This problem may be formulated as follows: is the sentence the highest existing syntactical unit, or are there higher syntactical units than the sentence — units of which a sentence is but a component part?
The traditional view of course is that the sentence is the highest syntactical unit and that whatever units we may find of a higher order will be not syntactical, but either stylistic, or literary. However, this traditional view has been questioned in a paper by Prof. N. Pospelov.1 Although Pospelov treats of the Russian language only, his views have a bearing on linguistics in general, since his reasonings are not based on any specifically Russian material, but on material found in other languages as well, and possibly in all languages, Therefore we will devote some space to the analysis of these ideas, as they may be applied to the English language.
What reasons are there, then, to suppose that there exists a grammatical, that is, a syntactical unit higher than the sentence, and how are the limits of this higher unit to be delineated?
The chief consideration that may be laid down in favour of this view is, that sometimes co-ordinating conjunctions establish some sort of connection between independent sentences, separated from each other by a full stop. The two conjunctions that are especially frequent in this function are, and and but. The conjunction and is found in this function often enough, and some writers seem to have a special predilection for it. Here is a typical example from Th. Dreiser's "An American Tragedy": They had been to all these wonderful places together. And now, without any real consciousness of her movements, she was moving from the chair to the edge of the bed, sitting with elbows on knees and chin in hands; or she was before the mirror or peering restlessly out into the dark to see if there were any trace of day. And at six, and six-thirty when the light was just breaking and it was nearing time to dress, she was still up — in the chair, on the edge of the bed, in the corner before the mirror. But she had reached but one definite conclusion and that was that in some way she must arrange not to have Clyde leave her.
It might perhaps be said that the higher unit established by co-ordinating conjunctions is somewhat like what we call a "paragraph". But a conjunction of this kind may even be found at the
1 See H. С. Поспелов, Проблема сложного синтаксического целого в современном русском языкe. Учёные записки МГУ, вып. 137, кн. 2, 1948
The Problem of Higher Syntactical Units 349
beginning of a paragraph. Thus, in the passage just quoted the sentence And now, without any real consciousness... stands at the opening of a new paragraph, and so does the sentence beginning with But she had reached...
Occasionally a conjunction of this kind may even stand at the opening of a new chapter, as is the case in the following extract from "An American Tragedy":
The beauty of the various houses along Wykeage Avenue and its immediate tributaries! The unusual and intriguing sense of movement and life there so much in evidence. Oh, if he were but of it!
Chapter XXIII
And then, one November evening as Clyde was walking along Wykeage Avenue, just west of Central, a portion of the locally celebrated avenue which, ever since he had moved to Mrs Peyton's he was accustomed to traverse to and from his work, one thing did occur which in so far as he and the Griffiths were concerned was destined to bring about a chain of events which none of them could possibly have foreseen.
If we were to take the view that a co-ordinating conjunction always forms some kind of grammatical unit, we should have to say, in this case, that the grammatical unit formed by the conjunction and embraces the end of one chapter and the beginning of another. That, however, would be most unnatural. If, therefore, we have to choose between the two alternatives: either to admit that a conjunction may join two independent sentences without forming any higher grammatical unit, or to say that a higher grammatical unit may include parts of two chapters, and so forth, we will decidedly prefer the first of them. This, however, will make it necessary to add something to the definition of a conjunction: a conjunction may unite words, parts of a sentence, clauses, and independent sentences as well (compare above, p. 31).
Let us study the means which are used to establish connections between sentences. And this leads on to a series of questions which may be said to lie on the border line of grammar.
What is meant is study of the structure of entire texts, such as short newspaper notices, poems, or novels. In this study it does not appear possible to stay strictly within the limits of grammar: some lexical phenomena will also have to be taken into consideration.
