6z THE LINGUISTIC HERITAGE OF SPAIN

Notes Chapter 3 r. The latter author claimed to be following the lead of cuervo (r9or) in his anti-Andalusian stance, but Guitarte (r958) has shown that Cuervo's position was much more cautious. Beforeand afterSpain - the NativeAmerican z. Sephardic (Judeo) Spanish, reflecting Peninsular speech trends of the turn of the sixteenth century, including castile and other non-Andalu- contribution sian zones, shows that seseowas already a viable process in Spain. It is not seseobut rather the particular realization of the former o = 1s1to [gJ in Castile which is the unusual development. 3. The simplistic notion that Castilians settled in the highlands and Andalu- sians stayed in coastal lowlands cannot be adequately substantiated, but many developing communities did receive a major proportion of immi- grants from a single region of Spain. For example Granda (r979a) has suggested,based on historical reconstruction, that the retention of l^t in Paraguay may be at least partially due to the high concentrations of settlers from the northern Basque provinces of Spain. Another regional distribution which merits further study is the use of the diminutive suffix -ico, characteristic of Arag6n, and widespread in cuba, costa Rica and lntroduction Colombia. to the Caribbean,native peoples.of Most dialects of Spanish allow non-inverted questions During Columbus'svoyages 4. when the inter- the first recordedlinguistic contacts with rogative element is in non-argument position (cf. Torrego 1984), i.e. as the Americasestablished 'How encounterednew flora, fauna, peo- an adverbial adjunct: iC6mo Juan pudo lograr eso? did John Europeans.Spanish speakers manage to do that?', qud momento ustedesse dieron cuenta que ples, cultures and meteorologicalphenomexal together.with the 'At iEn de lo pasaba? what point did you become aware of what was happening?' lerms used to describethem. Words of Caribbean origin were when the interrogative element occupies an argument position (subjeit, carriedthroughout Latin Americaby subsequentexplorers, where direct object, etc.), general spanish requires inversion. In the caribbean they often displacedlocal equivalents.Aii'pepper', hamaca'ham- 'hurricane', 'canoeo, ocorn', region, non-inverted questions routinely contain interrogative elements motk', huracdn canoa maiz mani in direct object position, but only when the subject is a pRoNoMTNAL; 'peanut' and many other words are now usedthroughout C_entral *LQui Juan quiere?'what does John want?'is not an acceptablesentence. and South Amerita, as well as in Spain. Writers such as Bernal There are additional constraints on non-inverted questions, even in the Diaz de Castillo,Garcilaso de la Vegaand Columbushimself were caribbean area. For example it is not usual for adverbs, negative items introducingresidents of Spainto Native American or other elements to intervene between the subject pronoun and the instrumentalin *iQui visited the verb: tu no quieres?'What don't you want?', *Leu6 trt veces lexical items, and the prestigeassociated with having 'what a piensas'! do you sometimes think?', etc. If object clitics intervene Americasinduced many Spaniardsto deliberatelyuse americanis' between the pronominal subject and the verb, the acceptability is usually mos. Cewantes,Lope de Vega and Quevedoare amongthe many higher, although not as high as when nothing separatesthe subject from Spanishwriters who brought suchwords to a wider public. 'what _tnJttg the verb: iQut til le dijiste? did you tell him/her?' These facts American colonies,the incorporationof Native Americanlexical suggest that in the Caribbean dialects, subject pronouns are acting as items was naturally more extensive,as blendingof Europeanand phonological clrrrcs, motiviated by the high rate of retention of Americancultural and socialpractices created the needfor words subject pronouns, in compensation for the loss of final consonants and to describehitherto unknown concepts.Spanish American place the concomitant erosion of person/number differentiation on verbs. namesalso reflectthis mixture; a frequentcombination consists of a Spanishsaint's name and an indigenousplace name: Santa Fe de Bogot6,San Francisco de Quito, SanMiguel de Tucum6n,etc. is no Aside from indigenouslexical items and toponyms'there ,/ consensuson the effectsof Native Americanlanguages on Spanish. The Spanishof Latin Americais widelyvaried, including configura- 64 I]EFoRE AND AFTER SPAIN - THE NA].IVE AMERICAN CoNTRIBUTIoN THE NATURE OF SPANISH INDIGENOUS CONTACTS 65

tions not attestedin Spain. ln pronunciation and syntax, many linguistic Latin American dialectspresent systematic innovations which are The nature of Spanish-indigenouscultural and not easy to explain away as linguistic:drift, the inheritance of contacts Spanish settlers,or borrowirrg from neighbouringdialects. Particu- During the sixteenthcentury (often regardedas the formative larly in areaswhere the indigenous population has remaineddemo- period for Latin American Spanish)and even indigenous graphically and prominent, _later, ethnically it is not unreasonableto populationsoften outnumberedEuropeans by hundredsto one, that some unique features lbuppose of regional Spanidhdialects are yet the natureof Spanishsettlement was not alwaysconducive attributable to prolonged and contact with indigenouslanguages. Hen- influences.In order for an indigenouslanguage,r' riquez (tgzr) to substratum Ur-e-na went so far as to divide all of Latin American permanentlyto influencecolonial Spanish,a specialset of condi- Spanishinto dialectzones based prevailing on indigenoussubstrata: tions was required,which werenot presentin all coloniesnor at all Nahuatl, carib/Arawak, Quechua,Mapuche/Araucano, and rupi/ times. Guarani. As discussedin chapter r r, this approach is inadequite For an indigenouslexical item to enter the Spanishvocabulary, for severalreasons, not the least of which is demographic iniccu- it is merelyneCessary for a colonistto askthe nameof an unknown racy. In most o[ the dialect zonespostulated by Henriquez urefra, object.This can be done regardlessof the linguisticskills of either differencesamong regionalvarieties of Spanishoutweigtrfundamen- individual and doesnot necessarilyentail any further interaction. tal similarities. In the caribbean, for example, tft indigenous Columbuslearned some indigenous words (not alwaysaccurately) populationsrapidly disappeared,and had little effecton the develop- by meansof gestures,and later through native.interpreters whose ment of Spanish. In venezuela, Argentina, uruguay and chile, commandof Spanishmay havebeen only partial. The sameeffect hostile indigenouspopulations were pushed ever further from Span- was producedwhen Spaniardsacquired the rudimentsof an indig- ish settlements.[n much of colombia and costa Rica, and in ..o,.tt language;Spanlsh missionary activity was a major veli.cl-e coastal Peru and Ecuador, Spanish settlershad minimal contacts this type of transfer.Finally, an indigenouspopulation which with the indigenous_population. paraguay, for This contrasts with had acquired fluency in Spanish could continue to contribute Mexico and the Andean countries, in which indigenous languagis lexicaliiems to Spaniihcolonists. without any other characteristics remain vigorous even today, and where many Spanish colonists of the nativelanguage being carried over. learnedand usedthe native tongues.Tracing the indigenousimprint Native Americani who use Spanishonly occasionally,having in Latin American Spanish requires a finer-grainedapproach ihan learnedit as a secondlanguage past childhood,speak an interlan- Henriquez urefra's broad categories,enriched by a deeperknowl- guage in which the phonology, morphology and syqlax of the edge of spanish dialect variation than that possessed6y the pio- iative languageare superimposedon Spanishpatterns. Today such neersof Latin American dialectology. speechcatt be heardin indigenousredoubts throughout the Amazon Few claims of indigenousinfluence have been accompaniedby a Iiasin, the Andesand Mesoamerica;in the past,it existedin nearly demonstration of the purported substrate patterns, nor of ihe everySpanish colonial settlement. Even when Spanishis usedon a I opportunity for bilingual interlanguage to percolate upward into daily b-asis,between workers and employers,or betweenrural regional dialects of Spanish. Too often, the mere demographic residentsand priests,fluency may neverrise above the level of a presenceof a large indigenousor mestizopopulation has unciitiially rough pidgin. Such indigenously-flavouredSpanish has no ready been taken as the source of 'peculiarities' of a given dialect zone, wai of .