the week of october 29

definite descriptions

Following closely Heim's handbook article "Articles and Definiteness", we looked at the pro's and con's of a Russell-style versus a Frege-style analysis of definite descriptions. The handout we looked at is here. Adapted to our present framework, a Russell-style translation of the definite article "the" into ETL looks as in (1) or (2), and a Frege-style translation looks as in (3).

(1) λP λQ ∃x (Q(x) & ∀y (P(y) ↔ x = y) )

(2) λP λQ ∃x (Q(x) & P(x) & ∀y (P(y) → x = y))

(3) λP ιx P(x)

On proposals (1) and (2), which are logically equivalent, definite DPs like "the cat" are treated like quantifier phrases: they have denotations of type <<et>t>. On proposal (3), definite descriptions have denotations of type e, that is, they denote individuals. The main challenge of a Frege-style analysis are cases like (4) and (5):

(4) The monster of Loch Ness does not exist.

(5) If there is a monster of Loch Ness, the monster of Loch Ness is very shy.

The main challenges for a Russell-style analysis are cases like (6) below. (6) does not mean the same as (7). Most contemporary linguistic analyses of definite descriptions have adopted some version of a Fregean analysis, and have tried to overcome the challenges via a more sophisticated theory of presuppositions.

(6) I am so sorry I missed the party.

(7) I am so sorry that there was a party and I missed it.

preparing for relative clauses

We had two sessions that prepared us for our upcoming case study of Hindi relative clauses empirically, technically, and "ideologically". What would a good analysis of the Hindi correlative construction look like? What are we trying to do when we craft an analysis for a linguistic construction? On Wednesday, we looked at Dayal's discussion of structural differences between Hindi left-adjoined, right-adjoined, and DP-internal relative clauses, and tentatively came up with a structure for left-adjoined relative clauses that accounted for the binding and Condition C effects we observed. We then evoked both a Montagovian and a Chomskyan superego to guide our work towards an insightful analysis of relatives. Wednesday's handout with Dayal's data and the two superegos is here. On Friday, you did group work and labored over a compositional derivation of the denotation of an English sentence with two stacked relative clauses and a bound variable pronoun. The task was to translate the English phrase structure tree into an expression of ETL (our Extended Type Logic), using translational analogues of the Heim & Kratzer composition rules. Since these computations have to get into your fingers, you should submit individual lab reports this time. The worksheet is here.