Limits of the Compound Verbal Predicate

Now we come to the second question, about the limits of the compound verbal predicate. It arises from the fact that a rather considerable number of verbs can be followed by an infinitive, some of them with, others without the particle to. Among such verbs are: shall, will, should, would, can, may, must (without to); ought, wish, want, desire, hate, fear, begin, start, continue, omit, forget, remember, etc. (with to).

The relation between these phrases and parts of the sentence is of course not the same in all cases. We can at once eliminate the phrases "shall, should, will, would + infinitive", which consti-

205 PARTS OP A SENTENCE. THE MAIN PARTS

tube tease or mood forms of the verb. Thus, the phrase shall write is a form of the verb write (as it does not differ from the forms write, writes, wrote in its lexical meaning) and, consequently, it is a simple verbal predicate. The phrases with the verbs can, may, must, ought (in the latter case with to) constitute a compound verbal predicate (this is almost universally recognised). But the phrases with the verbs wish, want, desire, hate, fear, begin, start, continue, etc. give rise to doubts and controversies. On the whole, there are two views expressed in this matter. According to one of them, all such phrases are also a compound predicate: the finite verb (wish, begin, etc.) does not denote any action of its own, it merely denotes the subject's attitude to the action expressed by the following infinitive (in the case of wish, fear, etc.), or a phase in the development of that action, namely, its beginning, continuation, etc. (in the case of begin, continue, etc.); consequently, it is argued, the phrase as a whole constitutes the predicate of the sentence: it is a compound verbal predicate, just as in the case of can, may, or ought. This argument, as will be easily seen, is based on purely semantic reasons: its decisive point is, that the finite verb does not denote any special action and only denotes the subject's attitude to it, or a phase of the action itself. But this is irrelevant from the grammatical viewpoint. What is more, this line of reasoning is dangerous: if we were to follow it to its logical consequences we should have to include into the predicate not only such phrases as stopped laughing, avoided meeting, and a number of other phrases including the gerund, but also such phrases as began his work, continued his speech, liked his job, and a number of other phrases containing a noun. Indeed, from the semantic viewpoint, on which the argument for began to work being the predicate is based, there is no difference between began to work and began his work.

Therefore, approaching phenomena from a grammatical viewpoint, which is the essential one here, we start from the assumption that in the phrase began his work the group his work is a separate (secondary) part of the sentence (an object).¹ This shows that the verb begin can be followed by a noun functioning as an object (the same of course applies to a number of other verbs). Since the verb begin can take an object there appears to be no reason to deny that an infinitive following this verb is an object as well. We might give here a table based on what is called transformation:

began to work — began his work continued to work — continued his work liked to sing — liked songs

etc.

¹ We are not discussing here the syntactic position of the word his (the attribute). For this problem, see p. 229 ff.

LIMITS OP THE PREDICATE 206

On the other hand, no table of this kind is possible with such verbs as can, may, must, ought: they cannot under any circumstances be followed by a noun, and this is an important difference on which syntactic analysis should be based.

Another question of a similar kind arises with reference to sentences containing idioms of the pattern "verb + noun", e.g. make a mistake, make one's appearance, have a look, have a smoke, take a glance, etc. Here two different approaches are possible, and the approach chosen will predetermine all conclusions to be arrived at in considering concrete examples.

One approach would be to say that if a phrase is a phraseological unit, that is, if its meaning is not equal to the sum of the meanings of its components, it cannot be divided into two parts of the sentence, and has to be taken as one part, namely, the predicate.

The other approach would be to say that such phraseological phenomena belong to the sphere of lexicology alone and are irrelevant for grammar, that is, for sentence analysis.

The choice between the two approaches entirely depends on the view one takes of grammar, its place in linguistics, and its relation to lexicology. It does not seem possible to prove that one of the approaches is right and the other wrong.

One of the arguments in favour of the view that phraseological units should be treated as one part of the sentence, is this. If the phrase "verb + noun" is not a phraseological unit, a separate question can be put to the noun, that is, a question to which the noun supplies an answer. For instance, if we take the sentence He makes toys the question would be, What does he make? and the answer would be supplied by the word toys, which, accordingly, is a separate part of the sentence, namely, an object. If, on the other hand, we take the sentence, He makes mistakes, it would not be possible to ask the question, What does he make? and to give mistakes as an answer to it. Consequently, according to this view, we cannot say that mistakes is a separate part of the sentence, and we must conclude that the phrase makes mistakes as a whole is the predicate.

However, this sort of argument is riot binding. The method of asking questions, though widely used in school language teaching, is not a scientifically valid method of syntactic study. In a number of cases the choice of the question is arbitrary, and there are even cases when no question at all can be asked.

Thus, the decision between the two alternatives presented above rests with the scholar. This is, and most probably will always be, a matter of opinion rather than of proved knowledge.

Before we go further in this matter, let us consider another case also belonging here, namely phrases of the type come in, bring up, put down, etc., which we discussed in Chapter XVII, when studying

207 PARTS OP A SENTENCE. THE MAIN PARTS

parts of speech. Should these phrases be taken as the predicate, or should the predicate be limited to the verb alone (come, bring, put, etc.)? This again is a matter of opinion. The phrase come in, for instance, can equally well be analysed as the predicate of the sentence, and as a combination of the predicate and a secondary part. On the other hand, the phrase bring up (as in the sentence, They brought up three children) would be taken to be the predicate, rather than a combination of the predicate with a secondary part, and this of course is due to the meaning of the phrase, which certainly is not equal to the sum of meanings of the verb bring and the adverb up. This semantic consideration is in favour of taking the whole phrase to be one part of the sentence (its predicate). But again, this argument is not binding. Whether such semantic considerations should or should not be taken into account in syntactic analysis is a matter of opinion. It is possible to argue that considerations of this kind should not weigh when we are engaged in syntactic studies. On the whole, we will adhere to the view that such considerations should be taken into account, and accordingly we will consider the phrases bring up, set in, etc., as the predicate of the sentence.

