavior-related verbs from nouns Fabienne Martin (Stuttgart) Joint work with Christopher Piñón (Lille) June 20, 2013 JeNom 5 Universitat Pompeu Fabra, Barcelona

1

Introduction

Behaviour-related verbs are intransitive unergative verbs that intuitively describe a way of behaving. These verbs can be derived from i. common nouns like lézard (Fr. lézarder) or diplomate (Fr. diplomatiser) or proper nouns (Fr. merkéliser, sarkozyser, sarkozer) that describe individuals associated with typical behavioural patterns ii. evaluative adjectives like bête (Fr. bêtifier) or pédant (Fr. pédantiser), known as adjectives of behaviour (e.g. by Marín (2010)), personality adjectives (mostly in social psychology) or propensity adjectives (Oshima (2009)). These predicates can be used in predicative sentences to describe (roughly) dispositions of individuals, cf. sentences (1). Verbs derived from these predicates describe (roughly) manifestations of these dispositions, cf. sentences (2)-(3). (1)

a. mon doudou est un vrai lézard. Il adore se prélasser sous le soleil. (Internet) My honey is a real lizard. He loves basking under the sun.’ b. A propos de sa fille, Carla Bruni déclare [...] ‘Elle est très Sarkozy. Nicolas a trouvé son maître’. (Internet) About her daughter, Carla Bruni declares [...] ‘She’s very Sarkozy. Nicolas found his master.’ c. je ne sais pas si je suis bête ou spirituel [...]. Comme tout le monde, je flotte entre tout cela. (Flaubert) I don’t know whether I’m stupid or spiritual [...]. Like everybody, I’m floating between all this. d. Pierre est un vrai diplomate! Pierre is a true diplomat! [i.e. has typical properties of a diplomat]

(2)

a. On le dit aux bains de mer, quelque part, où il lézarde et flirte. (Colette) One says he’s at a seaside resort, somewhere, where he’s basking/lazing about and is flirting.

1

b. Les spasmes financiers tenaillent l’Europe. Sarkozy sarkoze et Merkel merkélise. (Internet) Financial spasms are tormenting Europe. Sarkozy is sarkozing and Merkel is merkelizing. b’. Dominique de Villepin ‘sarkozyse’ régulièrement pendant le discours. (Internet) Dominique de Villepin regularly talks/behaves like Sarkozy during the speech. c. [Les mères de ce type] se donnent bonne conscience : ‘Mais je ne lui ai jamais parlé bébé, je n’ai jamais bêtifié, etc.’ Elles croient que c’est une question de vocabulaire. (Fr. Dolto) Mothers of this type give herselves a clean conscience: ‘But I never babytalked with him, I never dumbed things down, etc.’ They think it’s a question of vocabulary. (3)

a. Bazarova tergiverse, ruse, diplomatise pour tourner les points délicats. (Lénine) Bazarova tergiversates, goes around, diplomatizes in order to turn delicate points. b. On diplomatise, on discutaille, et les autres ils continuent d’implanter des colonies. One diplomatizes, one quibbles, while the other ones continue to set up colonies. c. ne moutonne pas ne diplomatise pas, hais - une haine d’acier, souple mais d’acier. (Céline) Don’t behave like a sheep, don’t behave like a diplomat, hate – a steel hate, flexible but of steel.

Note 1: none of these verbs is prefixed in French, cf. Martin and Mazziotta (2013), Martin and Tovena (2013). Note 2: There is sometimes hesitation between the ‘P-iser’ and the ‘P-er’ forms (cf. sarkozer/ sarkozyser above): – for some (deadjectival or denominal) behavioural verbs, both forms are/have been used (e.g. cabotiner/ cabotiniser, fainéanter/ fainéantiser, byzantiner/ byzantiniser, babouiner/ babouiniser); – native speakers do not all pick the same derivational suffix when asked to form a verb out of the adjective or the noun. Questions raised: How does the noun contribute semantically to the semantics of the verb? What are the semantic relations between lézard/ diplomate and lézarder/ diplomatiser? Merkel and merkéliser? What is the semantics of behavior-related unergative verbs? Plan: i. Section 2: être N vs N-ifier/-iser/-er: differences in use and entailment patterns ii. Section 3: Semantic analysis of denominal behavior-related verbs (1. from common nouns, 2. from proper nouns)

