Portal de Periódicos da CAPES

On the Permutation of the Vulgar Latin Tonic Front Mid-Vowels in Medieval Catalan

1987; Liverpool University Press; Volume: 64; Issue: 4 Linguagem: Inglês

10.1080/1475382872000364305

1469-3550

Thomas J. Walsh,

Linguistics and language evolution

Click to increase image sizeClick to decrease image sizeBSS Subject Index: CATALONIA/CATALUÑA — SPAIN — CATALÁN LANGUAGE & ITS HISTORY Notes 1. ‘L’é (= ē, i) latin en ancien français et en mayorquin’, Romania, XVII (1888), 89—95. 2. ‘Les e toniques du catalan’, RHi, XV (1906), 9–23. 3. ‘La diphtongaison en catalan’, BDC, XIII (1925), 1–46, at 29–30. 4. Das Katalanische, Seine Stellung zum Spanischen und Provenzalischen, Sammlung romanischer Elementar- und Handbücher (Heidelberg: Winter, 1925), 14–18. 5. Meyer-Lübke, Das Katalanische, 16, cited parallels from Eastern France, Western Gascony, Emilia and the Piedmont. 6. Das Katalanische, 16. 7. El dialecto de Alguer y su posición en la historia de la lengua catalana, Publicacions de l'Oficina Romànica (Barcelona: Balmes, 1934), 107 (appeared originally as a sequence of two articles in AORLL, V [1932], 121–77 and VII [1934], 41–112; references in the present paper are to the monographic version). 8. The essentials of Kuen's account were reproduced by A. Badia Margarit, Gramática histórica catalana (Barcelona: Noguer, 1951), 130–34 and 137–41. In the discussion to follow, the reader may find it useful to refer to the maps on pp. 51, 73, 139 and 157 of that work. 9. E.g., DOMINICU ‘Sunday’ > /diwmén /, ligna ‘wood’ > /λé /, and pisce ‘fish’ > /pé/. 10. ‘L'evolució de les ee tòniques en català’, in Mélanges de philologie et de toponymie romanes offerts à Henri Guiter (Perpignan: No publ., 1982), 81–103. This article opens with a detailed historique du problème. 11. Nonetheless, consultation of J. Coromines, ‘El parlar de Cardós i Vall Ferrera’, in Entre dos llenguatges, II (Barcelona: Curial, 1976), 29–67, at 36–37 (originally published, with appended glossary, in BDC, XXIII [1936], 241–331), the only source for Pallarès data cited by Gulsoy, discloses a more complex situation as regards the outcomes of /ϵ/ before /r/, /s/, /t/, /k/, and /g/. After noting that the Pallarès /ϵ/ is considerably less open than its Eastern counterpart and admitting that he may occasionally have confused it with /e/, Coromines sums up the development of V.L. tonic /ϵ/ in Pallars as follows: hi ha é devant oclusiva labial, devant LL latina, devant nasal, devant rm i devant w < V o B; hi ha devant el grup llatí CL i devant r seguida de vocal o de b o p: devant d, davant s, i davant oclusiva velar apareix adés ϵ adés e sense norma clara. For the environment /r/ + vowel, Coromines cites only /f / ‘wild’ < feru and /re/ ‘he was’, reporting that the latter word is occasionally pronounced with /e/. For the outcome before /d/, he cites only /pϵdra/ ‘rock’, which in the majority of locales shows /e/. For the outcome before velar consonants, he adduces five lexemes of which two regularly exhibit /ϵ/ and three /e/. Before /s/, two words show /ϵ/, while four offer /e/, with one instance of wavering. The foregoing, he points out, is valid only for more westerly districts (Areu, Ferrera, and Cardós). The eastern part of Pallars (Tor) invariably shows /e/ in all the above-mentioned contexts. 12. For details see G. Rohlfs, Historische Grammatik der italienischen Sprache und ihrer Mundarten, I, Lautlehre (Bern: Francke, 1949), §§49, 53, 61. 13. Logically speaking, under Gulsoy's push-chain analysis, tendential raising of /ϵ/ to /e/ (cause) must have predated, by at least a slight margin, retraction of /e/ to // (effect). 14. See Kuen, El dialecto, 123–24 and 128–30. One is further justified in assigning the change /e/ > / / a very early date insofar as it must have preceded the closing of /ϵ/ to /e/ before nasals, an innovation shared by Catalan, Gascon, and Provençal; cf. uĒndere ‘to sell’ > Centr. /bndr/, Bal. /vndr/ vs. incendere > Centr./ Bal. /nsndr/. Assuming the opposite chronology, E of incendere would have merged with Ē of uĒndere, with both ultimately undergoing retraction. No such arguments for assigning an equally early date to raising of /ϵ/are, to my knowledge, available. 15. See Meyer-Lübke, Grammatik der romanischen Sprachen, I, Lautlehre (Leipzig: Fues Reisland), §341 and H. Lausberg, Lingüística románica, I, trans. J. Pérez Riesco and E. Pascual Rodríguez, Biblioteca Románica Hispánica (Madrid: Gredos, 1970), §117. Other scholars, such as C. Grandgent, An Introduction to Vulgar Latin (Boston: Heath, 1907), 153–57, project on to the plateau of spoken Latin an alternating stress system with secondary stresses invariably separated from the main stress by one syllable. 