1.3.2 Predicates and arguments

1.3.3 Grammatical aspects of meaning

The predicate here is the verb marry and the arguments are Mary and Martin. Let us call the part of meaning expressed by a predicate and its arguments the basic proposition. But what role do the other words, may and that, have in the sentence? Clearly, they have no role in the basic proposition, being neither predicates nor arguments. But they do carry some meaning. May is a modal auxiliary verb and in this sentence it either expresses that the event described by the predicate and its arguments (Mary marrying Martin) is a possibility or that permission has been given for it to take place:

1.3.4 The Thematic categories

Let us focus our attention first on the thematic ([–F]) categories, returning to the functional ([+F]) categories towards the end of the chapter. Much of our discussion so far has concerned verbs. This perhaps reflects their centrality in many sentences, being typical predicates. It also seems that notions such as predicate and argument are more obviously expressed in relation to verbs. So it is right to start our discussion of categories with them.

1.3.4.1 Verbs

In (54) the verb has the same past tense inflection in all permutations of the sentence, but in (55) there is a difference between the first example and all the others. This corresponds to the fact that the argument which precedes the verb in the first case is third person and singular and in all other cases this argument is either plural or first or second person (I or you). This argument is called the subject and we will discuss its nature and properties in the next chapter. For now we will simply use the term to refer to the argument in front of the verb without further discussion. The morphological phenomenon shown in (55) is known as agreement. We say that the verb agrees with certain features (number and person) of the subject (later on, we will see that it is the inflection that agrees with the subject and that this is independent of the verb). English does not demonstrate much in the way of agreement inflection. For the vast majority of verbs it is only in the present tense and with a third person singular argument that the verb has an agreement form. The exception is the verb to be, for which there are three present tense forms (first person singular, third person singular and the rest) and two past tense forms (first and third person singular and the rest):

1.3.4.1 Verbs

To some extent, this is connected to the properties of the verb as a predicate: laugh is a one-place predicate and its only argument, an agent, tends to precede it, while defeat is a two-place predicate and takes its agent to the left and the patient to the right. If we consider a three-place predicate, a pattern begins to emerge:

1.3.4.1 Verbs

In this case, one of the arguments appears to the left and the others are on the right. It seems that there is always one argument on the left and any other argument must follow the verb. We call the arguments which follow the verb the verb’s complements. It appears that there is a special relationship that holds between a verb and its complements. Consider the following:

1.3.4.1 Verbs

What we see by these examples is that different verbs are followed by different complements. The verb await must be followed by a nominal complement (i.e. one expressed with a noun: his trial) whereas the verb wait must be followed by a prepositional complement (expressed with a preposition: for his trial). Although there is often a connection between the thematic interpretation of the complement argument and its category, patients tend to be nominal and locations tend to be expressed by prepositional complements for example, it is not always possible to predict the category of the complement from its thematic role. In (68) for example, the two complements seem to be interpreted fairly similarly, but still they are expressed by complements of different categorial statuses. It follows that the category of the complement should be stated as a separate piece of information in a verb’s lexical entry:

1.3.4.1 Verbs

What is represented in these lexical entries is that the two verbs are both two-place predicates taking agent and goal (something that an action is directed towards) arguments, but that the goal of await must be nominal while that of wait must be prepositional. The part of the lexical entry that states the categorial status of the complement is known as a subcategorisation frame. Thus a lexical entry for a typical verb will consist of a theta-grid and a subcategorisation frame in addition to phonological and semantic information.

1.3.4.1 Verbs

These verbs are one-place predicates with their arguments on the left. Their lexical entries might be represented as follows:

1.3.4.2 Nouns

It is clear from the examples given above that nouns formed from verbs have arguments in the same way that those verbs do: the noun wait may express the relationship between someone who is waiting and what they are waiting for. The argument that comes to the left of the verb is typically expressed by the possessor of the derived noun (his and my and their in (87)). In other instances, however, the possessor simply names the one who possesses the noun. The difference is made clear in the two interpretations of the following:

1.3.4.2 Nouns

This can be interpreted either as something that Ken did (he constructed a kite) or something that he possesses (the kite is his). Obviously the possessive interpretation is only available for the case of the noun, the related verb cannot have a possessive argument:

1.3.4.2 Nouns

In this example, Ken can only be interpreted as agent. The question arises as to whether the possessor is another thematic argument which nouns can have, in addition to agents, patients, themes, goals, etc., or whether it is something of a different nature. There is reason to believe that the possessor is not the same kind of element as a thematic argument. One thing that differentiates possessors from other arguments is that the possessor may appear with almost any noun and does not appear to be determined by the noun’s meaning:

1.3.4.2 Nouns

In this example, Emily has to be interpreted as the one who experiences the embarrassment rather than someone who possesses it outside of their emotions. But this is a general semantic fact: some things can be possessed and other things cannot. The fact remains, however, that of those things that are able to be possessed, the relationship between them and the possessor is uniform and is not affected by the meaning of the noun. This is very different from other argument–predicate relationships:

1.3.4.2 Nouns

Another difference between the possessor and arguments is that the semantic relationship that possessors express is rather vague in relation to those expressed by arguments. Consider the following:

1.3.4.2 Nouns

The relationship between Shufflebotham and the sheep could be almost anything, ranging from ownership to something far more distant such as the sheep that Shufflebotham selected in a sheep of the year contest. Thematic arguments, on the other hand, have very definite interpretations: an agent is someone who consciously performs an action and cannot be interpreted as anything else.

1.3.4.2 Nouns

A final difference between possessors and arguments is that the possessor relationship is restricted to nouns whereas thematic relationships seem to be available to all thematic categories: we can find themes, experiencers, etc. for verbs, nouns or adjectives.