2 être N vs. N-iser/-ifier/-er: differences in use 2.1

Predicates under study

Examples of denominal behavior-related verbs: 2

(4)

from proper nouns

Stendhal’ d. cicéroniser ‘imitate the language/style of Cicero’

a. ronsardiser ‘write like Ronsard’ b. aristotéliser ‘develop thoughts like Aristote’

e. sarkozyser ‘behave like Sarkozy/ express political views close of those of Sarkozy’

c. stendhaliser ‘write or behave like (5)

from common nouns

behaviours’ > girouette ‘weathercock’

a. lézarder ‘stay lazily in the sun’ > lézard ‘lizard’

j. babouiner ‘to monkey around’ > babouin ‘baboon’

b. robinsonner ‘live alone, like a Robinson, wander alone’ > robinson ‘person who lives alone into nature’

k. athéiser ‘to practise/teach atheism’ > athée ‘atheist’ l. byzantiner: ‘behave according to byzantinism’ > byzantin ‘Byzantine’

c. lambiner ‘act with slowness, languidity and nonchalance and lose one’s time’ > lambin ‘slowpoke’

m. chinoiser ‘discuss in a meticulous way’ > chinois ‘Chinese’

d. bateler ‘make acrobatics, buffooneries’ >bateleur ‘acrobat, buffoon’

n. guignoler ‘behave like a Guignol’ > guignol ‘clown’ [‘Guignol’ is the name of a famous puppet from Lyon]

e. paladiner ‘behave like a paladin’ > paladin ‘wandering knight’

o. hussarder ‘behave with courage, rapidity’ > hussard ‘hussar’

f. parisienner ‘behave like a Parisian’ > Parisien ‘inhabitant of Paris’

p. putasser ‘behave like a prostitute’ > pute ‘whore’

g. rossarder ‘move like a rossard’ > rossard ‘nasty guy’

q. renarder ‘behave like a fox’ > renard ‘fox’

h. somnambuler ‘act like a sleepwalker’ > somnambule ‘sleepwalker’

r. gaminer ‘behave in a youngster way’ > gamin ‘youngster/ kid’ s. diplomatiser ’behave like a diplomat’ > diplomate ‘diplomat’

i. girouetter ‘act like a weathercock, by changing one’s opinions or 2.2 être N and N-iser/-ifier/-er: entailment patterns 2.2.1

Does the verb entail the noun?

Characterizing sentences The absence of the entailment from the behavioural verb to être N is obvious when N is a proper noun, and also has been observed when it is a common noun for English, cf. e.g. Aronoff (1980), Acquaviva (2009): (6)

He nurses well (but he’s not a nurse).

3

◦ However, in languages like French and German where common nouns can sometimes be used predicatively with and without indefinite articles, things are less trivial:1 (7)

a. Pierre est 0/ diplomateN . Pierre is a diplomat by profession. b. Pierre est un diplomate. Pierre is a diplomat i. Pierre is a diplomat by profession. ii. Pierre has the typical properties of a diplomat.

(literal) (metaph.)

◦ The entailment towards the predicative nominal sentence is blocked when the noun of profession is used bare, because the sentence is true only if the subject is N by profession, cf. e.g. van Peteghem (1993), de Swart et al. (2007), von Heusinger and Wespel (2007)), see sentences (b) below. ◦ However, the entailment arguably does go through towards the predicative nominal sentence in the indefinite version, if the sentence receives its metaphorical reading; cf. sentences (c) below: (8)

a. Pierre hussarde. Pierre behaves like a Hussar.

(generic reading)

b. → 6 Pierre est 0/ hussard. Pierre is a Hussar. c. → Pierre est un hussard. Pierre is a Hussar. (9)

(metaphorical reading)

a. Marie putasse. Marie behaves like a whore.

(generic reading)

b. 6→ Marie est 0/ pute. Marie is a whore. c. → Marie est une pute. Marie is a whore. (10)

(metaphorical reading)

a. Tu diplomatises. You behave like a diplomat.

(generic reading)

b. 6→ Tu es 0/ diplomate. You are a diplomat. c. → Tu es un diplomate. You are a diplomat.