We will only give some hints as to the possible trends of investigation in this field, and we begin by studying some opening
850 Conclusion
paragraphs of a modern novel. Let this be Eyeless in Gaza by Aldous Huxley:
The snapshots had become almost as dim as memories. This young woman who had stood in a garden at the turn of the century was like a ghost at cockcrow. His mother, Anthony Beavis recognised. A year or two, perhaps only a month or two, before she died. But fashion, as he peered at the brown phantom, fashion is a topiary art. Those swan-like loins! That long slanting cascade of bosom — without any apparent relation to the naked body beneath! And all that hair, like an ornamental deformity of the skull! Oddly hideous and repellent it seemed in 1933. And yet, if he shut his eyes (as he could not resist doing), he could see his mother languidly beautiful on her chaise-longue, or, agile, playing tennis, or swooping like n bird across the ice of a far-off winter,
Now let us take a look at the elements in this passage which in some way or other tend to establish connections between sentences.
In the first sentence there is the past perfect form had become, which points to two time levels in the narration. In the second sentence, there is another past perfect form — had stood and this time it is correlated with the past indefinite form was in the same sentence.
In the third sentence the possessive pronoun his does not establish any connection with the preceding text, as there has so far been no mention of any man, to whom the possessive pronoun might refer. It refers to the name Anthony Beavis, which appears after the pronoun (this is not a frequent use). If there had been mention of a man in the preceding text this would be misleading. In the next sentence the pronoun she establishes a connection both with the second sentence (the phrase this young woman) and with the third (the phrase his mother).
In the next sentence, the conjunction but establishes a relation with the preceding text. So does the pronoun he, referring to the name Anthony Beavis, and also the phrase the brown phantom, which (as is clear from the context) refers to features of the woman in the photo. Then the pronoun it refers to the phrase all that hair, and would be unintelligible without this reference. Finally, the phrase his mother in the last sentence of the passage clearly refers back to the identical phrase his mother used in the third sentence.
Further investigation into such means of establishing connections between independent sentences should yield valuable conclusions about logical and semantic structure of larger text units. It is fairly obvious that here grammatical means go hand in hand with lexical ones, and the scholar's task should be to find out the precise part played by each of these, and the way they combine to produce the final result.
Representation and Substitution 851
REPRESENTATION AND SUBSTITUTION
It will often be found in Modern English, as in other languages, that some element of a sentence apparently necessary to its meaning is not actually there and its function is taken up by some other element. We will first illustrate this general statement by two examples which will at the same time show two different ways of expressing the function of an element which is not there: (1) I could not find him, though I wanted to. (2) He works more than you do. The full text of these sentences would evidently run like this: (1 ) I could not find him, though I wanted to find him. (2) He works more than you work. What we have to discuss is, in what way the meaning of the words find him and work respectively is suggested without their being actually used in the sentence. In the first of the two sentences, I could not find him, though I wanted to, the meaning of the missing infinitive to find with the adhering pronoun him is suggested by merely using the infinitival particle to (after wanted) which, as it were, does duty for the infinitive and the pronoun (or it might be a noun, or indeed any phrase denoting the object of the verb find). No extra word is used here, that is, no word that would not stand in the full text of the sentence as we have reconstructed it. The particle to may be said to represent the infinitive and the noun or pronoun denoting the object of the action.
This way of suggesting the meaning of words not actually used may be termed "representation".
In our other example, He works more than you do, things are somewhat different. If we compare the text as it stands with the full version: He works more than you work, we see that there is in our text a word that is not found in the full version, namely the verb do. It is quite obvious that the verb do in such cases may replace any verb except the auxiliaries be, have, etc., and the modal verbs can, may, etc. It should also be noted that the verb do in this function need not necessarily be in the same tense, or mood, as the verb which it replaces.
This case differs from the one considered above in that a word appears which would not have been used in a full version of the sentence. This way may be termed "substitution", as distinct from representation.
Having established the main facts concerning representation and substitution, we can now proceed to point out some typical phenomena of both kinds in Modern English.
