> 1-25 26-50 51-75> . . " : "

. . " : " ( 4)

: . . " : ".



As for predication proper, it embodies not any kind of modality, but only syntactic modality as the fundamental distinguishing feature of the sentence. It is the feature of predication, fully and explicitly expressed by a contextually relevant grammatical complex, that identi-fies the sentence in distinction to any other combination of words hav-ing a situational referent.
The centre of predication in a sentence of verbal type (which is the predominant type of sentence-structure in English) is a finite verb. The finite verb expresses essential predicative meanings by its cate-gorial forms, first of all, the categories of tense and mood (the category of person, as we have seen before, reflects the corresponding category of the subject). However, proceeding from the principles of sentence analysis worked out in the Russian school of theoretical syntax, in particular, in the classical treatises of V.V. Vinogradov, we insist that predication is effected not only by the

forms of the finite verb connecting it with the subject, but also by all the other forms and elements of the sentence establishing the connec-tion between the named objects and reality, including such means of expression as intonation, word order, different functional words. Be-sides the purely verbal categories, in the predicative semantics are in-cluded such syntactic sentence meanings as purposes of communica-tion (declaration interrogation inducement), modal probability, affirmation and negation, and others, which, taken together, provide for the sentence to be identified on its own, proposemic level of lin-gual hierarchy.
2. From what has been said about the category of predication, we see quite clearly that the general semantic content of the sentence is not at all reduced to predicative meanings only. Indeed, in order to establish the connection between some substance and reality, it is first necessary to name the substance itself. This latter task is effected in the sentence with the help of its nominative means. Hence, the sen-tence as a lingual unit performs not one, but two essential signemic (meaningful) functions: first, substance-naming, or nominative func-tion; second, reality-evaluating, or predicative function.
The terminological definition of the sentence as a predicative unit gives prominence to the main feature distinguishing the sentence from the word among the meaningful lingual units (signernes). However, since every predication is effected upon a certain nomination as its material semantic base, we gain a more profound insight into the dif-ference between the sentence and the word by pointing out the two-aspective meaningful nature of the sentence. The semantics of the sen-tence presents a unity of its nominative and predicative aspects, while the semantics of the word, in this sense, is monoaspective.
Some linguists do not accept the definition of the sentence through predication, considering it to contain tautology, since, allegedly, it equates the sentence with predication ("the sentence is predication, predication is the sentence"). However, the identification of the two aspects of the sentence pointed out above shows that this negative atti-tude is wholly unjustified; the real content of the predicative interpre-tation of the sentence has nothing to do with definitions of the "vi-cious circle" type. In point of fact as follows from the given exposi-tion of predication, predicative meanings

do not exhaust the semantics of the sentence; on the contrary, they presuppose the presence in the sentence of meanings of quite another nature, which form its deeper nominative basis. Predicative functions work upon this deep nominative basis, and as a result the actual utter-ance-sentence is finally produced.
On the other hand, we must also note a profound difference be-tween the nominative function of the sentence and the nominative function of the word. The nominative meaning of the syntagmatically complete average sentence (an ordinary proposemic nomination) re-flects a processual situation or event that includes a certain process (actional or statal) as its dynamic centre, the agent of the process, the objects of the process, and also the various conditions and circum-stances of the realisation of the process. This content of the pro-posemic event, as is known from school grammar, forms the basis of the traditional syntactic division of the sentence into its functional parts. In other words, the identification of traditional syntactic parts of the sentence is nothing else than the nominative division of the sen-tence. Cf.:
The pilot was steering the ship out of the harbour.
The old pilot was carefully steering the heavily loaded ship through the narrow straits out of the harbour.
As is easily seen, no separate word, be it composed of so many stems, can express the described situation-nominative semantics of the proposition. Even hyperbolically complicated artificial words such as are sometimes coined for various expressive purposes by au-thors of fiction cannot have means of organising their root compo-nents analogous to the means of arranging the nominative elements of the sentence.
Quite different in this respect is a nominal phrase a compound signemic unit made up of words and denoting a complex phenomenon of reality analysable into its component elements together with vari-ous relations between them. Comparative observations of predicative and non-predicative combinations of words have unmistakably shown that among the latter there are quite definite constructions which are actually capable of realising nominations of proposemic situations. These are word-combinations of full nominative value represented by expanded substantive phrases. It is these combinations that, by their nominative potential, directly correspond to sentences expressing typical proposemic situations. Cf.:
1G1499 241

... > The pilot's steering of the ship out of the harbour. ... > The old pilot's careful steering of the heavily loaded ship through the nar-row straits out of the harbour.
In other words, between the sentence and the substantive word-combination of the said full nominative type, direct transformational relations are established: the sentence, interpreted as an element of paradigmatics, is transformed into the substantive phrase, or "nominal-ised", losing its processual-predicative character. Thus, syntactic nominalisation, while depriving the sentence of its predicative aspect (and thereby, naturally, destroying the sentence as an immediate communicative unit), preserves its nominative aspect intact.
The identification of nominative aspect of the sentence effected on the lines of studying the paradigmatic relations in syntax makes it pos-sible to define more accurately the very notion of predication as the specific function of the sentence.
The functional essence of predication has hitherto been understood in linguistics as the expression of the relation of the utterance (sen-tence) to reality, or, in more explicit presentation, as the expression of the relation between the content of the sentence and reality. This kind of understanding predication can be seen, for instance, in the well-known "Grammar of the Russian Language" published by the Acad-emy of Sciences of the USSR, where it is stated that "the meaning and purpose of the general category of predication forming the sentence consists in referring the content of the sentence to reality".* Compare with this the definition advanced by A. I. Smirnitsky, according to which predication is understood as "referring the utterance to reality" [, (1), 102].
The essential principles of this interpretation of predication can be expressed even without the term "predication" as such. The latter ap-proach to the exposition of the predicative meaning of the sentence can be seen, for instance, in the course of English grammar by M. A. Ganshina and N. M. Vasilevskaya, who write: "Every sentence shows the relation of the statement to reality from the point of view of the speaker" [Ganshina, Vasilevskaya, 321].
Now, it is easily noticed that the cited and similar
* . M., 1960. T. 2, . I. . 79.80.
definitions of predication do not explicitly distinguish the two cardi-nal sides of the sentence content, namely, the nominative side and the predicative side. We may quite plausibly suppose that the non-discrimination of these two sides of sentence meaning gave the ulti-mate cause to some scholars for their negative attitude towards the notion of predication as the fundamental factor of sentence forming.
Taking into consideration the two-aspective character of the sen-tence as a signemic unit of language, predication should now be inter-preted not simply as referring the content of the sentence to reality, but as referring the nominative content of the sentence to reality. It is this interpretation of the semantic-functional nature of predication that discloses, in one and the same generalised presentation, both the unity of the two identified aspects of the sentence, and also their different, though mutually complementary meaningful roles.
1. The notional parts of the sentence referring to the basic ele-ments of the reflected situation form, taken together, the nominative meaning of the sentence. For the sake of terminological consistency, the division of the sentence into notional parts can be just so called the "nominative division" (its existing names are the "grammatical division" and the "syntactic division"). The discrimination of the nominative division of the sentence is traditional; it is this type of di-vision that can conveniently be shown by a syntagmatic model, in par-ticular, by a model of immediate constituents based on the traditional syntactic analysis (see Ch. XXIV).
Alongside of the nominative division of the sentence, the idea of the so-called "actual division" of the sentence has been put forward in theoretical linguistics. The purpose of the actual division of the sen-tence, called also the "functional sentence perspective", is to reveal the correlative significance of the sentence parts from the point of view of their actual informative role in an utterance, i.e. from the point of view of the immediate semantic contribution they make to the total information conveyed by the sentence in the context of con-nected speech. In other words,

the actual division of the sentence in fact exposes its informative per-spective.
The main components of the actual division of the sentence are the theme and the rheme. The theme expresses the starting point of the communication, i.e. it denotes an object or a phenomenon about which something is reported. The rheme expresses the basic informative part of the communication, its contextually relevant centre. Between the theme and the rheme are positioned intermediary, transitional parts of the actual division of various degrees of informative value (these parts are sometimes called "transition").
The theme of the actual division of the sentence may or may not coincide with the subject of the sentence. The rheme of the actual di-vision, in its turn, may or may not coincide with the predicate of the sentence either with the whole predicate group or its part, such as the predicative, the object, the adverbial.
Thus, in the following sentences of various emotional character the theme is expressed by the subject, while the rheme is expressed by the predicate:
Max bounded forward. Again Charlie is being too clever! Her ad-vice can't be of any help to us.
In the first of the above sentences the rheme coincides with the whole predicate group. In the second sentence the adverbial introducer again can be characterised as a transitional element, i.e. an element informationally intermediary between the theme and the rheme, the latter being expressed by the rest of the predicate group, The main part of the rheme the "peak" of informative perspective - is rendered in this sentence by the intensified predicative too clever. In the third sentence the addressee object to us is more or less transitional, while the informative peak, as in the previous example, is expressed by the predicative of any help.
In the following sentences the correlation between the nominative and actual divisions is the reverse: the theme is expressed by the predicate or its part, while the rheme is rendered by the subject:
Through the open window came the purr of an approaching motor car. Who is coming late but John! There is a difference of opinion be-tween the parties.
Historically the theory of actual division of the sentence is con-nected with the logical analysis of the proposition. The

principal parts of the proposition, as is known, are the logical subject and the logical predicate. These, like the theme and the rheme, may or may not coincide, respectively, with the subject and the predicate of the sentence. The logical categories of subject and predicate are pro-totypes of the linguistic categories of theme and rheme. However, if logic analyses its categories of subject and predicate as the meaning-ful components of certain forms of thinking, linguistics analyses the categories of theme and rheme as the corresponding means of expres-sion used by the speaker for the sake of rendering the informative content of his communications.
2. The actual division of the sentence finds its full expression only in a concrete context of speech, therefore it is sometimes referred to as the "contextual" division of the sentence. This can be illustrated by the following example: Mary is fond of poetry.
In the cited sentence, if we approach it as a stylistically neutral construction devoid of any specific connotations, the theme is ex-pressed by the subject, and the rheme, by the predicate. This kind of actual division is "direct". On the other hand, a certain context may be built around the given sentence in the conditions of which the order of actual division will be changed into the reverse: the subject will turn into the exposer of the rheme, while the predicate, accordingly, into the exposer of the theme. Cf.: "Isn't it surprising that Tim is so fond of poetry?" "But you are wrong. Mary is fond of poetry, not Tim."
The actual division in which the rheme is expressed by the subject is to be referred to as "inverted".
3, The close connection of the actual division of the sentence with the context in the conditions of which it is possible to divide the informative parts of the communication into those "already known" by the listener and those "not yet known" by him, gave cause to the recognised founder of the linguistic theory of actual division J. Mathesius to consider this kind of sentence division as a purely se-mantic factor sharply opposed to the "formally grammatical" or "purely syntactic" division of the sentence (in our terminology called its "nominative" division).
One will agree that the actual division of the sentence will really lose all connection with syntax if its components are to be identified solely on the principle of their being

"known" or "unknown" to the listener. However, we must bear in mind that the informative value of developing speech consists not only in introducing new words that denote things and phenomena not men-tioned before; the informative value of communications lies also in their disclosing various new relations between the elements of re-flected events, though the elements themselves may be quite familiar to the listener. The expression of a certain aspect of these relations, namely, the correlation of the said elements from the point of view of their immediate significance in a given utterance produced as a predi-cative item of a continual speech, does enter the structural plane of language. This expression becomes part and parcel of the structural system of language by the mere fact that the correlative informative significance of utterance components are rendered by quite definite, generalised and standardised lingual constructions. The functional purpose of such constructions is to reveal the meaningful centre of the utterance (i.e. its rheme) in distinction to the starting point of its con-tent (i.e. its theme).
These constructions do not present any "absolutely formal", "purely differential" objects of language which are filled with semantic content only in the act of speech communication. On the contrary, they are bilateral signemic units in exactly the same sense as other mean-ingful constructions of language, i.e. they are distinguished both by their material form and their semantics. It follows from this that the constructional, or immediately systemic side of the phenomenon which is called the "actual division of the sentence" belongs to no other sphere of language than syntax. And the crucial syntactic desti-nation of the whole aspect of the actual division is its rheme-identifying function, since an utterance is produced just for the sake of conveying the meaningful content expressed by its central informative part, i.e. by the rheme.
4. Among the formal means of expressing the distinction be-tween the theme and the rheme investigators name such structural elements of language as word-order patterns, intonation contours, con-structions with introducers, syntactic patterns of contrastive com-plexes, constructions with articles and other determiners, constructions with intensifying particles.
The difference between the actual division of sentences signalled by the difference in their word-order patterns can

be most graphically illustrated by the simplest type of transforma-tions. Cf.:
The winner of the competition stood on the platform in the middle of the hall. > On the platform in the middle of the hall stood the win-ner of the competition. Fred didn't notice the flying balloon. > The one who didn't notice the flying balloon was Fred. Helen should be the first to receive her diploma. > The first to receive her diploma should be Helen.
In all the cited examples, i.e. both base sentences and their trans-forms, the rheme (expressed either by the subject or by an element of the predicate group) is placed towards the end of the sentence, while the theme is positioned at the beginning of it. This kind of positioning the components of the actual division corresponds to the natural de-velopment of thought from the starting point of communication to its semantic centre, or, in common parlance, from the "known data" to the "unknown (new) data". Still, in other contextual conditions, the reversed order of positioning the actual division components is used, which can be shown by the following illustrative transformations:
It was unbelievable to all of them. > Utterly unbelievable it was to all of them. Now you are speaking magic words, Nancy. > Magic words you are speaking now, Nancy. You look so well! > How well you look!
It is easily seen from the given examples that the reversed order of the actual division, i.e. the positioning of the rheme at the beginning of the sentence, is connected with emphatic speech.
Among constructions with introducers, the there-pattern provides for the rhematic identification of the subject without emotive connota-tions. Cf.:
Tall birches surrounded the lake. > There were tall birches sur-rounding the lake. A loud hoot came from the railroad. > There came a loud hoot from the railroad.
Emphatic discrimination of the rheme expressed by various parts of the sentence is achieved by constructions with the anticipatory it. Cf.:
Grandma gave them a moment's deep consideration. > It was a moment's deep consideration that Grandma gave

them. She had just escaped something simply awful. ~* It was some-thing simply awful that she had just escaped. At that moment Laura joined them. > It was Laura who joined them at that moment.
Syntactic patterns of contrastive complexes are used to expose the rheme of the utterance in cases when special accuracy of distinction is needed. This is explained by the fact that the actual division as such is always based on some sort of antithesis or "contraposition" (see fur-ther), which in an ordinary speech remains implicit. Thus, a syntactic contrastive complex is employed to make explicative the inner con-trast inherent in the actual division by virtue of its functional nature. This can be shown on pairs of nominatively cognate examples of anti-thetic constructions where each member-construction will expose its own contrastively presented element. Cf.:
The costume is meant not for your cousin, but for you.
The costume, not the frock, is meant for you, my dear.
The strain told not so much on my visitor than on myself.
The strain of the situation, not the relaxation of it, was
what surprised me.
Determiners, among them the articles, used as means of forming certain patterns of actual division, divide their functions so that the definite determiners serve as identifiers of the theme while the indefi-nite determiners serve as identifiers of the rheme. Cf.:
The man walked up and down the platform. A man walked up and down the platform. The whole book was devoted to the descrip-tion of a tiny island on the Pacific.
A whole book is needed to describe that tiny island on the Pacific. I'm sure Nora's knitting needles will suit you. I'm sure any knitting needles will suit you.
Intensifying particles identify the rheme, commonly imparting emotional colouring to the whole of the utterance. Cf.:
Mr. Stores had a part in the general debate. > Even Mr. Stores had a part in the general debate. Then he sat down in one of the arm-chairs. > Only then did he sit down in one of the armchairs. We were impressed by what we heard and saw. > We were so impressed by what we heard and saw.

As for intonation as a means of realising the actual division, it might appear that its sphere is relatively limited, being confined to oral speech only. On closer consideration, however, this view of rheme-identifying role of intonation proves inadequate. To appreciate the true status of intonation in the actual division of the sentence, one should abstract oneself from "paper syntax" (description of written texts) and remember that it is phonetical speech, i.e. articulately pro-nounced utterances that form the basis of human language as a whole. As soon as the phonetical nature of language is duly taken account of, intonation with its accent-patterns presents itself not as a limited, but as a universal and indisputable means of expressing the actual divi-sion in all types and varieties of lingual contexts. This universal rheme-identifying function of intonation has been described in trea-tises on logic, as well as in traditional philological literature, in terms of "logical accent". The "logical accent", which amounts linguisti-cally to the "rhematic accent", is inseparable from the other rheme-identifying means described above, especially from the word-order patterns. Moreover, all such means in written texts in fact represent the logical accent, i.e. they indicate its position either directly or indi-rectly. This can be seen on all the examples hitherto cited in the pre-sent chapter.
5. While recognising the logical accent as a means of effecting the actual division, we must strictly distinguish between the elements immediately placed under the phonetical, "technical" stress, and the sentence segments which are identified as the informative centre of communication in the true sense of the term.
Technically, not only notional, but functional units as well can be phrasally stressed in an utterance, which in modern printed texts is shown by special graphical ways of identification, such as italics, bold type, etc. Cf.:
"I can't bring along someone who isn't invited." "But I am in-vited!" said Miss Casement (I. Murdoch). Moreover, being a highly intelligent young woman, she'd be careful not to be the only one af-fected (. Christie).
However, it would be utterly incorrect to think that in such in-stances only those word-units are logically, i.e. rhematically, marked out as are stressed phonetically. As a matter of fact, functional ele-ments cannot express any self-dependent nomination; they

do not exist by themselves, but make up units of nomination together with the notional elements of utterances whose meanings they specify. Thus, the phrasal phonetical stress, technically making prominent some functional element, thereby identifies as rhematic the corre-sponding notional part ("knot") of the utterance as a whole. It is such notional parts that are real members of the opposition "theme rheme", not their functional constituents taken separately. As for the said functional constituents themselves, these only set up specific se-mantic bases on which the relevant rhematic antitheses are built up.
6. The actual division, since it is effected upon the already pro-duced nominative sentence base providing for its contextually relevant manifestation, enters the predicative aspect of the sentence. It makes up part of syntactic predication, because it strictly meets the functional purpose of predication as such, which is to relate the nominative con-tent of the sentence to reality (see Ch. XXI). This predicative role of the actual division shows that its contextual relevance is not reduced to that of a passive, concomitant factor of expression. On the contrary, the actual division is an active means of expressing functional mean-ings, and, being organically connected with the context, it is not so much context-governed as it is |context-governing: in fact, it does build up concrete contexts out of constructional sentence-models cho-sen to reflect different situations and events.
One of the most important manifestations of the immediate con-textual relevance of the actual division is the regular deletion (ellipsis) of the thematic parts of utterances in dialogue speech. By this syntac-tic process, the rheme of the utterance or its most informative part (peak of informative perspective) is placed in isolation, thereby being very graphically presented to the listener. Cf.:
"You've got the letters?" "In my bag" (G. W. Target). "How did you receive him?" "Coldly" (J. Galsworthy).
In other words, the thematic reduction of sentences in the context, resulting in a constructional economy of speech, performs an informa-tive function in parallel with the logical accent: it serves to accurately identify the rheme of the utterance.

1. The sentence is a communicative unit, therefore the primary classification of sentences must be based on the communicative prin-ciple. This principle is formulated in traditional grammar as the "pur-pose of communication".
The purpose of communication, by definition, refers to the sen-tence as a whole, and the structural features connected with the ex-pression of this sentential function belong to the fundamental, consti-tutive qualities of the sentence as a lingual unit.
In accord with the purpose of communication three cardinal sen-tence-types have long been recognised in linguistic tradition: first, the declarative sentence; second, the imperative (inducive) sentence; third, the interrogative sentence. These communicative sentence-types stand in strict opposition to one another, and their inner proper-ties of form and meaning are immediately correlated with the corre-sponding features of the listener's responses.
Thus, the declarative sentence expresses a statement, either af-firmative or negative, and as such stands in systemic syntagmatic cor-relation with the listener's responding signals of attention, of appraisal (including agreement or disagreement), of fellow-feeling. Cf.:
"I think," he said, "that Mr. Desert should be asked to give us his reasons for publishing that poem." "Hear, hear!" said the . . (J. Galsworthy). "We live very quietly here, indeed we do; my niece here will tell you the same." "Oh, come, I'm not such a fool as that," answered the squire (D. du Maurier).
The imperative sentence expresses inducement, either affirmative or negative. That is, it urges the listener, in the form of request or command, to perform or not to perform a certain action. As such, the imperative sentence is situationally connected with the corresponding "action response" (Ch. Fries), and lingually is systemically correlated with a verbal response showing that the inducement is either com-plied with, or else rejected. Cf.:
"Let's go and sit down up there, Dinny." "Very well" (J. Gals-worthy). "Then marry me." "Really, Alan, I never met anyone with so few ideas" (J. Galsworthy). "Send him back!" he said again. "Nonsense, old chap" (J. Aldridge).

Since the communicative purpose of the imperative sentence is to make the listener act as requested, silence on the part of the latter (when the request is fulfilled), strictly speaking, is also linguistically relevant. This gap in speech, which situationally is filled in by the lis-tener's action, is set off in literary narration by special comments and descriptions. Cf.:
"Knock on the wood." Retan's man leaned forward and knocked three times on the barrera (E. Hemingway). "Shut the piano," whispered Dinny; "let's go up." Diana closed the piano without noise and rose (J. Galsworthy).
The interrogative sentence expresses a question, i.e. a request for information wanted by the speaker from the listener. By virtue of this communicative purpose, the interrogative sentence is naturally con-nected with an answer, forming together with it a question-answer dia-logue unity. Cf.:
"What do you suggest I should do, then?" said Mary help-lessly. "If I were you I should play a waiting game," he replied (D. du Maurier).
Naturally, in the process of actual communication the interroga-tive communicative purpose, like any other communicative task, may sporadically not be fulfilled. In case it is not fulfilled, the question-answer unity proves to be broken; instead of a needed answer the speaker is faced by silence on the part of the listener, or else he re-ceives the latter's verbal rejection to answer. Cf.:
"Why can't you lay off?" I said to her. But she didn't even notice me (R. P. Warren). "Did he know about her?" "You'd better ask him" (S. Maugham).
Evidently, such and like reactions to interrogative sentences are not immediately relevant in terms of environmental syntactic featur-ing.
2. An attempt to revise the traditional communicative classifica-tion of sentences was made by the American scholar Ch. Fries who classed them, as a deliberate challenge to the

"accepted routine", not in accord with the purposes of communication, but according to the responses they elicit [Fries, 29-53].
In Fries's system, as a universal speech unit subjected to commu-nicative analysis was chosen not immediately a sentence, but an utter-ance unit (a "free" utterance, i.e. capable of isolation) understood as a continuous chunk of talk by one speaker in a dialogue. The sentence was then defined as a minimum free utterance.
Utterances collected from the tape-recorded corpus of dialogues (mostly telephone conversations) were first classed into "situation ut-terances" (eliciting a response), and "response utterances". Situation single free utterances (i.e. sentences) were further divided into three groups:
1) Utterances that are regularly followed by oral responses only. These are greetings, calls, questions. E.g.:
Hello! Good-bye! See you soon! ... Dad! Say, dear! Colonel Howard! ... Have you got moved in? What are you going to do for the summer? ...
2) Utterances regularly eliciting action responses. These are re-quests or commands. E.g.:
Read that again, will you? Oh, wait a minute! Please have him call Operator Six when he comes in! Will you see just exactly what his status is?
3) Utterances regularly eliciting conventional signals of attention to continuous discourse. These are statements. E.g.:
I've been talking with Mr. D in the purchasing department about our type-writer. (Yes?). That order went in March seventh. However it seems that we are about eighth on the list. ( I see). Etc.
Alongside of the described "communicative" utterances, i.e. utter-ances directed to a definite listener, another, minor type of utterances were recognised as not directed to any listener but, as Ch. Fries puts it, "characteristic of situations such as surprise, sudden pain, disgust, anger, laughter, sorrow" [Fries, 53]. E.g.: Oh, oh! Goodness! My God! Darn! Gosh! Etc.
Such and like interjectional units were classed by Ch. Fries as "noncommunicative" utterances.
Observing the given classification, it is not difficult to

see that, far from refuting or discarding the traditional classification of sentences built up on the principle of the "purpose of communication", it rather confirms and specifies it. Indeed, the very purpose of commu-nication inherent in the addressing sentence is reflected in the listener's response. The second and third groups of Ch, Fries's "communicative" sentences-utterances are just identical imperative and declarative types both by the employed names and definition. As for the first group, it is essentially heterogeneous, which is recognised by the investigator him-self, who distinguishes in its composition three communicatively dif-ferent subgroups. One of these ("C") is constituted by "questions", i.e. classical interrogative sentences. The other two, viz. greetings ("A") and calls ("B"), are syntactically not cardinal, but, rather, minor inter-mediary types, making up the periphery of declarative sentences (greetings statements of conventional goodwill at meeting and part-ing) and imperative sentences (calls requests for attention). As re-gards "non-communicative" utterances interjectional units, they are devoid of any immediately expressed intellective semantics, which excludes them from the general category of sentence as such (see fur-ther).
Thus, the undertaken analysis should, in point of fact, be looked upon as an actual application of the notions of communicative sen-tence-types to the study of oral speech, resulting in further specifica-tions and development of these notions.
3. Alongside of the three cardinal communicative sentence-types, another type of sentences is recognised in the theory of syntax, namely, the so-called exclamatory sentence. In modern linguistics it has been demonstrated that exclamatory sentences do not possess any complete set of qualities that could place them on one and the same level with the three cardinal communicative types of sentences. The property of exclamation should be considered as an accompanying feature which is effected within the system of the three cardinal com-municative types of sentences.* In other words, each of the cardinal communicative sentence types can be represented in the two variants, viz. non-exclamatory and exclamatory. For instance, with the follow-ing
* See: . ., 1960. , 2. , . I, . 353; 365 .

exclamatory sentences-statements it is easy to identify their non-exclamatory declarative prototypes:
What a very small cabin it was! (K. Mansfield) It was a very small cabin. How utterly she had lost count of events! (J. Galsworthy) < She had lost count of events. Why, if it isn't my lady! (J. Erskine) It is my lady.
Similarly, exclamatory questions are immediately related in the syntactic system to the corresponding non-exclamatory interrogative sentences. E.g.:
Whatever do you mean, Mr. Critchlow? (A. Bennett) -What do you mean? Then why in God's name did you come? (K. Mansfield) - Why did you come?
Imperative sentences, naturally, are characterised by a higher gen-eral degree of emotive intensity than the other two cardinal communi-cative sentence-types. Still, they form analogous pairs, whose con-stituent units are distinguished from each other by no other feature than the presence or absence of exclamation as such. E.g.:
Francis, will you please try to speak sensibly! (E. Hemingway) - Try to speak sensibly. Don't you dare to compare me to common peo-ple! (B. Shaw) < Don't compare me to common people. Never so long as you live say I made you do that! (J. Erskine) < Don't say I made you do that.
As is seen from the given examples, all the three pairs of variant communicative types of sentences (non-exclamatory exclamatory for each cardinal division) make up distinct semantico-syntactic op-positions effected by regular grammatical means of language, such as intonation, word-order and special constructions with functional-auxiliary lexemic elements. It follows from this that the functional-communicative classification of sentences specially distinguishing emotive factor should discriminate, on the lower level of analysis, be-tween the six sentence-types forming, respectively, three groups (pairs) of cardinal communicative quality.
4. The communicative properties of sentences can further be exposed in the light of the theory of actual division of the sentence.
The actual division provides for the informative content of the ut-terance to be expressed with the due gradation of

its parts according to the significance of their respective role in the context. But any utterance is formed within the framework of the sys-tem of communicative types of sentences. And as soon as we compare the communication-purpose aspect of the utterance with its actual di-vision aspect we shall find that each communicative sentence type is distinguished by its specific actual division features, which are re-vealed first and foremost in the nature of the rheme as the meaningful nucleus of the utterance.
The strictly declarative sentence immediately expresses a certain proposition. By virtue of this, the actual division of the declarative sentence presents itself in the most developed and complete form. The rheme of the declarative sentence makes up the centre of some state-ment as such. This can be distinctly demonstrated by a question-test directly revealing the rhematic part of an utterance. Cf.: The next in-stant she had recognised him. > What had she done the next instant?
The pronominal what-question clearly exposes in the example the part "(had) recognised him" as the declarative rheme, for this part is placed within the interrogative-pronominal reference. In other words, the tested utterance with its completed actual division is the only an-swer to the cited potential question; the utterance has been produced by the speaker just to express the fact of "his being recognised".
Another transformational test for the declarative rheme is the logical superposition. The logical superposition consists in transform-ing the tested construction into the one where the rheme is placed in the position of the logically emphasised predicate. By way of example let us take the second sentence in the following sequence: And I was very uneasy. All sorts of forebodings assailed me.
The logical superposition of the utterance is effected thus: > What assailed me was all sorts of forebodings.
This test marks out the subject of the utterance "all sorts of fore-bodings" as the rheme, because it is just this part of the utterance that is placed in the emphatic position of the predicate in the superposi-tional transform.
Similar diagnostic procedures expose the layer-structure of the ac-tual division in composite syntactic constructions. For instance, in the following complex sentence rhematic question-tests easily reveal the three declarative rhemes on the three consecutive syntactic layers: I knew that Mr, Wade had been very excited by something that he had found out.

