The problem of a middle voice

This problem arises chiefly in connection with the possible double use of a number of verbs in Modern English. Compare, for instance, such pairs of sentences as these:

I opened the door The door opened

I burnt the paper The paper burnt

I boiled the water The water boiled

We resumed the conference The conference resumed

We apply the rule to many The rule applies to many

cases cases

First let us formulate what is established and does not depend on anybody's point of view or interpretation, and then we will proceed to analyse the questions which admit of different solutions.

The facts, then, are these. In the sentences of the first and in those of the second column we have verb forms sounding alike but differing from each other in two important points:

(1) In the first column, the verb denotes an action which is performed by the doer on an object in such a way that a change is brought about in that object, for instance, the door was closed and then I acted in such a way that the door became open; the paper was intact, but I subjected it to the action of fire, and it was reduced to ashes, etc.

In the second column a process is stated which is going on in the subject itself: the door opened (as if of its own will), the paper disappeared in flames, etc. Compare, e. g., His camp had filled. (LINKLATER) The teas making. (L. MITCHELL)

120 The Verb: Voice

This, of course, is a difference in the relation between the subject and the action (and, for the first column, the object).

(2) In the first column, the verb is followed by a noun (or pronoun) denoting the thing which is subjected to the action denoted by the verb. In the second column, the verb is not followed by any noun (or pronoun). In the first column the verb is transitive, in the second column the verb is intransitive.

What we have said so far is nothing but an objective description of the state of things found in these sentences, no matter what theory a scholar may prefer.

Now we must turn our attention to the possible theoretical interpretation of these facts, and here the problem of voice will arise.

One possible interpretation is this. In every line we have in the two columns two different verbs which may be represented in some such way as: open₁, verb transitive, open₂, verb intransitive; burn₁, verb transitive, burn₂, verb intransitive, etc. If this interpretation were adopted, the whole problem would be shifted into the sphere of lexicology, and from the grammatical viewpoint we should have to state that open₁ here stands in the active voice (correlative with was opened), and open₂ has no voice distinction at all (since from the intransitive verb open₂ no mutually opposed voice forms can be derived).

Another interpretation would run something like this. In both columns we have the same verb open, the same verb burn, etc. and the difference between the two is a difference of voice: in the first column it is the active voice (showing an action performed by the doer on the object), while in the second column it is the middle voice, denoting a process going on within the subject, without affecting any object. The difference between the voices, though not expressed by any morphological signs, would then be a difference in meaning and in syntactical construction, the active voice characterised by connection with a following noun or pronoun denoting the object of the action, and the middle voice characterised by the impossibility of connection with such a noun or pronoun. This interpretation would mean the admission of a special voice, the middle voice.

Still another interpretation would be the following. The verb in both columns is the same and the voice is the same, too, since there is no morphological difference between the two columns, and differences of meaning and of syntactical construction are not sufficient reason for establishing a difference of voice. If this view is accepted, we should have to define the category of active voice in such a way that it should include both the first-column and the second-column examples.

The choice between these interpretations depends on the principles which a scholar considers to be the most essential and the

The Problem of a Middle Voice 12t

most likely to yield an adequate picture of language facts. If, for instance, it is considered essential that a difference in grammatical categories should find its outward expression by some morpheme, etc., the second of the three suggested interpretations will have to be rejected. If, on the other hand, it is considered possible for two morphological categories to be distinguished in meaning and syntactical use without any special morphemes to show the distinction, that second interpretation will be found acceptable.

Without prejudice to the first or second interpretation, we will now follow up the third, which seems to present the greatest interest from a theoretical point of view. In doing so, we will assume that we do not accept either a reflexive or a reciprocal or a middle voice, so that only two voices are left, the active and the passive. If, then, we are to bring under the heading of the active voice such cases as the door opened, the paper burnt, the water boiled, etc., we shall have to give that voice a definition wide enough to include all uses of that kind as well (this may make it necessary to change the term for the voice, too).

Let us now consider the opposition between the voices: opened (in any sense)/ was opened; burnt (in any sense)/ was burnt from the point of view of meaning. It should at once be clear that the second member of the opposition (was opened, etc.) has a much more definite meaning than the first: the meaning of the type was opened is that the subject is represented as acted upon, whereas the meaning of the first member (opened, etc.) is much less definite. We could, then, say that opened is the unmarked, and was opened, the marked member of the opposition. The meaning of the unmarked member is, as has often been the case, hard to define. What seems the essential point in its meaning is, that the subject is represented as connected with the origin of the action, and not merely acted upon from the outside. Some such definition would seem to cover both the type he opened the door, and the type the door opened. Whether the subject produces a change in an object, or whether the action is limited to the sphere of the subject itself — all these and similar points would depend partly on the syntactical context (on whether the verb is followed by a noun / pronoun or not), partly on the lexical meaning of the verb and its relation to the lexical meaning of the noun expressing the subject (compare the old man opened... and the door opened), partly, probably, on a number of other factors which are yet to be studied. The question whether it is more advisable to keep the term "active voice" or to substitute another term for it would also have to be discussed.

If this view is adopted, all the special cases considered above: he shaved (in the reflexive meaning), they kissed (in the reciprocal meaning) would fall under the heading of the active voice (if this

122The Verb: Voice

term is kept) and their peculiarities would have to be referred to the context, the lexical meaning of the verb in question, etc.

