FRENCH LANGUAGE. I. Geography. - French is the general name of the north-north-western group of Romanic dialects, the modern Latin of northern Gaul (carried by emigration to some places - as lower Canada - out of France). In a restricted sense it is that variety of the Parisian dialect which is spoken by the educated, and is the general literary language of France. The region in which the native language is termed French consists of the northern half of France (including Lorraine) and parts of Belgium and Switzerland; its boundaries on the west are the Atlantic Ocean and the Celtic dialects of Brittany; on the north-west and north, the English Channel; on the north-east and east the Teutonic dialects of Belgium, Germany and Switzerland. In the south-east and south the boundary is to a great extent conventional and ill-defined, there being originally no linguistic break between the southern French dialects and the northern Provençal dialects of southern France, north-western Italy and south-western Switzerland. It is formed partly by spaces of intermediate dialects (some of whose features are French, others Provençal), partly by spaces of mixed dialects resulting from the invasion of the space by more northern and more southern settlers, partly by lines where the intermediate dialects have been suppressed by more northern (French) and more southern (Provençal) dialects without these having mixed. Starting in the west at the mouth of the Gironde, the boundary runs nearly north soon after passing Bordeaux; a little north of Angoulême it turns to the east, and runs in this direction into Switzerland to the north of Geneva.
II. External History. - (a) Political. - By the Roman conquests the language of Rome was spread over the greater part of southern and western Europe, and gradually supplanted the native tongues. The language introduced was at first nearly uniform over the whole empire, Latin provincialisms and many more or less general features of the older vulgar language being suppressed by the preponderating influence of the educated speech of the capital. As legions became stationary, as colonies were formed, and as the natives adopted the language of their conquerors, this language split up into local dialects, the distinguishing features of which are due, as far as can be ascertained (except, to some extent, as to the vocabulary), not to speakers of different nationalities misspeaking Latin, each with the peculiarities of his native language, but to the fact that linguistic changes, which are ever occurring, are not perfectly uniform over a large area, however homogeneous the speakers. As Gaul was not conquered by Caesar till the middle of the first century before our era, its Latin cannot have begun to differ from that of Rome till after that date; but the artificial retention of classical Latin as the literary and official language after the popular spoken language had diverged from it, often renders the chronology of the earlier periods of the Romanic languages obscure. It is, however, certain that the popular Latin of Gaul had become differentiated from that of central Italy before the Teutonic conquest of Gaul, which was not completed till the latter half of the 5th century; the invaders gradually adopted the language of their more civilized subjects, which remained unaffected, except in its vocabulary. Probably by this time it had diverged so widely from the artificially preserved literary language that it could no longer be regarded merely as mispronounced Latin; the Latin documents of the next following centuries contain many clearly popular words and forms, and the literary and popular languages are distinguished as latina and romana. The term gallica, at first denoting the native Celtic language of Gaul, is found applied to its supplanter before the end of the 9th century, and survives in the Breton gallek, the regular term for "French." After the Franks in Gaul had abandoned their native Teutonic language, the term francisca, by which this was denoted, came to be applied to the Romanic one they adopted, and, under the form française, remains its native name to this day; but this name was confined to the Romanic of northern Gaul, which makes it probable that this, at the time of the adoption of the name francisca, had become distinct from the Romanic of southern Gaul. Francisca is the Teutonic adjective frankisk, which occurs in Old English in the form frencise; this word, with its umlauted e from a with following i, survives under the form French, which, though purely Teutonic in origin and form, has long been exclusively applied to the Romanic language and inhabitants of Gaul. The German name franzose, with its accent on, and o in, the second syllable, comes from françois, a native French form older than français, but later than the Early Old French franceis. The Scandinavian settlers on the north-west coast of France early in the 10th century quickly lost their native speech, which left no trace except in some contributions to the vocabulary of the language they adopted. The main feature since is the growth of the political supremacy of Paris, carrying with it that of its dialect; in 1539 Francis I. ordered that all public documents should be in French (of Paris), which then became the official language of the whole kingdom, though it is still foreign to nearly half its population.
The conquest of England in 1066 by William, duke of Normandy, introduced into England, as the language of the rulers and (for a time) most of the writers, the dialects spoken in Normandy (see also Anglo-Norman Literature). Confined in their native country to definite areas, these dialects, following their speakers, became mixed in England, so that their forms were used to some extent indifferently; and the constant communication with Normandy maintained during several reigns introduced also later forms of continental Norman. As the conquerors learned the language of the conquered, and as the more cultured of the latter learned that of the former, the Norman of England (including that of the English-speaking Lowlands of Scotland) became anglicized; instead of following the changes of the Norman of France, it followed those of English. The accession in 1154 of Henry II. of Anjou disturbed the Norman character of Anglo-French, and the loss of Normandy under John in 1204 gave full play to the literary importance of the French of Paris, many of whose forms afterwards penetrated to England. At the same time English, with a large French addition to its vocabulary, was steadily recovering its supremacy, and is officially employed (for the first time since the Conquest) in the Proclamation of Henry III., 1258. The semi-artificial result of this mixture of French of different dialects and of different periods, more or less anglicized according to the date or education of the speaker or writer, is generally termed "the Anglo-Norman dialect"; but the term is misleading for a great part of its existence, because while the French of Normandy was not a single dialect, the later French of England came from other French provinces besides Normandy, and being to a considerable extent in artificial conditions, was checked in the natural development implied by the term "dialect." The disuse of Anglo-French as a natural language is evidenced by English being substituted for it in legal proceedings in 1362, and in schools in 1387; but law reports were written in it up to about 1600, and, converted into modern literary French, it remains in official use for giving the royal assent to bills of parliament.
(b) Literary. - Doubtless because the popular Latin of northern Gaul changed more rapidly than that of any other part of the empire, French was, of all the Romanic dialects, the first to be recognized as a distinct language, and the first to be used in literature; and though the oldest specimen now extant is probably not the first, it is considerably earlier than any existing documents of the allied languages. In 813 the council of Tours ordered certain homilies to be translated into Rustic Roman or into German; and in 842 Louis the German, Charles the Bald, and their armies confirmed their engagements by taking oaths in both languages at Strassburg. These have been preserved to us by the historian Nithard (who died in 853); and though, in consequence of the only existing manuscript (at Paris) being more than a century later than the time of the author, certain alterations have occurred in the text of the French oaths, they present more archaic forms (probably of North-Eastern French) than any other document. The next memorials are a short poem, probably North-Eastern, on St Eulalia, preserved in a manuscript of the 10th century at Valenciennes, and some autograph fragments (also at Valenciennes) of a homily on the prophet Jonah, in mixed Latin and Eastern French, of the same period. To the same century belong a poem on Christ's Passion, apparently in a mixed (not intermediate) language of French and Provençal, and one, probably in South-Eastern French, on St Leger; both are preserved, in different handwritings, in a MS. at Clermont-Ferrand, whose scribes have introduced many Provençal forms. After the middle of the 11th century literary remains are comparatively numerous; the chief early representative of the main dialects are the following, some of them preserved in several MSS., the earliest of which, however (the only ones here mentioned), are in several cases a generation or two later than the works themselves. In Western French are a verse life of St Alexius (Alexis), probably Norman, in an Anglo-Norman MS. at Hildesheim; the epic poem of Roland, possibly also Norman, in an A.-N. MS. at Oxford; a Norman verbal translation of the Psalms, in an A.-N. MS. also at Oxford; another later one, from a different Latin version, in an A.-N. MS. at Cambridge; a Norman translation of the Four Books of Kings, in a probably A.-N. MS. at Paris. The earliest work in the Parisian dialect is probably the Travels of Charlemagne, preserved in a late Anglo-Norman MS. with much altered forms. In Eastern French, of rather later date, there are translations of the Dialogues of Pope Gregory, in a MS. at Paris, containing also fragments of Gregory's Moralities, and (still later) of some Sermons of St Bernard, in a MS. also in Paris. From the end of the 12th century literary and official documents, often including local charters, abound in almost every dialect, until the growing influence of Paris caused its language to supersede in writing the other local ones. This influence, occasionally apparent about the end of the 12th century, was overpowering in the 15th, when authors, though often displaying provincialisms, almost all wrote in the dialect of the capital; the last dialect to lose its literary independence was the North-Eastern, which, being the Romanic language of Flanders, had a political life of its own, and (modified by Parisian) was used in literature after 1400.