There are some cases of representation highly characteristic of the English language. Among these we must mention, in the first place, representation by an auxiliary verb of an analytical verb form of which it is a part. The auxiliary verbs capable of performing this function are, be, have, shall, will, should, would, e. g. "Oh,
852 Conclusion
shes fainted again." "No I havnt." (SHAW) The auxiliary always represents the analytical verb form which was last used in the sentence. This indeed appears to be the only natural and idiomatic way of expressing the ideas in question: if the speaker had used the full form, this would in every case sound strikingly awkward and inappropriate, no matter what the stylistic sphere of the text may be. Compare also: "Which of us was the better fencer?" "I was." "Of course you were." (Idem)
This kind of representation is found within the limits of one sentence, as in the example already quoted: She didn't count with Stella, never had, and never would (WOODHILL) and also in short answers in dialogue, as in the following extracts: "I have a frightful feeling that I shall let myself be married because it is the world's will that you should have a husband." "I daresay I shall, someday." (SHAW) "Do you intend to tell him what you have been telling me to-night?" "I hadn't meant to. I had rather not." (R. MACAULAY)
Auxiliary and modal verbs, and the infinitival particle to are the chief means of representation in Modern English.
The other way of suggesting the meaning of a word that is not actually used in the sentence is substitution. Instead of repeating a word that has already been used in the sentence, or in the preceding one, another word is used, whose own lexical meaning is irrelevant and which serves as a means of "hinting" at the meaning of the word that is not repeated.
The two main words used in this function are the verb do and the pronoun one, each in its own sphere, of course. The verb do can substitute any verb except those enumerated OB page 351, in fact it can substitute all the verbs with which it is used to form their interrogative and negative forms. For instance, it can substitute the verb appreciate, as in the sentence Nobody can appreciate it more than I do (SHAW), just as it is used in its interrogative and negative forms: Do you appreciate it? He does not appreciate it, etc. But it cannot be used to substitute, for instance, the verb must, just as it is not used to derive interrogative and negative forms of that verb.
It will be readily seen that in the sphere of verbs representation and substitution complete each other: in some verbal forms (present indefinite and past indefinite) substitution by do is used, whereas in all other forms (the analytical ones) representation is the method used.
Occasionally the verb do in this function can even precede the verb which it replaces. This is the case in the following sentence: As he was accustomed to doing, Harry closed the sale and had the signed contract in his pocket within fifteen minutes. (E. CALDWELL) It may even be said that the verb do here replaces the whole phrase closed... fifteen minutes.
Representation and Substitution 353
As to the other substitution word, the pronoun one, it is of course used to substitute nouns. It is important to note that its use is limited. The noun to be substituted should be in its indefinite variety, that is, it should be accompanied by the indefinite article: otherwise its substitution by the pronoun one is not possible. Compare the two following bits of dialogue: (1) "Have you found an English teacher?" "Yes, I have found one," but (2) "Have you found the English teacher?" "Yes, I have found him (or her)," not "one". Or again: "Do you know a foreign language?" "Yes, I know one" but "Do you know the English language?" "Yes, I know it."
So the meaning of indefiniteness adheres to the pronoun one as it does to the indefinite article, whose doublet it actually is. However, the pronoun one differs from the indefinite article in that it has a plural form (ones), which the indefinite article of course has not.
On the other hand, however, the pronoun one can also be used with reference to a definite object, and in that case it is preceded by the definite article and some limiting attribute must come either before it (i. e. between the definite article and the pronoun) or after it, in the shape of an attributive clause with or without a relative pronoun. Hence the following types of groups are possible: (1) the green one, the larger one, (2) the one which you mentioned, the one he bought, etc. or in the plural, (1) the green ones, the ones you mentioned, (2) the ones which you mentioned, the ones he bought, etc.
Though the pronoun one is thus a very common substitute for a noun not repeated in the sentence, it by no means follows that the pronoun must be used wherever such repetition is avoided. Sentences are numerous enough in which the pronoun one is not used: we may say that in these cases it is the preceding attribute (which is usually, if not always, an adjective) that represents the omitted noun which is to be understood from a former part of the sentence, or from a preceding sentence. Here is a characteristic example from the beginning of a sketch by Jerome K. Jerome: "Now, which would you advise, dear? You see, with the red I shan't be able to wear my magenta hat." "Well, then, why not have the grey?" "Yes, yes, I think the grey will be more useful." "It's a good material." "Yes, and it's a pretty grey. You know what I mean, dear; not a common grey. Of course grey is always an uninteresting colour." "It's quiet." "And then again, what I feel about the red is that it is so warm-looking. Red makes you feel warm even when you're not warm. You know what I mean, dear." "Well, then, why not have the red? It suits you — red."
