1Nominalizations (i.e., the formation of nominals from deverbal and deadjectival bases) remain something of a puzzle for linguistic theory, in spite of the central place they have taken in linguistic investigation for about fifty years (starting with the seminal work of Lees 1960 and Chomsky 1970). One of the reasons for this is their obvious trans-categorial status, responsible for their having mixed properties of both nominals and the predicative items (either verbal or adjectival) they derive from. This mixed status still presents an interesting challenge to standard syntactic theories.
2If one thinks of the prototypical function of verbs as naming actions or events and that of nouns as naming things/individuals, one fundamental property that distinguishes deverbal nominalizations from other nouns is their ability to denote events. This ability is present with verbs and generally absent with nouns, with some possible exceptions, such as movie, game if we conceive them as denoting events (cf. below). Note that if one were to treat events as entities, one could plausibly suppose that the function of Ns is to name entities across the board (including things, individuals, and also events).
3Deverbal nominals such as destruction, proposition, achievement may denote events (e.g., the destruction of the towers by Godzilla), and therefore raise the issue of the source of the event interpretation inside nominals. One option is that the event interpretation is a lexical property of these nouns. This amounts, however, to multiplying lexical entries in order to account for the existence of the eventive nominals and homophonous non-eventive ones. This leads, for instance, to distinguishing examination1 (as in The examination of the students by the professor) and examination2 (as in The examination is on the table), with different lexical semantic specifications. Another option explored in recent works (Alexiadou 2001, Borer 2001, 2003, 2005) would be that the source of the event which is present in eventive deverbal nominals is indeed structural, and relates to the presence in such nominals of a true event-related (either verbal or aspectual) structure. In this case, the observed properties are mere results of linguistic computation: functional layers that typically characterize verbs are responsible for similar properties in both verbs and derived nominals.
4In this introduction, instead of providing a complete overview of the field (and for such an overview, we refer the reader to Alexiadou et al. 2007, Alexiadou & Rathert 2010, Kornfilt & Whitman (in press), in particular), we will rather focus on some major but often neglected issues underlying the reflection on nominalizations, present several new developments and insights that have recently emerged, and show how a range of open questions may find answers in the papers collected in the present issue.
5A classical problem in the study of nominalizations has always been their semantic ambiguity. Since Grimshaw’s (1990) seminal study, deverbal nominals are known to exhibit an important correlation between the obligatory presence of an argument structure (associated with the base verb) and the eventive interpretation (i.e., the event that caries the aspectual properties of the underlying verb). For instance, a nominal like examination can be eventive and realize obligatory argument structure in (1a), and it can be non-eventive and lack an argument structure in (1b); in turn, a nominal like exam unambiguously denotes an object and has no eventive properties (1c):
(1)
|
a.
|
the examination of the patients in one day/*was on the table.
|
|
b.
|
the examination *in one day/was long/was on the table.
|
|
c.
|
the exam *in one day/was long/was on the table.
|
- 1 But see section 3, below, for a third class of Simple Event nominals.
6Grimshaw relates this ambiguity to a distinction between Complex Event nominals and Result nominals. 1 Given the observed correlation between Argument Structure and eventive interpretation, in recent work on nominalizations (cf. Borer 2005, Alexiadou 2010a-b among others) the distinction has been restated in terms of Argument-Supporting (or AS)-nominals and Referential (R)-nominals. The table in (2) provides a map of their characteristic properties.
(2)
|
AS-Ns
|
R-Ns
|
i.
|
event reading
|
no event reading
|
ii.
|
obligatory arguments
|
arguments not obligatory
|
iii.
|
compatible with aspectual
|
not compatible with
|
|
modifiers like in three hours
|
aspectual modifiers
|
iv.
|
constant, frequent with
|
constant, frequent possible only
|
|
the singular
|
with the plural
|
v.
|
by-phrase is an argument
|
by-phrase is not an argument
|
7This map of acknowledged properties of the two main classes of deverbal nominalizations provides the basis for syntactic accounts of AS-Ns, known broadly as the “[NP [VP]] approach” (see Fu, Roeper & Borer 1991/2004 for a recent overview). The projection of argument structure inside deverbal nominals is seen in this type of analysis as a property of a verbal layer included in their structure. The compatibility with aspectual modifiers is related to the existence of verbal functional layers that have received different labels throughout the literature, from “Event Phrase” in Van Hout & Roeper (1998), to different flavors of AspP as in Borer (2005), Alexiadou et al. (2010).