*pu*ding beyond the group which createdit, and without verifying either the viability -has of the hypothesisin linguistic ordinarily fuavesno traceson nativelyspoken Spanish. In orderfor and historical terms, or the existence of alternative explanitions. an indiglnous interlanguageperrnanently to penetrateregional The case for an indigenous influence on non-lexical leatures of varietieJof Spanish,a major sociolinguisticshift must break the Latin American Spanish must presented be as in a court of law, equilibrium*hich sustainsthe interlanguage.Speakers of the intet- demonstratingmotive, method tr and opportunity. languageneed to occupypositions in which their speechbecomes the norm. Suchspeakers must be presentin greatenough numbers to makethe interlinguage demographically prominent. Theinterlan- guageitself, by definitionthe result of havinglearned Spanish as a - 66 BEFoRE AND AFTER sPAIN THE NATIVE AMERICAN CoNTRIBUTIoN THE NATURE OF SPANISH INDIGENOUS CONTACTS 67

secondlanguage, must graduallybecome a first language,without speakerlacking any knowledgeof Quechuastructures, and whose sheddingthe indigenousaccretions. This requiresinsulation from gru-.ut inclu?esbnly the combinationla casade Maria, will be normativestandards, or a socialenvironment in which suchstand- it a disadvantagein termsof rapid interpretation. ards are no longerrelevant. The permanentinsertion of indigenous As long as racial and cultural boundariesrigidly separateindlq- elementsinto regionalSpanish follows the samepattern by whictra enous and Eutopean communities,nativized indigenousSpanish pidgin, originallya survival-levelcontact language spoken natively existsas a stabtebut closedsociolect. Usage may be fluent, but by no memberof a linguisticallyhetereogenous population, evolvei contactbetween indigenous and EuropeanSpanish occurs only on into a creole,learned as a native language.As with creolization,a the peripheryof eachgroup, and the indigenousSpanish does not myriad of differentevents can leadto the sameresult. Holm (r9gg: boundariesimposed by the colonialadministra- 'semi-creole' ur yit trinscendthe has coinedthe term to refer to a arises,characterized by a rangeof socioeco- 9) 'both languagevaiiety tion. As a mestizoclass which has creoleand non-creolefeatures but . . .-doesnoi nomic opportunitiesnot generallyavailable to the indigenousJom- necessarilyimply that they wereever basilectalcreoles, since both munity,-Spanish. lndo-spanish comes into more intimatecontact with Euro- creolesand non-creoles. . . can becomesemi-creoles by borrowing pean When mestizaiebecomes defined by behavioural features.'A semi-creolehas neverundergone the radicalpidginizal patterni and economicpower rather than by simpleracial criteria, tion + renativizationwhich characterizescreole formation, but has *ore and more membersof the indigenouscommunity enter the been significantly restructuredwith respectto the original super- mestizolinguistic sphere,creating a fluid sociolinguisticspectrum 'white' stratelanguage. Black American Englishand vernacularBrazilian whosepolai extremisare still and'Indian',but themajority Portugueseare possibleexamples of semi-creoles. varieties.'If structuralchanges A similar result of whoie speakersuse intermediate 'Euro- would presumablyobtain when Europeanspanish was partially in societyor simpledemographic predominance undermines restructuredunder appropriate conditions of contactwith i Nativl pean' Spanishai the prestigestandard, features of the formerly Americanlanguage. In this vein, it is instructiveto considersome ;indigetrbus'sociolect may cometo be sociolinguisticallyunmarked, possiblescenarios. i.e.will be acceptedas the new standard. In a long-standinglanguage contact environmentsuch as that There exist ieveral ways for an originally marginal indigenous found in the Andean region of South America, the indigenous Spanishsociolect to achievepredominance as a standarddialect. A populati.ongradually acquiredspanish as a native language,not key ingredientis a discontinuityin the sociolinguisticevolution of necessarily through the loss of the first language,uut via an a liven region,whereby the speechof previouslyneglected sectors increasinglyspanish-dominant bilingualism. In a iacially and so- oisociety is moved into a more prestigiousposition. A social cially segregatedenvironment such as existedin colonial Latin upheavalcan have this effect: the Bolivian revolution of rg5z, America, Spanishis usednot only for essentialcontacts with the afthoughby no meansdisplacing the European-basedaristocracy population of Europeandescent, but also among membersof the and its monolingualSpanish norms, facilitated the penetratationof same indigenouscommunity. Mestizos provide a bridge between indigenous-influencedsociolects in higher spheresof government, the cultures, and facilitate languagetransfer and the dJvelopment eduJation,mass media and commerce.The Cuban Revolutionof of a stableethnic interlanguage.Indigenous patterns freely enter rg5g produced,among other effects,a linguistic shift away frgm the Spanishof thesebalanced bilinguals, and unimpededcommuni- sfiiied" European-soundingSpanish for official and mass media 'Cuban' cation is ensuredby the fact that all bilingual speakerswill implicitly purposesto a more speechpattern. A similareffect ensued draw on the same_indigenous linguistic patterns when interpreting Outitg the Sandinista Revolution of Nicaragua, which achieved innovative spanish structures.For example,the euechua speakei po*ei itt 1979 and whose legacy is still presentin Nicaraguan who says iducational,iiieraryand political language.In colonialtimes, social separatist movementsoften Mariya-x wasi -n revolutions were not viable, but 'Peninsular' speech,once admired and Maria-POSS house-POSS'Mary's house' achievedthe same results. imitated, fell out of favour, and nationalist sentimentsinfused in Quechuawill instantly recognizethe Spanishphrase de Maria su criollowords and expressionswith suddenvigour. casa,which follows an identicalorder of major constituents(G6mez The role of indigenouswomen as wivesof Spanishmen, and as Bacarrezaand Ar6valo Soto 1988:z9). The monolingualspanish domesticservants ind nursemaidsfor children of Europeanorigin, - INDIGENOUS INTERI-ANGUAGE. PAST AND PRESENT 69 68 BEFoRE AND AFTER SPAIN THE NATIVE AMERICAN CoNTRIBUTIoN exterminated,leaving only sketchylinguistic remains.In the Amazo- provides another means for indigenous-influenced Spanish to nian Basin, Spanish settlement has been relatively recent, and achieve wider circulation. For much of the colonial period, men linguistic assimilation of the indigenous population is partial at greatly outnumbered women among European settlers, especially best,while in many highland Andean regiotrs,as well as in isolated in zones where prolonged military action was required. This re- rural zonesof Mexico and Guatemala,the Spanish-speakingpopula- sulted in ethnically mixed unions, in which an indigenous woman tion was never large enough for Spanish to to be implanted as a whose command of Spanish may have been far from complete viable native language. Long-lasting and stable contacts between becamgthe primary care-giverand linguistic role model for'span- Spanishand indigenolrslanguages took placewith Native American ish' children. The extreme case occurred in paraguay, where the communities which had