In the whole of this extract the noun material, to which the words red and grey refer, has only been used once. It appears, too,
354 Conclusion
that the adjectives red and grey tend to be substantivised, as is seen from the use of the phrases a pretty grey and a common grey. Speaking of substitution in a wider sense, we might include personal pronouns of the third person, which more often than not perform this function. But this lies beyond that specific sphere or representation and substitution which we are considering here, and besides in this use of personal pronouns English does not appear to differ in any way from other languages.
GRAMMAR AND STYLE
Though problems of style as such are outside the scope of this book, some remarks concerning the stylistic value of grammatical categories and grammatical elements may prove appropriate to a thorough study of English grammatical structure.
From the stylistic viewpoint, it should first of all be noted that some grammatical categories and phenomena are neutral while others are not. To be more explicit, this means that some grammatical phenomena may appear in any sort of speech, whether oral or written, whether solemn or vulgar, etc., without in any way conflicting with the stylistic colouring of the text, whatever it may happen to be. Other grammatical phenomena, on the other hand, have a distinct stylistic colouring and will produce an effect of inappropriateness if applied outside their stylistic sphere.
To illustrate this general statement, we might say that the past indefinite tense is devoid of any stylistic colouring, it is stylistically neutral and it appears both in a solemn hymn and in a street song, and indeed in any kind of text without any exception whatsoever. On the other hand, the so-called absolute construction, as in the sentence She picked up a large split-oak basket and started down, the back stairs, each step jouncing her head until her spine seemed to be trying to crash through the top of her skull (M. MITCHELL) has a distinctly literary flavour. Constructions of this kind are not used in colloquial speech and if, say, an author were to put a construction of this kind into the mouth of a character in a comedy of modern English life, it would sound singularly inappropriate. To take a different example: the forms of the personal pronouns him, her, us, them, used in the function of a predicative after the subject it and the link verb is, or was, have a very distinct low colloquial tinge, and they would be completely inappropriate in a literary, still more so in a solemn context. A sentence like It was them that did it has that peculiar stylistic colouring which creates a certain atmosphere, even if nothing preceded that sentence (for example, if it were the opening sentence of some short story). All this has to be reckoned with in characterising the grammatical resources of the Modern English language.
Grammar and Style 855
We will now give a brief survey of the grammatical categories and the grammatical phenomena which bear (or tend to bear) some kind of stylistic colouring or other, first those of morphology, then those of syntax.
Morphology
In the sphere of nouns there is not much to be noted in the way of stylistic colouring.
In a very few cases where a noun has alternative plural forms, the irregular form (the one not in -s) naturally tends to have a high-flown, archaic, or poetic flavour. The very fact that there exists a plural form in -s alongside of it gives the other form the character of something unusual and restricted in use to special purposes. The only two words that have to be mentioned in this connection are, brother with its alternative plural form brethren differing from brothers not in stylistic colouring alone, and cow, with its alternative plural form kine having a very strong archaic and poetic tinge.
In the sphere of case it can be noted that the genitive in -'s tends to acquire a specific stylistic flavour when formed from a noun not denoting a living being. As a rule the of -phrase is used to express relation between the thing denoted by the noun and that denoted by another noun. For instance, if this sort of relation has to be expressed between England and history, the usual, stylistically neutral way of expressing it is to say the history of England, and this, indeed, is the title, for instance, of most textbooks on the subject. But alongside of it the variant England's history is also permissible. It has a poetic and possibly patriotic shade about it and it will do very well in an emotional context, but would be out of place in a strictly scientific one.
The exact sphere of nouns whose forms in -'s tend to acquire such a peculiar stylistic character is however extremely difficult, if not impossible, to define, as the forms in -'s tend to spread in recent times, as we noted in our chapter on case (see p. 43). Much concrete observation and analysis is necessary before anything more definite can be said on the subject.
There is little to be said about adjectives, too, which have only degrees of comparison as a morphological characteristic.