(metaphorical reading)

◦ That the entailment goes through from sentences (a) to sentences (c) is not a surprise, given that under the metaphorical reading, the indefinite version of the predicative sentence has been argued to be true as long as the subject x ‘behaves like N’ (von Heusinger and Wespel (2007)), or has the typical properties of a N (de Swart et al. (2007)), no matter whether x exercizes the corresponding profession. ◦ This suggests that the shift in meaning from the standard to the derived interpretation of the noun in the behavior-related verb is similar to the one induced by the indefinite article in the nominal sentence. 1 On

the differences between the two versions, see also a.o. van Peteghem (1993), Roy (2006), Matushansky and Spector (2005), Beyssade and Dobrovie-Sorin (2005), Le Bruyn (2010), Mari and Martin (2008, 2009).

4

Episodic sentences Intuitively, when the behavior-related verb is used in an episodic sentence to describe an event (the most salient reading with the present perfect in French), it ascribes some way to behave to the subject qua Agent of the event, for the event time only. Borrowing a term from Barker (2002), we will say that in that case, the ascription of a property to the subject is ‘relativized’ to the event, and that the predicate receives a relative episodic reading. When the sentence with the behavior-related verb is used that way, the entailment towards the nominal sentence is blocked for the simple reason that on both bare and indefinite versions, nominal sentences as a rule cannot get a relative episodic reading:2 (11) Pierre a hussardé/a putassé. Pierre behaved like a hussar/ a whore. a. → 6 # Pierre a été un hussard [en faisant cela]. Pierre has been a Hussar (while doing so).

(transitory relative reading)

b. 6→ # Marie a été une pute [en faisant cela]. Marie has been a whore (while doing so).

(transitory relative reading)

When nominal sentences (under the bare or indefinite version) have an episodic reading, it is normally absolute: the property is ascribed non-permanently to the subject, but without being relativized: (12) Cette semaine je suis (un) hussard/diplomate! This week I’m (a) Hussar/diplomate!

(transitory absolute reading)

Note however that with some contextual help and/or with some nouns, the metaphorical reading of the indefinite version can sometimes convey what we called a relative episodic reading: (13)

a. Gino, tu as été un vrai Gentleman hier. (Internet) Lit.: Gino, you have been a real gentleman yesterday Gino, you behaved like a real gentleman yesterday.

(transitory relative reading)

b. Tu as été une vraie brute envers cette Gelée Menthe. (Internet) Lit. You have been being a real brute towards this Mint Jelly. You behave like a real brute towards this Mint Jelly. But we found no natural examples with nouns that also appear in verbs. 2.2.2

Does the noun entail the verb?

◦ Intuitions about the reverse entailment from a nominal sentence with an indefinite to the corresponding characterizing sentence with a behavioural verb are fuzzy:3 (14)

a. Jean est un diplomate. Jean is a diplomat.

(metaphorical reading)

?

b. → Jean diplomatise. Jean behaves like a diplomat.

(generic reading)

2 The

term relative is taken from Barker (2002). the bare version, the entailment from the noun to the verb does not go through for the simple fact that one can be a diplomat by profession without behaving like a diplomat (see e.g. the case of atypical diplomats). 3 For

5

(15)

a. Marie est une pute. Marie is a whore (has the typical properties of a whore).

(metaphorical reading)

?

b. → Marie putasse. Marie behaves like a whore.

(generic reading)

◦ What seems clear is that while the characterizing sentence with the behavioural verb makes a generalization about concrete behaviours of the subject, the nominal sentence can also be understood as the ascription of a pure intensional property, i.e. it can be true even if the property has never been manifested. ◦ In other words, the differences between the (a) and (b) sentences remind one of those between the generic and habitual readings of dispositional sentences (Dahl (1975), Krifka et al. (1995), Menéndez-Benito (2013) among others). (16) John eats fish. a. John would not object to eating fish.

(generic reading)

b. John regularly eats/has the habit of eating fish.

(habitual reading)

◦ The generic reading normally does not entail actual event manifestations, while the habitual reading is an inductive generalisation inferred from observed behaviours and thus presupposes actual events. ◦ Our claim is therefore that behavior-related verbs differ from manger du poisson in that they only have the habitual reading: (17) Marie putasse/diplomatise. a. # Marie would not object to behaving like a whore/a diplomat.

(generic reading)

b. Marie regularly behaves like a whore/ a diplomat.

(habitual reading)

On the other hand, the nominal sentence with indefinite can in principle have the two readings, which explains the lack of clear intuitions about the entailment: it goes through only on one reading of the predicative nominal sentence). (18) Marie est une pute.

