filcat.uab.catfilcat.uab.cat/clt/activitats/Barcelona3.docx · Web viewAlso, gender and number ageement is typical of other distinct word classes, like quantifiers and determiners

Día 3Antonio Fábregas [email protected] - Universidad de Tromsø

Los adjetivos no existen

(1) El problema de la categorizaciónTo be of use for linguistic theories and crosslinguistic comparison, linguistic categories need to be rigorously defined with testable properties in such a way that one can determine unambiguously for any expression in any language whether it belongs to a particular category or not. The extension of a particular category should include all and only those expressions that should be compared across languages. Since the intension and the extension of a category are inversely related [...] categories that do not have enough defining properties [...] might have too heterogeneous an extension. [Tonhauser 2008: 335]

3.1. Qué (demonios) es un adjetivo

(2) ¿Podemos decir que el adjetivo está bien definido? No, incluso si nos limitamos a una sola lengua.

While one can, of course, contend that the positive characterisation of nouns and verbs is more complex than what these works suggest, it seems to us fair to say that, if there is a positive property that defines adjectives, we have not found it yet. This claim might not be controversial, but we would like to use a few paragraphs to show why we believe that it is right. In traditional grammars, there are a few properties that are used to define adjectives.

a) Syntactic properties: adjectives are used to modify nouns and can be the center of comparative constructions

b) Semantic properties: adjectives allow degree and are used to denote properties

c) Morphological properties: adjectives agree with nouns and combine with degree morphemes

Our immediate goal is to show that these properties do not define adjectives as a natural class (that is, that for any of these properties, it is not the case that all adjectives have it, and no word belonging to a different class lacks it), not even in one single language. Consider Spanish. It is true that virtually all words that have been classified as adjectives in this language can be used to modify nouns,1 but of course other nouns in appositive constructions (3a) and prepositional phrases (3b) can also modify nouns, sometimes without any apparent meaning difference (compare 3b and 3c).

(3) a. una oferta estrella an offer star ‘the best offer’b. un objeto de metal an object of metalc. un objeto metálico an object metalic

1 This is, by the way, not true of English. Bolinger (1967) noted that an adjective like asleep can only appear in predicative position: *an asleep boy; The boy is asleep.

1

mailto:[email protected]


With respect to comparative structures, in the same way that not all adjectives allow degree modification (cf. infra), not all adjectives can be the lexical core of a comparative structure (4a); it is obvious that mass nouns (4b) and other lexical classes can easily be the core of a comparative which on the surface uses the same set of elements as a gradable adjective (4c).

(4) a. *un objeto más metálico que otro an object more metalic than anotherb. tiene más amor que otro he.has more love than anotherc. es más guapo que otro he.is more handsome than another

In fact, not all adjectives allow degree modification (saliently, relational adjectives and non-subsective adverbial adjectives; cf. Bosque 1993, 2006; Fábregas 2007; Partee 2010) (5a); other lexical classes, like stative verbs (5b) allow combination with adverbs expressing degree.

(5) a. *muy {carbónico / textil / presunto} very carbonic / textile / allegedb. gustar mucho like a.lot ‘to love to a high degree’

With respect to their semantic denotation, we would like to remind the reader that in a model semantic analysis adjectives have the same type as nouns and monotransitive verbs (<e,t>) or PP modifiers (<<e,t>, <e,t>>) (Heim & Kratzer 1998). If the difference is not based on semantic type, then, it could be sortal. However, claiming that they denote properties is not of much help in the absence of a positive definition of property that differentiates it from states or other descriptive terms; but even if we take a common-sense, intuitive definition of property (as opposed, say, to ‘individual’ or ‘event’), it is not true either that adjectives invariably denote properties. Relational adjectives can, in fact, introduce participants of a situation (6a); non-subsective adjectives (6b) and privative adjectives (6c) instead of adding further properties to the denotation of the noun act as operators that shift the model world or cancel the entailments of the noun’s denotation.

(6) a. la invasión alemana de Italia the invasion German of Italy ‘the German invasion of Italy’ b. Esto es un posible problema this is a possible problemNote: It does not entail that this is a problem.c. Esto es una falsa pistola. this is a fake gunNote: It does entail that this is not a gun.

If we move to morphology, an area where considerable cross-linguistic variation is attested, not even in a highly inflectional language like Spanish can one say that all

2



adjectives, and nothing else, agree for gender and number with the noun (7a). Some adjectives agree only in number (7b), while others lack agreement in both dimensions (7c). Also, gender and number ageement is typical of other distinct word classes, like quantifiers and determiners (7d).

(7) a. pantalonesM.PL roj-o-s trousers red-M-PLb. pantalonesM.PL naranja-s trounsers orange-PLc. revistasF.PL porno magazines porn d. l-a-s much-a-s preocupaciones-F-PL the-F-PL many-F-PL worries

Finally, let us concentrate on the combination with degree morphology. Obviously, not all adjectives combine with it (8a), as the classes of relational and non-subsective adjectives mentioned before simply reject any degree modification. But adverbs in Spanish can also carry degree morphology (8b). One could claim that these adverbs are underlyingly adjectives, but note that appreciative morphology is in some sense similar to degree morphology (compare 8b with 8c) and nouns can, like adjectives, combine with it (8d). Finally, one way of interpreting Grandi (2008) is that verbs can combine with morphemes (traditionally classified as interfixes) that manipulate the degree or intensity with which the action takes place (9).

(8) a. *carboniqu-ísimo, *presunt-ísimo carbonic-SUPL alleged-SUPLb. lej-ísimo far-SUPL ‘very far away’c. lej-it-os far-DIM-WM ‘quite far away’d. actor-cillo actor-DIM ‘someone who works as an actor, but not very well’

(9) a. comer eatb. com-isc-ar eat-INFIX ‘to eat irregularly, in small portions’

Not even in one single language seems it possible to identify one property that defines adjectives as a natural class. There are two natural reactions: the first is to talk about ‘unmarked’ (Croft 1991) or ‘primary’ (Bhatt 1994) uses of adjectives –presumably, quality denoting gradable adjectives used as modifiers–, allowing grammar to alter this category in substantial ways that account for the wide range of non-prototypical cases. The second is simply to conclude that there are no positive properties of adjectives. Only the second proposal makes sense of the fact that all ‘primary’ properties of adjectives are attested in some members of other lexical classes.

3



Adjectives are defined as ‘something that cannot do what a noun or a verb do’

a) Adjectives are unable to introduce their subject (Baker 2002)

(10) vP (verb)

DP v

v VP

(11) PredP (adjective)

DP Pred

Pred AP

b) Adjectives are parasitic categories that cannot be defined inside one single projection (Hale & Keyser 2002)

(12) a. X (verb) b. h* (adjective)

X Y Y h*

h* X

c) Adjectives are not primitive categories: they are literally X + P (Mateu 2002)

d) Adjectives are unable to agree in person (Baker 2008)e) Adjectives are unable to define denotation classes before they combine

with degree (Kamp 1975)

(13) Consecuencia: el adjetivo no es una categoría universal y tampoco parece ser una categoría independiente

3.2. Los adjetivos como proyecciones de Place o Path: efectos comparados

Adjectives are built, perhaps drawing from features from a universal inventory (specifically, here we will argue that in Spanish adjectives are built by using the notion of PathP, while in English one uses the notion of PlaceP)

Consequence: in languages that have adjectives, nothing guarantees that adjectives will be projected exactly in the same way

Claim:

i. Semantically: both English and Spanish, perhaps all languages that differentiate adjectives as a grammatical category, associate scales to the denotation of adjectives

ii. Syntactically: only some languages project scales syntactically.