8The presence of eventive layers in the representation in syntactic approaches to nominalizations provides a nice interface with the neo-davidsonian semantic tradition, where the dynamic character of verbal constructions is related to the presence of an event variable in the semantics of these nominals as well. The representation of events inside nominals, however, raises a number of further questions. We will address some of these questions in what follows, as they are directly relevant to the papers presented in this volume.
9The first question relates to the idea of ‘events’inside nominals altogether. What does it mean for a nominal to be eventive? Is it the same thing as with a verb? Is it the same thing across nominals?
10In addition to the distinction between CENs and RNs, mentioned above, Grimshaw (1990) also distinguishes a class of Simple Event nominals, such as game, movie, crime, meeting. These nominals are possibly eventive, even though they are devoid of overt arguments. They differ from R-nominals in referring to an event rather than an individual entity (cf. table, book, boy).
- 2 The terms ‘strong’ and ‘weak’ eventivity should be understood descriptively, as meaning that the no (...)
11This view clearly implies two different senses of ‘eventivity’ and possibly the existence of a ‘strong eventivity’ and a ‘weak eventivity’. 2 There is a consensus that the projection of argument structure is responsible for the presence of a structural eventive layer in AS-Ns (Grimshaw’s CENs), determining the range of properties that are listed in the first column in table (2) in section 2. This strong eventivity is correlated to the presence of arguments, aspectual structure, etc. A SEN, which denotes a (simple) event but does not project argument structure is not eventive in the strong sense.
12We note, in this respect, that when by-phrases are present, they will have an argument status in CENs, as a reflect of their obligatory argument structure, while they will only be adjuncts in SENs (sometimes even taking a possessive or ‘author’ interpretation). This is supported by differences in prepositions in languages such as Romanian, in which “de către” by-phrases are unambiguously argumental (see Cornilescu 2001 among others):
(3)
|
a.
|
interpretarea acestei sonate de către Dinu Lipatti în 1960 a surprins pe toți
|
|
|
interpretation this. Gen sonata by Dinu Lipatti in 1960 has surprised ‘pe’ everyone
|
|
b.
|
*interpretarea de către
|
Dinu Lipatti a surprins
|
pe toți
|
|
|
|
interpretation by
|
Dinu Lipatti has surprised
|
‘pe’ everyone
|
|
|
b.
|
acest concert de (*către)
|
Dinu Lipatti a fost
|
difuzat
|
la radio
|
|
|
this concert by
|
Dinu Lipatti has been
|
transmitted
|
on radio
|
13What SENs are, however, remains unclear. From a purely semantic perspective, the split might seem unfounded: both SENs and CENs belong to the same ‘event-denoting’ class: meeting and destruction are, from this point of view, rather similar. There are, however, further arguments to distinguish them (besides the projection of obligatory argument structure already mentioned). First, CENs can inherit temporal adverbials associated with the underlying verb, while SENs cannot:
(4)
|
a.
|
the destruction of the house in 3 hours
|
|
b.
|
they destroyed the house in 3 hours
|
(5)
|
a.
|
#the meeting for 3 hours (in the main conference room)
|
|
b.
|
they met for 3 hours (in the main conference room)
|
14Another property that supports the distinction between the two classes, is the availability of control in purpose clauses in CENs but not in SENs (6):
(6)
|
a.
|
John’s playing cello in order to win a medal
|
|
b.
|
*John’s cello concert in order to win a medal
|
15Moreover, in Romanian, SENs, on a par with R-Ns, admit adjectival de ‘from’-time/place adjuncts – see (7a-b) vs. (8a), which are incompatible with CENs, as (8b) shows (cf. Cornilescu 2001).
