Grammar

Grammar deals with the usage of some one form of speech. It may be described as a section of the larger science of language (see article PHILOLOGY), which treats of the origin, development, and general character of the principal families of language and of human speech as a whole. In common use, however, grammar means not a branch of science, but a treatise on some one well-defined form of speech as used in the present day, as by French grammar we mean a book on the usage of Paris; by English grammar we mean an account of the language spoken and written by educated men throughout Great Britain, which language, however, is only one dialect of English speech, the East Midland. That dialect by favouring conditions has superseded the other dialects, southern and northern, which were once spoken and written, and are still in a lessening degree spoken, in different parts of the island.

Grammar has two parts. The first describes the forms of a language, the single words which occur in it, its nouns, verbs, &c.; and its modifications of such forms, the cases of its nouns, the persons and tenses of its verbs, &c., used to express modifications of the same idea, as 'child,' 'child's,' 'children,' 'spring,' 'sprang,' 'sprung,' in English. This is called the morphology of a language, or (more loosely) its etymology. The second part deals with the use of these forms in combination: their syntax—i.e. their arrangement in order of speech. The general principles of this will vary little in the different languages of the same family; but each language has its idioms, as we call them, its own special refinements of usage, and it is in the clear discrimination of these that the practical value of a grammar lies.

Grammar in this function may be called special. It does not enter into the history of the forms which it describes; it is sufficient if it sets forth what they are at a particular time, without showing how they became such. But it is possible to a considerable extent to trace the history of these forms—e.g. we can see how literary English has developed out of the English of Chaucer, and that from the English of an earlier day, how the forms have changed mostly in the direction of uniformity, and how (to a lesser degree) their syntax has altered. To trace this belongs to historical grammar, and some of the results of this science are now commonly given in each special grammar. Lastly we can compare together the forms and usage of cognate dialects. We can compare, e.g., the grammar of our literary English dialect and that of the speech of Dorset, as set forth by Mr Barnes; and, employing the results of historical grammar, we can trace back the varying development of English speech as a whole; or we can compare the development and trace the connection of English and of German speech, and the relation of each of these to Latin or to Greek, till we arrive at some knowledge of a common speech of which all these are only derived forms. This is the work of comparative grammar.

Naturally, we do not learn our own speech from a written grammar. A child learns his words and their use from those around him, not as a whole, but one by one; and he forms new words for himself on the analogy of those he has already acquired. When he finds that any of these formations are not used by others he rejects them, and so he assimilates his speech to that of those around him. It is when we have to deal with a speech which is not our own, either that of a foreign nation, or of our own language at some earlier period, or of some dialect of our own language, that we need a grammar. The earliest works on grammar were due to the second of these causes. At Alexandria, the great commercial and literary centre of Greece in the days when the separate Greek states had ceased to be autonomous, there was for the first time a huge collection of the works of earlier writers, especially the Homeric poems. The age was one destitute of original ability; the loss of freedom had caused the loss of the motives which had produced the literature of the past. But it contained a large number of literary men, whose activity was chiefly spent on the work of their predecessors. This was to them in language and in style archaic; it required glosses—as we should say, glossaries—and explanations of disused forms. Hence arose the first grammarians, men often of conspicuous ability in their own line, such as Zenodotus and Aristarchus. At a later time, Romans who wished to learn Greek had grammars based upon Greek models, compiled for them in Latin, and these have been the parents of all European grammars to the present day. The grammatical terms with which we are familiar are consequently in the main Latin translations of Greek originals, and because of this they are often less intelligible than they might be.

It is to the Greeks that we owe the number of the so-called 'parts of speech.' But their eight were not the same as ours. They had (1) the noun; (2) the verb (terms which go back to Aristotle, though in his use the 'verb' meant all that is logically called the predicate); (3) the participle, so called because it partook of the nature of both the noun and the verb—it was a noun in form, yet it governed a case like a verb; (4) the article; (5) the pronoun; (6) the preposition, so called not as being placed before a case, but as set before a verb or noun in composition; (7) the adverb—i.e. the 'additional predication,' not anything specially belonging to the verb, as the Latin name seems to imply; (8) the conjunction. The Romans modified this list. First, they rejected the participle, and supplied its place by dividing the noun into the substantive and the adjective; this is a gain to logic, but as a matter of history the two go back to the same origin. The thing and the quality of the thing were alike expressed by the noun, and the analogic feeling in man suggested that they should be represented when together by nouns of the same class—i.e. with the same terminations: hence we have the grammatical property called gender, which is alto- gether independent of natural gender. Secondly, they rejected the article in their grammar, not having it in their speech. Here they were historically right, for the Greek article was only a pronoun. Later Latin developed a new one out of a different pronoun, ille, seen in various forms in the different Romance languages—French, Italian, Spanish, &c. But having lost the article they felt bound to fill up its place: therefore they put in the interjection, which is the conventional stereotyped expression of the natural cries which, we may believe, in days before articulate speech existed, eked out the earliest and simplest means of communication—i.e. gestures (see article PHILOLOGY). The interjection is therefore no 'part of speech;' it is an imperfect undeveloped 'speech-whole;' and the Greeks rightly did not include it in their list.