1.3.4.3 Adjectives

The arguments of the adjectives are included as optional to allow for their non-predicative use. When an adjective is used to modify a noun, it does not typically appear with its arguments:

1.3.5.1 Inflections

The feature bundle [+F, –N, +V] defines a ‘functional verb’. Such an element would have verbal properties, but no thematic content: it would not be specified for taking arguments in its lexical entry and hence would have no theta-grid. The most obvious thing that fits this bill is the class of auxiliary verbs:

1.3.5.1 Inflections

This distinction is traditionally called a Case distinction, which has to do with the forms that certain nominal arguments appear in. In English there are not many Case distinctions to be seen as it is only the pronouns which have Case forms, but in other languages there can be more such distinctions made (think of Hungarian János, Jánost, Jánosnak, etc.). The he form of the pronoun (similarly, she, I, we, they) is called the nominative case form, while the him form (her, me, us, them) is the accusative case form. Note, finite clauses must have nominative elements in the relevant position, whereas, if the position is filled at all in non-finite clauses, it must be by an accusative element:

1.3.6 Functionally underspecified categories

The nouns in (190) do not function as the main semantic element of the expression as do those in (191). Note that the expressions in (191) refer to a picture, a disposal and a door respectively, but the referents of the expressions in (190) are wine, tea and tourists respectively. One can pour a bottle of wine and drink a cup of tea, but what is poured and drunk is not the bottle or the cup but the wine and the tea. On the other hand, if one breaks a picture of the president or deplores the disposal of the evidence, it is not the president that gets broken nor the evidence that is deplored. The kind of nouns in (190) are called measure or group nouns and they differ from other nouns in terms of their relationship to their complements. The complements of the nouns in (191) are arguments of those nouns and as such stand in a thematic relationship to them. In other words, these nouns are thematic elements which have Θ-grids in their lexical entries. Measure nouns do not stand in the same relationship to their complements at all and in fact they appear to have a similar relationship to their complements as quantifying determiners do to their nominal complements. This is not a thematic relationship and hence it appears that these nouns are not thematic nouns. Clearly they are not determiners either and hence they seem to be prime candidates to be analysed as non-thematic non-functional nouns.

2.1.1 The building blocks of sentences

In (a) we have the verb in the second position, with one of its arguments (the experiencer) to the left, in first position, and another of its arguments to the right in the third position. In (b) however, everything moves one step to the right to accommodate the word yesterday which now occupies first position. The point we made previously is that the ordinal placement of a word in a sentence is unimportant as it is clear that the words Sid, saw and Wendy are in the same grammatical positions in both sentences, even though they are in different ordinal positions in the string of words.

2.1.2 Phrases

We have said that a sentence can consist of a predicate and its arguments. So in a sentence such as (7):

2.1.2 Phrases

we have the verb pestered as the predicate which relates the two arguments Prudence, the agent and Dennis, the patient. Now consider a slightly more complex case:

2.1.2 Phrases

This could mean exactly the same thing as (7), on the assumption that Prudence is a postwoman and Dennis is a doctor. In this case the arguments seem to be the postwoman and the doctor, a sequence of words made up of a determiner followed by a noun. But what status do these sequences of words have in the sentence? It seems as though they function as single words do in (7), inasmuch as they constitute the same arguments as Prudence and Dennis do. Thus these two words seem to go together to make up a unit which is the functional equivalent of the proper nouns in the original sentence. This unit is called a phrase. We can represent this as follows:

2.2.1 The subject

In all the sentences we have looked at so far, there has been an argument of the verb which appears to its left. All of the other arguments have appeared after the verb. As we see by the following sentences, this is an essential fact about grammatical English sentences:

2.2.1 The subject

While there is a special way to pronounce these words in the order in (37c) that would make it grammatical (with a pause after Victor), this would have a special interpretation in which Victor is singled out from a set of possible referents and the rest of the sentence is taken to be something said particularly about him. However, without this special intonation and meaning the sentence is just as ungrammatical as the others: the ‘normal’ word order of English is as in (37a). Thus the basic word order of English has one and only one argument of the verb to its left and all the others to its right.

2.2.1 The subject

From a structural point of view, the argument that precedes the verb also differs from the other arguments. This argument is an immediate constituent of the sentence, whereas all other arguments are inside the verb phrase:

2.2.1 The subject

We call the argument that precedes the VP in the sentence the subject. Besides its privileged position in the sentence, the subject also plays an important role in a number of different phenomena. In a finite sentence, the verb may have a different form depending on properties of the subject:

2.2.1 The subject

Hungarian verbal morphology is a good deal more complex than this, though it is not my intention to go into it here. The point is that although English has less agreement morphology than Hungarian, the phenomenon is the same in that the form of the verb reflects person and number properties of the subject. In English, the other arguments have no effect on the form of the verb:

2.2.1 The subject

The verb seem has just one argument, the clause that Roger ran away, which acts as its complement. Thus from a semantic point of view there is no subject argument here. Yet there is a subject, the pronoun it, which in this case has no meaning. Note that this it is not the same as the one that refers to a third person non-human, as in the following:

2.2.1 The subject

In (49) the situation is exactly like the subject of the finite clause and the expletive subject must be present. In (50) however, the subject is obligatorily absent, though it is clear that the clause is interpreted as though Terry is the subject: the one who is escaping. We will investigate these observations later in this book. For now, however, what all this shows is that subjects are treated rather differently from other arguments from a grammatical point of view.

2.2.1 The subject

However, although this may be true for a lot of sentences, there are many occasions when it is not so. For example, sentences with expletive subjects could hardly be claimed to be about the subject as otherwise they would not be about anything at all. Moreover, other sentences can just as easily be said to be about arguments other than the subject:

2.2.1 The subject

The other semantic aspect of the subject concerns its interpretation as an argument of the verb. This is also very complex, but less doubtful than the claim that the subject is what the sentence is about. When there is a meaningful subject of a verb with two or more arguments, the subject is interpreted as a specific argument, and we do not just interpret it as any one of the possible arguments:

2.2.1 The subject

The verb hit has two arguments: the one who does the hitting, the agent, and the one who gets hit, the patient. But (53) is unambiguous: it must be interpreted with Henry as the agent and Thomas as the patient. Indeed, agent is a very typical Θ-role for a subject to have. Experiencer is also a typical subject Θ-role:

2.2.1 The subject

However, it might be claimed that these are special cases (the nature of their status will be discussed in a later chapter) and that the typical position for such arguments is not the subject.

2.2.1 The subject

In both these cases, the argument Fred is interpreted as experiencer and in (56a) the experiencer is the subject, as would be expected. However, this is not the case in (56b). We see then that the relationship between thematic interpretation and grammatical position is a complex business. We will not go into the matter here as we lack the means to do so. We will return to the issue in a subsequent chapter.