4



a. English does notb. Spanish does

(11) ScaleP (= PathP)

Scale AP (=PlaceP)

iii. This throws light on three previously unrelated properties that differentiate Spanish from English with respect to the morphosyntax of adjectivesa. One fact about their syntactic distributionb. One fact about their morphological rolec. One fact about the arguments they can introduce

Property I: resultative phrases in Spanish vs. English

Generalisation: If we have strong result state phrases (Washio 1997)...

PP APEnglish ✔ ✔Spanish ✔ *

(12) a. John broke the vase in one thousand pieces.b. Juan rompió el jarrón en mil pedazos. Juan broke the vase in one.thousand pieces

(13) a. John shot Mary dead.b. #Juan disparó a María muerta. Juan shot A María dead✔ ‘Juan shot Mary when she was already dead’ * ‘Juan shot Mary and as a result she died’

Note a crucial property: the prepositions used in (xx) above are locative (place) prepositions that only get interpreted as a result state by virtue of the syntactic context where they are introduced, as the following contrast shows:

(14) a. entrar en la casa [result location] enter in the house‘enter the house, move inwards until you are in the house’b. correr en la casa [process location] run in the house‘run while you are inside the house’

(Standard) analysis: two layers (at least) in P (Den Dikken 2006, Svenonius 2010, Pantcheva 2011, among many others):

(15) Place (16) Path

PlaceP PathP

Place ...DP Path PlaceP

5



Place ...DP

A path is a series of ordered points, while a place is a region which constitutes a point inside the set of points of a path.

(17) towards the bridge‘following a path that culminates at a point close enough to the bridge’

The result location reading of a place P depends on the structure introduced by the verb:

(18) ProcP

DP Proc Juan

Proc ResP

DP Res

Res PlaceP = Rheme of resultentrar

Place DPen la habitación

(19) ProcP

ProcP PlacePen la habitación

DP ProcJuan

Proc ...correr

Given this, we can diagnose whether ResP can take as complement Paths. It turns out it cannot.

(20) a. correr hacia la casa [No ResP --> PathP fine] run towards the houseb. *entrar hacia la casa [ResP --> *PathP] enter towards the house

(21) ProcP marries PathP; ResP marries PlaceP

a. ProcP b. ResP

Proc PathP Res PlacePrun towards the bridge break in 1000 pieces

6



Consider now this: scales are essentially paths: series of ordered values across the same dimension, with a defined directionality. So if we start from our hypothesis:

(22) a. Spanish APs b. English APs

ScaleP = PathP

Scale AP = PlaceP AP = PlaceP

Predictions:

a) English APs, lacking ScaleP, will be selectable by ResPb) Spanish APs, projecting ScaleP, will not be selectable by ResPc) Spanish APs, projecting ScaleP, can be selected by ProcP [hold until §4]

(23) *Juan disparó a María muerta.

InitP

Juan Init

Init ProcP

María Proc

Proc ResP

María Res

Res *ScaleP = PathP

Scale APmuerta

What Spanish can do is to project adjectives as spurious resultatives (Washio 1997), that in fact modify the manner in which the process takes place (and only by implication can be interpreted as affecting the object that receives the property entailed by the manner):

(24) a. Juan cortó [la cebolla] [fina] en rodajas. Juan cut the onion thin in slices‘Juan cut the onion thinly in slices’b. ⇏ La cebolla se quedó fina.

the onion became thin

But this is predicted, because manners are associated to ProcP and ProcP admits PathP as a modifier [the way in which adjuncts are introduced is orthogonal to our purposes]:

(25) ProcP

7



ProcP ScalePthin

the onion Proc

Proc ResP

the onion Res

Res PlacePcut in slices

Change of state with adjectives in English and Spanish

Generalisation: If we have a change of state verb derived from a morphologically irregular adjective...

Positive degree Suppletive comparativeEnglish * ✔Spanish ✔ ✔

(Bobaljik 2012: chapter 6; see, for degree achievements Dowty 1979, Hey, Kennedy & Levin 1999, Rothstein 2004, Winter 2006, Kearns 2007, Kennedy & Levin 2008)

(26) a. to worsen / *to baddenb. to belittle (not change of state)c. to widen (not suppletive comparative)d. *to small-en, *to good-(d)en

(27) Spanish also allows comparative degree derived verbsa. mejor-ar

better-V‘improve’

b. em-peor-arim-worse-V‘worsen’

c. a-minor-arto-smaller-V‘to reduce’

But there are change of state, even gradual change of state, deadjectival verbs from the positive degree:

(28) a-buen-arto-good-V‘to improve, to become better’

en el entretiempo me hice el gil y me puse a ver el entrenamiento un rato hasta que un poli me dijo que no se podia, despues se abuenó un poco

8



One could argue that here the base is a noun bueno ‘good person’, but it won’t work:

(29) *un bueno a goodIntended: ‘a good person’

One could also argue that here we have an analytic comparative construction más bueno, which in colloquial Spanish is used when ‘goodness’ is moral ‘goodness’. But this won’t work either: abuenarse can mean ‘become healthier’, from bueno ‘healthy’:

(30) a. Chepita que estaba maluca ya se está abuenando [Perú] Chepita that was sickly already SE is to-good-en-ing‘Chepita, who was a bit sick, is getting healthier already’b. #El paciente se puso más bueno.

the patient SE became more good*‘The patient became healthier’

(31) a-mal-ara-bad-V‘to make / become sick(er)’

(32) a-grand-ara-big-V‘to make / become big(ger)’

This contrast also follows from the same proposal: APs in Spanish project in syntax a ScaleP, which is a Path, but English APs lack ScaleP as a projection and are, for all intents and purposes, PlacePs of sorts. Therefore, if we want to define a process that defines a change across the scale of the adjective, we are faced with a problem in English: the AP does not syntactically project the path that ProcP would use to identify isomorphically with the change of state:

(33) #ProcP

Mary Proc

Proc AP(-en) good

*‘Mary goodened’

In Spanish, ScaleP projects, and that is why the positive degree is good enough:

(34) Proc

Chepita Proc

Proc ScaleP

9



a...a buen-

‘Chepita (se) abuenó’

The solution is to introduce the adjective dominated by the head Comparative, which does introduce a scalar object which (incrementally) can be used to define the change.

(35) Proc

Mary Proc

Proc ComparativeP-en worse

‘Mary worsened’

Nothing prevents Spanish from using the same structure:

(36) Proc

María Proc

Proc ComparativePem...ar peor

María empeoróMaría worsened

Property 3: syntactically-represented standards of comparison

Generalisation: Absolute adjectives are not compatible with overt standard of comparison PPs in Spanish; in English, they are allowed if they are accommodated in context

English SpanishAbsolute adjectives + PP ✔ *Relative adjectives + PP ✔ ✔

Rothstein & Winter (2004); McNally (2011); Sassoon & Toledo (2011)

(37) a. ?The gas tank is empty, but there are still a few drops left.b. The gas tank is full, but you can still top it off.

(38) a. This child is tall for a three year old. [Relative adjective]b. This woman is pretty for a witch.

(39) a. This glass is full for a wine glass. [Absolute adjective]b. This room is empty for a graduation-ceremony room.