The following phenomena would also belong here: the book sells well, the figures would not add, the rule does not apply in this case (as different from we do not apply the rule), and a number of others, which have been variously treated as ''absolute use", ¹ use of the active form in a passive meaning, etc.

As to form, it has been already said above (p. 115) that the passive is the marked, and the active the unmarked member of the opposition. Thus, then, the passive is marked both in meaning and in form and the active as unmarked both in meaning and in form.

This solution of the voice problem in Modern English appears to be convincing. However the other interpretations (mentioned above as first and second) ought also to be reasoned out to their logical conclusions.

¹ See M. Deutschbein, System der neuenglischen Syntax, S. 101, on such cases as the work does not pay.

Chapter XIII

THE VERB: PERSON AND NUMBER. OTHER MORPHOLOGICAL CATEGORIES

The categories of person and number must be considered in close connection with each other, since in language of the Indo-European family they are expressed simultaneously, i. e. a morpheme expressing person also expresses number, e. g. in Latin the morpheme -nt in such forms as amant, habent, legunt, amabant, habebunt, legerunt, etc., expresses simultaneously the 3rd person and the plural number.

We shall, however, start by considering the meaning of each of these categories, and then proceed to the analysis of their state in Modern English.

The category of person in verbs is represented by the 1st, 2nd, and 3rd person, and it expresses the relation between the speaker, the person or persons addressed, and other persons and things. The 1st person, of course, expresses the speaker or a group of which the speaker makes a part; the 2nd person, the person or persons spoken to, and the 3rd, that person or thing (or those persons or things) which are neither the speaker nor the person(s) spoken to. ¹

The category of number expresses the quantity of the subjects (one or more than one). Speaking deductively, we might build the following system of personal and numerical categories:

1st person singular — the speaker

2nd person singular—one person spoken to

3rd person singular — one person or thing (neither

speaker nor spoken to)

1st person plural — the speaker and another person or other persons

2nd person plural — more than one person spoken to 3rd person plural — more than one person or

— thing (neither speakers nor spoken to)

However, this system does not hold good for the Modern English verb, and this for two reasons,

First, there is no distinction of persons in the plural number. Thus, the form live may, within the plural number, be connected with a subject of any person (1st, 2nd, or 3rd).

¹ It will certainly not do to say that the 3rd person represents that which is spoken about. E.g. in the sentence You must come at once, you is the person spoken about but it is not the 3rd person.

124 The Verb: Person and Number

Second, there is no distinction of numbers in the 1st or 2nd person. Thus, the form live inthese persons may refer both to one and to more than one subject. ^l

So what we actually find in the Modern English verb is this:

3rd person singular — lives All the rest — live

If we analyse this state of things in the Modern English verb in exact terms we shall reach the following conclusion. The opposition lives I live, or, in general terms, stem + s / stem + Ø, expresses the relation: 3rd person singular / any person of both numbers except 3rd person singular.

It is quite clear that the first item of the opposition is marked both in meaning (3rd person sing.) and in form (-s), whereas the second item is unmarked both in meaning (everything except the 3rd person sing.) and in form (zero-inflection). We ought to add that the category of mood is implied in this opposition, the form lives belonging to the indicative mood only, whereas live may also be any person of both numbers in the subjunctive mood (as far as we recognise its existence at all). Another consequence of this analysis is, that the -s-inflection in verbs conveys 4 meanings: 1) 3rd person, 2) singular number, 3) present tense, 4) indicative mood. The present tense is of course characterised by other signs as well: by the absence of the -d (or -t) morpheme denoting the past tense in regular verbs, and by alternation of the root vowel (e. g. [ı] in drinks as against [ae] in drank) in irregular verbs. But in verbs of the type put the -s is the only distinctive sign of the present.

The ending -s having four meanings to express simultaneously is of course a synthetic feature, standing rather by itself in the general structure of Modern English.

Some verbs do not fit into the system of person and number described above and they must be mentioned separately both in a practical study of the language and in theoretical analysis. We will limit ourselves to the verb can (the verbs may, shall, and some others sharing some of its features) and the verb be, which stands quite apart and, of course, is very widely used.

The verb can, as is well known, takes no -s-inflection parallel to such forms as lives, writes, takes, etc. Hence it follows that this verb has no category of person or number at all.

¹ We do not consider here the forms livest, livedst, etc., which do not make part of the grammatical system of literary and colloquial English. See p. 125.

The verb be has a system of its own both in the present indicative and in the past. Its system in the present indicative is as follows:

1st person singular 3rd person singular	— am — is
2nd person (without distinction of number) Plural (without distinction of person)	— are
In the past tense the system is:
1st and 3rd person singular	— was
2nd person (without distinction of number) Plural (without distinction of person)	— were

In analysing the system of person and number we have so far bypassed the forms of the type livest, takest, livedst, tookest. These forms are associated with the personal pronoun thou and are only used in religions and occasionally in poetical texts and among Quakers. As they stand outside the received grammatical system we need not go into details concerning them. Suffice it to say that with these forms the category of number appears within the category of the 2nd person and the whole system of person and number (including the past tense) must be presented in a different shape.

OTHER MORPHOLOGICAL CATEGORIES Negative Forms

The English language has in its verbal system a peculiarity distinguishing it both from Russian, German, French, and other Indo-European languages. To express the notion that an action did not take place, the English verb does not always simply add a negative particle to the verb form, as in the example has come — has not come. In many cases a special auxiliary verb, namely the verb do, is used if the negative idea is to be expressed. ^lSince the negative has (at least partly) its own auxiliary verb, it must be acknowledged as a special morphological category of the English verb.

