In lieu of an abstract, here is a brief excerpt of the content:

  • Convergent Evidence for Rolling Up Catalan Adjectives
  • Andrew Nevins

1 Prenominal/Postnominal Asymmetries

As pointed out by Bonet, Lloret, and Mascaró (to appear), in the Northeastern Central (NC) dialect of Catalan, adjectives show an agreement asymmetry: prenominally, plural agreement can be omitted in the context C__C (between two flanking consonants), while postnominally, it is obligatory.

  1. 1. dos bon(s)  marcians
    two good.PL Martians

  2. 2. dos marcians bons
    two Martians good.PL

These facts are related to general postnominal/prenominal asymmetries in agreement, such as the "lazy concord" phenomenon of Friulian and Ladin (Haiman and Beninca` 1992, Rasom 2008) and those reported for count /mass agreement in Spanish dialects (Hualde 1992). NC Catalan has no general process of s-deletion before nouns, as evidenced by (3), where the /s/ is part of the adjectival stem.1

  1. 3. un fals  conseller
    a  false counselor [End Page 339]

Finally, as adjectives are prosodically phrased together with the noun whether prenominal or postnominal in Romance (Dehe´ and Samek-Lodovici 2009), the explanation for this pattern must be found in a difference in syntactic structure between (1) and (2).

2 Three Accounts of Postnominal Adjective Order

Many Catalan adjectives can either follow or precede the head noun. In this squib, I present evidence from NC Catalan based on both scopal interpretation and the inflectional agreement on adjectives in order to constrain the space of analytic possibilities for postnominal adjective word order. Let us begin with the structure in (4), where A1 and A2 stand for adjectives that are heads, which are ordered A1 before A2 before N in their prenominal variant. I omit possible additional functional structure for the purposes of simplicity of exposition, and I assume that A1 selects A2.

  1. 4. Prenominal adjectives

With (4) as a fixed point in the analysis, let us turn to the three main accounts for postnominal adjectives in the literature: one involving movement of the noun alone, one involving no movement at all, and one involving movement of both the noun and an adjective phrase.

The head-movement-of-N view of postnominal adjectives holds that N moves to a high head position above both A1 and A2 (see (5)). This view can potentially account for agreement asymmetries between prenominal and postnominal adjectives. However, it cannot account for the scope interpretation differences between prenominal and post-nominal position, since the adjectives are in the same hierarchical relation with respect to each other, whether the N moves or not.

The adjunction-of-Adj view holds that adjectives can be right-adjoined to NP to allow for the postnominal word order. It can thus account for the scope interpretation of adjectives according to the order of adjunction (where the lower postnominal adjective precedes the higher one). However, an adjunction account would not be able to account for agreement asymmetries between prenominal and post-nominal orders, since adjunction (direction) could not affect agreement realization. This approach, as shown in (6), is representative of a family of "symmetric" approaches to postnominal adjectives (see, e.g., Abels and Neeleman, to appear). [End Page 340]

  1. 5. Postnominal adjectives if derived by head movement of N

  2. 6. Postnominal adjectives if derived by adjunction

The roll-up-of-N view (see Cinque 2005) holds that N moves through Spec,A2 and pied-pipes A2 to Spec,A1 (see (7)). Scope interpretation is determined by c-command in the base order. Agreement asymmetries are determined by the principle that Spec-head agreement is obligatory, while downward Agree agreement is optional. (See Franck et al. 2006 for an explicit statement that the morphological realization of Agree is weaker than that of Spec-head agreement; see also Samek-Lodovici 2002.)

  1. 7. Postnominal adjectives if derived by roll-up

    1. a. Step 1

      [End Page 341]

    2. b. Step 2

Given the standard view on obligatory reconstruction of rolled-up phrases, the roll-up view predicts a scope asymmetry: prenominally, A1 before A2 will show a scope of A1 over A2, but postnominally, A1 before A2 will show a scope of A2 over A1. Similarly, given the standard view that Spec-head agreement is stronger than long-distance agreement, the roll-up view predicts an agreement asymmetry: pre-nominally, adjectives may...

pdf

Share