III. Internal History. - Though much has been done in recent years, in the scientific investigation of the sounds, inflexions, and syntax of the older stages and dialects of French, much still remains to be done, and it must suffice here to give a sketch, mainly of the dialects which were imported into England by the Normans - in which English readers will probably take most interest, and especially of the features which explain the forms of English words of French origin. Dates and places are only approximations, and many statements are liable to be modified by further researches. The primitive Latin forms given are often not classical Latin words, but derivatives from these; and reference is generally made to the Middle English (Chaucerian) pronunciation of English words, not the modern.
(a) Vocabulary. - The fundamental part of the vocabulary of French is the Latin imported into Gaul, the French words being simply the Latin words themselves, with the natural changes undergone by all living speech, or derivatives formed at various dates. Comparatively few words were introduced from the Celtic language of the native inhabitants (bec, lieue from the Celtic words given by Latin writers as beccus, leuca), but the number adopted from the language of the Teutonic conquerors of Gaul is large (guerre = werra; laid = laidh; choisir = kausjan). The words were imported at different periods of the Teutonic supremacy, and consequently show chronological differences in their sounds (haïr = hatan; français = frankisk; écrevisse = krebiz; échine = skina). Small separate importations of Teutonic words resulted from the Scandinavian settlement in France, and the commercial intercourse with the Low German nations on the North Sea (friper = Norse hripa; chaloupe = Dutch sloop; est = Old English eást). In the meantime, as Latin (with considerable alterations in pronunciation, vocabulary, etc.) continued in literary, official and ecclesiastical use, the popular language borrowed from time to time various more or less altered classical Latin words; and when the popular language came to be used in literature, especially in that of the church, these importations largely increased (virginitet Eulalia = virginitatem; imagena Alexis = imaginem - the popular forms would probably have been vergedet, emain). At the Renaissance they became very abundant, and have continued since, stifling to some extent the developmental power of the language. Imported words, whether Teutonic, classical Latin or other, often receive some modification at their importation, and always take part in all subsequent natural phonetic changes in the language (Early Old French adversarie, Modern French adversaire) . Those French words which appear to contradict the phonetic laws were mostly introduced into the language after the taking place (in words already existing in the language) of the changes formulated by the laws in question; compare the late imported laïque with the inherited lai, both from Latin laicum. In this and many other cases the language possesses two forms of the same Latin word, one descended from it, the other borrowed (meuble and mobile from mōbilem). Some Oriental and other foreign words were brought in by the crusaders (amiral from amir); in the 16th century, wars, royal marriages and literature caused a large number of Italian words (soldat = soldato; brave = bravo; caresser = carezzare) to be introduced, and many Spanish ones (alcôve = alcoba; hâbler = hablar). A few words have been furnished by Provençal (abeille, cadenas), and several have been adopted from other dialects into the French of Paris (esquiver Norman or Picard for the Paris-French eschiver). German has contributed a few (blocus = blochus; choucroute = surkrut); and recently a considerable number have been imported from England (drain, confortable, flirter). In Old French, new words are freely formed by derivation, and to a less extent by composition; in Modern French, borrowing from Latin or other foreign languages is the more usual course. Of the French words now obsolete some have disappeared because the things they express are obsolete; others have been replaced by words of native formation, and many have been superseded by foreign words generally of literary origin; of those which survive, many have undergone considerable alterations in meaning. A large number of Old French words and meanings, now extinct in the language of Paris, were introduced into English after the Norman Conquest; and though some have perished, many have survived - strife from Old French estrif (Teutonic strit); quaint from cointe (cognitum); remember from remembrer (rememorare); chaplet (garland) from chapelet (Modern French "chaplet of beads"); appointment (rendezvous) from appointement (now "salary"). Many also survive in other French dialects.
(b) Dialects. - The history of the French language from the period of its earliest extant literary memorials is that of the dialects composing it. But as the popular notion of a dialect as the speech of a definite area, possessing certain peculiarities confined to and extending throughout that area, is far from correct, it will be advisable to drop the misleading divisions into "Norman dialect," "Picard dialect" and the like, and take instead each important feature in the chronological order (as far as can be ascertained) of its development, pointing out roughly the area in which it exists, and its present state. The local terms used are intentionally vague, and it does not, for instance, at all follow that because "Eastern" and "Western" are used to denote the localities of more than one dialectal feature, the boundary line between the two divisions is the same in each case. It is, indeed, because dialectal differences as they arise do not follow the same boundary lines (much less the political divisions of provinces), but cross one another to any extent, that to speak of the dialect of a large area as an individual whole, unless that area is cut off by physical or alien linguistic boundaries, creates only confusion. Thus the Central French of Paris, the ancestor of classical Modern French, agrees with a more southern form of Romanic (Limousin, Auvergne, Forez, Lyonnais, Dauphiné) in having ts, not tsh, for Latin k (c) before i and e; tsh, not k, for k (c) before a; and with the whole South in having gu, not w, for Teutonic w; while it belongs to the East in having oi for earlier ei; and to the West in having é, not ei, for Latin a; and i, not ei, from Latin ĕ + i. It may be well to denote that Southern French does not correspond to southern France, whose native language is Provençal. "Modern French" means ordinary educated Parisian French.
(e) Phonology. - The history of the sounds of a language is, to a considerable extent, that of its inflections, which, no less than the body of a word, are composed of sounds. This fact, and the fact that unconscious changes are much more reducible to law than conscious ones, render the phonology of a language by far the surest and widest foundation for its dialectology, the importance of the sound-changes in this respect depending, not on their prominence, but on the earliness of their date. For several centuries after the divergence between spoken and written Latin, the history of these changes has to be determined mainly by reasoning, aided by a little direct evidence in the misspellings of inscriptions the semi-popular forms in glossaries, and the warnings of Latin grammarians against vulgarities. With the rise of Romanic literature the materials for tracing the changes become abundant, though as they do not give us the sounds themselves, but only their written representations, much difficulty, and some uncertainty, often attach to deciphering the evidence. Fortunately, early Romanic orthography, that of Old French included (for which see next section), was phonetic, as Italian orthography still is; the alphabet was imperfect, as many new sounds had to be represented which were not provided for in the Roman alphabet from which it arose, but writers aimed at representing the sounds they uttered, not at using a fixed combination of letters for each word, however they pronounced it.
The characteristics of French as distinguished from the allied languages and from Latin, and the relations of its sounds, inflections and syntax to those of the last-named language, belong to the general subject of the Romanic languages. It will be well, however, to mention here some of the features in which it agrees with the closely related Provençal, and some in which it differs. As to the latter, it has already been pointed out that the two languages glide insensibly into one another, there being a belt of dialects which possess some of the features of each. French and Provençal of the 10th century - the earliest date at which documents exist in both - agree to a great extent in the treatment of Latin final consonants and the vowels preceding them, a matter of great importance for inflections (numerous French examples occur in this section), (1) They reject all vowels, except a, of Latin final (unaccented) syllables, unless preceded by certain consonant combinations or followed by nt (here, as elsewhere, certain exceptions cannot be noticed); (2) they do not reject a similarly situated; (3) they reject final (unaccented) m; (4) they retain final s. French and Northern Provençal also agree in changing Latin ü from a labio-guttural to a labio-palatal vowel; the modern sound (German ü) of the accented vowel of French lune, Provençal luna, contrasting with that in Italian and Spanish luna, appears to have existed before the earliest extant documents. The final vowel laws generally apply to the unaccented vowel preceding the accented syllable, if it is preceded by another syllable, and followed by a single consonant - matin (matutinum), dortoir (dormitōrium), with vowel dropped; canevas (cannabaceum), armedure, later armëure, now armure (armaturam), with e = ǝ, as explained below.