Consider, in contrast, Spanish:

(40) a. Juan es alto para un chico de tres años.

10



Juan is tall for a boy of three yearsb. Morgana es guapa para una bruja.

Morgana is pretty for a witch(41) a. *Esta copa está llena para una copa de vino tinto.

this glass is full for a glass of wine redb. *Pedro estaba borracho para un astronauta.

Pedro was drunk for an astronaut

Proposal: if Spanish syntactically projects ScaleP, then it also projects syntactically the distinction between scales of different kinds. Thus, at a syntactic level, absolute and relative adjectives have a different syntax. In English, scales are semantic objects only: as syntactic objects, all adjectives are the same, even if one semantic class needs accommodation and the other one does not.

Option 1

(42) Two flavours of ScaleP

a. ScaleP b. ScaleP

Scale AP PP Scale

Scale AP absolute relative

Option 2

A syntactic decomposition of types of scale

(43) MaxP = final boundary

Max MinP = initial boundary

Min ScaleP = body

Scale AP

Generalisation of bodies and boundaries to scales, adapting Piñón (1997):

(44) Min :=vSP[Boundary(v) & Scale(v’) & Left-Boundary(v, v’) & P(v’) & ¬∃v’ [v’ > v’’ & P(v’’ ⊕ v’)]

(45) Max:=vSP[Boundary(v) & Body(S) & Right-Boundary(v, S) & P(S) & ¬∃v’ [v’ > v & P(v ⊕ v’)]

(44) reads as follows: a minimal value is a value (v) that acts as the left boundary of the scale (v’) and that it is such that there is no value v’’ lower than v’ that also belongs to the scale described as P. There are three substantive changes with respect to the previous formula: first, we have moved from the domain of happenings (e) to the domain of values (v), so the notions are translated to a different set of entities.

11



Second, instead of using the term ‘Begin’, which presupposes a temporal ordering, we are using the term ‘Min’, which presupposes a set of non-temporally ordered magnitudes inside the same domain. Third, as a consequence of not being anymore in the domain of happenings with temporal traces, the precedence relation << has been substituted by a mere ‘higher than’ relation ‘>’.

(46) Comparison class projects above ScaleP

CPP

PP CP

CP ScaleP

Scale AP

Problem: if a boundary is projected above, CPP intervenes between Scale and Boundary, that is, between the body and its boundary. There are many ways to implement this: maybe A0 has features [scale, boundary] that need to be checked through head movement and CP prevents this movement; maybe the adjective (à la Nanosyntax) spells out the constituent Min / MaxP, containing ScaleP, and projecting CPP breaks that constituent. Either way:

(47) * Min/MaxP

Min/Max CPP

PP CP

CP ScaleP

Scale AP

At the semantic level, perhaps there are no differences between English and Spanish when it comes to the properties of adjectives, but...

at the syntactic level, Spanish projects scalar structure English does not project scalar structure in the syntax Consequently, Spanish adjectives are PathP (thus, unselectable by ResP) Consequently, Spanish adjectives can be complements of ProcP Consequently, PPs are sentitive to the type of scale in Spanish

Potential extensions:

Spanish has ser and estar, two copulative verbs that are sensitive to the grammatical category of the attribute, and to aspectual and other properties of the adjective.

12



Maybe this is related to the proposal that adjectives (perhaps other categories) can project extra functional structure in the syntax

Spanish adjectives agree, English adjectives do not

Variation. If our analysis is on the right track, we are claiming that one source of variation is how many syntactic projections are drawn from the universal inventory in each language. Similarly to Amritavali & Jayaseelan (2005) in the domain of situations:

TP (Spanish) MoodP (Malayalam)

T MoodP Mood AspP

Mood AspP

3.3. ¿Qué es un adjetivo? Análisis interno de las clases de adjetivos sin SAWhen adjectives are morphologically derived from (eventive) verbs, the most productive readings are non-episodic –that is, readings which do not entail actual participation in a specific event at a particular occasion–. The following deverbal adjectives illustrate the basic range of non-episodic readings:

(1) a. a washable skirtb. a forgetful scholarc. a slippery road

(2) a. hui-dizo flee-DIZO‘elusive’ (‘that frequently flees’, habitual) b. quebra-dizo break-DIZO‘fragile’ (‘that can get broken easily’, dispostional)

3.3.1. DispositionalityConsider the following deverbal adjectives in Spanish:

(3) With -diz(o)a. un terreno move-dizo a ground move-DIZO‘an unsteady ground’b. un suelo resbala-dizo a floor slip-DIZO‘a slippery floor’c. un objeto quebra-dizo an object break-DIZO‘a fragile object’d. un techo llovedizoa roof rain-DIZO ‘a roof that lets water pass through it’

13



(4) With -ntea. una sustancia relaja-nte a substance relax-NTE‘a relaxing substance’b. un objeto corta-nte an object cut-NTE‘a cutting object’c. un pan crujie-nte a bread crunch-NTE‘crunchy bread’

The relevant semantic notion involved in these cases is dispositionality, defined as (5) (cf. Manley 2012: 321).

(5) Necessarily x is disposed to P in w if x would P in w

That is: dispositionality entails that, given the right facilitating conditions in the world (w) of discourse, an entity (x) would act as a particupant in situation P (described by the predicate). In an example like (4b) –the cutting object–, saying (4b) is saying that, necessarily, the knife would cut if the right conditions are found –that some other object gets in contact with its blade and some pressure is exerted, etc.–. If the object would not cut given a context, then (4b) would not correctly describe that object. Necessarily, dispositionality implies an assessment on the part of the speaker of the expected behaviour of an entity, and this assessment makes a judgement on the basis of its particular properties or its particular state.

3.3.2. HabitualityConsider the examples in (6), all derived by the suffix -ón.

(6) a. un chico abus-ón a boy abuse-ÓN‘a boy that is a bully’b. un hombre mir-ón watch-ÓN‘a man that is a voyeur’c. una chica destroz-ón-a a girl destroy-ÓN-f‘a girl that frequently destroys things’d. una mujer dorm-il-ón-a a woman sleep-inf-ÓN-f‘a woman that is fond of sleeping’e. un hombre com-il-ón a man eat-inf-ÓN‘a man that is a glutton’

It is clear by the meaning of these deverbal adjectives that a dispositional paraphrase will not be enough: take for instance (6c) ‘a girl that often destroysthings’. This nominal expression is a good description if and only if in the past that girl has destroyed things –eg., has ruined a number of electronic machines by using them in the wrong way–. If we are in a situation where she has never broken a machine but we

14



just expect that, by virtue of her properties (eg., her reduced knowledge of electronics), given the right circumstances (eg., if we let her manipulate the new tablet unsupervised), she would ruin it, the girl would protest the description as unfair. Similarly, if a boy has never bullied someone but has the typical properties that, in principle, could make him prone to being a bully, describing him as abusón ‘bully’ would simply not be the correct description. These adjectives, in fact, entail a habitual participation in the event expressed by the verbal base: we cannot say that entity X is abusón unless in our current world and time participates frequently in specific events of abusing someone. The semantic notion involved here is habituality, which we can define as follows (7).

(7) Regularly, x participates in P.