2.2.3

(metaphorical reading)

a. Marie would not object to behaving like a whore.

(generic reading)

b. Marie regularly behaves like a whore.

(habitual reading)

Intermediate conclusions

The analysis should capture the following observations: – x diplomatise (characterizing reading) entails x est un diplomate under their metaphorical reading; – the noun is reinterpreted in a similar way in both cases; – diplomatiser but not être (un) diplomate can be used to describe a certain property of the subject during the event time only; – While x diplomatise (characterizing reading) only has the habitual reading, x est diplomate (metaphorical reading) the habitual and the generic reading. 6

3 3.1

Semantic analyses Plag (1999) and Lieber (1998, 2004)

◦ Previous analyses of -ize/-ify verbs for English (despotize, boswellize) have argued that under the relevant reading, which is often called similative, the semantics of these verbs involve an unarticulated comparative component (cf. Plag (1999), Lieber (1998, 2004)): (19) despotize, boswellize, cannibalize = do/act/make in the manner of N (similative reading) ◦ Plag attempts to derive the similative reading from the causative meaning of -ize verbs, which he represents as in (20) using Jackendovian LCSs: (20) CAUSE ([ ] j , [GO ([ property,thing ]theme,base ; [TO[ property,thing ]base/theme ])]) ◦ But as already emphasized by Lieber (2004), similatives are not easily interpretable as causatives: despotize doesn’t mean ‘cause a despot to go to something’ or ‘cause to go to a despot’. ◦ Lieber (1998, 2004) assumes that the similative reading lies outside the core meaning of ize/-ify verbs, and corresponds to a sense extension of that core, arising when the second of the two subevents in the SR of these affixes is dropped. ◦ She replaces the primitive CAUSE by Pinker’s function ACT and explicitly introduces a comparative component LIKE in the structure. This component is supposed to provide the MANNER component encoded by this verb: (21) [Event ACT ([T hing ],[Manner LIKE ([T hing , Property base N])])] ◦ Problems: – where does the comparative component LIKE come from? From the meaning of the suffix? – it is not clear that (21) captures the fact that the actions denoted have to be ‘prototypical/stereotypical’ for N; for instance, if yesterday, I accidentally drove my car like Obama did three months ago, nothing in (21) impedes that the sentence yesterday I obamized is true. However, the intuition is that my driving does not suffice to make this sentence true (whereas it does for the sentence yesterday I drove like Obama). Lieber could reply that we only habits of Obama can be taken into account in a definition of a manner (e.g. his habitual way of driving for instance). But the sentence yesterday I obamized does not seem to be made true by my mimetic driving in this context either, at least if this habit is not part of the stereotypical/prototypical properties speakers commonly attribute to Obama. – the link observed between the meaning shift of N in these verbs and in indefinite nominal sentences is not captured. – how can we extend (21) to French deadjectival behavior-related verbs? 3.2

New proposal

◦ The first strategy proposed consists in trying to use the previous analyses of the meaning shift of N in nominal predicative sentences with an indefinite NP, where ‘an N’ receives a metaphorical/figurative reading, in order to apply it to the meaning shift of N in the verbs at study. We first look at common nouns and then at proper nouns. 7

3.2.1

Common nouns

Previous analyses Let us take again the following pair: (22)

a. Pierre est un diplomate! (mais il n’est pas 0/ diplomate) (metaphorical) Pierre is a diplomat! (but he’s not a diplomat). ‘Pierre has the prototypical/stereotypical properties of diplomats without being a diplomat by profession.’ b. Pierre est diplomateN . Pierre is diplomate (by profession).