If we exclude the interjection, we can prove by means of historical grammar that these different parts of speech run back to two, the noun and the verb; and the distinction even of these rests on the inability of our analysis to separate them completely. It is true that nouns are distinguished by 'case-suffixes'—lupus, lupum, lupi, lupō, &c. in Latin; and the verb by 'personal suffixes'—amo, amas, amat, &c.; but there was doubtless a time in our parent-speech when no such 'suffixes' existed, and all that lies behind them may have been in those earlier days identical for noun and verb. Our own language shows the possibility of using one form—e.g. 'love,' alike for noun and verb. The pronoun differs from the noun in meaning by its greater generality. 'This' includes all objects in our immediate neighbourhood, books, chairs, tables, &c.; 'he' includes all 'Johns,' 'Smiths,' &c. In form it differs only by the simpler and on the whole the more archaic character of its root or ultimate element. The term 'pronoun' expresses only one subordinate use—the anaphoric or 'reference' use, by virtue of which, having once uttered a man's differentiating name, 'John,' or the like, we refer to him afterwards, so long as clearness permits, only as 'he.' The origin of adverbs and prepositions out of nouns or pronouns is very obvious in our own language: 'once' is Old Eng. ānes, the genitive of ān ('one'); 'seldom' is an old dative plural of seld ('rare'); to go 'afoot' was to go 'on' foot; 'beside' is 'by side (of)'; and, if we are unable to reach the original form of prepositions like 'on' and 'by,' we do not doubt that in days beyond our analysis they were nouns modifying other words which then filled the place of the nouns and verbs of later times. Similarly, conjunctions are either noun-cases or condensed sentences; 'whil-s-t' is 'whiles,' the genitive of 'while' (time), with a final t, which may be analogous to that of 'lest' (another conjunction), originally 'thi less the,' then 'lesthe,' and 'leste'; 'howbeit,' 'because' (by cause of) explain themselves. Thus the eight parts of speech may be traced back to not more than two.

All language at all times of which we have any knowledge, and doubtless from the very beginning of human speech, is a modification of existing combinations of sound. Language probably began, as has been already suggested, with the use of cries to help out gestures. These cries were associated by use with particular ideas, and that most elementary language (or languages, for there is no need to suppose that language sprang up in one place only, the circumstances being everywhere similar) was subject to the same laws which mould our speech at the present day. Groups of sound expressing the required thought are combined together, as 'man' and 'kind,' or 'house' and 'top.' The combination may be such that the different parts are always separable; then each sound-group (or word, as we may now call it) remains intact, and the relation which one word bears to another in the expression of the entire thought depends on the position of the words, the stress, or the pitch of the voice with which each is pronounced, or other more minute conditions. A language of which this is the prevailing character is called 'isolating,' and Chinese is the best-known type. It seems inadequate, yet the facility with which ideas can be expressed in such a language may be seen from the different grammatical values which the same sound-group can have in our own language in phrases like 'love is sweet,' 'we feel love,' 'God is love,' 'I love you,' &c.