What matters here is the stylistic colouring of degrees of comparison in -er, -est of such adjectives as do not usually possess such forms. Where such forms do appear they tend to have a peculiar solemn stylistic quality which would make them unfit for any other context. The English nineteenth-century writer and philosopher Thomas Carlyle would use a superlative in -est of two-syllable
856 Conclusion
adjectives derived from present participles in -ing, as will be seen from the following example: With unabated bounty the land of England blooms and grows. Waving with yellow harvests, thick-studded with workshops, industrial implements, with fifteen millions of workers, understood to be the strongest, the cunningest and the willingest our Earth ever had... Neither of these forms occur in ordinary style: the analytic formations most cunning, most willing, etc. would be used instead.
In the sphere of pronouns, there is the use of the forms I or me, etc., which we have already considered in Chapter VI, and we need not dwell on it here.
Another point to be noted about pronouns in the morphological way is the form 'em in sentences like I'll show 'em alongside of I'll show them. Strictly speaking this is a morphological point if we consider 'em to be a different form, not merely a phonetically weakened variant of them. If we take it that way we will state that the morphological variant 'em for the objective case of the third person plural personal pronoun has a definite stylistic colouring of low colloquial style. It would be, for instance, entirely out of place in a serious scientific treatise. It is, however, quite appropriate in reproducing low colloquial (and possibly vulgar) speech.
The main bulk of stylistic remarks in the sphere of morphology belongs of course to the verb. There are a considerable number of details here which point to a peculiar stylistic colouring, either solemn and archaic, or low colloquial and eventually vulgar.
The first to be noted are the forms in -th for the third person singular, present indicative, that is, forms like liveth, knoweth, saith, doth, hath, etc. These have acquired (since the 17th century) a definite archaic and poetical flavour and cannot accordingly be used in any other, or in any neutral stylistic surroundings. Examples of their use in modern texts are rare indeed.
The same stylistic colouring as with the -th- formsis also inherent in forms in -st for the second person singular of both the present and the past indicative (that is, the forms livest, knowest, sayst, dost, livedst, knewest, saidst, didst, hadst, etc.) and also the forms shalt, wilt, art, wert (or wast) of the verbs shall, will, be. These forms are practically inseparable from the second person singular personal pronoun thou. In every other respect the - st -forms of the second person are exactly similar to the -th- formsof the third. They are quite rare in Modern English.
These, then, are forms which may, generally speaking, be derived from every verb.
The other forms with special stylistic colouring belong to definite individual verbs only, though some of them, belonging to verbs which are or may be auxiliary, can accordingly be brought into the system of all verbs which use the auxiliary.
Grammar and Style 857
Here we must first of all mention the form ain't pronounced [eint], or ain [ein] of the verb be, corresponding to the forms am not, is not, and are not of the stylistically neutral set. The essence of all of them is, of course, that the combination of a verb form with the negative particle not differs from the same form without the particle. The difference between am not, is not, and are not is in these cases neutralised. So this whole question also has some bearing on the categories of person and number in the verb be. The stylistic tinge of the form ain't is a very definite one: it is low colloquial with a clear tendency towards vulgarity, and of course it would be inadmissible in any serious literary style. Here are some examples: The house aint worth livin in since you left it Candy. (SHAW, Burgess's speech in "Candida") Our quarrel's made up now, ain it? (Idem) James and me is come to a nunnerstanding — a honorable unnerstandin. Ain we, James? (Idem)
A similar stylistic character attaches to the forms has, is, and was for the plural, e. g. Yes: limes 'as changed mor 'n I could a believed. (Idem ) I hused to wonder you was let preach at all. (Idem)
As the verb be is an auxiliary of the continuous aspect and of the passive voice, the form ain't can accordingly appear in every verb possessing either of these categories, or both, e. g. Ope you ain't lettin James put no foolish ideas into your ed? (Idem)
Besides, a certain number of verbs have, alongside of their normal and neutral forms, some special ones, differing from the usual by a distinct archaic or solemn colouring, e. g. spake for spoke (past tense of the verb speak); throve for thrived (past tense of the verb thrive); bare for bore (past tense of the verb bear).
In the opposite way, there are some forms having at present a very distinct vulgar or illiterate stylistic character and only used in writing to characterise an illiterate speaker. They are forms of the past tense and second participle on -ed of verbs regularly deriving these forms by ablaut (vowel change) or by adding the -n-suffix for the second participle, e. g. seed for saw (past tense) or seen (second participle of the verb see); knowed for knew (past tense) or known (second participle of the verb know). These forms are distinctly illiterate and in this they differ from the form ain't, for instance, which is somehow within the standard, though certainly at the lowest level of it.