This fact has of course been observed a long time since, and attempts have been made to tackle it. Academician A. Shakhmatov,

¹ Something broadly similar is found in some Finno-Ugrian languages, e.g. Finnish and Estonian.

126The Verb: Person and Number

comparing the Russian negation and the English, pointed out that there is a special auxiliary for the negative in English, and put forward the idea that in English there is a special negative mood. ^l This idea, however, cannot be accepted by modern linguistics, as the negative forms may be found in every mood: compare, for instance, does not take, do not take (!) (imperative). Since the negative is compatible with different moods, it cannot itself be a mood. In other words, if the opposition takes — take (!) is an opposition of mood, the opposition lakes — does not take cannot also be an opposition of mood. The opposition takes — does not take must be based on some other category, whose concrete manifestations are the affirmative and the negative. It is hard to find a name for this general category. Perhaps we might term it "quality". ² Of its two components (affirmative and negative) the former is unmarked and the latter is marked: its marks are the group "do + not" in some forms and the particle not alone in others.

As the auxiliary do appears in some negative forms only, it might be argued that the category of quality is found only in these forms. However, it seems preferable to state this category for the English verb as a whole, and to differentiate the means of expressing it into "do + not" and not (alone).

We need not give here a full list of forms in which the one or the other of these means is used to express negativity. The use of the pattern "do + not" is restricted to the forms which have no other auxiliary of any kind. That is, the auxiliary do is incompatible with any other auxiliary verb.

Interrogative Forms

An important question arises concerning the interrogative forms of the English verb. It is well known that the auxiliary do is used here in the same way as in the negative forms and that interrogative-negative forms exist, in which the auxiliary do is used on the same principle.

Since the verb do is an auxiliary to form the interrogative, we must conclude that the opposition between declarative and interrogative forms (e. g. takes — does .. . take?) is also based on some grammatical category, which is no less difficult to define and to give a name to. We might perhaps think that the interrogative should be included as a third item in the opposition "affirmative — negative", thus forming a triple grouping "affirmative — interrogative — negative". But this is rendered impossible by the fact that in-

¹ See А. А. Шахматов, Синтаксис русского языка, 1941, стр. 482.

² The term "quality" is used in logic to distinguish between affirmative and negative propositions.

Other Morphological Categories 127

terrogative and negative can be united in one form, as in does ... not take? Since interrogative and negative can be combined in one form, they cannot possibly belong to the same category but have to be assigned to different categories. We may put the four categories: affirmative, negative, interrogative, and interrogative-negative together in the following diagram:

Non-interrogative Interrogative

Non-negative takes does ... take?

Negative does not take does .. . not take?

The diagram, simple as it is, shows that we have here a system of 2 X 2 categories completing each other. The peculiar thing is, that only one of the four forms does not include the auxiliary do, and that the two items of the second line differ from each other only by word order, while the two items of the first line differ from each other by the use or non-use of the auxiliary verb.

The question may be asked: what is the meaning of the auxiliary do in the negative-interrogative form does ... not take? Is it an auxiliary of interrogation or an auxiliary of negation, or does it combine the two meanings? There seem to be no objective criteria in this matter, and if a somewhat subjective view may be expressed, we will say that the auxiliary do in the negative-interrogative form combines both meanings.

However, the whole problem of negative and interrogative forms of the English verb requires some deeper investigation.

Emphatic do-forms

Another question arises concerning the so-called emphatic do-forms, such as he does know, she did go, meaning more or less the same as he really knows, she really went, etc. The specific meaning of such formations is well known, but their status in the morphological system of the verb has not been clearly defined.

In the first place, we must find out whether the verb do does or does not introduce any lexical meaning of its own into the formation. Apparently it does not: it merely emphasises the meanings expressed by the infinitive following the form of the verb do. If this view is endorsed, we must conclude that these are analytical verb forms, that is, the verb do is an auxiliary verb here just as it is an auxiliary in the negative and interrogative formations of which it is a necessary component. If that is so, the opposition between knows and does know, or that between went and did go, etc., must be based on some grammatical category or other. It is also evident that the forms does know, did go, etc., are the marked members of the opposition, while the forms knows, went, etc., are its

128 The Verb: Person and Number

unmarked members. This is obvious both from the meaning and the form of each member: does know, did go, etc. are necessarily emphatic and they have the auxiliary as a means of expressing emphasis, that is, they cannot be used unemphatically; knows, went, etc., on the other hand, are not necessarily unemphatic: they may very well become emphatic if pronounced with the appropriate intonation, even though they have no special auxiliary or any other material sign to mark them off. The category which lies at the basis of this opposition may perhaps be briefly termed emphasis.

It should also be noted that the do-forms do not cover the entire field of the English verb: they are only found in the finite verb form (thus not in the infinitive, participle, or gerund), and only in those which have no auxiliary in the unemphatic form.

We may add that for all those forms of the verb which do not fall under this definition the way to express emphasis is purely phonetic: the verb form is pronounced with strong stress; in writing the form is usually underlined, and in print it is given in italics.

