scieee Open visual document viewer

Moras or Skeletal Units? A Question of Parametric Variation

Lloret, Maria-Rosa

Abstract

Una de les qüestions que més s'han discutit recentment en fonologia és si les representacions fonològiques han d'incloure com a primitius mores o unitats de l'esquelet. L'objectiu d'aquest treball és demostrar que aquests dos elements no són mútuament exclusius, ja que existeixen processos fonològics que han de fer referència a ambdós per poder donar compte de tots els fets d'una llengua. De fet, aquesta dicotomia sorgí erròniament en confrontar la teoria moraica de Hayes (1989) amb la resta de teories de l'estructura sil·làbica. Ara bé, si tenim en compte que ni podem prescindir de l'esquelet ni la quantitat sil·làbica és fonològicament rellevant en totes les llengües, podrem establir un nou ordre de primitius fonològics que no es presenti com a representació de teories prosòdiques alternatives sinó que posi de manifest la variació paramètrica que les llengües presenten entre llengües d'estructura moraica i llengües d'estructura sil·làbica.

Full text

Ca alan Wo king Pape s in Linguis ics (CWPL) 1991: 149-165 Uni e si a Au hnoma de Ba celona - MORAS OR SKELETAL UNITS? A QUESTION OF PARAMETRIC VARIATION Ma ia-Rosa Llo e Uni e si a Au 6noma de Ba celona One majo issue ha is being add essed in ecen phonological esea ch is whe he he phonological ep esen a ion has o ecognize he exis ence o mo as o skele ai uni s as p imi i es. The p esen pape a gues ha bo h app oaches a e no necessa ily mu ually exclusi e, since he e exis s he possibili y ha some phonological p ocesses ha e o e e o bo h phonologicai en i ies in o de o accoun o all he ac s o a language. In ac , bis dicho omy w ongly appea ed by acing Hayes' (1989) mo aic heo y wi h o he heo ies o syllable s uc u e. Howe e , i one akes in o amun ha nei he he skele on can be dis ega ded no syllable weigh is always a phonological a iable, a new o de ing o phonological p imi i es can be sugges ed, which does no cmespond o al ema i e p osodic heo ies bu e lec s he pa ame ic a ia ion ha languages show be ween mo aic and syllablic s uc u e. 1. In oduc ion. One majo con ibu ion o non-linea phonology o he gene al phonological heo y has been he in oduc ion o he skele on (McCa hy 1979) as an in e media e le el be ween syllabic cons i uen s and melodic segmen s. (1) O O R ( = syllable A I O =onse N C R = ~hp N = nucbw I i i " X 7 7 X = skele sl uni ) g a 1 a [gal The independence o his le el wi h espec o he o he s plays an impo an ole in he desc ip ion o some segmen s (long owels, gemina e consonan s, complex segmen s) as well as in he jus i ica ion o ce ain phonological p ocesses ( o al assimila ions, compensa o y leng henings, eduplica ions). Fo example, he ep esen a ion o long segmen s as wo skele al uni s associa ed wi h he same melody cap u es and sa is ac o ily exp esses he ambiguous beha io o long owels and consonan s, which some imes ac as i hey we e wo segmen s while o he s as i hey we e a single en i y. In he Cushi ic O omo language, o ins ance, phonological p ocesses ha a ec quan i y, such as epen hesis, ea long consonan s as a sequence o wo segmen s: he epen he ic owel [i], which is inse ed o a oid iconsonan ic sequences, occu s a e wo he e ogeneous consonan s (2a) as well as a e a gemina e consonan (2b). (2) a. a g+na -> a gina ' we see' gon + a -> gon i a 'you (sing.) ado n' b.kenn+na -> kennina 'we gi e' mudd+ a -> muddi a 'you p ess' XX X X XX XX X X XXX II /+ II-II~~III k e n na ke n i n a [kennina] Howe e , p ocesses ha in ol e melodic changes, such as labializa ion, ea long segmen s as a single en i y: [w] sp eads i s labial cha ac e is ic o a p eceding owel, whe he his is sho (3a) o long (3b). (3) a. ilbiisaw+na -> ilbiisowna -> ilbiiso na 'we a e simple' b. lakkaaw+na -> lakkoowna -> lakkoo na ("lakkao na) 'we coun ' II /VI+II-II I II- 1 a k a wna la k O yna XX X X X XXXX .I LAB The p esence o a skele al ie is also p o ed h ough he analysis o phonological p ocesses such as compensa o y leng henings. In La in, o ins ance, dele ion o /s/ be o e an e io sono an s en ails leng hening o a p eceding owel (4). I one assumes he exis ence o a skele al ie , compensa o y leng hening can be uni e sally jus i ied as a sp eading p ocess ha epai s ill- o med s uc u es esul ing om he appea ance o emp y X posi ions, which o hen ise could no be inco po a ed in o he p osodic s uc u e. (4) *kasnus -> kaanus 'g ay' * kosmis -> koomis 'cou eous' XXXXXX XXXXXX XXX XXX ka snu s ka nus ka nus [kaanw] Al hough he p e ious da a seem o adequa ely jus i y he p esence o a skele on, du ing he las decade an al e na i e p osodic ep esen a ion has been p oposed (Hayes 1989, among o he s), acco ding o which he skele al uni s (X) and he syllabic cons i uen s (O, R, N, C) a e eplaceú by mo as (5). Using his app oach, mo as do no me ely ep esen a segmen (as X did) bu play wo di e en oles: hey coun as phonological posi ions (i.e. hey a e phonological p imi i es) and also ep esen he con as be ween ligh (o monomo aic) and hea y syllables. The de ende s o such an app oach ega d his heo y as supe io o he one ha claims he exis ence o a skele on and syllabic s uc u e, because mo aic ep esen a ion does no only cap u e he duali y o some segmen s (e.g. long segmen s) bu jus i ies he phonological p ocesses ha in ol e syllable weigh (e.g. compensa o y leng hening) be e han p e ious app oaches. Fo exa nple, mo aic heo y explici ely cap u es he c oss-linguis ic gene aliza ion ha only segmen s dele eú in coda posi ion igge leng hening o an adjacen segmen . Thus, in La in dele ion o /s/ en ailed compensa o y leng hening in coda posi ion (4 and 6a) bu no wo d-ini ially (6b), because in his posi ion /s/ does no ha e mo aic alue. No e ha in o de ha he heo y ha assigns skele al posi ions o all he segmen s accoun s o he da a in (6b), one has o idiosync a ically s ipula e ha melodies only sp ead o syllable- inal emp y posi ions, so ha he emp y wo d-ini ial X posi ion is no illed by sp eading bu dele ed h ough con en ion because o i s s anded cha ac e (6b'). In mo aic heo y, howe e , his is independen ly mo i a ed because only syllable- inal consonan s make hei syllable hea y, i.e. hey ha e mo aic alue. (6) a. *kasnus -> kaanus 'g ay' * kosmis -> koomis 'cou eous' b. *snu us -> nu us 'daugh e -in-law' sme eo -> me e0 'I dese e' C . b' O n + A P " N N A c X XXXX IIIIY XXXX X I s n u u s 0 n u u s 9 1111 I 1 As a consequence o his eanalysis o he p osodic s uc u e, wo al e na i e p osodic heo ies a e p esen ed nowadays in he li e a u e: he heo y o syllable s uc u e , whose p imi i es a e he skele al uni s (X), he syllabic cons i uen s (O, R, N, C), and he syllables (O), and he mo aic heo y , which eplaces he skele al uni s and he syllabic cons i uen s by mo as (p). (7) P osodic heo ies: a. Theo y o syllable s uc u e (X; O, R, N, C; O) b. Mo aic heo y (p; O) The aim o his pape is o p o e ha bo h heo ies a e no necessa ily mu ually exclusi e bu hei