The Compound Nominal Predicate

The compound nominal predicate always consists of a link verb (also called copula) and a predicative, which may be expressed by various parts of speech, usually a noun, an adjective, also a stative, or an adverb (as in the sentence The lesson is over). Often enough the predicative is represented by a phrase, most usually of the pattern "preposition + noun", which may or may not be a phraseological unit.

Now we must find the characteristic features of a link verb. It should first of all be noted that the term "link verb" (as well as the term "copula", after which it appears to have been coined) is not a very happy one. The idea of "link" suggests that its function is to connect the predicative with the subject. This, however, is hardly intelligible. Why should the predicative need some special word to connect it with the subject? It could stand side by side with the subject without the help of any "link". Indeed it does not require any link in sentences with the simple nominal predicate which we have discussed on p. 208 ff., and this is still more usual in Russian, where no link verb as a rule appears in the present tense. The true function of a link verb is not aconnecting function. It expresses the tense and the mood in the predicate. The link verb be, which expresses these categories, and also those of number and person, is rightly considered to be the most abstract of all link verbs, that is, the one

THE COMPOUND NOMINAL PREDICATE 208

most devoid of any meaning of its own. Other link verbs have each some lexical meaning.

Though the term "link verb" is purely conventional, we will retain it, as it is in common use and an attempt to substitute another term would stand little chance of success.

Besides the verb be there are a number of other link verbs with different meanings which we need not discuss here, for instance become, get, continue, grow, turn, e. g. Then he grew thirsty and went indoors (LINKLATER); But presently the sea turned rough (Idem), etc. It will be readily seen that some of them do not always perform this function but may also be a predicate in themselves, for instance the verb grow in the sentences The child has grown, or, We grow potatoes. Of course it is only the meaning of the noun following the verb that shows whether the noun is a predicative or an object: compare the two sentences They have grown fine young men and They grow potatoes. So if we say that a verb is a link verb this need not necessarily mean that it is always a link verb and cannot perform any other function.

To approach the subject of link verb and predicative from another angle, we may say that if a verb is followed by a predicative it is, to some extent at least, a link verb. The restriction "to some extent at least" is necessary because there are sentences in which the finite verb is a predicate in itself, that is, it contains some information about the subject which may be taken separately, but at the same time the verb is followed by a predicative (a noun or an adjective) and is in so far a link verb. This is found in sentences like the following: He came home tired, She married young, He died a bachelor, etc. The finite verb in such sentences conveys a meaning of its own (he came, she married, he died), but the main point of the sentence lies in the information conveyed by the predicative noun or adjective. We might retell the meaning of these sentences in another way, namely: He was tired when he came home, She was young when she married, He was a bachelor when he died, etc. The finite verb, besides being a predicate in itself, also performs the function of a link verb.

Since such sentences have both a simple verbal predicate and a compound nominal predicate, they form a special or mixed type: predicates of this kind may be termed double predicates. ¹ Here are some examples:

¹ Corresponding phenomena in Russian have been treated by Academician A. Shakhmatov, who named such sentences "double-predicate sentences" (двусказуемые предложения). See А. А. Шахматов, Синтаксис русского языка, стр. 221 сл.

For a treatment of this type of predicate in English see M. M. Галинская, О двусказуемых предложениях в современном английском языке. Иностранные языки в школе, 1948, № 2.

209 PARTS OP A SENTENCE. THE MAIN PARTS

Sunlight seeped thick and golden through the high, oblong windows above the cages and fell in broad shafts to the linoleum floor where he dropped his bucket. (BUECHNER) Compare also the following sentence: Catherine's blood ran cold with the horrid suggestions which naturally sprang from these words. (J. AUSTEN) The lexical meaning of the verb run is here almost wholly obliterated, as will also be seen by translating the sentence into Russian, or, indeed, any other language. The essence of the predication is of course contained in the predicative adjective cold.

Let us now look at a few more examples of sentences with a predicative coming after a full predicate with secondary parts attached to it. She had set her feet upon that road a spoiled, selfish and untried girl, full of youth, warm of emotion, easily bewildered by life. (DREISER) A spoiled, selfish, and untried girl is a predicative, coming after a fully developed predicate group consisting of the predicate itself, an object and an adverbial modifier. That the group a spoiled, selfish and untried girl is a predicative, is clear, because no other syntactical tie between this group and the preceding words in the sentence can be imagined. It is a peculiarity of this sentence that the predicative has three loose attributes belonging to it: full of youth, warm of emotion, and easily bewildered by life. They make this predicative group very weighty indeed. It may also be noted that the predicative group a spoiled, selfish and untried girl, full of youth, warm of emotion, easily bewildered by life represents the rheme of the sentence, while the preceding words in the sentence represent its theme. Indeed, the contents, or the purpose of the sentence, is not to inform the reader that she had set her feet on that road, but what kind of person she was at the time she did so. If the predicative (with its secondary parts) were to be dropped, the communication value of the sentence would be basically changed, and in the context in which it stands its value would be reduced to nought.

The same is found in the following examples: You've come home such a beautiful lady. (TAYLOR) I sat down hungry, I was hungry while I ate, and I got up from the table hungry. (SAROYAN)

It should also be noted that the verb preceding the predicative and therefore performing (at least partly) the function of a link verb, may be in the passive voice. This is especially true of the verbs find, think, report, as in the sentences, He was found guilty, He was reported dead, etc.