achieveda dominant regional status,and extensivebilingualism among Paraguayanswho 'European' claim could maintain their demographic strength and some social and descenthas been attributed to the colonial family structure. The economic power. This includes part of Paraguay and much of the prototypical Paraguayan family consisted of a spanish-speaking former Inca and Aztec empires. father and a Gua':ani-spea-kingmother, with the latter originatin! Paraguayans' preference for Guarani in intimate, personal and familiar domains (cf. Service 1954).Moreover, many spanish men Indigenous interlanguage, past and present fathered children by several(indigenous) wives, who in turn raised Indigenous approximations to Spanish have varied over time, as the offspring, imparting Guarani as the dominant language (cf. have Spanish speakers'reactions to such interlects. In contempo- Roett and Sacks r99r). The Paraguayan case may havJ been rary Latin America, bilirrgual speakersstill produce configurations overstated, but families such as thJse just ', described were docu- which diverge sharply from monolingual Spanish usage.With few mented for many parts of colonial Spanish America. Children of exceptions,these contemporary interlanguagesdo not corresponrl mixed marriages,more often than not recognized by their fathers to any well-establishedmonolingual or fluent bilingual variety of and hence by the Spanish government and the church. were Spanish.The recessiveSpanish of bilinguals in, for example,Para- consideredcriollos and entered colonial society at a much higher guay, Mexico, Guatcmala and the Andean highlands contains level than was possiblefor membersof the indigenous community. syntactic and phonological traits which have not seepedinto the This provides an additional mechanism wheieby an ethnicaliy correspondingregional varietiesof Spanish.A continuum of Span- flavouredSpanish could becomethe native language of the..onorni- ish variants with iircreasinglyless indigenous influenceis found in callyand sociallydominant population. such bilingual areas, similar to the continuous variation whiclt Even when sanctionedmarriage between Spanishmen and indig- existedduring earlicr periods. enous women did not occur, the linguistic behaviour of servanis In order to judge the potential for indigenous penetration of and nursemaidscould exercisea powerful influence in reshaping regional Spanish phonology and syntax, it is useful to consider prestigenorms. Usually only the wealthiestfamilies entrusted fttrii specimensof Spanish interlanguagespoken by Native Americans. children to membersof the indigenous population, but the forma- Indians' approximations to Spanishhave beenrecorded, not always tive effects of the indigenous penetration can be considerable.If accurately, since the earliest colonial times. Most texts frorn the this top-down percolation is combined with influenceswhich have past were written by Europeans who rarely adopted a sympathetir: crept upward into the middle sociolinguistic strata from mixed- attitude toward indigenouscultures, which they regardedas inferior ethnic marriages, a_model is provided by which 'substratum' an indigenous to their own. They frequently equateddifficulties in learning Span- can affect a large community without the neid for ish with native intelligence. Indigenous writers, and those who eithera profound socialupheaval 'Indian' or an overwhelmingdemographic imitation of predominance. respected:native culture, carefully avoided Spanish,although their own command of Spanish nray have been In colonial Spanish America. the conditions described above influenced by a Native Americau language. As with the written were not always met. In the Antilles, for 'Indian' example, the native representationsof Africanized Spanish,imitations of Span- population disappearedafter only the briefest ethnolinguisticinter- ish are not necessarilyinaccurate just becausethey reflecta negative action with Spanish,providing for the transfer of lexicll items but attitude toward the speakers. having no other effect. Native Americans of Argentina, uruguay, Chile, and parts of central America, were also driven u*ui und - 70 BEFORE AND AFTER SPAIN TTIE NATIVE AMERICAN CONTRIBUTION ANDEAN SPANISH 7r

Andean Spanish more indicative of prejudicethan legitimateexamples of proto- In the Andean region, the Spanish conquerors came into contact AndeanSpanish. with a viit que.h.ru-tp.uking empire, whose constituents were The sevinteenthcentury Peruvianmestizo Felipe Guamin Poma sometimes allie'i,"Siiriietimesfoes, but never far removed from the de Ayala, in his own writings, givesindirect indicationsof vocalic Quechua language. Spanish religious orders adopted and standard- modificationsin indigenousSpanish, as well as occasionalerrors of ized Quechua as a lingua franca, extending it to peripheral areas concordance,use of uninflectedlo as direct objectmarker, elimina- and displacing still viable regional languages(cf. Mannheim r99r). tion of directarticles, and other proto-Andeantendencies (Rivarola From the outset, Quechua began absorbing large quantities of 1988):este mes esta la comidamaduro'this month the cropsare Hispanisms, and entered into a symbiotic los dichosyndios, que todo lo sabeDios y como relationship with Spanish. ripe', multiplicaron 'the The often imperfect mastery of Spanish by Quechua speakersdid podriroto to puede tener aparte esta gente de yndios aforesaid not escape the notice of Spanish settlers, who at an early date indians mu[tiplied, as known by almighty God who keepsthese 'Indian' began the imitations and parodies of Spanish that have Indians apart;,nosotros quisieramos ir a Chuquisacao Lima alcanqar continued up to the present. Rivarola (t987, 1988)has documented algfinproiisiiln para descansode pobres indios'We would like to go the early history of Andean Spanish, including the parodies of to Chuquisacaor Lima to obtain somefood to sustainthe poor Juan del valle caviedes, who imitated both criollo and indigenous Indians'.These unconsciously introduced hints of AndeanSpanish, Spanish of Peru towards the end of the seventeenthcentury. Cav- moreso than the deliberateimitations by Europeanwriters, indicate iedes observed,disapprovingly, the loss of the opposition ltl-lyl that Andean characteristicscould have been used by membersof among the Lima criollos, and mocked the lengua de indios: the indigenousintelligentsia, who employedSpanish as a vehicleof literary and political expression,and who becauseof their he- Balcail diablo,gorgobado gemony ovet large numbers of indigenousresidents may have queosasti tambi6n ti casas parallelindigenous Spanish dialect. ganga iowed ihe first seedsof a sin hallar in so doti patterns in interlanguage sinos6lo mojiganga. A study of the grammatical found givesa hint of the potentialfor large- jonto varietiesof AndeanSpanish Pariciosti al novia suchinterference has occurred only at the tan ridondoy ella scaletransfer. In reality, larga in areaswhere biling-ualismis comoin los trocosdi juego peripheryof monolingualSpanish; taco.bola in mismacama. ihe norm rather than the exception(i.e. outside of a few large urban areas),phenomena which were oncepart of the indigenous Ella dio el si con so tiple, sometimes osti con voz retumbada. interlanguagearc now found in middle-classSpanish, qui los gorgobadossiempre euenas usedby monolingualSpanish speakers. Cerr6n-Palomino hablandintro dil tinaja. (tg76), Hardman de Bautista(1982), Laprade (r98r), Rivarola 1r9aA;,Stratford (1988) and othershave demonstrated unequivocal These satirical verses do not exemplify the typically Quechua-influ- calquingfrom Quechuaand AYmara. enced word order frequently found in Andean Spanish, but do One iase is the use of the pluperfecttense to indicatereported demonstrate characteristically Quechua morphological and phono- eventsrather than personallyexperienced events, an innovativeuse logical transfer. In the Caviedes text, simplification of the Spanish which createsin Spanisha distinctionfound in Aymara and Que- five-vowel system to approximate the three vowels of Quechua is chua.This usageis found in bilingual Spanishof Bolivia and Peru. shown in osasti ) usastt, ti I te, parici < parece,jonto < junto, The use of postposednomds and siempre,common in Bolivian trocos highlandSpanish, is sometimesfound in other regions,for example poems is the erosion of noun-adjective concordance (al novia, dil Mexico. Postposedpero, however,as in ivas a tomar caft, pero? tinaja), and occasional lack of articles (in misma cama). This 'You're not going to have coffee, then?' is not found in non- stereotypical modification of mid and high vowels (known as Andean regions.Curiously, this leastSpanish-like of all Altiplano motosidad) has continued until the present time in the Andean 'particles' is also the least like its Aymara counterpart 'ra(i region, but when not accompanied by other indications of an (Laprade r98r: 2rg), and may come not from direct syntactic indigenous interlanguage, examples such as the ones just cited are transfer but rather from an origin as a topicalized or right- - 72 BEFOREAND AFTER SPAIN THE NATIVE AMERICAN CONTRIBUTION