2.2.2 The object

So far we have concentrated on the subject, but what about any other argument: do they have special statuses? One other argument, known as the object, might be claimed to have special features with regard to all other types of complement.

2.2.2 The object

Another fact about objects is that they are arguments which may undergo certain syntactic processes and so seem to be singled out by these. For example, in a passive sentence, the subject may go missing (it may be present inside a by-phrase, but we will not deal with this at the moment). In this case, the argument which would normally be interpreted as the object appears in the subject position. We may interpret this as a process which ‘moves’ the object into subject position:

2.2.3 Indirect object

The fact that people eat dinners and that dinners do not usually eat people is irrelevant as far as the interpretation of these sentences is concerned. What is important is which position each argument occupies and hence which grammatical function each argument has, and this alone is what determines how to interpret the sentence.

3.1.1 Rewrite rules and some terminology

This VP has many things in common with a clause and indeed it looks very much like one. We will discuss the difference between the two in a subsequent chapter. The important point to note is that the theme argument of the verb (the argument undergoing the process described by the verb – in this case, the bubbles) occupies the specifier position of the VP as defined by the rule in (1b).

3.1.4 Specifiers

In general we will find that specifiers are occupied by certain arguments of a predicate or by elements with a certain specified property which relates to the head. This second class of specifiers can only be discussed after a good deal more of the grammar has been established, so we will put these to one side for the moment.

3.1.4 Specifiers

The argument specifiers tend to be subjects, though again this statement will need much qualification as we proceed. One class of verb for which this is most straightforward are those which have theme subjects:

3.1.4 Specifiers

Simplifying somewhat, we might claim that these arguments sit in the specifier of the VP:

3.1.4 Specifiers

These arguments are nearly always DPs and so, unlike the complement they do not seem to be restricted differently by different heads in terms of their category. This is reflected in the lexical entries of the relevant heads in the fact that subcategorised elements are always complements and subjects are never subcategorised for. Of course, there are restrictions placed on these specifier arguments from the predicate, but of a more semantic nature. The verb assigns a Θ-role to these arguments and so the argument must be semantically compatible with the Θ-role it has to bear. For example:

3.1.5.1 Adjunction to X-bar

The other possibility is that the adjective functions as a specifier within the NP and as specifiers precede their heads, this seems more likely. Yet there are properties of the adjective that make it an unlikely specifier. As we saw, specifiers of thematic heads tend to be arguments of those heads. The adjective is obviously not an argument of the noun as it does not bear a thematic role assigned by the noun. Furthermore, specifiers are limited to a single occurrence and there cannot be more than one of them:

3.2 Theoretical Aspects of Movement

The verb hate typically has two arguments, experiencer and theme, and is transitive with the theme as its object:

3.2 Theoretical Aspects of Movement

But in (46) the object appears to be missing. This is not a case of an ‘understood’ object, where the argument is present at a semantic level, as it is fairly obvious that the interrogative pronoun who has the grammatical function of the object. Yet, this pronoun is not sitting in the canonical object position, the complement of the verb, directly after it. Indeed, the interrogative pronoun is occupying a position that no other object can occupy:

3.2.1 Move α

Here the main verb wasted takes its argument in the specifier position of its own phrase. This phrase is in turn the complement of the auxiliary verb was. The argument moves from the specifier of the lower verb to that of the higher one.

3.2.2.1 D-structure and Theta Theory

The principles that determine D-structure positions must therefore have something to do with thematic relationships. We saw in chapter 1 how Θ-roles are encoded in the lexical entry of predicates. Yet in a sentence it is the arguments that are interpreted as bearing these Θ-roles. It must be the case therefore that these Θ-roles are given from the predicate to the argument. We can refer to this process as Θ-role assignment. For example:

3.2.2.1 D-structure and Theta Theory

The verb arrive is a one-place predicate, having one Θ-role to assign which it assigns to the argument an unexpected package in (65a). The verb mend is a two-place predicate. It assigns the agent role to its subject and the patient role to its object.

3.2.2.1 D-structure and Theta Theory

Where can a predicate assign its Θ-roles to? If there were no restrictions on this then arguments would not have distributions at D-structure as they could appear anywhere. We are assuming that this is not so and hence there must be conditions which determine where Θ-roles can be assigned. One fairly clear condition on Θ-role assignment that can be seen in (64) is that Θ-roles are not assigned over long distances. For an argument to receive a Θ-role from a predicate it must be close to it. We can see this from the fact that the following sentence has just one interpretation:

3.2.2.1 D-structure and Theta Theory

We can only interpret this sentence with Sophie doing the suspecting, Linda doing the loving and Dwain getting loved and there is no way to get Sophie associated with love or Linda and Dwain with suspect. This is simply because Sophie is structurally closer to suspect and Linda and Dwain are close to love. If love could assign its Θ-roles over long distances, Sophie might be able to be interpreted as one of its arguments.

3.2.2.1 D-structure and Theta Theory

It is a long standing assumption that there is a uniformity in Θ-role assignment which links certain Θ-roles to certain positions. The reason why the object is assumed to move in a passive sentence is precisely because of this assumption. In an active sentence the object occupies the object position, following the verb, and so it is assumed that in the passive sentence the argument that is interpreted identically to the object in the active originates from the same position that we see it in in the active:

3.2.2.1 D-structure and Theta Theory

Thus, if we propose that the theme argument is assigned to the specifier of the verb it is related to in a structure such as:

3.2.2.1 D-structure and Theta Theory

There are other aspects of the assignment of Θ-roles than those to do with where they are assigned. We saw in chapter 1 that for some predicates an argument that they select as a lexical property does not have to be realised as a syntactic entity but may be present only at a semantic level. Such an argument would be understood, but unable to play any role in a sentence such as licensing a reflexive pronoun:

3.2.2.1 D-structure and Theta Theory

It is not well understood what determines when a Θ-role may be left understood, but it seems to be an idiosyncratic property of certain predicates. It is generally the case that Θ-roles must be assigned. The Θ-role assigned to the subject, for example, cannot be left as understood. Therefore we might propose that there is a general grammatical condition ensuring the assignment of Θ-roles, unless they are marked in the lexical entry of a predicate as being able to be understood. Moreover, a theta role can only be assigned to one argument and cannot be ‘shared out’ between more than one:

3.2.2.1 D-structure and Theta Theory

3.2.2.1 D-structure and Theta Theory

Now if we turn our attention to the arguments themselves we note that it is not possible to have an argument that is not assigned a Θ-role:

3.2.2.1 D-structure and Theta Theory

The verb smile is intransitive and therefore does not have a Θ-role to assign to an object. If we provide this verb with an object, we therefore have an argument that receives no Θ-role, which as we see from (76) is ungrammatical. Moreover, an argument cannot receive more than one Θ-role. So if a predicate must assign more than one Θ-role, it cannot assign them both to the same argument:

3.2.2.1 D-structure and Theta Theory

If it were possible for one argument to bear both Θ-roles of a predicate, (77b) would mean the same thing as (77a) which has a reflexive pronoun in one argument position taking its reference from the other argument. The unacceptability of (77b) can therefore not be a semantic fact.

3.2.2.1 D-structure and Theta Theory

This sentence is grammatical, but only with the interpretation that what Knut knows is that Dennis danced. In other words, the arguments of know are Knut, a DP, and Dennis danced, a sentence in which Dennis is the argument of danced:

3.2.2.1 D-structure and Theta Theory

Again this would involve one argument bearing more than one Θ-role, which appears to be impossible.

3.2.2.1 D-structure and Theta Theory

3.2.2.1 D-structure and Theta Theory

3.2.2.1 D-structure and Theta Theory

3.2.2.1 D-structure and Theta Theory

We have now reviewed three simple and basic principles which regulate the assignment of Θ-roles within a structure: the Locality Condition on Theta-role Assignment, the UTAH and the Theta Criterion. All of these apply to D-structures, restricting the distribution of arguments at this level of representation. Collectively, the principles which govern Θ-role assignment are often referred to as Theta Theory and this can be considered as a part of the grammar, similar to the principles of X-bar theory which regulate the general formation of structures. A final important contributor to the well-formedness conditions of D-structure is the lexicon which provides structures with categorial information and Theta theory with the Θ-roles to be assigned. We might represent this in the following way:

3.2.2.2 S-structure and Case Theory

So far we have looked at some of the principles governing the distribution of arguments at D-structure. In order to understand movement we must now consider some of the principles that apply at S-structure which determine the distribution of arguments at this level.

3.2.2.2 S-structure and Case Theory

While Theta theory accounts for the distribution of arguments at D-structure, it is the principles of Case theory that account for the distribution of DP arguments at S-structure.

3.2.2.2 S-structure and Case Theory

We are now in a position to be able to understand at least certain aspects of movement. Suppose that the principles of Theta theory determine that a DP argument must sit in position X. Suppose further that position X is not a position to which Case is assigned. If the DP remains in this position at S-structure, then the Case filter will be violated and the structure will be deemed ungrammatical. If on the other hand the DP can move to a position to which Case is assigned, then the movement will enable the Case Filter to be satisfied and the structure to be grammatical. This kind of movement might be said to be Case motivated and as we shall see, there are quite a few movements which follow this pattern. This is not the only motivation for movements, however, though we will not go into others at this point. The purpose of this section has been mainly to demonstrate how the interaction of grammatical principles applying at D-structure and S-structure can provide us with an understanding of movement phenomena.

3.2.3 Traces

In these examples, the trace and the pronoun sit in object positions and refer to the subject in much the same way. From this perspective, the trace can be seen as having properties of its own independent of the moved element. We will see that there is some truth to this idea. However, traces are not like pronouns in one important way. As we see in (98b), the pronoun and its antecedent represent two different arguments, though they both refer to the same individual. Thus Harry is interpreted as the one who does the helping and himself is interpreted as the one who gets helped. With the trace in (98b) however, there is only one argument interpreted here: the one who was cheated. Because the verb is passivised, the subject’s Θ-role is not present and so no element is interpreted as agent. Thus, it is as though the trace and the moved element share the same Θ-role, which strictly speaking should not be possible due to the Theta Criterion. From this perspective, the trace and the moved element seem to be interpreted as a single element, capable of bearing a single Θ-role. This single element is, however, spread out across a number of positions in a structure.

4.1 Why the Noun is not the Head of the DP

The phrase in (16a) is indefinite while that in (16b) is definite, obviously as a consequence of the determiner. The noun is the same in both cases and therefore does not seem to contribute to this distinction. But if the determiner is not the head of the phrase, how does it project this property to it? Projection, we showed in the previous chapter, is a property of heads, not adjuncts or specifiers so the fact that determiners do project properties to the phrase is an argument in favour of treating them as heads.

4.3 Multiple Determiners

The object of the verb is associated with accusative Case and hence must be in a Case position. But the object of the noun is not associated with any Case; indeed nouns in general cannot take bare object. We can account for these observations if we simply assume that the complement position of a noun is a Caseless position. Given the Case Filter introduced in the previous chapter, it follows that DPs are not allowed to occupy such a position at S-structure. There is nothing to prevent a noun from taking a DP complement at D-structure, however, and thematically it seems to be the case that many nouns do have DP arguments which all surface as PPs headed by the meaningless preposition of. We might therefore assume that this preposition is inserted into the structure at S-structure so that the Case Filter may be satisfied.

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.

5.2 Verb Types

Obviously, the Verb Phrase revolves around the head verb, the head being the central element of any phrase. Not only does the head project its categorial properties to the phrase, but also by restrictive selection it determines the categorial nature of its complements. Thematic heads also impose restrictions on arguments by assigning Θ-roles to them. The arguments of a thematic head, such as a verb, will appear either in complement or specifier positions according to the principles of Θ-role assignment detailed in chapter 2. It follows therefore that the lexical properties of individual verbs will play a large role in determining the structure of particular VPs. We will organise this central section of this chapter by focussing on different subcategories of verbs, starting with those that have the simplest lexical specifications.