di e ences co espond o some kind o pa ame ic a ia ion ha languages show. In gene al, he dis inc ion es ablished by mo aic heo y be ween ligh and hea y syllables u ns ou o be c ucial (i.e. phonologically ele an ) o languages ha show a phonemic con as be ween sho and long owels, as i was al eady poin ed ou by T oube zkoy (1939).1 Howe e , syllable weigh is no a phonological a iable in languages wi hou phonemic owel leng h. In hese languages, ins ead, he in e na1 s uc u e o he syllabic cons i uen s becomes ele an . The a icle is o ganized as ollows. We i s p o ide da a o suppo he claim ha he dicho omy p esen ed in (7) does no ma ch he eal ac s o he languages. Sec ion 2 jus i ies he le el o skele al ep esen a ion e en in languages wi h well-mo i a ed mo aic s uc u es while sec ion 3 shows ha he ole o he skele on can no be aken o e by melodic s uc u e ei he . 154 Wi h his done, we sugges a new o de ing o phonological p imi i es om which he pa ame ic a ia ion ha languages show can be conec ly es ablished (sec ion 4). 2. Skele on and Mo as. Hayes (1989) claims ha skele al uni s a e empi ically unnecessa y because compensa o y leng hening does no ake place sys ema ically whene e a segmen dele es lea ing an emp y p osodic posi ion behind, bu only comes along in ce ain posi ions, which ,no acciden ally, always ha e mo aic alue.2 In O omo, howe e , he e exis s a case o compensa o y leng hening ha can only be jus i ied wi h he p esence o he skele ~n.~ O omo shows a phonemic con as be ween sho and long owels and, hus, syllable weigh unc ions as a phonological a iable. The language ea s CV and CVC syllables as monomo aic o ligh , while CVV and CVVC syllables a e ea ed as hea y o bimo aic.4 (8) O om: Ligh o monomo aic syllables: [CVIp, [CVCIp Hea y o bimo aic syllables: [CVIp [VIp, [CVIp [VCIp This di ision be ween he wo di e en ypes o syllables u ns ou o be c ucial, o ins ance, o jus i y he beha io o such syllables in ela ion o he gene al cons ain ha he language shows on he syllable s uc u e o he wo ds, namely O omo does no allow long owels in wo successi e syllables unless he second is also wo d- inal. I conca ena ion o mo phs pu s oge he syllables wi h long owels (9a), he owel o he second syllable obliga o ily sho ens. This cons ain , howe e , does no conce n CVC syllables (9b). (9) a. gaal+oo a -> gaa 1 o a 'camels' (c . nam+oo a -> na moo a 'men') a caal+oom- -> a caa 10 m- 'be kind' (c . ko?+oom- -> ko ?oo m- 'be busy') b. a d+oo a -> a doo a 'ho ses' lugn+oom- -> lug noo m- 'be cowa d' Wi h hese p ecisions es ablished on he weigh o he O omo syllables, le us now look a he peculia compensa o y leng hening p ocess ha he language shows. O omo has a dele ion p ocess ha e ases he melodic con en o a glo al consonan (i.e. glo al s op /?/ and implosi e lc I) in s yllable- inal posi ion. This dele ion en ails leng hening o he p eceding owel ( 10). Acco ding o Hayes' mo aic heo y, his kind o dele ion should no yield compensa o y leng hening because he dele ed glo al consonan does no lea e an emp y mo a behlnd (10a). (Remembe ha in O omo inal consonan s do no add weigh o he syllable, i.e. hey ha e no mo aic alue.) Howe e , he da a p o e ha such leng hening occu s, which can only be jus i ied by he p esence o a skele on (10b). (10) ec +na -> eena 'we wish' a?