◦ Unfortunately, de Swart et al. (2007)’s analysis does not capture the metaphorical reading of (22a), although they observe the difference with the bare nominal sentences: for them, – in the bare version (b), diplomate denotes a capacity (of type e) and is then type-shifted under a set expression via their operator CAP. – in the indefinite version, the determiner triggers a coercion from capacity to kind (also of type e) and the type-shifting into a set expression via the application of Carlson’s operator REL (which originates from the determiner). ◦ Therefore (22a) is supposed to mean that Pierre is among the set of entities that realizes the kind diplomat’. But this seems to rightly capture the non-metaphorical reading of (22a) only (see von Heusinger and Wespel (2009) for the same point). ◦ Le Bruyn (2010):144 suggests that the metaphorical reading can, in fact, be seen as a kind reinterpretation of the noun (which he assumes to start as a capacity noun, following de Swart et al. (2007) on this point): “we look for inherent [i.e. prototypical] properties we associate with [diplomats] and predicate those of the subject.” ◦ However, he does not show how one could distinguish between the metaphorical and the nonmetaphorical reading of the indefinite version this way. It seems that we need another operation on the set of properties of the kind in order to separate the two readings. ◦ von Heusinger and Wespel (2007) sketch another analysis, but they don’t provide the details either. Their idea is that under the metaphorical reading, indefinite predicate nominals denote manifestations of the kind N. Therefore, sentences like (22a) assert that the subject’s referent is among the set of the manifestations of N. One can reconstruct their proposal as follows (‘R’ is the realization relation of Carlson). (23) Ja diplomatKmani f . = λ xm [R(xm , diplomat’k )] ◦ The trick is then to define manifestations of the kind diplomat’k as entities that have the prototypical or stereotypical properties of diplomats. ◦ But this definition of manifestations suspiciously sounds as an ad-hoc redescription of the ontological category of kind instantiations for the purpose of the analysis. What (23) actually does is simply to give the manifestations of the kind diplomat’k (in other words, individual (token) diplomats).

8

New proposal: basic idea Let us keep Carlson’s idea that diplomate primarily denotes a kind. Some observations we start with: – Intuitively, as many authors suggested, the relevant part of the noun meaning in the verb is the stereotypical/prototypical properties associated with this kind. – Therefore, we may assume a type-shifting operation from individuals to a set of properties of individuals. Let us suppose that this operation returns stereotypical/prototypical properties. – Another relevant observation is that in order to diplomatiser, the agent x does not have to have during the time of e all properties in the set of stereotypical/prototypical properties; one can ‘diplomatize’ simply in behaving discreetly. So probably, the next step consists in selecting one property S from the set of prototypical/stereotypical properties, arguably the contextually most salient one. Our basic idea is as follows: diplomatiser denotes the set of events e such that the subject x is the Agent of e and there is a stereotypical/prototypical property S (in the context c) of the kind diplomatk such that S holds of x at the time of e, and e is an event of behaving S. Analysis In order to define behavior-related verbs as just sketched, we first postulate a basic relation ‘behave’: hhe,he,tii, he,tii

(24) λ Sλ e [BEHAVE(e, S)]

is a relation between two-place properties S and events e, where S is a variable for relations between individuals and times: λ xλ t[S(t, x)]. The main principle for BEHAVE is the following: BEHAVE

(25) ∀e∀S [BEHAVE(e, S)→ ∃x[Agent(e,x)∧ S(τ(e), x)]] This principle says that if BEHAVE holds of an event e and a property S, then e has an Agent x and x has the property S at the time of e (=τ(e)). Now moving to diplomate, we assume that it as a kind-denoting term: (26) diplomate

diplomatk

type e

The next step is to postulate an operator in order to derive the stereotypical/prototypical properties of a kind: (27) OPst/pr λ kλ S[stereotype/prototypec (S, k)] ‘S is a stereotypical/prototypical property of kind k in context c’

he,hhe,he,tii,tii

Next, we define a principle for stereotype/prototypec : (28) ∀S∀k [stereotype/prototypec (S, k) → ∃x∃t [REL(x, k) ∧ S(t, x)]] In prose, if stereotype/prototypec holds between a kind k and a property S, then there is a realization x of k and there is a time t such that S holds of x and t. The next step is to apply OPst/pr to the kind diplomatk : (29) OPst/pr (diplomatk ) λ kλ S[stereotype/prototypec (S, k)] (diplomatk )= λ S[stereotype/prototypec (S, diplomatk )]

9

hhe,he,tii,ti

(29) denotes the set of stereotypical/prototypical properties S (in context c) of the kind diplomatk . Moving now to -iser, it introduces the BEHAVE relation and is also responsible for picking out a property from among the set of stereotypical/prototypical properties of the kind diplomatk denoted by the root. It is defined as a relation between properties Q, individuals x and events e, where Q is a variable of the same type as OPst/pr , i.e. a variable for sets of relations between individuals and times (hhe,he,tii,ti): (30) -iser