It would seem that no verb has archaic and vulgar variants at the same time, that is, no verb has three variants: the normal one, an archaic one, and an illiterate one. For instance, the verb speak has an archaic variant spake for its past tense spoke but it has no illiterate variant; on the other hand, the verb see has an illiterate variant seed for its past tense saw but it has no archaic variant, etc.
358 Conclusion
Some peculiarities in the sphere of stylistically coloured verb forms should also be noted in American English. The chief of these concerns forms of the present perfect tense. In low colloquial American style there is a very clear tendency to drop the auxiliary have (has) in the present perfect, so that only the second participle remains. Now, if the second participle is homonymous with the past tense, as is the case with most verbs, the result of the omission is a form not to be distinguished from the past tense, for instance, I have found > I found. If, however, the second participle is not homonymous with the past tense, the result of omitting the auxiliary is a new form, not coinciding with the usual past tense: I have taken > I taken, he has written > he written, etc. We may see this in the following quotation from an American author representing low colloquial speech: I been around to see her a coupla times since then, only Esta didn't want me to say anything about that either. (DREISER)
However, such forms may also be found in England, e. g. James: three year ago, you done me a hill turn. You done me hout of a contrac. (SHAW, Burgess's speech)
H. L. Mencken, the author of the well-known book, "The American language" (first published in 1919), treats such forms as I taken, he written as a past tense. He also asserts that with the auxiliary have preserved, the form of the second participle is took, wrote, etc., so that the British paradigms take, took, taken; write, wrote, written correspond to the American take, taken, took; write, written, wrote, and gives a list of irregular verbs arranged in this way. 1 Mencken's view appears to be an exaggeration not borne out by American narrative and dramatic literature. I taken is common enough in American colloquial style, but I have took does not appear to be so.
It is clear that forms like I taken have a stylistic tinge but their peculiarity is that they hardly appear outside the USA.
This is about all that can be said about stylistic values of morphological forms in present-day English.
Syntax
In the sphere of syntax we have to look for syntactical synonyms differing from each other by their stylistic colouring. We may look for two sets of cases: (1) each of the two syntactical synonyms has a peculiar stylistic colouring of its own, (2) of two syntactical synonyms one is stylistically neutral, that is, it may appear in every sort of style, while the other has a distinct stylistic colouring, that is to say, its use is limited to definite stylistic conditions.
1 See H. L. Mencken, The American Language, 3rd ed., 1929, pp. 279—283.
Grammar and Style 859
The first of these sets of cases can hardly be frequent, since it would imply that there is no neutral syntactical means available to express the idea in question.
As a rare example of the first kind we can point to the variants It is I and It is me. The difference between them is certainly one of style, and it seems that neither of them is really neutral stylistically. It is me has a very clear colloquial colouring, while It is I is stiff and formal. This of course is a state of affairs due to a historical development in the course of which It is me has been steadily gaming ground, and most probably it will in a near future lose that specific colouring of colloquial style, and become the normal, that is, the stylistically neutral variant, while It is I will be relegated to a distinctly archaic sphere.
Far more numerous are the cases when one of a pair of syntactical synonyms has a specific stylistic colouring while the other is stylistically neutral. This is the case, for example, with the absolute construction and its synonyms — subordinate adverbial clauses of time or cause. The absolute construction has practically always — with very few exceptions, phraseological units like all things considered, or weather permitting — a distinctly literary or even bookish character.
A distinctly literary or bookish colouring also attaches to non-defining attributive clauses. For instance, the following sentence would not be possible in colloquial style: Cathleen Calvert, who came out of the house at the sound of voices, met Scarlett's eyes above her brother's head and in them Scarlett read knowledge and bitter despair. (M. MITCHELL)
These notes on the stylistic values of some grammatical facts are no more than hints. They are meant to suggest that alongside of grammatical phenomena that are indifferent to style there are some which have a distinct stylistic colouring and are decidedly inappropriate outside a certain stylistic sphere. This is most essential both from a purely theoretical viewpoint and from the viewpoint of teaching the language to foreigners. A bookish grammatical construction appearing in a colloquial context, though "grammatically correct", is as serious an error against English usage as a mistake in grammatical construction. This should especially be remembered in giving exercises of the kind providing for changing one construction into another (such as replacing a subordinate clause by an absolute construction, and the like).