5.2.1 Unaccusative verbs

Perhaps the simplest verb type, seen from a lexical perspective, is a group known as unaccusative verbs. At first sight, these look like simple intransitive verbs, though we shall see that they are in fact simpler than intransitives (or at least, intransitives are more complex!). Unaccusatives take one DP argument to which they assign a theme Θ-role. They may also, optionally in most cases, take a location or path argument expressed by a PP:

5.2.1 Unaccusative verbs

In this usage, the verb is not unaccusative as it involves an agent argument: only something which is capable of volitional action can ‘sit’ in this sense, but virtually anything that is capable of being located can ‘sit’ in the unaccusative sense.

5.2.1 Unaccusative verbs

Unaccusative verbs have a certain range of properties by which we can identify them. One is that they may appear in there sentences, which we have mentioned several times in the previous chapters. These have a there subject and the theme argument sits behind the verb (and must be indefinite):

5.2.1 Unaccusative verbs

Another structure in which we find unaccusatives is known as the locative inversion construction. This involves the PP locative argument apparently sitting in subject position, while the DP theme again sits behind the verb:

5.2.1 Unaccusative verbs

Like the there construction, locative inversion seems to be available only for unaccusative verbs and cannot be used with other verbs which have locative arguments or adjuncts:

5.2.1 Unaccusative verbs

In the most straightforward case, ignoring the complexities of the there construction for the moment, the theme argument is always the subject. If there is a prepositional argument, this always appears behind the verb, presumably in complement position. Thus the simplest assumption we could make for the structure of a VP involving an unaccusative verb is:

5.2.1 Unaccusative verbs

The position to which the subject moves is typically a nominative position and so we might assume that the movement has something to do with placing this argument in a Case position. The reason why these verbs are called ‘unaccusative’ is because unlike with transitive verbs, which share the possibility of having theme arguments, the theme of the unaccusative cannot normally remain inside the VP to receive accusative Case.

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:

5.2.2 Light verbs

What light verbs actually contribute to the meaning of an expression is a complex and subtle issue. For example, it seems from (26) that they do have something to do with argument structure as the main difference here is to do with the number of arguments. The other examples in (25) demonstrate that the contribution of the light verb can affect aspect (do a dance verses dance) and duration (take a look verses look) of an event.

5.2.2 Light verbs

What is the structure of the VP containing a light verb? Let us concentrate on the cases in (27). In these we have the light verb itself with a subject to its left. To the right we appear to have a VP containing the main verb and its arguments. Suppose we assume that the main VP is a complement of the light verb. This would give us the structure:

5.2.2 Light verbs

The thematic relationships are straightforward. In the lower VP we have a situation fairly similar to the VP in the previous section. The theme argument, the vase, is in the specifier of the VP as we discovered previously. The verb break therefore looks fairly similar to an unaccusative verb (we will investigate the properties of this type of verb more fully in the next section). The specifier of the vP is interpreted as an agent and therefore the light verb is clearly not unaccusative. This is not surprising as unaccusative verbs either have no complement or prepositional ones, and here the light verb has a VP complement. In terms of the UTAH, we might therefore propose that the agent Θ-role is assigned to the specifier of a (light) verb which has a VP complement.

5.2.3 Ergative verbs

We have just seen that a verb like break can appear in a VP with a single theme argument which in the absence of a light verb will be the subject of the clause. This looks exactly like an unaccusative verb, yet there are differences between this class of verb and the unaccusatives. For one thing, these verbs are not movement or locative verbs, but typically involve a change of state:

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.

5.2.3.2 Light verbs and ergatives

The disadvantages of this analysis are: i) there is an empty light verb and ii) the wrong word order is predicted. The supposition of the empty verb is, of course, not a problem in itself. We have seen a number of instances of empty categories that are well justified and enable us to provide accounts for phenomena that would otherwise be mysterious. As long as we can independently justify the assumption of an empty element, given that language apparently makes use of such things, there is no problem in the assumption itself. There is both semantic and syntactic evidence of the existence of the empty light verb. We will return to the latter, but the semantic evidence is fairly obvious: the structure is interpreted as a causative and the presence of this meaning justifies the assumption of a light verb which provides it. Similarly, the presence of a ‘causer’ argument justifies the assumption of a predicate which assigns the relevant Θ-role. As there is no such visible predicate which can do such things in (48b) our conclusion is that this predicate is ‘invisible’.

5.2.3.3 Unaccusatives and ergatives

All of this demonstrates that ergative verbs can be analysed in exactly the same way as unaccusatives, in their ‘intransitive’ use, and as being part of a causative construction in their ‘transitive’ use. Indeed, ergative verbs themselves are identical to unaccusatives, even in causative constructions as it is the causative light verb which supplies the extra agent argument and the causative interpretation. For this reason, many linguists refer to these kinds of verbs as unaccusatives. However, it still remains that there are differences between the unaccusative verbs we reviewed above and the ergative verbs reviewed in this section. For a start, ergatives cannot appear in the there constructions and unaccusatives cannot appear in causative constructions:

5.2.3.3 Unaccusatives and ergatives

Of course, we also need to explain why the theme argument, that we have claimed to be generated in the specifier of the VP, sits behind the verb, not in front of it, in the there construction:

5.2.3.3 Unaccusatives and ergatives

However, we do not want to say that there is absolutely no connection between the abstract light verb and its subject, as there are restrictions placed on it: the subject must be there and not it. Thus, suppose that there is a special argument of this predicate, which receives no actual Θ-role from it but is restricted by it. A similar notion of ‘quasi-argument’ has been proposed for cases such as:

5.2.3.3 Unaccusatives and ergatives

The it subjects that accompany weather predicates are clearly not arguments as they have no referential content, but they are somehow not quite as empty as the expletive it in examples like (62b).

5.2.3.3 Unaccusatives and ergatives

The intended interpretation of (70b) is that Rob pretends to be rich in order to impress the neighbours, an thus it is the ‘seeming’ rather than the ‘being rich’ that is being modified by the purpose clause. The ungrammaticality of this sentence with this interpretation demonstrates that expletive it is unable to license this kind of modifier. When the thematic subject is raised, however, the purpose clause is grammatical. The quasi-argument weather predicate it appears to behave like a thematic argument in this respect as it does license a purpose clause. Obviously I do not want to claim that the there subject in there constructions is the same thing as a weather predicate’s quasi-argument subject.