It is important to differentiate habituality from iterativity, which is an instance of an episodic reading. With Carlson (2011) we assume that the truth conditions for a habitual depend on the assertion of the regularity of occurrence of the event, and not on any specific event taking place at a particular time and place. If we say John smokes, we do not imply that in a particular point in time John will be smoking a cigarette, that is, this cannot be denied by the fact that during a trip John did not smoke a single cigarette; we state that his characteristics are such that he is a smoker (habitual). Note that this contrasts sharply with John smoked five cigarettes yesterday, which is iterative: we locate a series of events in a specific time, world pair. In this sense, habituals are non-episodic. Iteratives are episodic in nature, and even though they also involve repetition –in an intuitive, non technical sense–, they behave as episodic sentences. The suffix -dizo (cf. 3) can also produce habitual readings.

(8) a. un niño olvida-dizo a child forget-DIZO‘a forgetful child’b. un cura enamora-dizo a priest fall.in.love-DIZO‘a priest that falls in love easily’c. un payaso enoja-dizo a clown anger-DIZO‘a clown that gets angry frequently’

(9) a. un hombre agobia-nte a man stress-NTE‘a stressing man’b. una mujer inquieta-nte a woman trouble-NTE‘a worrying woman’

3.3.3. ModalsThe final non-episodic class is modal in nature. Rainer (1999: 4600-4601, 4607-4609) identifies the groups of potential (9a) and deontic (9b) adjectives.

(10) a. una mesa transporta-ble a table transport-BLE‘a transportable table’b. un certificado paga-dero

15



a certificate pay-DERO‘a payable certificate’

These two classes are modal in nature, and as our dispositionals and habituals are non episodic; in fact, they are close to our dispositionals in the sense that they do not entail actual participation, past, present or future, in an event. (10a) is a truthful description in a situation where the table has never been transported, or perhaps will never be (eg., the table has just been put together in the shop and tomorrow an accidental fire will consume it); all that we need is that it is possible to transport the table. (10b) is equally truthful if nobody has paid the invoice and nobody will: all we need is that there is some obligation, through external law, to pay it.The question is in what sense these two readings are different from our dispositionals. Let us start with the deontic class, where the difference is easier to show. The minimal distinction here is that, in our dispositional reading, the hypothetical participation in an event is due to the particular properties of the entity, without necessary intervention of an external set of conditions. That is, if something is quebradizo ‘break-DIZO’, it is so because of its internal composition, the substance it is made of, its molecular structure, etc. However, if an invoice is pagadero ‘payable’, it is not so because its internal properties make it prone to be payed, or make it such that its natural state will tend towards payment, but because of a set of external circumstances –laws and regulations– that dictate it. There is, then, no necessary causal link between the properties of the invoice and its possible participation in a paying event. Note that the causal link is implicit in our definition of dispositionality, repeated here:

(11) Necessarily x is disposed to P in w if x would P in w

The causal connection required for dispositionality is evident when we consider counterfactuals, strongly tied to causality (particularly in some theories, like Lewis 1973). We say that someone is disposed to P if, given a set of facilitating circumstances, he would P in that context. Then, counterfactually:

(12) If we was not disposed to P in w, he would not P in w

Now, deontic deverbal adjectives do not involve a causal connection between the internal properties of the entity and the possible participation in an event, and thus there is no counterfactuality relation: an invoice is payable (eg., ‘must be paid’) even if in a facilitating context –where there is the obligation to pay it and there is money to do so– it is paid. That said, it is worth noting that in contemporary Spanish the deontic reading of adjectives is not productive. Rainer (1999: 4607) documents only some formations with -ble (eg., abominable ‘detestable’, from the unusual verb abominar ‘abhor’), -dero (substituted by -ble already in the 16th Century, cf. Clavería 2004) and a few exceptional suffixes which are not clearly adjectival (-dío as in rega-dío ‘water-DIO’, ‘field that must be watered’; -ndo, execra-ndo ‘loathe-NDO’, ‘something that must be loathed’).Consider now potentials, where the difference with dispositionals is perhaps more subtle. The crucial thing about potentials, vs. dispositionals, is that they allow participation in an event, but, crucially, are not inherently predisposed towards it because there is no necessary causal link between the entity’s internal properties and such event. Granted, there are some internal properties that a table has to have in order to be transportable –an upper limit to how heavy it can be, at least– but crucially

16



these internal properties do not by themselves force participation in the event, given the right situation: they just make it possible. There is, put more technically, no real causality between internal properties and event participation. For this reason, the counterfactual does not hold of potentials either: a table is transportable even if in a facilitating context it would not be transported, simply because –even though it can be transported– one choses not to transport it. Put in slightly more abstract terms: in a potential adjective, the entity has the necessary properties that allow it to participate in an event; in a dispositional adjective, those properties are sufficient to guarantee participation in that event.

(13) a. arroja-dizo throw-DIZO‘throwable’b. leva-dizo raise-DIZO‘raisable’c. rega-dizo water-DIZE‘that can be watered’

i. Habitual adjectives convert into nouns.ii. Adjectives in a habitual reading combine with animate nouns.

3.1. -dizo (15) a. roj-izo

red-IZO‘reddish, tending towards red’b. paj-izo straw-IZO‘like straw, of the colour of straw’

(16) a. mov-i-do move-ThV-ed‘moved’b. mov-e-dizo mov-ThV-DIZOc. mov-e-mos move-ThV-1pl‘we move’

This affix produces dispositional (17a) and habitual (17b) adjectives, and less productively, potentials (17c). (17) a. una sustancia quebradiza

a substance break-DIZO‘a substance whose properties make it prone to breaking’b. un hombre enamoradizo a man fall.in.love-DIZOc. un precinto quebradizo a seal break-DIZO‘a seal that can be broken’

17



(18) a. un soldado alquiladizo a soldier rent-DIZO‘a soldier whose services are rented’b. *un piso alquiladizo a flat rent-DIZOIntended: ‘a flat that is rented, that has been frequently rented’

In contrast, dispositionals typically prefer non animates as subjects. Por instance, in Mark Davies’s Corpus del español these are the nouns that appear combined with the adjective quebradizo ‘break-DIZO’.

(19) vanidad ‘vanity’; trasposiciones ‘transpositions’; tierra ‘earth, soil’; silencio ‘silence’; ramas ‘branches’; presentimiento ‘presentiment’; precinto ‘seal’; objeto ‘object’; moral ‘moral’; materiales ‘materials’; líneas ‘lines’; ídolos ‘idols’; fractura ‘fracture’; conchas ‘shells’; ánimos ‘courage’

(20) Alberto nunca fue quebrado.Alberto never was broken‘Alberto was never broken’

(21) *un prisionero quebradizo a prisoner break-DIZO

(22) a. un puente lev-a-dizo a bridge raise-ThV-DIZO‘a drawbridge’b. *un soldado levadizo a soldier raise-ThV-DIZOIntended: ‘a soldier that can be recruited’ (cf. levar soldados, ‘to recruit soldiers’)

(23) inútil > un inútiluseless a useless person

(24) un olvida-dizo ‘a forget-DIZO’; un enamora-dizo ‘a fall.in.love-DIZO’; un enoja-dizo ‘a anger-DIZO’2; un enfada-dizo ‘an anger-DIZO’3; un alquila-dizo ‘a rent-DIZO, someone who rents his services’; un espanta-dizo ‘a frighten-

2 Documented through google:

(i) Un enojadizo que siempre se sale con la suyaan anger-DIZO who always SE comes with his one‘An angry guy that always has it his way’


(i) Es un caprichoso, un cabezota y un enfadadizo.is a capricious, a stubborn and a anger-DIZO‘He is a capricious, stubborn, angry man’