On the other hand, French differs from Provençal: (1) in uniformly preserving (in Early Old French) Latin final t, which is generally rejected in Provençal - French aimet (Latin amat), Provençal ama; aiment (amant), Prov. aman; (2) in always rejecting, absorbing or consonantizing the vowel of the last syllable but one, if unaccented; in such words as angele (often spelt angle), the e after the g only serves to show its soft sound - French veintre (now vaincre, Latin vincere), Prov. vencer, with accent on first syllable; French esclandre (scandalum), Prov. escandol; French olie (dissyllabic, i = y consonant, now huile), Prov. oli (oleum); (3) in changing accented a not in position into ai before nasals and gutturals and not after a palatal, and elsewhere into é (West French) or ei (East French), which develops an i before it when preceded by a palatal - French main (Latin manum), Prov. man; aigre (acrem), agre; ele (alam), East French eile, Prov. ala; meitié (medietatem), East French moitieit, Prov. meitat; (4) in changing a in unaccented final syllables into the vowel ǝ, intermediate to a and e; this vowel is written a in one or two of the older documents, elsewhere e - French aime (Latin ama), Prov. ama; aimes (amas), Prov. amas; aimet (amat), Prov. ama; (5) in changing original au into ò - French or (aurum), Prov. aur; rober (Teutonic raubōn), Prov. raubar; (6) in changing general Romanic é, from accented ē and ĭ not in position, into ei - French veine (vēnam), Prov. vena; peil (pilum), Prov. pel.
As some of the dialectal differences were in existence at the date of the earliest extant documents, and as the existing materials, till the latter half of the 11th century, are scanty and of uncertain locality, the chronological order (here adopted) of the earlier sound-changes is only tentative.
(1) Northern French has tsh (written c or ch) for Latin k (c) and t before palatal vowels, where Central and Southern French have ts (written c or z) - North Norman and Picard chire (cēram), brach (brachium), plache (plateam); Parisian, South Norman, etc., cire, braz, place. Before the close of the Early Old French period (12th century) ts loses its initial consonant, and the same happened to tsh a century or two later; with this change the old distinction is maintained - Modern Guernsey and Picard chire, Modern Picard plache (in ordinary Modern French spelling); usual French cire, place. English, having borrowed from North and South Norman (and later Parisian), has instances of both tsh and s, the former in comparatively small number - chisel (Modern French ciseau = (?) caesellum), escutcheon (écusson, scutiōnem); city (cité, civitatem), place. (2) Initial Teutonic w is retained in the north-east and along the north coast; elsewhere, as in the other Romance languages, g was prefixed - Picard, etc., warde (Teutonic warda), werre (werra); Parisian, etc., guarde, guerre. In the 12th century the u or w of gu dropped, giving the Modern French garde, guerre (with gu = g); w remains in Picard and Walloon, but in North Normandy it becomes v - Modern Guernsey vâson, Walloon wazon, Modern French gazon (Teutonic wason). English has both forms, sometimes in words originally the same - wage and gage (Modern French gage, Teutonic wadi); warden and guardian (gardien, warding). (3) Latin b after accented a in the imperfect of the first conjugation, which becomes v in Eastern French, in Western French further changes to w, and forms the diphthong ou with the preceding vowel - Norman amowe (amabam), portout (portabat); Burgundian ameve, portevet. -eve is still retained in some places, but generally the imperfect of the first conjugation is assimilated to that of the others - amoit, like avoit (habēbat). (4) The palatalization of every then existing k and g (hard) when followed by a, i or e, after having caused the development of i before the e (East French ei) derived from a not in position, is abandoned in the north, the consonants returning to ordinary k or g, while in the centre and south they are assibilated to tsh or dzh - North Norman and Picard cachier (captiare), kier (carum), cose (causam), eskiver (Teutonic skiuhan), wiket (Teutonic wik + ittum), gal (gallum), gardin (from Teutonic gard); South Norman and Parisian chacier, chier, chose, eschiver, guichet, jal, jardin. Probably in the 14th century the initial consonant of tsh, dzh disappeared, giving the modern French chasser, jardin with ch = sh and j = zh; but tsh is retained in Walloon, and dzh in Lorraine. The Northern forms survive - Modern Guernsey cachier, gardìn; Picard cacher, gardin. English possesses numerous examples of both forms, sometimes in related words - catch and chase; wicket, eschew; garden, jaundice (jaunisse, from galbanum). (5) For Latin accented a not in position Western French usually has é, Eastern French ei, both of which take an i before them when a palatal precedes - Norman and Parisian per (parem), oiez (audiatis); Lorraine peir, oieis. In the 17th and 18th centuries close é changed to open è, except when final or before a silent consonant - amer (amarum) now having è, aimer (amare) retaining é. English shows the Western close é - peer (Modern French pair, Old French per), chief (chef, caput); Middle High German the Eastern ei - lameir (Modern French l'amer, l'aimer, la mer = Latin mare). (6) Latin accented e not in position, when it came to be followed in Old French by i unites with this to form i in the Western dialects, while the Eastern have the diphthongs ei - Picard, Norman and Parisian pire (pejor), piz (pectus); Burgundian peire, peiz. The distinction is still preserved - Modern French pire, pis; Modern Burgundian peire, pei. English words show always i - price (prix, pretium) spite (dépit, dēspectum). (7) The nasalization of vowels followed by a nasal consonant did not take place simultaneously with all the vowels. A and e before ṇ (guttural n, as in sing), ñ (palatal n), n and m were nasal in the 11th century, such words as tant (tantum) and gent (gentem) forming in the Alexis assonances to themselves, distinct from the assonances with a and e before non-nasal consonants. In the Roland umbre (ombre, umbram) and culchet (couche, collocat), fier (ferum) and chiens (canēs), dit (dictum) and vint (vēnit), ceinte (ciṇctam) and veie (voie, viam), brun (Teutonic brun) and fut (fuit) assonate freely, though o (u) before nasals shows a tendency to separation. The nasalization of i and u (= Modern French u) did not take place till the 16th century; and in all cases the loss of the following nasal consonant is quite modern, the older pronunciation of tant, ombre being tãnt, õmbrǝ, not as now tã, õbrh. The nasalization took place whether the nasal consonant was or was not followed by a vowel, femme (fēminam), honneur (honōrem) being pronounced with nasal vowels m the first syllable till after the 16th century, as indicated by the doubling of the nasal consonant in the spelling and by the phonetic change (in femme and other words) next to be mentioned. English generally has au (now often reduced to a) for Old French ã - vaunt (vanter, vanitare), tawny (tanné (?) Celtic). (8) The assimilation of ē (nasal e) to ã (nasal a) did not begin till the middle of the 11th century, and is not yet universal, in France, though generally a century later. In the Alexis nasal a (as in tant) is never confounded with nasal e (as in gent) in the assonances, though the copyist (a century later) often writes a for nasal e in unaccented syllables, as in amfant (enfant, infantem); in the Roland there are several cases of mixture in the assonances, gent, for instance, occurring in ant stanzas, tant in ent ones. English has several words with a for e before nasals - rank (rang, Old French renc, Teutonic hriṇga), pansy (pensée, pēnsatam); but the majority show e - enter (entrer, intrare), fleam (flamme, Old French fleme, phlebotomum). The distinction is still preserved in the Norman of Guernsey, where an and en, though both nasal, have different sounds - lànchier (lancer, laṇceare), but mèntrie (Old French menterie, from mentiri). (9) The loss of s, or rather z, before voiced consonants began early, s being often omitted or wrongly inserted in 12th century MSS. - Earliest Old French masle (masculum), sisdre (siceram); Modern French mâle, cidre. In English it has everywhere disappeared - male, cider; except in two words, where it appears, as occasionally in Old French, as d - meddle (mêler, misculare), medlar (néflier, Old French also meslier, mespilarium). The loss of s before voiceless consonants (except f) is about two centuries later, and it is not universal even in Parisian - Early Old French feste (festam), escuier (scutarium); Modern French fête, écuyer, but espérer (spērare). In the north-east s before t is still retained - Walloon chestai (château, castellum), fiess (fête). English shows s regularly - feast, esquire. (10) Medial dh (soft th, as in then), and final th from Latin t or d between vowels, do not begin to disappear till the latter half of the 11th century. In native French MSS. dh is generally written d, and th written t; but the German scribe of the Oaths writes adjudha (adjutam), cadhuna (Greek katá and unam); and the English one of the Alexis cuntretha (contratam), lothet (laudatum), and that of the Cambridge Psalter heriteth (hērēditatem). Medial dh often drops