λ Qλ xλ e[agent(e, x) ∧ ∃S[Q(S) ∧ S(τ(e),x) ∧ BEHAVE(e, S)]]

We can now apply the denotation of -iser to the result of the application of OPst/pr to the kind diplomatk : (31) -iser (OPst/pr (diplomatk )) (=diplomatiser) λ Qλ xλ e[agent(e, x) ∧∃S[Q(S) ∧ S(τ(e),x) ∧ BEHAVE(e, S)]] (λ S[stereotype/prototypec (S, diplomatk )])= λ xλ e[agent(e, x) ∧∃S[stereotype/prototypec (S, diplomatk ) ∧ S(τ(e), x) ∧ BEHAVE(e, S)]] (31) is the analysis of diplomatiser. It denotes a relation between individuals x and events e such that x is the agent of e and there is a stereotypical/prototypical property S (in context c) of the kind diplomatk such that S holds of x at the time of e and e is an event of behaving S. Advantages: – diplomatiser does not entail that its subject’s referent is an instantiation of the kind diplomat’k . – It entails that x has a stereotypical/prototypical property of diplomats during the event time. – The relevant property S can be chosen contextually. – It captures the fact that N is ‘non-referentially’ interpreted: at no moment do we shift from the kind to the instantiations of this kind. The predicative sentence with indefinite nominals For the metaphorical reading of the indefinite version of predicative nominal sentences, since the idea is that the meaning shift is of the same kind, some steps will be recycled. ◦ Note that as what we observed for the verb, for this metaphorical reading to be true, the subject’s referent arguably only has to satisfy one of the prototypical/stereotypical properties of the set. So again, we single out one of the properties of the set of stereotypical/prototypical properties. We therefore begin again with diplomate: (32) diplomate

diplomatk

type e

We apply again the operator OPst/pr (27) to this kind:

10

(41) OPst/pr (diplomatk ) λ kλ S[stereotype/prototypec (S, k)] (diplomatk )= λ S[stereotype/prototypec (S, diplomatk )]

hhe,he,tii,ti

We then define an operator (introduced by un) in order to derive a set of individuals satisfying one of the properties within the set (29): (33) OP

λ Qλ xλ t[∃S[Q(S)∧S(t, x)]]

We apply this operator to the result of OPst/pr (diplomatk ): (34) λ Qλ xλ t[∃S[Q(S)∧S(t, x)] (λ S[stereotype/prototypec (S, diplomatk )])]= λ xλ t[∃S[(stereotype/prototypec (S, diplomatk )∧S(t, x)]] The metaphorical reading of (22a) can therefore be defined by applying the result in (34) to pierre: (35)

a. Pierre

pierre

b. λ xλ t[∃S[(stereotype/prototypec (S, diplomatk )∧S(t, x)]](pierre) c. λ t[∃S[(stereotype/prototypec (S, diplomatk )∧S(t, pierre)]] This gives the set of times at which Pierre has the property S (which is a stereotypical/prototypical property of diplomatk ). ◦ Advantages: we capture – the link between the interpretation of N in un diplomate and diplomatiser; – the entailment from Pierre diplomatise to Pierre est un diplomate under the metaphorical reading; – the lack of entailment from Pierre est un diplomate (generic reading) to Pierre diplomatise (because the latter requires that Pierre be an Agent of an event of behaving). ◦ Still to explain: – the absence of temporary relative interpretation for predicative nominal sentences – the absence of generic (vs. habitual) reading for the behavior-related denominal verbs 3.2.2

Proper nouns

For behavior-related verbs derived from proper nouns, we can simply recycle the analysis proposed for those derived from common nouns by assuming that proper nouns like Merkel denotes a kind with one realization only. The verb merkéliser can then be defined on the model of the analysis proposed for diplomatiser: (36) -iser (OPst/pr (merkelk )) (=merkéliser) λ Qλ xλ e[agent(e, x) ∧∃S[Q(S) ∧ S(τ(e),x) ∧ BEHAVE(e, S)]] (λ S[stereotype/prototypec (S, merkelk )])= λ xλ e[agent(e, x) ∧∃S[stereotype/prototypec (S, merkelk ) ∧ S(τ(e),x) ∧ BEHAVE(e, S)]]

