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 />

16.4 Scope inversi<strong>on</strong><br />

That idea has been developed to apply to the data above in recent work by Beghelli and Stowell (1996) and<br />

Szabolcsi (1996).<br />

This strategy may seem odd, but Szabolcsi (1996) notes that we may be slightly reassured by the observati<strong>on</strong><br />

that in some languages, we seem to find overt counterparts <strong>of</strong> the “covert movements” we are proposing in<br />

English. For example, Hungarian exhibits scope ambiguities, but there are certain c<strong>on</strong>structi<strong>on</strong>s with “fr<strong>on</strong>ted”<br />

c<strong>on</strong>stituents that are scopally unambiguous:<br />

(5) a. Sok ember mindenkit felhívott<br />

many man every<strong>on</strong>e-acc up-called<br />

‘Many men ph<strong>on</strong>ed every<strong>on</strong>e’<br />

where many men < every<strong>on</strong>e<br />

b. Mindenkit sok ember felhívott<br />

every<strong>on</strong>e-acc many man up-called<br />

‘Many men ph<strong>on</strong>ed every<strong>on</strong>e’<br />

where every<strong>on</strong>e < many men<br />

(6) a. Hatnál több ember hívott fel mindenkit<br />

six-than more man called up every<strong>on</strong>e-acc<br />

‘More than six men ph<strong>on</strong>ed every<strong>on</strong>e’<br />

where more than six men < every<strong>on</strong>e<br />

b. Mindenkit hatnál több ember hívott fel<br />

every<strong>on</strong>e-acc six-than more man called up<br />

‘More than six men ph<strong>on</strong>ed every<strong>on</strong>e’<br />

where every<strong>on</strong>e < more than six men<br />

Certain other languages have scopally unambiguous fr<strong>on</strong>ted elements like this, such as KiLega and Palestinian<br />

Arabic. Scrambling in Hindi and some Germanic languages seems to depend <strong>on</strong> the “specificity” <strong>of</strong> the scrambled<br />

element, giving it a “wide scope.”<br />

To account for these and many other similar observati<strong>on</strong>s, Beghelli and Stowell (1996) and Szabolcsi (1996)<br />

propose that determiner phrases occupy different positi<strong>on</strong>s in structure according to (inter alia) the type <strong>of</strong><br />

quantifiers they c<strong>on</strong>tain. Furthermore, following the recent traditi<strong>on</strong> in transformati<strong>on</strong>al grammar, they assume<br />

that every language has structures with special positi<strong>on</strong>s for topicalized and focused elements, though<br />

languages will differ according to whether the elements in these positi<strong>on</strong>s are pr<strong>on</strong>ounced there or not.<br />

We can implement this kind <strong>of</strong> proposal quite easily. First, let’s distinguish five categories <strong>of</strong> determiners:<br />

wh-QPs (which, what)<br />

neg(ative)-QPs (no, nobody)<br />

dist(ributive)-QPs (every, each)<br />

count-QPs (few, fewer than five, six,…)<br />

group-QPs (opti<strong>on</strong>ally, the, some, a, <strong>on</strong>e, three, …)<br />

We assume that these categories can cause just certain kinds <strong>of</strong> quantified phrases to move.<br />

Both Beghelli and Stowell (1996) and Szabolcsi (1996) propose that the clause structure be elaborated with<br />

new functi<strong>on</strong>al heads: not because those heads are ever overt, but just in order to provide specifier positi<strong>on</strong>s<br />

for the various kinds <strong>of</strong> quantifier phrases. In our framework, multiple specifiers are allowed and so we do not<br />

need the extra heads. Furthermore, introducing extra heads between T and v would disrupt the affix-hopping<br />

analysis <strong>of</strong> English proposed in §10.2.1, since affix hopping is not recursive in the way that verb raising is: <strong>on</strong>e<br />

affix hop cannot feed another. Also, Beghelli and Stowell (1996, p81) propose that Dist can license any number<br />

252

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

Saved successfully!

Ooh no, something went wrong!