5.2.4.1 Evidence from passives

If the light verb ceased to be there, both the agent Θ-role and the accusative Case would be lost in one step. What would be left is the main verb with its patient argument which would lack Case and hence have to move to subject position:

5.2.4.1 Evidence from passives

In all these examples however, the passive verb has the passive morpheme. We can incorporate this into our analysis of the passive making use of an idea presented above, that certain bound morphemes can be analysed as light verbs. Suppose we assume that the passive morpheme is a light verb which replaces the agentive light verb of the active. As the passive light verb does not assign a Θ-role to its subject, it will not be able to assign a Case to the theme in the specifier of the VP and hence this argument will have to move to subject position. Moreover, the bound morpheme status of the passive light verb will force the main verb to move in order to support it:

5.2.4.2 Extended projections

In (82a) the verb give is used fully thematically and it contributes its full descriptive content to the whole sentence: the agent is in possession of the chocolates, and does something (i.e. gives) that results in the recipient in possession of the chocolates. But in (82b), where give is used as a light verb, it does not contribute its whole semantic content. For example, it cannot be claimed that anything has been given here and certainly Kevin does not end up in possession of a kick! Instead the main descriptive content comes from the deverbal noun and hence the similarity of meaning of (82b) and (c). It seems that semantically speaking, the complement of the light verb is the main contributor to the construction and although light verbs do contribute something, their contribution is often subtle and always dependent on the thematic complement. This shows a very different relationship between a light verb and its complement and a thematic verb and its complement. In the latter case, the thematic verb selects and imposes restrictions on its complement whereas in the former, the light verb is in some ways selected for and restricted by its complement: recall that unaccusative verbs do not appear with the abstract causative light verb, but ergatives do. Suppose then that the main semantic aspects of a light verb are determined by its thematic complement and that these are passed up to it by a process similar to projection – something which has been called extended projection, in fact. It would then depend on the thematic verb how the argument of the light verb was to be interpreted, as a causer, not directly seen as the agent of the thematic verb, or as a direct agent of that verb. We might visualise this in the following way:

5.2.4.2 Extended projections

If this is right, then the agent subject of the light verb involved with transitive verbs will receive its Θ-role indirectly from the main verb, via the light verb, and hence will be interpreted as the argument of the thematic verb. Of course the actual assignment of the Θ-role will be dependent on the presence of the light verb, as by the UTAH roles such as agent can only be assigned to the specifier of a light verb.

5.2.4.3 Agent and experiencer subjects

One observation that might be relevant here is that there are some verbs which take both agent and experiencer arguments. With these verbs, the agent always precedes the experiencer:

5.2.4.3 Agent and experiencer subjects

Assuming that the agent is in the specifier of a light verb, these observations suggest that the experiencer is in the specifier of the thematic verb, like theme arguments:

5.2.4.4 Multiple light verbs

If we assume that experiencers are assigned their Θ-roles in the specifier position of a light verb, we face a problem in analysing verbs with agent and experiencer arguments as in (88). What is puzzling about these verbs is how they can exist at all, given our assumption that agent and experiencer receive their Θ-roles in the same position. The only analysis available to us, if we wish to maintain the UTAH, is to assume that there are two light verbs in these constructions, one for the agent and one for the experiencer:

5.2.4.4 Multiple light verbs

The Θ-roles lower on the hierarchy have to be discharged on to an argument before those higher up. The UTAH ensures that Θ-roles can only be discharged in certain positions and in combination with (98) we get the following pattern. The first Θ-role to be assigned is the theme, if there is one. As this can be assigned to the specifier of the thematic verb it will be. Next the experiencer Θ-role must be assigned, providing there is one. This can only be assigned to the specifier of a light verb so the thematic verb will extend its projection to include a light verb and the experiencer Θ-role will be assigned to its specifier. Finally, if there is an agent, again this can only be assigned to the specifier of a light verb and hence will force the verb to extend its projection. If there already is an extended projection, a second light verb will be added to accommodate the agent. Thus, the agent will always be higher in the structure than the experiencer and theme.

5.2.5 Intransitive verbs

Intransitive verbs are verbs with one argument, but unlike unaccusatives this argument is either an agent or an experiencer, i.e. one of the Θ-roles assigned to the specifier of a light verb. Accordingly then, we may analyse them as involving the following structure:

5.2.5 Intransitive verbs

The one argument will move to the subject position in order to get Case and presumably the verb will move to support the light verb.

5.2.5 Intransitive verbs

Given the structure in (101) a number of possible analyses of cognate objects suggest themselves. One is to assume that these are like theme arguments, though obviously highly restricted by the thematic verb and hence they appear in the specifier of the VP and end up behind the verb when it raises to the light verb position:

5.2.6 Multiple complement verbs

So far we have been concerned with verbs that have either one or two arguments, but there are cases of verbs with more. In this section we will look at a number of verbs which have three arguments, again trying to maintain the UTAH and using this as a guide for the analysis of the VP’s structure.

5.2.6 Multiple complement verbs

Within the standard X-bar structure there are two positions in which we find arguments: specifier and complement:

5.2.6 Multiple complement verbs

Verbs with more than two arguments have therefore been considered as problematic. However, once we consider the role of light verbs as assigners of Θ-roles regulated by the thematic verb, we can see that it is possible to extend the Θ-roles assigning domain of a thematic verb to more than two positions. This is essentially the approach we will adopt here.