18



DIZO’4; un contenta-dizo ‘a satisfy-DIZO’5; un muda-dizo ‘a change-DIZO’6; un hui-dizo ‘a flee-DIZO’

(25) *un resbala-dizo ‘a slip-DIZO’; *un quebra-dizo ‘a break-DIZO’; *un desliza-dizo ‘a slide-DIZO’; *un move-dizo ‘a move-DIZO’; *un roba-dizo ‘a steal-DIZO’; *un desmorona-DIZO ‘a crumble-DIZO’; *un leva-dizo ‘a raise-DIZO’; *un leva-dizo ‘a raise-DIZO’; *un arroja-dizo ‘a throw-DIZO’; *un rega-dizo ‘a water-DIZO’; *un corre-dizo ‘a run-DIZO’ (Intended: an object that can slide)

3.2. -ón(26) busc-ón ‘search-ON, that is frequently looking for something’, abus-ón

‘bully’, critic-ón ‘criticise-ON, that frequently criticises’, respond-ón ‘reply-ON’, sob-ón ‘fondle-ON’, trag-ón ‘swallow-ON’, fisg-ón ‘spy-ON’, grit-ón ‘shout-ON’, zumb-ón ‘buzz-ON’, lig-ón ‘flirt-ON’, gruñ-ón ‘grumble-ON’, bail-ón ‘dance-ON’, llor-ón ‘cry-ON’, adul-ón ‘flatter-ON’, machac-ón ‘crush-ON’...

(27) a. problem-ón problem-ON‘big problem’b. una mujer cabez-on-a a woman head-ON-f‘a woman that has a big head’c. un niño pel-ón a child hair-ON‘a child with very little hair’

(28) a. La coreografía [...] abusó del uso de la danza sobre puntas. the coreography abused of-the use of the dance on points‘The coreography abused of the use of dancing sur les pointes’b. El viento les destrozó la casa. the wind them destroyed the house‘The wind destroyed their house’c. El Caañabé [...] es un agua de barro que chupa y traga todo. the Caañabé is a water of mud that sucks and swallows everything‘The [river] Caañabé is a flow of muddy water that sucks and swallows everything’


(i) Es un espantadizo que se acobarda con nadais a frighten-DIZO who SE gets.scared with nothing’He is an easy to frighten man who gets scared in no time’


(i) Quizá sea una virtud, ser un contentadizomaybe is a virtue, to.be a satisfy-DIZO’Maybe it is a virtue, to be someone easy to satisfy’


(i) un veleta, un veleidoso, un mudadizoa fickle, an inconstant, a change-DIZO’a fickle, an inconstant, a person that changes his mind’

19



The corresponding adjectives, always habitual, reject these inanimate subjects:7

(29) a. *una coreografía abusona (de la danza) a coreography abuse-ÓN of the danceIntended: ‘a coreography abusing of dance’b. *un viento destrozón a wind destroy-ÓNIntended: ‘a wind that destroys many things’ c. *un río chupón y tragón a river suck-ÓN and swallow-ÓNIntended: ‘a river that sucks and swallows many things’

(30) un abus-ón ‘an abuse-ÓN’; un adul-ón ‘an flatter-ÓN’; un burl-ón ‘a mock-ÓN’; un busc-ón ‘a search-ÓN’; un comil-ón ‘an eat-ÓN’; un critic-ón ‘a criticise-ÓN’; un destroz-ón ‘a destroy-ÓN’; un mir-ón ‘a watch-ÓN’; un dormil-ón ‘a sleep-ÓN’; un sob-ón ‘a flinger-ÓN’; un mam-ón ‘a suckle-ÓN’; un peg-ón ‘a hit-ÓN’; un besuc-ón ‘a kiss-ÓN’; un refunfuñ-ón ‘un grumble-ÓN’; un regañ-ón ‘a reprehend-ÓN’; un pid-ón ‘a demand-ÓN’; un chup-ón ‘a suck-ÓN’; un pregunt-ón ‘an ask-ÓN’...

3.3. -nte(31) a. un empresario contamina-nte

a businessman contaminate-NTE‘a businessman that has repeatedly contaminated something’b. un producto contamina-nte a product contaminate-NTE‘a product that has contaminating properties’

(32) a. una noticia preocupante8

a news worry-NTEb. *un político preocupante a politician worry-NTE

Similar effects are found in other -nte adjectives that are compulsorily dispositional: among many others, tranquilizante ‘soothing’, aislante ‘isolating’, subyacente ‘underlying’, secante ‘drying’, purgante ‘purgative’ or endulzante ‘sweetening’. As verbs, they have no trouble taking animate subjects.

(33) a. un calma-nte a soothe-NTE‘a tranquiliser’b. un desodora-nte a deodora-NTE

7 Orientatively, when used as an adjective, tragón ‘swallow-ÓN’ (which would be easy to conceptualise with personification of the subject, but can combine with a wide variety of non animate subjects) combines with the following nouns (according to CREA and Corpus del Español): niño ‘child’, indigentes ‘poor people’, griegos ‘greek people’, monstruos ‘monsters’, frailones ‘friars’, socialistas ‘socialists’, público ‘audience’, and flacas ‘girls’ (in some American varieties), in addition to proper names and person pronouns. 8 Orientatively, in Corpus del Español the nouns that combine with preocupante are the following: sitación ‘situation’, dimensión ‘dimension’, tema ‘topic’, tendencias ‘tendencies’, dato ‘datum’, atención ‘attention’, aspecto ‘aspect’, abandono ‘abandonment’, signos ‘signs’, paréntesis ‘parenthesis, break’, mal ‘sickness, problem’, indicios ‘traces’, impunidad ‘impunity’, forma ‘form’, episodios ‘episodes’, endeudamiento ‘debt’ and dudas ‘doubts’.

20



‘a deodorant’

(34) a. un ataca-nte ~ *un hombre ataca-nte a attack-NTE a man attack-NTE‘an attacker’b. un solicita-nte ~ *un alumno solicita-nte a request-NTE a student request-NTE‘an applicant’c. un preside-nte ~ *una institución preside-nte a preside-NTE an institution preside-NTE‘a president’

(35) Generalisations about non-episodic deverbal adjectivesa. When the adjective combines with an animate noun, it can be interpreted as habitual.b. If the adjective is habitual, it allows A-to-N conversion.

In this section we will argue that the link between the noun’s properties and the eventuality denoted by the base verb is crucial to differentiate the habitual reading from the rest. Specifically, we will propose this generalisation: Habitual readings are triggered when the properties of the modified noun include the capacity to have mental states and the link between these properties and the adjective’s eventuality is of weak causation. Consider first the role of the noun’s properties in defining a reading as habitual. We have already seen that habitual readings are linked to cases where the modified noun is animate. The crucial property of animates is that any animate entity is able to have mental states, and this is the first crucial ingredient for habituality. We find a grammatical relation between the presence of animate subjects and the notion of habituality –as opposed to dispositionality– in a wide variety of contexts. Take, as an illustration, English middle constructions, which imply dispositionality and not habituality (Ackema & Schoorlemmer 1995, Lekakou 2005, Klingvall 2007, among many others).

(37) a. This book reads well.b. This window breaks easily.c. This safe opens with difficulty.d. This car drives with ease.

(38) a. John reads well.b. The karate expert breaks [something] easily.c. Houdini opens [something] with difficulty.d. Sena drives with ease.