(7)
|
a.
|
casa
|
de
|
pe
|
deal
|
|
|
house.the
|
from
|
on
|
hill
|
|
|
‘the house on the hill’
|
|
b.
|
catastrofa
|
de
|
la
|
Fukushima
|
|
|
disaster.the
|
from
|
at
|
Fukushima
|
|
|
‘Fukushima disaster’
|
(8)
|
a.
|
discuția
|
din
|
acest
|
paragraf
|
|
|
|
|
discussion.the
|
from
|
this
|
paragraph
|
|
|
|
|
‘the discussion in this paragraph’
|
|
b.
|
discutarea
|
acestor
|
argumente
|
în/*din
|
acest paragraf
|
de către autor
|
|
|
discuss.Inf.the
|
these.Gen
|
arguments
|
in/*from
|
this paragraph
|
by the author
|
16There is, then, sufficient empirical basis to motivate a distinction between CENs and SENs, and it can be concluded that despite their having a certain common eventive semantics, the properties of SENs make them merely similar to R-Ns in the table in (2), rather than AS-Ns. The realization of argument structure, correlating with all other ‘strong’ event properties, seems to be the relevant discriminating factor, as originally described by Grimshaw ; it is therefore justified to distinguish ‘strong’ events from ‘weak’ events, the latter behaving on a par with individuals and things.
17Aside from CENs, names for participants in an event (e.g., -er nominals and their equivalent across languages) have also been claimed to be eventive, or episodic (cf. Alexiadou & Schäfer 2010, Roy & Soare (to appear), specifically). Their eventive component can be shown by their compatibility with event-modifying adjectives such as ‘frequent/occasional’ (9), as well as adjectives such as ‘beautiful/old’ in their event-oriented interpretation (10):
(9)
|
a
|
frequent flyer; the occasional buyer
|
(10)
|
a
|
beautiful dancer (= who dances beautifully); an old learner of Chinese (=who has been learning Chinese for a long time)
|
18We note, however, that compatibility with event-modifying adjectives is not unique to deverbal nominals (cf. an old friend, which can be interpreted as having a long lasting friendship), suggesting that the adjectival modification test does not signal the ‘strong event’ under discussion here specifically. This is confirmed by the impossibility of temporal modification illustrated in (11), as also pointed out by Alexiadou (2001) for Greek, examples in (12):
(11)
|
*the flyer in 5 hours (to Paris)
|
(12)
|
a.
|
*i
|
damastes
|
ton
|
fotonion
|
mesa se/gia
|
enan
|
eona
|
|
|
|
|
the
|
tamers
|
the.Gen
|
photons
|
within/for
|
a
|
century
|
|
|
|
b.
|
*o
|
katharistis
|
tu
|
ktiriu
|
epi
|
ena
|
mina
|
telika
|
apolithike
|
|
|
the
|
cleaner
|
the.Gen
|
building
|
for
|
a
|
month
|
finally
|
got fired
|
19Nevertheless, we must be able to distinguish these (‘weak’) eventive -er nominals from, for example, instruments (which are non-eventive altogether). As shown in Roy & Soare (to appear) for French, the distribution of event-oriented adjectives big/happy/old distinguishes instrumental nominals (incompatible with the event-oriented reading of adjectives), from dispositional nominals (which admit them), as in (13a-b):
(13)
|
a.
|
Les gros consommateurs/vendeurs font tourner la machine économique.
|
|
|
the big consumers/sellers run the economy
|
|
b.
|
#un gros broyeur est toujours utile
|
|
|
a big grinder is always useful(with the event-oriented reading, meaning which grinds much’)
|
20In turn, Roy & Soare (to appear) show that, even though both episodic (14a) and dispositional (14b) participant nominals involve an event, which is absent from instrumental nominals, they are different in that the former, but not the latter, are compatible with frequent modifiers. Episodic nominals involve a particular event, while dispositional nominals do not. The combination of the two tests (big/happy/old in their event-oriented reading, and frequent-type modifiers) therefore provides a three-fold classification of participant nominals that better describes their properties (see also below).
(14)
|
a.
|
un consommateur fréquent de plusieurs drogues douces
|
|
|
a frequent user of many soft drugs
|
|
b.
|
#un consommateur fréquent;
|
#un vendeur fréquent de voitures
|
|
|
a frequent consumer
|
a frequent seller of cars
|
21The idea of an event inside -er nominals is therefore defensible. In German this goes even further, inasmuch as in this language -er nominals not only function as ‘agent’-nominals, but may also function as truly event-denoting nouns (with specific aspectual properties): ein Piepser ‘a beep’, ein Klopfer ‘a knock’. This provides an interesting ground for exploring events inside -er nominals, if one is to assume that there is only one -er morpheme (cf. Schäfer, this volume).