even in the last-named MSS., and soon disappears; the same is true for final th in Western French - Modern French contrée, loué. But in Eastern French final th, to which Latin t between vowels had probably been reduced through d and dh, appears in the 12th century and later as t, rhyming on ordinary French final t - Picard and Burgundian pechiet (peccatum) apeleit (appellatum). In Western French some final ths were saved by being changed to f - Modern French soif (sitim), mœuf (obsolete, modum). English has one or two instances of final th, none of medial dh - faith (foi, fidem); Middle English carite (charité, caritatem), drut (Old French dru, Teutonic drud); generally the consonant is lost - country, charity. Middle High German shows the Eastern French final consonant - moraliteit (moralité, mōralitatem). (11) T from Latin final t, if in an Old French unaccented syllable, begins to disappear in the Roland, where sometimes aimet (amat), sometimes aime, is required by the metre, and soon drops in all dialects. The Modern French t of aime-t-il and similar forms is an analogical insertion from such forms as dort-il (dormit), where the t has always existed. (12) The change of the diphthong ai to èi and afterwards to èè (the doubling indicates length) had not taken place in the earliest French documents, words with ai assonating only on words with a; in the Roland such assonances occur, but those of ai on è are more frequent - faire (facere) assonating on parastre (patraster) and on estes (estis); and the MS. (half a century later than the poem) occasionally has ei and e for ai - recleimet (reclamat), desfere (disfacere), the latter agreeing with the Modern French sound. Before nasals (as in laine = lanam) and ié (as in payé = pacatum), ai remained a diphthong up to the 16th century, being apparently ei, whose fate in this situation it has followed. English shows ai regularly before nasals and when final, and in a few other words - vain (vain, vanum), pay (payer, pacare), wait (guetter, Teutonic wahtēn); but before most consonants it has usually èè - peace (pais, pacum), feat (fait, factum). (13) The loss or transposition of i (= y-consonant) following the consonant ending an accented syllable begins in the 12th century - Early Old French glorie (glōriam), estudie (studium), olie (oleum); Modern French gloire, étude, huile. English sometimes shows the earlier form - glory, study; sometimes the later - dower (douaire, Early Old French doarie, dōtarium), oil (huile). (14) The vocalization of l preceded by a vowel and followed by a consonant becomes frequent at the end of the 12th century; when preceded by open è, an a developed before the l while this was a consonant - 11th century salse (salsa), beltet (bellitatem), solder (solidare); Modern French sauce, beauté, souder. In Parisian, final èl followed the fate of èl before a consonant, becoming the triphthong èau, but in Norman the vocalization did not take place, and the l was afterwards rejected - Modern French ruisseau, Modern Guernsey russé (rivicellum). English words of French origin sometimes show l before a consonant, but the general form is u - scald (échauder, excalidare), Walter (Gautier, Teutonic Waldhari); sauce, beauty, soder. Final èl is kept - veal (veau, vitellum), seal (sceau, sigillum). (15) In the east and centre éi changes to òi, while the older sound is retained in the north-west and west - Norman estreit (étroit, strictum), preie (proie, praedam), 12th century Picard, Parisian, etc., estroit, proie. But the earliest (10th century) specimens of the latter group of dialects have éi - pleier (ployer, plicare) Eulalia, mettreiet (mettrait, mittere habēbat) Jonah. Parisian òi, whether from ei or from Old French òi, ói, became in the 15th century uè (spellings with oue or oe are not uncommon - mirouer for miroir, miratōrium), and in the following, in certain words, è, now written ai - français, connaître, from françois (franceis, franciscum), conoistre (conuistre, cognōscere); where it did not undergo the latter change it is now ua or wa - roi (rei, rēgem), croix (cruis, crucem). Before nasals and palatal l, ei (now = è) was kept - veine (vēna), veille (vigila), and it everywhere survives unlabialized in Modern Norman - Guernsey ételle (étoile, stēlla) with é, ser (soir, sērum) with è. English shows generally ei (or ai) for original ei - strait (estreit), prey (preie); but in several words the later Parisian oi - coy (coi, qviētum), loyal (loyal, lēgalem). (16) The splitting of the vowel-sound from accented Latin ō or u not in position, represented in Old French by o and u indifferently, into u, o (before nasals), and eu (the latter at first a diphthong, now = German ö), is unknown to Western French till the 12th century, and is not general in the east. The sound in 11th century Norman was much nearer to u (Modern French ou) than to ó (Modern French ô), as the words borrowed by English show uu (at first written u, afterwards ou or ow), never óó; but was probably not quite u, as Modern Norman shows the same splitting of the sound as Parisian. Examples are - Early Old French espose or espuse (spōnsam), nom or num (nōmen), flor or flur (flōrem); Modern French épouse, nom, fleur; Modern Guernsey goule (gueule, gulam), nom, flleur. Modern Picard also shows u, which is the regular sound before r - flour; but Modern Burgundian often keeps the original Old French ó - vo (vous, vōs). English shows almost always uu - spouse, noun, flower (Early Middle English spuse, nun, flur); but nephew with éu (neveu, nepōtem). (17) The loss of the u (or w) of qu dates from the end of the 12th century - Old French quart (qvartum), quitier (qviētare) with qu = kw, Modern French quart, quitter with qu = k. In Walloon the w is preserved - couâr (quart), cuitter; as is the case in English - quart, quit. The w of gw seems to have been lost rather earlier, English having simple g - gage (gage, older guage, Teutonic wadi), guise (guise, Teutonic wisa). (18) The change of the diphthong òu to uu did not take place till after the 12th century, such words as Anjou (Andegavum) assonating in the Roland on fort (fortem); and did not occur in Picardy, where òu became au caus from older còus, còls (cous, collōs) coinciding with caus from calz (chauds, calidōs). English keeps òu distinct from uu - vault for vaut (Modern French voûte, volvitam), soder (souder, solidare). (19) The change of the diphthong ié to simple é is specially Anglo-Norman, in Old French of the Continent these sounds never rhyme, in that of England they constantly do, and English words show, with rare exceptions, the simple vowel - fierce (Old French fiers, ferus), chief (chief, caput), with ie = ee; but pannier (panier, panarium). At the beginning of the modern period, Parisian dropped the i of ie when preceded by ch or j - chef, abréger (Old French abregier, abbreviare); elsewhere (except in verbs) ie is retained - fier (ferum), pitié (pietatem). Modern Guernsey retains ie after ch - ap'rchier (approcher, adpropeare).(20) Some of the Modern French changes have found their places under older ones; those remaining to be noticed are so recent that English examples of the older forms are superfluous. In the 16th century the diphthong au changed to ao and then to ó, its present sound, rendering, for instance, maux (Old French mals, malōs) identical with mots (muttōs). The au of eau underwent the same change, but its e was still sounded as ǝ (the e of que); in the next century this was dropped, making veaux (Old French vëels, vitellōs) identical with vaux (vals, vallēs). (21) A more general and very important change began much earlier than the last; this is the loss of many final consonants. In Early Old French every consonant was pronounced as written; by degrees many of them disappeared when followed by another consonant, whether in the same word (in which case they were generally omitted in writing) or in a following one. This was the state of things in the 16th century; those final consonants which are usually silent in Modern French were still sounded, if before a vowel or at the end of a sentence or a line of poetry, but generally not elsewhere. Thus a large number of French words had two forms; the Old French fort appeared as fòr (though still written fort) before a consonant, fòrt elsewhere. At a later period final consonants were lost (with certain exceptions) when the word stood at the end of a sentence or of a line of poetry; but they are generally kept when followed by a word beginning with a vowel. (22) A still later change is the general loss of the vowel (written e) of unaccented final syllables; this vowel preserved in the 16th century the sound ǝ, which it had in Early Old French. In later Anglo-Norman final ǝ (like every other sound) was treated exactly as the same sound in Middle English; that is, it came to be omitted or retained at pleasure, and in the 15th century disappeared. In Old French the loss of final ǝ is confined to a few words and forms; the 10th century saveiet (sapēbat for sapiēbat) became in the 11th saveit, and ore (ad hōram), ele (illam) develop the abbreviated or, el. In the 15th century ǝ before a vowel generally disappears - mûr, Old French mëur (maturum); and in the 16th, though still written, ǝ after an unaccented vowel, and in the syllable ent after a vowel, does the same - vraiment, Old