Linguistics 550
8. Grammatical function changing movement

Object idiom chunks

In addition to the subject idiom chunks discussed in Notes 7, English also has idiom chunks that are restricted to the complement position of particular verbs. As in the case of subject idiom chunks, we say that the verbs in question license (= are the condition for the availability of) these object idiom chunks. Some examples are given in (1) (Radford 1988:422).

(1) a.   The government keeps tabs on his operations.
b. They will pay heed to her proposal.
c. The Prime Minister paid homage to the dead.
d. The Chief of Staff pays lip service to the President.
e. Let's take advantage of the situation.
f. She took note of what I said.

The restriction to the object position of a particular licensing verb is shown by the contrast between the nearly synonymous attention and heed in (2) (Radford 1988:423).

(2) a.   He's always trying to attract my attention/*heed.
b. You can't expect to have my attention/*heed all the time.
c. He's a child who needs a lot of attention/*heed.
d. I try to give him all the attention/*heed he wants.

For completeness, (3) shows that object idiom chunks are generally also impossible in subject position (not surprisingly; after all, they're called object idiom chunks).

(3) a. Some attention/*heed from time to time would make them feel more important.
b. Attention/*heed is an important precondition for learning.

However, despite the distributional restrictions just illustrated, there is one environment in which object idiom chunks are able to occur in subject position (cf. Radford 1988:423). This environment is the passive–specifically, the passive of the licensing verb.

(4) a.   Close tabs were kept on his operations.
b. Little heed was paid to her proposal.
b. Due homage was paid to the dead.
c. Little advantage was taken of the situation.
d. Little note was taken of what I said.
(5) a. * My heed was attracted immediately.
b. * Your close heed is required.

A movement analysis of the passive

In Notes 7, it was argued that special subjects are always the clausemates of their licensing verbs, even when they end up functioning as the subjects of a higher clause. In order to implement this view, we introduced subject raising, which allows the original clausemate relationship between the special subject and its licenser to be maintained via the special subject's trace in the embedded clause. Subject raising allows the link between the special subject and its licenser to be stretched, as it were, with the trace indicating the position where the special subject is licensed.

Notice now that object idiom chunks have a similar close relationship with their licensers, which persists even when passivization turns the object idiom chunk into a subject. Not surprisingly, therefore, it has been proposed that the passive involves movement. However, in contrast to the instances of movement discussed so far (subject movement and subject raising), where the grammatical function of the moved noun phrase remains unchanged, passive movement is associated with a change in a noun phrase's grammatical function from object to subject.

Note: Grammatical functions are the set of relations that complements can have with a verb. In English, they include subject, direct object, indirect object and, according to some syntacticians, prepositional object (= PP complement).

The movement analysis of the passive is based on the premise that active and passive verbs are associated with identical treelets.

(6) a.       b.  

However, active and passive verbs do differ in two ways. First, the agent of a passive verb (= the doer of the action) is (optionally) expressed by an adjunct by-phrase rather than by a subject. This means that Spec(VP) is occupied by a noun phrase in the treelet of an active verb, but remains empty in the treelet of its passive counterpart.

Note: The agent by-phrase is omitted in the structure in (7b). If it were present, it would adjoin at V'. For simplicity, the internal structure of DPs and PPs is omitted as well.

(7) a.  
b.

Second, passive verbs cannot assign case to their complements. The evidence for this is that even if expletive it occupies subject position in (7b) so that the sentence has a subject and I has something to assign nominative case to, the resulting sentence is ungrammatical.

(8)   * It was paid little heed to her proposal.

Since the noun phrase in object position cannot be assigned case in its original position, it must move to a position where it can, and so it moves to Spec(IP). The resulting structure is shown in (9).

(9)    

Here's a question to ponder: does the object move from object position to Spec(IP) directly or via Spec(VP)?

Transitivity alternations

Data

Consider the data in (10) (from Radford 1988:446).