5.2.6 Multiple complement verbs

Again, the arguments are similar, involving an agent, a theme and a PP complement expressing the goal or beneficiary and so we can expect the structure to be similar. This structure is sometimes called the dative construction. The interesting thing about these verbs is that they can often enter into another construction which means virtually the same thing as the dative, only involving two DP complements:

5.2.6 Multiple complement verbs

In this structure the theme is sitting in the complement position of the thematic verb, not the specifier, and the goal is in the specifier. The indirect object is obviously interpreted in the same way as the PP is in the dative construction and so we should expect it to appear in the complement position if the UTAH holds. We might try to account for the properties of the double object construction via a movement analysis, using the dative construction as the underlying arrangement as this seems relatively unproblematic. The question is, what moves and where does it move to? A minimal assumption is that besides the verb moving to the light verb position, one of the arguments moves to change their order. Thus, either the theme moves backwards or the goal moves forwards. If the theme moves backwards, it isn’t clear what position it would move to and moreover it isn’t clear why it would move, given that the position it occupies seems to be a Case position in virtually all other cases we have looked at. The goal argument is slightly different however. In the dative construction there is a preposition and this we might assume is what is responsible for providing the argument with its Case. In the double object construction, however, this preposition is not present and hence the argument cannot be assigned Case in the same way. This would then provide the motivation for the argument to move to a position in which it could get case. Considering the problem more closely the goal must move to a phrasal position between the specifier of the VP, occupied by the theme, and the light verb to which the main verb moves. The only possibility is that there is another specifier position between the two:

5.2.6 Multiple complement verbs

The remaining problems to solve are the identity of X and how the theme argument gets Case if the goal argument gets the Case assigned by the light verb. The obvious answer to the latter is that X provides the theme with its Case, which in turn suggests that X is a Case assigning head, i.e. a verb or a preposition. If X is a verb, we have a structure which is identical to those involving multiple light verbs:

5.2.7 Phrasal verbs

It seems that it is only when the verb has a PP complement which consists only of a prepositional head that the preposition is allowed to move out of the PP. If the preposition itself has a complement, or if it is modified, then it is not allowed to move. It is not entirely clear why this should be, as other heads can move out of their own phrases when there are other elements in other positions within them. For example, we have seen many cases of a verb moving out of the VP when its specifier or complement are filled by its arguments. Another observation from (138) might help to shed some light on the problem. Note that when the verb has a simple PP complement, it has a different interpretation: to put something off does not mean the same as to put something somewhere. Similarly, put down, put on, put back, put over, etc. all have somewhat idiosyncratic meanings that are not simply related to the meaning of put as a verb of placement. So, put down can mean ‘to kill’ (of animals), put on ‘to fake’, put back ‘to delay’ and put over ‘to convey’. This might suggest that it is not the same verb we are looking at in all these cases and especially they are not the same verb as in (138b). If this is true then it could be that the ability of the preposition to move might be lexically restricted by the verb: some verbs allow it, others do not. Of course, this still does not explain why those that do allow the preposition to move only take ‘simple’ PP complements, which contain just the preposition and so we cannot be said to have solved all the mysteries of phrasal verbs here. In fact we have probably only just scratched the surface and it has to be admitted that phrasal verbs present many very difficult problems for analysis under any set of assumptions. We will therefore leave this topic at this point and be content with the meagre understanding of them that we have gained.

5.2.8 Verbs with clausal complements

If we consider the thematic roles assigned to these arguments, typically the DP arguments receive a goal Θ-role: the one to whom the event described by the verb is directed. The clause has a theme Θ-role. We saw with dative/double object verbs, the goal argument sits in the complement position of the thematic verb, but may move in order to get Case. If this is what is going on here, then the structure should 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

A second problem is why the PP argument would undergo the same movement as the DP as PPs do not need Case and do not normally undergo this kind of movement. Hence it appears that there is not much to recommend this analysis.

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.

5.4.1 Adverbs

If an adjunct is placed between the head and its sister, i.e. the complement, the branches of the structure cross and this is not a possible configuration. The problem with this account, however, besides its reliance on the assumption that complements are all sisters to the head, is that it is not at all clear why various movement phenomena would not separate the head from its complement. Another account, due to Stowell (1981) assumes that the verb is responsible for assigning Case to the object and that there is an adjacency requirement on Case assigners and assignees. As we have assumed that the theme gets its Case from the light verb, we cannot use Case adjacency to account for why the verb and its theme argument cannot be separated. Even if we assume that Case assigners must be adjacent to the element they Case mark, this will not prevent the verb moving to a higher light verb position allowing an adverb to come between the two:

5.4.1 Adverbs

This structure has the adverb phrase adjoined to the lower vP and the verb moving to the higher light verb. Such a structure would be possible either when there is both an agent and an experiencer argument, or if the top light verb is an aspectual morpheme. The structure that would be produced however would be ungrammatical as the adverb would appear between the verb and its theme argument.

5.4.2 PP modifiers

The only way for (176a) to have been generated would be to adjoin the PP to the left of the lower VP. However, PPs never adjoin to the left, only to the right, and moreover this would necessitate the verb moving over the PP adjunct. As this is impossible for AP modifiers, we can assume that it is impossible for PP modifiers as well. In (176b), assuming the locative PP to be the complement of the verb, the only way for this to get behind the PP adjunct would be for it to move. And hence we can assume that there is a backwards movement that PP arguments may undergo which is similar to the movement that clausal complements undergo, as discussed in section 3.8. That PP complements may undergo such a movement is supported by the following data:

5.4.3 Clausal modifiers

Although these clauses seem to lack a subject, it is immediately obvious that a subject is interpreted: in (180a) it is Bert who will be doing the impressing and in (180b) it is they who will collect the insurance. We call this phenomenon control. There is an element in the main clause who is interpreted as, or who ‘controls’ the missing subject of the modifying clause. There are restrictions, however, on which argument can act as the controller:

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.

7.2 The Clause as CP

The generalisation is now that all verbs which select for a finite declarative complement select for a CP. This is fairly easy to capture in terms of the notion of canonical structural realisation principles. The idea behind this is quite simple. Basically, certain arguments are canonically realised by certain categories. For example, themes are typically realised as DPs and locations as PPs. This is their ‘canonical realisation’. It may be that a certain degree of non-canonical realisation of arguments is possible, for example the nominal home can realise a goal argument usually realised by a PP:

7.3.2 Wh-movement

The movement of the wh-element to spec CP therefore seems to have an interpretative motivation, which contrasts with the grammatical motivation of the movement to spec IP. There are other differences between the movements, which we will look at in the next chapter. We may, for now, simply identify the kind of Case filter/grammatically motivated movement as A-movement (A stands for ‘argument’ as it is only argument DPs which undergo it) and the kind of semantically motivated movement, such as wh-movement, as Ā-movement (? means ‘not argument’).