Let us move now to the link between the properties of the noun and the properties of the eventuality embedded under the derived adjective. The habitual interpretation is triggered whenever the noun, by virtue of its properties, can be interpreted as the (internal or external) causer of the eventuality, the entity that initiates it, be it volitional or not. Animates combine with habitual readings because, being able to have mental states, can act easier as causers of that eventuality; as causers, their properties are enough to trigger (often, if necessary) that situation, making the facilitating circumstances secondary. Animate entities have inherent causation

21



properties (they can initiate processes by themselves) and moreover they can even control how some event progresses: fast, slow, in an orderly or disorderly fashion, etc., but non-animate entities cannot (Alexiadou & Schäfer 2006). The deep connection between animates and habituals (which in a sense are more ‘verby’ than dispositionals, because they presuppose that the event has actually happened) can be understood from this perspective.It is important to note that, when we use the notion of ‘weak causation’ we are not implying ‘volitionality’. In a case like huidizo ‘flee-DIZO, that often escapes’ it is very clear that the animate subject decides, at will, to escape, but other cases show that volitionality is not a necessary part of the description of the habitual class. It is clear from the entailments of many of our habitual adjectives that volitional control by the modified noun is not necessary, and sometimes plainly impossible. Consider the following verbs, that produce habitual adjectives in -dizo, as an example.

(39) olvidar ‘forget’, asustar ‘frighten’, enfadar ‘anger’

Traditional accounts would assign an experiencer role to the subject of predication of these adjectives (eg., Belletti & Rizzi 1988). In what way can we say that the experiencer acts as a ‘causer’ of the eventuality in these cases?The correct generalisation, in our view, is the following: the noun acts as a weak causer whenever the event can be started by the entity denoted by the noun, without the intervention of other external participants: intuitively, the subject can be construed as the only entity responsible for the event getting initiated, even if that entity is not responsible for the way in which the situation progresses, or even, does not play a role in whether the eventuality will progress or be maintained through time. This implies deconstructing what the initiator of an event is along the lines of what Van Valin (1990) or Dowty (1991), among others, have proposed: prototypical initiators (‘agents’) involve volitionality, but there is a second, more crucial for the definition of an initiator: weak causation of the event. By weak causation we mean that the fact that the event is initiated is attributed to the subject, and not to other participants. For explicitness, consider now how this causal link explains the two readings of contaminante ‘contaminating’. Remember (cf. 31) that this derived adjective gets a habitual reading with an animate subject (un empresario contaminante ‘a contaminating buisinessman’) and a dispositional reading with non animates (un producto contaminante ‘a contaminating product’). Why? The question is specially relevant because Rifón (1996) has characterised non-animates with these adjectives as ‘causers’. The businessman is the entity responsible for the contamination, but there is no sense of responsibility affecting the product when it contaminates. The businessman triggers the contamination event, but then he does not play a role in how the event progresses: he can disappear and the contamination could go on. In contrast, the product has a much weaker relation with respect to the initiation, as it does not determine when the contamination takes place: if lead contaminates the waters, lead must have some properties (eg., ‘being poisonous’) in order to keep the waters contaminated –that is, to guarantee the result or to maintain the progression of the ongoing contamination eventuality–, but it does not initiate it. It needs to be placed, though facilitating circumstances, in a position where it can contaminate in order to start the contamination eventuality. Thus, the businessman initiates the situation, but the situation can go on without him; the product acts in the opposite way: the contamination stops when it disappears, but it cannot initiate the eventuality without facilitating circumstances.

22



(41) a. Juan pulsó el botón. Juan pressed the button

b. El botón está pulsado (*por Juan). the button isestar pressed (by Juan)c. El botón fue pulsado (por Juan). the button wasser pressed by Juan

(42) a. Moriarty contaminó Londres. Moriarty contaminated Londonb. *Londres está contaminada por Moriarty. London isestar contaminated by Moriarty

(43) a. El plomo contaminaba Londres.9

the lead contaminated Londonb. Londres está contaminada por el plomo. London isestar contaminated by the lead

Similarly, if something is quebradizo ‘break-DIZO, fragile’, that something does not play a role in determining when the breaking event would take place, if ever. It must have a set of properties that allow it to be broken (eg., a solid entity with some physical unity), and that are relevant for the breaking process and for the result achieved after breaking, but that are not responsible for causing that breaking event.10 Given the central role of the causal link in habitual readings, the facilitating circumstances play, at best, a secondary role in habitual constructions. These facilitating circumstances do have some influence: if we say that someone is a smoker (that is, that someone often smokes), this is not denied by the fact that during a three hour long trip on a plane the person does not smoke. Crucially (Carlson 2011), we evaluate if something happens often attending only to situations where the event can happen, given our assumptions about the normal behaviour of the participants in that event. However, note that in habitual readings, unlike what we will see in §4.3. with respect to the dispositional readings, the facilitating circumstances do not lead deterministically to the event taking place. John can be a smoker even if in a situation where smoking is allowed (eg., an outdoors party) John decides not to smoke. Similarly, we say that someone is forgetful (olvidadizo) even if in facilitating circumstances –a particularly stressful week– that person happens not to forget his wife’s birthday.Let us see now how dispositonal readings are differentiated from modal readings, potential or deontic. We will contend that in order to distinguish dispositionals from potentials and deontics, crucial reference has to be made to the properties of the noun: dispositional readings are available when the properties of the noun are sufficient to guarantee participation in the event if the facilitating circumstances are in place;

9 There is a second difference between (42) and (43): external aspect (Smith 1991) is perfective in the first case and imperfective in the second. The reason is that (42a) refers to a change of state, given that the animate subject starts the event at some point –defining an initial point–, while the second (42b) refers to a stative situation where there is no defined starting or final point. This emphasises that the non-animate subject does not define the causation of the eventuality, while the animate subject defines a causative event that starts at a particular point.10 An anonymous reviewer, to whom we are grateful, points out that the anticausative construction with verbs like break is generally taken as internal causation. We follow Schäfer (2008: 176 and folls.) in the proposal that the anticausative construction is actually non-homogeneous typologically and language internally: there are at least three distinct anticausative constructions, but what they have in common is that no agent or causer theta-role is assigned to the subject. Thus, even under anticausative construals, the nouns would not be categorised as the initatiors of an eventuality.

23



modal readings emerge if the properties of the noun are necessary, but not sufficient, to guarantee participation in the event, even if there are facilitating circumstances. Consider (46) as an illustration.