EL SALVADOR IJ dislocated comment in the indigenous interlanguage. The sanre may be true of postposed dice, which rather thin i calque from were already undergoing shifting to Spanish. A sample of the Aymara or Quechua may simply represent a fossilization of an Gilegilence,including Brinton's translation is (Brinton I883: 4z-j, originally independentconjoined comment. combinations of two 5q-5r): or three particles,usually in the order nomds-pues-pero,are almost cuecueNsr: Puesmas ha sidocarpintero, hacedor de yugosaunque impossible to accept as having tponfan.ously in .arisen spuniJ seande papayo,hacedor de arados,aunque sean de without an indigenous contribution, althbugh these combinations tecomajocheya pachigiiemayule Sor. Gobo'Tastuanes {o n9t always correspondon a morpheme-by-morphemebasis with Quechuaor Aymara patterns(cf. Laprade r9At: zzo). }ial;*'"::::Lxt::iTf#:,i'i:,i';lxl wood. This should satisfy the clever Governor Tastuanes.' Nicaragua: the Giiegilense coBERNADoR: Ya pachigiiete no pachigtiete,pues Griegiienceasanese colonial central America contained large indigenous groups, but palparesiamo Don Forcico timaguas y verdad tin oficios surviving documents make little referetr.i to thiir approximuiion, No, I am not yet satisfied.Let Giiegiiense tell his son, to Spanish.The m9s! important extant text is the anonymous skit Don Forcico, to give a truthful account of his trades.' ctinci;nNcr: Pues si cana amigo Capn Alg' Mo'Mayague nistipampa El gilegilence,Aepicting Nahuatr rife in colonial Nicaraeua and Sres.principales, sones,mudanzas, velancicos,necana y evidently written in that cbuntry. The date of Comf6iition is palparesia Don Forcico timaguas y verdad tin oficios. unknown, but is probably ro*.*h.re towards the beginning or tn Then, if friend Captain Chief Alguacil will, in my eighteenthcentury. The time period covered by the events,ls well presence,cause the leading men to suspendthe music, as the language of the text, appear to centre on the seventeenth dances,songs and ballets, Don Forcico will give a century. Among the earliestcommentators of the Gilegi)ense, '... Brin_ truthful account of his trades.' lon_(r883:xvi) speaksof a mixeddialect ... .o-posed of a broken-down Nahuatl and a corrupt Spanish, which, at first, If the Gilegilenseis a legitimatesample of Hispanic-Nahuatlspeech servedas a means of communication betweenthe conqu.ro., und in colonialNicaragua, it would suggesta very fluent and penetrat- their subjects,and later became,to some degree,the uiual tonlu. ing bilingualism,rather than a haltinginterlanguage. of the latter'. The author goesfurther (p. xvii-),to suggestthat sich a mixed language once formed a pan-Mesoamerican creole; .This El Salvador jargon was carried into the various nations who came into contact Another Central American area for which indigenoussubstrata with. the Spaniards and half-breeds, antl hence we may find scat- havebeen implicated in phonologicaland morphologicalevolution tered words traceableto it in many of their tongues . . .' tn reality, is El Salvador.No creolestage can be postulatedfor El Salvador, the Gilegiiensecontains very few deviations frori standard Spanish sincea fundamentalingredient was lacking: the predominanceof 91 1t. time period. It is written in a combination of Spanish anJ speakersof mutually unintelligiblelanguages, who were forcedto Nahuatl, with extendedpassages in the latter languageihtercalated useimperfectly learned Spanish for essentialcommunication, in the in the form so often found among fluent bilingu-'als.iTh. spunirt absenceof extensivenative-speaker linguistic models. Even during of the Gilegilenceis not obviously influencedby a Nahuatl substra- the most intensiveHispanization, indigenous Salvadorans could tum; rather, the two languagesare superimposedon each other. and did use their own native languagesto communicateamong Flenriquez Urefra (r938: 326) speculatesthaf the wiitten text was themselves.There is no accuratedocumentation of the language(s) only meant to be a guide for the actors, who in turn would adopt a usedamong different ethnic groupsin precolonialand colonial El mix of Spanish and Nahuatl appropriate to their jndividuat auoi- Salvador,although linguistic contact probably took place only e-nces:The language of the text is colloquial and archaic, as befits alongthe edgesof the Pipil- and Lenca-speakingregions, either via the play's protagonists,a group of muleteers.There is some indica- bilingualismor through useof a singlelingua franca.Pipil, repre- tion that it is not spanish but rather Nahuatl which has suffered sentingthe culturally dominant group, is the most likely choice. from this bilingual contact. This could reflect the fact that low- More recentinterethnic contacts have adopted Spanish, and while ranking Spanishmilitary figures, as well as artisansand teamsters, there have alwaysbeen indigenous Salvadorians whose command also spoke Nahuatl, or that the Nahuatl speakers in Nicaragua of Spanishfalls below that of native speakers,nothing suggests progressivereinforcement of non-nativepatterns. 74 BEFORE AND AFTER SPAIN - THE NATIVE AI{ERICAN CONTRIBUTION EL SALVADOR 75 The Lencalanguage remainslittle studied,and Lencacontribu- closedsyllables, the phoneme lt"/ is not common syllable-finally. tions to Central American Spanishare an enigma.In the caseof Much more frequentis the groovefricative /5/ (orthographicallyx Pipil, claims of substratephonological influeice on salvadorian or sh),both word-internallyand word-finally.The replacementof Spanishare not lacking. For example,Geoffroy Rivas eg78: t7) Spanishsyllable-final /sl by 13/ is a recognizedcharacteristic of offers a number of assertions whicfi, while at odds wittr-our.*"0 Pipil-influencedSpanish in El salvador, and has given rise to linguistic behaviour,are representativeof substratisttheories of innovativeSpanish words: maestro'teacher'vs. maishtro'artisan.' central Americanspanish. with respectto pronunciationof /s/, he Salvadorianpopular literaturefrequently represents the shift /s/ > notes.that Pipil speakers'suprimieron toda diferenciaentre s, c y z, l3l as found among rural peasantsof Nahuatl-speakingback- substituy6ndolaspor el fonema nahua velar, fricativo, no ,onoro ground,as shownby the followingexcerpts from Salarru6(1969): que suenacomo unaj suave,Que airn usamosal decir,por ejemplo, nojotros,y que tan notorio es en el habla de los niiaragtieor.r.i No te resbal6s,ioishte'l (44o) (ryppressedthe differencesamong s,.c and z, replacingtheil uyit, Nahuatl Llego a la escuelay busc6 al maishtro, pero el maishtro se habia voicelessvelar fricativewhich soundsiit e Jroft/ ;hich acostado. . .(++r) we still use, for example,when we say nojotros,and which is so Pero el mal estabaen su querencia egoishta(436) noteworthyin the speechof Nicaraguans.)The example nosotros Aishtd esebaboso (4r9)