But nearly all languages admit of combination more complete than this, whereby two or more words can be joined together, so that a single sound-complex expresses two or more ideas in combination—e.g. 'free-man,' 'black-bird,' 'thank-ful,' 'high-born,' 'back-bite,' 'ill-treat,' &c. Each of these may form the model for numerous copies; thus, 'thankful' can produce 'youthful,' 'healthful,' which are later English compounds. Then came hybrid compounds, where the first member is of Latin origin (of course through the Norman), as 'merciful,' 'masterful.' In this last we see that the exact nature of the original compound is obscured, and that 'ful' gives merely the additional sense of 'like,' as though the compound had been 'masterlike,' which does indeed occur in a briefer form, and with a secondary sense, as 'masterly.' This example throws light on the history of all word-formation. A word may cease to be felt as a compound commonly through change of form in one or both of its parts, as 'masterly,' where the idea of the skill of a master in some art alone remains; or 'hussy' (house-wife), where both parts of the compound are lost. Sometimes only one syllable may remain, as in 'lord' (loafward). Often some great change of idea joins with phonetic change in obscuring the nature of a compound, as in fortnight (fourteen-night). Now, when the last part of the compound fulfils certain conditions, it may be used in the formation of countless other words: -lic (like), which is found in O. E. in 'earth-lic,' 'cyne-lic' (earthly, kingly), passes on in its simpler form -ly in 'daily,' 'princely,' &c.; and -ly is then what grammarians call a suffix, an element which cannot be used alone, but can be added on at pleasure to another word to modify its meaning. The conditions are (1) that the form of the so-called suffix must be a convenient one phonetically; (2) that it must have been in use in a considerable number of compounds at the same time: for 'bridegroom' (bride-man), 'nightingale' (nihte-gale, night-singer), 'gossip' (God-sib, God-related) have produced no analogous forms in English owing to the rarity of the use of their second member; (3) that the last member must be general in its sense, or at least acquire some general sense in composition. A suffix is especially favoured which can be mentally referred to some common word of general sense, though it may really have nothing to do with that word. Thus, in 'credible,' 'invincible,' &c. the original suffix -ble (-bili in Latin) is seen; but in many words which come to us through the French, 'probable,' 'amiable,' a preceded the last syllable: thus these words seemed to mean 'able' to be proved, or to be loved; and so words like 'knowable,' 'lovable,' 'reliable' sprang up in abundance. Independently of these conditions of the origin of suffixes, it is also necessary that the first member of a compound remain unobscured. Thus, no words have been formed on the model of 'orchard' (wort-yard), though -ard as a Norman-French suffix has produced derivatives like 'drunkard,' on the analogy of 'bastard,' 'wizard.'

We are justified in inferring from the English suffixes which can be explained as remnants of words (-ful, -ly, -dom, -hood, and the like) that the others whose history can no longer be traced had a similar origin; and even in extending this principle to those formative suffixes which reach back to the earliest period of language. It is a sound axiom that what is in language has been and will be; it is only by dealing with spoken languages that we can infer the nature of those known to us by tradition only. It cannot be said with certainty that we should assign the same origin to those other suffixes—which we call inflectional—to which we owe the cases of our nouns, and the persons, tenses, and voices of our verbs. The persons, indeed, of the verbs were, it is most probable, pronouns. The m in 'am' represents original 'I,' so 'am' meant 'exist I,' and was a compound of two words, originally as separate as 'I exist;' s represented 'thou,' and t (Eng. th in 'loveth,' &c.) was 'he.' But we cannot say exactly what the tense-suffixes were, though we believe they are the remnants of words; nor what were the case-suffixes of the nouns—what, for example, was the s which still marks our genitive case, or the s of our plurals. But we know that we can make a 'noun of multitude' by making such a compound as 'man-kind,' and there is no reason why -es (the original form of -s, our plural suffix) may not once have been some such word as 'kind,' and compounded in the same manner. Such a history is in accordance with all we know of the processes of language.

It will be apparent from what has been said that there never was in any language some one period in which its suffixes were made, succeeded by a period in which there was no more growth but only decay. Formation is always going on, though more slowly in languages which are stereotyped by literature. In English we have almost ceased to use our second personal suffix -st, in 'lovest,' &c. But that st is itself an English growth: the older English form was s: in the old Mercian Psalter (edited by Mr Sweet in his Oldest English Texts) we find both 'thu dydes' and 'thu dydest,' 'thu bis' and 'thu bist,' &c. Other Teutonic languages show the same (independent) development. Still more do 'formative suffixes' go on growing. One of our commoner English suffixes (used to make a diminutive) is -let, seen in comparatively recent words, like 'brooklet,' 'streamlet,' &c. But the form is really a development of the older -et (the French -ette) in 'helmet,' 'banneret,' 'cygnet.' Several of these forms, like 'islet,' 'circlet,' and 'eaglet,' were formed out of nouns which ended in l; and so new ones were formed—'ring-let,' &c., as though the l had always belonged to the suffix. We are getting a new suffix in -nist, seen in 'tobacconist,' &c. This is an extension of the old suffix (Greek, through Latin into French) -ist, in 'jurist,' 'dentist,' &c.; this seems to be due to words where the n belongs to the root-part, as 'mechan-ist,' 'pian-ist,' and other late forms.