(46) a. un sello quebra-dizo a seal break-DIZO‘a breakable seal’b. cabello quebra-dizo hair break-DIZO‘fragile hair’

The conceptual knowledge we have about seals is that they have a function, to close something. We know also, through our world knowledge, that things that break easily do not perform well when used to keep something closed. Hair, on the other hand, is not conceptualised as an instrument used with a particular function. In (46a), imagine that we try to break the seal of an envelope and we do not manage to break it in our first try: we would not say that the seal is not breakable, because the adjective never implied that the properties it had would make the breaking event unavoidable once there is facilitating circumstances; we would just think that we did not try hard enough. We would say that (46a) is not a good description of a seal, however, if after repeated attempts and with facilitating circumstances, the seal would not break: this is a modal reading (specifically, potential). In contrast, (46b) is dispositional: imagine that we pull the hair and it does not break. In this case, we might want to conclude that the hair was not fragile, because we facilitated the conditions for it to break and it did not break. What let us decide between these readings? Interpreting (46a) in a dispositional way would imply that a seal has been designed with a set of properties that would necessarily lead to it breaking if there are facilitating conditions. These facilitating conditions are not restricted to human intervention: violent movement, contact with other objects or changes in temperature could also set conditions where the unity of an object can be put at risk. However, out of context, this goes against our world knowledge about the use of this instrument: they are supposed to be entities whose function is precisely to keep something close even when there are facilitating conditions for it to open. Then, it is unexpected that a seal is designed with properties that would make it break, because this would mean that its properties are sufficient to guarantee that it will break; in contrast, it is expected that it is designed in such a way that breaking it is at least possible, meaning that it has the necessary properties to participate in a breaking event.If we were to pin down the specific property that defines the potential reading, it would be the notion of instrument. Instruments are designed with some function, and as such they have to be designed in a way that allows an external participant to control when and for how long they will satisfy that function. If we go back to the potential adjectives with -dizo, a striking property is that beyond being non animate entities, they are also names of tools and instruments: arma ‘weapon’, puente ‘bridge’, campo ‘field to grow plants’, etc.11

11 Even though we have left the suffix -ble ‘-able’ outside this investigation, note that the characterisation can be extended to its derived adjectives. Leaving aside highly lexicalised adjectives (amable ‘nice’), the nouns they combine with tend to be interpreted as instruments designed to satisfy the functions denoted by the base verb: un objeto legible ‘a book readable’ forces interpretation of the object, whatever it is, as a something intended for reading.

24



What we see here is that, in the absence of the causal link that defines habituals, the interaction between the facilitating circumstances and the semantic properties of the noun become crucial. Dispositionality is the situation where the properties of the noun guarantee that the facilitating circumstances will lead to the event taking place. These facilitating circumstances can be very complex: they include the possible conscious or involuntary involvement of external participants (force dynamics, for instance), the location of the object, the natural passing of time, etc. Modal readings emerge when the noun’s properties do not conduct, necessarily, to participation in the event if there are facilitating circumstances.

4.3. Potentials vs. deonticsThus, dispositionals are differentiated from modals because in the first, the combination of the properties of the noun and the facilitating circumstances guarantee that the event will take place. Our final contrast is the one within the class of modals: potentials and deontics. What they have in common is that the properties of the noun are not sufficient to ensure that the noun would participate in the event under the right conditions. The difference is the following: in the potential class, the facilitating circumstances play a crucial role, but they are irrelevant in the deontic one. If a weapon is arrojadiza ‘throw-DIZO, throwable’, it means that given reasonable facilitating circumstances that do not require the intervention of unusual instruments or helping tools, the weapon would be thrown. A spear is an arrojadiza weapon because reasonable circumstances (a warrior of normal strength, with enough space to take a run-up) guarantee that the weapon can be thrown. A tank would not be a throwable weapon, even if we can throw it under very special circumstances, because its design as a tool is such that under reasonable facilitating circumstances it could not be thrown (but unusual circumstances, like using a huge catapult, could make it possible).There are no documented adjectives in -dizo with a deontic meaning in contemporary Spanish (as in this use it was replaced historically by -ble, cf. Clavería 2004), and to the best of our knowledge there are no cases of deontic adjectives with -nte or with -ón (specialised, as was noted, in habitual readings); deontic adjectives in -dero are not frequent either: (48) and (49) are the only two documented cases with a deontic meaning with this suffix.

(48) paga-deropay-DERO‘that must be paid’

(49) vence-deroexpire-DERO‘that must expire’

In the deontic interpretation, even in the presence of facilitating circumstances, the event would not necessarily take place. The facilitating circumstances, in fact, would not play any role: something can be pagadero (48) even if there are reasonable circumstances that facilitate payment but payment does not take place. Moreover, we could argue that if something must be paid, it still must be paid even in a context where there are no facilitating circumstances (eg., the person does not have money to pay it), so the facilitating circumstances become essentially irrelevant for the deontic reading.

25



(50) Non-episodicity

weak causation No causation

habitual

Facilitating circumstances Facilitating circumstances

are relevant are irrelevant

deontic

The properties of the noun The properties of the noun

are sufficient to guarantee are necessary, but not sufficient, to

participation in a guarantee participation

facilitating context in a facilitating context

dispositional potential

3.4. Sincretismos: hacia una estructura

Habitual Disposicional Potencial Deóntico-dero

(plañidero, casadero, pagadero)-ble

(legible, punible)-dizo

(olvidadizo, quebradizo, levadizo)-ón

(mirón, picón)-nte

(contaminante, preocupante)

Además:- si algo tiene disposición a romperse cuando hay circunstancias

facilitadoras, es que puede romperse- si alguien tiene el hábito de fumar, es que tiene disposición a fumar

cuando hay circunstancias facilitadoras

26



Propuesta:

ScaleP

Scale XP = habitual

X YP = disposicional [uAnim]

[cont. var] Y

Y ZP = potencial / deóntico

(Op∃/∀) Z

Z EvtP = propiedades, sin ligar a T

ReferencesAckema, Peter and M. Schoorlemmer. 1995. Middles and movement. Linguistic Alexiadou, Artemis & Florian Schäfer 2006. Instrument subjects are agents or

causers. Proceedings of WCCFL 25, 40-48.and its crosslinguistic realization. Ph.D. Dissertation, University of London.

Apresjan, Ju. D. 1973. Regular polysemy. Linguistics 142, pp. 5-32. Arche, María J. 2006. Individuals in time. Amsterdam: John Benjamins.Asher, Nicholas 2007. A type driven theory of predication with complex types.

Fundamentae Informaticae 84 (2), 151-183.Baker, Mark C. 2003. Lexical categories: verbs, nouns and adjectives. Cambridge:

Cambridge University Press. Baker, Mark C. 2008. The syntax of agreement and concord. Cambridge: Cambridge

University Press.Bauer, Laurie & Salvador Valera 2005. Approaches to conversion / zero-derivation.

Münster: Waxman.Beard, Robert 1990. The nature and origins of derivational polysemy. Lingua 81, pp.

101-140. Belletti, Adriana & Luigi Rizzi 1988. Psych-verbs and theta-roles. Natural Language

and Linguistic Theory 3, 291-352.Bernstein, Judy 1993. The syntactic role of word markers in null nominal

constructions. Probus 5 (1-2), 5-38.Booij, Geert & Rochelle Lieber. 2004. On the paradigmatic nature of affixal

semantics in English and Dutch. Linguistics 42, pp. 327-357.Booij, Geert 2010. Construction morphology. Oxford: Oxford University Press.Bosque, Ignacio 1989. Las categorías gramaticales. Madrid: Síntesis.Bowers, John 1993. The syntax of predication. Linguistic Inquiry 24: 591-656.Bowers, John 2000. Predication. In M. Baltin & C. Collins (eds.), The handbook of

contemporary syntactic theory. Oxford: Blackwell. 299-334.Bréal, Michel 1897. The history of words. In G. Wolof (ed. / trans.), The beginnings

of semantics: essays, lectures and reviews. London: Duckworth, pp. 152-175. Cano Cambronero, María de los Ángeles 2013. Las derivaciones en ‘-nte’ y ‘-dor’:

estructura argumental y complejidad sintáctica en una morfología

27



neoconstruccionista. Ph. D. dissertation. Madrid: Universidad Autónoma de Madrid.