The auxiliary do is also occasionally used as a kind of homogeneous part parallel to a modal verb and marking the reality of the action denoted by the following infinitive, as distinct from, and opposed to, its mere possibility or necessity, etc., expressed by the modal verb. Here is an example of this use: Life could and did go on almost as usual. (M. MITCHELL)

Hierarchy of Verbal Categories

It is natural to assume that in the system of verbal categories there is some hierarchy, that is, some categories are above others, determining their possibilities. To give a clear example: the category of voice to some extent dominates that of aspect, as there are fewer continuous forms in the passive than in the active voice: such continuous forms as shall be writing, have been writing, had been writing find no counterpart in the passive. We could also say that the passive voice limits the possibilities of the continuous aspect. ¹

The category of mood, as we take it, dominates the category of tense. In the indicative mood there are (at least) three tenses, whereas in the "oblique" moods there are at any rate not more than two, and the imperative mood has no tenses at all.

¹ In this the English language fundamentally differs from Russian, where the category of aspect dominates. As Academician V. Vinogradov puts it, the category of aspect dissects the entire system of the Russian verb (see В. В. Виноградов, Русский язык. М., 1947, стр. 493). Thus, in the imperfective aspect in the indicative mood there are three tenses, while in the perfective there are only two.

Other Morphological Categories 129

A peculiar relation obtains between the categories of number and of person. Leaving aside for the moment the verb be with its individual system of forms, number and person of English verbs have a positive (that is, non-zero) expression only in the -s-ending of the 3rd person singular present indicative. We might even suppose that in Modern English there are not two separate categories, number and person, but one "combined" number-person category. It is, however, doubtful whether such interpretation of phenomena would in any way yield a clearer and more consistent view of the verbal system. The notions of "number", that is, the difference between one and more-than-one doer, on the one hand, and that of "person", that is, distinction between the speaker, the one spoken to, and that which is neither speaker nor spoken to, seems too far apart, to be united under a common heading.

In pursuing this subject further, it should be possible to work out a system of verbal categories, something of a "pyramid"; however, there would probably arise some doubts and difficulties in assigning a place to this or that category.

б в. A. Ильиш

Chapter XIV

THE VERB: VERBALS

In so far as the verbals (infinitive, gerund, and participle) make up a part of the English verb system, they have some features in common with the finite forms, and in so far as they are singled out amid the forms of the verb, they must have some peculiarities of their own.

Let us first consider the system of verbal categories which are expressed in the English verbals. They have some of them, and they lack some others. We must also observe that it is by no means certain in advance that all the verbals are in the same position as regards the verb categories.

It is clear that none of the verbals has any category of person or mood. The English verbals have no category of number either, though this is not so in some other languages. What we must examine is the categories of aspect, tense, correlation, and voice.

With reference to aspect we shall have to examine each of the verbals separately.

In the infinitive, we find an opposition between two sets of forms:

(to) speak — (to) be speaking

(to) have spoken— (to) have been speaking,

which is obviously the same as the opposition in the sphere of finite forms between:

speak — am speaking spoke — was speaking

etc.

The conclusion here is quite obvious: the infinitive has the category of aspect, viz. there is a distinction between the common and the continuous aspect. The continuous infinitive is found, for example, in the following sentence: He seems to be enjoying himself quite a lot. (R. WEST)

In our next example the continuous infinitive of the verb love is used: I can recollect yet how I loved him; and can dimly imagine I could still be loving him if — No, no! (E. BRONTE) The variant with the simple infinitive would be: I can recollect yet how I loved him; and can dimly imagine I could still love him, if — The difference in this case seems to be that the continuous infinitive gives more prominence to the idea of the continuity of her love, and this is obviously much stronger than the mere statement that love might still be there now. The stylistic difference is thus unquestionable, but there would seem to be also a grammatical difference. The meaning of the continuous aspect is well brought out here, though the lexical meaning of the verb love would seem to go against it.

Tense and Correlation 131

With the gerund and the participle, on the other hand, things are different. Generally speaking, they exhibit no such distinction. Neither in the one nor in the other do we find continuous forms.

Occasionally, however, a continuous participle is found, as in the following sentence from a novel by Jane Austen: The younger Miss Thorpes being also dancing, Catherine was left to the mercy of Mrs Thorpe and Mrs Allen, between whom she now remained. It is not clear here what exactly is added to the meaning of the sentence by using the continuous participle being dancing rather than the usual participle dancing. Be that as it may, this example shows that a continuous first participle is at least potentially a part of the morphological system of the English verb. But this use appears to be obsolete.

In the following sentence there are even three continuous participles, with one auxiliary common to all of them: Catherine had no leisure for speech, being at once blushing, tying her gown, and forming wise resolutions with the most violent dispatch. (J. AUSTEN) The word order (the phrase at once coming after the auxiliary being) clearly shows that the auxiliary belongs to all three participles (blushing, tying, and forming). The use of the continuous participles seems to be a means of giving prominence to the fact that the actions indicated were actually happening at that very moment.

TENSE AND CORRELATION

The problem of the category of tense and that of correlation have to be considered together, for reasons which will become clear immediately.

In the infinitive, we find the following oppositions:

(to) speak — (to) have spoken

(to) be speaking— (to) have been speaking,

and in the gerund and the participle the oppositions

speaking — having spoken being spoken — having been spoken

The question now is, what category is at the base of these oppositions?

The considerations which can be put forward in this matter might be compared to those which were applied to similar phenomena in the forms should speak — should have spoken, but here everything is much simpler. If we start from the way these forms are derived we shall say that it is the category of correlation which finds its expression here, the first-column forms having no pattern "have + second participle" and the second-column forms having this very pattern. If we turn to the meaning of the second-column

5*

132 The Verb: Verbals

forms, we shall find that they express precedence, whereas the first-column forms do not express it. Once again we see that in each pair one item is unmarked both in meaning and in form whereas the other (the perfect) is marked both in meaning (expressing precedence) and in form (consisting of the pattern "have + second participle").

