Chapter5 Verb Phrases

In this chapter we are going to take a detailed look at the structure of the phrase that is in some ways the core of the clause. In this phrase the basic argument structure of the clause is formed which is made up of the verb, acting as the predicate, its arguments and adjoined modifiers. Thus, it is within the VP that Θ-role assignment takes place. We will also see that it is within the VP that other aspects of semantic interpretation are represented, such as event structure and aspect. Other aspects of clausal interpretation, such as tense, modality and force are introduced in structures built on top of the VP and we will deal with these in the following chapters.

Chapter5 Verb Phrases

The notions of event structure and aspect are new and we will spend a little time introducing them in the next section.

5.1 Event Structure and Aspect

5.2.1 Unaccusative verbs

As far as the event structure is concerned, unaccusative verbs have a very simple interpretation involving a simple state or relationship between the theme argument and the location. To see this, compare the unaccusative and agentive use of sit again:

5.2.1 Unaccusative verbs

The simple event structure corresponds with the simple VP structure of the unaccusative. We will see that more complex event structures lead to more complex VPs.

5.2.2 Light verbs

Consider the event structure expressed by this verbal complex. It is fairly clear that there is one (complex) event described by the light verb and thematic verb complex: there is just one clause here with one subject. The event, however, is made up of two sub-events: she does something and this causes the vase to break:

5.2.2 Light verbs

Note that the complex event structure is mirrored by the complex VP structure. There are two sub-events and two parts to the VP, an upper vP and a lower VP. Moreover, the vP corresponds to the first sub-event and the causative connection between the sub-events. The VP corresponds to the sub-event that results from the first. This indicates that there is a connection between event structure and syntactic structure, specifically the more complex the event structure, the more complex the syntactic structure used to represent it.

5.2.3 Ergative verbs

Moreover, the structure of the VP in (45) is simple, in comparison to that of verbal complexes involving light verbs, as in (32), for example. Yet the event structure expressed here is not simple. In the butler opened the door, there is an event involving the butler doing something and an event involving the door being open and clearly the first event causes the second. Hence the event structure is:

5.2.3 Ergative verbs

If (45) is the correct analysis, then there is a mismatch here between event structure and syntactic structure whereas in other cases we have seen there has been an isomorphism between the two.

5.2.3.1 Potential problems

How come these structures mean virtually the same thing, especially as, as we have seen, light verbs are not without meaning? Note that the subject of the ergative verb in (48b) is interpreted as the causer of the ball’s bouncing, which is exactly the same interpretation given to the subject of make, a causative verb. The event structure of both examples is also the same:

5.2.3.1 Potential problems

But while the syntactic structure of (48a) is isomorphic with the event structure in (49), if we analyse the sentence in (48b) as having a structure like (45) then the syntactic isomorphism with the event structure is completely lost.

5.2.3.2 Light verbs and ergatives

Under this analysis, the UTAH can be maintained as each argument sits in exactly the position it should according to our previous analyses: the theme is the specifier of the main verb and the agent is the specifier of the abstract light verb. Moreover the event structure is represented in an isomorphic way with there being two parts to the syntactic structure each of which relate to the relevant sub-event.

5.2.4.3 Agent and experiencer subjects

What about the event structure of a transitive construction? Above we argued for an isomorphism between the structure of the VP and the structure of the event it describes such that each part of the VP corresponds to a separate sub-event. If transitive verbs involve an agentive light verb, and hence there are two parts to the verbal complex, we should expect that the event described by a transitive verb should consist of two sub-events. But we have just seen that transitives are not causative: Harry hit Ron does not mean that Harry does something that causes Ron to get hit. However, there is not necessarily a direct relationship between what the subject does and the object getting hit. Consider the following:

5.2.4.3 Agent and experiencer subjects

We can see from these examples there is a sense in which there are two parts to a hitting event: somebody does something and somebody or something gets hit as a more or less direct result of this. There is obviously a very subtle difference between this interpretation and a causative one, which we will not attempt to describe here. The point is that the event structure of the transitive can be represented in a similar way to that of a causative:

5.2.4.3 Agent and experiencer subjects

A second observation that seems to support (87b) concerns the event structure of transitive verbs with experiencer subjects. Certainly there does not appear to be a causative relationship between what the subject is interpreted as doing and what happens to the object. (86c), for example, does not appear to be interpretable as Fred doing something that results in the dark being feared. Instead it seems that these sentences express a simple state of affairs with no sub-events:

5.2.4.3 Agent and experiencer subjects

If there is an isomorphism between event structure and VP structure and (90) is the correct analysis of the event structure involving an experiencer subject transitive verb, then (87b) appears to be the correct structure of the VP.