French vraiement (vēraca mente); avoient two syllables, as now (avaient), in Old French three syllables (as habēbant). These phenomena occur much earlier in the anglicized French of England - 13th century aveynt (Old French aveient). But the universal loss of final e, which has clipped a syllable from half the French vocabulary, did not take place till the 18th century, after the general loss of final consonants; fort and forte, distinguished at the end of a sentence or line in the 16th century as fòrt and fòrtǝ, remain distinguished, but as fòr and fòrt. The metre of poetry is still constructed on the obsolete pronunciation, which is even revived in singing; "dîtes, la jeune belle," actually four syllables (dit, la zhœn bèl), is considered as seven, fitted with music accordingly, and sung to fit the music (ditǝ, la zhœna bèlǝ). (23) In Old French, as in the other Romanic languages, the stress (force, accent) is on the syllable which was accented in Latin; compare the treatment of the accented and unaccented vowels in latrō amas, giving lére, áime, and in latrōnem, amatis, giving larón, améz, the accented vowels being those which rhyme or assonate. At present, stress in French is much less marked than in English, German or Italian, and is to a certain extent variable; which is partly the reason why most native French scholars find no difficulty in maintaining that the stress in living Modern French is on the same syllable as in Old French. The fact that stress in the French of to-day is independent of length (quantity) and pitch (tone) largely aids the confusion; for though the final and originally accented syllable (not counting the silent e as a syllable) is now generally pronounced with less force, it very often has a long vowel with raised pitch. In actual pronunciation the chief stress is usually on the first syllable (counting according to the sounds, not the spelling), but in many polysyllables it is on the last but one; thus in caution the accented (strong) syllable cau, in occasion it is ca. Poetry is still written according to the original place of the stress; the rhyme-syllables of larron, aimez are still ron and mez, which when set to music receive an accented (strong) note, and are sung accordingly, though in speech the la and ai generally have the principal stress. In reading poetry, as distinguished from singing, the modern pronunciation is used, both as to the loss of the final ǝ and the displacement of the stress, the result being that the theoretical metre in which the poetry is written disappears. (24) In certain cases accented vowels were lengthened in Old French, as before a lost s; this was indicated in the 16th century by a circumflex - bête, Old French beste (bestiam), âme, Old French anme (anima). The same occurred in the plural of many nouns, where a consonant was lost before the s of the flection; thus singular coc with short vowel, plural cos with long. The plural cos, though spelt cogs instead of cô (= kóó), is still sometimes to be heard, but, like other similar ones, is generally refashioned after the singular, becoming kòk. In present French, except where a difference of quality has resulted, as in côte (Old French coste, costam) with ò and cotte (Old French cote), with ò, short and long vowels generally run together, quantity being now variable and uncertain; but at the beginning of this century the Early Modern distinctions appear to have been generally preserved.
(d) Orthography. - The history of French spelling is based on that of French sounds; as already stated, the former (apart from a few Latinisms in the earliest documents) for several centuries faithfully followed the latter. When the popular Latin of Gaul was first written, its sounds were represented by the letters of the Roman alphabet; but these were employed, not in the values they had in the time of Caesar, but in those they had acquired in consequence of the phonetic changes that had meantime taken place. Thus, as the Latin sound u had become ó (close o) and u had become y (French u, German ü), the letter u was used sometimes to denote the sound ó, sometimes the sound y; as Latin k (written c) had become tsh or ts, according to dialect, before e and i, c was used to represent those sounds as well as that of k. The chief features of early French orthography (apart from the specialities of individual MSS., especially the earliest) are therefore these: - c stood for k and tsh or ts; d for d and dh (soft th); e for é, è, and ǝ; g for g and dzh; h was often written in words of Latin origin where not sounded; i (j) stood for i, y consonant, and dzh; o for ó (Anglo-Norman u) and ò; s for s and z; t for t and th; u (v) for ó (Anglo-Norman u), y and v; y (rare) for i; z for dz and ts. Some new sounds had also to be provided for: where tsh had to be distinguished from non-final ts, ch - at first, as in Italian, denoting k before i and e (chi = ki from qvi) - was used for it; palatal l was represented by ill, which when final usually lost one l, and after i dropped its i; palatal n by gn, ng or ngn, to which i was often prefixed; and the new letter w, originally uu (vv), and sometimes representing merely uv or vu, was employed for the consonant-sound still denoted by it in English. All combinations of vowel-letters represented diphthongs; thus ai denoted a followed by i, ou either óu or òu, ui either ói (Anglo-Norman ui) or yi, and similarly with the others - ei, eu, oi, iu, ie, ue (and oe), and the triphthong ieu. Silent letters, except initial h in Latin words, are very rare; though MSS. copied from older ones often retain letters whose sounds, though existing in the language of the author, had disappeared from that of the more modern scribe. The subsequent changes in orthography are due mainly to changes of sound, and find their explanation in the phonology. Thus, as Old French progresses, s, having become silent before voiced consonants, indicates only the length of the preceding vowel; e before nasals, from the change of ē (nasal e) to ã (nasal a), represents ã; c, from the change of ts to s, represents s; qu and gu, from the loss of the w of kw and gw, represent k and g (hard); ai, from the change of ai to è, represents è; ou, from the change of òu and óu to u, represents u; ch and g, from the change of tsh and dzh to sh and zh, represent sh and zh; eu and ue, originally representing diphthongs, represent œ (German ö); z, from the change of ts and dz to s and z, represents s and z. The new values of some of these letters were applied to words not originally spelt with them: Old French k before i and e was replaced by qu (evesque, eveske, Latin episcopum); Old French u and o for ó, after this sound had split into eu and u, were replaced in the latter case by ou (rous, for ros or rus, Latin russum); s was accidentally inserted to mark a long vowel (pasle, pale, Latin pallidum); eu replaced ue and oe (neuf, nuef, Latin novum and novem); z replaced s after é (nez, nes, nasum). The use of x for final s is due to an orthographical mistake; the MS. contraction of us being something like x was at last confused with it (iex for ieus, oculōs), and, its meaning being forgotten, u was inserted before the x (yeux) which thus meant no more than s, and was used for it after other vowels (voix for vois, vōcem). As literature came to be extensively cultivated, traditional as distinct from phonetic spelling began to be influential; and in the 14th century, the close of the Old French period, this influence, though not overpowering, was strong - stronger than in England at that time. About the same period there arose etymological as distinct from traditional spelling. This practice, the alteration of traditional spelling by the insertion or substitution of letters which occurred (or were supposed to occur) in the Latin (or supposed Latin) originals of the French words, became very prevalent in the three following centuries, when such forms as debvoir (dēbēre) for devoir, faulx (falsum) for faus, autheur (auctōrem, supposed to be authōrem) for auteur, poids (supposed to be from pondus, really from pēnsum) for pois, were the rule. But besides the etymological, there was a phonetic school of spelling (Ramus, in 1562, for instance, writes èime, èimates - with e = é, è = è, and ę = ǝ - for aimai, aimastes), which, though unsuccessful on the whole, had some effect in correcting the excesses of the other, so that in the 17th century most of these inserted letters began to drop; of those which remain, some (flegme for flemme or fleume, Latin phlegma) have corrupted the pronunciation. Some important reforms - as the dropping of silent s, and its replacement by a circumflex over the vowel when this was long; the frequent distinction of close and open e by acute and grave accents; the restriction of i and u to the vowel sound, of j and v to the consonant; and the introduction from Spain of the cedilla to distinguish c = s from c = k before a, u and o - are due to the 16th century. The replacement of oi, where it had assumed the value è, by ai, did not begin till the last century, and was not the rule till the present one. Indeed, since the 16th century the changes in French spelling have been small, compared with the changes of the sounds; final consonants and final e (unaccented) are still written, though the sounds they represent have disappeared.
Still, a marked effort towards the simplification of French orthography was made in the third edition of the Dictionary of the French Academy (1740), practically the work of the Abbé d'Olivet. While in the first (1694) and second (1718) editions of this dictionary words were overburdened with silent letters, supposed to represent better the etymology, in the third edition the spelling of about 5000 words (out of about 18,000) was altered and made more in conformity with the pronunciation. So, for instance, c was dropped in beinfaicteur and object, ç in sçavoir, d in advocat, s in accroistre, albastre, aspre and bastard, e in the past part. creu, deu, veu, and in such words as alleure, souilleure; y was replaced by i in cecy, celuy, gay, joye, etc. But those changes were not made systematically, and many pedantic spellings were left untouched, while many inconsistencies still remain in the present orthography (siffler and persifler, souffler and boursoufler, etc). The consequence of those efforts in contrary directions is that French orthography is now quite as traditional and unphonetic as English, and gives an even falser notion than this of the actual state of the language it is supposed to represent. Many of the features of Old French orthography, early and late, are preserved in English orthography; to it we owe the use of c for s (Old English c = k only), of j (i) for dzh, of v (u) for v (in Old English written f), and probably of ch for tsh. The English w is purely French, the Old English letter being the runic . When French was introduced into England, kw had not lost its w, and the French qu, with that value, replaced the Old English c (queen for cen). In Norman, Old French ó had become very like u, and in England went entirely into it; o, which was one of its French signs, thus came to be often used for u in English (come for cume). U, having often in Old French its Modern French value, was so used in England, and replaced the Old English y (busy for bysi, Middle English brud for brŷd), and y was often used for i (day for dai). In the 13th century, when ou had come to represent u in France, it was borrowed by English, and used for the long sound of that vowel (sour for sur); and gu, which had come to mean simply g (hard), was occasionally used to represent the sound g before i and e (guess for gesse). Some of the Early Modern etymological spellings were imitated in England; fleam and autour were replaced by phlegm and authour, the latter spelling having corrupted the pronunciation.
(e) Inflections. - In the earliest Old French extant, the influence of analogy, especially in verbal forms, is very marked when these are compared with Latin (thus the present participles of all conjugations take ant, the ending of the first, Latin antem), and becomes stronger as the language progresses. Such isolated inflectional changes as saveit into savoit, which are cases of regular phonetic changes, are not noticed here.
(i.) Verbs. - (1) In the oldest French texts the Latin pluperfect (with the sense of the perfect) occasionally occurs - avret (habuerat), roveret (rogaverat); it disappears before the 12th century. (2) The u of the ending of the 1st pers. plur. mus drops in Old French, except in the perfect, where its presence (as ǝ) is not yet satisfactorily explained - amoms (amamus, influenced by sumus), but amames (amavimus). In Picard the atonic ending mes is extended to all tenses, giving amomes, etc. (3) In the present indicative, 2nd person plur., the ending ez of the first conjugation (Latin atis) extends, even in the earliest documents, to all verbs - avez, recevez, oez (habetis, recipĭtis, auditis) like amez (amatis); such forms as dites, faites (dicĭtis, facĭtis) being exceptional archaisms. This levelling of the conjugation does not appear at such an early time in the future (formed from the infinitive and from habētis reduced to ētis); in the Roland both forms occur, portereiz (portare habētis) assonating on rei (roi, rēgem), and the younger porterez on citet (cité, civitatem), but about the end of the 13th century the older form -eiz, -oiz, is dropped, and -ez becomes gradually the uniform ending for this 2nd person of the plural in the future tense. (4) In Eastern French the 1st plur., when preceded by i, has e, not o, before the nasal, while Western French has u (or o), as in the present; posciomes (posseamus) in the Jonah homily makes it probable that the latter is the older form - Picard aviemes, Burgundian aviens, Norman aviums (habēbamus). (5) The subjunctive of the first conjugation has at first in the singular no final e, in accordance with the final vowel laws - plur, plurs, plurt (plōrem, plōrēs, plōret). The forms are gradually assimilated to those of the other conjugations, which, deriving from Latin am, as, at, have e, es, e(t); Modern French pleure, pleures, pleure, like perde, perdes, perde (perdam, perdas, perdat). (6) In Old French the present subjunctive and the 1st sing. pres. ind. generally show the influence of the i or e of the Latin iam, eam, iō, eō - Old French muire or moerge (moriat for moriatur), tiegne or tienge (teneat), muir or moerc (moriō for morior), tieng or tienc (teneō). By degrees these forms are levelled under the other present forms - Modern French meure and meurs following meurt (morit for moritur), tienne and tiens following tient (tenet). A few of the older forms remain - the vowel of aie (habeam) and ai (habeō) contrasting with that of a (habet). (7) A levelling of which instances occur in the 11th century, but which is not yet complete, is that of the accented and unaccented stem-syllables of verbs. In Old French many verb-stems with shifting accent vary in accordance with phonetic laws - parler (parabolare), amer (amare) have in the present indicative parol (parabolō), paroles (parabolas), parolet (parabolat), parlums (parabolamus), parlez (parabolatis), parolent (parabolant); aim (amō), aimes (amas), aimet (amat), amums (amamus), amez (amatis), aiment (amant). In the first case the unaccented, in the second the accented form has prevailed - Modern French parle, parler; aime, aimer. In several verbs, as tenir (tenēre), the distinction is retained - tiens, tiens, tient, tenons, tenez, tiennent. (8) In Old French, as stated above, ié instead of é from a occurs after a palatal (which, if a consonant, often split into i with a dental); the diphthong thus appears in several forms of many verbs of the 1st conjugation - preier (= prei-ier, precare), vengier (vindicare), laissier (laxare), aidier (adjutare). At the close of the Old French period, those verbs in which the stem ends in a dental replace ie by the e of other verbs - Old French laissier, aidier, laissiez (laxatis), aidiez (adjutatis); Modern French laisser, aider, laissez, aidez, by analogy of aimer, aimez. The older forms generally remain in Picard - laissier, aidier. (9) The addition of e to the 1st sing. pres. ind. of all verbs of the first conjugation is rare before the 13th century, but is usual in the 15th; it is probably due to the analogy of the third person - Old French chant (cantō), aim (amō); Modern French chante, aime. (10) In the 13th century s is occasionally added to the 1st pers. sing., except those ending in e (= ǝ) and ai, and to the 2nd sing. of imperatives; at the close of the 16th century this becomes the rule, and extends to imperfects and conditionals in oie after the loss of their e. It appears to be due to the influence of the 2nd pers. sing. - Old French vend (vendō and vende), vendoie (vendēbam), parti (partivi), ting (tenui); Modern French vends, vendais, partis, tins; and donne (dōna) in certain cases becomes donnes. (11) The 1st and 2nd plur. of the pres. subj., which in Old French were generally similar to those of the indicative, gradually take an i before them, which is the rule after the 16th century - Old French perdons (perdamus), perdez (perdatis); Modern French perdions, perdiez, apparently by analogy of the imp. ind. (12) The loss in Late Old French of final s, t, etc., when preceding another consonant, caused many words to have in reality (though often concealed by orthography) double forms of inflection - one without termination, the other with. Thus in the 16th century the 2nd sing. pres. ind. dors (dormis) and the 3rd dort (dormit) were distinguished as dòrz and dòrt when before a vowel, as dòrs and dòrt at the end of a sentence or line of poetry, but ran together as dòr when followed by a consonant. Still later, the loss of the final consonant when not followed by a vowel further reduced the cases in which the forms were distinguished, so that the actual French conjugation is considerably simpler than is shown by the customary spellings, except when, in consequence of an immediately following vowel, the old terminations occasionally appear. Even here the antiquity is to a considerable extent artificial or delusive, some of the insertions being due to analogy, and the popular language often omitting the traditional consonant or inserting a different one. (13) The subsequent general loss of e = ǝ in unaccented final syllables has still further reduced the inflections, but not the distinctive forms - perd (perdit) and perde (perdat) being generally distinguished as pèr and pèrd, and before a vowel as pèrt and pèrd.
(ii.) Substantives. - (1) In Early Old French (as in Provençal) there are two main declensions, the masculine and the feminine; with a few exceptions the former distinguishes nominative and accusative in both numbers, the latter in neither. The nom. and acc. sing, and acc. plur. mas. correspond to those of the Latin 2nd or 3rd declension, the nom. plur. to that of the 2nd declension. The sing, fem. corresponds to the nom. and acc. of the Latin 1st declension, or to the acc. of the 3rd; the plur. fem. to the acc. of the 1st declension, or to the nom. and acc. of the 3rd. Thus masc. tors (taurus), lere (latrō); tor (taurum), laron (latrōnem); tor (tauri), laron (latrōni for -nēs); tors (taurōs), larons (latrōnēs); but fem. only ele (ala and alam), flor (flōrem); eles (alas), flors (flōrēs nom. and acc.). About the end of the 11th century feminines not ending in e = ǝ take, by analogy of the masculines, s in the nom. sing., thus distinguishing nom. flors from acc. flor. A century later, masculines without s in the nom. sing. take this consonant by analogy of the other masculines, giving leres as nom. similar to tors. In Anglo-Norman the accusative forms very early begin to replace the nominative, and soon supersede them, the language following the tendency of contemporaneous English. In continental French the declension-system was preserved much longer, and did not break up till the 14th century, though acc. forms are occasionally substituted for nom. (rarely nom. for acc.) before that date. It must be noticed, however, that in the current language the reduction of the declension to one case (generally the accusative) per number appears much earlier than in the language of literature proper and poetry; Froissart, for instance, c.1400, in his poetical works is much more careful of the declension than in his Chronicles. In the 15th century the modern system of one case is fully established; the form kept is almost always the accusative (sing. without s, plural with s), but in a few words, such as fils (filius), sœur (soror), pastre (pastor), and in proper names such as Georges, Gilles, etc., often used as vocative (therefore with the form of nom.); the nom. survives in the sing. Occasionally both forms exist, in different senses - sire (senior) and seigneur (seniōrem), on (homō) and homme (hominem). (2) Latin neuters are generally masculine in Old French, and inflected according to their analogy, as ciels (caelus for caelum nom.), ciel (caelum acc.), ciel (caeli for caela nom.), ciels (caelōs for caela acc.); but in some cases the form of the Latin neuter is preserved, as in cors, now corps, Lat. corpus; tens, now temps, Lat. tempus. Many neuters lose their singular form and treat the plural as a feminine singular, as in the related languages - merveille (mirabilia), feuille (folia). But in a few words the neuter plural termination is used, as in Italian, in its primitive sense - carre (carra, which exists as well as carri), paire (Lat. paria); Modern French chars, paires. (3) In Old French the inflectional s often causes phonetic changes in the stem; thus palatal l before s takes t after it, and becomes dental l, which afterwards changes to u or drops - fil (filium and filii) with palatal l, filz (filius and filiōs), afterwards fiz, with z = ts (preserved in English Fitz), and then fis, as now (spelt fils). Many consonants before s, as the t of fiz, disappear, and l is vocalized - vif (vivum), mal (malum), nominative sing. and acc. plur. vis, maus (earlier mals). These forms of the plural are retained in the 16th century, though often etymologically spelt with the consonant of the singular, as in vifs, pronounced vis; but in Late Modern French many of them disappear, vifs, with f sounded as in the singular, being the plural of vif, bals (formerly baux) that of bal. In many words, as chant (cantus) and champs (campōs) with silent t and p (Old French chans in both cases), maux (Old French mals, sing. mal), yeux (oculōs, Old French œlz, sing. œil) the old change in the stem is kept. Sometimes, as in cieux (caelōs) and ciels, the old traditional and the modern analogical forms coexist, with different meanings. (4) The modern loss of final s (except when kept as z before a vowel) has seriously modified the French declension, the singulars fort (fòr) and forte (fòrt) being generally undistinguishable from their plurals forts and fortes. The subsequent loss of ǝ in finals has not affected the relation between sing. and plur. forms; but with the frequent recoining of the plural forms on the singular present Modern French has very often no distinction between sing. and plur., except before a vowel. Such plurals as maux have always been distinct from their singular mal; in those whose singular ends in s there never was any distinction, Old French laz (now spelt lacs) corresponding to laqveus, laqveum, laqvei and laqveōs.
(iii.) Adjectives. - (1) The terminations of the cases and numbers of adjectives are the same as those of substantives, and are treated in the preceding paragraph. The feminine generally takes no e if the masc. has none, and if there is no distinction in Latin - fem. sing. fort (fortem), grant (grandem), fem. plur. forz (fortēs), granz (grandēs), like the acc. masc. Certain adjectives of this class, and among them all the adjectives formed with the Latin suffix -ensis, take regularly, even in the oldest French, the feminine ending e, in Provençal a (courtois, fem. courtoise; commun, fem. commune). To these must not be added dous (Mod. Fr. dolz, dous), fem. douce, which probably comes from a Low Latin dulcius, dulcia. In the 11th century some other feminines, originally without e, begin in Norman to take this termination - grande (in a feminine assonance in the Alexis), plur. grandes; but other dialects generally preserve the original form till the 14th century. In the 16th century the e is general in the feminine, and is now universal, except in a few expressions - grand'mère (with erroneous apostrophe, grandem, matrem), lettres royaux (literas rēgalēs), and most adverbs from adjectives in -ant, -ent - couramment (currante for -ente mente), sciemment (sciente mente). (2) Several adjectives have in Modern French replaced the masc. by the feminine - Old French masc. roit (rigidum), fem. roide (rigidam); Modern French roide for both genders. (3) In Old French several Latin simple comparatives are preserved - maiur (majōrem), nom. maire (major); graignur (grandiōrem), nom. graindre (grandior); only a few of these now survive - pire (pejor), meilleur (meliōrem), with their adverbial neuters pis (pejus), mieux (melius). The few simple superlatives found in Old French, as merme (minimum), pesme (pessimus), proisme (proximum), haltisme (altissimum), this last one being clearly a literary word, are now extinct, and, when they existed, had hardly the meaning of a superlative. (4) The modern loss of many final consonants when not before vowels, and the subsequent loss of final ǝ, have greatly affected the distinction between the masc. and fem. of adjectives - fort and forte are still distinguished as fòr and fòrt, but amer (amarum) and amère (amaram), with their plurals amers and amères, have run together.
(f) Derivation. - Most of the Old French prefixes and suffixes are descendants of Latin ones, but a few are Teutonic (ard = hard), and some are later borrowings from Latin (arie, afterwards aire, from arium). In Modern French many old affixes are hardly used for forming new words; the inherited ier (arium) is yielding to the borrowed aire, the popular contre (contra) to the learned anti (Greek), and the native ée (atam) to the Italian ade. The suffixes of many words have been assimilated to more common ones; thus sengler (singularem) is now sanglier.
(g) Syntax. - Old French syntax, gradually changing from the 10th to the 14th century, has a character of its own, distinct from that of Modern French; though when compared with Latin syntax it appears decidedly modern.
(1) The general formal distinction between nominative and accusative is the chief feature which causes French syntax to resemble that of Latin and differ from that of the modern language; and as the distinction had to be replaced by a comparatively fixed word-order, a serious loss of freedom ensued. If the forms are modernized while the word-order is kept, the Old French l'archevesque ne puet flechir li reis Henris (Latin archiepiscopum nōn potest flectere rex Henricus) assumes a totally different meaning - l'archevêque ne peut fléchir le roi Henri. (2) The replacement of the nominative form of nouns by the accusative is itself a syntactical feature, though treated above under inflection. A more modern instance is exhibited by the personal pronouns, which, when not immediately the subject of a verb, occasionally take even in Old French, and regularly in the 16th century, the accusative form; the Old French je qui sui (ego qvi sum) becomes moi qui suis, though the older usage survives in the legal phrase je soussigné.... (3) The definite article is now required in many cases where Old French dispenses with it - jo cunquis Engleterre, suffrir mort (as Modern French avoir faim); Modern French l'Angleterre, la mort. (4) Old French had distinct pronouns for "this" and "that" - cest (ecce istum) and cel (ecce illium), with their cases. Both exist in the 16th century, but the present language employs cet as adjective, cel as substantive, in both meanings, marking the old distinction by affixing the adverbs ci and là - cet homme-ci, cet homme-là; celui-ci, celui-là. (5) In Old French, the verbal terminations being clear, the subject pronoun is usually not expressed - si ferai (sic facere habeō), est durs (durus est), que feras (quid facere habēs)? In the 16th century the use of the pronoun is general, and is now universal, except in one or two impersonal phrases, as n'importe, peu s'en faut. (6) The present participle in Old French in its uninflected form coincided with the gerund (amant = amantem and amandō), and in the modern language has been replaced by the latter, except where it has become adjectival; the Old French complaingnans leur dolours (Latin plaṇgentēs) is now plaignant leurs douleurs (Latin plaṇgendō). The now extinct use of estre with the participle present for the simple verb is not uncommon in Old French down to the 16th century - sont disanz (sunt dicentēs) = Modern French ils disent (as English they are saying). (7) In present Modern French the preterite participle when used with avoir to form verb-tenses is invariable, except when the object precedes (an exception now vanishing in the conversational language) - j'ai écrit les lettres, les lettres que j'ai écrites. In Old French down to the 16th century, formal concord was more common (though by no means necessary), partly because the object preceded the participle much oftener than now - ad la culur muée (habet colōrem mutatam), ad faite sa venjance, les turs ad rendues. (8) The sentences just quoted will serve as specimens of the freedom of Old French word-order - the object standing either before verb and participle, between them, or after both. The predicative adjective can stand before or after the verb - halt sunt li pui (Latin podia), e tenebrus e grant. (9) In Old French ne (Early Old French nen, Latin nōn) suffices for the negation without pas (passum), point (puṇctum) or mie (micam, now obsolete), though these are frequently used - jo ne sui lis sire (je ne suis pas ton seigneur), autre feme nen ara (il n'aura pas autre femme). In principal sentences Modern French uses ne by itself only in certain cases - je ne puis marcher, je n'ai rien. The slight weight as a negation usually attached to ne has caused several originally positive words to take a negative meaning - rien (Latin rem) now meaning "nothing" as well as "something." (10) In Old French interrogation was expressed with substantives as with pronouns by putting them after the verb - est Saul entre les prophètes? In Modern French the pronominal inversion (the substantive being prefixed) or a verbal periphrasis must be used - Saul est-il? or est-ce que Saul est?
(h) Summary. - Looking at the internal history of the French language as a whole, there is no such strongly marked division as exists between Old and Middle English, or even between Middle and Modern English. Some of the most important changes are quite modern, and are concealed by the traditional orthography; but, even making allowance for this, the difference between French of the 11th century and that of the 20th is less than that between English of the same dates. The most important change in itself and for its effects is probably that which is usually made the division between Old and Modern French, the loss of the formal distinction between nominative and accusative; next to this are perhaps the gradual loss of many final consonants, the still recent loss of the vowel of unaccented final syllables, and the extension of analogy in conjugation and declension. In its construction Old French is distinguished by a freedom strongly contrasting with the strictness of the modern language, and bears, as might be expected, a much stronger resemblance than the latter to the other Romanic dialects. In many features, indeed, both positive and negative, Modern French forms a class by itself, distinct in character from the other modern representatives of Latin.
IV. Bibliography. - The few works which treat of French philology as a whole are now in many respects antiquated, and the important discoveries of recent years, which have revolutionized our ideas of Old French phonology and dialectology, are scattered in various editions, periodicals, and separate treatises. For many things Diez's Grammatik der romanischen Sprachen (4th edition - a reprint of the 3rd - Bonn, 1876-1877; French translation, Paris, 1872-1875) is still very valuable; Burguy's Grammaire de la Langue d'Oïl (2nd edition - a reprint of the 1st - Berlin, 1869-1870) is useful only as a collection of examples. Schwan's Grammatik des Altfranzösischen, as revised by Behrens in the 3rd edition (Leipzig, 1898; French translation, Leipzig and Paris, 1900), is by far the best old French grammar we possess. For the history of French language in general see F. Brunot, Histoire de la langue française des origines à 1900 (Paris, 1905, 1906, etc.). For the history of spelling, A. F. Didot, Observations sur l'orthographe ou ortografie française suivies d'une histoire de la réforme orthographique depuis le XVe siècle jusqu'à nos jours (2nd ed., Paris, 1868). For the history of French sounds: Ch. Thurot, De la prononciation française depuis le commencement du XVIe siècle, d'après les témoignages des grammairiens (2 vols., Paris, 1881-1883). For the history of syntax, apart from various grammatical works of a general character, much is to be gathered from Ad. Tobler's Vermischte Beiträge zur französischen Grammatik (3 parts, 1886, 1894, 1899, parts i. and ii. in second editions, 1902, 1906). G. Paris's edition of La Vie de S. Alexis (Paris, 1872) was the pioneer of, and retains an important place among, the recent original works on Old French. Darmesteter and Hatzfeld's Le Seizième Siècle (Paris, 1878) contains the first good account of Early Modern French. Littré's Dictionnaire de la langue française (4 vols., Paris, 1863-1869, and a Supplement, 1877); and Hatzfeld, Darmesteter and Thomas, Dict. général de la langue française, more condensed (2 vols., Paris, 1888-1900), contain much useful and often original information about the etymology and history of French words. For the etymology of many French (and also Provençal) words, reference must be made to Ant. Thomas's Essais de philologie française (Paris, 1897) and Nouveaux essais de philologie française (Paris, 1904). But there is no French dictionary properly historical. A Dictionnaire historique de la langue française was begun by the Académie française (4 vols., 1859-1894), but it was, from the first, antiquated. It contains only one letter (A) and has not been continued. The leading periodicals now in existence are the Romania (Paris), founded (in 1872) and edited by P. Meyer and G. Paris (with Ant. Thomas since the death of G. Paris in 1903), and the Zeitschrift für romanische Philologie (Halle), founded (in 1877) and edited by G. Gröber. To these reference should be made for information as to the very numerous articles, treatises and editions by the many and often distinguished scholars who, especially in France and Germany, now prosecute the scientific study of the language. It may be well to mention that, Old French phonology especially being complicated, and as yet incompletely investigated, these publications, the views in which are of various degrees of value, require not mere acquiescent reading, but critical study. The dialects of France in their present state (patois) are now being scientifically investigated. The special works on the subject (dictionaries, grammars, etc.) cannot be fully indicated here; we must limit ourselves to the mention of Behren's Bibliographie des patois gallo-romans (2nd ed., revised Berlin, 1893), and of Gilliéron and Edmont's Atlas linguistique de la France (1902 et seq.), a huge publication planned to contain about 1800 maps.
(H. N.; P. M.)
Note - this article incorporates content from Encyclopaedia Britannica, Eleventh Edition, (1910-1911)