If this view is accepted it follows that the category of correlation is much more universal in the Modern English verb than that of tense: correlation appears in all forms of the English verb, both finite and non-finite, except the imperative, while tense is only found in the indicative mood and nowhere else.

Since the verbals are hardly ever the predicate of a sentence, they do not express the category of tense in the way the finite verb forms do. Thus, it seems pointless to argue that there is a present and a past tense in the system of verbals.

We will therefore endorse the view that the opposition between (to) speak and (to) have spoken, and that between speaking and having spoken is based on the category of correlation.

VOICE

Like the finite forms of the verb, the verbals have a distinction between active and passive, as will readily be seen from the following oppositions:

(to) read — (to) be read

(to) have read — (to) have been read reading — being read

having read — having been read

As to other possible voices (reflexive, reciprocal, and middle) there is no reason whatever to treat the verbals in a different way from the finite forms. Thus, if we deny the existence of these voices in the finite forms, we must also deny it in the verbals.

To sum up, then, what we have found out concerning the categories in the verbals, we can say that all of them have the categories of correlation and voice; the infinitive, in addition, has the category of aspect. None of the verbals has the categories of tense, mood, person, or number.

THE SECOND PARTICIPLE

The second participle, that is, forms like invited, liked, written, taken, etc., presents many peculiar difficulties for analysis. In analysing the category of correlation and that of voice in the participle and in stating that the participle has no category of tense, we have so far not mentioned these forms at all.

The Second Participle 188

Now we must give them some special consideration.

First of all we must emphasise that we will analyse the meaning and the use of the second participle when it does not make part of an analytical verb form, whether it be the perfect (have invited, have taken), or the passive voice (was invited, was taken). When the second participle makes part of an analytical form, it loses some of its own characteristics, and indeed we may doubt whether it should still bear the name of participle in those cases.

Again, in analysing the meaning and the functions of the second participle, we must exclude the cases where it has been adjectivised, that is, changed into an adjective, and is no longer a participle, for example, in such phrases as written work, which is used as the opposite of oral work, or devoted friend, where devoted does not designate an action, or, indeed, the result of an action, but a property.

The use of the second participle outside the analytical formations is comparatively limited. We find it either as a predicative in such cases as The door is shut, when it does not denote an action (compare, The door is shut at nine p. m. every day) but a state of things, or as an objective predicative, e. g. He found the door shut, or as an attribute following a noun, more often with some words accompanying it, as in This is the new machine invented by our engineers, and less often an attribute preceding the noun, as in "The Bartered Bride" (the title of Smetana's opera). We can note that the use of second participles as prepositive attributes is on the whole limited in English. For example, the title of the opera just mentioned could not be rendered in English with the help of the participle sold, as this participle cannot be used in that way.

Analysis of the grammatical categories expressed in the second participle is a matter of great difficulty, and so is the problem of finding its place among the other participles.

Let us first consider the problems of aspect, tense, and correlation with reference to this participle. Let us take our examples with intransitive verbs, so that the problem of voice may be left aside for the moment.

It was pointed out long ago that many intransitive verbs have no second participle that could be used outside the analytical forms of the perfect. For instance, such forms as been, laughed, run, sat, lain, wept, etc. can only appear within a perfect form and do not exist as separate participles. A few second participles of intransitive verbs can, however, be used as attributes, e. g. retired in expressions like a retired colonel, or a retired teacher. We may also compare the word runaway (spelt as one word, from the phrase run away), for example, in the expression a runaway horse.

On the whole, then, with intransitive verbs the second participle does not constitute an integral part of the verbal system at all,

134 The Verb: Verbals

and it may be left out of consideration when we analyse that system.

Things are different with transitive verbs. Here, though the use of the second participle as an attribute is limited, there can be no doubt that it exists as a separate form of the verb and not merely as a component of the analytical perfect or passive.

It is also clear that as far as the category of voice goes the past participle of transitive verbs belongs to the passive. We need not illustrate this by examples, since this is common knowledge. It is only necessary to mention the few special cases in which the second participle has no passive meaning in the usual sense, e. g. a well-read man 'one who has read much', not 'one who has been read', or he was drunk, and a few more. These are separate phenomena restricted to a few verbs.

As to aspect, tense, and correlation, the problem appears to be this: Which of these categories find expression in the form of the second participle itself, i. e. do not depend either on the lexical meaning of the verb or on the context? This proviso is necessary, because differences in meaning can be found which do depend on lexical peculiarities of the verb and on the context. We can, for instance, compare such phrases as the following: (1) a young man liked by everybody, (2) a young man killed in the war. It is clear at once that the action denoted by the participle liked is going on, whereas that denoted by the participle killed is finished. This certainly should not be interpreted as two different meanings of the participle as a grammatical form, since it depends on the lexical meaning of the verb (the verb like denotes an emotional attitude, which can last indefinitely, while the verb kill denotes an action which reaches its end and does not last after that). We must then say that the meaning of the form as such is not affected by these differences.

The conclusion about the grammatical categories in the second participle (of transitive verbs) is, then, this. The only category which is expressed in it is that of voice (namely, the passive voice); the other categories, namely, aspect, tense, and correlation (and, of course, mood, person, and number) find no expression in it. Owing to these peculiarities, the second participle occupies a unique position in the verbal system, and it is impossible to find for it a place in a table where special columns or lines are allotted to aspect, tense, and correlation.