spanishthroughout the world, and is transpa-rentlythe ,.*tt'oiu Few speakers of Pipil remain in El Salvador (cf. Campbell 1985), bimorphemicinterpretation + as nos otros. Assuming that Geoi- but this phonetic trait remains to signal the last vestiges froy Rivas is referring of what to aspirationof syllable-final-/s/,as weil was once a more widespread ethnolinguistic trait. perhapsas to aspirationof word-initial and occasionalword-inier- Representing earlier stages of indigenous interlanguage in El nal intervocalic/s/, this flatly contradictsthe oft-mentionedclaim Salvador are several texts for folklore which imitate the speech that Nahuatl influen:€ wll responsible for RErBNrroNof syllable- of Pipil- and Lenca-speaking salvadorians. The following Pipil- final /s/ in MexicanSpanish. pipil Evenif as spokenin El salvador influenced texts give a hint of earlier ethnolinguistic varieties: had replacedSpanish /s/by [x] or [h], GeoffroyRivus'claim leaves unexplainedthe retentionof [sJin positionsother than syllablel Cuandolo cant6lagayu, word-final. Little accurateinformation existson the pronunciation lo peg6un pugido el mula, of Pipil in colonial El Salvador, but descriptionsirom the late Casisporque lo ha nacido nineteenthand early twentieth century give no support to claims La Ninguio que esti en el cuna . . . that indigenousphonetic transfer lies-behind weakining of Irli" Esodigu yo tambi6n, SalvadorianSpanish.2 Porquelus dosli hemosoydo, Dicen que lo nacio Next to nothin-gis known about the Lenca languageas en Bel6n used Y lu queremosconocer (Baratta rgsr-z;v. z,6tz). during the time of the colonization of El salvador. bon"te-po.ury Lenca-(as spokenduring the early decadesof the presentcentury) No lu triiigo nadaqu6 dar, Me lu robaronlos lagronis, provides no basis for reduction of /s/. Indeed, Lenca contains instances Com6teeste tu yuca, -of syllable-final/sl, although closed syllable, urc noi particularly common (cf. JrentiI'templu isti I'Ceiba Mendoza rglg). Lenca influenceis thus centenariu, ruled out in hermusu,bayu, accountingfor modificaiion of /s/ in Salvadorian centineladel mi pueblu Spanish. almiradu pur el cerrus _ Despitethe fragilitv-of the precedingclaims, pipil influencemay y culinasvigilantis have.indeed provoked a phonetic change in rural salvadorian undeI'sul cun eljulgures Spanish,a trait which is fading.as moiolingual spanish urugu lust6lumbrandonel tardes. . . encom-passesthe entire nation. pipil retains /t"/ in Lu hay barrancusy lu hay burdus, syllable-" aid word-final positions, although deaffrication to is attested piegrasgrandis y pefiunis, [s] y (Baratta rgsr-2: vol r, 27i. However, although iripil permits nel altu campanariu. . . 76 BEFoREAND AF.TERsPAIN - THE NATIVEAMERICAN CoNTRIBUTION PARAGUAY 77

Huy l' tiempulu estestigu Models of indigenous phonological penetration: cuandolu ruedal'caminu the vieju riyu Chanazigua most likely candidates cuanduI'so aguava in carreta.. 'ln . The claim is often macle that indigenous languages permanently front of the temple is the centenarian and beautiful ceibatree, influenced regional Spanish pronunciation. Such theories range sentinelof my village,admired by the vigilant mountains and hills, and betweentwo contradictory viervpoints.According to one position, lit by the afternoon sun . . . there are hoilows and valleys,large ani Spanish-Amerindian linguistic contacts extracted phonological small rocks. and in the tall belfry . . . today the weathei witnJssesthe common denominatorsfor twisting path of the river Chanazigua,whose waters go racing by . . .' the developingregional dialect. Spanish Deodanes(rg7z. r*4) sounds,phonotactic combinations or oppositions not found in the indigenous language were neutralized or eliminated. opposecl to These texts coincide with known facts of Nahuatl-influenced this view of phonological levelling is the assertionthat indigenous Spanish. Nahuatl/Pipil is noted for its predominantly three-vowel speakersfound certain Spanish sounds so difficult to articulate that system ([e] is assumed to be an allophonic variant of li/, and[o] is a their extraordinary efforts resulted in retention in the local Spanish variantof lul). Also observableis difficultyin articulating'lil, dialect even when other regions had eliminated the sound. Neither suggestingat timesa velar nasal(singuior) and at other ti-mesa theory completely excludesthe penetration of new sounds or opposi- consonantcluster lnyl (lenyita).The pleonasticclitic /o is also a tions from one language to the other, provided that certain unspeci- recurringtrait of Spanish-Nahuatlcontact. fied conditions are met. A brief sampling of proposed casesof Te.rtsimitating Lenca-influencedspanish are few, but demon- indigenous influence on Latin American Spanish pronunciation slrate similar phoneticcharacteristics. Unlike pipil, Lencahas a will illustrate the complexity of the issuesinvolved. five-vowelsystem, containing the samevowels is Spanish.The preceding textsshow some raising of unstressedfinal io/ to [uJ,a Paraguay phenonrenonwhich has its roots in popular peninsularspanish, If indigenous phonetic influence is to be found in any variety of but which could have beenreinfbrced by congruentpattirns in Spanish, it must surely appear in the nation where a Native Lenca.The remainingphonetic traits are commdn in ruial Spanish American language has captured the hearts and minds of even its of otherareas: inhabitants of European descent.The use of Guarani in Paraguay. and the fluent code-switchingexhibited by ,b_ililigualParaguayans" SeirorSan Pegru. has given rise to sweepingclaims of indigenous influenceon Para- que me llenen mi huacal guayan Spanish pronunciation. At one time or another, virtually y para otro afro every phonetic detail of Paraguayan Spanish has been implicated, venirlua llevar. including alveolar ltl and I dl affricated ltr Santo ScfrorSan Pegru. , I ,loss of syllable-final te lu estamoscelebrando /s/, presenceof phrase-initialfricative[f], and certain vowel reduc- tu clia,dejarnos llegar tions. Most of theseclaims are tenuous, and have been refuted by otro afro, te lu haremosmejor. careful research(e.g. Cassano rg7ral r97rb, r97rc, rg72a. rg7zb, (Barattatg5t-2: vol r., r972c, t97zd; Granda rg7ga, 198o,l98za). Sometimesa self-fulfill- 343) 'discovery' ing prophecy is at work, the of substratum influencein Both Pipil- and Lenca-based Spanish interlanguage reveal morpho- a region where the indigenous language is so obviously vigorous. syntaclic traits currently found in other regions wh.te Mesoameri- Inaccurate descriptionsof Paraguayan Spanish are also to blanre. can l-allsuagesare spoken. This consisteniy suggeststhat literary including early studies based on expatriate inforrnants or hasty antl folklore texts like the onesjust given aciuraiely reflectinterme- visits which resultedin non-representativespeech samples.3 diate.stagesof Spanis.hinterlanguage, which in some regions may One of the more tenaciousclairns of Guarani influenceinvolves havebecome nativized. retention of the palatal lateral l[,] ,which, from Malmberg (rt)+Z;S) onwards, many have attributed to Guarani. Guarani itself doesnot have this phoneme,and apparently neverdid; early Spanishborrow- ings into Guarani used either the affricate [y] or a simple vocalic BEFORE AND AFTER SPAIN - THE NATIVE AMERICAN 78 CONTRIBUTION MEXICO 79

hiatus.This fact has not deterredproponents of a substrateinflu- indistinguishablefrom that of other Southern Cone nations. A ence;as put by Cottonand Sharp(1988: 2n-4'when the Guarani reasonableexplanation is not hard to find. In order for Guarani- masteredthe difficult foreign sound,they made a point of distin- inducedsyntactic patterns to havepenetrated Paraguayan Spanish, qeating guishingit from ly I , an islandof lleismoin contrastto the a Guarani-inducedSpanish interlanguage would havehad to exist yelsmo of the surroundingareas.' This is an astoundingclaim, in colonial Paraguay.Such language would be usedby indigenous which if substantiatedwould be without precedentin the history of subjectsto Spanishcolonists, who spokelittle or no Guarani and languagecontact. It is not necessaryto hold out such extreme who maintainedphysical and socialdistance from the indigenous views,for the simplegeographical and socialisolation of Paraguay, community.In Paraguay,however, bilingualism was the rule from as comparedwith the rapidly developingmetropolitan area at the the very outset,and codeswitching into Guarani took the placeof mouth of the River Plate,provides a more than adequateexplana- Guarani-inducedcalques in Spanish.It is feasibleto assumethat at tion. Islandsof I [,] arealso found in otherremote regions of South the beginningthe proportion was heavily weightedin favour of America whose histories are not much different from that of Guarani,with Spanishwords introduced only to fill lexicalgaps. It 'Guarani-like' Paraguay, as well as in rural enclavesof Spain and the Canary was during this stage that the earliest and most Islands.a Spanishwords wereborrowed. In subsequentperiods, when Span- The feature of ParaguayanSpanish pronunciation which most ish was spokenfluently by largegroups of (urban) speakers,code vigorously resistsa non-substrateanalysis is the glottal stop fre- switchinginto Guarani assumedmore stylisticand emotivefunc- quently heard betweenwords, particularlywhen the secondword tions. Glottalization,extended across both Spanishand Guarani, beginswith a vowel. Two facts combineto suggesta legitimate underscoredthe bilingualcontact situation. substratecontribution, if not direct cause,for this pronunciation. The first is that amongregional Spanish dialects of SbuthAmerica, Mexico glottal this constriction coincidesalmost exactly with Guarani Although dozensof indigenouslanguages were spoken in colonial presence: Paraguay,extreme northeastern Argentina, and the east- Mexico,the eventsleading to the Spanishconquest of Tenochtitl6n ern extremesof Bolivia. Guarani insertsa similar glottal element, and subsequentinteraction with the well-organizedand politically patromonial both in Guarani forms and in borrowingsfrom span- strongAztecs gavea specialprominence to Nahuatl. In the Yuca- ish. StronglyGuarani-dominant bilinguals exhibit the glott alization tan, Maya was also a major language,although accordedlittle greater to an even degreethan spanish-dominantspeakers, further importanceby the Spaniards.Throughout Mexico and Central implicatingGuarani prosodic patterns. In peripheralareas such as America,Nahuatl was usedby Spaniardsas a lingua franca,even easternBolivia, bilingualismwith Guaraniwas not prolonged,and in some areas where Nahuatl had not previouslybeen spoken is currently limited to a small indigenouspopulation (Schuchard Spanish-Nahuatllinguistic contactswere intenseand penetrating rgTil.Glottal occlusions,while occurringin easternBolivian span- (cf. Heath ry72).Not only did Spanishreligious personnel adopt ish, are considerablyless frequent than in areaswhere Guarani Nahuatl as the languageof evangelizing,but Spanishcivilians and continuesto be spoken.In a contextof extendedbilingualism such military also acquiredsome fluency in the language,as witnessed asParaguay, where children were raised by Guarani-spJakingmoth- by the number of grammar books publishedduring the colonial erswho spokeSpanish as a secondlanguage, such a global phono- period.Indeed, the first bookspublished in Mexico werewritten in logical constraint could feasibly be carried over to the evolving Nahuatl. WhereasSpain never treated Nahuatl-speakingNative ParaguayanSpanish dialect. Americansas equals,the political strength of the former Aztec If Spanish-Guaranilanguage contact was strong enoughto im- empire and the need to maintain cordial relationswith its heirs plant a Guarani prosodictrait in the regionalspanish dialect, we resultedin a greater diffusion of the Nahuatl languageamong might wonder why more non-Spanishsyntactic patterns are not Spaniards.As in Paraguay,mixed-ethnic marriages, producing found with comparableregularity. Granda (tg79b, 1988),Meli6 childrenraised by Nahuatl-speakingmothers, were instrumental in (t974), Welti (tgll and othershave described Guarani-influenced fosteringa symbioticrelation betweenSpanish and Nahuatl. Na- syntacticconfigurations, or simpleresults of imperfectlearning of huatl disappearedas an urban home languagemuch faster in Spanish,but such patterns predominateonly among Spanish- Mexico than Guarani in Paraguay,since the demographicpropor- recessivebilinguals; educated Paraguayans use a syntax which is tions and socialbarriers were different in eachcountry. No trust- THE ANDEAN HIGHLANDS BEFOREAND AFTER SPAIN - THE NA'TIVE AMERICAN CONTRIBUTION 8t

worthy documentation of earlier stagesof Hispano-Nahuatl bilin- side. Given that speakers of Maya are employed as domestic gualismsurvive to document what were surelyintermediate configu- servants, labourers and vendors, almost no urban yucatecos lack rations of code-switchingand languageadmixture, but the languige contact with Mayan-influenced varieties of Spanish. The social of the Gilegilenseis probably representative.Nothing suggesistfie caste system of the Yucatan is still in effect, so that upward spread prior existenceof a stable Nahuatl-induced interlanguage,which of linguistic features is inhibited by the nearly total lack of upward could have allowed extra-I{ispanic grammatical patterns to coa- socioeconomicmobility of the indigenouspopulation. The linguistic lesce. while there were always Nahuatl-speaking residents who traits associatedwith the Mayan substratum (glottalized conso- spoke little spanish, the Hispanization of the Nahuatl community nants, glottal stops, and certain syntactic features) have remained soon created a dominant society in which Spanish was the first in stasis,found among large numbers of indigenous subjectsbut language. within the urban ron., where Mexican Spanish was not penetrating into the monolingual Spanish of urban dwellers. formed, Nahuatl became the home language of an increasingly For the latter changeto occur, a fundamental shift in the sociodemo- smaller population. The lack of a transitional interlanguage ii- graphics of Yucatan would have to take place. counts for the nonexistenceof Nahuatl syntactic structuies in Mexican Spanish. what can still be found in Nahuatl-speaking The Andean highlands areasis a transitory interlanguage,which is usually overcomeafter The contemporary existence of Quechua- and Aymara-influenced a single generation. and which may evolve in the lifetime of interlanguages in the Andean region, and the unusual characteris- tics of the Andean Spanish dialect cluster, have resulted in numer- individual speakers(Hill 1987,Hill and Hill r986, Siade ry10. Nahuatl speakerswho acquire Spanishas a native languagedo not ous claims of indigenous influence in segmentaland suprasegmental usesuch structures, phonology and syntax. From extremesouthern Colombia to north- which.leaveno traceson Mexican Spanish. 'Andean' easternChile and northwesternArgentina, Spanishexhib- . F.y ph-onetic traits of Mexican Spanish leqd themselvesto theoriesof an indigenoussubtrate. some have suggeiieda Nahuatl its a recurring cluster of phonetic traits, including unreduced and influencein loss of I [,] , assibilaiionof final /r/ (Malmberg 1965: frequently voiced /s/ combined with unstressedvorvel reduction, a t z4), intonational patterns, and reduction of unstressedvoweli, groove fricative lr I , a sibilant syllable-final lr I , quasi-affricate pro- but theseclaims are not substantiatedby empirical data (cf. Lope nunciation of the cluster ltr I , and retention of the phoneme I Kl as Blanch tg67a). Most frequently associated with an indigenous opposed to lyl, usually assigningsome type of lateral articulation transfer is the extraordinary resistance of syllable-final lsl (cf. to the former. Among Quechua and Aymara speakers for whom Malmberg 1965: 123-4,HenriquezlJrefia 1938:336,Alonso r93g: Spanish is a weak second language, effective reduction of the 336). Early investigatorshypothesized that Nahuatl had no /sl, at Spanish five-vowel system to a three-vowel system also occurs, but feastword-finally, but only an affricate lt"L This sound replaced/s/ no traces of such reduction are found in monolingual Andean in indigenous pronunciation of Spanish, and when eventual deaf- Spanish dialects. A possible exception is the recurring tendency to 'stronger' frication occurred, the original articulation persisted in pronounce the diphthong liel, as in tierra, as [i]. 'Andalu- the form of a sibilant [sJ.According to this viewpoint, the Contacts between Spanish and Quechua speakers, especially sian' reduction of /sl to ttrt run unlh.cked in r.gionr where /s/had those belonging to the dominant Inca culture, parallel Hispano- not been previously reinforced to an affricate, thus explaining the Nahuatl contacts in Mexico. Spanish clerics immediately seized prevalenceof sibilant [s] in the Mexican highlandsvs. loss of /i/ in upon Quechua as a lingua franca for religious and political pur- the caribbean. canfield (r934), Lope Blanch (r967b) and others poses, and while extending Quechua to areas where it was not indigenous, also created a pool of native Spanish speakersproficient have demonstrated that Nahuatl did indeed contain a final ls/ , in From the earliest contact with Europeans, which was_equatedwith Spanish c and z, already reduced to [s] by Quechua. Quechua Hispanisms. There is some indication that those the turn of the sixteen century. The Castilian apical [S]was equated began to absorb with Nahuatl [5], but Nahuatl /t" / never consistently replaced Span- Quechua dialects which currently possessthe palatal lateral l[,1 ish /s/. acquired it from contact with Spanish, rather than contributing to In the Mexican Yucatan, Mayan languages have prevailed over the retention of this phoneme in Andean Spanish. At the same Spanish for a longer time period; cities and towns have been time, high-ranking Incas or mestizossuch as Garcilaso de la Vega Spanish-speakingislands in the midst of a Mayan-speakingcountry- obtained European education, and became proficient writers in 8z BEFoRE AND AFTER SPAIN - TIIE NATIVE AMERIcAN CoNTRIBUTIoN A SYNTACTIC EPIPHENOMENON: CLITIC DOUBLING 83 literary Spanish.In betweenthe two poles, i.e. Spaniardswho Hispaniccombinations, or Spanishverb forms suchas the gerund spoke Quechua(possibly with interferencefrom Spanish),and or infinitive may be usedin ungrammaticalfashion, reflecting the Quechuaswho were completelyfluent in Spanish,were the indig- behaviour of equivalentforms in the indigenouslanguage. The enousmasses, who at first had no contactwith Spanishat all, and resulting constructions are transparent carryovers from the consequentlyno influenceon its evolutionin the Andeancolonies. speaker'snative language.Other putative instancesof indigenous As spanishrecruitment of indigenouslabour for minesand agricul- syntacticinfluence are more resistantto a simple morpheme-by- ture cameinto full swing,the inevitableinterlanguage arose, always morphemeanalysis, causing sceptics to reject a substrateorigin. mediatedby a layerof mestizooverseers, foremen, artisans, and the Among the mostinteresting of the recurrentLatin Americansyntac- like. Unlike what happenedin Paraguay,where there was never an tic phenomenastrongly correlatedto an indigenouspresence is 'clitic attractiveaccumulation of easywealth, the mineral richesof Peru, doubling' of inanimatedirect object nouns, particularlyby Bolivia and later Ecuadorand Colombia attractedlarge numbers invariant lo. Examplesof this usagetypify the bilingual Spanishof of Spaniardsof all classesand professions,including stable family Quechuaand Aymara speakersthroughout the Andes,and arealso nucleiand singlewomen. Hispano-indigenous marriages, while not found in Nahuatl-influencedSpanish dialects of Mexico, and in 'spanish' uncommon,did not automaticallyproduce offspringand bilingual CentralAmerican Spanish. In the CentralAmerican dia- did not always receivethe tacit acceptancefound in less affluent lects and occasionallyin Mexico, but not in Andean Spanish, areas.A more rigid castesystem was maintained,creating fewer pleonastic/o is also usedin locativeconstructions and with some opportunitiesfor upward penetrationof indigenous-basedstruc- intransitiveverbs. In none of the dialectsdoes the use of lo in turesinto urbancolonial Spanish. bilingual Spanishcorrespond in a one-to-onefashion with any At the interfaceof the Spanish-speakingelite and the indigenous singleelement in the indigenouslanguage, much lesswith an object underclassesin the Andeanregion was a mestizopopulation which pronoun,and yet the fact that theseuses of lo areexclusively found was eventuallyto becomenumerically dominant in urban areas, in bilingual contactzones encourages the searchfor a substratum and whoselinguistic traits would graduallylose the stigmaassoci- influence. atedwith the underprivileged.It is impossibleto determinewhether In the caseof clitic doubling with inanimatedirect objects(/o 'I the soundsassociated with lrl, lnl, l[.] , etc.in AndeanSpanish tengoel carro have the cart'), and invarant lo (e.g.lo pongola 'I havebeen influenced by Quechuaor Aymara, evenwhen the same caja place the box'), a purely Hispanic solution is not readily soundsare found in the latter languages.This is becausethe apparent,and no parallels can be found in dialectslacking an influenceof Spanishpronunciation on Andeanlanguages, including indisputableindigenous substrate.s In Puno,Peru, for example,the the incorporationof new phonenies(such as le/, /ol andperhaps sociodemographicsstrongly favour the possibility of indigenous /,(/), must also be reckonedwith. The almost exact geographical linguistictransfer to Spanish,since as much as go'/oof the popula- 'Andean coincidencebetween the clusterof Spanish'phonetic traits tion speaksQuechua or Aymara. Benavente(1988) found that and the area where Quechuaor Aymara was spokenseems too among university studentsin Puno, acceptanceof clitic-doubled greatto be dueto merechance, but the factsof Hispano-indigenous constructionsoccurred at levelsof and evenhigher, includ, 7o-8oo/o 'who cultural and linguisticcontacts in the Andean zone do not easily ing non-agreeinglo as in iquiin lo tiene la llsve? has the explainsuch coincidences. It remainsbeyond dispute that in most key?', impossiblein other Spanishdialects. Moreover, bilingual of the region, extreme forms of consonantalmodification and speakersaccepted these combinations more readilythan did mono- vowelreduction carry a sociolinguisticstigma, and are found more lingual Spanishspeakers. Godenzzi (1988), also studyingthe Span- frequentlyamong bilingual members of the indigenouscommunity. ish of Puno, obtained comparableresults. Clitic doubling was preferredamong the lowest socioeconomicsectors (among which A syntactic epiphenomenon:clitic doubling indigenousspeakers are overrepresented), as wereredundant posses- siveconstructions. In Mexico,clitic doublingwith /o is found only Syntactictransfer from indigenouslanguages to Spanishis usually in the Spanishinterlanguage of Nahuatl-speakingbilinguals (Hill confinedto the non-nativeinterlanguage of non-fluentbilinguals. 