11

Another (perhaps more orthodox) option is to assume that Merkel denotes an individual constant merkel. (37) Merkel

merkel

type e

The next step is to redefine the operator OPst.pr. in (27) in order to derive the stereotypical/prototypical properties of kinds or token (ordinary) individuals. So let’s assume that the variable α ranges over kinds or individuals:4 (38) OP’st/pr λ xα λ S[stereotype/prototypec (S, xα )] he,hhe,he,tii,tii ‘S is a stereotypical/prototypical property of the individual xα in context c’ Next, we can define a new version of the principle for stereotype/prototype’c so that it also applies to kinds or token individuals: (39) ∀S∀xα [stereotype/prototype’c (S, xα ) → ∃x∃t [REL(x, xα ) ∧ S(t, x)]] In prose, if stereotype/prototype’c holds between an individual xα and a property S, then there is a realization x of xα and there is a time t such that S holds of x and t. PS: We are therefore committed to the idea that not only kinds, but also token individuals can have instantiations. But this has already been proposed independently by e.g. von Heusinger and Wespel (2007); von Heusinger and Wespel (2009), in order to handle sentences like (40a), where proper names are preceded by the indefinite article. Roughly, their idea is that in this use, the article introduces a variable over manifestations of (token) individuals, which are instantiations of this individual, cf. (40c): (40)

a. We need another Merkel. b. J Merkel K = merkel

c. J a (MOD+PN)Kmani f . =λ xm [REL(xm , m) ∧ J a (MOD+PN)K(xm )

The next step is to apply OP’st/pr to the individual merkel : (41) OP’st/pr (merkel) λ xα λ S[stereotype/prototypec (S, xα )] (merkel)= λ S[stereotype/prototype’c (S, merkel)]

hhe,he,tii,ti

(41) denotes the set of stereotypical/prototypical properties S (in context c) of the individual merkel. We can keep the same meaning for -iser: (30) -iser

λ Qλ xλ e[agent(e, x) ∧ ∃S[Q(S) ∧ S(τ(e),x) ∧ BEHAVE(e, S)]]

We apply the denotation of -iser to the result of the application of OP’st/pr to the token merkel: (42) -iser (OP’st/pr (merkel)) λ Qλ xλ e[agent(e, x) ∧∃S[Q(S) ∧ S(τ(e),x) ∧ BEHAVE(e, S)]] (λ S[stereotype/prototype’c (S, merkel)])= λ xλ e[agent(e, x) ∧∃S[stereotype/prototype’c (S, merkel) ∧ S(τ(e),x) ∧ BEHAVE(e, S)]] (42) denotes a relation between individuals x and events e such that x is the agent of e and there is a stereotypical/prototypical property S (in context c) of the individual merkelk such that S holds of x at the time of e and e is an event of behaving S. 4 Notation

adopted by Gehrke and McNally (2012).

12

Comparison with the derived reading of PNs in other constructions A recent study investigated a construction which seems to trigger a similar meaning shift of the proper noun, cf. Bylinina (2011). Her English examples all have a French correspondent: (43)

a. Matching shirt and hat is so McDonalds. (≈ cheap, unfashionable) b. Buying DVDs is so 2004! (≈ out of date) c. Yeah, this is so Obama! (≈ cool)

There are two interesting common points between the processes she assumes to be involved in the meaning derivation of N in this structure and what we assumed for the verb. 1. In order to arrive at the derived meaning, she assumes two type-shifts, first from individuals to the set of their properties, second to a single property. (44)

a. e → he,ti

(Harvey → polite, smart, 1.80m tall, blond, 35yo, etc.)

b. he,ti → et

(polite, smart, 1.80m tall, blond, 35yo, etc. → smart)

c. J Harvey K = λ xλ d. smart(x)>d 2. Trying to account for how the final property is singled out, she proposes that the property gets picked because ‘the individual referred to possesses this property to a significantly high extent, and probably to the highest extent compared to all the other (conversationally active) individuals, so that the individual in question serves as some kind of maximum relative to all other available candidates.’ (p. 9). She therefore does not mention stereotypicality. But: – the property picked up seems to be ‘stereotypical/prototypical’; – it is interesting to note that some studies devoted to stereotypicality suggest that stereotypes emphasise differences between individuals. → there might be a link between her proposal and ours that the property singled out has to be picked among the stereotypical properties of the individual. In sum, there are some similarities between the two meaning shifts, although the details have to be spelled-out.