As far as voice is concerned, the second participle of transitive verbs (e. g. invited) joins the other passive participles (e. g. being invited and having been invited) as against the active participles inviting and having invited. However, from the formal point of view we run into difficulties here. In all other passive forms, whether finite or non-finite, the category of the passive voice is ex-

The ing-Forms < 185

pressed by the group "be + second participle", whereas the second participle itself, of course, goes without the verb be. We have to choose between accepting this state of things and excluding the second participle from the passive system (that is, if we insist that every passive form must contain the verb be). As this latter alternative appears to be still more undesirable, we shall have to recognise this peculiar position of the second participle among the forms of the passive voice.

THE ing-FORMS

So far we have spoken of the ing-forms as of two different sets of homonymous forms: the gerund (with its distinctions of correlation and voice) and the participle (with its distinctions of correlation and voice). As there is no external difference between the two sets (they are complete homonyms), the question may arise whether there is reason enough to say that there are two different sets of forms, that is, whether it could not be argued that there is only one set of forms (we might then call them ing-forms), which in different contexts acquire different shades of meaning and perform different syntactical functions. Such a view (though without detailed argumentation) was indeed put forward by the Dutch scholar E. Kruisinga.¹ In some passages of his book he merely speaks of "the ing", though in other parts he uses the terms "gerund" and "participle".

It must be said that this is one of the questions which do not admit of a definite solution. The solution largely depends on what view we take of the unity of a grammatical form and on the extent to which we are prepared to allow for shades of meaning in one form (or one set of forms). If we are prepared to admit any amount of variety in this sphere rather than admit the existence of grammatical homonyms, we shall have to develop a detailed theory of the mutual relations between the various shades of meaning that the form (or set of forms) can have. If, on the other hand, we are prepared to admit homonymy rather than let the unity of the form (or set of forms) disintegrate, as it were, in a variety of "shades", we shall be justified in keeping to the traditional view which distinguishes between gerund and participle as between two different, though homonymous, sets of grammatical forms.

The difference between the gerund and the participle is basically this. The gerund, along with its verbal qualities, has substantival qualities as well; the participle, along with its verbal qualities, has adjectival qualities. This of course brings about a corresponding difference in their syntactical functions: the gerund

¹ See E. Kruisinga, A Handbook of Present-Day English, vol. II, p. 55 II,

136The Verb: Verbals

may be the subject or the object in a sentence, and only rarely an attribute, whereas the participle is an attribute first and foremost.

We should also bear in mind that in certain syntactical contexts the difference tends to be obliterated. For instance, if in the sentence Do you mind my smoking? (where smoking is a gerund) we substitute me for my, in the resulting sentence Do you mind me smoking? the form smoking may, at least, be said to be the participle. Again, in the sentence Do you mind her smoking? where her may be the possessive pronoun, corresponding to my, or the objective case of the personal pronoun, corresponding to me, the gerund and the participle are practically indistinguishable. We may say, in terms of modern linguistics, that the opposition between them is neutralised.¹

If, on the other hand, we prefer to abandon the distinction and to speak of the ing-form, we shall have to formulate its meaning and its functions in such a way as to allow for all the cases of the ing-forms to be included. For instance, instead of distinguishing between substantival and adjectival qualities, we shall speak, in a more general way, of nominal qualities, so as to embrace both the substantival and the adjectival ones, and so forth. Such a view seems also quite possible, and the decision to be taken will, as we have seen above, depend on the general attitude one adopts in matters of this kind.

¹ The notion of neutralisation was first introduced by N. Trubetzkoy in his book on essentials of phonology (Grundzuge der Phonologie, Prague, 1939; the book also appeared in a Russian translation in 1960).

The essential idea at the bottom of neutralisation in phonology may be briefly stated as follows. An opposition existing between two phonemes may under certain circumstances (which are to be strictly defined in each case) disappear, that is, it may lose its validity and become irrelevant. Such cases probably occur in every language. It will perhaps be best to give an example of neutralisation in Russian phonology. The sounds [t] and [d] are certainly different phonemes in Russian, as the difference between them may be the only means of distinguishing between two words. Compare, e.g., том 'volume' and дом 'house', or там 'there' and дам 'I shall give'. However, the difference between the two phonemes disappears at the end of a word (and also in some 'other cases). Thus, for example, the words рот 'mouth' and род 'genus' sound alike, a voiced [d] being impossible at the end of a word in Russian. Trubetzkoy says, accordingly, that the opposition between [t] and [d] is neutralised in those conditions. To put it more exactly, whereas in the word том the relevant features of the initial phoneme are three, namely, it is (a) a forelingual consonant, (b) a stop, and (c) voiceless, and the initial consonant of дом also has three relevant features, namely, it is (a) a forelingual consonant, (b) a stop, (c) voiced, the final consonant in рот or род has only two relevant features: it is (a) a forelingual consonant, and (b) a stop. No third relevant feature is found here. The consonant is of course phonetically voiceless, but the voicelessness is phonologically irrelevant, as the corresponding voiced consonant cannot appear in this position.

The notion of neutralisation has since been applied to grammar as well.

Chapter XV

THE VERB: POLYSEMANTIC AND HOMONYMOUS FORMS

Modern lexicology has in many cases to solve the problem whether we have to deal with two or more meanings of one word or with two or more different words sounding the same. Such questions have arisen concerning, for example, the nouns hand, head, board, the verbs draw, bear, and a number of other words.