22In this line of reasoning, very little attention has been paid to nominals derived from non-event predicates. Two important cases that do not involve a dynamic event, but plausibly an eventuality in a broader sense, nevertheless, have often been neglected: nominalizations of stative verbs (e.g., knowledge, hatred) and deadjectivals (e.g., redness, pride, length). These two classes of nominals have little in common, however, and we shall postpone a discussion of deadjectivals until section 6 below, focusing here on nominals derived from stative verbs.
23Nominals formed from stative verbs are not eventive per se; although they may be argued to involve an ‘eventuality’ in a broader sense, here a state. Recent work on stativity has claimed that stative verbs may in fact involve a state variable. Parsons (1990), and more recently Roy (2006) and Martin (2009) have argued that states do not differ from events in having a davidsonian argument. Maienborn (2004) and Rothmayr (2009) claim that two groups of stative verbs must be distinguished, those involving a davidsonian state (a concrete state variable), and those involving a kimian state (an abstract state variable). Without entering into the details of the argumentation, these positions open the way to a new reflection about nominals derived from statives (as in Fábregas & Marín 2011, for instance). How do nominals derived from statives behave with respect to the three classes distinguished by Grimshaw? Is there a correlation between an eventuality reading for nominals derived from statives and argument structure as found with event nominals?
24One issue concerns the putative inheritance of the stative aspect, and/or the stative verb properties inside the related nominal. Recent works (cf. Barque, Fábregas & Marín (to appear), among others) have shown, for instance, that psychological verbs, may form nominalizations alternating between a (psychological) state-noun (which allows argument structure) and an object/cause of the state (which lacks arguments). The psychological state interpretation is related to the presence of an experiencer, suggesting that fine-grained aspectual and argument structure properties of the stative verb are visible for nominalizations (and see also Fábregas, Marín and McNally (to appear)).
25The question of the argument structure with stative verbs is, however, far from trivial. The external argument of a stative is normally seen as a Holder/Experiencer (and we may remain agnostic as to whether the distinction is needed at all; see Landau 2010, for instance). Most views on statives accept that their complements, when realized, are in fact part of the description of the state (i.e., ‘rheme’ objects for Ramchand 2008; ‘fusionned’ predicates for Krifka et al. 2005). This difficult question transfers to the nominalizations as well, and very few reliable tests can tell us whether (15) has an argument structure (as opposed to a genitive of possession):
(15)
|
John’s knowledge of music
|
26In most cases a direct transfer of the verb’s arguments into the nominal construct is simply impossible:
(16)
|
a.
|
Mon
|
fils
|
croit
|
au
|
Père
|
Noël.
|
|
|
my
|
son
|
believes
|
in
|
Santa
|
Claus
|
|
b.
|
*la
|
croyance de
|
mon
|
fils au
|
Père
|
Noël
|
|
|
the
|
belief of
|
my
|
son in
|
Santa
|
Claus
|
(17)
|
a.
|
Sarah
|
connaît
|
le
|
résultat
|
de
|
l’opération.
|
|
|
|
sarah
|
knows
|
the
|
result
|
of
|
the operation
|
|
|
b.
|
*la
|
connaissance
|
de
|
Sarah
|
du
|
résultat
|
de l’opération
|
|
|
the
|
knowledge
|
of
|
Sarah
|
of the
|
result
|
of the operation
|
27Example (17) demonstrates the fact that a double genitive is generally banned, a generalization that holds cross-linguistically.
28The correct typology of stative nominals needs also to be investigated. In that perspective, an interesting path would be to consider forms that alternate between an active and a stative reading: John decorates the house/The flowers decorate the table, as opposed to verbs that do not allow such alternation (cf. Alexiadou, this volume). As Fradin (2010) pointed out, in the vast majority of cases, with such alternating verbs, the nominal can only pick the active base (John’s decoration of the house/*The flowers’ decoration of the house). On the other hand, as discussed by Fradin (this volume), we also find that nominals interpreted as stative (emprisonnement, lit. im-prison-ment ‘detaining’) do not necessarily derive from a stative verb (emprisonner, lit. im-prison ‘put in jail’). The interpretational/structural relationship between stativity and dynamicity, and/or agentivity, needs to be better understood if we are to understand the inheritance properties of derived nominals generally. The inheritance of stative aspect in derived nominals remains an open issue; as does the role of (overt vs. zero derived) morphology in the building of stative aspect inside these nominals.