From such sentences there is an easy transition to sentences in which the finite verb is followed by an infinitive, as in He was known to have arrived, etc.

It may be the infinitive of the verb be, which is then in its turn followed by a predicative (a noun or an adjective), for instance, He was said to be a great actor, He was reported to be dead, etc.

THE DOUBLE PREDICATE 210

As far as meaning is concerned, there seems to be no difference between the sentences He was reported dead, and He was reported to be dead, or between the sentences He seemed clever and He seemed to be clever. As far as structure is concerned, the second variant in each case is somewhat more complicated, in that the finite verb is first followed by an infinitive, which apparently is bound to be a predicative (since it comes after the link verb), but which is itself the infinitive of a link verb and therefore followed by another predicative.

Besides the combinations of different predicates, already mentioned various other combinations are possible and actually occur in texts. However, finding out all these possibilities is of no particular scientific interest. ¹

¹ We shall have to touch on another question connected with the predicate after examining the secondary parts of the sentence (see p. 237 ff,).

Chapter XXVII

SECONDARY PARTS

The theory of the secondary parts is one of the last developed sections of linguistics. The usual classification of these parts into objects, attributes, and adverbial modifiers is familiar to everyone, no matter what his mother tongue may be and what foreign language he may study. Yet it has many weak points. The characteristic features of each of the three types are not clearly defined, and describing a given word or phrase as an object or an attribute in some cases, or again describing it as an object or an adverbial modifier, in others, often proves to be a matter of personal opinion or predilection. Such statements then lack all scientific value. Indeed, with some scholars the theory of secondary parts of the sentence as usually given has been discredited to such an extent that they have attempted to discard it altogether. For instance, Prof. A. Peshkovsky, in his very valuable book on Russian syntax, does not use the notions of object, attribute, and adverbial modifier at all. Instead, he classifies the secondary parts of a sentence into those which are governed and those which are not.¹ The notion of government, however, properly belongs to the level of phrases, not to that of the sentence and its parts. So we will not follow Peshkovsky in this method of classification. But his decision to do without the traditional categories of secondary parts is very instructive as an attempt to do away with a method whose weaknesses he realised well enough.

We will now look at the three traditional secondary parts of a sentence (object, attribute, and adverbial modifier) and try to find out on what principles their distinction is based and what objective criteria can be found to identify them. Apart from that there will be a special problem concerning the attribute.

The object is usually defined in some such way as this: It is a secondary part of the sentence, referring to a part of the sentence expressed by a verb, a noun, a substantival pronoun, an adjective, a numeral, or an adverb, and denoting a thing to which the action passes on, which is a result of the action, in reference to which an action is committed or a property is manifested, or denoting an action as object of another action.²

If we take a closer look at this definition, which is typical in its way, we shall find that it is based on two principles, namely (1) the relation of the object to a certain part of speech, (2) the meaning of the object, that is, the relation between the thing denoted and the action or property with which it is connected.

¹ See A. M. Пешковский, Русский синтаксис в научном освещении, гл. XIII.

² See, for example, Грамматика русского языка, т. II, ч. 1, стр. 523.

212 Secondary Parts

The first of these principles is syntactical, based on morphology (morphologico-syntactical), the second is semantic.

The first item of the definition practically means that an object can refer to any part of speech capable of being a part of the sentence. The second item enumerates certain semantic points in the relation between the thing denoted by the object and the action (or the property) with which it is connected.

We can at once note that the second item of the definition is incomplete, as it does not include a very important case, namely an object denoting the doer of the action with the predicate verb in the passive voice, as in the sentences He (Dickens) was asked by Chapman and Hall to write the letterpress for a series of sporting plates to be done by Robert Seymour who, however, died shortly after, and was succeeded by Hablot Browne (Phiz), who became the illustrator of most of Dickens's novels. (COUSIN) If Charles had inherited any of the qualities of the stern, fearless, not-tempered soldier who had been his father, they had been obliterated in childhood by the ladylike atmosphere in which he had been reared. (M. MITCHELL) (Compare similar cases in Russian.) This part of a sentence is usually called object and certainly is neither an attribute nor an adverbial modifier of any kind. And so, the list of meanings which the object may have should be enlarged by adding one more, which might be put like this — the thing (or person) which is the origin of the action (with the predicate verb in the passive voice).

However, this addition will not make the second item of the definition satisfactory. Formulated as it is, it shows that the object can mean one thing and another (the number of these meanings is 5 or 6), but the essential question remains unanswered: what is the meaning of an object, or, what is that which unites all those 5 or 6 meanings into one category, called the object? If the definition is to be satisfactory it is bound to formulate this invariable, as we may call it, which will appear in different shapes in each particular case where an object is found in a sentence. It is certainly far from easy to give a definition of this invariable, but probably it should run something like this — "a thing (or person) connected with a process or a property". Before, however, adopting any definition of this kind, we should carefully check it in as many sentences as possible. This difficulty will become clearer after we have considered the definitions of the other secondary parts — the attribute and the adverbial modifier.

We will now take up the definition of the attribute and consider it as we have considered that of the object. Besides, as we have already hinted, there is a special question concerning the status of the attribute in relation to other parts of a sentence.

_______________ Definition of the Adverbial Modifier______________ 218

The usual kind of definition of the attribute is this: It is a secondary part of the sentence modifying a part of the sentence expressed by a noun, a substantival pronoun, a cardinal numeral, and any substantivised word, and characterising the thing named by these words as to its quality or property. ¹

This definition, as well as that of the object, contains two items: (1) its syntactical relations to other parts of the sentence expressed by certain parts of speech, (2) its meaning. As far as meaning is concerned, the definition of the attribute is stricter than that of the object, as it practically mentions one meaning only (the terms "quality" and "property" may well be said to denote closely connected varieties of one basic notion). As far as the first item is concerned, the definition also gives a narrower scope of syntactic connections than was the case with the object: all parts of speech enumerated here are united by the notion of substantivity common to all of them.