In this book we have considered a number of problems presented by the grammatical structure of Modern English. In doing so we have applied certain methods developed by modern linguistics, which allow of a more exact analysis and evaluation of lan-
360 Conclusion
guage facts and reduce the sphere of subjective opinions incompatible with one another and admitting of no proof.
In this connection it has proved essential to distinguish as carefully as possible between two kinds of problems. On the one hand there are those which admit of a definite solution, so that an answer to the problem can be found and the problem need not be reconsidered unless and until some new facts emerge which may necessitate a change in the solution. On the other hand, there are the problems which, as far as we can see, do not admit of such a solution, but must remain a field of differing opinions, with the solution depending on a student's basic views of language phenomena in general, or of some narrower language sphere in particular. Some of the latter problems had better be abandoned altogether, since they offer no ground for any truly scientific analysis and only give rise to useless and unpromising dispute. Some of the problems connected with parts of a sentence clearly belong here: a typical example is the so-called problem as to whether a word or phrase accompanying a noun can be an object or must always be an attribute.
It is no use whatever to discuss such problems: the right way to deal with them is to adopt a certain definition (for example, a word or phrase accompanying a noun is termed an attribute) and then act according to the definition accepted. But there are other problems belonging to this category of doubtful cases, which cannot and should not be discarded. Among these is, for example, the problem of the category of voice, which has been causing lively discussion for a considerable length of time. We can by no means say that it has been solved, but in fact it deserves close attention, and its solution may be brought nearer by careful application of more exact and objective methods. It is essential for a student of English to bear in mind these various aspects of linguistic study, if a right perspective of this study is not to be lost sight of.
Another essential point to emerge from a careful scientific study of English grammatical structure is, the necessity of a very concrete approach to the individual structure of this one language, whose structure, such as it is, is probably not to be found in any other language on the globe. This fact tends sometimes to be obscured by a somewhat superficial application of the notions of "synthetic structure", "analytical structure", and the like. It is of course quite right to say that Modern English is a language mainly analytical in its structure but this general statement, true as it is, does not give us any clue to particular questions of grammatical structure, and it cannot replace careful study of these particular questions.
We may as well illustrate this point by an example or two. Both Modern English and Modern French are analytical languages, and that statement is certainly true. But it does not include some essential points of difference between the two languages. Thus, in
Grammar and Style 881
Modern English, adjectives have neither distinctions of gender nor any of number: for instance, the form fine, as it is, will do for all cases. Now, in Modern French, though also an analytical language, adjectives do have those categories, so that here we distinguish between four separate forms: masculine singular fin, masculine plural fins, feminine singular fine, feminine plural fines. Another point of difference between the two languages: English has only analytical forms for the future tenses (shall write, shall have written, shall be writing, shall have been writing in the active voice), while French, analytical as it is, has one synthetic future tense (écrirai) and one analytical (aurai écrit). The same may be said about forms expressing unreal action (whatever terms we may prefer to denote them): English has only analytical forms here (should write, should have written, should be writing, should have been writing in the active voice), whereas French has one synthetic form (écrirais) and one analytical (aurais écrit).
Similar caution is required when comparing English', a language basically analytical, with Russian, a language basically synthetic. These characteristics, though essentially true, should not be pressed too close.
In concluding our survey of English grammatical structure, we shall do well to emphasise that there remains much to be investigated in the future. To say nothing of the theory of phrases, which is still in its infancy, even those parts of grammar which have been studied for a hundred years or more present a number of unsolved problems where much energy and patient effort will have to be applied. The new methods aiming at a more exact and objective study of language facts should enable scholars to overcome outdated ideas and prejudices, which often constitute a formidable obstacle in the way of fundamental scientific research work, and further a complete and unbiassed view of Modern English grammatical structure as it presents itself to-day and as it tends to develop in the foreseeable future,