Test for the first syntactic layer: What did I know?
Test for the second syntactic layer: What state was Mr. Wade in?
Test for the third syntactic layer: What made him excited? (By what was he excited?)
The strictly imperative sentence, as different from the strictly de-clarative sentence, does not express by its immediate destination any statement of fact, i.e. any proposition proper. It is only based on a proposition, without formulating it directly. Namely, the proposition underlying the imperative sentence is reversely contrasted against the content of the expressed inducement, since an urge to do something (affirmative inducement) is founded on the premise that something is not done or is otherwise not affected by the wanted action, and, con-versely, an urge not to do something (negative inducement) is founded on the directly opposite premise. Cf.:
Let's go out at once! (The premise: We are in.) Never again take that horrible woman into your confidence, Jerry! (The premise: Jerry has taken that horrible woman into his confidence.)
Thus, the rheme of the imperative utterance expresses the infor-mative nucleus not of an explicit proposition, but of an inducement a wanted (or unwanted) action together with its referential attending elements (objects, qualities, circumstances).
Due to the communicative nature of the inducement addressed to the listener, its thematic subject is usually zeroed, though it can be represented in the form of direct address. Cf.:
Don't try to sidetrack me (J. Braine). Put that dam* dog down, Fleur; I can't see your face (J. Galsworthy). Kindly tell me what you meant, Wilfrid (J. Galsworthy).
Inducements that include in the address also the speaker himself, or are directed, through the second person medium, to a third person (persons) present their thematic subjects explicit in the construction. E.g.:
I say, Bob, let's try to reconstruct the scene as it developed. Please don't let's quarrel over the speeds now. Let her produce the document if she has it.
The whole composition of an ordinary imperative utterance is usually characterised by a high informative value,
171499 257

so that the rheme proper, or the informative peak, may stand here not so distinctly against the background information as in the declarative utterance. Still, rhematic testing of imperative utterances does dis-close the communicative stratification of their constituents. Compare the question-tests of a couple of the cited examples:
Put that dam' dog down, Fleur. > What is Fleur to do with the dog? Kindly tell me what you meant, Wilfrid. > What is Wilfrid to tell the speaker?
As for the thematic, and especially the subrhematic (transitional) elements of the imperative utterance, they often are functionally charged with the type-grading of inducement itself,i.e.-with making it into a command, prohibition, request, admonition, entreaty, etc. Compare, in addition to the cited, some more examples to this effect:
Let us at least remember to admire each other (L. Hellman). Oh, please stop it... Please, please stop it (E. Hemingway). Get out before I break your dirty little neck (A. Hailey).
The second-person inducement may include the explicit pronomi-nal subject, but such kind of constructions should be defined as of secondary derivation. They are connected with a complicated infor-mative content to be conveyed to the listener-performer, expressing, on the one hand, the choice of the subject out of several persons-participants of the situation, and on the other hand, appraisals render-ing various ethical connotations (in particular, the type-grading of in-ducement mentioned above). Cf.:
"What about me?" she asked. "Nothing doing. You go to bed and sleep" (A. Christie). Don't you worry about me, sir. I shall be all right (B..K. Seymour).
At a further stage of complication, the subject of the inducement may be shifted to the position of the rheme. E.g.:
"...We have to do everything we can." "You do it," he said. "I'm tired" (E. Hemingway).
The essentially different identifications of the rheme in the two imperative utterances of the cited example can be proved by trans-formational testing: ... > What we have to do is (to do) everything we can. ... > The person who should do it is you.
The inducement with the rhematic subject of the latter

type may be classed as the "(informatively) shifted inducement".
5. As far as the strictly interrogative sentence is concerned, its actual division is uniquely different from the actual division of both the declarative and the imperative sentence-types.
The unique quality of the interrogative actual division is deter-mined by the fact that the interrogative sentence, instead of conveying some relatively self-dependent content, expresses an inquiry about information which the speaker (as a participant of a typical question-answer situation) does not possess. Therefore the rheme of the inter-rogative sentence, as the nucleus of the inquiry, is informationally open (gaping); its function consists only in marking the rhematic po-sition in the response sentence and programming the content of its filler in accord with the nature of the inquiry.
Different types of questions present different types of open rhemes.
In the pronominal ("special") question, the nucleus of inquiry is expressed by an interrogative pronoun. The pronoun is immediately connected with the part of the sentence denoting the object or phe-nomenon about which the inquiry ("condensed" in the pronoun) is made. The gaping pronominal meaning is to be replaced in the answer by the wanted actual information. Thus, the rheme of the answer is the reverse substitute of the interrogative pronoun: the two make up a rhematic unity in the broader question-answer construction. As for the thematic part of the answer, it is already expressed in the question, therefore in common speech it is usually zeroed. E.g.:
"Why do you think so?" "Because mostly I keep my eyes open, miss, and I talk to people" (A. Hailey).
The superpositional rhematic test for the pronominal question may be effected in the following periphrastic-definitional form: The question about your thinking so is: why?
For the sake of analytical convenience this kind of superposition may be reduced as follows: > You think so why?
Compare some more pronominal interrogative superpositions:
What happens to a man like Hawk Harrap as the years go by? (W. Saroyan). > To a man like Hawk Harrap, as

the years go by what happens? How do you make that out, mother? (E. M. Forster) > You make that out, mother, how? How's the weather in the north? (D. du Maurier) > The weather in the north how is it? What's behind all this? (A. Hailey) > Behind all this is what?
The rheme of non-pronominal questions is quite different from the one described. It is also open, but its openness consists in at least two semantic suggestions presented for choice to the listener. The choice is effected in the response; in other words, the answer closes the sug-gested alternative according to the interrogative-rhematic program inherent in it. This is clearly seen in the structure of ordinary, explicit alternative questions. E.g.: Will you take it away or open it here? (Th. Dreiser)
The superposition of the utterance may be presented as follows: > You in relation to it will take (it) away, will open (it) here?
The alternative question may have a pronominal introduction, emphasising the open character of its rheme. Cf.: In which cave is the offence alleged, the Buddhist or the Jain? (E. M. Forster)
The superposition: > The offence is alleged in the Buddhist cave, in the Jain cave?
Thus, in terms of rhematic reverse substitution, the pronominal question is a question of unlimited substitution choice, while the al-ternative question is a question of a limited substitution choice, the substitution of the latter kind being, as a rule, expressed implicitly. This can be demonstrated by a transformation applied to the first of the two cited examples of alternative questions: Will you take it away or open it here? > Where will you handle it take it away or open it here?
The non-pronominal question requiring either confirmation or ne-gation ("general" question of yes-no response type) is thereby implic-itly alternative, though the inquiry inherent in it concerns not the choice between some suggested facts, but the choice between the exis-tence or non-existence of an indicated fact. In other words, it is a question of realised rhematic substitution (or of "no substitution choice"), but with an open existence factor (true to life or not true to life?), which makes up its implicitly expressed alternative. This can be easily shown by a superposition; Are they going to stay long? > They are going to stay long, not long?

The implicit alternative question can be made into an explicit one, which as a rule is very emphatic, i.e. stylistically "forced". The nega-tion in the implied alternative part is usually referred to the verb. Cf.: > Are they going to stay long, or are they not going to stay long?
The cited relation of this kind of question to interrogative reverse substitution (and, together with it, the open character of its rheme) is best demonstrated by the corresponding pronominal transformation: > How long are they going to stay long (or not long)?
As we see, the essential difference between the two types of al-ternative questions, the explicit one and the implicit one, remains valid even if the latter is changed into an explicit alternative question (i.e. into a stylistically forced explicit alternative question). This dif-ference is determined by the difference in the informative composi-tion of the interrogative constructions compared.
In general terms of meaning, the question of the first type (the normal explicit alternative question) should be classed as the alterna-tive question of fact, since a choice between two or more facts is re-quired by it; the question of the second type (the implicit alternative question) should be classed as the alternative question of truth, since it requires the statement of truth or non-truth of the indicated fact. In terms of actual division, the question of the first type should be classed as the polyperspective alternative question (biperspective, triperspective, etc.), because it presents more than one informative perspectives (more than one actual divisions) for the listener's choice; the question of the second type, as opposed to the polyperspective, should be classed as the monoperspective alternative question, be-cause its both varieties (implicit and explicit) express only one infor-mative perspective, which is presented to the listener for the existen-tial yes-no appraisal.
6. The exposition of the fundamental role of actual division in the formation of the communicative sentence types involves, among other things, the unequivocal refutation of recognising by some lin-guists the would-be "purely exclamatory sentence" that cannot be re-duced to any of the three demonstrated cardinal communicative types.*
* The existence of the "purely exclamatory sentence" is de-fended, in particular, by B. A. Ilyish in his cited book (pp. 186-187).

Indeed, by "purely exclamatory sentences" are meant no other things than interjectional exclamations of ready-made order such as "Great Heavens!", "Good Lord!", "For God's sake!" "Fiddle-dee-dee!", "Oh, I say!" and the like, which, due to various situational con-ditions, find themselves in self-dependent, proposemically isolated positions in the text. Cf.:
"Oh, for God's sake!" "Oh, for God's sake!" the boy had re-peated (W. Saroyan). "Ah!" said Lady Mont. "That reminds me" (J. Galsworthy).
As is seen from the examples, the isolated positions of the inter-jectional utterances do not make them into any meaningfully articu-late, grammatically predicated sentences with their own informative perspective (either explicit, or implicit). They remain not signals of proposemically complete thoughts, not "communicative utterances" (see above), but mere symptoms of emotions, consciously or uncon-sciously produced shouts of strong feelings. Therefore the highest rank that they deserve in any relevant linguistic classification of "sin-gle free units of speech" is "non-sentential utterances" (which is just another name for Ch. Fries's "noncommunicative utterances").
Of quite another nature are exclamatory sentences with emphatic introducers derived on special productive syntactic patterns. Cf.:
Oh, that Mr. Thornspell hadn't been so reserved! How silly of you! If only I could raise the necessary sum! Etc.
These constructions also express emotions, but they are meaning-fully articulate and proposemically complete. They clearly display a definite nominative composition which is predicated, i.e. related to reality according to the necessary grammatical regularities. And they inevitably belong to quite a definite communicative type of sentences, namely, to the declarative type.
7. The vast set of constructional sentence models possessed by language is formed not only by cardinal, mono-functional communi-cative types; besides these, it includes also intermediary predicative constructions distinguished by mixed communicative features. The true nature of such intermediary constructions can be disclosed in the light of the

actual division theory combined with the general theory of paradig-matic oppositions.
Observations conducted on the said principles show that interme-diary communicative sentence models may be identified between all the three cardinal communicative correlations (viz., statement question, statement inducement, inducement question); they have grown and are sustained in language as a result of the transfer-ence of certain characteristic features from one communicative type of sentences to another.
8. In the following dialogue sequence the utterance which is de-clarative by its formal features, at the same time contains a distinct pronominal question:
"I wonder why they come to me about it. That's your job, sweet-heart." I looked up from Jasper, my face red as fire. "Darling," I said, "I meant to tell you before, but but I forgot" (D. du Maurier).
Semantic-syntactic comparison of the two utterances produced by the participants of the cited dialogue clearly shows in the initial utter-ance the features inherently peculiar to the interrogative communica-tive type, namely, its open rhematic part ("why they come to me about it") and the general programming character of its actual divi-sion in relation to the required response.
Compare some more examples of a similar nature:
"But surely I may treat him as a human being." "Most cer-tainly not" (B. Shaw), "I don't disturb you, I hope, Mr Cokane." "By no means" (B. Shaw). "Wait a second, you haven't told me your address." "Oh, I'm staying at the Hotel du Phare" (A. Christie), "I should like to hear your views on that," replied Utterson (R. L. Ste-venson).
As is seen from the examples, utterances intermediary between statements and questions convey meanings and connotations that sup-plement the direct programming of the answer effected by strictly monofunctional, cardinal interrogative constructions. Namely, they render the connotation of insistency in asking for information, they express a more definite or lass definite supposition of the nature of information possessed by the listener, they present a suggestion to

the listener to perform a certain action or imply a request for permis-sion to perform an action, etc.
On the other hand, in the structural framework of the interrogative sentence one can express a statement. This type of utterance is classed as the "rhetorical question" an expressive construction that has been attracting the closest attention of linguistic observers since an-cient times.
A high intensity of declarative functional meaning expressed by rhetorical questions is best seen in various proverbs and maxims based on this specifically emphatic predicative unit. Cf.:
Can a leopard change his spots? Can man be free if woman be a slave? O shame! Where is thy blush? Why ask the Bishop when the Pope's around? Who shall decide when the doctors disagree?
Compare rhetorical questions in stylistically freer, more common forms of speech:
That was my mission, you imagined. It was not, but where was I to go? (O. Wilde) That was all right; I meant what I said. Why should I feel guilty about it? (J. Braine) How could I have ever thought I could get away with it! (J. Osborne)
It should be noted that in living speech responses to rhetorical questions exactly correspond to responses elicited by declarative sen-tences: they include signals of attention, appraisals, expressions of fellow feeling, etc. Cf.:
"How can a woman be expected to be happy with a man who in-sists on treating her as if she were a perfectly rational being?" "My dear!" (O. Wilde)
A rhetorical question in principle can be followed by a direct an-swer, too. However, such an answer does not fill up the rheme of the rhetorical question (which, as different from the rheme of a genuine question, is not at all open), but emphatically accentuates its intensely declarative semantic nature. An answer to a rhetorical question also emphasises its affirmative or negative implication which is opposite to the formal expression of affirmation or negation in the outer struc-ture of the question. Cf.: "What more can a gentleman desire in this world?" "Nothing more, I am quite sure" (O. Wilde).
Due to these connotations, the answer to a rhetorical

question can quite naturally be given by the speaker himself: Who, being in love, is poor? Oh, no one (O. Wilde).
The declarative nature of the rhetorical question is revealed also in the fact that it is not infrequently used as an answer to a genuine question namely, in cases when an expressive, emphatic answer is needed. Cf.: "Do you expect to save the country, Mr Mangan?" "Well, who else will?" (B. Shaw)
Rhetorical questions as constructions of intermediary communi-cative nature should be distinguished from such genuine questions as are addressed by the speaker to himself in the process of deliberation and reasoning. The genuine quality of the latter kind of questions is easily exposed by observing the character of their rhematic elements. E.g.: Had she had what was called a complex all this time? Or was love always sudden like this? A wild flower seeding on a wild wind? (J. Galsworthy)
The cited string of questions belongs to the inner speech of a lit-erary personage presented in the form of non-personal direct speech. The rhemes of the questions are definitely open, i.e. they are typical of ordinary questions in a dialogue produced by the speaker with an aim to obtain information from his interlocutor. This is clearly seen from the fact that the second question presents an alternative in rela-tion to the first question; as regards the third question, it is not a self-dependent utterance, but a specification, cumulatively attached to the foregoing construction.
Genuine questions to oneself as part of monologue deliberations can quite naturally be followed by corresponding responses, forming various kinds of dialogue within monologue. Cf.:
Was she tipsy, week-minded, or merely in love? Perhaps all three! (J. Galsworthy). My God! What shall I do? I dare not tell her who this woman really is. The shame would kill her (O. Wilde).
9. The next pair of correlated communicative sentence types be-tween which are identified predicative constructions of intermediary nature are declarative and imperative sentences.
The expression of inducement within the framework of a declara-tive sentence is regularly achieved by means of constructions with modal verbs. E.g.:

You ought to get rid of it, you know (C. P. Snow). "You can't come in," he said. "You mustn't get what I have" (E. Hemingway). Well, you must come to me now for anything you want, or I shall be quite cut up (J. Galsworthy). "You might as well sit down," said Ja-votte (J. Erskine).
Compare semantically more complex constructions in which the meaning of inducement is expressed as a result of interaction of dif-ferent grammatical elements of an utterance with its notional lexical elements:
"And if you'll excuse me, Lady Eileen, I think it's time you were going back to bed." The firmness of his tone admitted of no parley (A. Christie). If you have anything to say to me, Dr Trench, I will listen to you patiently. You will then allow me to say what I have to say on my part (B. Shaw).
Inducive constructions, according to the described general ten-dency, can be used to express a declarative meaning complicated by corresponding connotations. Such utterances are distinguished by es-pecially high expressiveness and intensity. E.g.: The Forsyte in him said: "Think, feel, and you're done for!" (J. Galsworthy)
Due to its expressiveness this kind of declarative inducement, similar to rhetorical questions, is used in maxims and proverbs. E.g.:
Talk of the devil and he will appear. Roll my log and I will roll yours. Live and learn. Live and let live.
Compare also corresponding negative statements of the formal imperative order: Don't count your chickens before they are hatched. Don't cross the bridge till you get to it.
10. Imperative and interrogative sentences make up the third pair of opposed cardinal communicative sentence types serving as a frame for intermediary communicative patterns.
Imperative sentences performing the essential function of inter-rogative sentences are such as induce the listener not to action, but to speech. They may contain indirect questions. E.g.:
"Tell me about your upbringing." "I should like to hear about yours" (E. J. Howard). "Please tell me what I can do. There must be something I can do." "You can take the leg off and that might stop it..." (E. Hemingway).

The reverse intermediary construction, i.e. inducement effected in the form of question, is employed in order to convey such additional shades of meaning as request, invitation, suggestion, softening of a command, etc. E.g.:
"Why don't you get Aunt Em to sit instead, Uncle? She's younger than I am any day, aren't you, Auntie?" (J. Galsworthy) "Would would you like to come?" "I would," said Jimmy heartily. "Thanks ever so much, Lady Coote" (A. Christie).
Additional connotations in inducive utterances having the form of questions may be expressed by various modal constructions. E.g.:
Can I take you home in a cab? (W. Saroyan) "Could you tell me," said Dinny, "of any place close by where I could get something to eat?" (J. Galsworthy) I am really quite all right. Perhaps you will help me up the stairs? (A. Christie)
In common use is the expression of inducement effected in the form of a disjunctive question. The post-positional interrogative tag imparts to the whole inducive utterance a more pronounced or less pronounced shade of a polite request or even makes it into a pleading appeal. Cf.:
Find out tactfully what he wants, will you? (J. Tey) And you will come too, Basil, won't you? (O. Wilde)
11. The undertaken survey of lingual facts shows that the com-bination of opposite cardinal communicative features displayed by communicatively intermediary sentence patterns is structurally sys-temic and functionally justified. It is justified because it meets quite definite expressive requirements. And it is symmetrical in so far as each cardinal communicative sentence type is characterised by the same tendency of functional transposition in relation to the two other communicative types opposing it. It means that within each of the three cardinal communicative oppositions two different intermediary communicative sentence models are established, so that at a further level of specification, the communicative classification of sentences should be expanded by six subtypes of sentences of mixed communi-cative features. These are, first, mixed sentence patterns of declaration (interrogative-declarative, imperative-declarative); second, mixed sentence patterns of interrogation (declarative-interrogative, impera-tive-interrogative); third,

mixed sentence-patterns of inducement (declarative-imperative, interrogative-imperative). All the cited intermediary communi-cative types of sentences belong to living, productive syntactic means of language and should find the due reflection both in theoretical linguistic description and in practical language teaching.
1. The basic predicative meanings of the typical English sentence, as has already been pointed out, are expressed by the finite verb which is immediately connected with the subject of the sentence. This predicative connection is commonly referred to as the "predicative line" of the sentence. Depending on their predicative complexity, sentences can feature one predicative line or several (more than one) predicative lines; in other words, sentences may be, respectively, "monopredicative" and "polypredicative". Using this distinction, we must say that the simple sentence is a sentence in which only one predicative line is expressed. E.g.:
Bob has never left the stadium. Opinions differ. This may happen any time. The offer might have been quite fair. Etc.
According to this definition, sentences with several predi-cates referring to one and the same subject cannot be considered as simple. E.g.: I took the child in my arms and held him.
It is quite evident that the cited sentence, although it in-cludes only one subject, expresses two different predicative lines, since its two predicates are separately connected with the subject. The content of the sentence reflects two closely con-nected events that happened in immediate succession: the first "my taking the child in my arms"; the second "my holding him".
Sentences having one verb-predicate and more than one subject to it, if the subjects form actually separate (though in-terdependent) predicative connections, cannot be considered as simple, either. E.g.: The door was open, and also the front win-dow.

Thus, the syntactic feature of strict monopredication should serve as the basic diagnostic criterion for identifying the simple sentence in distinction to sentences of composite structures of various systemic standings.
2. The simple sentence, as any sentence in general, is or-ganised as a system of function-expressing positions, the con-tent of the functions being the reflection of a situational event. The nominative parts of the simple sentence, each occupying a notional position in it, are subject, predicate, object, adverbial, attribute, parenthetical enclosure, addressing enclosure; a spe-cial, semi-notional position is occupied by an interjectional en-closure. The parts are arranged in a hierarchy, wherein all of them perform some modifying role. The ultimate and highest object of this integral modification is the sentence as a whole, and through the sentence, the reflection of the situation (situ-ational event).
Thus, the subject is a person-modifier of the predicate. The predicate is a process-modifier of the subject-person. The ob-ject is a substance-modifier of a processual part (actional or statal). The adverbial is a quality-modifier (in a broad sense) of a processual part or the whole of the sentence (as expressing an integral process inherent in the reflected event). The attribute is a quality-modifier of a substantive part. The parenthetical en-closure is a detached speaker-bound modifier of any sentence-part or the whole of the sentence. The addressing enclosure (address) is a substantive modifier of the destination of the sen-tence and hence, from its angle, a modifier of the sentence as a whole. The interjectional enclosure is a speaker-bound emo-tional modifier of the sentence.
All the said modifiers may be expressed either singly (sin-gle modifiers) or collectively, i.e. in a coordinative combina-tion (co-modifiers, in particular, homogeneous ones).
The traditional scheme of sentence parsing shows many es-sential traits of the said functional hierarchy. On the scheme presented graphically, sentence-parts connected by bonds of immediate domination are placed one under the other in a suc-cessive order of subordination, while sentence-parts related to one another equipotently are placed in a horizontal order. Di-rect connections between the sentence-parts are represented by horizontal and vertical lines.
By way of example, let us take an ordinary English sentence featuring the basic modifier connections, and see its

traditional parsing presentation (Fig. 3): The small lady listened to me attentively.
Fig. 3
The scheme clearly shows the basic logical-grammatical connections of the notional constituents of the sentence. If nec-essary, it can easily be supplemented with specifying linguistic information, such as indications of lexico-grammatical features of the sentence-parts the same as their syntactic sub-functions.
However, observing the given scheme carefully, we must note its one serious flaw. As a matter of fact, while distinctly exposing the subordination ranks of the parts of the sentence, it fails to consistently present their genuine linear order in speech.
This drawback is overcome in another scheme of analysis called the "model of immediate constituents" (contractedly, the "IC-model").
The model of immediate constituents is based on the group-parsing of the sentence which has been developed by traditional grammar together with the sentence-part parsing scheme. It consists in dividing the whole of the sentence into two groups: that of the subject and that of the predicate, which, in their turn, are divided into their sub-group constituents according to the successive subordinative order of the latter. Profiting by this type of analysis, the IC-model explicitly exposes the binary hi-erarchical principle of subordinative connections, showing the whole structure of the sentence as made up by binary immedi-ate constituents. As for equipotent (coordinative) connections, these are, naturally, non-binary, but, being of a more primitive character than subordinative connections, they are included in the

analysis as possible inner subdivisions of subordinative con-nections.
Thus, structured by the IC-model, the cited sentence on the upper level of analysis is looked upon as a united whole (the accepted symbol S); on the next lower level it is divided into two maximal constituents the subject noun-phrase (NP-subj) and the predicate verb-phrase (VP-pred); on the next lower level the subject noun-phrase is divided into the determiner (det) and the rest of the phrase to which it semantically refers (NP), while the predicate noun-phrase is divided into the adver-bial (DP, in this case simply D) and the rest of the verb-phrase to which it semantically refers; the next level-stages of analysis include the division of the first noun-phrase into its adjective-attribute constituent (AP, in this case A) and the noun constitu-ent (N), and correspondingly, the division of the verb-phrase into its verb constituent (V or Vf finite verb) and object noun-phrase constituent (NP-obj), the latter being, finally, di-vided into the preposition constituent (prp) and noun constitu-ent (N). As we see, the process of syntactic IC-analysis contin-ues until the word-level of the sentence is reached, the words being looked upon as the "ultimate" constituents of the sen-tence.
The described model of immediate constituents has two ba-sic versions. The first is known as the "analytical IC-diagrarn", the second, as the "I-derivation tree". The analytical IC-diagram commonly shows the groupings of sentence constitu-ents by means of vertical and horizontal lines (see Fig. 4). The IC-derivation tree shows the groupings of

det NP VP D
NP-subj VP-pred
Fig. 4
sentence constituents by means of branching nodes: the nodes symbolise phrase-categories as unities, while the branches mark their division into constituents of the corresponding sub-categorial standings (see Fig. 5).

3. When analysing sentences in terms of syntagmatic con-nections of their parts, two types of subordinative relations are exposed: on the one hand, obligatory relations, i.e. such as are indispensable for the existence of the syntactic unit as such; on the other hand, optional relations, i.e. such as may or may not be actually represented in the syntactic unit. These relations, as we have pointed out elsewhere, are at present interpreted in terms of syntactic valency (combining power of the word) and are of especial importance for the characteristic of the verb as the central predicative organiser of the notional stock of sen-tence constituents. Comparing the IC-representation of the sen-tence with the pattern of obligatory syntactic positions directly determined by the valency of the verb-predicate, it is easy to see that this pattern reveals the essential generalised model of the sentence, its semantico-syntactic backbone. For instance, in the cited sentence this pattern will be expressed by the string "The lady listened to me", the attribute "small" and the adver-bial "attentively" being the optional parts of the sentence. The IC-model of this key-string of the sentence is logically trans-parent and easily grasped by the mind (see Fig. 6).


Thus, the idea of verbal valency, answering the principle of dividing all the notional sentence-parts into obligatory and op-tional, proves helpful in gaining a further insight into the struc-ture of the simple sentence; moreover, it is of crucial impor-tance for the modern definition of the simple sentence.
In terms of valencies and obligatory positions first of all the category of "elementary sentence" is to be recognised; this is a sentence all the positions of which are obligatory. In other words, this is a sentence which, besides the principal parts, in-cludes only complementive modifiers; as for supplementive modifiers, they find no place in this type of predicative con-struction.
After that the types of expansion should be determined which do not violate the syntactic status of the simple sentence, i.e. do not change the simple sentence into a composite one. Taking into consideration the strict monopredicative character of the simple sentence as its basic identification predicative feature, we infer that such expansions should not complicate the predicative line of the sentence by any additional predica-tive positions.
Finally, bearing in mind that the general identification of obligatory syntactic position affects not only the principal parts of the sentence but is extended to the complementive secondary parts, we define the unexpanded simple sentence as a monopre-dicative sentence formed only by obligatory notional parts. The expanded simple sentence will, accordingly, be defined as a monopredicative sentence which includes, besides the obliga-tory parts, also some optional parts, i.e. some supplementive modifiers which do not constitute a predicative enlargement of the sentence.
Proceeding from the given description of the elementary sentence, it must be stressed that the pattern of this construction presents a workable means of semantico-syntactic analysis of sentences in general. Since all the parts of the elementary sen-tence are obligatory, each real sentence of speech should be considered as categorially reducible to one or more elementary sentences, which expose in an explicit form its logical scheme of formation. As for the simple sentence, however intricate and expanded its structure might be, it is formed, of necessity, upon a single-elementary sentence-base exposing its structural key-model. E.g.: The tall trees by the island shore were shaking vio-lently in the gusty wind.
This is an expanded simple sentence including a number

of optional parts, and its complete analysis in terms of a syn-tagmatic parsing is rather intricate. On the other hand, applying the idea of the elementary sentence, we immediately reveal that the sentence is built upon the key-string "The trees were shak-ing", i.e. on the syntagmatic pattern of an intransitive verb.
As we see, the notions "elementary sentence" and "sentence model" do not exclude each other, but, on the contrary, supple-ment each other: a model is always an abstraction, whereas an elementary sentence can and should be taken both as an abstract category (in the capacity of the "model of an elementary sen-tence") and as an actual utterance of real speech.
4. The subject-group and the predicate-group of the sen-tence are its two constitutive "members", or, to choose a some-what more specific term, its "axes" (in the Russian grammatical tradition ). According as both members are present in the composition of the sentence or only one of them, sentences are classed into "two-member" and "one-member" ones.
Scholars point out that "genuine" one-member sentences are characterised not only as expressing one member in their outer structure; in addition, as an essential feature, they do not imply the other member on the contextual lines. In other words, in ac-cord with this view, elliptical sentences in which the subject or the predicate is contextually omitted, are analysed as "two-member" sentences [Ilyish, 190, 252].
We cannot accept the cited approach because, in our opin-ion, it is based on an inadequate presupposition that in the sys-tem of language there is a strictly defined, "absolute" demarca-tion line between the two types of constructions. In reality, though, each one-member sentence, however pure it might ap-pear from the point of view of non-association with an ellipsis, still, on closer observation, does expose traits of this associa-tion.
For instance, the sentence "Come on!" exemplifying one of the classical one-member sentence varieties, implies a situ-ational person (persons) stimulated to perform an action, i.e. the subject of the event. Similarly, the construction "All right!" rendering agreement on the part of the speaker, is a representa-tive unit standing for a normal two-member utterance in its con-textual-bound implication plane, otherwise it would be sense-less.

Bearing in mind the advanced objection, our approach to the syntactic category of axis part of the sentence is as follows.
All simple sentences of English should be divided into two-axis constructions and one-axis constructions.
In a two-axis sentence, the subject axis and the predicate axis are directly and explicitly expressed in the outer structure. This concerns all the three cardinal communicative types of sentences. E.g.:
The books come out of the experiences. What has been happening here? You better go back to bed.
In a one-axis sentence only one axis or its part is explicitly expressed, the other one being non-presented in the outer struc-ture of the sentence. Cf.:
"Who will meet us at the airport?" "Mary." The response utterance is a one-axis sentence with the subject-axis expressed and the predicate-axis implied: > *Mary will meet us at the airport. Both the non-expression of the predicate and its actual implication in the sub-text are obligatory, since the complete two-axis construction renders its own connotations.
"And what is your opinion of me?" "Hard as nails, abso-lutely ruthless, a born intriguer, and as self-centred as they make 'em." The response utterance is a one-axis sentence with the predicate-axis expressed (partially, by its predicative unit) and the subject-axis (together with the link-verb of the predi-cate) implied: > *You are hard as nails, etc.
"I thought he might have said something to you about it." "Not a word." The response utterance is a one-axis sentence with the predicate-axis partially expressed (by the object) and the subject-axis together with the verbal part of the predicate-axis implied: > *He said not a word to me.
"Glad to see you after all these years!" The sentence is a one-axis unit with the predicate-axis expressed and the subject-axis implied as a form of familiarity: > *I am glad to see you ...
All the cited examples belong to "elliptical" types of utter-ances in so far as they possess quite definite "vacant" positions or zero-positions capable cf being supplied with the correspond-ing fillers implicit in the situational contexts. Since the restora-tion of the absent axis in such sentences is,

So to speak, "free of avail", we class them as free one-axis sentences. The term "elliptical" one-axis sentences can also be used, though it is not very lucky here; indeed, "ellipsis" as a sentence-curtailing process can in principle affect both two-axis and one-axis sentences, so the term might be misleading.
Alongside of the demonstrated free one-axis sentences, i.e. sen-tences with a direct contextual axis-implication, there are one-axis sentences without a contextual implication of this kind; in other words, their absent axis cannot be restored with the same ease and, above all, semantic accuracy.
By way of example, let us read the following passage from S. Maugham's short story "Appearance and Reality";
Monsieur Le Sueur was a man of action. He went straight up to Lisette and smacked her hard on her right cheek with his left hand and then smacked her hard on the left cheek with his right hand. "Brute," screamed Lisette.
The one-axis sentence used by the heroine does imply the you-subject and can, by association, be expanded into the two-axis one "You are a brute" or "You brute", but then the spontaneous "scream-style" of the utterance in the context (a cry of indigna-tion and revolt) will be utterly distorted.
Compare another context, taken from R. Kipling's "The Light that Failed":
"...I'm quite miserable enough already." "Why? Because you're going away from Mrs Jennett?" "No." "From me, then?" No answer for a long time. Dick dared not look at her.
The one-axis sentence "No answer for a long time" in the narra-tive is associated by variant lingua! relations with the two-axis sentence "There was no answer...". But on similar grounds the association can be extended to the construction "He received no answer for a long time" or "No answer was given for a long time" or some other sentence supplementing the given utterance and rendering a like meaning. On the other hand, the peculiar position in the text clearly makes all these associations into re-mote ones: the two-axis version of the construction instead of the existing one-axis one would destroy the expressive property of the remark conveying Dick's strain by means of combining the author's line of narration with the hero's inner perception of events.
Furthermore, compare the psychologically tense description

of packing up before departure given in short, deliberately dis-connected nominative phrase-sentences exposing the heroine's disillusions (from D. du Maurier's "Rebecca"):
Packing up. The nagging worry of departure. Lost keys, unwrit-ten labels, tissue paper lying on the floor. I hate it all.
Associations referring to the absent axes in the cited sentences are indeed very vague. The only unquestionable fact about the relevant implications is that they should be of demonstrative-introductory character making the presented nominals into predicative names.
As we see, there is a continuum between the one-axis sentences of the free type and the most rigid ones exemplified above. Still, since all the constructions of the second order differ from those of the first order just in that they are not free, we choose to class them as "fixed" one-axis sentences.
Among the fixed one-axis sentences quite a few subclasses are to be recognised, including nominative (nominal) constructions, greeting formulas, introduction formulas, incentives, excuses, etc. Many of such constructions are related to the corresponding two-axis sentences not by the mentioned "vague" implication, but by representation; indeed, such one-axis sentence-formulas as affirmations, negations, certain ready-made excuses, etc., are by themselves not word-sentences, but rather sentence-representatives that exist only in combination with the full-sense antecedent predicative constructions. Cf.:
"You can't move any farther back?" "No." (I.e. "I can't move any farther back"). "D'you want me to pay for your drink?" "Yes, old boy." (I.e. "Yes, I want you to pay for my drink, old boy"). Etc.
As for the isolated exclamations of interjectional type ("Good Lord!", "Dear me!" and the like), these are not sentences by virtue of their not possessing the inner structure of actual divi-sion even through associative implications (see Ch. XXII).
Summing up what has been said about the one-axis sentences we must stress the two things: first, however varied, they form a minor set within the general system of English sentence pat-terns; second, they all are related to two-axis sentences either by direct or by indirect association.

5. The semantic classification of simple sentences should be effected at least on the three bases: first, on the basis of the subject categorial meanings; second, on the basis of the predi-cate categorial meanings; third, on the basis of the subject-object relation.
Reflecting the categories of the subject, simple sentences are divided into personal and impersonal. The further division of the personal sentences is into human and non-human; hu-man into definite and indefinite; non-human into animate and inanimate. The further essential division of impersonal sen-tences is into factual (It rains, It is five o'clock) and percep-tional (It smells of hay here).
The differences in subject categorial meanings are sustained by the obvious differences in subject-predicate combinability.
Reflecting the categories of the predicate, simple sentences are divided into process-featuring ("verbal") and, in the broad sense, substance-featuring (including substance as such and substantive quality "nominal"). Among the process-featuring sentences actional and statal ones are to be discriminated (The window is opening The window is glistening in the sun); among the substance-featuring sentences factual and percep-tional ones are to be discriminated (The sea is rough The place seems quiet).
Finally, reflecting the subject-object relation, simple sen-tences should be divided into subjective (John lives in London), objective (John reads a book) and neutral or "potentially" ob-jective (John reads), capable of implying both the transitive ac-tion of the syntactic person and the syntactic person's intransi-tive characteristic.
1. Traditional grammar studied the sentence from the point of view of its syntagmatic structure: the sentence was ap-proached as a string of certain parts fulfilling the corresponding syntactic functions. As for paradigmatic relations, which, as we know, are inseparable from syntagmatic relations, they were explicitly revealed only as part of morphological descriptions, because, up to recent times, the idea of the sentence-model with its functional variations was not

developed. Moreover, some representatives of early modern linguistics, among them F. de Saussure, specially noted that it was quite natural for morphology to develop paradigmatic (as-sociative) observations, while syntax "by its very essence" should concern itself with the linear connections of words.
Thus, the sentence was traditionally taken at its face value as a ready unit of speech, and systemic connections between sentences were formulated in terms of classifications. Sentences were studied and classified according to the purpose of com-munication, according to the types of the subject and the predi-cate, according to whether they are simple or composite, ex-panded or unexpanded, compound or complex, etc.
In contemporary modern linguistics paradigmatic structur-ing of lingual connections and dependencies has penetrated into the would-be "purely syntagmatic" sphere of the sentence. The paradigmatic approach to this element of rendering communi-cative information, as we have mentioned before, marked a new stage in the development of the science of language; in-deed, it is nothing else than paradigmatic approach that has provided a comprehensive theoretical ground for treating the sentence not only as a ready unit of speech, but also and above all as a meaningful lingual unit existing in a pattern form.
2. Paradigmatics finds its essential expression in a system of oppositions making the corresponding meaningful (func-tional) categories. Syntactic oppositions are realised by corre-lated sentence patterns, the observable relations between which can be described as "transformations", i.e, as transitions from one pattern of certain notional parts to another pattern of the same notional parts. These transitions, being oppositional, at the same time disclose derivational connections of sentence-patterns. In other words, some of the patterns are to be ap-proached as base patterns, while others, as their transforms.
For instance, a question can be described as transformation-ally produced from a statement; a negation, likewise, can be presented as transformationally produced from an affirmation. E.g.:
You are fond of the kid. > Are you fond of the kid? You are fond of the kid. > You are not fond of the kid.
Why are the directions of transitions given in this way

and not vice versa? Simply because the ordinary affirmative statement presents a positive expression of a fact in its purest form, maximally free of the speaker's connotative appraisals.
Similarly, a composite sentence, for still more evident rea-sons, is to be presented as derived from two or more simple sentences. E.g.:
He turned to the waiter.+The waiter stood in the door. > He turned to the waiter who stood in the door.
These transitional relations are implicitly inherent in the syntagmatic classificational study of sentences. But modern theory, exposing them explicitly, has made a cardinal step for-ward in so far as it has interpreted them as regular derivation stages comparable to categorial form-making processes in mor-phology and word-building.
And it is on these lines that the initial, basic element of syn-tactic derivation has been found, i.e. a syntactic unit serving as a "sentence-root" and providing an objective ground for identi-fying syntactic categorial oppositions. This element is known by different names, such as the "basic syntactic pattern", the "structural sentence scheme", the "elementary sentence model", the "base sentence", though as the handiest in linguistic use should be considered the "kernel sentence" due to its termino-logical flexibility combined with a natural individualising force.
Structurally the kernel sentence coincides with the elemen-tary sentence described in the previous chapter. The difference is, that the pattern of the kernel sentence is interpreted as form-ing the base of a paradigmatic derivation in the corresponding sentence-pattern series.
Thus, syntactic derivation should not be understood as an immediate change of one sentence into another one; a pro-nounced or written sentence is a finished utterance that thereby cannot undergo any changes. Syntactic derivation is to be un-derstood as paradigmatic production of more complex pattern-constructions out of kernel pattern-constructions as their struc-tural bases. The description of this production ("generation") may be more detailed and less detailed, i.e. it can be effected in more generalised and less generalised terms, depending on the aim of the scholar. The most concrete presentation concerns a given speech-utterance analysed into its derivation history on the level of the word-forms.

By way of example let us take the following English sen-tence: I saw him come.
This sentence is described in school grammar as a sentence with a complex object, which is syntagmatically adequate, though incomplete from the systemic point of view. The syn-tagmatic description is supplemented and re-interpreted within the framework of the paradigmatic description presenting the sentence in question as produced from the two kernel sen-tences: I saw him. + He came. > I saw him come.

The same may be given in terms of the IC-derivation tree diagrams (see Fig. 7). The indices specifying the basic sym-

In a more generalised, categorial-oriented paradigmatic presentation the sentence will be shown as a transformational combination of the two kernel pattern-formulas:
bols can vary in accord with the concrete needs of analysis and demonstration.
3. The derivation of genuine sentences lying on the "sur-face" of speech out of kernel sentences lying in the "deep base" of speech can be analysed as a process falling into sets of ele-mentary transformational steps or procedures. These procedures make up six major classes.
The first class includes steps of "morphological arrange-ment" of the sentence, i.e. morphological changes expressing syntactically relevant categories, above all, the predicative categories of the finite verb: tense, aspect, voice, mood. The syntactic role of these forms of morphological change (sys-tematised into morphological paradigms) consists in the fact

that they make up parts of the more general syntactico-paradigmatic series. E.g.:
John+start (the kernel base string) > John starts. John will be starting. John would be starting. John has started. Etc.
The second class of the described procedures includes vari-ous uses of functional words (functional expansion). From the syntactic point of view these words are transformers of syntac-tic constructions in the same sense as the categorial morphemes {e.g. inflexions) are transformers of lexemes, i.e. morphological constructions. E.g.:
He understood my request. > He seemed to understand my request. Now they consider the suggestion. > Now they do consider the suggestion.
The third class of syntactic derivational procedures includes the processes of substitution. Among the substitutes we find personal pronouns, demonstrative-substitute pronouns, indefi-nite-substitute pronouns, as well as substitutive combinations of half-notional words. Cf.:
The pupils ran out of the classroom. > They ran out of the classroom. I want another pen, please. > I want another one, please.
The fourth class of the procedures in question is formed by processes of deletion, i.e. elimination of some elements of the sentence in various contextual conditions. As a result of dele-tion the corresponding reduced constructions are produced. E.g.:
Would you like a cup of tea? > A cup of tea? It's a pleas-ure! > Pleasure!
The fifth class of syntactic derivational procedures includes processes of positional arrangement, in particular, permuta-tions (changes of the word-order into the reverse patterns). E.g.:
The man is here. > Is the man here? Jim ran in with an ex-cited cry. In ran Jim with an excited cry.
The sixth class of syntactic derivational procedures is formed by processes of intonational arrangement, i.e. applica-tion of various functional tones and accents. This arrangement is represented in written and typed speech by

punctuation marks, the use of different varieties of print, the use of various modes of underlining and other graphical means. E.g.:
We must go. > We must go? We? Must go?? You care nothing about what I feel. > You care nothing about what I feel!
The described procedures are all functionally relevant, i.e. they serve as syntactically meaningful dynamic features of the sentence. For various expressive purposes they may be applied either singly or, more often than not, in combination with one another. E.g.: We finish the work. > We are not going to finish it.
For the production of the cited sentence-transform the fol-lowing procedures are used: morphological change, introduc-tion of functional words, substitution, intonational arrangement. The functional (meaningful) outcome of the whole process is the expression of the modal future combined with a negation in a dialogue response. Cf.:
Are we ever going to finish the work? > Anyway, we are not going to finish it today!
4. The derivational procedures applied to the kernel sen-tence introduce it into two types of derivational relations in the sentential paradigmatic system: first, the "constructional" rela-tions; second, the "predicative" relations. The constructional derivation effects the formation of more complex clausal struc-tures out of simpler ones; in other words, it provides for the ex-pression of the nominative-notional syntactic semantics of the sentence. The predicative derivation realises the formation of predicatively different units not affecting the constructional volume of the base; in other words, it is responsible for the ex-pression of the predicative syntactic semantics of the sentence. Both types of derivational procedures form the two subsystems within the general system of syntactic paradigmatics.
5. As part of the constructional system of syntactic para-digmatics, kernel sentences, as well as other, expanded base-sentences undergo derivational changes into clauses and phrases.
The transformation of a base sentence into a clause can be called "clausalisation". By way of clausalisation a

sentence is changed into a subordinate or coordinate clause in the process of subordinative or coordinative combination of sentences. The main clausalising procedures involve the use of conjunctive words subordinators and coordinators. Since a composite sentence is produced from minimum two base sen-tences, the derivational processes of composite sentence pro-duction are sometimes called "two-base transformations". For example, two kernel sentences "They arrived" and "They re-lieved me of my fears" (> I was relieved of my fears), com-bined by subordinative and coordinative clausalising, produce the following constructions:
> When they arrived I was relieved of my fears. > If they arrive, I shall be relieved of my fears. > Even though they ar-rive, I shan't be relieved of my fears. Etc. > They arrived, and I was relieved of my fears. > They arrived, but I was not re-lieved of my fears. Etc.
The transformation of a base sentence into a phrase can be called "phrasalisation". By phrasalisation a sentence is trans-formed either into a semi-predicative construction (a semi-clause), or into a nominal phrase.
Nominal phrases are produced by the process of nominalisa-tion, i.e. nominalising phrasalisation which we have analyzed before (see Ch. XX). Nominalisation may be complete, consist-ing in completely depriving the sentence of its predicative as-pect, or partial, consisting in partially depriving the sentence of its predicative aspect. Partial nominalisation in English pro-duces infinitive and gerundial phrases. By other types of phra-salisation such semi-clauses are derived as complex objects of infinitive and participial types, various participial constructions of adverbial status and some other, minor complexes. The re-sulting constructions produced by the application of the cited phrasalising procedures in the process of derivational combina-tion of base sentences will be both simple expanded sentences (in case of complete nominalisation) and semi-composite sen-tences (in case of various partial nominalisations and other phrasalisations). Cf.:
On their arrival I was relieved of my fears. They arrived to relieve me of my fears. > They arrived relieving me of my fears. > Having arrived, they did relieve me of my fears. Etc.

As is seen from the examples, each variety of derivational combination of concrete sentences has its own semantic pur-pose expressed by the procedures employed.
6. As part of the predicative system of syntactic paradig-matics, kernel sentences, as well as expanded base-sentences, undergo such structural modifications as immediately express the predicative functions of the sentence, i.e. the functions re-lating the nominative meanings of the sentence to reality. Of especial importance in this respect is the expression of predica-tive functions by sentences which are elementary as regards the set of their notional constituents: being elementary from the point of view of nominative semantics, these sentences can be used as genuine, ordinary utterances of speech. Bearing in mind the elementary nominative nature of its constructional units, we call the system of sentences so identified the "Primary Syntac-tic System" (Lat. "Prima Systema Syntactica").
To recognise a primary sentence in the text, one must use the criteria of elementary sentence-structure identification ap-plied to the notional constituents of the sentence, irrespective of the functional meanings rendered by it. For instance, the no-tionally minimal negative sentence should be classed as pri-mary, though not quite elementary (kernel) in the paradigmatic sense, negation being not a notional, but a functional sentence factor. Cf.:
I have met the man. > I have not met the man. > I have never met the man.
Any composite (or semi-composite) sentence is analysable into two or more primary sentences (i.e. sentences elementary in the notional sense). E.g.:
Is it a matter of no consequence that I should find you with a young man wearing my pyjamas? - Is it a matter of no con-sequence?+I should find you with a (young) man.+ The (young) man is wearing my pyjamas.
The kernel sentence can also have its representation in speech, being embodied by the simplest sentential construction not only in the notional, but also in the functional sense. In other words, it is an elementary sentence which is non-interrogative, non-imperative, non-negative, non-modal, etc. In short, in terms of syntactic oppositions, this is the "weakest"

construction in the predicative oppositional space of the pri-mary syntactic system.
7. The predicative functions expressed by primary sen-tence patterns should be divided into the two types: first, lower functions; second, higher functions. The lower functions in-clude the expression of such morphological categories as tenses and aspects; these are of "factual", "truth-stating" semantic character. The higher functions are "evaluative" in the broad sense of the word; they immediately express the functional se-mantics of relating the nominative content of the sentence to reality.
The principal predicative functions expressed by syntactic categorial oppositions are the following.
First, question as opposed to statement. Second, inducement as opposed to statement. Third, negation as opposed to affirma-tion. Fourth, unreality as opposed to reality. Fifth, probability as opposed to fact. Sixth, modal identity (seem to do, happen to do, prove to do, etc.) as opposed to fact. Seventh, modal sub-ject-action relation as opposed to fact (can do, may do, etc.). Eighth, specified actual subject-action relation as opposed to fact. Ninth, phase of action as opposed to fact. Tenth, passive action as opposed to active action. Eleventh, specialised actual division (specialised perspective) as opposed to non-specialised actual division (non-specialised perspective). Twelfth, emphasis (emotiveness) as opposed to emotional neutrality (unemotive-ness).
Each opposition of the cited list forms a categorial set which is rather complex. For instance, within the framework of the question-statement opposition, pronominal and alternative ques-tions are identified with their manifold varieties; within the sys-tem of phase of action, specialised subsets are identified render-ing the phase of beginning, the phase of duration, the phase of end, etc. The total supersystem of all the pattern-forms of a given sentence base constitutes its general syntactic paradigm of predicative functions. This paradigm is, naturally, extremely complicated so that it is hardly observable if presented on a dia-gram. This fact shows that the volume of functional meanings rendered by a sentence even on a very high level of syntactic generalisation is tremendous. At the same time the derivation of each functional sentence-form in its paradigmatically deter-mined position in the system is simple enough in the sense that it is quite explicit. This shows the dynamic essence of the para-digm

in question; the paradigm exactly answers the needs of expres-sion at every given juncture of actual communication.
8. All the cited oppositions-categories may or may not be represented in a given utterance by their strong function-members. In accord with this oppositional regularity, we ad-vance the notion of the "predicative load" of the sentence. The predicative load is determined by the total volume of the strong members of predicative oppositions (i.e. by the sum of positive values of the corresponding differential features) actually repre-sented in the sentence.
The sentence, by definition, always expresses predication, being a predicative unit of language. But, from the point of view of the comparative volume of the predicative meanings actually expressed, the sentence may be predicatively "loaded" or "non-loaded". If the sentence is predicatively "non-loaded", it means that its construction is kernel elementary on the accepted level of categorial generalisation. Consequently, such a sentence will be characterised in oppositional terms as non-interrogative, non-inducive, non-negative, non-real, non-probable, non-modal-identifying, etc., down to the last of the recognised predicative oppositions. If, on the other hand, the sentence is predicatively "loaded", it means that it renders at least one of the strong op-positional meanings inherent in the described categorial system. Textual observations show that predicative loads amounting to one or two positive feature values (strong oppositional mem-bers) may be characterised as more or less common; hence, we consider such a load as "light" and, correspondingly, say that the sentence in this case is predicatively "lightly" loaded. As for sentences whose predicative load exceeds two positive feature values, they stand out of the common, their functional semantics showing clear signs of intricacy. Accordingly, we consider such loads as "heavy", and of sentences characterised by these loads we say that they are "heavily" loaded. Predicative loads amount-ing to four feature values occur but occasionally, they are too complicated to be naturally grasped by the mind.
To exemplify the cited theses, let us take as a derivation sentence-base the construction "The thing bothers me". This sentence, in the above oppositional sense, is predicatively "non-loaded", or has the "zero predicative load". The predicative structure of the sentence can be expanded by the expression of the modal subject-action relation, for instance,

the ability relation. The result is: > "The thing can bother me"; the predicative load of the sentence has grown to 1. This con-struction, in its turn, can be used as a derivation base for a sen-tence of a higher predicative complexity; for instance, the fea-ture of unreality can be added to it: > "The thing could bother me (now)". The predicative load of the sentence has grown to 2. Though functionally not simple, the sentence still presents a more or less ordinary English construction. To continue with our complicating it, we may introduce in the sentence the fea-ture of passivity: > "I could be bothered (by the thing now)". The predicative semantics expressed has quite clearly changed into something beyond the ordinary; the sentence requires a special context to sound natural. Finally, to complicate the pri-mary construction still further, we may introduce a negation in it: > "I could not be bothered (by the thing now)". As a result we are faced by a construction that, in the contextual conditions of real speech, expresses an intricate set of functional meanings and stylistic connotations. Cf.:
"...Wilmet and Henrietta Bentworth have agreed to differ already." "What about?" "Well, I couldn't be bothered, but I think it was about the P.M., or was it Portulaca? they differ about everything" (J. Galsworthy).
The construction is indeed semantically complicated; but all its meaningful complexity is linguistically resolved by the demonstrated semantico-syntactic oppositional analysis show-ing the stage-to-stage growth of the total functional meaning of the sentence in the course of its paradigmatic derivation.
1. The composite sentence, as different from the simple sentence, is formed by two or more predicative lines. Being a polypredicative construction, it expresses a complicated act of thought, i.e. an act of mental activity which falls into two or more intellectual efforts closely combined with one another. In terms of situations and events this means that the composite sentence reflects two or more elementary

situational events viewed as making up a unity; the constitutive connections of the events are expressed by the constitutive con-nections of the predicative lines of the sentence, i.e. by the sen-tential polypredication.
Each predicative unit in a composite sentence makes up a clause in it, so that a clause as part of a composite sentence cor-responds to a separate sentence as part of a contextual se-quence. E.g.:
When I sat down to dinner I looked for an opportunity to slip in casually the information that I had by accident run across the Driffields; but news travelled fast in Blackstable (S. Maugham).
The cited composite sentence includes four clauses which are related to one another on different semantic grounds. The sentences underlying the clauses are the following:
I sat down to dinner. I looked for an opportunity to slip in casually the information. I had by accident run across the Drif-fields. News travelled fast in Blackstable.
The correspondence of a predicative clause to a separate sentence is self-evident. On the other hand, the correspondence of a composite sentence to a genuine, logically connected se-quence of simple sentences (underlying its clauses) is not evi-dent at all; moreover, such kind of correspondence is in fact not obligatory, which is the very cause of the existence of the com-posite sentence in a language. Indeed, in the given example the independent sentences reconstructed from the predicative clauses do not make up any coherently presented situational unity; they are just so many utterances each expressing an event of self-sufficient significance. By way of rearrangement and the use of semantic connectors we may make them into a more or less explanatory situational sequence, but the exposition of the genuine logic of events, i.e. their presentation as natural parts of a unity, achieved by the composite sentence will not be, and is not to be replaced in principle. Cf.:
I ran by accident across the Driffields. At some time later on I sat down to dinner. While participating in the general con-versation, I looked for an opportunity to slip in casually the in-formation about my meeting them. But news travelled fast in Blackstable.
The logical difference between the given composite sen-tence and its contextually coherent de-compositional
19-1499 289

presentation is, that whereas the composite sentence exposes as its logical centre, i.e. the core of its purpose of communication, the intention of the speaker to inform his table-companions of a certain fact (which turns out to be already known to them), the sentential sequence expresses the events in their natural tempo-ral succession, which actually destroys the original purpose of communication. Any formation of a sentential sequence more equivalent to the given composite sentence by its semantic status than the one shown above has to be expanded by addi-tional elucidative prop-utterances with back-references; and all the same, the resulting contextual string, if it is intended as a real informational substitute for the initial composite, will hardly be effected without the help of some kind of essentially composite sentence constructions included in it (let the reader himself try to construct an equivalent textual sequence meeting the described semantic requirements).
As we see, the composite sentence in its quality of a struc-tural unit of language is indispensable for language by its own purely semantic merits, let alone its terseness, as well as intel-lectual elegance of expression.
2. As is well known, the use of composite sentences, espe-cially long and logically intricate ones, is characteristic of liter-ary written speech rather than colloquial oral speech. This un-questionable fact is explained by the three reasons: one relating to the actual needs of expression; one relating to the possibili-ties of production; and one relating to the conditions of percep-tion.
That the composite sentence structure answers the special needs of written mode of lingual expression is quite evident. It is this type of speech that deals with lengthy reasonings, de-scriptions, narrations, all presenting abundant details of intricate correlations of logical premises and inferences, of situational foreground and background, of sequences of events interrupted by cross-references and parenthetical comments. Only a com-posite sentence can adequately and within reasonable bounds of textual space fulfil these semantic requirements.
Now, the said requirements, fortunately, go together with the fact that in writing it is actually possible to produce long composite sentences of complicated, but logically flawless structure (the second of the advanced reasons). This is possible here because the written sentence, while in the process of being

produced, is open to various alterations: it allows corrections of slips and errors; it can be subjected to curtailing or expanding; it admits of rearranging and reformulating one's ideas; in short, it can be prepared. This latter factor is of crucial importance, so that when considering the properties of literary written speech we must always bear it in mind. Indeed, from the linguistic point of view written speech is above all prepared, or "edited" speech: it is due to no other quality than being prepared before its presentation to the addressee that this mode of speech is structurally so tellingly different from colloquial oral speech. Employing the words in their broader sense, we may say that literary written speech is not just uttered and gone, but is al-ways more carefully or less carefully composed in advance, be-ing meant for a future use of the reader, often for his repeated use. In distinction to this, genuine colloquial oral speech is ut-tered each time in an irretrievably complete and final form, each time for one immediate and fleeting occasion.
We have covered the first two reasons explaining the com-posite sentence of increased complexity as a specific feature of written speech. The third reason, referring to the conditions of perception, is inseparable from the former two. Namely, if writ-ten text provides for the possibility for its producer to return to the beginning of each sentence with the aim of assessing its form and content, of rearranging or re-composing it altogether, it also enables the reader, after he has run through the text for the first time, to go back to its starting line and re-read it with as much care as will be required for the final understanding of each item and logical connection expressed by its wording or implied by its construction. Thus, the length limit imposed on the sentence by the recipient's immediate (operative) memory can in writing be practically neglected; the volume of the writ-ten sentence is regulated not by memory limitations as such, but by the considerations of optimum logical -balance and sty-listic well-formedness.
3. Logic and style being the true limiters of the written sentence volume, two dialectically contrasted active tendencies can be observed in the sentence-construction of modern printed texts. According to the first tendency, a given unity of reasons in meditation, a natural sequence of descriptive situations or narrative events is to be reflected in one composite sentence, however long and structurally complicated
19 291

it might prove. According to the second, directly opposite ten-dency, for a given unity of reflected events or reasons, each of them is to be presented by one separate simple sentence, the whole complex of reflections forming a multisentential para-graph. The two tendencies are always in a state of confronta-tion, and which of them will take an upper hand in this or that concrete case of text production has to be decided out of vari-ous considerations of form and meaning relating to both contex-tual and con-situational conditions (including, among other things, the general purpose of the work in question, as well as the preferences and idiosyncrasies of its users).
Observe, for instance, the following complex sentence of mixed narrative-reasoning nature:
Once Mary waved her hand as she recognised her driver, but he took no notice of her, only whipping his horses the harder, and she realised with a rather helpless sense of futility that so far as other people were concerned she must be consid-ered in the same light as her uncle, and that even if she tried to walk to Boduin or Launceston no one would receive her, and the door would be shut in her face (D. du Maurier).
The sentence has its established status in the expressive context of the novel, and in this sense it is unrearrangeable. On the other hand, its referential plane can be rendered by a mul-tisentential paragraph, plainer in form, but somewhat more natural to the unsophisticated perceptions:
Once Mary recognised her driver. She waved her hand to him. But he took no notice of her. He only whipped his horses the harder. And she realised that so far as other people were concerned she must be considered in the same light as her un-cle. This gave her a rather helpless sense of futility. Even if she tried to walk to Boduin or Launceston no one would receive her. Quite the contrary, the door would be shut in her face.
One long composite sentence has been divided into eight short sentences. Characteristically, though, in our simplification we could not do without the composite sentence structure as such: two of the sentential units in the adaptation (respectively, the fourth and the sixth) have retained their compositive fea-tures, and these structural properties seem

to be indispensable for the functional adequacy of the rear-ranged passage.
The cited example of syntactic re-formation of text will help us formulate the following composition rule of good non-fiction (neutral) prose style: in neutral written speech each sen-tence construction should be as simple as can be permitted by the semantic context.
4. We have emphatically pointed out in due course (see Ch. I) the oral basis of human language: the primary lingual matter is phonetical, so that each and every lingual utterance given in a graphic form has essentially a representative charac-ter, its speech referent being constructed of so many phones organised in a rhythmo-melodical sequence. On the other hand, and this has also been noted before, writing in a literary lan-guage acquires a relatively self-sufficient status in so far as a tremendous proportion of what is actually written in society is not meant for an oral reproduction at all: though read and re-read by those to whom it has been addressed, it is destined to remain "silent" for ever. The "silent" nature of written speech with all its peculiarities leads to the development of specifi-cally written features of language, among which, as we have just seen, the composite sentence of increased complexity oc-cupies one of the most prominent places. Now, as a natural consequence of this development, the peculiar features of writ-ten speech begin to influence oral speech, whose syntax be-comes liable to display ever more syntactic properties directly borrowed from writing.
Moreover, as a result of active interaction between oral and written forms of language, a new variety of speech has arisen that has an intermediary status. This type of speech, being ex-plicitly oral, is at the same time prepared and edited, and more often than not it is directly reproduced from the written text, or else from its epitomised version (theses). This intermediary written-oral speech should be given a special linguistic name, for which we suggest the term "scripted speech", i.e. speech read from the script. Here belong such forms of lingual com-munication as public report speech, lecturer speech, preacher speech, radio- and television-broadcast speech, each of them existing in a variety of subtypes.
By way of example let us take the following passage from President Woodrow Wilson's address to the Congress urging it to authorise the United States' entering the World War (1917):

But the right is more precious than peace, and we shall fight for the things which we have always carried nearest our hearts, for democracy, for the right of those who submit to authority to have a voice in their own governments, for the rights and liberties of small nations, for a universal dominion of right by such a concert of free peoples as shall bring peace and safety to all nations and make the world itself at last free.
The text presents a typical case of political scripted speech with a clear tinge of solemnity, its five predicative units being complicated by parallel constructions of homogeneous objects (for-phrases) adding to its high style emphasis.
Compare the above with a passage from President Franklin D. Roosevelt's second inaugural address (1937):
In this nation I see tens of millions of its citizens a sub-stantial part of its whole population who at this very mo-ment are denied the greater part of what the very lowest stan-dards of today call the necessities of life.
The sentence is not a long one, but its bookish background, although meant for oral uttering before an audience, is most evident: a detached appositional phrase, consecutive subordina-tion, the very nature of the last appositional clausal complex of commenting type, all these features being carefully prepared to give the necessary emphasis to the social content of the utter-ance aimed at a public success.
Compare one more example a passage from Bernard Shaw's paper read before the Medico-Legal Society in London (1909):
Nevertheless, trade in medical advice has never been for-mally recognised, and never will be; for you must realise that, whereas competition in ordinary trade and business is founded on an elaborate theoretic demonstration of its benefits, there has never been anyone from Adam Smith to our own time who has attempted such a demonstration with regard to the medical pro-fession. The idea of a doctor being a tradesman with a pecuni-ary interest in your being ill is abhorrent to every thoughtful person.
The scripted nature of the cited sentential sequence is clearly seen from its arrangement as an expressive climax built upon a carefully balanced contrastive composite construction.

5. We have hitherto defended the thesis of the composite sentence of increased complexity being specifically characteris-tic of literary written speech. On the other hand, we must clearly understand that the composite sentence as such is part and par-cel of the general syntactic system of language, and its use is an inalienable feature of any normal expression of human thought in intercourse. This is demonstrated by cases of composite sen-tences that could not be adequately reduced to the correspond-ing sets of separate simple sentences in their natural contexts (see above). Fictional literature, presenting in its works a reflec-tion of language as it is spoken by the people, gives us abundant illustrations of the broad use of composite sentences in genuine colloquial speech both of dialogue and monologue character.
Composite sentences display two principal types of con-struction: hypotaxis (subordination) and parataxis (coordina-tion). Both types are equally representative of colloquial speech, be it refined by education or not. In this connection it should be noted that the initial rise of hypotaxis and parataxis as forms of composite sentences can be traced back to the early stages of language development, i. e. to the times when language had no writing. Profuse illustrations of the said types of syntactic rela-tions are contained, for instance, in the Old English epic "Beo-wulf" (dated presumably from the VII A. D.). As is known, the text of the poem shows all the basic forms of sentential composition including the grammatically completed presenta-tion of reported speech, connection of clauses on various nomi-nal principles (objective, subjective, predicative, attributive), connection of clauses on various adverbial principles (temporal, local, conditional, causal, etc.). E. g.:

* From: Beowulf/Ed. by A. J. Wyatt. New edition revised with introduc-tion and notes by R. W, Chambers. Cambr., 1933, verses 590- 597.

Compare the tentative prose translation of the cited text into Modern English (with the corresponding re-arrangements of the word-order patterns):
Truly I say onto thee, oh Son Egglaf, that never would Grendel, the abominable monster, have done so many terrible deeds to your chief, (so many) humiliating acts in Heorot, if thy soul (and) heart had been as bold as thou thyself declarest; but he has found that he need not much fear the hostile sword-attack of your people, the Victorious Skildings.
Needless to say, the forms of composite sentences in pre-writing periods of language history cannot be taken as a proof that the structure of the sentence does not develop historically in terms of perfecting its expressive qualities. On the contrary, the known samples of Old English compared with their modern rendering are quite demonstrative of the fact that the sentence does develop throughout the history of language; moreover, they show that the nature and scope of the historical structural change of the sentence is not at all a negligible matter. Namely, from the existing lingual materials we see that the primitive, not clearly identified forms of subordination and coordination, without distinct border points between separate sentences, have been succeeded by such constructions of syntactic composition as are distinguished first and foremost by the clear-cut logic of connections between their clausal predicative parts. However, these materials, and among them the cited passage, show us at the same time that the composite sentence, far from being ex-traneous to colloquial speech, takes its origin just in the oral colloquial element of human speech as such: it is inherent in the very oral nature of developing language.
6. The two main types of the connection of clauses in a composite sentence, as has been stated above, are subordination and coordination. By coordination the clauses are arranged as units of syntactically equal rank, i. equipotently; by subordi-nation, as units of unequal rank, one being categorially domi-nated by the other. In terms of the positional structure of the sentence it means that by subordination one of the clauses (sub-ordinate) is placed in a notional position of the other (principal). This latter characteristic has an essential semantic implication clarifying the difference

between the two types of polypredication in question. As a mat-ter of fact, a subordinate clause, however important the infor-mation rendered by it might be for the whole communication, presents it as naturally supplementing the information of the principal clause, i.e. as something completely premeditated and prepared even before its explicit expression in the utterance. This is of especial importance for post-positional subordinate clauses of circumstantial semantic nature. Such clauses may of-ten shift their position without a change in semantico-syntactic status. Cf.:
I could not help blushing with embarrassment when I looked at him. > When I looked at him I could not help blush-ing with embarrassment. The board accepted the decision, though it didn't quite meet their plans. > Though the decision didn't quite meet their plans, the board accepted it.
The same criterion is valid for subordinate clauses with a fixed position in the sentence. To prove the subordinate quality of the clause in the light of this consideration, we have to place it in isolation and see that the isolation is semantically false. E.g.:
But all the books were so neatly arranged, they were so clean, that I had the impression they were very seldom read.> *But all the books were so neatly arranged, they were so clean. That I had the impression they were very seldom read. I fancy that life is more amusing now than it was forty years ago. > *I fancy that life is more amusing now. Than it was forty years ago.
As for coordinated clauses, their equality in rank is ex-pressed above all in each sequential clause explicitly corre-sponding to a new effort of thought, without an obligatory fea-ture of premeditation. In accord with the said quality, a sequen-tial clause in a compound sentence refers to the whole of the leading clause, whereas a subordinate clause in a complex sen-tence, as a rule, refers to one notional constituent (expressed by a word or a phrase) in a principal clause [Khaimovich, Rogov-skaya, 278]. It is due to these facts that the position of a coordi-nate clause is rigidly fixed in all cases, which can be used as one of the criteria of coordination in distinction to subordination. Another probe of rank equality of clauses in coordination is a potential possibility for any coordinate sequential clause to take either the copulative

conjunction and or the adversative conjunction but as introdu-cers. Cf.:
That sort of game gave me horrors, so I never could play it. > That sort of game gave me horrors, and I never could play it. The excuse was plausible, only it was not good enough for us. > The excuse was plausible, but it was not good enough for us.
7. The means of combining clauses into a polypredicative sentence are divided into syndetic, i. e. conjunctional, and asyn-detic, i. e. non-conjunctional. The great controversy going on among linguists about this division concerns the status of syn-deton and asyndeton versus coordination and subordination. Namely, the question under consideration is whether or not syndeton and asyndeton equally express the two types of syn-tactic relations between clauses in a composite sentence.
According to the traditional view, all composite sentences are to be classed into compound sentences (coordinating their clauses) and complex sentences (subordinating their clauses), syndetic or asyndetic types of clause connection being specifi-cally displayed with both classes. However, this view has of late been subjected to energetic criticism; the new thesis formu-lated by its critics is as follows: the "formal" division of clause connection based on the choice of connective means should be placed higher in the hierarchy than the "semantic" division of clause connection based on the criterion of syntactic rank. That is, on the higher level of classification all the composite sen-tences should be divided into syndetic and asyndetic, while on the lower level the syndetic composite sentences (and only these) should be divided into compound and complex ones in accord with the types of the connective words used. The cited principle was put forward by N. S. Pospelov as part of his syn-tactic analysis of Russian, and it was further developed by some other linguists.
But the new approach to coordination and subordination has not been left unchallenged. In particular, B. A. Ilyish with his characteristic discretion in formulating final decisions has pointed out serious flaws in the non-traditional reasoning result-ing first of all from mixing up strictly grammatical criteria of classification with general semantic considerations [Ilyish, 318 ff.].
Indeed, if we compare the following asyndetic composite

sentences with their compound syndetic counterparts on the ba-sis of paradigmatic approach, we shall immediately expose un-questionable equality in their semantico-syntactic status. E. g.:
My uncle was going to refuse, but we didn't understand why.> My uncle was going to refuse, we didn't understand why. She hesitated a moment, and then she answered him. > She hesitated a moment, then she answered him.
The equality of the compound status of both types of sen-tences is emphatically endorsed when compared with the corre-sponding complex sentences in transformational constructional paradigmatics. Cf.:
... > We didn't understand why my uncle was going to re-fuse. ... > After she hesitated a moment she answered him.
On the other hand, bearing in mind the in-positional nature of a subordinate clause expounded above, it would be alto-gether irrational to deny a subordinate status to the asyndetic attributive, objective or predicative clauses of the commonest order. Cf.:
They've given me a position I could never have got without them. > They've given me a position which I could never have got without them. We saw at once it was all wrong. > We saw at once that it was all wrong The fact is he did accept the invita-tion. > The fact is that he did accept the invitation.
Now, one might say, as is done in some older grammatical treatises, that the asyndetic introduction of a subordinate clause amounts to the omission of the conjunctive word joining it to the principal clause. However, in the light of the above para-digmatic considerations, the invalidity of this statement in the context of the discussion appears to be quite obvious: as regards the "omission" or "non-omission" of the conjunctive introducer the compound asyndetic sentence should be treated on an equal basis with the complex asyndetic sentence. In other words, if we defend the idea of the omission of the conjunction with asyndetic subordinate clauses, we must apply this principle also to asyndetic coordinate clauses. But the idea of the omission of the conjunction expounded in its purest, classical form has al-ready been demonstrated in linguistics as fallacious, since asyndetic

connection of clauses is indisputably characterised by its own functional value; it is this specific value that vindicates and supports the very existence of asyndetic polypredication in the system of language. Moreover, many true functions of asyn-detic polypredication in distinction to the functions of syndetic polypredication were aptly disclosed in the course of investiga-tions conducted by the scholars who sought to refute the ade-quacy of coordinate or subordinate interpretation of clausal asyndeton. So, the linguistic effort of these scholars, though not convincing in terms of classification, has, on the whole, not been in vain; in the long run, it has contributed to the deeper insight into the nature of the composite sentence as a polypre-dicative combination of words.
8. Besides the classical types of coordination and subordi-nation of clauses, we find another case of the construction of composite sentence, namely, when the connection between the clauses combined in a polypredicative unit is expressly loose, placing the sequential clause in a syntactically detached posi-tion. In this loosely connected composite, the sequential clause information is presented rather as an afterthought, an idea that has come to the mind of the speaker after the completion of the foregoing utterance, which latter, by this new utterance-forming effort, is forcibly made into the clausal fore-part of a composite sentence. This kind of syntactic connection, the traces of which we saw when treating the syntagmatic bonds of the word, comes under the heading of cumulation. Its formal sign is often the tone of sentential completion followed by a shorter pause than an inter-sentential one, which intonational complex is rep-resented in writing by a semi-final punctuation mark, such as a semicolon, a dash, sometimes a series of periods. Cf.-.
It was just the time that my aunt and uncle would be coming home from their daily walk down the town and I did not like to run the risk of being seen with people whom they would not at all approve of; so I asked them to go on first, as they would go more quickly than I (S. Maugham).
Cumulation as here presented forms a type of syntactic con-nection intermediary between clausal connection and sentential connection. Thus, the very composite sentence

(loose composite) formed by it is in fact a unit intermediary be-tween one polypredicative sentence and a group of separate sentences making up a contextual sequence.
There is good reason to interpret different parenthetical clauses as specific cumulative constructions, because the basic semantico-syntactic principle of joining them to the initially planned sentence is the same, i. e. presenting them as a de-tached communication, here of an introductory or comment-ing-deviational nature. E.g.:
He was sent for very suddenly this morning, as I have told you already, and he only gave me the barest details before his horse was saddled and he was gone (D. du Maurier). Unprece-dented in scale and lavishly financed (? 100,000 was collected in 1843 and 9,000,000 leaflets distributed) this agitation had all the advantages that the railways, cheap newspapers and the penny post could give (A. L. Morton).
If this interpretation is accepted, then the whole domain of cumulation should be divided into two parts: first, the continu-ative cumulation, placing the cumulated clause in post-position to the expanded predicative construction; second, the" paren-thetical cumulation, placing the cumulated clause in inter-position to the expanded predicative construction. The inter-position may be made even into a pre-position as its minor par-ticular case (here belong mostly constructions introduced by the conjunction as: as we have seen, as I have said, etc.). This paradox is easily explained by the type of relation between the clauses: the parenthetical clause (i. e. parenthetically cumu-lated) only gives a background to the essential information of the expanded original clause. And, which is very important, it can shift its position in the sentence without causing any change in the information rendered by the utterance as a whole. Cf.:
He was sent for very suddenly this morning, as I have told you already. > He was sent for, as I have told you already, very suddenly this morning. > As I have told you already, he was sent for very suddenly this morning.
9. In the composite sentences hitherto surveyed the consti-tutive predicative lines are expressed separately and explicitly: the described sentence types are formed by minimum two clauses each having a subject and a predicate of its own. Alongside of these "completely" composite sentences,

there exist constructions in which one explicit predicative line is combined with another one, the latter being not explicitly or completely expressed. To such constructions belong, for in-stance, sentences with homogeneous predicates, as wall as sen-tences with verbid complexes. Cf.:
Philip ignored the question and remained silent. I have never before heard her sing. She followed him in, bending her head under the low door.
That the cited utterances do not represent classical, explic-itly constructed composite sentence-models admits of no argu-ment. At the same time, as we pointed out elsewhere (see Ch. XXIV), they cannot be analysed as genuine simple sentences, because they contain not one, but more than one predicative lines, though presented in fusion with one another. This can be demonstrated by explanatory expanding transformations. Cf.:
... > Philip ignored the question, (and) he remained silent. ... > I have never before heard how she sings. ... > As she fol-lowed him in, she bent her head under the low door.
The performed test clearly shows that the sentences in ques-tion are derived each from two base sentences, so that the sys-temic status of the resulting constructions is in fact intermedi-ary between the simple sentence and the composite sentence. Therefore these predicative constructions should by right be analysed under the heading of semi-composite sentences.
It is easy to see that functionally semi-composite sentences are directly opposed to composite-cumulative sentences: while the latter are over-expanded, the former are under-expanded, i. e. they are concisely deployed. The result of the predicative blend is terseness of expression, which makes semi-composite constructions of especial preference in colloquial speech.
Thus, composite sentences as polypredicative constructions exist in the two type varieties as regards the degree of their predicative explicitness: first, composite sentences of complete composition; second, composite sentences of concise composi-tion. Each of these types is distinguished by its own functional specification, occupies a permanent place in the syntactic sys-tem of language and so deserves a separate consideration in a grammatical description.

1. The complex sentence is a polypredicative construction built up on the principle of subordination. It is derived from two or more base sentences one of which performs the role of a matrix in relation to the others, the insert sentences. The matrix function of the corresponding base sentence may be more rig-orously and less rigorously pronounced, depending on the type of subordinative connection realised.
When joined into one complex sentence, the matrix base sentence becomes the principal clause of it and the insert sen-tences, its subordinate clauses.
The complex sentence of minimal composition includes two clauses a principal one and a subordinate one. Although the principal clause positionally dominates the subordinate clause, the two form a semantico-syntactic unity within the framework of which they are in fact interconnected, so that the very exis-tence of either of them is supported by the existence of the other.
The subordinate clause is joined to the principal clause ei-ther by a subordinating connector (subordinator), or, with some types of clauses, asyndetically. The functional character of the subordinative connector is so explicit that even in traditional grammatical descriptions of complex sentences this connector was approached as a transformer of an independent sentence into a subordinate clause. Cf.:
Moyra left the room. > (I do remember quite well) that Moyra left the room. > (He went on with his story) after Moyra left the room. > (Fred remained in his place) though Moyra left the room. > (The party was spoilt) because Moyra left the room. > (It was a surprise to us all) that Moyra left the room...
This paradigmatic scheme of the production of the subordi-nate clause vindicates the possible interpretation of contact-clauses in asyndetic connection as being joined to the principal clause by means of the "zero"-connector. Cf.: (How do you know) 0 Moyra left the room?
Needless to say, the idea of the zero-subordinator simply stresses the fact of the meaningful (functional) character of the asyndetic connection of clauses, not denying the actual absence of connector in the asyndetic complex sentence.

The minimal, two-clause complex sentence is the main vol-ume type of complex sentences. It is the most important type, first, in terms of frequency, since its textual occurrence by far exceeds that of multi-clause complex sentences; second, in terms of its paradigmatic status, because a complex sentence of any volume is analysable into a combination of two-clause complex sentence units.
2. The structural features of the principal clause differ with different types of subordinate clauses. In particular, vari-ous types of subordinate clauses specifically affect the principal clause from the point of view of the degree of its completeness. As is well known from elementary grammatical descriptions, the principal clause is markedly incomplete in complex sen-tences with the subject and predicative subordinate clauses. E.g.:
And why we descend to their level is a mystery to me. (The gaping principal part outside the subject clause: " is a mys-tery to me".) Your statement was just what you were expected to say. (The gaping principal part outside the predicative clause: "Your statement was just ")
Of absolutely deficient character is the principal clause of the complex sentence that includes both subject and predicative subordinate clauses: its proper segment, i. e. the word-string standing apart from the subordinate clauses is usually reduced to a sheer finite link-verb. Cf.: How he managed to pull through is what baffles me. (The principal clause representation: " is ")
A question arises whether the treatment of the subject and predicative clauses as genuinely subordinate ones is rational at all. Indeed, how can the principal clause be looked upon as syn-tactically (positionally) dominating such clauses as perform the functions of its main syntactic parts, in particular, that of the subject? How can the link-verb, itself just a little more than an auxiliary element, be taken as the "governing predicative con-struction" of a complex sentence?
However, this seeming paradox is to be definitely settled on the principles of paradigmatic theory. Namely, to understand the status of the "deficiently incomplete and gaping" principal clause we must take into consideration the matrix nature of the principal clause in the sentence: the matrix presents the upper-level positional scheme which is to be completed by predicative constructions on the lower level.

In case of such clauses as subject and predicative, these are all the same subordinated to the matrix by way of being its embed-ded elements, i. e. the fillers of the open clausal positions intro-duced by it. Since, on the other hand, the proper segment of the principal clause, i. e. its "nucleus", is predicatively deficient, the whole of the clause should be looked upon as merged with the corresponding filler-subordinate clauses. Thus, among the principal clauses there should be distinguished merger principal clauses and non-merger principal clauses, the former character-ising complex sentences with clausal deployment of their main parts, the latter characterising complex sentences with clausal deployment of their secondary parts.
3. The principal clause dominates the subordinate clause positionally, but it doesn't mean that by its syntactic status it must express the central informative part of the communication. The information perspective in the simple sentence does not repeat the division of its constituents into primary and secon-dary, and likewise the information perspective of the complex sentence is not bound to duplicate the division of its clauses into principal and subordinate. The actual division of any con-struction, be it simple or otherwise, is effected in the context, so it is as part of a continual text that the complex sentence makes its clauses into rheme-rendering and theme-rendering on the complex-sentence information level.
When we discussed the problem of the actual division of the sentence, we pointed out that in a neutral context the rhematic part of the sentence tends to be placed somewhere near the end of it (see Ch. XXII, 4). This holds true both for the simple and complex sentences, so that the order of clauses plays an impor-tant role in distributing primary and secondary information among them. Cf.: The boy was friendly with me because I al-lowed him to keep the fishing line.
In this sentence approached as part of stylistically neutral text the principal clause placed in the front position evidently expresses the starting point of the information delivered, while the subordinate clause of cause renders the main sentential idea, namely, the speaker's explanation of the boy's attitude. The "contraposition" presupposed by the actual division of the whole sentence is then like this: "Otherwise the boy wouldn't have been friendly". Should the clause-order of the utterance

be reversed, the informative roles of the clauses will be re-shaped accordingly: As I allowed the boy to keep the fishing line, he was friendly with me.
Of course, the clause-order, the same as word-order in gen-eral, is not the only means of indicating the correlative informa-tive value of clauses in complex sentences; intonation plays here also a crucial role, and it goes together with various lexical and constructional rheme-forming elements, such as emphatic particles, constructions of meaningful antithesis, patterns of logical accents of different kinds.
Speaking of the information status of the principal clause, it should be noted that even in unemphatic speech this predicative unit is often reduced to a sheer introducer of the subordinate clause, the latter expressing practically all the essential informa-tion envisaged by the communicative purpose of the whole of the sentence. Cf.:
You see that mine is by far the most miserable lot. Just fancy that James has proposed to Mary! You know, kind sir, that I am bound to fasting and abstinence.
The principal clause-introducer in sentences like these per-forms also the function of keeping up the conversation, i.e. of maintaining the immediate communicative connection with the listener. This function is referred to as "phatic". Verbs of speech and especially thought are commonly used in phatic principals to specify "in passing" the speaker's attitude to the information rendered by their rhematic subordinates:
I think there's much truth in what we hear about the matter. I' sure I can't remember her name now.
Many of these introducer principals can be re-shaped into parenthetical clauses on a strictly equivalent basis by a mere change of position:
I can't remember her name now, I' sure. There's much truth, I think, in what we hear about the matter.
4. Of the problems discussed in linguistic literature in con-nection with the complex sentence, the central one concerns the principles of classification of subordinate clauses. Namely, the two different bases of classification are considered as competi-tive in this domain: the first is functional, the second is cate-gorial.

In accord with the functional principle, subordinate clauses are to be classed on the analogy of the positional parts of the simple sentence, since it is the structure of the simple sentence that underlies the essential structure of the complex sentence (located on a higher level). In particular, most types of subordi-nate clauses meet the same functional question-tests as the parts of the simple sentence. The said analogy, certainly, is far from being absolute, because no subordinate clause can exactly re-peat the specific character of the corresponding non-clausal part of the sentence; moreover, there is a deep difference in the functional status even between different categorial types of the same parts of the sentence, one being expressed by a word-unit, another by a word-group, still another by a substitute. Cf.:
You can see my state. > You can see my wretched state. > You can see my state being wretched. > You can see that my state is wretched. > You can see that. What can you see?
Evidently, the very variety of syntactic forms united by a central function and separated by specific sub-functions is brought about in language by the communicative need of ex-pressing not only rough and plain ideas, but also innumerable variations of thought reflecting the ever developing reality.
Furthermore, there are certain (and not at all casual) clauses that do not find ready correspondences among the non-clausal parts of the sentence at all. This concerns, in particular, quite a number of adverbial clauses.
Still, a general functional analogy (though not identity) be-tween clausal and lexemic parts of the sentence does exist, and, which is very important, it reflects the underlying general simi-larity of their semantic purpose. So, the functional classification of subordinate clauses on the simple sentence-part analogy does reflect the essential properties of the studied syntactic units and has been proved useful and practicable throughout many years of application to language teaching.
Now, in accord with the categorial principle, subordinate clauses re to be classed by their inherent nominative properties irrespective of their immediate positional relations in the sen-tence. The nominative properties of notional words are reflected in their part-of-speech classification. A question
20* 307

arises, can there be any analogy between types of subordinate clauses and parts of speech?
One need not go into either a detailed research or heated ar-gument to see that no direct analogy is possible here. This is made clear by the mere reason that a clause is a predicative unit expressing an event, while a lexeme is a pure naming unit used only as material for the formation of predicative units, both in-dependent and dependent.
On the other hand, if we approach the categorial principle of the characterisation of clauses on a broader basis than drawing plain part-of-speech analogies, we shall find it both plausible and helpful.
As a matter of fact, from the point of view of their general nominative features all the subordinate clauses can be divided into three categorial-semantic groups. The first group includes clauses that name an event as a certain fact. These pure fact-clauses may be terminologically defined as "substantive-nominal". Their substantive-nominal nature is easily checked by a substitute test:
That his letters remained unanswered annoyed him very much. > That fact annoyed him very much. The woman knew only too well what was right and what was wrong. > The woman knew those matters well.
The second group of clauses also name an event-fact, but, as different from the first group, this event-fact is referred to as giving a characteristic to some substantive entity (which, in its turn, may be represented by a clause or a phrase or a substantive lexeme). Such clauses, in compliance with our principle of choosing explanatory terminology, can be tentatively called "qualification-nominal"'. The qualification-nominal nature of the clauses in question, as is the case with the first group of clauses, is proved through the corresponding replacement pat-terns:
The man who came in the morning left a message. > That man left a message. Did you find a place where we could make a fire? > Did you find such kind of place?
Finally, the third group of clauses make their event-nomination into a dynamic relation characteristic of another, event or a process or a quality of various descriptions. In keep-ing with the existing practices, it will be quite natural to call these clauses "adverbial". Adverbial clauses are best

tested not by a replacement, but by a definitive transformation. Cf.:
Describe the picture as you see it. > Describe the picture in the manner you see it. All will be well if we arrive in time. > All will be well on condition that we arrive in time.
5. When comparing the two classifications in the light of the systemic principles, it is easy to see that only by a very su-perficial observation they could be interpreted as alternative (i. e. contradicting each other). In reality they are mutually com-plementary, their respective bases being valid on different lev-els of analysis. The categorial features of clauses go together with their functional sentence-part features similar to the cate-gorial features of lexemes going together with their functional characteristics as parts of the simple sentence.
Subordinate clauses are introduced by functional connective words which effect their derivation from base sentences. Cate-gorially these sentence subordinators (or subordinating clausalisers) fall into the two basic types: those that occupy a notional position in the derived clause, and those that do not occupy such a position. The non-positional subordinators are referred to as pure conjunctions. Here belong such words as since, before, until, if, in case, because, so that, in order that, though, however, than, as if, etc. The positional subordinators are in fact conjunctive substitutes. The main positional subordi-nators are the pronominal words who, what, whose, which, that, where, when, why, as. Some of these words are double-functional (bifunctional), entering also the first set of subordina-tors; such are the words where, when, that, as, used both as con-junctive substitutes and conjunctions. Together with these the zero subordinator should be named, whose polyfunctional status is similar to the status of the subordinator that. The substitute status of positional subordinators is disclosed in their function as "relative" pronominals, i. e. pronominals referring to syn-tagmatic antecedents. Cf.:
That was the day when she was wearing her pink dress. Sally put on her pink dress when she decided to join the party downstairs.
The relative pronominal "when" in the first of the cited sen-tences syntagmatically replaces the antecedent "the day",

while the conjunction "when" in the second sentence has no relative pronominal status. From the point of view of paradig-matics, though, even the second "when" cannot be understood as wholly devoid of substitute force, since it remains associated systemically with the adverb "then", another abstract indicator of time. So, on the whole the non-substitute use of the double-functional subordinators should be described not as utterly "non-positional", but rather as "semi-positional".
On the other hand, there is another aspect of categorial dif-ference between the subordinators, and this directly corresponds to the nature of clauses they introduce. Namely, nominal clauses, being clauses of fact, are introduced by subordinators of fact (conjunctions and conjunctive subordinators), while ad-verbial clauses, being clauses of adverbial relations, are intro-duced by subordinators of relational semantic characteristics (conjunctions). This difference holds true both for monofunc-tional subordinators and bifunctional subordinators. Indeed, the subordinate clauses expressing time and place and, correspond-ingly, introduced by the subordinators when and where may be used both as nominal nominators and adverbial nominators. The said difference is quite essential, though outwardly it remains but slightly featured. Cf.:
I can't find the record where you put it yesterday. I forget where I put the record yesterday.
It is easy to see that the first place-clause indicates the place of action, giving it a situational periphrastic definition, while the second place-clause expresses the object of a mental effort. Accordingly, the subordinator "where" in the first sentence in-troduces a place description as a background of an action, while the subordinator "where" in the second sentence introduces a place description as a fact to be considered. The first "where" and the second "where" differ by the force of accent (the first is unstressed, the second is stressed), but the main marking differ-ence between them lies in the difference between the patterns of their use, which difference is noted by the chosen terms "nomi-nal" and "adverbial". This can easily be illustrated by a ques-tion-replacement test: ... > Where can't I find the record? ...> What do I forget?
Likewise, the corresponding subdivision of the nominal

subordinators and the clauses they introduce can be checked and proved on the same lines. Cf.:
The day when we met is unforgettable. > Which day is un-forgettable? When we met is of no consequence now. > What is of no consequence now?
The first when-ttrn is clearly disclosed by the test as a qualification-nominal, while the second, as a substantive-nominal.
Thus, the categorial classification of clauses is sustained by the semantic division of the subordinators which are distin-guished as substantive-nominal clausalisers, qualification-nominal clausalisers and adverbial clausalisers. Since, on the other hand, substantive nomination is primary in categorial rank, while qualification nomination is secondary, in terms of syntactic positions all the subordinate clauses are to be divided into three groups: first, clauses of primary nominal positions to which belong subject, predicative and object clauses; second, clauses of secondary nominal positions to which belong at-tributive clauses; third, clauses of adverbial positions.
6. Clauses of primary nominal positions subject, predi-cative, object are interchangeable with one another in easy reshufflings of sentence constituents. Cf.:
What you saw at the exhibition is just what I want to know. > What I want to know is just what you saw at the exhibition. > I just want to know what you saw at the exhibition.
However, the specific semantic functions of the three re-spective clausal positions are strictly preserved with all such interchanges, so that there is no ground to interpret positional rearrangements like the ones shown above as equivalent.
The subject clause, in accord with its functional position, regularly expresses the theme on the upper level of the actual division of the complex sentence. The thematic property of the clause is well exposed" in its characteristic uses with passive constructions, as well as constructions in which the voice oppo-sition is neutralised. E.g.:
Why he rejected the offer has never been accounted for. What small reputation the town does possess derives from two things.

It should be noted that in modern colloquial English the formal position of the subject clause in a complex sentence is open to specific contaminations (syntactic confusions on the clausal level). Here is one of the typical examples: Just because you say I wouldn't have (seen a white elephant M. B.) doesn't prove anything (E.Hemingway).
The contamination here consists in pressing into one con-struction the clausal expression of cause and the expression of the genuine theme-subject to which the predicate of the sen-tence refers. The logical implication of the statement is, that the event in question cannot be taken as impossible by the mere reason of the interlocutor's considering it as such. Thus, what can be exposed of the speaker's idea by way of "de-contaminating" the utterance is approximately like this: Your saying that I wouldn't have doesn't prove anything.
Another characteristic type of syntactic contamination of the subject-clause pattern is its use as a frame for an independ-ent sentence. E. g.: You just get yourselves into trouble is what happens (M. Bradbury).
The cited contamination presents a feature of highly emo-tional speech. The utterance, as it were, proves to be a living illustration of the fact that where strong feelings are concerned the logic of lingual construction is liable to be trespassed upon. The logic in question can be rehabilitated by a substitution pat-tern: You just get yourselves into trouble, this is what happens.
As is known, the equivalent subject-clausal function can be expressed by the construction with an anticipatory pronoun (mostly the anticipatory it). This form of expression, emphasis-ing the rheme-clause of the sentence, at the same time presents the information of the subject clause in a semantically stronger position than the one before the verb. Therefore the anticipatory construction is preferred in cases when the content of the sub-ject clause is not to be wholly overbalanced or suppressed by the predicate of the sentence. E. g.: How he managed to pull through is a miracle. It is a miracle how he managed to pull through.
Some scholars analyse the clause introduced by the anticipa-tory construction as presenting two possibilities of interpreta-tion which stand in opposition to each other. Accord-ing to the first and more traditional view, this is just a subject clause in-troduced by the anticipatory it, while in the light of the second, the clause introduced by it is appositive,
In our opinion, the latter explanation is quite rational; how-ever, it cannot be understood as contrary to the "anticipatory" theory. Indeed, the appositive type of connection between the introducer it and the introduced clause is proved by the very equivalent transformation of the non-anticipatory construction into the anticipatory one; but the exposition of the appositive character of the clause does not make the antecedent it into something different from an introductory pronominal element. Thus, the interpretation of the subject clause referring to the introducer it as appositive, in fact, simply explains the type of syntactic connection underlying the anticipatory formula.
The predicative clause, in conformity with the predicative position as such, performs the function of the nominal part of the predicate, i. e. the part adjoining the link-verb. The link-verb is mostly expressed by the pure link be, not infrequently we find here also the specifying links seem and look; the use of other specifying links is occasional. E. g.:
The trouble is that I don't know Fanny personally. The question is why the decision on the suggested innovation is still delayed. The difficulty seems how we shall get in touch with the chief before the conference. After all those years of travel-ling abroad, John has become what you would call a man of will and experience.
Besides the conjunctive substitutes, the predicative clause, the same as other nominal clauses, can be introduced by some conjunctions (that, whether, as if, as though). The predicative clause introduced by the conjunctions as if, as though has an adverbial force, which is easily shown by contrast: She looks as though she has never met him. > She behaves as though she has never met him.
While considering subordinate clauses relating to the finite be in the principal clause, care should be taken to strictly dis-criminate between the linking and non-linking (notional) rep-resentations of the verb. Indeed, the linking be is naturally fol-lowed by a predicative clause, while the notional be, featuring verbal semantics of existence, cannot join a predicative. Cf.:
It's because he's weak that he needs me. This was because, he had just arrived.
The cited sentences have been shown by B. A. Ilyish as examples of predicative clauses having a non-conventional

nominal-clause conjunction (Ilyish, 276-2771. However, the analysis suggested by the scholar is hardly acceptable, since the introducing be in both examples does not belong to the class of links.
The predicative clause in a minimal complex sentence regu-larly expresses its rheme. Therefore there is an essential infor-mative difference between the two functional uses of a cate-gorially similar nominal clause: that of the predicative and that of the subject. Cf.:
The impression is that he is quite competent. That he is quite competent is the impression.
The second sentence (of an occasional status, with a sen-tence-stress on the link-verb), as different from the first, sug-gests an implication of a situational antithesis: the impression may be called in question, or it may be contrasted against an-other trait of the person not so agreeable as the one mentioned, etc.
The same holds true of complex sentences featuring subor-dinate clauses in both subject and predicative positions. Cf.:
How she gets there is what's troubling me (> I am trou-bled). What's troubling me is how she gets there (> How is she to get there?).
The peculiar structure of this type of sentence, where two nominal clauses are connected by a short link making up all the outer composition of the principal clause, suggests the scheme of a balance. For the sake of convenient terminological dis-crimination, the sentence may be so called a "complex bal-ance".
The third type of clauses considered under the heading of clauses of primary nominal positions are object clauses.
The object clause denotes an object-situation of the process expressed by the verbal constituent of the principal clause.
The object position is a strong substantive position in the sentence. In terms of clausal relations it means that the substan-tivising force of the genuine object-clause derivation is a strongly pronounced nominal clause-type derivation. This is revealed, in particular, by the fact that object clauses can be in-troduced not only non-prepositionally, but also, if not so freely, prepositionally. Cf.;

They will accept with grace whatever he may offer. She stared at what seemed a faded photo of Uncle Jo taken half a century before. I am simply puzzled by what you are telling me about the Car fairs.
On the other hand, the semantic content of the object clause discriminates three types of backgrounds: first, an immediately substantive background; second, an adverbial background; third, an uncharacterised background of general event. This differentiation depends on the functional status of the clause-connector, that is on the sentence-part role it performs in the clause. Cf.:
We couldn't decide whom we should address. The friends couldn't decide where they should spend their vacation.
The object clause in the first of the cited sentences is of a substantive background (We should address whom), whereas the object clause in the second sentence is of adver-bial-local background (They should spend their vacation where).
The plot of the novel centred on what might be called a far-fetched, artificial situation. The conversation centred on why that clearly formulated provision of international law had been violated.
The first object clause in the above two sentences is of sub-stantive background, while the second one is of an adverbial-causal background.
Object clauses of general event background are introduced by conjunctions: Now he could prove that the many years he had spent away from home had not been in vain.
The considered background features of subordinate clauses, certainly, refer to their inner status and therefore concern all the nominal clauses, not only object ones. But with object clauses they are of especial contrastive prominence, which is due to immediate dependence of the object clause on the va-lency of the introducing (subordinating) verb.
An extremely important set of clause-types usually included into the vast system of object clauses is formed by clauses pre-senting chunks of speech and mental-activity processes. These clauses are introduced by the verbs of speech and mental activ-ity (Lat. "verba sentiendi et declarandi"), whose contextual content they actually expose. Cf.:

Who says the yacht hasn't been properly prepared for the voyage? She wondered why on earth she was worrying so much, when obviously the time had come to end the incident and put it out of mind.
The two sentences render by their subordinate clauses speech of the non-author (non-agent) plane: in the first one ac-tual words of some third person are cited, in the second one a stream of thought is presented which is another form of the ex-istence of speech (i. e. inner speech). The chunk of talk ren-dered by this kind of presentation may not necessarily be actu-ally pronounced or mentally produced by a denoted person; it may only be suggested or imagined by the speaker; still, even in the latter case we are faced by lingually (grammatically) the same kind of non-author speech-featuring complex construc-tion. Cf.: Do you mean to say that the story has a moral?
Not all the clauses introduced by the verbs in question be-long to this type. In principle, these clauses are divided into the ones exposing the content of a mental action (as shown above) and the ones describing the content of a mental action, such as the following: You may tell me whatever you like. Will you tell me what the matter is?
The object clauses in the cited sentences, as different from the foregoing examples, describe the information allowed by the speaker-author (the first sentence) or wanted by the speaker-author (the second sentence), thereby not differing much from non-speech-rendering clauses. As for the speech-rendering object clauses, they are quite special, and it is by right that, as a rule, they are treated in grammar books under the separate heading of "rules of reported speech". Due to their se-mantic nature, they may be referred to as "reportive" clauses, and the same term will helpfully apply to the corresponding sentences as wholes. Indeed, it is in reportive sentences that the principal clause is more often than not reduced to an introduc-tory phrase akin to a parenthesis of additionally specifying se-mantics, so that the formally subordinate clause practically ab-sorbs all the essential information rendered by the sentence. Cf.:
Wainright said that Eastin would periodically report to him. > Periodically, Wainright said, Eastin would report to him (A. Hailey),

1. Subordinate clauses of secondary nominal positions in-clude attributive clauses of various syntactic functions. They fall into two major classes: "descriptive" attributive clauses and "restrictive" ("limiting") attributive clauses.
The descriptive attributive clause exposes some characteris-tic of the antecedent (i. e., its substantive referent) as such, while the restrictive attributive clause performs a purely identi-fying role, singling out the referent of the antecedent in the given situation. The basis of this classification, naturally, has nothing to do with the artistic properties of the classified units: a descriptive clause may or may not possess a special expres-sive force depending on the purpose and mastery of the respec-tive text production. Moreover, of the two attributive clause classes contrasted, the restrictive class is distinguished as the more concretely definable one, admitting of the oppositional interpretation as the "marked element": the descriptive class then will be oppositionally interpreted as the "non-restrictive" one, which precisely explains the correlative status of the two types of subordinate clauses.
It should be noted that, since the difference between de-scriptive and restrictive clauses lies in their functions, there is a possibility of one and the same clausal unit being used in both capacities, depending on the differences of the contexts. Cf.:
At last we found a place where we could make a fire. The place where we could make a fire was not a lucky one.
The subordinate clause in the first of the cited examples in-forms the listener of the quality of the place (> We found such a place) thereby being descriptive, while the same clause in the second example refers to the quality in question as a mere mark of identification (> The place was not a lucky one) and so is restrictive.
Descriptive clauses, in their turn, distinguish two major sub-types: first, "ordinary" descriptive clauses; second, "continu-ative" descriptive clauses.
The ordinary descriptive attributive clause expresses various situational qualifications of nounal antecedents. The qualifica-tions may present a constant situational feature or a temporary situational feature of different contextual relations and implica-tions. Cf.:
It gave me a strange sensation to see a lit up window in a big house that was not lived in. He wore a blue shirt the

collar of which was open at the throat. They were playing such a game as could only puzzle us.
The continuative attributive clause presents a situation on an equal domination basis with its principal clause, and so is at-tributive only in form, but not in meaning. It expresses a new predicative event (connected with the antecedent) which some-how continues the chain of situations reflected by the sentence as a whole. Cf.:
In turn, the girls came singly before Brett, who frowned, blinked, bit his pencil, and scratched his head with it, getting no help from the audience, who applauded each girl impartially and hooted at every swim suit, as if they could not see hundreds any day round the swimming pool (M. Dickens).
It has been noted in linguistic literature that such clauses are essentially not subordinate, but coordinate, and hence they make up with their principal clause not a complex, but a com-pound sentence. As a matter of fact, for the most part such clauses are equal to coordinate clauses of the copulative type, and their effective test is the replacement of the relative subor-dinator by the combination and + substitute. Cf.:
I phoned to Mr. Smith, who recognised me at once and in-vited me to his office. > I phoned to Mr. Smith, and he recog-nised me at once...
Still, the form of the subordinate clause is preserved by the continuative clause, the contrast between a dependent form and an independent content constituting the distinguishing feature of this syntactic unit as such. Thus, what we do see in continu-ative clauses is a case of syntactic transposition, i. e. the trans-ference of a subordinate clause into the functional sphere of a coordinate clause, with the aim of achieving an expressive ef-fect. This transpositional property is especially prominent in the which-continuative clause that refers not to a single nounal an-tecedent, but to the whole principal clause. E. g.:
The tower clock struck the hour, which changed the train of his thoughts. His pictures were an immediate success on the varnishing day, which was nothing to wonder.
The construction is conveniently used in descriptions and reasonings.
To attributive clauses belongs also a vast set of appositive

clauses which perform an important role in the formation of complex sentences. The appositive clause, in keeping with the general nature of apposition, does not simply give some sort of qualification to its antecedent, but defines or elucidates its very meaning in the context. Due to this specialisation, appositive clauses refer to substantive antecedents of abstract semantics. Since the role of appositive clauses consists in bringing about contextual limitations of the meaning of the antecedent, the status of appositive clauses in the general system of attributive clauses is intermediary between restrictive and descriptive.
In accord with the type of the governing antecedent, all the appositive clauses fall into three groups: first, appositive clauses of nounal relation; second, appositive clauses of pro-nominal relation; third, appositive clauses of anticipatory rela-tion.
Appositive clauses of nounal relation are functionally nearer to restrictive attributive clauses than the rest. They can intro-duce information of a widely variable categorial nature, both nominal and adverbial. The categorial features of the rendered information are defined by the type of the antecedent.
The characteristic antecedents of nominal apposition are ab-stract nouns like fact, idea, question, plan, suggestion, news, information, etc. Cf.:
The news that Dr. Blare had refused to join the Antarctic expedition was sensational. We are not prepared to discuss the question who will chair the next session of the Surgical Society.
The nominal appositive clauses can be tested by transform-ing them into the corresponding clauses of primary nominal po-sitions through the omission of the noun-antecedent or translat-ing it into a predicative complement. Cf.:
... > That Dr. Blare had refused to join the Antarctic expe-dition was sensational. That Dr. Blare had refused to join the Antarctic expedition was sensational news.
The characteristic antecedents of adverbial apposition are abstract names of adverbial relations, such as time, moment, place, condition, purpose, etc. Cf.:
We saw him at the moment he was opening the door of his Cadillac. They did it with the purpose that no one else might share the responsibility for the outcome of the venture.

As is seen from the examples, these appositive clauses serve a mixed or double function, i. e. a function constituting a mix-ture of nominal and adverbial properties. They may be tested by transforming them into the corresponding adverbial clauses through the omission of the noun-antecedent and, if necessary, the introduction of conjunctive adverbialisers. Cf.:
... > We saw him as he was opening the door of his Cadil-lac. ... > They did it so that no one else might share the respon-sibility for the outcome of the venture.
Appositive clauses of pronominal relation refer to an ante-cedent expressed by an indefinite or demonstrative pronoun. The constructions serve as informatively limiting and attention-focusing means in contrast to the parallel non-appositive con-structions. Cf.:
I couldn't agree with all that she was saying in her irrita-tion. > I couldn't agree with what she was saying in her irrita-tion. (Limitation is expressed.) That which did strike us was the inspector's utter ignorance of the details of the case. > What did strike us was the inspector's utter ignorance of the details of the case. (The utterances are practically equivalent, the one with a clausal apposition being somewhat more intense in its delimi-tation of the desired focus of attention.)
Appositive clauses of anticipatory relation are used in con-structions with the anticipatory pronoun (namely, the anticipa-tory it, occasionally the demonstratives this, that). There are two varieties of these constructions subjective and objective. The subjective clausal apposition is by far the basic one, both in terms of occurrence (it affects all the notional verbs of the vo-cabulary, not only transitive) and functional range (it possesses a universal sentence-transforming force). Thus, the objective anticipatory apposition is always interchangeable with the sub-jective anticipatory apposition, but not vice versa. Cf.:
I would consider it (this) a personal offence if they didn't accept the forwarded invitation. > It would be a personal of-fence (to me) if they didn't accept the forwarded invitation. You may depend on it that the letters won't be left unanswered. > It may be depended on that the letters won't be left unanswered.

The anticipatory appositive constructions, as is widely known, constitute one of the most peculiar typological features of English syntax. Viewed as part of the general appositive clausal system here presented, it is quite clear that the exposure of their appositive nature does not at all contradict their antici-patory interpretation, nor does it mar or diminish their "idio-matically English" property so emphatically pointed out in grammar books.
The unique role of the subjective anticipatory appositive construction, as has been stated elsewhere, consists in the fact that it is used as a universal means of rheme identification in the actual division of the sentence.
8. Clauses of adverbial positions constitute a vast domain of syntax which falls into many subdivisions each distinguish-ing its own field of specifications, complications, and difficul-ties of analysis. The structural peculiarities and idiosyncrasies characterising the numerous particular clause models making up the domain are treated at length in grammatical manuals of various practical purposes; here our concern will be to discuss some principal issues of their functional semantics and classifi-cation.
Speaking of the semantics of these clauses, it should be stressed that as far as the level of generalised clausal meanings is concerned, semantics in question is of absolute syntactic relevance; accordingly, the traditional identification of major adverbial clause models based on "semantic considerations" is linguistically rational, practically helpful, and the many at-tempts to refute it in the light of the "newly advanced, objective, consistently scientific" criteria have not resulted in creating a comprehensive system capable of competing with the tradi-tional one in its application to textual materials.
On the other hand, it would be a mistake to call in question the usefulness of the data obtained by the latest investigations. Indeed, if their original negative purpose has failed, the very positive contribution of the said research efforts to theoretical linguistics is not to be overlooked: it consists in having studied the actual properties of the complicated clausal system of the sentence, above all the many-sided correlation between struc-tural forms and functional meanings in the making of the sys-temic status of each clausal entity that admits of a description as a separate unit subtype.

Proceeding from the said insights, the whole system of ad-verbial clauses is to be divided into four groups.
The first group includes clauses of time and clauses of place. Their common semantic basis is to be defined as "local-isation" respectively, temporal and spatial. Both types of clauses are subject to two major subdivisions, one concerning the local identification, the other concerning the range of func-tions.
Local identification is essentially determined by subordina-tors. According to the choice of connector, clauses of time and place are divided into general and particularising. The general local identification is expressed by the non-marking conjunc-tions when and where. Taken by themselves, they do not intro-duce any further specifications in the time or place correlations between the two local clausal events (i.e. principal and subordi-nate). As for the particularising local identification, it specifies the time and place correlations of the two events localising the subordinate one before the principal, parallel with the principal, after the principal, and possibly expressing further subgrada-tions of these correspondences.
With subordinate clauses of time the particularising localisa-tion is expressed by such conjunctions as while, as, since, be-fore, after, until, as soon as, now that, no sooner than, etc. E.g.:
We lived here in London when the war ended. While the war was going on we lived in London. We had lived in London all through the war until it ended. After the war ended our fam-ily moved to Glasgow. Etc.
With clauses of place proper the particularising localisation is expressed but occasionally, mostly by the prepositional con-junctive combinations from where (bookish equivalent whence) and to where. E.g.:
The swimmers gathered where the beach formed a small promontory. The swimmers kept abreast of one another from where they started.
For the most part, however, spatial specifications in the complex sentence are rendered not by place-clauses proper, but by adverbial-appositive clauses. Cf.: We decided not to go back to the place from where we started on our journey.
From the functional point of view, clauses of localisation

should be divided into "direct" (all the above ones) and "trans-ferred", the latter mostly touching on matters of reasoning. E.g.:
When you speak of the plain facts there can't be any ques-tion of argument. But I can't agree with you where the princi-ples of logic are concerned.
A special variety of complex sentence with a time clause is presented by a construction in which the main predicative in-formation is expressed in the subordinate clause, the actual meaning of temporal localisation being rendered by the princi-pal clause of the sentence. E.g.:
Alice was resting in bed when Humphrey returned. He brought his small charge into the room and presented her to her "aunt" (D. E. Stevenson).
The context clearly shows that the genuine semantic accents in the first sentence of the cited passage is to be exposed by the reverse arrangement of subordination: it is Humphrey's actions that are relevant to the developing situation, not Alice's resting in bed: > Humphrey returned when Alice was resting in bed...
This type of complex sentence is known in linguistics as "inversive"; what is meant by the term, is semantics taken against the syntactic structure. The construction is a helpful sty-listic means of literary narration employed to mark a transition from one chain of related events to another one.
The second group of adverbial clauses includes clauses of manner and comparison. The common semantic basis of their functions can be defined as "qualification", since they give a qualification to the action or event rendered by the principal clause. The identification of these clauses can be achieved by applying the traditional question-transformation test of the how-type, with the corresponding variations of specifying character (for different kinds of qualification clauses). Cf.:
He spent the Saturday night as was his wont. > How did he spend the Saturday night? You talk to people as if they were a group. > How do you talk to people? I planned to give my mother a length of silk for a dress, as thick and heavy as it was possible to buy. > How thick and heavy the length of silk was intended to be?

All the adverbial qualification clauses are to be divided into "factual" and "speculative", depending on the real or unreal propositional event described by them.
The discrimination between manner and comparison clauses is based on the actual comparison which may or may not be ex-pressed by the considered clausal construction of adverbial qualification. The semantics of comparison is inherent in the subordinators as if, as though, than, which are specific introdu-cers of comparison clauses. On the other hand, the subordinator as, both single and in the combinations as ... as, not so ... as, is unspecific in this sense, and so invites for a discrimination test to be applied in dubious cases. It should be noted that more of-ten than not a clausally expressed manner in a complex sen-tence is rendered by an appositive construction introduced by phrases with the broad-meaning words way and manner. E.g.: Mr. Smith looked at me in a way that put me on the alert.
Herein lies one of the needed procedures of discrimination, which is to be formulated as the transformation of the tested clause into an appositive that- or which-clause: the possibility of the transformation marks the clause of manner, while the impossibility of the transformation (i.e. the preservation of the original as-clause) marks the clause of comparison. Cf.:
Mary received the guests as nicely as Aunt Emma had taught her > ... in a (very) nice way that Aunt Emma had taught her. (The test marks the clause as that of manner.) Mary received the guests as nicely as Aunt Emma would have done. > ... in as nice a way as Aunt Emma would have done. (The test marks the clause as comparative.)
Clauses of comparison are subdivided into those of equality (subordinators as, as ... as, as if, as though) and those of ine-quality (subordinators not so ... as, than). The discontinuous introducers mark, respectively, a more intense rendering of the comparison in question. Cf.:
That summer he took a longer holiday than he had done for many years. For many years he hadn't taken so long a holiday as he was offered that summer.
With clauses of comparison it is very important to distin-guish the contracted expression of predication, i.e. predicative zeroing, especially for cases where a clause of comparison as such is combined with a clause of time. Here 324

predicative zeroing may lead to the rise of peculiarly fused constructions which may be wrongly understood. By way of example, let us take the sentence cited in B. Ilyish's book: Do you find Bath as agreeable as when I had the honour of making the enquiry before? (J. Austen)
B. Ilyish analyses the construction as follows: "The when-clause as such is a temporal clause: it indicates the time when an action ("his earlier enquiry") took place. However, being introduced by the conjunction as, which has its correlative, an-other as, in the main clause, it is at the same time a clause of comparison" [Ilyish, 299].
But time and comparison are absolutely different character-istics, so that neither of them can by definition be functionally used for the other. They may go together only in cases when time itself forms the basis of comparison (I came later than Mr. Jerome did). As far as the analysed example is concerned, its clause of time renders no other clausal meaning than temporal; the clausal comparison proper is expressed reductionally, its sole explicit representative being the discontinuous introducer as ... as. Thus, the true semantics of the cited comparison is to be exposed by paradigmatic de-zeroing: > Do you find Bath as agreeable as it was when I had the honour of making the en-quiry before?
The applied principle of analysis of contamination time-comparison clauses for its part supports the zero-conception of other outwardly non-predicative comparative constructions, in particular those introduced by than. Cf.: Nobody could find the answer quicker than John. > Nobody could find the answer quicker than John did (could do).
The third and most numerous group of adverbial clauses in-cludes "classical" clauses of different circumstantial semantics, i.e. semantics connected with the meaning of the principal clause by various circumstantial associations; here belong clauses of attendant event, condition, cause, reason, result (consequence), concession, purpose. Thus, the common seman-tic basis of all these clauses can be defined as "circumstance". The whole group should be divided into two subgroups, the first being composed by clauses of "attendant circumstance"; the second, by clauses of "immediate circumstance".
Clauses of attendant circumstance are not much varied in structure or semantics and come near to clauses of time. The difference lies in the fact that, unlike clauses of time, the event described by a clause of attendant circumstance

is presented as some sort of background in relation to the event described by the principal clause. Clauses of attendant circum-stance are introduced by the conjunctions while and as. E.g.: As (while) the reception was going on, Mr. Smiles was engaged in a lively conversation with the pretty niece of the hostess.
The construction of attendant circumstance may be taken to render contrast; so all the clauses of attendant circumstance can be classed into "contrastive" (clauses of contrast) and "non-contrastive". The non-contrastive clause of circumstance has been exemplified above. Here is an example of contrastive at-tendant circumstance expressed clausally:
Indeed, there is but this difference between us that he wears fine clothes while I go in rags, and that while I am weak from hunger he suffers not a little from overfeeding (O. Wilde).
As is clear from the example, a complex sentence with a contrastive clause of attendant circumstance is semantically close to a compound sentence, i.e. a composite sentence based on coordination.
Clauses of immediate circumstance present a vast and com-plicated system of constructions expressing different explana-tions of events, reasonings and speculations in connection with them. The system should relevantly be divided into "factual" clauses of circumstance and "speculative" clauses of circum-stance depending on the real or unreal predicative denotations expressed. This division is of especial significance for complex sentences with conditional clauses (real condition, problematic condition, unreal condition). Other types of circumstantial clauses express opposition between factual and speculative se-mantics with a potential relation to some kind of condition in-herent in the deep associations of the syntactic constructions. E.g.:
Though she disapproved of their endless discussions, she had to put up with them. (Real concession) > Though she may disapprove of their discussions, she will have to put up with them. (Speculative concession) If she disapproved (had disapproved) of their discussions, why would she put up (have put up) with them? (Speculative condition)
The argument was so unexpected that for a moment Jack lost his ability to speak. (Real consequence) > The argument was so unexpected that it would have frustrated Jack's

ability to speak if he had understood the deep meaning of it. (Speculative consequence, based on the speculative condition)
Each type of clauses of circumstance presents its own prob-lems of analysis. On the other hand, it must be pointed out that all the types of these clauses are inter-related both semantically and paradigmatically, which may easily be shown by the corre-sponding transformations and correlations. Some of such corre-lations have been shown on the examples above. Compare also:
He opened the window wide that he might hear the conver-sation below. (Purpose) > Unless he wanted to hear the con-versation below he wouldn't open the window. (Condition) > As he wanted to hear the conversation below, he opened the window wide and listened. (Cause) > Though he couldn't hear properly the conversation below, he opened the window and listened. (Concession) > The voices were so low that he couldn't hear the conversation through the open window. (Con-sequence) > If he hadn't opened the window wide he couldn't have heard the conversation. (Condition)
Certain clausal types of circumstance are closely related to non-circumstantial clausal types. In particular, this kind of con-nection is observed between conditional clauses and time clauses and finds its specifically English expression in the rise of the contaminated if- and when-clauses: If and when the dis-cussion of the issue is renewed, both parties will greatly benefit by it.
Another important variety of clauses of mixed syntactic se-mantics is formed by concessive clauses introduced by the con-nectors ending in -ever. E.g.:
Whoever calls, I'm not at home. However tempting the offer might be, Jim is not in a position to accept it.
Clauses of mixed adverbial semantics present an interesting field of paradigmatic study.
The fourth group of adverbial clauses is formed by paren-thetical or insertive constructions. Parenthetical clauses, as has been stated elsewhere, are joined to the principal clause on a looser basis than the other adverbial clauses; still, they do form with the principal clause a syntactic sentential unity, which is easily proved by the procedure of diagnostic elimination. Cf.:

Jack has called here twice this morning, if I am not mis-taken. > (*) Jack has called here twice this morning.
As is seen from the example, the elimination of the paren-thesis changes the meaning of the whole sentence from prob-lematic to assertive: the original sense of the utterance is lost, and this shows that the parenthesis, though inserted in the con-struction by a loose connection, still forms an integral part of it.
As to the subordinative quality of the connection, it is ex-pressed by the type of the connector used. In other words, par-enthetical predicative insertions can be either subordinative or coordinative, which is determined by the contextual content of the utterance and exposed by the connective introducer of the clause. Cf. a coordinate parenthetical clause: Jim said, and I quite agree with him, that it would be in vain to appeal to the common sense of the organisers.
Cf. the subordinate correlative of the cited clause: Jim said, though I don't quite agree with him, that it would be in vain to appeal to the common sense of the organisers.
Parenthetical clauses distinguish two semantic subtypes. Clauses of the first subtype, illustrated by the first example in this paragraph, are "introductory", they express different modal meanings. Clauses of the second subtype, illustrated by the lat-ter example, are "deviational", they express commenting inser-tions of various semantic character. Deviational parenthesis marks the loosest possible syntactic connection of clauses com-bined into a composite sentence.
9. Clauses in a complex sentence may be connected with one another more closely and less closely, similar to the parts of a simple sentence. The intensity of connection between the clauses directly reflects the degree of their proposemic self-dependence and is therefore an essential characteristic of the complex sentence as a whole. For instance, a predicative clause or a direct object clause are connected with the principal clause so closely that the latter cannot exist without them as a com-plete syntactic unit. Thus, this kind of clausal connection is obligatory. Cf.:
The matter is, we haven't received all the necessary instruc-tions yet. > (*) The matter is I don't know what Mike is go-ing to do about his damaged bike. > (*)I don't know

As different from this, an ordinary adverbial clause is con-nected with the principal clause on a looser basis, it can be de-leted without destroying the principal clause as an autonomous unit of information. This kind of clausal connection is optional. Cf.:
The girl gazed at him as though she was struck by some-thing extraordinary in his appearance. > The girl gazed at him.
The division of subordinative clausal connections into obligatory and optional was employed by the Russian linguist N. S. Pospelov (1950) for the introduction of a new classifica-tion of complex sentences. In accord with his views, all the complex sentences of minimal structure (i.e. consisting of one principal clause and one subordinate clause) should be classed as "one-member" complex sentences and "two-member" com-plex sentences. One-member complex sentences are distin-guished by an obligatory subordinative connection, while two-member complex sentences are distinguished by an optional subordinative connection. The obligatory connection is deter-mined both by the type of the subordinate clause (subject, predicative, object clauses) and the type of the introduction of the clause (demonstrative correlation). The optional connection characterises adverbial clauses of diverse functions and attribu-tive clauses of descriptive type. Semantically, one-member complex sentences are understood as reflecting one complex logical proposition, and two-member complex sentences as re-flecting two logical propositions connected with each other on the subordinative principle.
The rational character of the advanced conception is quite obvious. Its strong point is the fact that it consistently demon-strates the correlation between form and meaning in the com-plex sentence structure. Far from rejecting the traditional teach-ing of complex sentences, the "member conception" is based on its categories and develops them further, disclosing such prop-erties of subordinative connections which were not known to the linguistic science before.
Speaking not only of the complex sentence of minimal composition, but in terms of complex sentences in general, it would be appropriate to introduce the notions of "monolythic" and "segregative" sentence structures. Obligatory subordinative connections underlie monolythic complexes, while optional subordinative connections underlie segregative complexes.

Monolithic complex sentences fall into four basic types.
The first of them is formed by merger complex sentences, i.e. sentences with subject and predicative subordinate clauses. The subordinate clausal part of the merger monolythic com-plex, as has been shown above (see 2), is fused with its prin-cipal clause. The corresponding construction of syntactic an-ticipation should also be considered under this heading. Cf.: It was at this point that Bill had come bustling into the room. > (*) It was at this point
The second subtype of complex sentences in question is formed by constructions whose subordinate clauses are depend-ent on the obligatory right-hand valency of the verb in the prin-cipal clause. We can tentatively call these constructions "va-lency" monolith complexes. Here belong complexes with object clauses and valency-determined adverbial clauses: from the point of view of subordinative cohesion they are alike. Cf.:
I don't know when I'm beaten. (*) I don't know Put the book where you've taken it from. > (*) Put the book Her first shock was when she came down. > (*) Her first shock was
The third subtype of monolythic complex sentences is formed by constructions based on subordinative correlations "correlation" monolith complexes. E.g.:
His nose was as unkindly short as his upper lip was long. You will enjoy such a sight as you are not likely to see again. The more I think of it, the more I'm convinced of his innocence.
Restrictive attributive clauses should be included into this subtype of correlation monoliths irrespective of whether or not their correlation scheme is explicitly expressed. Cf.:
This is the same report as was submitted last week. This is the report that was submitted last week.
Finally, the fourth subtype of monolithic complex sentences is formed by constructions whose obligatory connection be-tween the principal and subordinate clauses is determined only by the linear order of clausal positions. Cf.: If he comes, tell him to wait. >(*) If he comes
As is easily seen, such "arrangement" monolith complexes are not "organically" monolithic, as different from the first three monolith subtypes; positional re-arrangement deprives

them of this quality, changing the clausal connection from obligatory into optional: Tell him to wait if he comes. > Tell him to wait.
The rest of the complex sentences are characterised by seg-regative structure, the maximum degree of syntactic option be-ing characteristic of subordinative parenthetical connection.
10. Complex sentences which have two or more subordi-nate clauses discriminate two basic types of subordination ar-rangement: parallel and consecutive.
Subordinate clauses immediately referring to one and the same principal clause are said to be subordinated "in parallel" or "co-subordinated". Parallel subordination may be both ho-mogeneous and heterogeneous. For instance, the two clauses of time in the following complex sentence, being embedded on the principle of parallel subordination, are homogeneous they depend on the same element (the principal clause as a whole), are connected with each other coordinatively and perform the same function: When he agrees to hear me, and when we have spoken the matter over, I'll tell you the result.
Homogeneous arrangement is very typical of object clauses expressing reported speech. E.g.: Mrs. Lewin had warned her that Cadover was an extraordinary place, and that one must never be astonished by anything (A. Huxley).
By heterogeneous parallel subordination, co-subordinate clauses mostly refer to different elements in the principal clause. E.g.: The speakers who represented different nations and social strata were unanimous in their call for peace which is so ardently desired by the common people of the world.
As different from parallel subordination, consecutive subor-dination presents a hierarchy of clausal levels. In this hierarchy one subordinate clause is commonly subordinated to another, making up an uninterrupted gradation. This kind of clausal ar-rangement may be called "direct" consecutive subordination. E.g.: I've no idea why she said she couldn't call on us at the time I had suggested.
Alongside of direct consecutive subordination there is an-other form of clausal hierarchy which is formed without an immediate domination of one subordinate clause over another. For instance, this is the case when the principal clause of a complex multi-level sentence is built up on a merger basis, i.e. includes a subject or predicative clause.

E.g.: What he saw made him wince as though he had been struck.
In the cited sentence the comparative subordinate clause is dominated by the whole of the principal clause which includes a subordinate propositional unit in its syntactic position of the subject. Thus, the subordinative structure of the sentence is in fact consecutive, though not directly consecutive. This type of hierarchical clausal arrangement may be called "oblique" con-secutive subordination; it is of minor importance for the system of subordination perspective as a whole.
The number of consecutive levels of subordination gives the evaluation of the "depth" of subordination perspective one of the essential syntactic characteristics of the complex sentence. In the first three examples cited in the current paragraph this depth is estimated as 1; in the fourth example (direct consecu-tive subordination) it equals 3; in the fifth example (oblique consecutive subordination) it equals 2. The subordination per-spective of complex sentences used in ordinary colloquial speech seldom exceeds three consecutive clausal levels.
1. The compound sentence is a composite sentence built on the principle of coordination. Coordination, the same as subordination, can be expressed either syndetically (by means of coordinative connectors) or asyndetically.
The main semantic relations between the clauses connected coordinatively are copulative, adversative, disjunctive, causal, consequential, resultative. Similar semantic types of relations are to be found between independent, separate sentences form-ing a continual text. As is known, this fact has given cause to some scholars to deny the existence of the compound sentence as a special, regular form of the composite sentence.*
The advanced thesis to this effect states that the so-called "compound sentence" is a fictitious notion developed under
* See: . . . ., 1968.

the school influence of written presentation of speech; what is fallaciously termed the "compound sentence" constitutes in re-ality a sequence of semantically related independent sentences not separated by full stops in writing because of an arbitrary school convention.
To support this analysis, the following reasons are put for-ward: first, the possibility of a falling, finalising tone between the coordinated predicative units; second, the existence, in writ-ten speech, of independently presented sentences introduced by the same conjunctions as the would-be "coordinate clauses"; third, the possibility of a full stop-separation of the said "coor-dinate clauses" with the preservation of the same semantic rela-tions between them.
We must admit that, linguistically, the cited reasons are not devoid of a rational aspect, and, which is very important, they appeal to the actual properties of the sentence in the text. How-ever, the conception taken as a whole gives a false presentation of the essential facts under analysis and is fallacious in princi-ple.
As a matter of fact, there is a substantial semantico-syntactic difference between the compound sentence and the correspond-ing textual sequence of independent sentences. This difference can escape the attention of the observer when tackling isolated sentences, but it is explicitly exposed in the contexts of contin-ual speech. Namely, by means of differences in syntactic distri-butions of predicative units, different distributions of the ex-pressed ideas is achieved, which is just the coordinative syntac-tic functions in action; by means of combining or non-combining predicative units into a coordinative polypredicative sequence the corresponding closeness or looseness of connec-tions between the reflected events is shown, which is another aspect of coordinative syntactic functions. It is due to these functions that the compound sentence does not only exist in the syntactic system of language, but occupies in it one of the con-stitutive places.
By way of example, let us take a textual sequence of inde-pendent monopredicative units:
Jane adored that actor. Hockins could not stand the sight of him. Each was convinced of the infallibility of one's artistic judgment. That aroused prolonged arguments.
Given the "negative" theory of the compound sentence is correct, any coordinative-sentential re-arrangements of the cited sentences must be indifferent as regards the sense

rendered by the text. In practice, though, it is not so. In particu-lar, the following arrangement of the predicative units into two successive compound sentences is quite justified from the se-mantico-syntactic point of view:
> Jane adored that actor, but Hockins could not stand the sight of him. Each was convinced of the infallibility of one's judgment, and that aroused prolonged arguments.
As different from this, the version of arranging the same ma-terial given below cannot be justified in any syntactic or seman-tic sense:
> *Jane adored that actor. But Hockins could not stand the sight of him, each was convinced of the infallibility of one's judgment. And that aroused prolonged arguments.
On the other hand, some subordinate clauses of a complex sentence can also be separated in the text, thus being changed into specific independent sentences. Still, no one would seek to deny the existence of complex sentence patterns based on op-tional subordinative connections. Cf.:
Suddenly Laura paused as if she was arrested by something invisible from here. > Suddenly Laura paused. As if she was arrested by something invisible from here.
As for the factor of intonation, it should indeed be invariably taken into account when considering general problems of sen-tence identification. The propositional intonation contour with its final delimitation pause is one of the constitutive means of the creation and existence of the sentence as a lingual phe-nomenon. In particular, the developing intonation pattern in the process of speech sustains the semantic sentence strain from the beginning of the sentence up to the end of it. And there is a pro-found difference between the intonation patterns of the sentence and those of the clause, no matter how many traits of similarity they may possess, including finalising features. Moreover, as is known, the tone of a coordinate clause, far from being rigor-ously falling, can be rising as well. The core of the matter is that the speaker has intonation at his disposal as a means of forming sentences, combining sentences, and separating sentences. He actively uses this means, grouping the same syntactic strings of words now as one composite sentence, now as so many simple sentences, with the corresponding more

essential or less essential changes in meanings, of his own choice, which is determined by concrete semantic and contex-tual conditions.
Thus, the idea of the non-existence of the compound sen-tence in English should be rejected unconditionally. On the other hand, it should be made clear that the formulation of this negative idea as such has served us a positive cause, after all: its objective scientific merit, similar to some other inadequate ideas advanced in linguistics at different times, consists in the very fact that it can be used as a means of counter-argumentation in the course of research work, as a starting point for new insights into the deep nature of lingual phenomena in the process of theoretical analysis sustained by observation.
2. The compound sentence is derived from two or more base sentences which, as we have already stated above, are connected on the principle of coordination either syndetically or asyndetically. The base sentences joined into one compound sentence lose their independent status and become coordinate clauses parts of a composite unity. The first clause is "lead-ing" (the "leader" clause), the successive clauses are "sequen-tial". This division is essential not only from the point of view of outer structure (clause-order), but also in the light of the se-mantico-syntactic content: it is the sequential clause that in-cludes the connector in its composition, thus being turned into some kind of dependent clause, although the type of its depend-ence is not subordinative. Indeed, what does such a predicative unit signify without its syntactic leader?
The coordinating connectors, or coordinators, are divided into conjunctions proper and semi-functional clausal connectors of adverbial character. The main coordinating conjunctions, both simple and discontinuous, are: and, but, or, nor, neither, for, either ... or, neither ... nor, etc. The main adverbial coordi-nators are: then, yet, so, thus, consequently, nevertheless, how-ever, etc. The adverbial coordinators, unlike pure conjunctions, as a rule can shift their position in the sentence (the exceptions are the connectors yet and so). Cf.:
Mrs. Dyre stepped into the room, however the host took no notice of it. > Mrs. Dyre stepped into the room, the host, how-ever, took no notice of it.

The intensity of cohesion between the coordinate clauses can become loose, and in this case the construction is changed into a cumulative one (see Ch. XXVI). E.g.: Nobody ever dis-turbed him while he was at work; it was one of the unwritten laws.
As has been stated elsewhere, such cases of cumulation mark the intermediary status of the construction, i.e. its place in syntax between a composite sentence and a sequence of inde-pendent sentences.
3. When approached from the semantico-syntactic point of view, the connection between the clauses in a compound sentence should be analysed into two basic types: first, the un-marked coordinative connection; second, the marked coordina-tive connection.
The unmarked coordinative connection is realised by the coordinative conjunction and and also asyndetically. The un-marked semantic nature of this type of connection is seen from the fact that it is not specified in any way and requires a diag-nostic exposition through the marked connection. The exposi-tion properly effected shows that each of the two series of com-pound predicative constructions falls into two principal subdivi-sions. Namely, the syndetic and-constructions discriminate, first, simple copulative relations and, second, broader, non-copulative relations. The asyndetic constructions discriminate, first, simple enumerative relations and, second, broader, non-enumerative relations. Cf. examples of the primary connective meanings of the constructions in question:
You will have a great deal to say to her, and she will have a great deal to thank you for. She was tall and slender, her hair was light chestnut, her eyes had a dreamy expression.
The broader connective meanings of the considered con-structions can be exposed by equivalent substitutions:
The money kept coming in every week, and the offensive gossip about his wife began to be replaced by predictions of sensational success. > The money kept coming in every week, so the offensive gossip about his wife began to be replaced by predictions of sensational success. The boy obeyed, the request was imperative. > The boy obeyed, for the request was impera-tive.

The marked coordinative connection is effected by the pure and adverbial coordinators mentioned above. Each semantic type of connection is inherent in the marking semantics of the connector. In particular, the connectors but, yet, stilt, however, etc. express different varieties of adversative relations of clauses; the discontinuous connectors both ... and, neither ... nor express, correspondingly, positive and negative (exclusive) copulative relations of events; the connectors so, therefore, con-sequently express various subtypes of clausal consequence, etc.
In order to give a specification to the semantics of clausal relations, the coordinative conjunction can be used together with an accompanying functional particle-like or adverb-like word. As a result, the marked connection, as it were, becomes doubly marked. In particular, the conjunction but forms the conjunctive specifying combinations but merely, but instead, but also and the like; the conjunction or forms the characteristic coordinative combinations or else, or rather, or even, etc. Cf.:
The workers were not prepared to accept the conditions of the administration, but instead they were considering a mass demonstration. She was frank with him, or rather she told him everything concerning the mere facts of the incident.
The coordinative specifiers combine also with the conjunc-tion and, thus turning the unmarked coordinative connection into a marked one. Among the specifiers here used are included the adverbial coordinators so, yet, consequently and some oth-ers. E.g.: The two friends didn't dispute over the issue after-wards, and yet there seemed a hidden discord growing between them.
It should be specially noted that in the described semantic classification of the types of coordinative relations, the asyn-detic connection is not included in the upper division of the sys-tem, which is due to its non-specific functional meaning. This fact serves to sustain the thesis that asyndetic connection of clauses is not to be given such a special status in syntax as would raise it above the discrimination between coordination and subordination.
4. It is easily seen that coordinative connections are corre-lated semantically with subordinative connections so that a compound sentence can often be transformed into
22149!) 337

a complex one with the preservation of the essential relational semantics between the clauses. The coordinative connections, as different from subordinative, besides the basic opposition to the latter by their ranking quality, are more general, they are semantically less discriminatory, less "refined". That is why the subordinative connection is regularly used as a diagnostic model for the coordinative connection, while the reverse is an exception rather than a rule. Cf.:
Our host had rung the bell on our entrance and now a Chi-nese cook came in with more glasses and several bottles of soda. > On our entrance, as our host had rung the bell, a Chi-nese cook came in with more glasses and several bottles of soda. There was nothing else to do, so Alice soon began talking again. > Alice soon began talking again because there was nothing else to do.
Speaking of the diagnostic role of subordinative construc-tions in relation to coordinative ones, it should be understood that this is of especial importance for the unmarked construc-tions, in particular for those realised by the conjunction and.
On the other hand, the coordinative connection of clauses is in principle not reducible to the subordinative connection, which fact, as in other similar cases of correlations, explains the separate and parallel existence of both types of clausal connec-tion in language. This can be illustrated by the following exam-ple: I invited Mike to join us, but he refused.
It would appear at first sight that the subordinative diagnos-tic-specifying exposition of the semantic relations between the clauses of the cited sentence can be achieved by the concessive construction: "Though I invited Mike to join us, he refused". But the proper observation of the corresponding materials shows that this diagnosis is only valid for part of the possible contexts. Suffice it to give the following two contextual expan-sions to the sentence in question, of which only one corresponds to the cited diagnosis.
The first expansion: You are mistaken if you think that Mike was eager to receive an invitation to join us. I invited him, but he refused.
The given concessive reading of the sentence is justified by the context: the tested compound sentence is to be replaced here by the above complex one on a clear basis of equivalence.
The second expansion: It was decided to invite either Mike or Jesse to help us with our work. First I invited Mike, but he refused. Then we asked Jesse to join us.
338 '

It is quite clear that the devised concessive diagnosis is not at all justified by this context: what the analysed construction does render here, is a stage in a succession of events, for which the use of a concessive model would be absurd.
5. The length of the compound sentence in terms of the number of its clausal parts (its predicative volume), the same as with the complex sentence, is in principle unlimited; it is de-termined by the informative purpose of the speaker. The com-monest type of the compound sentence in this respect is a two-clause construction.
On the other hand, predicatively longer sentences than two-clause ones, from the point of view of semantic correlation be-tween the clauses, are divided into "open" and "closed" con-structions. Copulative and enumerative types of connection, if they are not varied in the final sequential clause, form "open" coordinations. These are used as descriptive and narrative means in a literary text. Cf.:
They visited house after house. They went over them thor-oughly, examining them from the cellars in the basement to the attics under the roof. Sometimes they were too large and some-times they were too small; sometimes they were too far from the center of things and sometimes they were too close; some-times they were too expensive and sometimes they wanted too many repairs; sometimes they were too stuffy and sometimes they were too airy; sometimes they were too dark and some-times they were too bleak. Roger always found a fault that made the house unsuitable (S. Maugham).
In the multi-clause compound sentence of a closed type the final part is joined on an unequal basis with the previous ones (or one), whereby a finalisation of the expressed chain of ideas is achieved. The same as open compound sentences, closed compound constructions are very important from the point of view of a general text arrangement. The most typical closures in such compound sentences are those effected by the conjunc-tions and (for an asyndetic preceding construction) and but (both for an asyndetic and copulative syndetic preceding con-struction). Cf., respectively:
His fingernails had been cleaned, his teeth brushed, his hair combed, his nostrils cleared and dried, and he had been dressed in formal black by somebody or other (W. Saroyan).



Pleasure may turn a heart to stone, riches may make it callous, but sor-row oh, sorrow cannot break it (O. Wilde).
The structure of the closed coordinative construction is most conven-ient for the formation of expressive climax.
1. In accord with the principles laid down in the introductory de-scription of composite sentences (Ch. XXVI), the semi-composite sen-tence is to be defined as a sentence with more than one predicative lines which are expressed in fusion. For the most part, one of these lines can be identified as the leading or dominant, the others making the semi-predicative expansion of the sentence. The expanding semi-predicative line in the minimal semi-composite sentence is either wholly fused with the dominant (complete) predicative line of the construction, or partially fused with it, being weakened as a result of the fusing derivational trans-formation.
The semi-composite sentence displays an intermediary syntactic character between the composite sentence and the simple sentence. Its immediate syntagmatic structure ("surface" structure) is analogous to that of an expanded simple sentence, since it possesses only one completely expressed predicative unit. Its derivational structure ("deep" structure), on the other hand, is analogous to that of a composite sentence, because it is derived from two or more completely predicative units its base sentences.
There are two different causes of the existence of the semi-composite sentence in language, each of them being essentially impor-tant in itself.
The first cause is the tendency of speech to be economical. As a re-sult of this tendency, reductional processes are developed which bring about semi-blending of sentences. The second cause is that, apart from being economical, the semi-composite sentence fulfils its own purely semantic function, different from the function of the composite sentence proper (and so supplementing it). Namely, it is used to show that the events described in the corresponding sentence parts are more closely connected than the events described in the

parts of the composite sentence of complete composition. This function is inherent in the structure it reflects the speaker's view of reality, his presentation of it. Thus, for different reasons and purposes the same two or several events can be reflected now by one type of structure, now by another type of structure, the corresponding "pleni"- and semi-constructions existing in the syntactic system of language as pairs of re-lated and, for that matter, synonymically related functions. E.g.:
The sergeant gave a quick salute to me, and then he put his squad in motion. > Giving a quick salute to me, the sergeant put his squad in motion. > With a quick salute to me, the sergeant put his squad in mo-tion.
The two connected events described by the cited sentences are, first, the sergeant's giving a salute to the speaker, and, second, the sergeant's putting his squad in motion. The first sentence, of the pleni-composite type, presents these situationally connected events in separate processual descriptions as they happened one after the other, the successive order being accentuated by the structural features of the construction, in par-ticular, its sequential coordinate clause. The second sentence, of the semi-composite participial-expanded type, expresses a semantic ranking of the events in the situational blend, one of them standing out as a dominant event, the other as a by-event. In the presentation of the third construction, belonging to the primitivised type of semi-composition (maximum degree of blending), the fusion of the events is shown as con-stituting a unity in which the attendant action (the sergeant's salute) forms simply a background detail in relation to the immediately reflected occurrence (the sergeant's putting the squad in motion).
According to the ranking structure of the semi-composite sentences, they should be divided into semi-complex and semi-compound ones. These constructions correspond to the complex and compound sentences of complete composition (i.e., respectively, pleni-complex and pleni-compound sentences).
2. The semi-complex sentence is a semi-composite sentence built up on the principle of subordination. It is derived from minimum two base sentences, one matrix and one insert. In the process of semi-complexing, the insert sentence is transformed into a partially depredi-cated construction which is embedded in one of the syntactic positions of the

matrix sentence. In the resulting construction, the matrix sentence be-comes its dominant part and the insert sentence, its subordinate semi-clause.
The semi-complex sentences fall into a number of subtypes. Their basic division is dependent on the character of predicative fusion: this may be effected either by the process of position-sharing (word-sharing), or by the process of direct linear expansion. The sentences based on position-sharing fall into those of subject-sharing and those of object-sharing. The sentences based on semi-predicative linear expan-sion fall into those of attributive complication, adverbial complication, and nominal-phrase complication. Each subtype is related to a definite complex sentence (pleni-complex sentence) as its explicit structural pro-totype.
3. Semi-complex sentences of subject-sharing are built up by means of the two base sentences overlapping round the common subject. E.g.:
The man stood. + The man was silent. > The man stood silent. The moon rose. + The moon was red. > The moon rose red.
From the syntagmatic point of view, the predicate of these sentences forms the structure of the "double predicate" because it expresses two essential functions at once: first, the function of a verbal type (the verb component of the predicate); second, the function of a nominal type (the whole combination of the verb with the nominal component). The para-digmatic analysis shows that the verb of the double predicate, being on the surface a notional link-verb, is in fact a quasi-link.
In the position of the predicative of the construction different cate-gorial classes of words are used with their respective specific meanings and implications: nouns, adjectives, participles both present and past. Cf.:
Sam returned from the polar expedition a grown-up man. They waited breathless. She stood bending over the child's bed. We stared at the picture bewildered.
Observing the semantic, content of the given constructions, we sec that, within the bounds of their functional differences, they express two simultaneous events or, rather, the simultaneity of the event described by the complicalor expansion with that described by the dominant part. At the

same time the construction gives informative prominence not to its dominant, but to the complicator, and corresponds to the pleni-complex sentence featuring the complicator event in the principal clause placed in post-position. Cf.:
The moon rose red. > As the moon rose it was red. She stood bend-ing over the child's bed. > As she stood she was bending over the child's bed.
In the subject-sharing semi-composites with reflexivised dominant verbs of intense action the idea of change is rendered. E.g.:
He spoke himself hoarse. > As he spoke he became hoarse. (Fur-ther diagnosis: He spoke and spoke until he became hoarse.)
Apart from the described types of subject-sharing sentences there is a variety of them featuring the dominant verb in the passive. E.g.:
The idea has never been considered a wise one. The company was ordered to halt.
These sentences have active counterparts as their paradigmatic deri-vation bases which we analyse below as semi-complex sentences of ob-ject sharing.
4. Semi-complex sentences of object-sharing, as different from those of subject-sharing, are built up of two base sentences overlapping round the word performing different functions in them: in the matrix sentence it is the object, in the insert sentence it is the subject. The com-plicator expansion of such sentences is commonly called the "complex object". E.g.:
We saw him.-\-He approached us. > We saw him approach us (ap-proaching us). They painted the fence.-\-The fence was (became) green. > They painted the fence green.
Some dominant verbs of such constructions are not used in the same essential meaning outside the constructions, in particular, some causative verbs, verbs of liking and disliking, etc. Cf.: *I made him.+He obeyed. ~ I made him obey.
This fact, naturally, reflects a very close unity of the constituents of such constructions, but, in our opinion, it can't be looked upon as exclud-ing the constructions from

the syntactic subsystem in question; rather, the subsystem should be di-vided into the subsets of "free" object-sharing and "bound" object-sharing.
The adjunct to the shared object is expressed by an infinitive, a pre-sent or past participle, an adjective, a noun, depending on the structural type of the insert sentence (namely, on its being verbal or nominal).
As is seen from the above, the paradigmatic (derivational) explana-tion of the sentence with a "complex object" saves much descriptive space and, which is far more important, is at once generalising and prac-ticable.* As for the relations between the two connected events expressed by the object-sharing sentence, they are of the three basic types: first, re-lations of simultaneity in the same place; second, relations of cause and result; third, relations of mental attitude towards the event (events thought of, spoken of, wished for, liked or disliked, etc.). All these types of relations can be explicated by the corresponding transformations of the semi-complex sentences into pleni-complex sentences.
Simultaneity in the same place is expressed by constructions with dominant verbs of perceptions (see, hear, feel, smell, etc.). E.g.:
He felt the morning breeze gently touching his face. > He felt the morning breeze as it was gently touching his lace. I never heard the word pronounced like that. > I never heard the word as it was pronounced like that.
Cause and result relations are rendered by constructions with domi-nant causative verbs taking three types of complex objects: an unmarked infinitival complex object (the verbs make, let, get, have, help); a nounal or adjectival complex object (the verbs call, appoint, keep, paint, etc.); a participial complex object (the verbs set, send, keep, etc.). Cf.:
I helped Jo find the photo. > I helped Jo so that he found the photo. The cook beat the meat soft. The cook beat the meat so that it was (became) soft.
Different mental presentations of the complicator event are effected, respectively, by verbs of mental perceptions and thinking (think, believe, expect, find, etc.); verbs of speech
* Cf. the classical "syntagmatic" explanation of constructions with com-plex objects in the cited 13. A. llyish's book, p. 257 ff.

(tell, ask, report, announce, etc.); verbs of wish; verbs of liking and dis-liking. Cf.:
You will find many things strange here. > You will find that many things are strange here. I didn't mean my words to hurt you. > I didn't mean that my words should hurt you.
Semi-complex sentences of the object-sharing type, as we have stated above, are closely related to sentences of the subject-sharing type. Structurally this is expressed in the fact that they can be transformed into the passive, their passive counterparts forming the corresponding sub-ject-sharing constructions. Cf.:
We watched the plane disappear behind the distant clouds. > The plane was watched to disappear behind the distant clouds. They washed the floor clean. > The floor was washed clean.
Between the two series of constructions, i.e. active and passive, equivalence of the event-relations is observed, so that the difference in their basic meaning is inherent in the difference between the verbal ac-tive and passive as such.
5. Semi-complex sentences of attributive complication are derived from two base sentences having an identical element that occupies the position of the subject in the insert sentence and any notional position in the matrix sentence. The insert sentence is usually an expanded one. By the semi-complexing process, the insert sentence drops out its subject-identical constituent and is transformed into a semi-predicative post-positional attribute to the antecedent element in the matrix sentence. E.g.:
The waves sent out fine spray. + The waves rolled over the dam. > The waves rolling over the dam sent out fine spray. I came in late for the supper. + The supper was served in the dining-room. > I came in late for the supper served in the dining-room.
The analogy between post-positional attributes (especially of a de-tached type) and attributive subordinate clauses has always been pointed out in grammar-books of various destination. The common pre-positional attribute is devoid of a similar half-predicative character and is not to be considered as forming a semi-composite construction with the

dominant predicative unit. Cf.: The bored family switched off the TV. The family, bored, switched off the TV.
As for the possible detachment of the defining element (construc-tion) in pre-position, this use is rather to be analysed as adverbial, not attributive, the circumstantial semantic component prevailing over the attributive one in this case. Cf.: Bored, the family switched off the TV. > As the family was bored, it switched off the TV.
, Naturally, the existence of some intermediary types cannot be ex-cluded, which should be exposed in due course by the corresponding contextual observation.
As is seen, the base syntactic material for producing attributively complicated semi-composites is similar to the derivation base of posi-tion-sharing semi-composites. The essential difference between the con-structions, though, lies in the character of joining their clausal parts: while the process of overlapping deprives the position-sharing expansion of any self-dependent existence, however potential it might be, the proc-ess of linear expansion with the attributive complication preserves the autonomous functional role of the semi-clause. The formal test of it is the possibility of inserting into the construction a relative conjunctive plus the necessary verbal element, changing the attributive semi-clause into the related attributive pleni-clause. E.g.:' This is a novel translated from the French. > This is a novel which has been translated from the French,
This test resembles a reconstruction, since an attributive complica-tion in many respects resembles a reduced clause. The position-sharing expansion does not admit of this kind of procedure: the very process of overlapping puts it out of the question. The other factor of difference is the obligatory status of the position-sharing expansion (even in construc-tions of'"free"''object-sharing) against the optional status of the attribu-tive complicator.
The attributive semi-clause may contain in its head position a present participle, a past participle and an adjective. The present participial at-tributive semi-clause corresponds to the attributive subordinate clause with a verbal predicate in the active. E.g.: We found dry ground at the base of a tree looking toward the sun. > We found dry ground at the base of a tree that looked toward the sun.
Naturally, the present participial semi-clause of the attributive type cannot express an event prior to the event

of the dominant clause. So, an attributive clause of complete predicative character expressing such an event has no parallel in a participial attribu-tive semi-clause. E.g.: The squad that picked me up could have been scouts. > (*) The squad picking me up...
The past participial attributive semi-clause corresponds to the pas-sive attributive subordinate clause. E.g.: You can never rely on the in-formation received from that office. > You can never rely on the infor-mation which is received from that office.
The adjectival attributive semi-clause corresponds to the nominal at-tributive subordinate clause. E.g.: We admired the lilies white against the blue water. > We admired the lilies which were white against the blue water.
Semi-complex sentences of participial attributive complication formed by introducer constructions resemble subject-sharing semi-complex sentences. Cf.:
There is a river flowing through the town. > There is a river which flows through the town. This is John speaking. > This is John who is speaking.
Still closer to the subject-sharing semi-composite sentence stands the peculiar introducer or demonstrative construction whose attributive semi-clause has a finite verb predicate. This specific semi-complex sen-tence, formed much on the pattern of common subject overlapping, is called the "apo-koinou" construction (Greek "with a common element"). E.g.:
It was you insisted on coming, because you didn't like restaurants (S. O'Casey), He's the one makes the noise at night (E. Hemingway). And there's nothing more can be done (A. Christie).
The apo-koinou construction is considered here under the heading of the semi-complex sentence of attributive complication on the ground of its natural relation to the complex sentence with an attributive subordi-nate clause, similar to any common semi-complex sentence of the type in question. The apo-koinou construction should be classed as a familiar colloquialism of occasional use.
6. Semi-complex sentences of adverbial complication are derived from two base sentences one of which, the insert

sentence, is predicatively reduced and embedded in an adverbial position of the other one, the matrix sentence. E.g.:
The task was completed. + The task seemed a very easy one. > The task, when completed, seemed a very easy one. The windows were closed.-\-She did not hear the noise in the street. The windows being closed, she did not hear the noise in the street.
The subject of the insert sentence may be either identical with that of the matrix sentence (the first of the above examples) or not identical with it (the second example). This feature serves as the first fundamental basis for classifying the semi-complex sentences in question, since in the de-rived adverbial semi-clause the identical subject is dropped out and the non-identical subject is preserved. It will be reasonable to call the adver-bial semi-clause of the first type (i.e. referring to the subject of the domi-nant clause) the "conjoint" semi-clause. The adverbial complicator ex-pansion of the second type (i.e. having its own subject) is known under the name of the "absolute construction" (it will further be referred to as "absolutive").
The given classification may be formulated for practical purposes as the "rule of the subject", which will run as follows: by adverbialising scmi-complexing the subject of the insert sentence is deleted if it is iden-tical with the subject of the matrix sentence,
The other classificational division of adverbial semi-clauses concerns the representation of the predicate position. This position is only partially predicative, the role of the partial predicate being performed by the parti-ciple, either present or past. The participle is derived from the finite verb of the insert sentence; in other words, the predicate of the insert sentence is participialised in the semi-clause. Now, the participle-predicate of the adverbial semi-clause may be dropped out if the insert sentence, presents a nominal or existential construction (the finite verb be). Thus, in accord with this feature of their outer structure, adverbial semi-clauses are di-vided into participial and non-participial. E.g.:
One day Kitty had an accident. + She was swinging in the garden. > One day Kitty had an accident while swinging in the garden. (The parti-ciple is not to be deleted, being of an actional character.) He is very young.+ He is quite competent in this field. Though being very young, he is

quite competent in this field. > Though very young, he is quite compe-tent in this field. (The participle can be deleted, being of a linking na-ture.) She spoke as if being in a dream. > She spoke as if in a dream. (The predicate can be deleted, since It is expressed by the existential be.)
The two predicate types of adverbial semi-clauses, similar to the two subject types, can be briefly presented by the "rule of the predicate" as follows: by adverbialising semi-complexing the verb-predicate of the insert sentence is participialised, and may be deleted if it is expressed by be.
Conjoint adverbial semi-clauses are either introduced by adverbial subordinated conjunctions or joined to the dominant clause asyndeti-cally. The adverbial semantics expressed is temporal, broader local, causal, conditional, comparative. Cf. syndetic introduction of adverbial semi-clauses:
He was silent as if not having heard the call. > ...as if he had not heard the call. Read on unless told otherwise. > ... unless you are told otherwise. Although kept out of the press, the event is widely known in the diplomatic circles. > Although it is kept out of the press... When in London, the tourists travelled in double-deckers. > When they were in London...
Asyndetic introduction of adverbial semi-clauses is characteristic of temporal and causal constructions. Cf.:
Working on the book, the writer travelled much about the country. > When working on the book... Dialling her number, she made a mis-take. > While dialling her number... Being tired, I could not accept the invitation. > As I was tired...
As for the absolutive adverbial semi-clauses, they are joined to the dominant clause either asyndetically, or, mostly for the purpose of em-phasis, by the conjunction with. The adverbial semantics of the absolut-ive complicator expansion is temporal, causal, and attendant-circumstantial. E.g.:
Everything being settled, Moyra felt relieved. > As everything was settled... Two days having elapsed, the travellers set out on their way. When two days had elapsed...With all this work waiting for me, I can't afford to join their Sunday outing. > As all this work is waiting for me... "

The rule of the predicate is observed in absolulive complicators the same as in conjoint adverbial complicators. Its only restriction concerns impersonal sentences where the link-verb is not to be deleted. Cf.:
The long luncheon over, the business friend would bow and go his way. > When the long luncheon was over... It being very hot, the chil-dren gladly ran down to the lake. > As it was very hot...
7. Semi-complex sentences of nominal phrase complication are de-rived from two base sentences one of which, the insert sentence, is par-tially norninalised (changed into a verbid phrase of infinitival or gerun-dial type) and embedded in one of the nominal and prepositional adver-bial positions of the other sentence serving as the matrix. The nominal verbid constructions meet the demands both of economy and expressive-ness, and they are widely used in all the functional orders of speech. The gerundial phrase is of a more substantive semantic character, the infiniti-val phrase, correspondingly, of a more processual semantic character. The gerundial nominalisalion involves the optional change of the noun subject into the possessive, while the infinitival nominalisation involves the use of the preposition for before the subject. E.g.
Tom's coming late annoyed his mother. > The fact that Tom came late annoyed his mother. For him to come so late was unusual. > It was unusual that he came so late.
The rule of the subject exposed in connection with the adverbial semi-complexing (see above) applies also to the process of partial nomi-nalisation and is especially important here. It concerns the two types of subject deletion; first, its contextual identification; second, its referring to a general (indefinite) person. Thus, the rule can be formulated in this way: the subject of the verbid phrase is deleted when it is either identi-fied from the context (usually, but not necessarily, from the matrix sen-tence) or denotes an indefinite person. Cf. the contextual identification of the subject:
We are definite about it. > Our being definite about it. > Let's postpone being definite about it. Mary has recovered so soon. For Mary to have recovered so soon Mary is happy to have recovered so soon.

Cf. the indefinite person identification of the subject:
One avoids quarrels with strangers. One's avoiding quarrels with strangers. > Avoiding quarrels with strangers is always a wise pol-icy. One loves spring. For one to love spring.>It's but natural to love spring.
A characteristic function of the infinitive phrase is its use with sub-ordinative conjunctions in nominal semi-clauses. The infinitive in these cases implies modal meanings of obligation, admonition, possibility, etc. E.g.:
I wondered where to go. I wondered where I was to go. The question is, what to do next. > The question is, what we should do next.
In contrast with nominal uses of infinitive phrases, gerundial phrases are widely employed as adverbial semi-clauses introduced by prepositions. Semi-clauses in question are naturally related to the corre-sponding adverbial pleni-clauses. Cf.:
In writing the letter he dated it wrong. > White he was writing the letter he dated it wrong. She went away without looking back. > As she went away she didn't look back. I cleaned my breast by telling you eve-rything. > I cleaned my breast because I told you everything.
The prepositional use of gerundial adverbial phrases is in full accord with the substantival syntactic nature of the gerund, and this feature dif-ferentiates in principle the gerundial adverbial phrase from the particip-ial adverbial phrase as a positional constituent of the semi-complex sen-tence.
1. The semi-compound sentence is a semi-composite sentence built up on the principle of coordination. Proceeding from the outlined grammatical analysis of the composite sentence, the structure of the semi-compound sentence is derivationally to be traced back to minimum two base sentences having an identical element belonging to one or both of their principal syntactic positions, i.e. either the subject,

or the predicate, or both. By the process of semi-compounding, the sentences overlap round the identical element sharing it in coordinative fusion, which can be either syndetic or asyndetic. Thus, from the formal point of view, a sentence possessing co-ordinated notional parts of immediately sentential reference (directly related to its predicative line) is to be treated as semi-compound. But different structural types of syntactic coordina-tion even of direct sentential reference (coordinated subjects, predicates, objects, adverbial modifiers) display very different implications as regards semi-compounding composition of sen-tences.
By way of a general statement we may say that, other things being equal, the closer the coordinative group is related to the verb-predicate of the sentence, the more directly and explicitly it functions as a factor of sentence semi-compounding.
For instance, coordinated subjects connected asyndetically in an enumerative sequence or forming a plain copulative syn-detic string can hardly be taken as constituting so many shared though separately identified predicative lines with the verbal constituent of the sentence. As different from this, two subject-groups connected adversatively or antithetically are more "live" in their separate relation to the predicative centre; the derivative reference of such a sentence to the two source predicative con-structions receives some substantiality. E.g.:
There was nothing else, only her face in front of me. > There was nothing else in front of me.+There was only her face in front of me.
Substantially involved in the expression of semi-compounding is a combination of two subjects relating to one predicate when the subjects are discontinuously positioned, so that the first starts the utterance, while the second concludes it with some kind of process-referred introduction. Cf.:
The entrance door stood open, and also the door of the liv-ing-room. The entrance door stood open.+ The door of the living-room stood also open.
However, if we turn our attention to genuine coordinations of predicates (i.e. coordinations of non-repetitive or otherwise primitivising type), both verbal and nominal, we shall immedi-ately be convinced of each element of the group presenting its own predicative centre relating to the one

subject axis of the sentence, thereby forming a strictly com-pounding fusion of the predicative lines expressed. This fact is so trivially clear that it does not seem to require a special dem-onstration.
Hence, we will from now on treat the corresponding sen-tence-patterns with coordinate predicate phrases as featuring classes of constructions that actually answer the identifying definition of semi-compound sentence; in our further exposition we will dwell on some structural properties and functional se-mantics of this important sentence-type so widely represented in the living English speech in all its lingual divisions, which alone displays an unreservedly clear form of sentential semi-compounding out of the numerous and extremely diversified patterns of syntactic coordination.
2. The semi-compound sentence of predicate coordination is derived from minimum two base sentences having identical subjects. By the act of semi-compounding, one of the base sen-tences in most cases of textual occurrence becomes the leading clause of complete structure, while the other one is transformed into the sequential coordinate semi-clause (expansion) referring to the same subject. E.g.:
The soldier was badly wounded. +The soldier stayed in the ranks. > The soldier was badly wounded, but stayed in the ranks. He tore the photograph in half. + He threw the photo-graph in the fire. > He tore the photograph in half and threw it in the fire.
The rare instances contradicting the given rule concern in-verted constructions where the intense fusion of predicates in overlapping round the subject placed in the end position de-prives the leading clause of its unbroken, continuous presenta-tion. Cf.:
Before him lay the road to fame. + The road to fame lured him. > Before him lay and lured him the road to fame.
In case of a nominal predicate, the sequential predicative complement can be used in a semi-compound pattern without its linking part repeated. E.g.:
My manner was matter-of-fact, and casual. The savage must have been asleep or very tired.
The same holds true about coordinated verbids related
23-1499 ' 353

to a common finite verb in the function of an auxiliary or oth-erwise. E.g.:
The tiger was at large and burning with rage. He could not recall the face of the peasant girl or remember the feel of her.
By the number of bases joined, (and predicate phrases rep-resenting them) semi-compound sentences may be two-base (minimal) or multi-base (more than minimal two-base). The coordinated expansion is connected with the leading part either syndetically or asyndetically.
The syndetic formation of the semi-compound sentence ex-presses, first, copulative connection of events; then contrast, either comparative or adversative; furthermore, disjunction (al-ternation), consequence, limitation, elucidation. The conjunc-tive elements effecting this syndetic semi-compounding of sen-tences are both pure conjunctions and also words of adverbial nature. The pure conjunction and, the same as with pleni-compound sentences, expresses the unmarked semantic type of semi-compounding; the rest of the connectors render various marked types of it. The pure conjunctions used for semi-compounding, besides the copulative and, are monoconjunc-tions but, or, nor, and double (discontinuous) conjunctions both ... and, not only ... but also, either ... or, neither ... nor. The conjunctive adverbials are then, so, just, only.
Here are some examples of double-conjunctional forma-tions expressing, respectively, disjunction, simple copulative relation, copulative antithesis, copulative exclusion:
They either went for long walks over the fields, or joined in a quiet game of chess on the veranda. That great man was both a soldier and a born diplomat. Mary not only put up with his presence, but tried to be hospitable. I am neither for the pro-posal, nor against the proposal; nor participating in that sham discussion of theirs at all.
Cf. instances of conjunctive-adverbial introduction of predi-cate expansion rendering the functional meanings of action or-dering (then), of adversative-concessive relation (yet), of con-sequence (so), of limitation (just):
His beady eyes searched the clearing, then came back to my face. He was the tallest and bravest, yet was among those to give up life. I knew then that she was laughing,

so laughed with her. The Colonel didn't enlarge on the possible outcome of their adventure, just said a few words of warning against the abrupt turns of the mountain-pass.
With semi-compound sentences, similar to pleni-compound sentences, but on a larger scale, conjunctions combine with par-ticle-like elements of modal-adverbial description. These ele-ments supplement and specify the meaning of the conjunction, so that they receive the status of sub-conjunction specifiers, and the pairs "conjunction plus sub-conjunctive" become in fact regular conjunctive-coordinative combinations. Here belong such combinations as and then, and perhaps, and probably, and presently, and so, and consequently, etc; but merely, but only, but instead, but nevertheless, etc.; or else, or even, or rather, etc. The specifications given by the sub-conjunctives are those of change of events, probability evaluation, consequence in rea-soning, concessive contrast, limiting condition, intensity grada-tion, and many others, more specific ones. E.g.:
He waited for some moments longer and then walked down to the garden to where, on the terrace, the jeep was parked (H. E. Bates). She lived entirely apart from the contemporary liter-ary world and probably was never in the company of anyone more talented than herself (J. Austen). To his relief, she was not giving off the shifting damp heat of her anger, but instead was cool, decisive, material (J. Updike). For several hours I dis-cussed this with you, or rather vented exhaustive rewordings upon your silent phantom (J. Updike).
3. Of all the diversified means of connecting base sen-tences into a semi-compound construction the most important and by far the most broadly used is the conjunction and. Effect-ing the unmarked semi-compounding connection of sentences, it renders the widest possible range of syntactic relational meanings; as for its frequency of occurrence, it substantially exceeds that of all the rest of the conjunctives used for semi-compounding taken together.
The functional meanings expressed by the and-semi-compound patterns can be exposed by means of both coordina-tive and subordinative correlations. Here are some basic ones:
The officer parked the car at the end of the terrace and went into the Mission. > The officer parked the car ...,

then went into the Mission. (Succession of events, inviting a coordinative exposition) Suddenly the door burst open and Tommy rushed in panting for breath.> As the door burst open, Tommy rushed in ...("Successive simultaneity" of actions, in-viting a subordinative exposition) Patterton gavelled for atten-tion and speedily disposed of several routine matters. > Patter-ton gavelled for attention so that he could dispose and did dis-pose of several routine matters. (Purpose in successive actions, inviting a subordinative exposition) Her anger and emotion grew, and finally exploded. > Her anger and emotion grew to the degree that they finally exploded. (Successive actions in gradation, inviting a subordinative exposition) He just miscal-culated and won't admit it. Though he miscalculated, he won't admit it. (Concession in opposition, inviting a subordina-tive exposition) Mary promised to come and he was determined to wait. > He was determined to wait because Mary had prom-ised to come. (Cause and consequence, inviting a subordinative exposition)
Among the various connective meanings expressed by the conjunction and in combination with the corresponding lexemic constituents of the sentence there are two standing very promi-nent, due to the regular correlations existing between such con-structions and semi-complex patterns with verbid phrases infinitival and participial.
The first construction expresses a subsequent action of inci-dental or unexpected character:
He leaped up in time to see the Colonel rushing out of the door (H. E. Bates). > He leaped up in time and saw the Colo-nel rushing out of the door. Walker woke in his bed at the bourbon house to hear a strange hum and buzz in the air (M. Bradbury). > Walker woke in his bed at the bourbon house and heard a strange hum and buzz in the air.
In these constructions the leading clause, as a rule, includes verbs of positional or psychological change, while the expan-sion, correspondingly, features verbs of perceptions. As is seen from the examples, it is the semi-compound pattern that diag-noses the meaning of the pattern with the infinitive, not the re-verse. The infinitive pattern for its part makes up an expressive stylistic device by virtue of its outward coincidence with an infinitive pattern of purpose: the unexpectedness of the referent action goes together with the contextual unexpectedness of the construction.

The participial construction expresses a parallel attendant event that serves as a characteristic to the event rendered by the leading clause:
He sat staring down the gardens, trying to remember whether this was the seventh or eighth day since the attack had begun (H. E. Bates). > He was sitting and staring down the gardens, and was trying to remember... Rage flamed up in him, contorting his own face (M. Puzo). >Rage flamed up in him and contorted his own face.
With the participial pattern, the same as with the infinitival one, the diagnostic construction is the semi-compound sen-tence, not vice versa.
The nature of the shown correlations might be interpreted as a reason for considering the relations between the head-verb and the verbid in the tested patterns as coordinative, not subor-dinative. However, on closer analysis we must admit that diag-nosis of this kind is called upon to expose the hidden meanings, but not to level up the differences between units of opposed categorial standings. The verbid patterns remain part of the sys-tem of semi-complex sentences because of the hierarchical ranking of their notional positions, while the correlation with semi-compound sentences simply explain their respective se-mantic properties.
4. The asyndetic formation of the semi-compound sen-tence stands by its functional features close to the syndetic and-formation in so far as it does not give a rigorous characterisa-tion (semantic mark) to the introduced expansion. At the same time its functional range is incomparably narrower than that of the and-formation.
The central connective meaning distinguishing the asyndetic connection of predicative parts in semi-compound sentences is enumeration of events, either parallel or consecutive. In accord with the enumerative function, asyndetic semi-compounding more often than not is applied to a larger set of base sentences than the minimal two. E.g.:
He closed the door behind him with a shaking hand, found the old car in its parking place, drove along with the drifting lights. They talked, laughed, were perfectly happy late into the night.
Asyndetic semi-compound sentences are often used to

express gradation of intensity going together with a general emphasis. E.g.:
He would in truth give up the shop, follow her to Paris, fol-low her also to the chateau in the country (D. du Maurier). He never took the schoolbag again, had refused to touch it (J. Up-dike).
Characteristic of enumerative and gradational semi-compound sentences is the construction where the first two parts are joined asyndetically, and the third part syndetically, by means of the conjunction and. In such three-base constructions the syndetic expansion finalises the sentence both structurally and semantically, making it into an intensely complete utter-ance. E.g.:
He knows his influence, struts about and considers himself a great duellist. They can do it, have the will to do it, and are actually doing it.
Of the meanings other than enumerative rendered by the construction in question, the most prominent is elucidation combined with various connotations, such as consequence, pur-pose, additional characteristics of the basic event. Cf.:
The sight of him made me feel young again: took me back to the beaches, the Ardennes, the Reichswald, and the Rhine. I put an arm round her, tried to tease her into resting.
5. The number of predicative parts in a semi-compound sentence is balanced against the context in which it is used, and, naturally, is an essential feature of its structure. This number may be as great as seven, eight, or even more.
The connection-types of multi-base semi-compound sen-tences are syndetic, asyndetic, and mixed.
The syndetic semi-compound sentences may be homo-syndetic (i.e. formed by so many entries of one and the same conjunctive) and heterosyndetic (i.e. formed by different con-junctives). The most important type of homosyndetic semi-compounding is the and-type. Its functional meaning is enu-meration combined with copulation. E.g.:
A harmless young man going nowhere in particular was knocked down and trodden on and rose to fight back and was punched in the head by a policeman in mistake for someone else and hit the policeman back and ended in more trouble than if he had been on the party himself (M. Dickens).

A series of successive events is intensely rendered by a ho-mosyndetic construction formed with the help of the conjunc-tive then. E.g.: You saw the flash, then heard the crack, then saw the smoke ball distort and thin in the wind (E. Heming-way).
Another conjunctive pattern used in homosyndetic semi-compounding is the or-type in its different variants. E.g.:
After dinner we sat in the yard of the inn on hard chairs, or paced about the platform or stumbled between the steel sleepers of the permanent way (E. Waugh). Babies never cried or got the wind or were sick when Nurse Morrison fed them (M. Dickens).
By heterosyndetic semi-compounding the parts of the sen-tence are divided into groups according to the meanings of the conjunctives. Cf.:
A native woman in a sarong came and looked at them, but vanished when the doctor addressed her (S. Maugham). Ugly sat in the bow and barked arrogantly at passing boats, or stood rockily peering in the river (M. Dickens).
The asyndetic connections in semi-compound sentences, within their range of functions, are very expressive, especially when making up long enumerations-gradations. E.g.:
He had enjoyed a sharp little practice in Split, had meddled before the war in anti-Serbian politics, had found himself in an Italian prison, had been let out when the partisans briefly "lib-erated" the coast, had been swept up with them in the retreat (E. Waugh).
In the mixed syndetic-asyndetic semi-compound sentence various groupings of coordinated parts are effected. E.g.: He spun completely round, then fell forward on his knees, rose again and limped slowly on (E. Waugh).
In cases where multi-base semi-compound sentences are formed around one and the same subject-predicate combination, they are very often primitivised into a one-predicate sentence with coordinated secondary parts. Of these sentences, a very characteristic type is presented by a construction with a string of adverbial groups. This type of sentence expresses an action (usually, though not necessarily, a movement) or a series of ac-tions continued through a sequence of consecutive place- and time situations. E.g.: Then she took my hand, and we went down the steps of the tower

together, and through the court and to the walls of the rock-place (D. du Maurier).
The construction is very dynamic, its adverbial constituents preserve clear traces of the corresponding predications, and therefore it approaches the genuine semi-compound sentence of predicate coordination by its semantic nature.
6. The semi-compound sentence of predicate coordination immediately correlates with a compound sentence of complete composition having identical subjects. Both constructions are built upon the same set of base sentences, use the same connec-tive means and reflect the same situation, E.g.:
She looked at him and saw again the devotion, the humility in his eyes. > She looked at him and she saw again the devo-tion, the humility in his eyes (The latter sentence from D. du Maurier). The officer received the messengers, took their letters, and though I stood with them, completely ignored me. The officer received the messengers, took their letters, and though I stood with them, he completely ignored me (The latter sen-tence from H. E. Stover).
A question arises whether the compared sentences are abso-lutely the same in terms of functions and semantics, or whether there is some kind of difference between them which causes them to be used discriminately.
In an attempt to expose the existing functional difference between the two constructions, it has been pointed out that base sentences with identical subjects are connected not in a semi-compound, but into a compound sentence (of complete compo-sition) in the three main cases: first, when the leading sentence is comparatively long; second, when the finite verbs in the two sentences are of different structure; third, when the second sen-tence is highly emotional.* These tentative formulations should rather be looked upon as practical guides, for they do corre-spond to the existing tendencies of living speech. But the ten-dencies lack absolute regularity and, which is far more signifi-cant, they do not present complete lingual facts by themselves, but rather are particular manifestations of a general and funda-mental mechanism at work. This mechanism is embodied in the actual division of the
* Irtenyeva N. F., Shapkin A. P., Blokh M. Y. The Structure of the English Sentence. M., 1969, p. 110.

sentence: as a matter of fact, observations of the relevant con-texts show that the structure of the actual division in the two types of sentences is essentially different. Namely, whereas the actual division of the compound sentence with identical sub-jects presents two (or more) separate informative perspectives characterised by identical themes and different rhemes, the ac-tual division of the semi-compound sentence presents only one perspective, analysed into one theme and one, though complex, rheme; the latter falls into two or more constituent rhemes (sub-rhemes) in various concrete contexts.
The sub-rhemes may be of equal importance from the in-formational point of view, as in the following example: We were met by a guide who spoke excellent English and had a head full of facts.
The sub-rhemes may be of unequal informative importance, the predicative expansion rendering the basic semantic content of the sentence. E.g.: She gave us her address and asked us to come and see her.
The coordinated predicate groups may also be informatively fused into an essentially simple rheme, i.e. into a phrase mak-ing up a close informative unity. E.g.: He took out his diary and began to write. The man looked up and laughed.
As different from the semi-compound construction with its exposed informative properties, the very identity of the subject themes in a compound sentence of complete composition is a factor making it into a communicatively intense, logically ac-cented syntactic unit (compare the examples given at the begin-ning of the paragraph).
1. We have repeatedly shown throughout the present work that sentences in continual speech are not used in isolation; they are interconnected both semantically-topically and syntac-tically.
Inter-sentential connections have come under linguistic in-vestigation but recently. The highest lingual unit which was approached by traditional grammar as liable to syntactic study was the sentence; scholars even specially stressed

that to surpass the boundaries of the sentence was equal to sur-passing the boundaries of grammar.
In particular, such an outstanding linguist as L. Bloomfield, while recognising the general semantic connections between sentences in the composition of texts as linguistically relevant, at the same time pointed out that the sentence is the largest grammatically arranged linguistic form, i.e. it is not included into any other linguistic form by a grammatical arrangement.*
However, further studies in this field have demonstrated the inadequacy of the cited thesis. It has been shown that sentences in speech do come under broad grammatical arrangements, do combine with one another on strictly syntactic lines in the for-mation of larger stretches of both oral talk and written text.
It should be quite clear that, supporting the principle of syn-tactic approach to arrangement of sentences into a continual text, we do not assert that any sequence of independent sen-tences forms a syntactic unity. Generally speaking, sentences in a stretch of uninterrupted talk may or may not build up a co-herent sequence, wholly depending on the purpose of the speaker. E.g.:
Barbara. Dolly: don't be insincere. Cholly: fetch your con-certina and play something for us (B. Shaw).
The cited sequence of two sentences does not form a unity in either syntactic or semantic sense, the sentences being ad-dressed to different persons on different reasons. A discon-nected sequence may also have one and the same communica-tion addressee, as in the following case:
Duchess of Berwic... I like him so much. I am quite de-lighted he's gone! How sweet you're looking! Where do you get your gowns? And now I must tell you how sorry I am for you, dear Margaret (O. Wilde).
But disconnected sequences like these are rather an excep-tion than the rule. Moreover, they do not contradict in the least the idea of a continual topical text as being formed of gram-matically interconnected sentences. Indeed, successive sen-tences in a disconnected sequence mark the corresponding tran-sitions of thought, so each of them can potentially be expanded into a connected sequence bearing on one
* See: Bloomfield L. Language. N.-Y., 1933, p. 170. 362

unifying topic. Characteristically, an utterance of a personage in a work of fiction marking a transition of thought (and break-ing the syntactic connection of sentences in the sequence) is usually introduced by a special author's comment. E.g.:
"You know, L.S., you're rather a good sport." Then his tone grew threatening again. "It's a big risk I'm taking. It's the big-gest risk I've ever had to take" (C. P. Snow).
As we see, the general idea of a sequence of sentences forming a text includes two different notions. On the one hand, it presupposes a succession of spoken or written utterances ir-respective of their forming or not forming a coherent semantic complex. On the other hand, it implies a strictly topical stretch of talk, i.e. a continual succession of sentences centering on a common informative purpose. It is this latter understanding of the text that is syntactically relevant. It is in this latter sense that the text can be interpreted as a lingual element with its two distinguishing features: first, semantic (topical) unity, second, semantico-syntactic cohesion.
2. The primary division of sentence sequences in speech should be based on the communicative direction of their com-ponent sentences. From this point of view monologue se-quences and dialogue sequences are to be discriminated.
In a monologue, sentences connected in a continual se-quence are directed from one speaker to his one or several lis-teners. Thus, the sequence of this type can be characterised as a one-direction sequence. E.g.:
We'll have a lovely garden. We'll have roses in it and daffo-dils and a lovely lawn with a swing for little Billy and little Barbara to play on. And we'll have our meals down by the lily pond in summer (K. Waterhouse and H. Hall).
The first scholars who identified a succession of such sen-tences as a special syntactic unit were the Russian linguists N. S. Pospelov and L. A. Bulakhovsky. The former called the unit in question a "complex syntactic unity", the latter, a "super-phrasal unity". From consistency considerations, the corre-sponding English term used in this book is the "supra-sentential construction" (see Ch. I).
As different from this, sentences in a dialogue sequence are uttered by the speakers-interlocutors in turn, so that they are directed, as it were, to meet one another; the sequence

of this type, then, should be characterised as a two-direction sequence. E.g.: "Annette, what have you done?" "I've done what I had to do" (S. Maugham).
It must be noted that two-direction sequences can in prin-ciple be used within the framework of a monologue text, by way of an "inner dialogue" (i.e. a dialogue of the speaker with himself). E.g.: What were they jabbering about now in Parlia-ment? Some two-penny-ha'penny tax! (J. Galsworthy).
On the other hand, one-direction sequences can be used in a dialogue, when a response utterance forms not a rejoinder, but a continuation of the stimulating utterance addressed to the same third party, or to both speakers themselves as a collective self-addressee, or having an indefinite addressee. E.g.:
St. Erth. All the money goes to fellows who don't know a horse from a haystack. Canynge (profoundly). And care less. Yes! We want men racing to whom a horse means some-thing (J. Galsworthy). Elt. I'm glad we didn't go out tonight. Amanda. Or last night. El-yt. Or the night before. Amanda. There's no reason to, really, when we're cosy here (N. Cow-ard).
Thus, the direction of communication should be looked upon as a deeper characteristic of the sentence-sequence than its outer, purely formal presentation as either a monologue (one man's speech) or a dialogue (a conversation between two parties). In order to underline these deep distinguishing fea-tures of the two types of sequences, we propose to name them by the types of sentence-connection used. The formation of a one-direction sequence is based on syntactic cumulation of sentences, as different from syntactic composition of sentences making them into one composite sentence. Hence, the supra-sentential construction of one-direction communicative type can be called a cumulative sequence, or a "cumuleme". The formation of a two-direction sequence is based on its sentences being positioned to meet one another. Hence, we propose to call this type of sentence-connection by the term "occursive", and the supra-sentential construction based on occursive con-nection, by the term "occurseme".
Furthermore, it is not difficult to see that from the hierar-chical point of view the occurseme as an element of the system occupies a place above the cumuleme. Indeed, if the cu-muleme is constructed by two or more sentences joined by cumulation, the occurseme can be constructed by two

or more cumulemes, since the utterances of the interlocutors can be formed not only by separate sentences, but by cumula-tive sequences as well. E.g.:
"Damn you, stop talking about my wife. If you mention her name again I swear I'll knock you down." "Oh no, you won't. You're too great a gentleman to hit a feller smaller than yourself" (S. Maugham).
As we see, in formal terms of the segmental lingual hierar-chy, the supra-proposemic level (identified in the first chapter of the book) can be divided into two sublevels: the lower one "cumulemic", and the higher one "occursemic". On the other hand, a fundamental difference between the two units in question should be carefully noted lying beyond the hierar-chy relation, since the occurseme, as different from the cu-muleme, forms part of a conversation, i.e. is essentially pro-duced not by one, but by two or several speakers, or, linguisti-cally, not by one, but by two or several individual sub-lingual systems working in an intercourse contact.
As for the functional characteristic of the two higher seg-mental units of language, it is representative of the function of the text as a whole. The signemic essence of the text is exposed in its topic. The monologue text, or "discourse", is then a topi-cal entity; the dialogue text, or "conversation", is an exchange-topical entity. The cumuleme and occurseme are component units of these two types of texts, which means that they form, respectively, subtopical and exchange-sub-topical units as re-gards the embedding text as a whole. Within the framework of the system of language, however, since the text as such does not form any "unit" of it, the cumuleme and occurseme can simply be referred to as topical elements (correspondingly, topical and exchange-topical), without the "sub "-specification.
3. Sentences in a cumulative sequence can be connected either "prospectively" or "retrospectively".
Prospective ("epiphoric", "cataphoric") cumulation is ef-fected by connective elements that relate a given sentence to one that is to follow it. In other words, a prospective connector signals a continuation of speech: the sentence containing it is semantically incomplete. Very often prospective connectors are notional words that perform the cumulative function for the nonce. E.g.:

I tell you, one of two things must happen. Either out of that darkness some new creation will come to supplant us as we have supplanted the animals, or the heavens will fall in thunder and destroy us (B. Shaw).
The prospective connection is especially characteristic of the texts of scientific and technical works. E.g.:
Let me add a word of caution here. The solvent vapour drain enclosure must be correctly engineered and constructed to avoid the possibility of a serious explosion (From a technical journal).
As different from prospective cumulation, retrospective (or "anaphoric") cumulation is effected by connective elements that relate a given sentence to the one that precedes it and is semantically complete by itself. Retrospective cumulation is the more important type of sentence connection of the two; it is the basic type of cumulation in ordinary speech. E.g.:
What curious "class" sensation was this? Or was it merely fellow-feeling with the hunted, a tremor at the way things found one out? (J. Galsworthy).
4. On the basis of the functional nature of connectors, cumulation is divided into two fundamental types: conjunctive cumulation and correlative cumulation.
Conjunctive cumulation is effected by conjunction-like connectors. To these belong, first, regular conjunctions, both coordinative and subordinative; second, adverbial and paren-thetical sentence-connectors (then, yet, however, consequently, hence, besides, moreover, nevertheless, etc.). Adverbial and parenthetical sentence-connectors may be both specialised, i.e. functional and semi-functional words, and non-specialised units performing the connective functions for the nonce. E.g.:
There was an indescribable agony in his voice. And as if his own words of pain overcame the last barrier of his self-control, he broke down (S. Maugham). There was no train till nearly eleven, and she had to bear her impatience as best she could. At last it was time to start, and she put on her gloves (S. Maugham).
Correlative cumulation is effected by a pair of elements one of which, the "succeedent", refers to the other, the

"antecedent", used in the foregoing sentence; by means of this reference the succeeding sentence is related to the preceding one, or else the preceding sentence is related to the succeeding one. As we see, by its direction correlative cumulation may be either retrospective or prospective, as different from conjunc-tive cumulation which is only retrospective.
Correlative cumulation, in its turn, is divided into substitu-tional connection and representative connection. Substitutional cumulation is based on the use of substitutes. E.g.:
Spolding woke me with the apparently noiseless efficiency of the trained housemaid. She drew the curtains, placed a can of hot water in my basin, covered it with the towel, and retired (E. J. Howard).
A substitute may have as its antecedent the whole of the preceding sentence or a clausal part of it. Furthermore, substi-tutes often go together with conjunctions, effecting cumulation of mixed type. E.g.:
And as I leaned over the rail methought that all the little stars in the water were shaking with austere merriment. But it may have been only the ripple of the steamer, after all (R. Kipling).
Representative correlation is based on representative ele-ments which refer to one another without the factor of replace-ment. E.g.:
She should be here soon. I must tell Phipps, I am not in to any one else (O. Wilde). I went home. Maria accepted my de-parture indifferently (E. J. Howard).
Representative correlation is achieved also by repetition, which may be complicated by different variations. E.g.:
Well, the night was beautiful, and the great thing not to be a pig. Beauty and not being a pig\ Nothing much else to it (J. Galsworthy).
5. A cumuleme (cumulative supra-sentential construction) is formed by two or more independent sentences making up a topical syntactic unity. The first of the sentences in a cumuleme is its "leading" sentence, the succeeding sentences are "sequen-tial".
The cumuleme is delimited in the text by a finalising into-nation contour (cumuleme-contour) with a prolonged pause

(cumuleme-pause); the relative duration of this pause equals two and a half moras ("mora" the conventional duration of a short syllable), as different from the sentence-pause equalling only two moras.
The cumuleme, like a sentence, is a universal unit of lan-guage in so far as it is used in all the functional varieties of speech. For instance, the following cumuleme is part of the au-thor's speech of a work of fiction:
The boy winced at this. It made him feel hot and uncom-fortable all over. He knew well how careful he ought to be, and yet, do what he could, from time to time his forgetfulness of the part betrayed him into unreserve (S. Butler).
Compare a cumuleme in a typical newspaper article:
We have come a long way since then, of course. Unem-ployment insurance is an accepted fact. Only the most die-hard reactionaries, of the Goldwater type, dare to come out against it (from Canadian Press).
Here is a sample cumuleme of scientific-technical report prose:
To some engineers who apply to themselves the word "prac-tical" as denoting the possession of a major virtue, applied re-search is classed with pure research as something highbrow they can do without. To some business men, applied research is something to have somewhere in the organisation to demon-strate modernity and enlightenment. And people engaged in applied research are usually so satisfied in the belief that what they are doing is of interest and value that they are not particu-larly concerned about the niceties of definition (from a techni-cal journal).
Poetical text is formed by cumulemes, too:
She is not fair to outward view, | As many maidens be; | Her loveliness I never knew | Until she smiled on me. |Oh, then I saw her eye was bright, | A well of love, a spring of light (H. Coleridge).
But the most important factor showing the inalienable and universal status of the cumuleme in language is the indispensa-ble use of cumulemes in colloquial speech (which is reflected in plays, as well as in conversational passages in works of vari-ous types of fiction).

The basic semantic types of cumulemes are "factual" (narra-tive and descriptive), "modal" (reasoning, perceptive, etc.), and mixed. Here is an example of a narrative cumuleme:
Three years later, when Jane was an Army driver, she was sent one night to pick up a party of officers who had been test-ing defences on the cliff. She found the place where the road ran between a cleft almost to the beach, switched off her engine and waited, hunched in her great-coat, half asleep, in the cold black silence. She waited for an hour and woke in a fright to a furious voice coming out of the night (M. Dickens).
Compare this with modal cumulemes of various topical standings:
She has not gone? I thought she gave a second performance at two? (S. Maugham) (A reasoning cumuleme of perceptional variety)
Are you kidding? Don't underrate your influence, Mr. O'Keefe. Dodo's in. Besides, I've lined up Sandra Straughan to work with her (A. Hailey). (A remonstrative cumuleme)
Don't worry. There will be a certain amount of unpleasant-ness but I will have some photographs taken that will be very useful at the inquest. There's the testimony of the gunbearers and the driver too. You're perfectly all right (E. Hemingway). (A reasoning cumuleme expressing reassurance) Etc.
6. Cumuleme in writing is regularly expressed by a para-graph, but the two units are not wholly identical.
In the first place, the paragraph is a stretch of written or typed literary text delimited by a new (indented) line at the be-ginning and an incomplete line at the close. As different from this, the cumuleme, as we have just seen, is essentially a feature of all the varieties of speech, both oral and written, both literary and colloquial.
In the second place, the paragraph is a polyfunctional unit of written speech and as such is used not only for the written representation of a cumuleme, but also for the introduction of utterances of a dialogue (dividing an occurseme into parts), as well as for the introduction of separate points in various enu-merations.
In the third place, the paragraph in a monologue speech can contain more than one cumuleme. For instance, the
241499 369

following paragraph is divided into three parts, the first formed by a separate sentence, the second and third ones presenting cumulemes. For the sake of clarity, we mark the borders be-tween the parts by double slash:
When he had left the house Victorina stood quite still, with hands pressed against her chest. // She had slept less than he. Still as a mouse, she had turned the thought: "Did I take him in? Did I?" And if not what? // She took out the notes which had bought or sold their happiness, and counted them once more. And the sense of injustice burned within her (J. Galsworthy).
The shown division is sustained by the succession of the forms of the verbs, namely, the past indefinite and past perfect, precisely marking out the events described.
In the fourth place, the paragraph in a monologue speech can contain only one sentence. The regular function of the one-sentence paragraph is expressive emphasis. E.g.:
The fascists may spread over the land, blasting their way with weight of metal brought from other countries. They may advance aided by traitors and by cowards. They may destroy cities and villages and try to hold the people in slavery. But you cannot hold any people in slavery.
The Spanish people will rise again as they have always risen before against tyranny (E. Hemingway).
In the cited passage the sentence-paragraph marks a transi-tion from the general to the particular, and by its very isolation in the text expressively stresses the author's belief in the invin-cible will of the Spanish people who are certain to smash their fascist oppressors in the long run.
On the other hand, the cumuleme cannot be prolonged be-yond the limits of the paragraph, since the paragraphal border-marks are the same as those of the cumuleme, i.e. a characteris-tic finalising tone, a pause of two and a half moras. Besides, we must bear in mind that both multicumuleme paragraphs and one-sentence paragraphs are more or less occasional features of the monologue text. Thus, we return to our initial thesis that the paragraph, although it is a literary-compositional, not a purely syntactic unit of the text, still as a rule presents a cumuleme; the two units, if not identical, are closely correlative.

7. The introduction of the notion of cumuleme in linguis-tics helps specify and explain the two peculiar and rather im-portant border-line phenomena between the sentence and the sentential sequence.
The first of these is known under the heading of "parcella-tion". The parcellated construction ("parcellatum") presents two or more collocations ("parcellas") separated by a sentence-tone but related to one another as parts of one and the same sentence. In writing the parts, i.e., respectively, the "leading parcella" and "sequential parcella", are delimited by a full stop (finality mark). E.g.:
There was a sort of community pride attached to it now. Or shame at its unavoidability (E.Stephens). Why be so insistent, Jim? If he doesn't want to tell you (J. O'Hara). ...I realised I didn't feel one way or another about him. Then. I do now (J. O'Hara).
Having recourse to the idea of transposition, we see that the parcellated construction is produced as a result of transposing a sentence into a cumuleme. This kind of transposition adds topi-cal significance to the sequential parcella. The emphasising function of parcellation is well exposed by the transformation of de-transposition. This transformation clearly deprives the sequential parcella of its position of topical significance, chang-ing it into an ordinary sentence-part. Cf.:
... > There was a sort of community pride attached to it now or shame at its unavoidability. ...> Why be so insistent, Jim, if he doesn't want to tell you? ... > I didn't feel one way or another about him then.
With some authors parcellation as the transposition of a sen-tence into a cumuleme can take the form of forced paragraph division, i.e. the change of a sentence into a supra-cumuleme. E.g.:
... It was she who seemed adolescent and overly concerned, while he sat there smiling fondly at her, quite self-possessed, even self-assured, and adult.
And naked. His nakedness became more intrusive by the second, until she half arose and said with urgency, "You have to go and right now, young man" (E. Stephens).
The second of the border-line phenomena in question is the opposite of parcellation, it consists in forcing two
24- 371

different sentences into one, i.e. in transposing a cumuleme into a sentence. The cumuleme-sentence construction is charac-teristic of uncareful and familiar speech; in a literary text it is used for the sake of giving a vivid verbal characteristic to a personage. E.g.:
I'm not going to disturb her and that's flat, miss (A. Christie). The air-hostess came down the aisle then to warn passengers they were about to land and please would everyone fasten their safety belts (B. Hedworth).
The transposition of a cumuleme into a sentence occurs also in literary passages dealing with reasoning and mental percep-tions. E.g.:
If there were moments when Soames felt cordial, they were such as these. He had nothing against the young man; indeed, he rather liked the look of him; but to see the last of almost anybody was in a sense a relief; besides, there was this question of what he had overheard, and to have him about the place without knowing would be a continual temptation to compro-mise with one's dignity and ask him what it was (J. Galswor-thy).
As is seen from the example, one of the means of transpos-ing a cumuleme into a sentence in literary speech is the use of half-finality punctuation marks (here, a semicolon).
8. Neither cumulemes, nor paragraphs form the upper limit of textual units of speech. Paragraphs are connected within the framework of larger elements of texts making up different paragraph groupings. Thus, above the process of cu-mulation as syntactic connection of separate sentences, supra-cumulation should be discriminated as connection of cumule-mes and paragraphs into larger textual unities of the corre-spondingly higher subtopical status. Cf.:
... That first slip with my surname was just like him; and af-terwards, particularly when he was annoyed, apprehensive, or guilty because of me, he frequently called me Ellis.
So, in the smell of Getliffe's tobacco, I listened to him as he produced case after case, sometimes incomprehensibly, be-cause of his allusive slang, often inaccurately. He loved the law (C. P. Snow).

In the given example, the sentence beginning the second paragraph is cumulated (i.e. supra-cumulated) to the previous paragraph, thus making the two of them into a paragraph grouping.
Moreover, even larger stretches of text than primary para-graph groupings can be supra-cumulated to one another in the syntactic sense, such as chapters and other compositional divi-sions. For instance, compare the end of Chapter XXIII and the beginning of Chapter XXIV of J. Galsworthy's "Over the River":
Chapter XXIII. ... She went back to Condaford with her fa-ther by the morning train, repeating to her Aunt the formula: "I'm not going to be ill."
Chapter XXIV. But she was ill, and for a month in her con-ventional room at Condaford often wished she were dead and done with. She might, indeed, quite easily have died...
Can, however, these phenomena signify that the sentence is simply a sub-unit in language system, and that "real" informa-tive-syntactic elements of this system are not sentences, but various types of cumulemes or supra-cumulemes? In no wise.
Supra-sentential connections cannot be demonstrative of the would-be "secondary", "sub-level" role of the sentence as an element of syntax by the mere fact that all the cumulative and occursive relations in speech, as we have seen from the above analysis, are effected by no other unit than the sentence, and by no other structure than the inner structure of the sentence; the sentence remains the central structural-syntactic element in all the formations of topical significance. Thus, even in the course of a detailed study of various types of supra-sentential con-structions, the linguist comes to the confirmation of the classi-cal truth that the two basic units of language are the word and the sentence: the word as a unit of nomination, the sentence as a unit of predication. And it is through combining different sen-tence-predications that topical reflections of reality are achieved in all the numerous forms of lingual intercourse.

! . - .



2005-2024. ! homeenglish@mail.ru