If we now compare the definition of the attribute with that of the object we shall see at once that there are two main differences between them: (1) the attribute, as distinct from the object, cannot modify a verb, an adjective, or an adverb, and (2) the attribute expresses a property while the object expresses a thing.

They also have something in common: they both can modify a noun, a pronoun, and a numeral.

We shall see further on that the two definitions are not always a safe guide in distinguishing an attribute from an object.

Now let us consider the definition of an adverbial modifier. It may sound like this: It is a secondary part of the sentence modifying a part of the sentence expressed by a verb, a verbal noun, an adjective, or an adverb, and serving to characterise an action or a property as to its quality or intensity, or to indicate the way an action is done, the time, place, cause, purpose, or condition, with which the action or the manifestation of the quality is connected.² This definition is based on the same principles as two other definitions we have discussed: (1) the syntactical connection of an adverbial modifier with parts of the sentence expressed by certain parts of speech (and the list in this case is shorter than in either of the first two). (2) the meanings, which in this case are extremely varied, comprising no less than eight different items (and the list is not exhaustive at that). It appears to be particularly necessary here to look for an answer to the question: what does an adverbial modifier express, after all, so that the answer should not be a list, but just one notion — its invariable. It could probably be argued that all the meanings enumerated in the definition amount to one, viz.

¹ See Грамматика русского языка, т. II, ч. 1, стр. 522.

² See Грамматика русского языка, т. II, ч. 1. стр. 523.

214 Secondary Parts

the characteristic feature of an action or property. If we choose to put it that way, there remains only one "or" (action or property), which we might perhaps try to get rid of by looking for a term which would cover both actions and properties.

As to the first item of the definition, it has some peculiarities worth notice. In the list of parts of speech there is a point unparalleled in the other definitions which we have discussed, namely not every kind of noun can have an adverbial modifier, but only a verbal noun, that is, a noun expressing an action presented as a thing (see above, p. 28). So the characteristic features of a noun as a part of speech are not sufficient to enable it to have an adverbial modifier: it must belong to a certain lexical class.

Let us now proceed to compare the parts of speech enumerated in the definition of an adverbial modifier with those enumerated in the definitions of the other secondary parts of a sentence, and let us first take the list given in the definition of the object. There all the parts of speech capable of being parts of the sentence were given. So the parts of speech which the two lists have in common are, a verb, a verbal noun, an adjective, and an adverb.

If we now compare the list given for the adverbial modifier with that given for the attribute, we shall find that the only point which they have in common is the verbal noun: for the attribute it says "noun", which of course includes verbal nouns, and for the adverbial modifier it expressly says "verbal noun". Thus the sphere of overlapping between attributes and adverbial modifiers is very limited.

Summing up these comparisons we find that the first item of the definitions leaves room for ambiguity in the following cases: (1) if the part of the sentence which is modified is expressed by a noun, its modifier may be either an object or an attribute; (2) if it is expressed by a verbal noun, the modifier may be either an object, or an attribute, or an adverbial modifier; (3) if it is expressed by an adjective, the modifier may be either an object or an adverbial modifier; (4) if it is expressed by a cardinal numeral, the modifier may be either an object or an attribute; (5) if it is expressed by a verb,- the modifier may be either an object or an adverbial modifier; (6) if it is expressed by an adverb the modifier may be either an object or an adverbial modifier, too.

The above classification does not take into account the stative. If we add it to our list we shall get one more point: (7) if the part modified is expressed by a stative, the modifier may be either an object, or an adverbial modifier.

Since in these cases the first item of the definition does not lead to unambiguous results, we shall have to apply its second item, namely, the meaning of the modifier: a property, a thing, or whatever it may happen to be. For instance, if there is in the sentence a secon--

Doubtful Cases 215

dary part modifying the subject which is expressed by a noun, this secondary part may be either an object or an attribute. (It cannot be an adverbial modifier, which cannot modify a part of the sentence expressed by a non-verbal noun.) Now, to find out whether the secondary part in question is an object or an attribute we shall have to apply the second test and see whether it expresses a thing or a property. This would seem to be simple enough, but is far from being always so. In a considerable number of cases, the answer to the question whether a secondary part expresses a thing or a property will, strangely enough, be arbitrary, that is, it will depend on the scholar's opinion, and not on any binding objective facts.

Take, for instance, the following sentence: The dim gloom of drawn blinds and winter twilight closed about her. (M. MITCHELL) Here the phrase of drawn blinds and winter twilight modifies the noun gloom, which is the subject of the sentence. Since it modifies a noun it may be either an object or an attribute, and the choice between the two has to be made by deciding whether it denotes a thing (of whatever kind) or a property. How are we to decide that? On the one hand, it may be argued that it denotes a thing and its relation to the other thing, called gloom, is indicated. Then the phrase is an object. This view can hardly be disproved. On the other hand, however, it is also possible to regard the matter differently, and to assert that the phrase expresses a property of the gloom and is therefore an attribute. That view is quite plausible, too, and there is nothing in the facts of the language to show that it is wrong. So we shall have to choose the answer that seems to us to be the "better" one, i. e. to apply personal taste and opinion. The result thus gained will inevitably be subjective. The matter may also be settled by convention, that is, we may declare that we shall consider, say, every prepositional phrase modifying a noun to be an attribute (this is actually done in most English grammars).