Carlson, Greg 2011. Habitual and generic aspect. In R. Binnick (ed.), The Oxford Handbook of Tense and Aspect. Oxford: Oxford University Press. 828-852.

Condoravdi, Cleo. 1989. The middle: where semantics and morphology meet. MIT Working Papers in Linguistics 11: 18-30.

Croft, William 1991. Syntactic categories and grammatical relations. Chicago: University of Chicago Press.

Croft, William 2001. Radical construction grammar: syntactic theory in typological perspective. Oxford: Oxford University Press.

Cruse, David A. 1992. Monosemy vs. polysemy. Linguistics 30, pp. 577-599. Dixon, Robert M. W. 1982. Where have all the adjectives gone? Berlin: De Gruyter.Dowty, David 1991. Thematic proto-roles and argument selection. Language 67 (3):

547-619.E. Reuland & W. Abraham (eds.), Knowledge of Language II, pp. 183-220. Dordrecht, Kluwer.

Folli, Raffaela & Heidi Harley 2008. Teleology and animacy in external arguments. Lingua 118, 190-202.

Hale, Kenneth & Samuel J. Keyser 1993. On argument structure and the lexical expression of syntactic relations. In K. Hale & S. J. Keyser (eds.), The view from building 20. Cambridge (Mass.): MIT Press. 53-110.

Hale, Kenneth & Samuel J. Keyser 2002. Prolegomenon to a theory of argument structure. Cambridge (Mass.): MIT Press.

Harley, Heidi & Rolf Noyer 2000. Formal versus encyclopedic properties of vocabulary: evidence from nominalisations. In B. Peeters (ed.), The lexicon-encyclopedia interface. Amsterdam: Elsevier. 349-375.

Hoekstra, Tuen and Ian Roberts. 1993. Middle constructions in Dutch and English. In: Hollósy, Béla 2008. The heterogeneous treatment of homonyms in dictionaries. In J.

Andor, B. Hollósy, T. Laczkó & P. Pelyvás (eds.), When grammar minds language and literature: Festschrift for Prof. Béla Korponay on the occasion of his 80th Birthday. Debrecen: University of Debrecen, pp. 207-216.

in English and Swedish. Ph.D. dissertation, University of Lund.Inquiry 26: 173-197.

Jackendoff, Ray 1997. The architecture of the language faculty. Cambridge (Mass.): MIT Press.

Janda, Laura. 2011. Metonymy in word formation. Cognitive Linguistics 22 (2): 359-392.

Katz, Jerrold J. & Jerry A. Fodor. 1963. The structure of a semantic theory. Language 39, pp. 170-210.

Kilgarriff, Adam 1992. Polysemy. Doctoral dissertation, University of Sussex.Klingvall, Eva 2007. (De)composing the middle. A minimalist approach to middles Krifka, Manfred 1999. Manner in dative alternation. WCCFL 18, 260-271.Lakoff, George 1987. Women, fire and dangerous things. Chicago: The University of

Chicago Press.Landau, Idan 2006. Ways of being rude. Ms., Ben Gurion University.Lehrer, Adrienne 2003. Polysemy in derivational affixes. In B. Nehrlich, Z. Todd, V.

Herman & D. D. Clarke (eds.), Trends in linguistics: Polysemy. Flexible patterns of meaning in mind and language. Berlin: Mouton de Gruyter, pp. 217-232.

28



Clavería, Gloria. 2004. Los caracteres de la lengua en el siglo XIII: el léxico. In R. Cano Aguilar (coord.), Historia de la lengua española. Barcelona: Ariel, pp. 473-507.

Lekakou, Marika. 2005. In the middle, somewhat elevated. The semantics of middles Levin, Beth & Malka Rappaport 1995. Unaccusativity. Cambridge (Mass.): MIT

Press.Lewis, David 1973. Counterfactuals. Oxford: Blackwell.Manley, David 2012. Dispositionality: beyond the biconditionals. Australasian

Journal of Philosophy 90 (2): 321-334.Marantz, Alec. 1997. No escape from syntax: don’t try morphological analysis in the Marchand, Hans 1969. The categories and types of present-day English word-

formation: A synchornic-diachronic aproach. München: Beck. Mel’cuk, Igor 1994. Cours de morphologie générale. Deuxième partie: significations

morphologiques. Montréal: Presses de l’Université de Montréal. Mendikoetxea, Amaya 1999. Construcciones con se: medias, pasivas e impersonales.

In I. Bosque & V. Demonte (dirs.), Gramática descriptiva de la lengua española. Madrid: Espasa. 1631-1723.

Norrick, Neal R. 1981. Semiotic principles in semantic theory. Amsterdam: John Benjamins.

Oltra-Massuet, Isabel 2013. Deverbal adjectives at the interface. Boston / Berlin: De Gruyter.

Roy, Isabelle & Elena Soare. 2012. L’enquêteur, le surveillant et le détenu: les noms déverbaux de participant aux événements, lectures événementielles et structure argumentale. Lexique 20, 207-231.

Panther, Klaus-Uwe & Linda L. Thornburg. 2002. The roles of metaphor and metonymy in English -er nominals. In R. Dirven & R. Pörings (eds.), Metaphor and metonymy in comparison and contrast. Berlin & New York: Mouton de Gruyter, pp. 279-319.

privacy of your own lexicon. In: A. Dimitriadis, et al. (eds.), Penn Working Papers in Linguistics 4.2, 201-225.

Pustejovsky, James 1995. The generative lexicon. Cambridge (Mass.): MIT Press.RAE & ASALE 2009. Nueva gramática de la lengua española. Madrid: Espasa.Rainer, Franz. 1999. La derivación adjetival. In I. Bosque & V. Demonte (dirs.),

Gramática descriptiva de la lengua española. Madrid: Espasa. 4595-4645.Rainer, Franz. 1993. Spanische Wortbildungslehre. Tübingen: Niemeyer.Rainer, Franz. 2011. The agent-instrument-place polysemy of the suffix –TOR in

Romance. Language Typology and Universals 64.1, 8-32.Rainer, Franz. 2013. Instrument and place nouns: a typological and diachronic

perspective. Linguistics 51.6, 1301-1359. Rifón, Antonio 1996. Sinonimia y polisemia de los sufijos -dor y -nte. Revista de

Lexicografía III, 95-109.Stowell, Tim 1991. The alignment of arguments in adjective phrases. In S. Rothstein

(ed.), Perspectives on phrase structure: heads and licensing. New York: Academic Press. 105-135.

Ullman, Stephen 1959. Semantics: an introduction to the science of meaning. Oxford: Blackwell.

Van Hout, Angeliek & Tom Roeper 1998. Events and aspectual structure in derivational morphology. MIT Working papers in linguistics 32, 175-220.

Van Valin, Robert 1990. Semantic parameters in split intransitivity. Language 66, 221-260.

29



Laca, Brenda. 1993. Las nominalizaciones orientadas y los derivados españoles en -dor y -nte. In S. Varela (ed.), La formación de palabras. Madrid: Taurus. 180-204.

Wasow, Thomas 1977. Transformations and the lexicon. In P. Culicover, T. Wasow & J. Bresnan (eds.), Formal syntax. New York: Academic Press. 327-360.

30


Documents

filcat.uab.catfilcat.uab.cat/clt/activitats/Barcelona3.docx · Web viewAlso, gender and number ageement is typical of other distinct word classes, like quantifiers and determiners