References Acquaviva, P. (2009). Roots, categories, and nominal concepts. Lingue e linguaggio, VIII(1):25–51. Aronoff, M. (1980). Contextuals. Language, pages 744–758. Barker, C. (2002). The Dynamics of Vagueness. Linguistics and Philosophy, 25:1–36. Beyssade, C. and Dobrovie-Sorin, C. (2005). A syntax-based analysis of predication. In Georgala, E. and Howell, J., editors, Proceedings of Salt XV, pages 44–61. Cornell University, Ithaca. Bylinina, E. (2011). This Is So NP! In Partee, B., Glanzberg, M., and Skilters, J., editors, Formal Semantics and Pragmatics: Discourse, Context and Models, volume 6 of The Baltic International Yearbook of Cognition, Logic and Communication, pages 1–29. New Prairie Press, New-York. Dahl, Ö. (1975). On generics. In Keenan, E., editor, Formal Semantics of Natural Language, pages 99–112. Cambridge University Press, London & New-York. 13

de Swart, H., Winter, Y., and Zwarts, J. (2007). Bare nominals and reference to capacities. Natural Language and Linguistic Theory, 25:195–222. Gehrke, B. and McNally, L. (2012). Frequency adjectives as distributional modifiers. Ms, Universitat Pompeu Fabra. Krifka, M., Pelletier, F., Carlson, G., ter Meulen, A., Chierchia, G., and Link, G. (1995). Introduction. In Carlson, G. and Pelletier, G., editors, The Generic Book, pages 1–24. Chicago University Press, Chicago. Le Bruyn, B. (2010). Indefinite articles and beyond, volume 236 of LOT Dissertation Series. LOT. Lieber, R. (1998). The suffix-ize in english: Implications for morphology. In Morphology and its relation to phonology. CSLI Publications, Stanford. Lieber, R. (2004). Morphology and Lexical Semantics. Cambridge University Press, Cambridge. Mari, A. and Martin, F. (2008). Bare and Indefinites NPs in Predicative Position in French. In Schäfer, F., editor, SinSpec, Working Papers of the SFB 732, pages 119–144. Universität Stuttgart, Stuttgart. Mari, A. and Martin, F. (2009). To be (an) idiot in french. on the systematic polysemy of copular sentences. Abstract for the Workshop Indefiniteness, DGFS 2010. Martin, F. and Mazziotta, N. (2013). Transitivizing Affixes in French. Manuscript, University of Stuttgart. Martin, F. and Tovena, L. (2013). How deadjectival verbs entail their adjectival basis. Manuscript in preparation, Universität Stuttgart and Université Paris Diderot. Marín, R. (2010). Spanish adjectives within bounds. In Cabredo, P. and Matushansky, O., editors, Adjectives: Formal analyses in syntax and semantics, pages 307–331. John Benjamins, Amsterdam. Matushansky, O. and Spector, B. (2005). Tinker, tailor, soldier, spy. In Maier, E., Bary, C., and Huitink, J., editors, Proceedings of Sinn und Bedeutung 9, pages 241–255. University of Nijmegen, The Netherlands. Menéndez-Benito, P. (2013). On dispositional sentences. In Mari, A., Beyssade, C., and del Prete, F., editors, Genericity. Oxford University Press, Oxford. Oshima, D. (2009). Between being wise and acting wise: A hidden conditional in some construction with propensity adjectives. Journal of Linguistics, 45:363–393. Plag, I. (1999). Morphological Productivity: Structural Constraints in English Derivation. de Gruyter, Berlin. Roy, I. (2006). Non-verbal Predications. A Syntactic Analysis of Predicational Copular Sentence. PhD thesis, University of South California. van Peteghem, M. (1993). La détermination de l’attribut nominal. Etude comparative de quatre langues romanes. Paleis der Akademie, Brussels. von Heusinger, K. and Wespel, J. (2007). Indefinite proper names and quantification over manifestations. In Puig-Waldmüller, E., editor, Proceedings of Sinn und Bedeutung 11, pages 332–345. Universitat Pompeu Fabra, Barcelona. von Heusinger, K. and Wespel, J. (2009). Indefinite Eigennamen in generischen Satzen: Quantifikation uber Manifestationen von Individuen. Linguistische Berichte, 2009(217):3–35.

14