5.2.4.3 Agent and experiencer subjects

However, an obvious disadvantage of (87b) is that the theme is placed in the complement position which is counter to what we have previously discovered. If the theme goes in the specifier of the thematic verb, then there is no alternative than to include the experiencer in a higher position which would mean adding an abstract light verb. A further disadvantage of (87b) is that transitive verbs with experiencer subjects can be passivised. We have analysed passivisation as a process which removes the light verb responsible for the assignment of the Θ-role to the subject and the Case to the object, replacing it with the passive morpheme. If there is no light verb responsible for assigning the experiencer Θ-role, it is not at all clear how these verbs could undergo passivisation: what would the passive morpheme replace and why would the experiencer Θ-role and accusative Case go missing? Moreover, the passivisation of these verbs casts doubt on the assumption that they have a simple event structure. Passivisation of agentive verbs by getting rid of the agentive light verb turns a verb with a complex event structure into one with a simple one:

5.2.4.3 Agent and experiencer subjects

But if experiencer transitive verbs have a simple event structure and we remove the experiencer, what are we left with? Surely we cannot be left with half an event! This would argue that the event structure of experiencer transitives is similar to that of agentive transitives:

5.2.5 Intransitive verbs

The event structure of intransitives is also a little problematic as we predict it to be complex if intransitives involve light verbs, but a sentence like Sam smiled does not obviously express a complex event structure. However, it is not impossible to think of this as involving a situation in which Sam does something which results in a smile, which is made more plausible by comparison with the overt light verb construction:

5.2.5 Intransitive verbs

If intransitives are in fact formed from an underlying structure involving a ‘cognate object’ and a light verb as suggested above, then the parallel between (108a) and (b) is even stronger. We might therefore propose the following analysis of the event structure:

5.2.6 Multiple complement verbs

The event structures of these verbs however indicate that the analysis might be a little more complex than we have indicated in (112). For example, consider what is involved in ‘putting’. There is an agent who performs some action and there is a theme which undergoes a change of position and there is a location where the theme ends up. Thus the event structure seems to be:

5.2.8 Verbs with clausal complements

In this, the verb moves from light verb to light verb and the DP moves to the specifier of the first light verb to get Case from the higher one. A similar structure would have to be supposed for the PP arguments. However, this structure does not seem to reflect the event structure of such verbs, which seem to consist of just two events:

5.2.8 Verbs with clausal complements

If we want to maintain the UTAH we cannot just assume that the arguments start off in different positions, however. So we want to keep the basic structure of the VP as it is in (145). We need to simplify the light verb structure, getting rid of one of these to match with the event structure and finally we need to find a way of getting the PP in front of the clause that does not assume that it undergoes a similar movement to DPs. The structure is as follows:

5.2.9 Summary

To summarise this fairly long discussion of the structure of the VP, we can conclude that strict adherence to the UTAH and the assumption that there is isomorphism between event structure and VP structure leads us to sometime quite surprising but enlightening analyses of the central part of the clause. The VP itself seems a hive of activity, with verbs and arguments moving about from position to position which obviously complicates its description. However, the reasons for the movements themselves are fairly straightforward. The verb moves to support the abstract light verbs which have a bound morpheme status, DP arguments move to Case positions and clausal arguments move away from Case positions. Once such things are understood, some rather mysterious properties of VPs become demystified. Phenomena such as passivisation, double object alternation, the there construction, cognate objects and phrasal verb constructions are given a fairly satisfying analysis which we can take as encouraging for this approach.

Chapter7 Complementiser Phrases

In this chapter we continue to present the parts of the English clause, extending it further upwards. So far we have seen that the clause has a number of layers to it, relating to certain syntactic and semantic properties. The lower layer consists of the thematic part of the VP, including the Θ-assigning light verbs, which concern argument and event structure. Above this we have the non-thematic part of the verb phrase where the morphemes of aspect and tense are introduced. Finally above this we have the IP in which agreement and modal auxiliaries are situated. The IP also provides the surface Case position for the grammatical subject.