29The idea of aspectual distinctions inside nominals goes again back to Grimshaw (1990), according to which the difference in properties between ASNs (her CENs), on the one hand, and R-Ns (her SENs and Result Nominals), on the other, – such as the projection of obligatory argument structure, the availability of adverbial aspectual modifiers, and control into purpose clauses, discussed above – is connected to the presence of an internal (complex) aspectual structure (leading to ‘strong’ eventivity) in these nominals (cf. for example, (4)/(5) above). If we take this claim seriously, it means that at some level deverbal AS-Ns must not only encode the inner aspect of the verb base (i.e., lexical aspect/Aktionsart), but also, and maybe more importantly, outer aspect as well (i.e., grammatical aspect, such as imperfectivity, for instance).
30Recent studies of the fine-grained distinctions among AS-Ns point to the possibility that this is indeed the case. Cornilescu (2001) has argued, for instance, that the Romanian infinitive and supine nominalizations encode different aspectual values (telicity vs. atelicity, respectively). This difference is plausibly situated at the level of outer aspect (i.e., imperfectivity; see Iordăchioaia & Soare 2008, 2009). As another illustration, Ferret et al. (2010) have argued that French event nominals in -age vs. -ée encode imperfectivity (in (a)) and perfectivity (in (b)), respectively:
(18)
|
a.
|
(??après)
|
l’arrivage
|
de
|
la marchandise
|
(est
|
en
|
cours).
|
|
|
after
|
the. arriving
|
of
|
the merchandise
|
is
|
in
|
process
|
|
b.
|
(après)
|
l’arrivée
|
de la marchandise
|
(*est en cours)
|
|
|
|
|
|
after
|
the. arrival
|
of the merchandise
|
is in process
|
|
|
|
31Whether nominalizing morphemes themselves may contribute an (outer) aspectual value, or such value is provided by the internal make-up of the nominalizations is subject to some debate. In the light of minimal pairs such as arrivage/arrivée (from arriver ‘arrive’) (18), perçage/percée (from percer ‘drill’), where the only difference seems to be the actual suffix involved, the role of morphology appears rather straightforward. Knittel (this volume) presents arguments, however, in support of the opposite view, according to which the aspectual properties of derived nominals are related to the inflectional system inside the nominal, and more particularly number, addressing the relationship between semantic/morphological plurality.
32The idea that aspect inside event-nominals is correlated to number, and more specifically, plurality, has been explored in various works in recent years. One generalization put forward by Grimshaw (1990) for CENs (our AS-Ns) states that they are unable to pluralize (19a-b) (which is often seen as a consequence of their being [–count], and is confirmed by the general ban on discrete determiners and quantifiers (19c)):
(19)
|
a.
|
The shooting of rabbits is illegal.
|
(AS-Ns)
|
|
b. *
|
The shootings of rabbits are illegal.
|
|
|
c. *
|
A/*One/*That shooting of rabbits is illegal.
|
|
33Recent works on Romance languages have challenged this generalization; see Roodenburg (2006) for French, Iordăchioaia & Soare (2008, 2009) for Romanian and Alexiadou et al. (2010) for a cross-linguistic perspective. This forces a redefinition of the issue, the problem arising from these studies being that AS-Ns display a non-uniform behavior with respect to plurality; what precise property determines the blocking of plural in some – but not all – event nominals across languages? How does this relate to the mass/count distinction among nominals across the board? (see Huyghe, this volume).
34From a structural perspective, one possibility that has been seriously explored is to relate it to the correlation between inner aspect (telicity) and number (as put forward by Mourelatos 1978, Borer 2005; cf. also Knittel, this volume); only telic deverbal nominals are predicted to pluralize. Another possibility has been to relate this property to the competition between the projection of (outer) Aspect and Number in these nominals (cf. Iordăchioaia & Soare 2008, 2009 and Alexiadou et al. 2010). What seems to be shared by these views is the general agreement on the fact that aspect (either the Aktionsart contributed by the verbal base or built-in aspect inside nominals) determines the count properties of derived nominals, and this again shows that typically verbal information is accessible inside derived nominals.