A common method of inflection in language is, not by suffix, but by change of the original vowel: thus, we have 'man,' but plural 'men;' and in verbs we find present 'drink,' preterite 'drank,' past participle 'drunken.' These can, however, be traced to the influence in different ways of lost suffixes. Thus, the old declension of 'man' was nom. 'mann;' gen. 'mannes;' dat. 'menn(i)'; plur. nom. 'menn(i)'; gen. 'manna;' dat. 'mannum.' It is clear that the change of a to e had at first nothing to do with the plural, for it is found in singular and plural alike when i followed: this vowel had the property of modifying a in a preceding syllable to e. But when the cases were lost, as happened in English mainly through Norman influence, 'man' remained as the only singular form, and 'men' as the only plural; so, for grammatical purposes, the plural might truly be said to be made by changing a to e. Similar is the history of 'mouse,' plur. 'mice;' 'goose,' plur. 'geese,' &c. The verb-change, i, a, u, has a most symmetrical look, and seems as though it must have been devised to express the change of relation. As a fact, however, in this and all similar cases, i and a represent in all Germanic languages original e and o; and these two vowels probably represent developments of a minute variation in pitch-accent (e being higher than o), dating from beyond the historic period of the parent-speech; and this variation marks indeed tense distinctions—e.g. in Greek, pres. dérkomai, perf. dédorka; but it is also found in nouns such as génos, gónos, and it seems to have had nothing to do with tenses at first. The second change, that in 'drank,' 'drnken,' has quite a different origin, but one equally removed from tense-formation. Like the first variation, it represents a very ancient change—due to the fact that in the parent language the syllables immediately preceding or following that which bore the stress-accent were weakened: no language shows better than English how to slur a syllable immediately preceding or following a stressed one—e.g. in 'alone' (where the last syllable is stressed) the a, originally the full a of 'all,' is sounded like the u of 'but,' or the o of 'son;' the same sound is commonly heard—e.g. in such a word as 'liberty,' instead of the ev of the middle syllable, the stress being on the first. Now in the past participle the stress was on the suffix -no (seen as -en in 'drunken'), and hence the vowel-change in the root. But it oddly happens that just the same change took place in the plural of the perfect itself, owing to the plural personal suffixes being stressed in the parent language; and so the Old English singular third person was 'drank,' but the third plural was 'druncon' (a precisely parallel case is the Greek sing. oida, plur. idmen, orig. idmen). So there was a time when it was right to say 'I drank' and 'we drunk;' but a meaningless distinction like this could not be maintained: one form was bound to supplant the other, and 'drank' won; but 'won,' the plur. of 'winnan,' supplanted the sing. 'wann;' 'stung' beat 'stang;' 'sprang' and 'sprung' were used indifferently at the beginning of this century, as by Scott and Byron, to help their rhymes; and here and in other verbs there is still some fluctuation of use, even among educated men. These examples may suffice to show that vowel-change, though extremely useful to mark grammatical distinctions, was not in any way designed for this end, which has been reached by unconscious differentiation: for we may infer from what we can observe in languages whose history can be traced that the prehistoric distinctions in the earliest recorded languages had a like accidental origin.