Similar problems confront us in the field of grammar as well. In quite a number of cases we are faced with a choice between two possible interpretations of established linguistic facts, notably in the sphere of verb morphology: is a certain form one grammatical form with two or more different meanings, or two or more different grammatical forms sounding alike?

We have dealt with each of these problems as they arose in the course of our study of the verb system. Now it may prove expedient to cast a look at the problem in its entirety. We will first take up those cases in which there has been a general discussion and both, varying views have found more or less wide support, and then we will pass on to the problems in which one view is more or less prevailing, and only a few dissenting voices are heard.

(1) Is the form knew in the sentence He knew it all along and the form knew in the sentence If he knew this, he would be here the same form, or are they two different forms sounding the same? The question also applies to forms of the type lived, stopped, told, etc.

(2) Is the form had known in the sentence He had known it all along and the form had known in the sentence If he had known this, he would have come the same form, or are they two different forms sounding the same? The question also applies to forms of the type had lived, had stopped, had told, etc.

(3) Is the form should come in the sentence I said I should come soon and the form should come in the sentence If I were you I should come at once the same form, or are they two different forms sounding the same? Is the form would come in the sentence She said she would come soon and the form would come in the sentence If she knew this she would come at once the same form, or are they two different forms sounding the same?

(4) Is the form should have come in the sentence I thought 1 should have come before he rang up and the form should have come in the sentence If I had known this I should have come at once the same form, or are they two different forms sounding the same? Is the form would have come in the sentence He thought he would have come before you rang up and the form would have come in the sentence If he had known this he would have come at once the same form, or are they two different forms sounding the same?

188 The Verb: Polysemantic and Homonymous Forms

(5) Is the form would come in the sentence If he knew this he would come at once and the form would come in the sentence In those days he would come and sit with us for hours, and tell us about his life the same form, or are they two different forms sounding the same?

(6) Is the form laughing in the sentence I found a laughing little boy and the form laughing in the sentence He answered by laughing the same form, or are they two different forms sounding the same?

(7) Is the form having found in the sentence Having found the solution of the problem, he published a paper on the subject and the form having found in the sentence He was proud of having found the solution of the problem the same form, or are they two different forms sounding the same?

Those were questions that had been answered in different ways by different scholars. Now comes a question that has had no special attention focused upon it:

(8) Is the form wrote in the sentence He wrote it and the form wrote in the sentence They wrote it the same form, or are they two homonymous forms sounding the same?

There is a consideration in favour of the view that they are two different homonymous forms; the verb be has different forms for the singular and the plural in the past tense (was, were) — from this fact the inference may be drawn that in Modern English there is the category of number (singular and plural) in the past tense, and consequently in the verbs where no distinction in sound is found between singular and plural, we have to recognise homonymous forms. It may be further argued that in all verbs which admit of a past continuous form, or of a past passive, or of a past continuous passive, that is, of forms derived by means of the verb be, the category of number is found in the past tense: compare, e. g., was writing, were writing, was written, were written, was being written, were being written, or, was driving, were driving, was driven, were driven, was being driven, were being driven, etc.

(9) If the argument laid down in (8) is followed up, it may also be asked whether the forms know (1st person singular) and know (plural) are one form, or different forms sounding alike. In favour of the latter view it may be argued that in the verb be the corresponding forms do not sound the same: am, are, so this verb has a material distinction along these lines, and, consequently, all verbs in which no material distinction is found have homonymous forms. It may further be argued that verbs which have a present continuous, or a present passive, or a present passive continuous, or two, or all of these forms, also show that distinction: compare am driving, are driving; am driven, are driven, am being driven, are being driven, etc.

Invariable Meaning 189

In proceeding now to consider different arguments referring to the nine questions enumerated here, we will first of all point out the problem of various structural meanings inherent in a grammatical form and of an invariable meaning, that is, one to be found in every possible single application of a form. This has been treated in different ways with reference to such questions as the general meaning of a case, for instance of the genitive case in Russian or Latin, etc.

Whether we think it necessary to find an invariable structural meaning which manifests itself in different ways in different applications of a grammatical category, or whether we deny the necessity of such an invariable meaning, is a matter which largely depends on a scholar's theoretical views on the meaning of grammatical categories and grammatical forms in general. We can hardly expect either of these views (for or against an invariable structural meaning for every category and every form) to be definitely proved as the only right one. We will assume that an invariable meaning does exist, and then try to find out what it is in every case.

Starting, then, with the question of polysemy or homonymy of forms like knew or lived, which may either denote a real action in the past, or an unreal supposition in the present or future, we may look for an invariable meaning comprising both these concrete applications. This meaning would seem to be something like "an action not actually happening in the present", or an action removed from present reality, that is, either having occurred in the past and in this way cut off from present reality or else only vaguely supposed, or even definitely unreal, and in this second way cut off from present reality.¹ This would justify the view that knew or lived in all its applications is one and the same form, which we may call past indicative, and which is used in certain syntactical contexts to denote an unreal action in the present or future.

While this way of interpreting facts will probably never be proved to be the only correct one, -there are many arguments in its favour and we will here endorse this view of forms like knew and lived.

If that is accepted, a similar reasoning will hold good concern? ing the forms had known and had lived. The common element of meaning, that is the invariable to be found both in the sentence I had known this before and in the sentence If I had known this 1 should have come may be defined as follows: an action not really

¹ This idea was propounded, in a somewhat different context, by Prof. A. Potebnia with reference to some facts of Slavonic languages; e.g. in Russian the form of the conditional mood сказал бы is a combination of the past tense form сказал with the particle бы, which itself is by origin a past tense form of the verb быть.

