2.- El mètode de Foucault
2.2. L’historiador
Campos and Stavrou (2004) provide a unified account of polydefinites in both Greek and Aromanian, a Balkan Roman language. Crucially, they argue that only one definite adjective can appear in a polydefinite construction and that any additional adjectives are instances of parallel modification as defined by Sproat and Shih (1991:578–579). As was discussed in section 3.2 of chapter 3, the claim is
that adjectives in parallel modification modify the noun in a separate phonological and syntactic phrase, and do not follow the fixed order that is observed cross-linguistically.
Campos and Stavrou analyse polydefinites with a single adjective in a pred-icative structure, which involves the functional category PredP. The fact that adjectives are unambiguous and only come with a single reading with collective nouns, follows directly from the assumption that polydefinites have a predica-tive source. What we see below is that only one reading is available when the adjective is found in a postcopular position:
(39) a. o the
fititis student
ine is
ftoxos poor
‘impoverished’/*‘pitiable’
b. to the
zevGari couple
ine is
oreo beautiful
‘beautiful as a couple’/*‘beautiful independently’
The base structure Campos and Stavrou propose is given in (40).
(40) FocP
Foc0 DP
D0 i pena
‘the pen’
PredP pro Pred′
Pred0 i
‘the’
AP asimeña
‘silver’
Campos and Stavrou assume that the second definite article, which is found with the adjective, is the realisation of the predication operator and is, therefore, merged under Pred0. The subject of Pred0 is a silent pronoun, which can be optionally spelled out as the anaphoric demonstrative afto ‘this’, and the com-plement of Pred0 is the AP.12
12The same demonstrative can be deictic when it is prenominal, but must be anaphoric when postnominal. This is illustrated in the examples below, taken from Campos and Stavrou (2004:159, (42)):
(41) a. afto this
(eDo) here
to the
vivlio book
As witnessed in the structure in (40), Campos and Stavrou consider the defi-nite article and the noun in polydefidefi-nites to be a complex head, under D0. Their motivation for analysing D+N as a complex head is twofold: a) in Balkan lan-guages N+D seems to form a word, and b) nothing seems to be able to intervene between D+N in Greek polydefinites. Nevertheless, the judgements that Campos and Stavrou give in favour of the second motivation are not shared across Greek speakers. The first piece of evidence they provide comes from the fact that the noun in polydefinites cannot be preceded by a numeral or a quantifier:
(42) *ta
‘the three/many black cars’
This judgement is shared across native speakers. The controversy, however, oc-curs when the same claim is extended to include adjectives. Campos and Stavrou state that adjectives, just like numerals, are forbidden from appearing between the definite article and the noun, but most speakers find the phrase in (43) ac-ceptable.13
‘the nice black car
Given the grammaticality of the above phrase, it is reasonable to conclude that the definite article and the noun do not form a complex head, even if there is variation among speakers. Another argument against the complex head D+N is the fact that it is possible to find a genitive with the noun in a polydefinite. Under this analysis we expect that the sequence D+N+genitive would be ungrammatical as it is not clear where the genitive would be merged. However, as is evident from the example below, this sequence is grammatical:
b. to
13Another example of this was given in (6b) and (6c), repeated below:
(6) b. to
(44) to
So far we have only seen how polydefinites with postnominal adjectives are de-rived. In order to derive polydefinites with prenominal adjectives, Campos and Stavrou propose that A0 moves to Pred0, picks up the definite article, and then there is subsequent head movement of Pred+A to a Focus head above the DP.
Building on this point, the authors suggest that the reason polydefinites with prenominal adjectives are unavailable to Aromanian or Ancient Greek might be because a Focus position is not available to these languages.
Be that as it may, implementing movement of the adjective to a Focus position has unwelcome results. It predicts that prenominal adjectives are necessarily focussed, something that is not borne out. Consider the examples in (45). In all of them we see that there is a polydefinite with a prenominal adjective, that is
‘the chubby the alien’. In the first example, the speaker continues the sentence by adding that aside from kissing the chubby alien, Sarah also kissed the thin alien.
In the second example, however, where the only difference from the first example is that ‘chubby’ is intonationally marked, the same continuation of the sentence is infelicitous. An appropriate continuation in this case must be contrastive, as in (45c). Taking into account the fact that focus in Greek is marked by stress, it becomes obvious that if the prenominal adjective in the polydefinite in (45a) was focussed as it is in (45b), then it would also require a contrastive continuation.
(45) a. i
‘Sarah kissed the chubby alien, as well as the thin one.’
b. i
‘Sarah kissed the chubby alien, #as well as the thin one.’
c. i
‘Sarah kissed the chubby alien, not the thin one.’
The proposal that the prenominal order is derived via head movement is also
problematic for this analysis. Campos and Stavrou claim that only postnominal adjectives in polydefinites can be modified or take a complement as the exam-ples in (46) and (47) indicate. Drawing from these data, they conclude that prenominal adjectives in polydefinites must be heads.
(46) a. *to
‘the very nice/nicer book’
(Campos and Stavrou 2004:140, (7) & fn. 2) (47) a. *i
‘the mother proud of her son’
(Campos and Stavrou 2004:140, (8) & fn. 3) However, the judgements in (46a) and (47a) are again not shared by other Greek speakers, a point that is also brought up by Panagiotidis and Marinis (2011). For many speakers, the prenominal adjective can be modified and it can also appear with a complement, which suggests that the prenominal adjective is not a head, but an AP.
Finally, Campos and Stavrou’s analysis does not make any predictions about where numerals or nonpredicative adjectives are merged in the instances where speakers accept them in prenominal polydefinites. A way forward would be to propose that, at least numerals, which are predicative, are the complement of Pred0 when the context permits this. If this was the case, however, we would expect that numerals would be free to appear either prenominally or postnomi-nally in polydefinites. Yet, as was already mentioned, numerals are only allowed prenominally when they appear in a polydefinite.
Even if there is a way to derive numerals in polydefinites, the presence of non-predicative adjectives still remains a puzzle, as the Campos and Stavrou analysis only allows predicative modifiers to enter PredP.