8.1.2 Clauses without IP

If some clauses can lack CPs, the question naturally arises as to how small clauses can get. Recall that clauses are structured in various layers, each of which adds something to the interpretation of the whole clause. But the basic proposition is expressed by a predicate and its arguments. As the predicate does not have to be a verb a basic proposition can be expressed with no verbal element at all:

8.1.2 Clauses without IP

Both expressions on either side of the hyphen state exactly the same relationships between the predicate and its arguments. The difference is that while the expressions on the right are grammatical English sentences, those on the left are not. Or at least, one might think so. But consider the following:

8.2 Raising and Control

The verb seem is one which takes a clausal complement but it has no thematic subject. In this case the subject position is filled by a meaningless it, known as a pleonastic subject. This subject, like all subjects in English finite clauses is obligatory. This suggests that the obligatory nature of the subject is more than a semantic condition that arguments need to be realised. In fact there seems to be a grammatical requirement that clauses have subjects. This condition has been called the Extended Projection Principle (EPP). Recall from chapter 3 that the Projection Principle ensures that the lexical properties of heads are projected into the structure at all levels of syntactic representation. Thus if a verb requires an object as a lexical property, it must have an object at D-structure and at S-structure. The Extended Projection Principle claims not only this, but that the subject position must be present at all levels of structural representation and moreover that it must be filled by something at S-structure. Of course, under usual circumstances there will be something in the subject position at S-structure as an argument of the verb will move there for Case reasons. But even if there is no argument inside the VP in need of Case, the subject position must be filled by the insertion of a pleonastic subject:

8.2 Raising and Control

Recall also that the Θ-Criterion requires that Θ-roles be assigned to arguments. While some verbs may take implicit arguments which are not actually present in the structure but are ‘understood’ at a semantic level, these arguments are always complements and never subjects:

8.2 Raising and Control

Consider the two sentences in (44). They both appear to have missing subjects and in other ways they seem to be similar. However, even at this point we can see that the two missing subjects are not entirely equivalent. In (44a), the missing subject is referentially identical to the subject of the higher clause, Tim. This subject is not semantically related to the verb of its own clause: Tim is not the one doing the seeming. We have already seen that verbs like seem do not have a subject of their own and often have pleonastic subjects, as we see in (45a). Thus the missing subject of the non-finite clause and the overt subject of the higher clause share a single Θ-role assigned from the lower predicate tall. In other words, they represent a single argument. This contrasts starkly with the situation in (44b), where the missing subject and the overt subject of the higher clause bare completely different Θ-roles: Robin is the thematic subject of want and the missing subject is the thematic subject of rich. The two subjects are coreferential, but they are independent elements in exactly the same way that a pronoun and its referent are independent elements:

8.2 Raising and Control

In this case, Henry is the one doing the thinking and he is the one who is happy. If the pronoun refers to Henry then the interpretation is ultimately that Henry is happy (or at least this is what he thinks). But if the pronoun refers to someone else, then the interpretation is not that Henry is happy. However, the two elements are independent, regardless of what their referential properties are. The same is true of the overt subject and the missing subject in (44b). (45b) demonstrates that there really are two independent arguments in this construction as the subject of the higher predicate cannot be spelled out as a pleonastic element. In contrast, this is exactly what is possible in (45a), demonstrating that there really is only one argument here. The same point is made the other way round in (46). In this case we see that with a verb like seem, a different argument cannot be realised in the two different subject positions as there is only one Θ-role involved. A verb like want, on the other hand, can realise arguments overtly in both subject positions as there are two independent Θ-roles. So, one kind of missing subject shares a Θ-role with another element in the sentence while the other kind of missing subject has a Θ-role all of its own. The next three examples demonstrate that the two different kinds of missing subject have different referential properties. The non-independent type of missing subject which shares a Θ-role with its antecedent, must be lower in the structure than its antecedent. Hence it cannot be part of a structure which is raised to a higher position and the ungrammaticality of (47a) follows. The independent type of missing subject on the other hand can, under certain circumstances be higher in the structure than its antecedent, hence the grammaticality of (47b). The contrast in (48) again shows a difference in the referential properties of the two missing subjects. In (48a) we see that it is impossible for the dependent missing subject to refer to an object: they are always associated with subjects. In (48b) we have an independent missing subject, it being the one who is doing the looking after. As we can see, it is capable of referring to the object as the ultimate meaning is that Larry will be the one ‘looking after himself’. Finally, (49) shows that the dependent type of missing subject cannot refer out of the subject clause of another clause, where as the independent missing subject can.

8.2 Raising and Control

Of course, the fact that the dependent missing subject shares a Θ-role with its antecedent can lead to only one conclusion, given the theory of Θ-role assignment discussed in chapter 2: the missing subject and its antecedent are a single argument. In other words, this kind of phenomenon is the result of movement and the missing subject is a trace:

8.2 Raising and Control

In reality then, the subject is not ‘missing’, it has just moved. By the same conditions, we cannot treat the independent missing subject as the result of movement: it bears a Θ-role different from its antecedent and so they do not represent a single argument, but two different ones. Moreover, at D-structure these two arguments must also be in different positions as different Θ-roles are assigned to different positions at D-structure. Hence we conclude that this kind of missing subject is present at D-structure. As we see in (51), this is not true of the other ‘missing’ subject. Because this element has many properties in common with a pronoun, i.e. it bears an independent Θ-role but can be referentially dependent on something else, it is often referred to as PRO.

8.3 The Gerund

A number of consequences follow from this analysis. First, in (91a) as there is no light verb, the DP in the specifier will be Caseless and hence the preposition of will have to be inserted to salvage the structure. The presence of the light verb in (91b) renders this unnecessary as the DP will get Case in the normal way. More interestingly, the subject is present in the structure in (91b) but not in (91a). This does not mean that the subject cannot be introduced in (91a) as it is perfectly possible to have a possessive argument introduced into the NP/DP structure under normal circumstances, so there is no reason why it could not happen here. However, the point is that, as is usually the case with nouns, this will be optional and therefore the possessor/agent may or may not appear. When there is no possessor, there can be a determiner and when there is a possessor there can be no determiner, as the two are in complementary distribution: