THE SYSTEM OF GRAMMAR Most grammars, at any rate most of those dealing with our own family of speech, are built up in the traditional way with the following main divisions:— I. Phonology. This treats of the general theory of the sounds and sound-combinations of the language concerned, and expounds the orthography, where there is occasion.
II. Accidence or Morphology, the theory of forms (German Formenlehre is a better term than those used in English). This generally treats of the traditional "parts of speech" in their usual order, substantives, adjectives, etc. The main subject is the changes words undergo in flexion, paradigms being given which show all the forms of one and the same typical word; but the point of view is not pursued consistently, for under "numerals" we generally find an enumeration of all these words in their natural order, though most of them are subject to no formal changes.
III. Word-formation, dealing with prefixes, suffixes and other Iii. Word-formation, dealing with prefixes, suffixes and other means of forming one word from another.
IV. Syntax, generally in its first part taking the parts of speech separately as in II. and stating the rules for the use of each case, tense, mood, etc. A second part then deals with word-order, etc.
This system, which varies a good deal in details, has been re peatedly criticized (by J. Ries, Noreen, Jespersen), but no other system has been universally accepted. In France, F. Brunot has proposed basing the teaching of grammar not, as is usually done, on the forms from which the pupils proceed to their syntactical use, but on the inner meanings expressed by grammatical phe nomena, stating in each case secondarily the external forms, etc., which are used to express them.
As a matter of fact, grammatical phenomena can be viewed from two different angles: one that of the hearer (reader), to whom a certain series of sounds (letters) is presented, the inner meaning of which it is his task to understand : he begins from the outside and moves inwards ; the other that of the speaker (writer) : he has certain ideas which he wants to communicate to others; he therefore has to choose the forms (sounds, etc.) that serve best to express these ideas : he moves from within to without.
We are thus led to the following two main divisions of gram mar : I The theory of Forms. II. The theory of Notions. Both deal with the same grammatical facts but from opposite points of view.
I. The Theory of Forms.—The following is a systematic survey of the external means used in languages for grammatical purposes.
(r) A simple sequence of words. This is seen, for instance, in compound substantives like post-office. The importance of the order in which words are arranged, is seen in cases like garden flower and flower-garden, where the first element limits and de fines the meaning of the second, and in the distinction between Paul loves Ann and Ann loves Paul, where word-order shows which is subject and which object.
(2) "Empty words," i.e., words which have no proper mean ing of their own, but merely serve to indicate the relations of other words. Examples: of in "the father of the boy" (=the boy's father), "the City of Rome," "that scoundrel of a servant"; to in "I want to hear," "he refused food to the poor" ; that in "I saw that he came," etc. There is no hard-and-fast distinction between full and empty words; to in "I give food to the poor" has still something of its local meaning found in "go to London,' etc.
(3) Prefixes, e.g., for- in forbid, be- in besiege.
(4) Infixes, e.g., n in Latin vinco, cf. the perfect vici, English stand, cf. stood, messenger, cf. message.
(5) Suffixes, e.g., -ness in goodness, -en in blacken; these can not be separated from such "inflexional endings" as -s in kings, -en in oxen.
(3) (4) (5) together may be termed affixes. The origin of some of these is quite obvious : they were at one time independent words joined to other words like those in (r ). A word may easily be accentually subordinated to another with which it is continu ally combined, especially if the combination acquires a meaning of its own, independent of that of each element, as in blackbird; in postman, the vowel of the second element is obscured, and in other cases further phonetic changes take place ; gentlemanlike and gentlemanly show two stages in the development of a suffix from what was originally an independent word ; for- in forgive is an old preposition, though perhaps not exactly identical with the ordinary for; be- is a weakened form of by. But it is not all affixes that originate in this manner from independent words : -en in oxen originally belonged to the stem of the word in all its forms, and it was only through the accident of this syllable hav ing been lost in the singular, but not in the plural, that it came to be felt as an affix to denote the plural number. The origin of most of our affixes is hopelessly obscure.
(6) Change in intonation, stress or quantity, e.g., Yes? with a rising tone in a question, Yes with a falling tone as an affirma tive answer; object with varying stress according as it is a sub stantive or a verb.
(7) Consonantal changes, e.g., send, sent; half, halve; use as a substantive with unvoiced, as a verb with voiced consonant.
(8) Vocalic changes, e.g., feed, fed; see, saw; man, men; drink, drank, drunk.
(6) (7) (8) are phonetic changes, which may be due to the most different causes; some are recent, others go back to the most remote times; some have only in various circuitous ways acquired significative grammatical importance.
(9) Combined changes, affixes like those in (3) (4) (5) being joined to phonetic changes like those in (6) (7) (8) . Examples: forgot (3) (8), forgotten (3) (8) (5), drunken (5) (8), halves (5) (7), men's (5) (8), won't (5) (7) (ro) Supplementing with different stems: I, me, we, us; am, is, was, been; good, better.
Some languages make a more extensive use of some of these grammatical means than of others. Chinese uses scarcely any thing but word order and empty words; some languages are pre dominantly prefix-languages, as for instance, the Bantu family ; others predominantly suffix-languages, e.g., Eskimo and Turkish. As will be seen from the examples, English uses all these means freely, though there are few examples of infixes.
We must here mention a classification of all the languages of the world according to their morphological system, which played a great role in the discussions of the i 9th century, but has now been given up as superficial, namely, into (r) isolating languages or root-languages like Chinese, (2) agglutinative languages like Finnish and Turkish, which use affixes, but have no internal changes in the roots, (3) flexional languages like those of the Ar yan and Semitic families. The last were also supposed to have gone through the isolating and agglutinative stages in their pre historical development, while Chinese was thought to represent the earliest childlike linguistic structure. The latter supposition has been shown to be wrong, as the earliest Chinese in some re spects was "flexional," and those hundreds of languages that were formerly classed together as "agglutinative" represent the most diverse types of morphological structure. The world is more com plex than our ancestors imagined.
(I ) Parts of speech. It is usual to divide words grammatically into the following classes and to define them somewhat as is here (very succinctly) indicated : (a) Substantives—denoting "persons" and "things." (b) Adjectives—showing qualities.
Substantives and adjectives are often classed together as "nouns," but many grammarians make the term "noun" equiva lent to "substantives," and do not comprise under it adjectives.
(c) Pronouns—used instead of nouns "to designate a per son or thing already mentioned or known or forming the subject of inquiry." Various well-known subclasses: per sonal, demonstrative, relative, interrogative, indefinite. The so-called articles, as well as numerals, are best treated as subdivisions of pronouns.
(d) Verbs—denoting actions, states or happenings.
(e) Adverbs—serving to modify adjectives or verbs.
(f) Prepositions—marking relations between words.
(g) Conjunctions—used to connect clauses or to co ordinate words in the same clause.
(h) Interjections—ejaculations, standing outside ordinary sentences.
This division and the definitions usually given have, however, been subjected to severe criticism and should not be taken at their face value. One of the chief difficulties with substantives is the existence of such words as arrival and kindness, which are un doubtedly substantives and are treated grammatically as such, but cannot be termed names of "things" ; they represent "nexus" (see p. 615). Adverbs, prepositions and conjunctions are best classed together as "particles." (2) Rank. While the division under (I) concerns words sep arately, we here have a distinction that has regard to words or word-groups in combinations, namely into : (a) Primary (b) Secondary (c) Tertiary—words or word-groups.
The three ranks to some extent, but only to some extent, correspond to substantives, adjectives and adverbs respectively. Secondary elements serve to modify or delimit primaries, tertiary elements to modify or delimit secondaries, as will be seen from the following examples in which those words or groups are itali cized which belong to the rank under which they are classed : (a) The King's palace. The King arrived. I know when he arrived.
(b) The King's palace. A big palace. The palace that he built.
(c) A really big palace. I was present when he arrived.
The combination of a primary and a secondary element in the way exemplified under (b) is termed junction; the adjective or (relative) clause standing as secondary is called an adjunct. Tertiaries are also termed (3) Other classifications. Many languages classify words in such a way that a class is indicated either in the word itself or in the form required in those adjuncts, etc., which belong to it. Sometimes the distinction is into animate and inanimate, some times into big and small things, sometimes into male, female and sexless ; but such distinctions are rarely indicated with what we should call consistency; some languages, for instance, that have the main distinction "animate : inanimate," reckon certain parts of the human body as animate, others as inanimate. Sometimes it is impossible to see what is really the notional basis of a classi fication. When the distinction is connected with sex, as in most of the Aryan languages, we speak of gender; but the actual dis tinction between masculine, feminine and neuter gender does not correspond at all exactly with that between male and female beings and sexless things ; very often it is impossible to discern why one word belongs to one gender rather than to another. In Old English, for instance, stan stone, daeg day, finger, wi f mann woman were masculine; niht night, ecg edge, hand feminine ; and treow tree, gear year, blod blood, wif wife neuter. This word gender, which is still found in German, and which influenced the flexion of the words and the form of the article and adjective belonging to them, disappeared gradually from English in the Middle English period. In Semitic languages, the sex of the subject influences the form of the verb.
(4) Number. On the dual number see above. The distinction between "one" and "more than one" is very easy from a no tional point of view, but not always so easy grammatically, partly because some things may be looked upon either as units (as German brille) or as composite (Engl. spectacles), partly from other causes. A collective is a word which though singular in form denotes a plurality, hence such anomalies as twenty police; cf. also "my family is an old one" and "my family are early risers." Number properly belongs to primaries only, but many languages require secondaries to agree in number with their primaries, e.g., those trees, and German die hohen bbume the big trees, where in English the article and adjective are invariable. In English verbs the distinction has been given up in all past tenses, e.g., he went, they went (except he was, they were) , in the present tense it is preserved in the third person only : I go, we go; he goes, they go.
(5) Person, i.e., the distinction between the speaker, the person (or persons) addressed, and what is neither speaker nor spoken to. The distinction is shown in pronouns and in many languages also in the verb. The plural "we" does not mean two or more "first persons," but "I+you" or "I+some one or more persons besides," and some languages make a distinction accord ing as the second person is included or not. Such a pronoun as French on, Engl. one, may be considered a "common person." (On reflexive pronouns, see above.) (6) Space. Some languages have different forms according to distance from the speaker, etc. Case-f orms denoting existence in or at a place and movement to, towards or from a place are very frequent. In our languages, with a view to greater pre cision, such case-forms were frequently supplemented by ad verbs, and these in time became prepositions governing the cases which at first were sufficient in themselves to denote the spatial relation; eventually the case-endings were often dropped as super fluous.
(7) Time. With substantives the same means (case-forms, adverbs, prepositions) as are used to indicate spatial relations are as a rule also used to denote time relations. But with verbs many, or perhaps most, languages have separate means of denot ing time-relations, which cannot surprise us, as the idea of time is naturally associated with that of action or happening. But while the notional division of time into past, present and future is quite simple, mankind has not, as a rule, found correspondingly simple grammatical expressions for time and its subdivisions, such distinctions as that between permanent and transitory, or between finished and unfinished, or between once and repeatedly, or between stability and change, or between resultative and non resultative action being often inextricably connected with real time-indications in the "tenses" of verbs. Expressions for the future are often much more vague than those for the past, and frequently expressions which at first had and still to some extent have the meaning of volition or obligation or motion are made to do duty as a kind of future tense, as in "he will come," "I shall come," "they are going to start" (French "on va partir"), etc. Some languages have very elaborate tense systems with separate forms for imperfect, aorist, perfect, pluperfect, future in the past, etc., others rely more on the context or on adverbs for such nuances, if they are conceived at all.
(8) Comparison. The superlative ("strongest," etc.) is really a kind of comparative : "he is the strongest of the boys" means the same thing as "he is stronger than the other boys," the dif ference being only that the result in the former sentence is stated with regard to all boys, himself included, while in the latter he is excluded. A comparison results in expression of in equality or equality, as in "he is stronger than X" (a), "he is as strong as X" (b), "he is less strong than X" (c) ; of these (a) and (c) are closely connected as they both denote inequality and therefore use the comparative. Many languages even for this sense use the positive form and say "strong from X" or the like.
(9) Nexus. This is a comprehensive term for the combination of two words (or word-groups) which stand to another in the relation of subject to predicate. The simplest case is a sentence with a subject and a verb, as "the doctor arrived" or with a subject, an "empty" verb ("copula") and a predicative, as "the doctor is clever." Compare also sentences without a verb like "Happy the man who. . ." and "He a doctor!" There are other cases of nexus, in which the nexus does not in the same way as here form a whole sentence, but only part of one, as in "the doctor's arrival," "the doctor's cleverness," "(I saw) the doctor arrive," "(we thought) the doctor clever," "(we count on) the doctor to arrive," "(he slept with) the window open," "every thing considered (he must be clever)." A nexus of a different kind exists between a verb and its object, as in "we saw the doctor," or its two objects, as in "we offered the doctor money." Further, the theory of nexus leads to a con templation of the relation between the active and the passive expression for one and the same thought : what in the active turn is an object, is made into a subject in the passive turn: "the doctor was seen (by us)," "money was offered (to) the doctor," "the doctor was offered money." (To) Affirmation and negation. In some languages the verb has special forms for negation : this is to a certain extent true of English, especially in its colloquial form : won't, cf. will, shan't, cf. shall; note also the use of the auxiliary do in most negative sentences which contain no other auxiliary : "The doctor did not arrive." (I I) Subjective attitude of the speaker. By the side of simple ("flat") assertions we find others in which the speaker does not want to commit himself, but speaks with a certain hesi tation, doubt, hope or fear, and such emotional repressions often manifest themselves grammatically, either in particles like Greek av or in special forms of the verb (chiefly the subjunctive mood) . The same means are frequently applied in conditioned clauses, which range from those in which doubt is not expressed at all or slightly hinted at, to those in which unreality is expressly indi cated. In the latter kind some languages use a special conjunc tion, while others show that "the condition is rejected" by shifting the mood into the subjunctive and the tense into the preterite or by the latter means alone. In connection with this must be mentioned the expression of diffidence or modesty in questions like "Could you (Would you) lend me a pound?" as against the simple and direct "Can you (Will you) . . . ?" and the difference between the unrealizable wish in "Would he were still alive!" and the realizable wish in "May he be still alive!" (I 2) Relation to the will of the hearer. In one class of utter ances (ordinary statements and exclamations, for example) the speaker does not want to influence the will of the hearer. The aim of another class is to influence the will of the hearer, that is, to make him do something. This may be effected by requests, which range from brutal commands or orders through demands, implorations, invitations, etc., to the most humble entreaty or supplication. One of the linguistic forms for requests is the im perative, other forms are seen in "One minute!" and "Hands off ! " Questions belong to requests, as they imply a request (command, prayer, etc.) to give the original speaker a piece of information. They are of two distinct kinds according to the existence or non existence of an "unknown quantity" expressed by means of an interrogative pronoun or adverb: "Who said that?" "What did he say?" and "When did he say that?" are examples of one kind, "Did he say that?" of the other kind. Questions, and requests generally, are naturally liable to those influences which were dealt with under (I I) ; questions are likewise notionally related to negations, whence they often employ similar grammatical means; this is seen, for instance, in the English use of the auxiliary do in both kinds of sentences.
The system here given shows how a notional arrangement leads to the separation of things which in the ordinary grammatical system are placed together. Under case, accordingly, we have things which have relation to junction (the chief use of the gen itive is to make a word the adjunct of another word), to space (the so-called locative cases) and to nexus (nominative to denote the subject, accusative and dative, for various kinds of objects). Here, as elsewhere, we see that linguistic phenomena are capable of being viewed from different angles and that they present all kinds of intersections and overlappings.
In the treatment of each particular language we meet with units which are units neither from the purely formal nor from the purely notional point of view, but which nevertheless must be taken together as what might be called functional units. Take the English preterite : it is not a formal unit, because it is formed in different ways: ended from end, sent from send, thought from think, put from put, saw from see, was from be, etc. Neither is it a notional unit, for sometimes it indicates the past time pure and simple, sometimes unreality ("if he came"), or modesty ("Could you . . . . ?") or even future time ("it is time you went to bed"), and it has even more spheres of application. Yet all these formal and notional things go together and form one separate unit in English grammar, which is different from such units in any foreign grammar as in some ways correspond to it : in French, for instance, we have two or three tenses (je finissais and je finis or colloquially j'ai fini corresponding to I ended), each of which is a unit in the same way as the English preterite is. But all the units we arrive at through our analysis of grammatical phenomena are at best symbols or shadowings of the innermost notional categories.
Sweet, A New English Grammar, Logical and Historical (1892) ; F. Brunot, La Pensee et la Langue (1922) ; A. Noreen, Einf uhrung in die Wissenschaftliche Betrachtung der Sprache (1923) ; O. Jespersen, The Philosophy of Grammar (1924) ; R. Lenz, La Oracion y sus Partes (1925) ; A. Sechehaye, Essai sur la Structure Logique de la Phrase (1926). (O. J.)