35Additional properties of nominalizations seem to go beyond the (verbal) phrasal and inflectional layers (i.e., Tense, Aspect or Number), grounding another parallel, namely between nominalizations and clauses. We know, that clausal nominalizations do exist across languages (as for instance, in Turkish and Greek (cf. ex (20) from Alexidou et. al 2007), Malagassy (Ntelitheos 2006), but they seem to be absent in others (Romance languages, Germanic languages, etc.), where nominals involving a complementizer (CP) and/or Tense (TP) layer seem to be lacking:
(20)
|
To oti
|
irthe
|
me
|
stenaxori
|
|
the that
|
he-came
|
me
|
upsets
|
|
‘The fact
|
that he
|
came
|
upsets me’
|
36However, looking at names of event participants, Roy & Soare (to appear) have argued that they may involve an inner genericity, necessarily contributed by a full clause upon which the nominalization is built. Genericity can be built at the clausal level; similarly to characterizing sentences such as Lions have bushy tails (Carlson & Pelletier 1995). It may also be associated with differences in the specificity of the object (the episodic reading being associated with specific arguments, and the generic reading with non-specific ones): Ben likes this book/Ben likes (good) books. For Roy & Soare (to appear), the role of the internal argument in bringing about an episodic vs. generic reading for nominalizations is therefore identical to what is found at the clausal level. They point out the neat mapping between the specificity of the internal arguments, the episodicity of the underlying event and the episodicity of the derived nominal, such that [–specific] arguments give rise to a generic interpretation inside the derived nominals (also sometimes referred to as a ‘dispositional’ reading, cf. Alexiadou and Schäfer 2010), as is commonly the case at the clausal level.
(21)
|
a.
|
le vendeur de voitures
|
> dispositional/generic ‘agent’ N
|
|
|
‘the car-seller’
|
|
|
|
i. /vendre des voitures/
|
> generic event
|
|
|
‘sell the cars’
|
|
|
|
des voitures
|
> non-specific DP object
|
|
|
‘PART. cars’
|
|
|
b.
|
le vendeur des voitures
|
> episodic ‘agent’ N
|
|
|
‘the seller of the cars’
|
|
|
|
i. vendre les voitures/
|
> episodic event
|
|
|
‘sell the cars’
|
|
|
|
ii. les voitures
|
> specific DP object
|
|
|
‘the cars’
|
|
|
c.
|
le vendeur de cette voiture
|
> episodic ‘agent’ N
|
|
|
‘the seller of this car’
|
|
|
|
i. vendre cette voiture/
|
> episodic event
|
|
|
‘sell the car’
|
|
|
|
ii. cette voiture
|
> specific DP object
|
|
|
‘this cars’
|
|
37The parallelism between the semantic effects at the clausal level and the ones found with nominalizations suggests strongly that -er nominals are derived from full clauses, thereby possibly expressing genericity vs. episodicity. This type of nominalization illustrates new dimensions of the correlation between eventive (sentential) layers and argument structure in deverbal nominals.
38The other large group of non-eventive predicates that has often been neglected is that of deadjectival nominalizations. Some issues here are similar to the ones introduced by nominalizations based on stative verbs (section 3.3.), while others are more specific: (i) are all occurrences of deadjectival nominals the same? And what is the right typology of deadjectival nominalizations? (ii) what is their relationship to event nominals across the board?, and (iii) do deadjectival nominals have arguments?
39The issue of typology breaks down into two different, though related, questions. First, one might wonder if deadjectival nominals form a unified class, with unified interpretational and morpho-syntactic properties. We are not particularly concerned here with whether deadjectival nominals all express a unique type of object (e.g., they may express qualities and properties, as in Van de Velde 1995 and Flaux & Van de Velde 2000), but rather whether they are the same item in all of their occurrences. Is kindness always the same nominal, or are further distinctions needed, reminiscent at some level of Grimshaw’s distinctions? Roy (2010) has shown that deadjectival nominals with and without the realization of the argument structure of the base adjective behave differently, in particular in their compatibility with temporal modifiers. This finding suggests that, as for ‘complex-event’ nominals, deadjectival nominals contain an underlying eventuality in (22) but not in (23). The contrast between the two types is confirmed by the obligatory presence of the article in (22) and its absence in (23):
(22)
|
The kindness of John towards his mother was greatly appreciated.