16. See C. Appel, Provenzalische Lautlehre (Leipzig: Reisland, 1918), 39–53; M. K. Pope, From Latin to Modern French, with Especial Consideration of Anglo-Norman: Phonology and Morphology (Manchester: Manchester U.P., 1934), 113; E. Richter, Beiträge zur Geschichte der Romanismen, I, Chronologische Phonetik des französischen bis zum Ende des 8. Jahrhunderts, ZRPh, Suppl. 82 (Halle: Niemeyer, 1934), 230–34; and, more recently, Lausberg, Lingüística románica, §295. Richter assigned such reduction to the sixth and seventh centuries. 17. See relevant sections of works mentioned in the previous note. 18. See Pope, From Latin to Modern French, 107–08. 19. Most experts view such words as Prov. altre ‘other’ < alteru,asne ‘ass’ < asinu,oncle ‘uncle’ < AuuNCULU, cobde ‘elbow’ < cubitu,segle ‘century’ < saeculu,sofre ‘I suffer’ < sufferô, etc., as confirming the assumption that atonic /o/, and hence presumably /e/ as well, must have passed through a stage //. Further examples are cited by O. Schultz-Gora, Altprovenzalisches Elementarbuch, Sammlung romanischer Elementar-bücher (Heidelberg: Winter, 1906), 34. 20. Characteristic verbs of this class are anomenar ‘to name’, bategar ‘to beat’, carregar ‘to load’, doblegar ‘to fold’, empedregar ‘to pave’, estossegar ‘to cough’, gemegar ‘to groan’, llampegar ‘to lightning’, mastegar ‘to chew’, mossegar ‘to bite’, pedregar ‘to hail’, rebregar ‘to squeeze’, and termenar ‘to end, limit’. See Fouché, Phonétique historique du roussillonais (Toulouse: Privat, 1924), 82–83; J. Huber, Katalanische Grammatik: Laut- und Formenlehre, Syntax, Wortbildung, Sammlung romanischer Elementar- und Handbücher (Heidelberg: Winter, 1929), 216; and F. de B. Moll, Gramática histórica catalana (Madrid: Gredos, 1960), 100. 21. See Rohlfs, Le Gascon: Etudes de philologie pyrénéenne, 2nd ed., ZRPh, Beiheft 85 (Tübingen: Niemeyer, 1970), 426, who, echoing J. Ronjat, Grammaire [h]istorique des parlers provençaux modernes, I (Montpellier: Société des Langues Romanes, 1930), 133, called the change of /e/ > /æ/ in Landais ‘une évolution sûrement récente’, without, however, producing any supporting arguments for that belief. 22. I prefer in the present context simply to note the opening and closing power of certain consonants and syllable configurations, while leaving the ultimate motivation for such effects to phoneticians. 23. The syllabification of such words as heretge, metge and petge is less clear-cut than one might be tempted to believe at first glance. For while the medieval graph tg (less frequently tj) is echoed by the palatal affricate in all modern dialects, j (as in roja ‘red [fem. sg.]’ and pluja ‘rain’) is regularly reflected by the palatal fricative in certain varieties (incl. the standard), but by the affricate in others (see map showing geographic distribution in Badia, Gramática, 185). Thus, while it seems highly probable that a distinction was consistently made in Old Catalan, its precise phonetic nature, i.e., // vs. // or / / vs. //, remains obscure. What is important in the present context is that raising could have been expected under either assumption. The tonic vowel of */m. /, in free syllable, would naturally have yielded /e/. Assuming */m . /, raising is also predictable owing to the palatal quality of the tautosyllablic consonant. Finally, the same reasoning may be applied even if the affricate straddled the syllable boundary (*/md. / or */md. /), since the syllable-final stop would surely have assimilated in point of articulation to the following syllable-initial palatal segment. Such a process is known to occur in the modern language, even across word boundaries (e.g., /st ardíns/ ‘seven gardens’ [s ríns], cited by Badia, Gramática, 99). 24. The only systematic set of exceptions to this generalization is constituted by /ϵ/ followed by /r/ + labial consonant, which unexpectedly yields /e/: e.g., cervu > /sérbu(l)/ ‘deer’, eremu > /érm/ ‘wasteland’, herba > /érb/ ‘grass’, *serpe > /sérp/ ‘snake’, seruat > /sérb/ ‘he serves’, and termine > /térm/ ‘limit’. No convincing explanation for this surprising development has ever, to my knowledge, been put forth, though Fouché, ‘La diphtongaison’, 34, and Kuen, EI dialecto, 113, made halfhearted attempts. 25. ‘El paper de les fronteres polítiques carolíngies en la gènesi dels primitius dialectes catalans.’ In Actes del quart col·loqui d’estudis catalans a Nord-Amèrica, Washington, D.C., 1984 (Barcelona: Publicacions de l'Abadia de Montserrat), 25–39. 26. The same phenomenon was noticed, at least for word-final //, by J. Séguy, Le Français parlé à Toulouse (Toulouse: Privat, 1978), 28, in the French of uneducated Toulousains (cf. /dóne/ for donne and /súfle/ for souffle). 