Sie sind auf Seite 1von 2

Allomorph Selection and Lexical Preferences in Judeo-Spanish Travis G.

Bradley, UC Davis Migration is known to favor the creation of weak ties within social networks, which can in turn promote linguistic change (Milroy & Milroy 1985, Milroy 1987). Despite conventional descriptions as conservative and archaic, Judeo-Spanish (JS) shows many innovations vis--vis Modern Spanish (MS) due to dialect mixing and competition among variant linguistic forms that arose in Sephardic communities after their expulsion from Spain in 1492 (Penny 1992, 2000). The present study analyzes two such cases of variation in phonologically conditioned allomorphy. One consequence of dialect mixing was competition between variants from central areas of the Peninsula in which Latin tonic // and // had given rise to diphthongs [je] and [we] and those from lateral areas which preserved the monophthongs (Penny 1992:253, 2000:188-9). In many cases, variation can be observed within the same lexeme and the same dialect. The JS forms in (1a,b) show free variation between diphthongs and mid vowels, where the corresponding MS forms show only the diphthong. In other cases, as seen in (2a,b), an innovative alternation has been introduced in JS words whose Latin etymon did not contain tonic // or // and whose MS correspondents show only mid vowels. For some speakers, the vacillation between diphthongs and mid vowels has led to the analogical extension of [je] and [we] to morphologically derived words. The JS forms in (3a,b) have the diphthong regardless of stress, whereas the MS forms show an alternation between stressed diphthongs and unstressed mid vowels. Recent work in Optimality Theory has proposed that surface allomorphs are derived not from a common underlier but from an ordered set of allomorphs in the input (Bonet et al 2007). The constraint in (4a) demands faithfulness to the lexically specified ordering and favors the least marked allomorph. I adopt the constraint in (4b) as an informal label for whatever contextsensitive markedness constraints favor diphthongs in tonic syllables (Bermdez-Otero 2007:285; see also Holt 2007:384-5). I propose an account of the free variation in JS (1) in which (i) mid vowels are ordered above diphthongs in the set of input allomorphs, e.g., /d{e1>je2}nt-i/, and (ii) the constraints PRIORITY and TONICDIPHTHONGAL are variably ranked. When markedness dominates, the diphthong is optimal in stressed syllables (5c), while the opposite ranking favors the mid vowel (5e). High-ranking IDENT ensures phonological faithfulness within allomorphs by ruling out the e1je1 and je2e2 mappings in (5b,f) and (5d,h), respectively. The same analysis covers the data in (2) on the assumption that the JS speakers in question have extended the {mid vowel>diphthong} ordering to novel lexemes whose etyma lacked tonic // and //, e.g., /k{e1>je2}z-u/. Finally, the fixed ranking of TONICDIPHTHONGAL PRIORITY in MS derives the alternation between [djnte] (6c) and [dentsta] (6e). Although not shown here, the ranking of IDENT TONICDIPHTHONGAL is necessary in MS to preserve input mid vowels in non-alternating forms, e.g., terco, *tierco stubborn ~ terquedad stubbornness. JS speakers who level the alternation in (3a,b) have simply eliminated mid vowels from the set of input allomorphs for the relevant lexemes, rendering PRIORITY irrelevant. IDENT preserves the input diphthongs in [djnti] (7b) and [djentsta] (7d). The second case of phonologically conditioned allomorphy in JS involves diminutive suffixation. There was competition among variant suffixes in the 16th century, but by the 18th century iko/a had come to predominate, subject to certain phonological restrictions (Bunis 2003). For example, the preferred suffix iko/a appears when the final consonant of a stem is non-dorsal (8a), but after dorsals ito/a appears instead (8b). I propose that these allomorphic preferences are encoded in the lexical ordering {ik1>it2} and that the alternation in present-day JS is driven by the fixed ranking of (9) above PRIORITY. The preferred allomorph surfaces in (10a) but is blocked in (10e) by OCP(DOR), whereby the next best allomorph emerges as optimal in (10g). The ranking of IDENT OCP(DOR) is necessary to preserve input /k/ in non-alternating suffixes, e.g., JS monarka monarch ~ monarkiko, *monarkito monarchist.

(1) (2) (4)

a. b. a. b. a.

JS MS JS djnti ~ dnti djnte tooth (3) a. djnti kwpu ~ kpu kwpu body djentsta kzu ~ kjzu kso cheese b. bwnu plvu ~ pwlvu plo dust bwend PRIOR(ITY) Bonet et al (2007:906) Respect lexical priority (ordering) of allomorphs. b. T(ONIC)D(IPHTHONGAL) Bermdez-Otero (2007:285) Diphthongs are favored in stressed syllables.

MS djnte dentsta bwno bond

tooth dentist good goodness

(5) /d{e1>je2}nt-i/ IDENT TD PRIOR a. d1nti *! b. dj1nti *! c. dj2nti * d. d2nti *! * * (6) /d{e1>je2}nt-e/ IDENT TD PRIOR a. d1nte *! b. dj1nte *! c. dj2nte * d. d2nte *! * * (7) /djent-i/ IDENT TD PRIOR a. dnti *! * (n/a) b. djnti (n/a) (8)

/d{e1>je2}nt-i/ IDENT PRIOR TD e. d1nti * f. dj1nti *! g. dj2nti *! h. d2nti *! * * /d{e1>je2}nt-ista/ IDENT TD PRIOR e. de1ntsta * f. dje1ntsta *! * g. dje2ntsta * *! h. de2ntsta *! * * /djent-ista/ IDENT TD PRIOR c. dentsta *! * (n/a) d. djentsta * (n/a)

(9)

a. lto altko tall kamno kaminko path rza rizka laugh pao paako bird b. blko blakto, *blakko white ma mita, *mika crumb fwa fawta, *fawka forge OCP(DOR) Avoid adjacent segments on the consonantal tier that are identical in DORSAL place. /blak-{ik1>it2}-o/ IDENT OCP PRIOR e. blakk1o *! f. blakt1o *! g. blakt2o * h. blakk2o *! * *

(10) /alt-{ik1>it2}-o/ IDENT OCP PRIOR a. altk1o b. altt1o *! c. altt2o *! d. altk2o *! *

Selected References Bonet, Eullia, Maria-Rosa Lloret, & Joan Mascar. 2007. Allomorph selection and lexical preferences: Two case studies. Lingua 117.903-927. Bunis, David. 2003. Ottoman Judezmo diminutives and other hypocoristics. Linguistique des langues juives et linguistique gnrale, 193-246. Paris: CNRS ditions. Penny, Ralph. 1992. La innovacin fonolgica del judeoespaol. Actas del II Congreso Internacional de Historia de la Lengua Espaola. Tomo II, 251-257. Madrid: Pabelln de Espaa.

Das könnte Ihnen auch gefallen