From a theoretical point of view the distinction is of no particular interest, and a convention may be accepted without prejudice to the scientific level of our study. We might even say that in such circumstances the distinction between object and attribute is neutralised, and propose some new term, which would be more general than either "object" or "attribute". This indeed would probably be the best way of making syntactic theory agree with the actual facts. ¹

¹ We may suppose that that is the case in other spheres of syntax, too: if two views conflict in defining the essence of a certain phenomenon, and appear to be both of them right in their way, it is very likely that the distinction has been neutralised.

216Secondary Parts

A similar situation is also possible with the object and the adverbial modifier. This is the case, for instance, in a sentence like this: In her face were too sharply blended the delicate features of her mother, a coast aristocrat of French descent, and the heavy ones of her florid Irish father. (M. MITCHELL) The question is, what part of the sentence is in her face. As it modifies a verb it can be either an object or an adverbial modifier. To decide between these, we apply the second item of the definitions, and find out whether the secondary part expresses a thing or characterises the action. Now, obviously it is possible to take it in two ways; we might say that the secondary part of the sentence expresses an object affected by the action, or that it expresses a characteristic of the action itself. Accordingly whoever said that it was an object would have something on his side, and so would he who said that it was an adverbial modifier. In this case it seems rather more difficult than in the preceding case (see above) to settle the matter by convention. We could not possibly set down that any prepositional phrase modifying a verb is an object, since in many cases that would be clearly untrue. So the choice between object and adverbial modifier is bound to be arbitrary and subjective. The right way out of this is to say that in these positions the distinction between object and adverbial modifier is neutralised, and to propose some new term which would be more general than either "object" or "adverbial modifier". Since it is always difficult to invent a new term that will stand a reasonable chance of being generally accepted, it might perhaps be the best solution to use the term "secondary part" for all cases when the distinction between object and attribute, or that between object and adverbial modifier, is neutralised. If this view were accepted, we should have in some cases an object, in others an attribute, in others again an adverbial modifier, and last, sometimes a secondary part as such. ¹

In a few cases a similar doubt may be possible when we try to decide whether a phrase is an object without a preposition or an adverbial modifier, as in the sentence, He walks the streets of London. (THACKERAY , quoted by Poutsma) However, such cases appear to be very rare indeed.

After this preliminary survey of the three secondary parts, we will now proceed to examine each of them and their types.

¹ But see also below (p. 221 ff.) on the problems of the attribute.

Chapter XXVIII

SECONDARY PARTS IN DETAIL

THE OBJECT

It is well known that there are several types of objects and some kind of classification has to be found for them. Objects differ from one another, on the one hand, by their morphological composition, that is, by the parts of speech or phrases which perform the function of object, and on the other hand, in some cases objects modifying a part of the sentence expressed by a verb form (and that is most usually the predicate) differ by the type of their relation to the action expressed by the verb (it is to this difference that the terms "direct object" and "indirect object" are due).

Since the latter distinction applies only to a certain morphological type of objects, it will be convenient to take first the classification according to morphological differences.

From this point of view we must draw a distinction between non-prepositional and prepositional objects. Under the latter heading we will include every object of the type "preposition + noun or pronoun", no matter what preposition makes part of it, whether it be a preposition with a very concrete meaning, such as between, or a most abstract one, such as of or to. In establishing the two types of objects (non-prepositional and prepositional) we do not ask the question whether a prepositional object can or cannot be synonymous with a non-prepositional (as is the case with some objects containing the preposition to).

Both non-prepositoinal and prepositional objects (more especially the latter) may sometimes be hard to distinguish from adverbial modifiers (see above, pp. 215 and 216).

We will not attempt to give an exhaustive list of possible morphological types of non-prepositional objects but we will content ourselves with pointing out the essential ones.

These, then, are the important morphological types. An object may be expressed by a noun, a pronoun (of different types), a substantivised adjective, an infinitive, and a gerund. In some few special cases an object may be expressed by an adverb (as in the sentence We will leave here next week). ¹

The classification of objects into direct and indirect ones applies only to objects expressed by nouns or pronouns (and occasionally substantivised adjectives). It has no reference whatever to objects expressed by an infinitive, a gerund, or a phrase. With objects of these kinds the question whether they are direct or indirect would

¹ The so-called complex object and related phenomena will be studied in Chapter XXXII (see p. 257 ff.).

218 Secondary Parts in Detail

be meaningless. But even with objects expressed by nouns or pronouns the distinction is far from being always clear.

We will begin our study of direct and indirect objects by a type of sentence in which both objects are found simultaneously and no other interpretation of the facts seems possible.

A case in point are sentences in which the predicate is expressed by the verbs send, show, lend, give, and the like. These verbs usually take two different kinds of objects simultaneously: (1)an object expressing the thing which is sent, shown, lent, given, etc., and (2) the person or persons to whom the thing is sent, shown, lent, given, etc. The difference between the two relations is clear enough: the direct object denotes the thing immediately affected by the action denoted by the predicate verb, whereas the indirect object expresses the person towards whom the thing is moved. This is familiar in sentences like We sent them a present, You showed my friend your pictures, etc. It is well known that when the two objects occur together in a sentence, they are distinguished by their relative places in the sentence, that is, by word order: the indirect object stands first, and the direct object comes after it. ¹

However, even in sentences in which there are two objects simultaneously the distinction between direct and indirect objects is not always clear. With some verbs, and owing to their peculiar meanings, there are not sufficient objective facts to prove that one object is direct, and the other indirect. This is the case with the verbs tell and teach. They can take simultaneously two objects, one denoting the person addressed and the other the news told or the subject taught, as in the sentences, He told me the whole story, or She taught the children geography. So far the structure seems to be the same as in the above sentences with the verbs send, show, etc., and we might call the objects me and the children indirect, and the objects the whole story and geography direct. There is, however, something to be said against that view. The verbs tell and teach can also be used in a different way, as will be seen from the following sentences, He told me about his success, and She taught children. In the former sentence the first object denotes the person addressed but the second is expressed by a prepositional phrase and cannot be called a direct object; in the latter sentence there is no second object at all. Under these circumstances there would seem to be no reasonable objective ground for calling the first object in each of these sentences an indirect object.

