20.07.2013 Views

Notes on computational linguistics.pdf - UCLA Department of ...

Notes on computational linguistics.pdf - UCLA Department of ...

Notes on computational linguistics.pdf - UCLA Department of ...

SHOW MORE
SHOW LESS

Create successful ePaper yourself

Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.

Stabler - Lx 185/209 2003<br />

m. The idea which is elusive to my friends will suffice.<br />

n. * The frightened <strong>of</strong> my mother friends will not visit.<br />

o. The friends who are frightened <strong>of</strong> my mother will not visit.<br />

6.<br />

The given rules also do not fit well with the semantic idea that modifiers take the modifiees as argument<br />

(Keenan, 1979; Keenan and Faltz, 1985) – which is usually indicated in the syntax by a head with its arguments<br />

in complement positi<strong>on</strong>. Abney (1987) proposes just this alternative idea about prenominal adjective<br />

modifiers: they are functi<strong>on</strong>al categories inside DP that obligatorily select nominal NP or AP complements.<br />

Since NP or another AP is the complement in these c<strong>on</strong>structi<strong>on</strong>s, they cannot take another complement<br />

as well, as we see in examples l-o. Modify the proposed phrase structure rules al<strong>on</strong>g these lines. Turn in<br />

a listing <strong>of</strong> the resulting grammar and a sessi<strong>on</strong> log showing this grammar being used with td.pl, witha<br />

brief commentary <strong>on</strong> what you have d<strong>on</strong>e and how it works in the implementati<strong>on</strong>.<br />

There are various reas<strong>on</strong>s that our measure <strong>of</strong> space complexity is not really a very good measure <strong>of</strong> the<br />

computati<strong>on</strong>al resources required for top-down recogniti<strong>on</strong>. Describe some <strong>of</strong> them.<br />

7. a. Write a propositi<strong>on</strong>al c<strong>on</strong>text free grammar for tdp.pl that generates the following tree from the input<br />

axioms [john,praises,john], and test it with either pp_tree or wish_tree (or both) before submitting<br />

the grammar. (Get exactly this tree.)<br />

DP<br />

john<br />

V<br />

IP<br />

praises<br />

b. Notice that in the previous tree, while the arcs descending from IP, VP and V’ are c<strong>on</strong>necting these c<strong>on</strong>stituents<br />

to their parts, the arcs descending from the “pre-lexical” categories DP and V are c<strong>on</strong>necting<br />

these c<strong>on</strong>stituents to their ph<strong>on</strong>etic/orthographic forms. It is perhaps c<strong>on</strong>fusing to use arcs for these<br />

two very different things, and so it is sometimes proposed that something like the following would be<br />

better:<br />

IP<br />

VP<br />

V’<br />

DP:john VP<br />

V’<br />

DP<br />

john<br />

V:praises DP:john<br />

Is there a simple modificati<strong>on</strong> <strong>of</strong> your grammar <strong>of</strong> the previous questi<strong>on</strong> that would produce this tree<br />

from the input axioms [john,praises,john]? If so, provide it. If not, briefly explain.<br />

56

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!