| (1) | Ordinary pronouns must be free. |
The condition in (1) is satisfied in both sentences in (2) (= (4) in Chapter 1). Yet the pronoun and the full noun phrase can be coindexed in (2a), but not in (2b).
Note: Some speakers find coindexing difficult in (2a). Even these speakers usually detect a difference between (2a) and (2b), however.
| (2) | a. | [ Her ]1 brother helped [ Zelda ]1,2. | |
| b. | [ She ]1 helped [ Zelda ]2,*1. |
This suggests that hierarchical structure plays a role not only in the interpretation of ordinary pronouns, but in that of full (= nonprononimal) noun phrases as well.
A. Using the grammar tool in English binding 1, build trees for the sentences in (2). (Disregard the indices when building your trees, as the grammar tool doesn't allow you to include them.)
B. Using the terminology introduced in Chapter 1, formulate a
structural condition concerning full noun phrases that accounts for the
facts in (2).
Exercise 2
A. Using the grammar tool in English
binding 1, build trees for the sentences in (3) and (4). (Since the
grammar tool doesn't allow you to include indices, a single tree is
sufficient for the two subcases of (4a).)
| (3) | a. | [ Zelda ]1 helped [ herself ]1. | ||
| b. | * | [ Herself ]1 helped [ Zelda ]1. | ||
| (4) | a. | i. | * | [ Zelda's ]1 sister helped [ herself ]1. |
| ii. | [ Zelda's sister ]1 helped [ herself ]1. | |||
| b. | * | [ Herself's ]1 sister helped [ Zelda ]1. |
B. Using the terminology introduced in Chapter 1, formulate a structural condition concerning reflexive pronouns that accounts for the data in (3) and (4).
C. Why is (3b) ungrammatical? Your answer should be brief, but
exhaustive.
The sentences in (5) are Malagasy, a VOS language spoken in Madagascar
(Reinhart 1983:47).
Exercise 3
The languages of the world differ in their basic constituent order. Most
languages (about 60%) are SOV (that is, subjects precedes objects, and both
precede the verb). There are also many languages (about 40%) that are SVO,
like English. Very rarely, languages are VOS. However, despite their
rarity, VOS languages are very important for linguistic theory. Because
their linear order is so different from that of the more common languages,
they provide an excellent opportunity to test putative universals
concerning the role of hierarchical structure.
| (5) | a. | namono | azy1 | ny | anadahin | d | Rakoto1. | |
| killed | him | the | sister | of | Rakoto | |||
| 'Rakoto's1 sister killed him1.' | ||||||||
| b. | * | namono | ny | anadahin | d | Rakoto1 | izy1. | |
| killed | the | sister | of | Rakoto | he | |||
| Intended interpretation: 'Rakoto1 killed his1 sister.' | ||||||||
A. Using the grammar tool in Malagasy binding, build trees for the sentences in (5).
B. Explain the grammaticality contrast in (5). Assume that azy and izy are ordinary pronouns in Malagasy.
C. Using the grammar tool in Malagasy binding, build trees for the sentences in (6).
| (6) | a. | namono | azy | Rakoto. | |
| killed | him | Rakoto | |||
| b. | namono | Rakoto | izy. | ||
| killed | Rakoto | he | |||
D. Can either azy or izy be coindexed with Rakoto
in (6)? That is, can either of the sentences in (6) be interpreted as
'Rakoto killed himself'? Explain.
Exercise 4
Show that each of the statements in (7) is false. Feel free to use
nonlinguistic trees, where the terminal (= bottom) nodes of the trees
are alphabetic symbols rather than words.
| (7) | a. | If A c-commands B, then A precedes B. | |
| b. | If A precedes B, then A c-commands B. | ||
| c. | If A c-commands B, then B c-commands A. |
Given the definition of c-command in (8) (= (11) in Chapter 1), a node c-commands itself.
| (8) | c-command: A c-commands B iff (= if and only if) the first branching node that dominates A also dominates B. |
Given the data discussed so far, explain why this is problematic. How can the problem be eliminated?