There is another consideration here which rather tends toward the same conclusion. In studying different kinds of objects it is also essential to take into account the possibility of the correspond-

¹ We are not at the moment speaking of objects expressed by prepositional phrases.

The Object 219

ing passive construction. It is well known that in English there is a greater variety of possible passive constructions than in many other languages. For instance, the sentence We gave him a present can have two passive equivalents: A present was given to him (here the subject corresponds to the direct object in the active construction), and He was given a present (here the subject corresponds to the indirect object of the active construction). However, the second passive variant is only possible if the direct object is there, too. The sentence He was given in this sense (without the direct object) would not be possible. Now, with the verbs tell and teach things are different. It is quite possible to say The story has been told many times and I have been told about it (in this case the subject corresponds to the indirect object of the active construction, and there is no direct object in the sentence). In a similar way, it is possible to say Geography is taught by a new teacher and also Children are taught by a new teacher (without any direct object and indeed without any object corresponding to "geography"). From this point of view the sentences with the verbs tell and teach are different from those with the verbs send, show, give, etc. With the former there are not sufficient objective grounds for saying that one object is direct, and the other indirect.

As to sentences containing one object only, there are no grounds at all for saying that the object is "indirect". Sentences with the verb help are a case in point. In the sentence We will help our friends, for instance, there is nothing to show whether the object is direct or indirect. ¹

The object with verbs meaning 'to call by telephone or telegraph' is another case in point. We might suppose that the object with such verbs is indirect. The usual type of sentence, with the verb in the active voice, does not give any clue to this. For instance, in the following sentence there is nothing to show whether the object of telegraph is direct or indirect: "That's fine" she replied. "I'll telegraph Lee right away that I'm coming." (BUECHNER) But there are cases in which a verb of this category is used in the passive voice, e. g. Three days later, I was surprised to be rung up by Charles (SNOW) , that is, in the corresponding sentence with the predicate verb in the active voice, he rang me up, the object might equally be said to be a direct one.

Now, moreover, this question of direct and indirect objects is also connected with one type of object expressed by a phrase, namely the one of the pattern "to + noun or pronoun". It is common knowledge that the thought expressed in the sentence He gave me

¹ The fact that in Russian the corresponding verb помогать takes an indirect object, that is, a noun or pronoun in the dative case, is of course totally irrelevant here.

220Secondary Parts in Detail

a present can also be expressed in a slightly different way, namely, He gave a present to me. 1 We may call the first of the two objects direct because it stands in the same relation to the predicate verb as in the sentence He gave me a present. As to the second object, which includes a preposition, it is doubtful whether it will serve any useful purpose to call it an indirect object, since objects of the pattern "preposition + noun or pronoun" cannot be direct,² so that for objects of this kind there is no opposition of direct and indirect. If, however, we insist that the function of to me in this sentence is the same as that of me in the sentence He gave me a present, we shall have to include all prepositional objects under the heading of indirect objects and to change the system of classification which we have so far followed, in accordance with this view. It must be admitted that either way entails difficulties. If we follow the line adopted, we have to separate to me in the sentence He gave a present to me from me in the sentence He gave me a present; but then we can restrict the division of objects into direct and indirect to noun and pronoun objects (without preposition). If, on the other hand, we take to me to be an indirect object, we are obliged to extend the category of indirect objects to the prepositional ones; by way of compensation, we can keep up the connection between me in the sentence He gave me a present and to me in the sentence He gave a present to me. It would seem that, on the whole, the first alternative is preferable.

There is another question to be discussed concerning prepositional objects. Let us compare the following two sentences: We spoke about recent events, and We bought about twenty books. In the first sentence, the preposition denotes a relation between the action denoted by the verb and the thing denoted by the noun. The sentence is based on the pattern "speak about something". In the second sentence, the verb buy is not associated with a preposition: there is no pattern "buy about something". The word about does not denote any relation between the action and the thing, and bears in fact no relation at all to the verb. It is connected with the numeral only and shows that the number denoted by the numeral is not here given as exact. It is even doubtful whether the word about is here a preposition, as both its meaning and function are different from those of prepositions. ³ If we take this view, the object in this case will not be prepositional, and this is perhaps the best way out of the difficulty. If, however, we insist on the word about being a preposition we shall have to distinguish between two different types

¹ There is a difference in emphasis between the two sentences, but we Deed not dwell on it here.

² See, however, the next paragraph.

³ See above (p. 152) on the possibility of taking about as a particle.

The Attribute 221

of objects corresponding to the pattern "preposition + noun or pronoun"; a necessary feature of the type we are now considering would be the numeral preceding the noun, so that the pattern would be this: "preposition (about, over, under) + cardinal numeral + noun". A decisive point of difference between the types would be this. In type 1 (as in the sentence We spoke about recent events) the preposition cannot be left out: a sentence We spoke recent events is impossible. In type 2 the preposition can be left out without affecting the grammatical correctness of the sentence; only the idea of approximation conveyed by the word about in this context will disappear.

