• Aucun résultat trouvé

Chapter 2: fi insertion in Tunisian Arabic

2.7. Causativized stative predicates

Besides progressive and SVCs, aspectual fi is systematically inserted before the theme argument of a causative verb derived from a dyadic stative verb (see also Brahim 2007). This is for instance the case of the causative forms of experiencer-subject/theme-object verbs such as ‘to hate’ and

‘to love’:

(180) a. semi karreh-nī fī-l-kosksi Semi made-hate.perf-me fi-the couscous

‘Semi made me hate couscous.’

b. *semi karreh-nī l-kosksi Semi make_hate.perf-me the couscous

(181) a. le-ktēb hadaya ḥabb semi fī-l-mutālʕa the.book this make_love.perf Semi fi-the-literature

‘This book is making Semi love literature.’

b. *le-ktēb hadaya ḥabb semi l-mutālʕa the.book this make_love.perf Semi the-literature

The sentences in (180) and (181) are only grammatical if aspectual fi precedes the theme-object of the sentence as illustrate by the starred examples.

The experiencer-object, in contrast, must not be fi marked:

(182) *le-ktēb hadaya ḥabb fī-semi fī-l-mutālʕa the.book this make_love fi-Semi the-literature

Brahim (2007) notices that both arguments of a causativized stative predicate are fi marked in progressive contexts, although the resulting construction is marginal:

(183) ??t-karreh fī-yyia fī-l-kosksi make_have.imp fi-me fi-the couscous

‘She is making me hate couscous.’

(Brahim 2007, 99: (15)b)

The arguments of a causativized stative predicate are both unmarked in generic contexts, which also requires the predicate to occur in the imperfective form:

(184) le-ktob kif haḏaya yeḥabbebu-nī l-mutālʕa the.books as this make_love.imp-me the-literature

‘Books like this make me love literature.’

The insertion of aspectual fi interacts with causativization but not with serialization as illustrated by examples (185)a-b where the stative verb ‘to hate’ occurs in an SVC embedded under the inceptive verb bdē. As the example under discussion illustrates, the theme of the stative verb must not be fi marked in constructions of that type:

(185) a. marwa bdit tkarhe l-kosksi Marwa start.perf hate.imp the couscous

‘Marwa started hating couscous.’

b. *marwa bdit tkarhe fī-l-kosksi Marwa start.perf hate.imp fi-the couscous

Example (185)a is an inchoative construction, meaning that the sentence denotes the “entry into a state” (Tatevosov 2002) and, as the example illustrates, inchoative constructions do not require a marked embedded object, unlike the inceptive SVCs discussed in the previous section (cf. 2.6.5).

A peculiarity of causativized stative predicates is that they require the insertion of aspectual fi in perfective contexts. This is at odds with the data discussed earlier (cf. sections 2.4.1 and 2.5) since fi insertion is attested with imperfective transitive predicates in progressive contexts, as in (186), but not with perfective transitive predicates, as in (187):

(186) semi (qāʕed) yedez fī-yya Semi prog push.imp fi-me

‘Semi is pushing me.’

(187) semi dez-ni Semi push.perf-me

‘Semi pushed me.’

Causativization, however, is not an independent trigger for aspectual fi insertion since the pattern of insertion illustrated by examples (186) and (187) is also attested with causativized dynamic verbs, as we can see from the examples in (188) and (189):

(188) semi (qāʕed) yeḍawweq fī-marwa fī-šoklata Semi prog make_taste.imp fi-Marwa fi-chocolate

‘Semi is making Marwa taste chocolate.’

(189) ḏawweq-ni l-kosksi made_taste.perf-me the-kosksi He made me taste the couscous.’

Perfective causativized dynamic predicates do not allow their direct object to be fi marked, while imperfective causativized dynamic predicates require both internal arguments to be fi marked in progressive contexts.

Causativized stative verbs, in contrast, require a marked object when they occur in the perfective form, as already illustrates by examples (180) and (181) making them a case on their own. Thus, causativized stative predicates in the perfective form require a fi marked object.

In addition to the facts just discussed, the following set of examples illustrates that stative predicates do not independently require the insertion of aspectual fi:

(190) a. semi yeḥebb-ni Semi love.imp-me

‘Semi loves me.’

b. *semi (qāʕed) yḥebb fī-yya Semi prog love.imp fi-me (191) a. semi ḥebb-ni

Semi love.perf-me

‘Semi loved me.’

b. *semi ḥebb fī-yya Semi love.perf fi-me

Hence, causativized stative predicates require a marked theme object in perfective contexts, (180) and (181), but transitive states, perfective morphology and causativized verbs taken in isolation do not independently trigger the presence of a marked object. These facts jointly indicate that the insertion of aspectual fi in the contexts under discussion must originate from the intersection of the three mentioned factors. I will argue at length in favor of this conclusion in the course of chapter 4.

Finally, examples (192) and (193) illustrate that causativized dynamic and stative predicates have similar morphosyntactic properties since they can both undergo movement above the enclitic negation -š:

(192) a. ma-ḏawweq-ni-š l-kosksi neg-make_taste.perf-me-neg the-kosksi

‘He did not make me taste the couscous.’

b. *ma-hu-š ḏawweq-nī l-kosksi neg-he-neg make_taste.perf-me the-kosksi (193) a. ma-karrhet-ni-š fī-l-kosksi

neg.make_hate.perf-me-neg fi-l-couscous

‘She did not make me hate couscous.’

b. *mi-he-š karrhet-ni fī-kosksi neg-she-neg make_hate-me fi-couscous

Under the assumption that verb raising to negation can be blocked in the presence of a null syntactic operator which interacts with the insertion of aspectual fi, i.e. optional qāʕed; the availability of verb movement in (193) points at the conclusion already proposed, namely: a causativized stative predicate requires the insertion of a marked theme for independent reasons and not because of the presence of a null operator, since an operator would possibly interfere with the movement.

One last point to consider is that not all causative verbs derived from stative predicates allow the presence of fi marked objects. Some verbs like, for instance, ʕarāf ‘to know’, do not allow the insertion of aspectual fi before the direct object, see example (194):

(194) a. ʕarraft semi ʕam li fēt meet.perf Semi year that pass.perf

‘I met Semi last year.’

b. *ʕarraft fī-semi ʕam li fēt meet.perf fi-Semi year that pass.perf

Importantly, however, this verb displays an idiosyncratic property, meaning that its causative form, i.e. ʕarrāf, does not mean ‘to make someone know’ ergo ‘to inform’ someone, but it translates as the dynamic predicate ‘to meet someone’. The unavailability of the “cause to know”

interpretation is clearly shown by the following ungrammatical examples which show that [-human] DPs are not suitable objects in ʕarrāf constructions irrespective of their type:

(195) a. *marwa ʕarrefet-ū fī-l-ḥqīqa Marwa make.know.perf-he fi-the-truth b. *marwa ʕarrefet-ū l-ḥqīqa

Marwa make.know.perf-he the-truth

Thus, causativized stative predicates require the insertion of a fi marked theme when they occur in the perfective form. The insertion of fi appears to be triggered jointly by the properties displayed by perfective causativized stative verbs, since each property taken in isolation on verbs of other kinds does not require a fi marked object.

With the exception of this context, the other constructions involving verbs of the same kind, i.e. progressive and generic sentences with a causativized stative predicate, display the pattern of fi insertion already encountered with other transitive and ditransitive predicates.