The history of grammatical forms may then be roughly sketched thus. They arose probably always from composition. Such compounds were subject to phonetic corruption, and the unstressed syllables were slurred and lost their individuality; or one member of the compound ceased to be used independently, some other word having superseded it, the result being the same as in the first case—viz. the loss of special significance in one part of the compound; and when the part so generalised is the final syllable, that syllable becomes a mere suffix, and can express relation, as the -ly in 'fatherly,' or the -s in 'fathers.' Furthermore, the cases of the nouns and the persons of the verbs thus formed were liable to variations of form in the same noun or verb, due to the incidence of stress or the influence of one syllable on another. The irregularities thus produced were again levelled in process of time by the natural tendency to do away with differences which are no longer significant; hence came symmetry of inflection, which is not the earliest stage in grammar, but rather the result of long unconscious play of physical and mental forces. Again, inflections constantly perished, either by simple phonetic decay, or more commonly through change of nationality, as, for example, when the Tenthic and other races adopted the Latin of the conquered Roman provinces, or when the descendants of the Normans began to use the national speech of England. Thus arises much simplification of what is to the speakers a foreign grammar; also there is a great growth of hybrid forms, Norman-French words combining with English suffixes, and rice versá. With the dying out of inflections arises a great growth of indeclinable words—adverbs, conjunctions, and prepositions: some cases, as the locative or the ablative in Greek, or the instrumental in Latin, became almost extinct; the few surviving forms, as Greek locatives in -ei and ablatives in -ós, belonging to nouns of the o class, lost their connection with those nouns; they remained isolated forms, freed from the levelling tendencies which affected the other cases of the same noun, because no longer felt to be in connection with them. Thus they could become the origin each of a new group of forms, extending (as did the so-called Greek adverbs in -ei and -ós) to many other classes of nouns besides that which gave them birth. Very commonly this isolation of some particular form may arise while the case is still in full use, through some accidental break of connection. In English our one surviving case-form in the genitive is -s, yet this very form has been the parent of numerous adverbs: 'ānes' (already mentioned) was the genitive of 'ān' (one): the connection was lost, and the adverb 'once' arose, and produced 'twice' (older form 'twi-es'), 'thrice' by mere analogy, no such genitives having ever existed; so, too, 'forward-s,' 'alway-s,' and many others are analogical forms—no true genitives, but copies of the model set by an isolated genitive. It has been well said by one of the greatest of modern German philologists, Professor H. Paul, that isolation is the essential condition of all speech-development.

Lastly, even while cases survive in use, it is necessary to supplement them by prepositions, because (except perhaps in languages which, like the Finnish, have fifteen cases) there are not enough case-forms to express the numerous relations in space ('to,' 'from,' 'in,' 'upon,' 'by,' 'near,' 'with,' &c.) in which one person or thing may stand to another. As cases die out this need increases, and modern European languages express practically all relations by prepositions. This principle is sometimes called analysis, as contrasted with the combinatory 'synthetic' principle of older forms of languages. Naturally no language is ever completely analytic: even in English words like 'father's' and 'love's' still attest that the language was once synthetic.

Those who desire fuller insight into the principles of grammar (as seen in languages of the Indo-European type) may consult the well-known works of Prof. Max Müller; A. H. Sayce's Principles of Comparative Philology, and his Introduction to the Science of Language, which treat the subject from a different standpoint; W. D. Whitney's Life and Growth of Language, and his Linguistic Studies; H. Paul's Principien der Sprachgeschichte, an invaluable but difficult work, translated, though not made materially easier, by Prof. Strong. A synoptic view of the relation of the Indo-European languages will be found in the still unfinished Grundriss der vergleichenden Grammatik of Karl Brugmann (vol. i. trans. by Wright). Excellent works on special languages are Whitney's Sanskrit Grammar and Delbrück's Alt-indische Syntax; for Greek, may be mentioned (out of many) Brugmann's Grammar in J. Müller's Handbuch der klassischen Altertumswissenschaft, and D. B. Monro's Homeric Grammar—a most suggestive book; for Latin, Stolz's Grammar (also in Müller's Handbuch), and Dräger's Historical Latin Syntax, which, though old, is still the most systematic work on the subject; innumerable valuable articles bearing on both Greek and Latin are to be found in Kuhn's Zeitschrift, Bezzenberger's Beiträge, the Mémoires de la Société de Linguistique, the Cambridge and the American Journals of Philology; for Keltic, the Grammars of Zeuss and Windisch (Old Irish, trans. by Dr N. Moore), and Prof. Rhys's Lectures on Welsh Philology; for the Romance languages collectively, the Grammar of Diez and the (still unfinished) Grundriss der romanischen Philologie of G. Gröber; special works on these languages are too numerous to mention. For Teutonic languages there is an excellent series of grammars published by Niemeyer (Halle), on Icelandic by Noreen, on Old High German by Braune, and on Middle High German by Paul (Strong and K. Meyer's History of the German Language may also be found useful); on Gothic, Braune; see also Douse's Ulfilas, and Prof. Skeat's little edition of the Gospel of St Mark in Gothic. For English, Sweet's Anglo-Saxon Reader and Cook's trans. of Sievers' Grammar of Old English are the best; Prof. Skeat's Principles of English Etymology should also be consulted; Storm's Englische Philologie is excellent, but still a fragment; A. J. Ellis' Early English Pronunciation (5 vols. 1869-89) is a mine of information on the history of the English language.