THE ATTRIBUTE

As we have already discussed the cases where the distinction between object and attribute is neutralised, so that a secondary part can equally be termed the one or the other (see above, p. 215), we need not dwell on these cases here but we can turn to the attribute as such.

An attribute can either precede or follow the noun it modifies. Accordingly we use the terms "prepositive" and "postpositive" attribute. The position of an attribute with respect to its head word depends partly on the morphological peculiarities of the attribute itself, and partly on stylistic factors.

We will discuss this question at some length in the chapter on word order (see pp. 246—247).

The size of a prepositive attributive phrase can be large in Modern English. This is mainly due to the fact that whatever is included between the article (definite or indefinite) and the noun, is apprehended as an attribute to the noun. Examples of attributes reaching considerable length are met with in usual literary (though not in colloquial) style. This is what we can see in the following sentence: The younger, Leander, was above all young, it seemed to him, charmingly, crashingly so, with only a slightly greater than usual grace and a deep reserve to distinguish him from any of his friends who had joined them. (BUECHNER) The phrase slightly greater than usual is characterised as an attribute by its position between the indefinite article and the noun grace, so that no. misunderstanding is possible here. Compare the following example: . . . her courage was not equal to a wish of exploring them after dinner, either by the fading light of the sky between six and seven o'clock, or by the yet more partial though stronger illumination of a treacherous lamp. (J. AUSTEN) The attributive group here is rather long (yet more partial though stronger) but it is held together by being placed between the definite article and the noun illumination. It is essential that no other noun appears between the article and

222 Secondary Parts in Detail

illumination. In this example we have even the subordinating conjunction though introducing the second attributive adjective stronger, so that the structure of the attributive group almost oversteps the limits of a clause. Compare also the following sentence from a modern novel: He was relieved when I motioned to him and started to wrap the by now almost insensible figure of Melissa in the soft Bokhara rug. (DURRELL)

Such attributes can acquire enormous proportions in humorous writings, so that whole sentences with subordinate clauses are squeezed into them, as in the following example (from an article containing criticism of the most common types of British crime films): Here are two possibilities only, and the threadbare variations are endlessly woven around them: the "I-ain't-askin'-no-questions-just-tell-me-what-to-do" kind and the "My-God,-Henry,-you-must-believe-me" kind (which can also be described as the "Why-the-devil-can't-you-leave-my-wife-alone-Can't-you-see-she's-distraught" kind). The hyphens connecting the various elements do not of course mean that the whole has coalesced into one monstrous word: they merely serve to show the unity of the syntactical formation functioning as an attribute. It goes without saying that such possibilities are due to the absence of inflections for number, gender, and case in the part of speech which most usually performs the function of an attribute, namely, the adjective.

This consideration brings us to what is the most difficult question in the study of the attribute, its position in the general system of parts of the sentence. The question is briefly this: is the attribute a secondary part of the sentence standing on a footing of equality with the object and the adverbial modifier, or is it a unit of a lower rank? Approached from another angle, the question would be this: is the attribute a constituent of the sentence, or does it belong to the level of phrases? This is of course a problem of general linguistics, and it has been discussed with reference to different languages. Here we will treat it taking into account the specific conditions of Modern English.

The problem can best be approached in the following way. If we take the sentence: History only emerged in the eighteenth century as a literary art. . . (MOULTON) and if we want to state the parts of the sentence, we shall stop at the phrase in the eighteenth century. We shall have to choose between two views: (1) in the century is an adverbial modifier of time; eighteenth is an attribute; the two secondary parts of the sentence stand on the same syntactical level; (2) in the eighteenth century is an adverbial modifier of time and is (as a whole) a secondary member of the sentence, modifying the predicate verb emerged; eighteenth is part of that adverbial modifier, which is expressed by a phrase, and it is part of the phrase, not of the sentence: it stands on a lower level than the

The Attribute 228

sentence with its parts, i. e. it stands on the phrase level, being an attribute to the noun century.

The same reasoning and the same choice would of course apply to the phrase as a literary art. The two possible views of its syntactic function would be these: (1) as a(n) art is a part of the sentence, namely a predicative; literary is another part, namely an attribute, standing syntactically on the same level with it; (2) as a literary art as a whole is a part of the sentence, namely, a predicative; literary is part of the predicative, and thus not a separate part of the sentence: it is part of the phrase, namely an attribute to the noun art, and stands on a lower level than the sentence and its parts: it stands on the phrase level.

To give another example, let us take the sentence In the rich brown atmosphere peculiar to back rooms in the mansion of a Forsyte, the Rembrandtesque effect of his great head, with its white hair, against the cushion of his high-backed seat, was spoiled by the moustache, which imparted a somewhat military look to his face. (GALSWORTHY) We will consider the following phrases: in the rich brown atmosphere; the Rembrandtesque effect; of his great head; with its white hair; a somewhat military look: With all of these the following two ways of analysis are possible: (1) in the atmosphere is an adverbial modifier of place, rich and brown are attributes — secondary parts of the sentence, on the same level as the adverbial modifier; the effect is the subject of the sentence, Rembrandtesque is an attribute — a secondary part of the sentence; with hair is an object, its and white are attributes; a look is an object, military an attribute, somewhat an adverbial modifier of degree, the last two being separate secondary parts and outside the object; (2) in the rich brown atmosphere is an adverbial modifier of place, rich and brown are parts of the phrase and, being attributes, stand on a lower level than secondary parts of the sentence; the Rembrandtesque effect is the subject of the sentence; Rembrandtesque, the attribute, is part of the phrase, not of the sentence as such; with its white hair is an object; white, the attribute, a part of the phrase; a somewhat military look is an object, military and somewhat are parts of the phrase, not of the sentence as such, military being an attribute to the noun look. ¹

There obviously is much to be said in favour of the view that the attribute in each case is a part of a phrase, rather than of the sentence. For one thing, it should be noted that in some cases the attribute cannot be left out without making the text grammatically incorrect. For instance, if we leave out the attributes his and great in the phrase of his great head, we shall get the impossible

¹ The function of the adverb somewhat and of other words in a similar position requires special discussion. See below, p. 224.