27. See U. Weinreich, Languages in Contact (The Hague/Paris: Mouton, 1974), 18. 28. The Llibre del repartiment, compiled in Valencia in 1239, reports that 597 houses in the newly reconquered city were assigned to Aragonese, 630 to eastern Catalans and 388 to western Catalans; see A. Badia Margarit, ‘Sobre els elements medievals en la formació de la llengua catalana’, in Actes del tercer col·loqui d’estudis catalans a Nord-Amèrica, Toronto 1982, eds. P. Bohne, J. Massot i Muntaner and N. B. Smith (Barcelona: Publicacions de l'Abadia de Montserrat, 1983). 29. It is quite possible that the merger conjectured in Sect. 4.4.2 had already been completed in the western part of Old Catalonia by the time of the reconquest of Valencia. Be that as it may, the upshot of mixture of the three vocalic systems (Eastern Catalan, Western Catalan, and Aragonese/Mozarabic) would have been identical to that postulated in the present paragraph. 30. One encounters the open mid-vowel in these words in the East as well, but there it evolved, like most instances of V.L. /e/, through an intermediate stage //, preserved in the Balearic cognates (see forms cited at end of present paragraph). 31. The same process would presumably have been extended to instances of V.L. /e/ before /rr/, but no sure cases exist in Catalan. On the etymologies of cerro ‘tuft’ (showing /ϵ/ in all dialects) and cerra ‘bristle’ (in use only in the Balearic Islands, where it exhibits //) see relevant entries in Alcover-Moll, Diccionari and the DECLC. 32. See Kuen, El dialecto, 117–19, and Badia, Gramática, 137–38 and map on 139. 33. This dating, first proposed by J. B. Alart, ‘Études historiques sur quelques particularités de la langue catalane’, RLR, XII (1877), 109–32 (reprinted as an Appendix to the author's Documents sur la langue catalane des anciens comtés de Roussillon et de Cerdagne [Paris, 1881]), went unchallenged for almost a century, until J. Rafel i Fontanals, ‘La u catalana d'origen consonàntic’, ER, XII (1963—68), 179—211, at 200, argued for a distinctly earlier date of origin. But the crucial piece of documentary evidence adduced by the latter scholar has lately been impugned by G. Hilty, ‘Alguns aspectes de l'evolució històrica de la e i de la e en català’, Estudis Universitaris Catalans, XXIV (1980), Estudis de llengua i literatura catalanes oferts a R. Aramon i Serra, II, 231–37, at 233–35. 34. This assumption is confirmed by such frequent twelfth-century spellings as ded, ped, and pred: see Gulsoy, ‘El desenvolupament de la semivocal -w en català’, in Catalan Studies in Memory of Josephine de Boer, eds. J. Gulsoy and J. Solà-Solé (Barcelona: Borràs, 1977), 71-98, at 85. 35. Veny, Els parlars, 62–63, credits such archaic forms to Rossellonès and Balearic, while Coromines, El parlar, 49, chanced upon them in Pallars as well as in other Pyrenean dialects. See, further, entry for divendres in DECLC (s.v. dia) as well as relevant entries in Alcover–Moll, Diccionari. 36. The evolution of incendere and intendere in Catalan calls for further comment. After operation of syncope, an early change shared with French and Provençal, /nd/ of */ensndre/ and */entndre/ would have failed to undergo the specifically Catalan reduction to /n/, which applied only in intervocalic (and later in word-final) position. Thus, at the medieval stage, the root-medial clusters of encendre and entendre consistently contrasted with those of diven(r)res, gen(r)re and ten(r)re (see Badia, Gramática, 215, as well as relevant entries in Alcover-Moll, Diccionari). That contrast has been lost in most modern dialects, which either delete the /d/ in the former pair of words (Rossellonès, Mallorquí) or epenthesize a /d/ in the latter triad (Central, Western, Valencian). 37. Pope, From Latin to Modern French, 79, assumes that effacement of atonic -e predated geminate reduction in French. I see no reason to adopt a different chronology for Catalan. 38. Cf. Fr. /st/ ‘seven’ showing nondiphthongization of stressed vowel. 39. See Fabra, ‘Les e toniques’, 10 and Kuen, El dialecto, 111n. 40. French dialects vacillate between expected outcomes of E and E, while cognates in Italian dialects unanimously presuppose E, a fact that held considerable weight with Meyer-Lübke (REW), who projected the long vowel into the Celto-Latin base. See, in addition, C. Battisti and G. Alessio, Dizionario etimologico italiano, I (Firenze: Barbèra, 1950); and M. Cortelazzo and P. Zolli, Dizionario etimologico della lingua italiana (Bologna: Zanichelli, 1979–). 41. The analogical origin of the masculine form is inferrable from the existence of only the feminine in certain areas of Romania (e.g., Sardinia, Provence). The masculine is attested solely in the Iberian peninsula. See Alcover-Moll, Diccionari and REW 3 .