+ a -> aa a 'you push' b. XXXXX XXXXX XX XXX ec a e a e 1 a [ ee a] Since O omo seems o equi e he p esence o a skele on, one could hink ha he language does no need o e e o i s mo aic s uc u e bu , ins ead, i s syllable s uc u e would unc ion as a phonological a iable. We will nex show, howe e , ha syllable s uc u es a e unable o accoun o he O omo ac s. This will p o e ha in O omo, as in any language wi h phonemic owel leng h, syllable weigh is a phonological a iable, which is be e exp essed h ough mo aic ep esen a ion. I we assume ha O omo has he syllable s uc u e shown in (1 l), we co ec ly cap u e he O omo syllable di ision be ween ligh CV(C) and hea y CVV(C) syllables: he hyme o ligh syllables consis s only o nucleus while he hyme o hea y syllables consis s o nucleus and coda. Howe e , he ep esen a ion o (1 1) canno jus i y he compensa o y leng hening p ocess (CL) ha he language shows, because he dele ed inal consonan occupies a posi ion di e en om ha occupied by he second pa o a long owel, na nely he o me is included wi hin he nucleus while he la e cons i u es he coda, and he e is no eason o claim ha he language has wo di e en ypes o long owels. ligh syiiables ka- syllables The syllable s uc u es ep esen ed in (12) and (13) a e also un enable o O omo. In his case he compensa o y leng hening p ocess can be adequa ely ep esen ed because he owel leng hened as a consequence o he dele ion o a inal consonan esul s in a long owel wi h he same s uc u e o unde lying long owels (i.e. a complex nucleus in (12) o a b anching hyme in (13)). No e, howe e , ha in hese cases he di ision be ween ligh and hea y syllables is inco ec ly es ablished. A i s e sion o his pape was p esen ed a he Pnme Coloquio de G amA ica Gene a i a held a Mi a lo es de la Sie a, Mad id, Spain, 20-22 Ma ch 1991. The wo k bene i ed om a Pos doc o al Resea ch G an and he PB 89-0344 C.1.C.Y .T. G an . No es 1 See Newman (1972) o a gene al s a emen on he impo ance o he syllable weigh concep . 2 See he p e ious La in example in (6). 3 The ollowing O omo da a a e aken om Llo e (1988). 4 No e ha O omo con as s wi h La in in ea ing CVC syllables as ligh . 5 The e a e only wo excep ions, which ha e o be lexically ma ked: (i) k1i i '+ii -> k'i i ' ii (*k i i ' 'ii) 'chisel' ( ' = ejec i e o glo alized ) (ii) hiik+ii -> hiik ii (*hiikkii) 'ex ension' 6 This phenomenon is adi ionally e e ed o as liaison . 7 The o he p osodic heo ies a e segmen al in he sense ha he numbe o p osodic segmen s in ui i ely w esponds o he numbe o melodic segmen s. Re e ences Clemen s, Geo ge N. and Jay Keyse (1983) CV Phonology , M.I.T. P ess, Ca nb idge, Massachuse s. Hayes, B uce (1989) "Compensa o y Leng hening in Mo aic Phonology", Linguis ic Znqui y , 20, 253-306. Hyman, La y (1985) A Theo y o Phonological Weigh , Fo is Publica ions, Do d ech . I 6, Junko (1989) "A P osodic Theo y o Epen hesis", NLLT ,7, 217-59. Kaye, Jona han D. and Jean Lowens amm (1984) "De la syllabici é" in F. Dell, D. Hi s , and J .-R. Ve gnaud (eds.) Fome sono e du langage: s uc u e des ep ésen a ions en phonologie , He mann, Pa is. Llo e , Ma ia-Rosa (1988) Gemina ion and Vowel Leng h in O omo Mo phophonology , Ph.D. disse a ion, Indiana Uni e si y, Blooming on, Indiana. McCa hy, John (1979) Fo mal P oblems in Semi ic Phonology and Mo phology , IULC, Blooming on, Indiana. McCa hy, John and Alan P ince ( o hcoming) P osodic Mo phology , M.I.T. P ess, Ca nb idge, Massachuse s. Newman, Pau1 (1972) "Syllable Weigh as a Phonological Va iable", S udies in Apican Linguis ics ,3,301-23. T oube zkoy, N. S. (193911967) P incipes de phonologie , Klincksieck, Pa is. UAB Depa amenl de Filologia Ca alana 081 93 Bella ema (Ba celona)