224 Secondary Parts in Detail

expression the effect of head. Then, in some cases, though the omission of an attribute would not make the construction wrong, it would deprive it of any reasonable sense, as in the end of our example, which would then run like this: . . .the moustache, which imparted a look to his face.

Against this latter point it may be argued that this is a semantic consideration which should have no influence on syntactic analysis, so that the point seems to remain doubtful. The first point seems more compelling, because it is strictly grammatical: the sentence without the attribute in question proves to be syntactically impossible.

Speaking more generally, the very fact that an attribute often comes within a part of the sentence (whether a main or a secondary one), for example, between the article and the noun to which the article belongs, and that in a number of cases it cannot be "extracted" without damaging the grammatical structure of the sentence, speaks strongly in favour of the view that the attribute stands on a lower level than the usual parts of the sentence (including the secondary ones) and that it should be considered a part of a phrase, not of a sentence. This view also gives the structure of the sentence a deeper perspective, as it opens up a syntactical sphere beyond that of parts of the sentence.

However, this view of the attribute also entails difficulties. To illustrate these, we may turn to the sentence from Galsworthy's "Man of Property" which we have just been considering. The end of the sentence runs like this: .. .which imparted a somewhat military look to his face. If we agree that the attribute military is not a separate part of the sentence but makes part of the phrase object whose centre is the noun look, this has its consequences for the adverb somewhat, which modifies the adjective military. If military is not a separate part of the sentence, somewhat obviously cannot be one either, as it is syntactically subordinate to a word which itself is not a part of the sentence. This leads to the conclusion that somewhat also makes part of the phrase of which look is the centre, and has to be treated accordingly. On the other hand, somewhat would seem to perform in this sentence a function similar to that which it performs in a sentence like His look was somewhat military, where military is the predicative, and somewhat an adverbial modifier belonging to it, and in this much a secondary part of the sentence. The functions of the word somewhat in the two sentences, though similar as far as its relation to its head word military is concerned, are different, according as the word military itself is a predicative or an attribute. It would seem to follow from this that a kind of double syntactic analysis is necessary. This question is a very difficult one indeed and a satisfactory solution has not so far been found.

The Adverbial Modifier 225

THE ADVERBIAL MODIFIER

We must begin by stating that the term "adverbial modifier" cannot be said to be a very happy one, as it is apt to convey erroneous ideas about the essence of this secondary part. The word "adverbial" may give rise to two notions, both of them wrong. For one thing, we may suppose that an adverbial modifier is always expressed by an adverb, which of course is not true: an adverbial modifier may be expressed by different morphological means. Secondly, the term "adverbial" may give rise to the notion that an adverbial modifier always modifies a verb, which is also wrong! an adverbial modifier may modify a part of the sentence expressed by an adjective or by an adverb, as well as by a verb. As the term "adverbial modifier" is firmly established, it would be futile to try and substitute another term in its place. So we will keep the term, bearing in mind what has been said about its meaning.

There are several ways of classifying adverbial modifiers: (1) according to their meaning, (2) according to their morphological peculiarities, (3) according to the type of their head word.

Of these, the classification according to meaning is not in itself a grammatical classification. For instance, the difference between an adverbial modifier of place and one of time is basically semantic and depends on the lexical meaning of the words functioning as adverbial modifiers. However, this classification may acquire some grammatical significance, especially when we analyse word order in a sentence and one semantic type of adverbial modifier proves to differ in this respect from another. Therefore the classification of adverbial modifiers according to their meaning cannot be ignored by syntactic theory.

Classification according to morphological peculiarities, i. e. according to the parts of speech and to phrase patterns, is essential: it has also something to do with word order, and stands in a certain relation to the classification according to meaning.

Classification according to the element modified is the syntactic classification proper. It is of course connected in some ways with the classification according to meaning; for instance, an adverbial modifier can modify a part of the sentence expressed by a verb only if the type of meaning of the word (or phrase) acting as modifier is compatible with the meaning of a verb, etc.

A complete classification of adverbial modifiers according to their meaning, i. e. a list of all possible meanings they can have, is impossible to achieve, and it would serve no useful purpose. A certain number of meanings can be found quite easily, such as place, time, condition, manner of an action, degree of a property, etc., but whatever list we may compile along these lines, there are bound to be special cases which will not fit in. For instance, in the sentence

226 Secondary Parts in Detail

I saw him at the concert it is hard to tell whether the adverbial modifier at the concert expresses place or time; and the dilemma appears to be futile. Since all this depends on the lexical meanings of words, possibilities here are practically boundless. We must therefore content ourselves with establishing some main categories and abstain from trying to squeeze every single adverbial modifier that may occur in a sentence into a "pigeonhole" prepared for it.

As to the classification according to morphological peculiarities, it can probably be made exhaustive, although some of the morphological types are met with very seldom indeed.

The most usual morphological type seems to be the adverb. This is testified, among other things, by the fact that the very term for this part of the sentence is derived (in English, and also, for instance, in German) from the term "adverb". In some grammar books the two notions are even mixed up. Occasionally an author speaks of adverbs, where he obviously means adverbial modifiers. ¹

Another very frequent morphological type of adverbial modifier is the phrase pattern "preposition + noun" (also the type "preposition + adjective + noun" and other variations of this kind). This typ