|
(23)
|
Kindness is a quality that is valued by all persons.
|
40If a distinction between quality-nominals and property-nominals is needed, it is super-imposed on the distinction in (22) and (23). The contrast in (22) and (23) shows also that early views that exclude deadjectivals from the set of nominalizations (and therefore claim the absence of a derivational relationship between kind and kindness) are problematic, as further confirmed by the well-described fact that only adjectives that have a predicative use can form the base of a nominalization (see Fradin & Kerleroux 2003, for the French suffix -ité in particular, and Roy 2010 for a more general discussion). If there were no derivational relationship between the base adjective and the noun, such a restriction should not exist, and there should not be a ban on any class of adjectives in particular.
41The second aspect of the issue is how best to describe the properties of the expressions in (22) and (23), and the ontology of these expressions, i.e., whether these properties are derivable from other principles or primitives. Moltmann (2004) and Villalba (2009) argue that deadjectival nominals such as those in (22) are ‘tropes’ (which for Villalba can come in different flavours, i.e., property-tropes and quality-tropes, thereby reintroducing the earlier distinction by van de Velde). Taking the opposite stand, deadjectival nominals may express properties. Taking seriously the idea that properties are predicates, and that this is reflected in the structures, it can be argued that deadjectival nominals are derived from properties, and that there may be no linguistic relevance for a further enrichment of the ontology (cf. Arsenijević, this volume).
42Another issue that bears also on the question of typology, in the domain of deadjectival nominals in particular, comes from the conjunction of the fact that the nominal expression in (23) can generally be combined with a genitive (e.g., John’s kindness) and the across-the-board absence of reliable tests for distinguishing arguments (introduced by a genitive) and mere possessives. The status of nominal complements with such nominals must be further investigated, namely whether they are N modifiers or real arguments. We would like to suggest that French may offer such a test; or at least, offers a way to think about this issue. Genitives of possession are expressed by de-phrases in French, but colloquial French tends to prefer the dative to express the possessor:
(24)
|
la voiture de Jean ;
|
le problème de Jean
|
|
the car of John
|
the problem of John
|
|
‘John’s car’
|
‘John’s problem’
|
|
|
(can potentially express all sorts of relations. It may mean: the problem that belongs to John, the problem that John created, or the problem that John represents, etc.)
|
(25)
|
la voiture à Jean ;
|
le problème à Jean
|
|
the car at John
|
the problem at John
|
|
‘John’s car’
|
‘John’s problem’
|
|
|
(only means possession)
|
43When a deadjectival nominal is accompanied by a de-phrase (e.g., la gentillesse de Paul; ‘Paul’s kindness’ (26)) it could either be a possessive genitive or a holder argument inherited from the adjective. If a possessive, it can be replaced by the dative:
(26)
|
la gentillesse de Paul
|
la gentillesse à Paul
|
|
the kindness of Paul
|
the kindness at Paul
|
|
‘Paul’s kindness’
|
‘Paul’s kindness’
|
- 3 Two notable issues would have to be considered before offering a firm test: (i) the relatively low (...)
44However, when the argument structure is fully realized, the substitution of de with à becomes impossible, which can only be shown with transitive adjectives. Pending further investigation, the realization of a full argument structure seems, in French, to block the dative replacement: 3
(27)
|
la gentillesse de Paul envers sa mère ;
|
#la gentillesse à Paul envers sa mère
|
|
the kindness of Paul towards his mother
|
the kindness of Paul towards his mother
|
45This point illustrates the difficulty to find appropriate tests for argumenthood, with deadjectival Ns and Ns derived from stative predicates more generally.
46Recent and ongoing research on nominalizations point in the direction of refined verbal/aspectual distinctions inside derived nominals, once the right classes of nominals are considered, teasing apart the cases where inheritance of the properties of the base are expected from cases where it is not. This introduction has presented current issues in the domain of nominalizations to which the papers gathered in this volume offer a ground for discussion. Among these issues are the notion of events inside nominals, the typologies of derived nominals, the correlation between morpho-syntactic and semantic properties in the domain of number, argument realization and aspectual distinctions, which we hope open the way to a better understanding of what Nominalizations are.