348 SPAIN [LANGUAGE. very early period, earlier doubtless than the oldest exist- ing monuments of those languages, a labio-palatal pro- nunciation (?<). It is not to be supposed that the separa- tion of Catalan from the Gallo-Koman family should have occurred before the transformation had taken place ; there is good reason to believe that Catalan possessed the u at one time, but afterwards lost it in its contact with the Spanish dialects. The question, however, is one for further examination. Catalan being a variety of the lan^ie d'oc, it will be convenient to note the peculiarities of its phonetic and inflexion as compared with ordinary Provencal. Tonic Vowels. With regard to a, which is pronounced alike in open and close syllables (amar, amare ; abre, arbor), there is nothing to remark. The Latin e, which is treated like I, gives e, sometimes close, sometimes open. On this point Catalan is more hesitating than Provci^al ; it does not distinguish so clearly the pronunciation of e according to its origin ; while e (I) is capable of yielding an open e, the is often pronounced close, and the poets have no difficulty in making words in e close and in c open rhyme together, which is not the case in Proven?al. The Latin never yields ie in Catalan as it does in French and occasionally in Pro- vengal ; sedet becomes sen (where u represents the final d), pedem makes pen, and ego eu; in some words where the tonic e is followed by a syllable in which an i occurs, it may become i (ir, heri ; mig, medius ; 'mils, melius) ; and the same holds good for e in a similar situation (cm, cerius, cereus ; fira, feria), and for c in a close syllable before a nasal (eximpli, exemplum ; mintr6 for mentire, gint for gent), /tonic long and i short, when in hiatus with another vowel, produce i (amich, amicus ; via, via). tonic long and o short are represented by o close and o open (amor, amorem; poble, populus). short is never diphthongized into uo or ue ; such a treatment is as foreign to Catalan as the diphthongization of into ie. Just as e before a syllable in which an i occurs is changed into i, so in the same circumstances o becomes u (full, folium ; vail, volio for voleo) and also when the accented vowel precedes a group of consonants like cl, pi, and the like (ull, oc'lus; cscull, scop'lus). Latin u persists with the Latin pronunciation, and, as already said, does not take the Franco-Provensal pronunciation u. Latin au becomes o (cosa, causa; or, aurum); Old-Catalan has kept the diphthong better, but possibly we should attribute the examples of au which are met with in texts of the 13th and 14th centuries to the literary influence of Provence. Latin ua tends to become o (cor, quare). Atonic Vowels. As for the Latin post-tonic vowels already spoken of, it remains to be noted that a is often represented in writing by e, especially before s ; in old Catalan, the substantives, adjectives, and participles readily form their singular in a and their plural in es: arma, armes (anima, animas); bona, bones (bona, bonas); amada, amades (amata, amatas). This e is neither open nor close, but a surd e the pronunciation of which comes very near a. In the same way the supporting vowel, which is regularly an e in Catalan, is often written a, especially after r (abra, arborem ; astra, astrum ; para, patrem); one may say that in the actual state of the language post-tonic e and a become indistinguishable in a surd sound inter- mediate between the French a and mute e. Before the tonic the same change between a and e constantly takes place ; one finds in manuscripts enar, emor for anar, amor (the same manner of writing extends even to the case of the tonic syllable, ten and sent from tantum and sanctum being far from rare), and, on the other hand, antre, arrar, for entre, errar. I atonic is often represented by e even when it is long (vehi, vicinus). atonic close, which in genuine Catalan exists only before the tonic, has become u ; at the present day truvar, cuntradir is the real pronunciation of the words spelt trovar, contradir, and in the final syllables, verbal or other, where under Castilian influence an o has come to be added to the normal Catalan form, this o has the value of a u : trovo (genuine Catalan, trop) is pronounced trovu ; bravo (genuine Catalan, bran) is pronounced bravu. U atonic keeps its ground. The only strong diphthongs of the spoken language are di, au (rather rare), ei, &u, iu, 6i, 6u, ui, uu. Ai produced by a + i or by a + a palatal consonant has for the greater part of the time become an e in the modern language ; factum has yielded fait, feit, and then fet, the last being the actual form ; arius has given er alongside of aire, ari, which are learned or semi-learned forms. Of the two weak diphthongs id and ud, the latter, as has been seen, tends to become o close in the atonic syllable, and is pronounced u : quaranta has become coranta, then curanta. After the tonic ua often becomes a in the Catalan of the mainland (ayga, aqua, llenga, lingua), while in Majorca it becomes o (aygo, llcngo). Consonants. Final t readily disappears after n or I (tan, tantum ; aman, venin, partin, for amant, venint, &c. ; mol, multum; ocul, ocultum) ; the t reappears in composition before a vowel (/em, fontem, but Font-alba). On the other hand, a t without etymo- logical origin is frequently added to words ending in r (cart for car, quare ; mart for mar, mare ; amart, ohirt, infinitive for amar, ohir), and even to some words terminating in a vowel (genit, ingenium ; premit, premium), or the addition of the t has taken place by assimilation to past participles in it. The phenomenon occurs also in Provenal (see Romania, vii. 107, viii. 110). Median intervocal d, represented by s (z) in the first stage of the language, has dis- appeared : fidelis gavo/eseZ, then feel, and finally fcl ; videtis became vezets, then reels, vets, and vcu. Final d after a vowel has produced u (peu, pedem; niu, nidnm; mott, modum); but, when th~ d, in consequence of the disappearance of the preceding vowel, rests upon a consonant, it remains and passes into the correspond- ing surd : frigidus gives fred (pronounced fret}. The group dr, when produced by the disappearance of the intermediate vowel, becomes ur (crcurc, credere; oc'iure, occidere; veure, videre ; seure, seder e). Final n, if originally it stood between two vowels, falls away (bo, bonum ; vi, vinum), but not when it answers to mn (thus donum makes do, but doninum don; sonum makes so, but somnum son). Nd is reduced to n (demanar, comanarfor demandar, comandar). Assibilated c before e, i is treated like d ; within a word it disappears after having been represented for a while by 5 (lucere gives llusir, lluhir; recipere gives rezebrc, reebre, rcbrc) ; at the end of a word it is replaced by u, (veu, vicem; feu, fecit). The group c'r gives ur, just like d'r (jaure, jacere; naure, nocere; plaure, placere ; but facere, dicere, ducere, make far (fcr), dir, dur. Initial I has been preserved only in certain monosyllables (the article lo, los) ; everywhere else it has been replaced by I mouillee (Prov. Ih), which in the present ortho- graphy is written II as in Castilian, but formerly used to be repre- sented by ly or yl (lletra, litera; llengua, lingua). P readily disappears after m, like t after n (cam, campum ; terns, tempus). B is replaced by the surd p at the end of a word (trobar in the in- finitive, but trop in the present tense) ; so also in the interior of a word when it precedes a consonant (supvenir, sub venire, sopte, sub'to). Median intervocalic / gives v (Estece, Stephanus); it has disappeared from profundus, which yielded the form preon, then pregon (g being introduced to obviate the hiatus). V, wher- ever it has been preserved, has the same pronunciation as b ; at the end of a word and between vowels it becomes vocalized into u (suau, suavis ; viure, vivere). C guttural, written qu before e and i, keeps its ground as a central and as a final letter ; in the latter position it is generally written ch (amich, amicum ; jock, jo cum). G guttural is replaced as a final letter by surd c (longa, but lone ; trigar, but trich). Tj after a consonant gives ss (cassar, captiare) ; between vowels, after having been represented by soft s, it has disappeared (ration em gave razd, raysd, then rah6) ; at the end of every word it behaves like ts, that is to say, changes into u (preu, pretium); instead of ts the second person plural of the verb at(i)s, et(i)s, it(i)s now has au, cu, iu after havinghad ats, ets, its. t>j gives $ between vowels (verger, viridiarum), and c as a terminal (written either ig or tx~. goig, gaudium, 'mig, mitx, medium). Stj and sc before e and i, as well as x anH ps, yield the sound sh, represented in Catalan by x (angoxa, angustia ; coneixer, cognoscere ; dix, dixit; matcix, metipse). ^almost everywhere has taken the sound of the French j (jutge, &c. ). Lj and II give I mouillee (II in the present orthography: fill, filium ; consell, con- silium; null, nullum). In the larger portion of the Catalan domain this I mouillee has become y ; almost everywhere Jz?/ is pro- nounced for fill, consey for consell. Nj and nn give n mouillee (ny in both old and modern spelling: scnyor, seniorem; any, annum). Sometimes the ny becomes reduced to y ; one occasionally meets in manuscripts with seyor, ay, for scnyor, any, but this pronunciation has not become general, as has been the case with the y having its origin in II. Lingual r at the end of a word has a tendency to dis- appear when preceded by a vowel : thus the infinitives amare, temere, *legire are pronounced amd, tcme, llcgl. It is never pre- served except when protected by the non-etymological t already spoken of (llcgirt or llegl, but never Ucglr) ; the r reappears, never- theless, whenever the infinitive is followed by a pronoun (donarme, dirho). Ms is reduced to s (cos for cors, corpus). H is merely an orthographic sign ; it is used to indicate that two consecutive vowels do not form a diphthong (vehi, raho), and, added to c, it denotes the pronunciation of the guttural c at the end of a word (amich). Inflexion. Catalan, unlike Old Provencal and Old French, has never had declensions. It is true that in certain texts (especially metrical texts) certain traces of case-endings are to be met with, as for example Deus and Deu, amors and amor, clars and clar, forts and/or<, tuyt and tots, abduy and abdos, senyer and scnyor, empcrairc and cmperador; but, since these forms are used convertibly, the nominative form when the word is in the objective, and the accusative form when the word is the subject, we can only recog- nize in these cases a confused recollection of the Provencal rules known only to the literate but of which the transcribers of manu- scripts took no account. Catalan, then, makes no distinctions save in the Bender and the number of its nouns. As regards the forma- tion of the plural only two observations are necessary. (1) Words which have their radical termination in n but which in the singular
