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

  • The Nonreality of Doubly Filled Comps
  • Mark Baltin

Many languages have what look like “doubly filled Comps,” in which a wh-phrase is said to occur in the specifier position of CP (Spec,CP) and an invariant “complementizer” is posited as the head of CP. An example from Belfast English is given in (1).1

  1. 1. They discussed a certain model, but they didn’t know which model that they discussed.

Other so-called doubly filled Comp languages include Norwegian and Bavarian German (Lobeck 1995), as well as the Italian Bellinzonese dialect (Andrea Cattaneo, pers. comm.).

Without exception, such languages disallow the overt complementizer in sluicing, as noted by Lobeck (1995) and discussed by Merchant (2001). This phenomenon is exemplified in (2).

  1. 2. They discussed a certain model, but they didn’t know which model (*that) ___.

Additionally, even in non–doubly filled Comp languages, so-called T-to-C movement is incompatible with sluicing, as noted by Lasnik (1999).

  1. 3. Speaker A: He visited somebody.

    Speaker B: Oh, really. {Who did he visit ___?}

        {Who (*did) ___?}

Since the moved T in, for example, English question formation is thought to occupy the same position as the invariant particle in so-called doubly filled Comp languages (i.e., C), a unified account of these two phenomena seems appropriate. They both pose a problem for the view that (a) clauses are bifurcated into a C that takes a CP complement, and (b) sluicing is TP-deletion, as assumed by Lasnik (1999), among others. This pair of assumptions would lead us to predict [End Page 331] that the overt C would survive sluicing, and would furthermore render mysterious the disappearance of the overt C; C is, after all, by hypothesis outside the TP that is deleted.

Both of these problems are solved if we adopt the view of the left periphery advanced by Rizzi (1997). He proposes that C is decomposed into four separate projections, given in (4).

  1. 4. [Force [Topic [Focus [Finite . . . ]

Force and Topic will not concern me here, but Foc(us) and Fin(ite) will play a role. Rizzi (1999) posits Spec,FocP as the landing site for wh-phrases.2 Henry (1995) provides some evidence that Fin, in Rizzi’s system, may be the position for the elements that and for in English, given the acceptability of (5), courtesy of Alison Henry (pers. comm.).

  1. 5. They wanted to do something, but they weren’t sure what for to do.

Putting (5) together with (1), it seems clear that English that and for introduce finite and nonfinite complements, respectively. Taking sluicing to be deletion of the complement of Foc, a structure that has the characteristics in (5) would cause any element in Fin to be deleted, be it a base-generated particle such as that and for or their cognates, or a T that fronts to Fin (as we might more definitively specify T-to-C movement).

Additional evidence for sluicing as deletion of the complement of Foc is found in certain African languages of the Kwa family such as Gungbe, analyzed by Aboh (2004). Aboh provides a detailed analysis showing that certain clause-initial particles in this language directly instantiate Rizzi’s projections. In particular, he takes the particle w as marking Focus, and notes that it obligatorily follows questioned wh-phrases.

  1. 6. Éti *(w) Sná xìá ti?

    what foc Sena read-perf

    ‘What did Sena read?’

    (Aboh 2004:279, (7.82a))

Aboh’s analysis of w as instantiating Foc, together with my analysis of sluicing as deletion of the complement of Foc, leads to the prediction that w must be retained in Gungbe sluicing. (7) would seem to point [End Page 332] in this direction,3 under the assumption that w is indeed the Foc marker.4

  1. 7.

  1. a. Ùn sè ɗ Súrù x mótò àmn mă nyn mótò 1sg hear that Suru buy car but 1sg.neg know car ɗě w.

    det.rel foc

    ‘I heard that Suru bought a car but I don’t know which car.’

  2. b. Súrù nù àmàsìn ɗé àmn íyà étn má nyn Suru drink medicine det but mother his neg know àmàs...

pdf

Share