140 The Verb: Polysemantic and Homonymous Forms

happening at a given period in the past; the two applications of that common invariable would then result in the following meaning: (a) an action happening before that period of the past which is being considered, and (b) an action merely supposed, and not actually happening in the past. The additional difficulty in this second item is, that everything has to be treated as belonging to the past (in some way or other), whereas with the first item the distinction was between the past and the present.

This approach to things is also possible in the case of our item (3), where the forms in question are, should come and would come, respectively. In trying to arrive at an invariable meaning for these forms, we will look for something which might establish a connection between an action unreal in the present and an action expected to happen at some moment future from the point of view of past time. The invariable in this case may be defined something like this: an action not really happening either in the present or in the specified period of the past; that idea is then substantiated either

(a) as something merely supposed for the present or future, or

(b) as an action viewed from a past viewpoint as happening in the future.

A similar reasoning would of course have to be applied to forms like should have come and would have come, with everything shifted, as it were, one step further back into the past: the invariable in this case would be something like "an action not actually happening either in the past, present, or future", and the applications would be, (a) an unreal action in the past, and (b) an action viewed from a past viewpoint as completed at a certain time in the future. In this last type of forms the past dominates throughout.

Similar considerations will hold good with reference to forms like should be coming, would be coming, should have been coming, would have been coming, which, however, are rarely found in their temporal application (future-continuous-in-the-past, future-perfect-continuous-in-the-past). Everything stated so far would also apply to the corresponding forms of the passive voice, wherever a verb admits of passive forms at all.

The next item, which we gave under number 5, is of a somewhat different character, and presents us with new difficulties. Besides being used to denote an unreal action in the present, and an action expected in the future from a past viewpoint, the phrase would come (in this particular case the verb would is completely dissociated from the verb should) can also express a repeated action in the past. For this problem, there seems to be no convincing way of finding an invariable meaning able to cover both the meaning of unreality in the present and expectation in the future from a past viewpoint. So, unless and until such common ground for an invariable is found, it will be well to say that would come denoting un-

Gerund and Participle 141

reality in the present and expected action in the future from a past viewpoint, on the one hand, and would come denoting a repeated action in the past are two different formations sounding the same. ¹

Now we come to items (6) and (7), concerning the ing-form or ing-forms in their different applications. The traditional view is, that we have here two homonymous forms: the participle (present or perfect) and the gerund (present or perfect). A more recent view, put forward by E. Kruisinga, is that there are not two different forms sounding the same but one form, which he shortly terms "the ing", being used in various ways in the sentence.

It is a peculiar feature of this ing-problem that in some contexts the two "ings" come very close together and additional factors are required to draw a distinction between them. The two "ings" coincide in such sentences as, He was afraid of her knowing the truth,² where the "ing" is a gerund if her is a possessive pronoun, and a participle if her is a personal pronoun in the objective case; also in the sentence He was glad at John's coming the "ing" is a gerund, but if John's is replaced by John, the "ing" seems to be a participle, though this is not acknowledged by all scholars: M. Deutschbein believed the "ing" to be a gerund in both cases.

The question is a very difficult one. Since up to now it has not been possible to find a convincing invariable meaning to cover both participle and gerund, we shall do well again, until such an invariable is discovered, to hold to the traditional view which has it that the participle and the gerund are two essentially different forms sounding the same. This of course applies equally to present and perfect, active and passive participles and gerunds.

The last two items of our list include questions connected with the whole system of grammar and the principles of stating grammatical categories. In item (8) the essence of the problem is this. All Modern English verbs, however many they may be, have no distinction of number in the past tense, with one exception only, the verb be, which distinguishes in the past tense between the singular (was) and the plural (were).³ Should this peculiarity of the verb be bring us to the conclusion that the category of number in the past tense exists in all English verbs, and that, accordingly, all verbs

¹ From the viewpoint of synchronic analysis of Modern English the fact that the source of the auxiliary would is in both cases the same (past tense of will) is of course irrelevant.

² The example is taken from M. Deutschbein, System der neuenglischen Syntax, S. 154.

³ We will for the moment overlook the fact that in non-standard English there is a strong tendency to do away with the distinction and to use the form was without regard to number: I was, he was, we was, you was, they was.

142 The Verb: Polysemantic and Homonymous Forms

except be have here homonymous forms? This is the view held by L. Bloomfield. Bloomfield thinks that the existence of one word of a certain category, which has a certain grammatical distinction, is sufficient reason for stating that all words of that category have that distinction, and all of them but the one in question have homonymous forms. In his own words, "The existence of even a single over-differentiated paradigm implies homonymy in the regular paradigms." ¹

This view, however, is completely arbitrary and unacceptable. If we were to endorse it, we should arrive at very strange conclusions indeed. For example, starting from the fact that two English words which may be used as attributes to a noun, namely the words this and that, have a distinction between singular and plural (they agree in number with their head word, e. g. this street, but these streets, that street, but those streets), we might infer that all words thus used also have the category of number; for example, we should have to say that the word new has a distinction of singular and plural: in the phrase new house the word new is in the singular, but in the phrase new houses it is in the plural; the singular and the plural forms would be homonyms.

Besides being queer in itself, such a view would lead to a very peculiar interpretation of the development of a language. We interpret the development of