Ling-001-601, Fall 2007

Homework 5, due October 11 at the beginning of class.

Note: You have the opportunity to earn a total of 11 points on this homework (i.e. one extra-credit point).

Part 1: Structural ambiguity (5 points) Before you start this exercise, draw two trees for the sentence I fed the cats with a shovel from the bottom of p. 6 of your Week 5 handout, and write an unambiguous paraphrase for each tree. Then check your answer against mine to make sure you were on the right track. If you have questions, review the Santorini and Kroch reading and/or email me.

Here are some more examples of structural ambiguity:

1. I found a doctor wearing my pajamas.
2. Marjorie wants to speak badly.
3. big circles and squares
4. ice cream truck driver
5. (your example)
6. (your example)

The first part of the assignment is to collect two more examples of structurally ambiguous compounds, phrases, or sentences for (5) and (6) above. They can be written or spoken examples.

Then do the following:

• For each of the 6 examples above, identify at least two readings (there may be more) and write an unambiguous paraphrase of each. You can supplement your paraphrases with pictures or diagrams to help make sure they're unambiguous. Number the paraphrases (a) and (b).
• Select any 4 of the 6 examples and draw a tree for each of the two readings you identified. Number the trees (a) and (b) so that I can tell you know which tree goes with which paraphrase.
  • For full sentences like (1) and (2), you can use node labels like NP, PP, VP, etc. if you wish, but you don't have to. If you decide to use these labels, you can use the label 'PartP' (participle phrase) for the constituent headed by the present-participle wearing in (1), and 'IP' (inflectional phrase) for the constituent headed by the to-infinitive in (2).
  • For compounds like (4), you can either show the internal structure of each word (e.g. drive + er) or treat each word within the compound as a terminal/leaf - either way is fine with me.
  • In general, your trees should be binary-branching. When two constituents are linked with and), though, ternary-branching nodes may be used:
    
    You can use this as a model for (3).

Part 2: Constituency (4 points). Do Exercise 2.1 (#1-2 only), and Exercise 2.2 (#1-2) from Chapter 2 of this week's Santorini/Kroch reading.

Part 3: Evidence for structure-(in)dependence (2 points). We talked briefly in class about a phenomenon called contraction in English, where two adjacent words undergo some irregular phonological changes so that they end up looking 'fused.' For this exercise, we'll use you'll (derived from you + will) as our example.

You'll actually has two pronunciations: (i) it can rhyme with full, pull, bull, or (ii) it can retain the /u/ vowel so that it rhymes with mule, school, tool. I'll refer to the (i) pronunciation as full contraction. This is the type we're going to be focusing on here.

The basic question you're trying to answer here is: Is full contraction structure-dependent or structure-independent? In other words, which of the following hypotheses is more likely to be correct?

• Structure-independent hypothesis: In order for full contraction to take place, the only requirement is that the two words in question - e.g. you and will - be adjacent to each other, with you immediately preceding will. The constituent structure of the sentence doesn't matter.
• Structure-dependent hypothesis: The two words do have to be adjacent, but string-adjacency isn't the only condition on contraction. The constituent structure of the sentence also plays a role. Sometimes contraction will be impossible even when the two words are string-adjacent, because they are not in the right structural relationship.

Now consider the following examples. For each example, I've written 'yes' if the two underlined words can undergo full contraction, and 'no' if full contraction is ungrammatical.

	Full contraction okay?
a. You will never walk alone.	yes
b. Moira says you will never walk alone.	yes
c. Brendan and you will never walk alone.	no
d. The guy you will meet tomorrow hasn't arrived yet.	yes
e. The guy who likes you will leave soon.	no
f. He won't sit next to you, will he?	no

• Why is full contraction okay in (a), (b) and (d), but unavailable in (c), (e) and (f)? Write a short paragraph describing what might be going on in (c), (d) and (f) that makes full contraction impossible. Which of the two hypotheses above do these facts lend support to? You may wish to supplement your answer with partial trees or bracketings to make your structural descriptions clear.
• Now go through the same exercise with wanna. Provide examples of sentences that demonstrate whether wanna-contraction is possible whenever want and to are adjacent, or whether it is structure-dependent. If you're not a native speaker of English, you can ask me or another speaker to provide judgments for your example sentences.