(10) a. i.   John broke the door.
ii.   The door broke.
b. i.   John might drown the kittens.
ii.   The kittens might drown.
c. i.   John rolled the ball down the hill.
ii.   The ball rolled down the hill.
d. i.   The artillery will sink the ship.
ii.   The ship will sink.

From a superficial point of view, the difference between the verbs break, drown, roll and sink in the (i) and (ii) examples is that they are transitive in the (i) examples and intransitive in the (ii) examples. But this is not all that sets the two examples apart. To see this, compare (10) with (11).

(11) a. i.   John is eating dinner.
ii.   John is eating.
b. i.   John is reading the paper.
ii.   John is reading.

In (11), the transitivity alternation simply deletes the argument affected by the action, without any effect on the grammatical function of the verb's agent, which is expressed as a subject in the (i) and (ii) examples alike. In (10), on the other hand, the transitivity alternation changes the the affected argument's grammatical function. In the transitive sentences in (i), this argument is expressed as the object, but in the intransitive sentences in (ii), it is expressed as the subject. As a result, the sentences in (10) are reminiscent of passive sentences (even though they contain no passive auxiliary or passive morphology on the verb).

Analysis

Given the movement analysis of the passive, it is not surprising that an analogous analysis has been suggested for transitivity alternations of the type in (10). Again, the fundamental idea is that the verb's treelet is identical regardless of whether it surfaces as a transitive or an intransitive verb. And again, the passive-like intransitives differ from the transitives in two respects. As with the passive, these differences concern the way the agent of the action is expressed and whether the verb assigns oblique case. In the transitive variant, as in the active member of an active-passive pair, the agent is expressed in the Spec(VP) position, and the verb assigns oblique case under government. In the intransitive variant, on the other hand, the agent is not expressed (it cannot even be expressed in a by-phrase, unlike the passive), and the verb is defective in not being able to assign case. Just as in the passive, the result is that the object is forced to move to Spec(IP) to be assigned case.

According to the movement analysis just presented, there is a major syntactic difference between the intransitive verbs in (10) and those in (11). The apparently intransitive verbs in (10) as actually transitive verbs in disguise, since they are associated with an object position. The intransitive verbs in (11), on the other hand, are true intransitives without a syntactic complement.

An alternative approach to grammatical function changing

According to the movement analysis (or syntactic analysis) of the passive presented earlier, the subcategorization frames of active verbs and their passive counterparts (= the part of the treelet dominated by V') are identical.
Note: An analogous argument could be made for the transitivity alternations of the type in (10). For simplicity, the discussion focuses on the passive.
However, in contrast to active verb forms, their passive counterparts are incapable of (i) assigning oblique case to their objects and (ii) assigning the semantic role of agent to the subject. As a result, Spec(VP) is empty in the passive. Furthermore, the subject of a passive, though assumed to originate in object position, is forced to move to Spec(IP) to receive case.

According to an alternative view (the lexical analysis), the subcategorization frames of active and passive verbs are not in fact identical. Rather, the lexical analysis holds that passive is a process that modifies an active verb's treelet by eliminating the subject as an argument and turning the erstwhile object into a subject, thereby changing the active verb's subcategorization frame from transitive to intransitive. Since the process affects the shape of the passive verb's treelet, it must take place not in the syntax, but in the lexicon–hence, the name of the analysis.

The difference between the syntactic and the lexical analysis of the passive is expressed in (13). (13a) is the treelet for an active verb form under both analyses. (13b) and (13c) are the treelets for the corresponding passive verb form under the syntactic and the lexical analyses of the passive, respectively.

(13) a.       b.       c.  

There has been a long-standing debate in generative grammar as to which view of the passive is correct. The syntactic view is associated with Chomsky, whereas the lexical view is associated with Joan Bresnan's Lexical-Functional Grammar (LFG) and other West Coast schools of syntax. You will be asked to evaluate potential evidence concerning this debate in Assignment 8, Exercise 3.