1987),and is unknown among monolingualSpanish speakers of In addition to the instability of morphologicalendings, the word the sameregions. order of the native languagemay be transferred,resulting in non- A comparisonof Spanishconstructions involving doubled lo and 84 BEFoRE AND AFTER SPAIN - THE NATIVE AMERIcAN CoNTRIBUTIoN A SYNTACTIC EPIPHENOMENON: CLITIC DOUBLING 85 the equivalent combinations in and Nahuatl suggests Quechua Nahuatl/Pipil or Lenca to Spanish occurred slowly, and possible modes of transfer, more subtle than a word-by-word indigenous-influenced syntactic structures at times became nativ- translation. Two models of indigenous penetration will be briefly ized, and persistedfor severalgeneratioos,6 sketched,by way of illustration. Even the most cursory glance at the grammar of Nahuatl reveals Nahuatl-influencedSpanish is currently found in some parts of 'non-Spanish' that no one elementcorresponds to the usesof /o as Mexico. among Spanish-recessivebilinguals. Such usage has re- shown above (cf. Andrews 1975, Sullivan more cently disappearedfrom El Salvador,and vestigesof clitic-doubling ry76).However, subtle forms of interlanguage transfer suggestthemselves. Nahuatl which may indicate a Nahuatl substrate are found in Honduras forms sentenceswith a SVO word order similar to Spanish, but and Nicaragua. The following examples from El Salvador illustrate Nahuatl transitive verbs take a direct object prefix (qui in the third the range of syntactic possibilitiesexhibited by lo, where the item 'doubled'by person singular and quin in the third person plural), which is also the clitic is given in brackets: combined with direct object nouns, even when questioned: /n alist6stus caites, te /rupon6s tu sombrerovieju'you tie up your sandals andput on yourold hat.' Ni- qui-ttain cihuatl IS seethe woman lo veoa la mujer} {definite direct object} 3s {Yo yo Tle ti- qui-tta? no /u tengumilpa'l don't havea cornfield.'{indefinite direct = object) What 2s 3ssee What do you see?{iQui lo ues?} No /r.rtraigo nada qu6 dar'I don't bringanything to give'{negative directobject) Here the correspondence between qui(n) and Spanish /o is even no /u sabemusqui6n es Esa Persona que est6 en luscielus'we don't closer, since the corresponding Nahuatl element is also a clitic, knowwho theperson in heavenis' albeit with different distributional characteristics. Like Spanish /o, i nterrogativedirect object} { 'now qui(n) is invariable for gender. Moreover, the minimal differences ya mekt voya mi casa I'm goinghome' {locative} between the singular and plural forms in Nahuatl might addition- dicenque Io nacioen Bel6n'They saythat he was born in Bethlehem' ally prompt a single choice in Spanish. {intransitiveverb} yo /o heyvenido'[ havecome' verb} Nahuatl- and Pipil-influenced Spanish, unlike Andean varieties, 'you {intransitive frequently uses invariant lo for intransitive verbs and locative tan bonitoque te /o sois areso pretty' {intransitive verb} constructions. Nearly all documented cases involve verbs in the preterite tense; Io accompanies few intransitive verbs in other Similar constructions are found in Nahuatl-speaking areas of tenses. In Nahuatl, a frequent morphological indicator of the Mexico (e.g. Hill rg8Z). Clitic doubling with inanimate and indefi- preterite is the prefix o-, placed before the proclitic subject pronoun nite direct objects is frequent in Mexicano speech,as is use of lo and the verb stem. This morpheme is invariable, combined with with intransitiveverbs: frequent stem changesdepending upon the classof verb: Lo compramosla harina'we buy the flour' 'he o- ni-coch lo traeun chiquihuite bringsa basket' PRETIs sleep= 'I slept'(lo dormi) la mam6/o est6mirando la novia'themother is lookingat thebride' 'they o- O tlacat No /o sabenhablar en castilla don't knowhow to speakSpanish' 'He/she 'he nnnr{he/she} born = wasborn' @naci6) /o compraun mediokilo . . . buyshalf a kilo' o- ti- nen qui6n quiso?'Who ;A /crs did shelove?' pRerzS live = 'You lived' /o ponenabajo los plitanos'they put thebananas underneath' @viviste\ 'now ya lo llevala novia . . . he takesthe bride. . .' The preterite marker o- occupies the same relative position as comida/o vamosa dar'we aregoing givefood' to Spanish clitic lo (assuming a null subject pronoun in normal Spanish usage), and moreover bears a phonetic similarity with /o. In Mexico, this interlanguageis unstable,and usually evolveseven Given the relative paucity of Spanishclitics which could be appropri- during the lifetime of individual speakers.Apparently the same ated in representation of Nahuatl clitics, lo appears to have been occurred in El Salvador, but in rural regions the shift from pressedinto duty with several distinct functions, corresponding to 86 BEFoREAND AFTER sPAIN - TIIENATIVE AMERIcAN CoNTRIBUTIoN A SYNTACTIC EPIPHENOMENON: CLITIC DOUBLING 87 Spanish.Postnominal -ta may also separate in Nahuatl. Quechuadoes not have a single be followed by other enclitic -particles particlesin non-dative particle for either intransitiveor preteritecorresponding in both constructions,in effectbeing'buried' among the clitics position and consistency,and use of Spanishinvariant lo is not and not correspondingin any clear way with a Spanisfi -ta normallyfound in AndeanSpanish in combinationwith intransitive element.only in the case of accusative is the linear order verbs. convergentenough with Spanishcurrc + vERBcombinations to make transfer Quechua-influencedspanish exhibits a more restrictedrange of feasible.It is not irrelevant that Spanish/o itself clitic doubling, at the least fluent levels marks an accusativerelationship, albeit not in the fashion of involving the invariable -ta. element/o. clitic doubling occurs only with direct objects;/o is Quechua A speakerof the developingindigenous interlanguage, never combinedwith intransitiveverbs, in locativeconstructions, encounteringpreverbal lo only in clearlytransitive sentences linclud- ing the possibility ' or in other combinationswhere no direct objectis involved.Unlike of clitic doubling with human Dos, as in the southern Nahuatl, wherea direct object marker is attachedto the verb, in cone), would be all the more likely to overgeneralizethe Quechuait is the direct object noun itself which is inflected(cf. 1e9d_for lo to appearin all transitiveclauses. Since the quintessen- tial interlanguage Catta 1985,Cole 1985,Cusihuamiln ry76, GiilvezAstorayme r99o, Quechua-influenced maintainsan o-v word order, Spanish/o Lastra 1968).Quechua marks direct object nouns with the suffix would at first be misanalysedas a casemarker attached -ta: -ta (or -manif followinga verb of motion). This suffixis invariable, to the noun,in a directcalque of Quechua cliticizesto all direct object nouns whether definiteor indefinite, elponcho-lo tengo and even attachesto questionsand relativeclauses, as shown by the following (Peruvian)examples (an approximation in 'Andean' Spanishis givenin parentheses): As-interlanguagespeakers develop greater fluency in spanish,word order gravitatesto the more usual o-v for non-clitic Dos.It is at T'ika -ta kuchu-ni this stage that lo, now recognizedas an object clitic, remains 'I Flower-acccut rs = cut the flower' @corto laflor) bglind in proclitic position, yielding the stable Andean Spanish ima- ta kuchi-nf'! clitic-doubledpattern. This patternof eventsis admittedlyspecula- = 'What What-,c,cccut rs do I cut?'(aqu6lo corto?) tive, but it doescorrelate well with observationson the devetopment Challwa-taapa -nki of proficiency -.q.cc 'You spanish amongQuechua speakers (cf. also Muysken Fish carry zs (FUT) = will carry fish'@ llevardspescado) r984). Asta-ni unu-ta 'I The consistencyof clitic Carry ts water-acc = carry water' (lo acarreoagua) doubling and use of invariant lo in regionalvarieties of Spanishinfluenced by widely separatedand unrelatedindigenous The accusativemarker -ta does not occupy the identical syntactic languagesdoes not have a unique origin, whatever position as the invariable Io of the corresponding Andean Spanish the contemporarymanifestations of such clitics. ftre clitics in question sentences,which would be roughly as indicated above. However, it are identicalwith normal Spanishdirect object clitics, be, easy for a speaker of spanish interlanguage to interpret albeit at times exhibiting apparent loss of agreement.In ryould addition, the clitic /o, statistically the most frequent, as some sort of transitiv- Spanishdialects for which no indigenoussyntactic pat- ity marker comparable to -ra. Although in ternscan be postulated,e.g. in the southerncone, allow wider Quechua euechua this patterns 'clitic element is always attached to the direct object noun, in a canonical of doubling' of direct objectsthan dialectsof Spain or elsewhere Quechua sov transitive sentencewhere the direct object immedi- in Latin America. For example,personal definite di.rgctobjects can ately precedes the verb, -ta coincidentally comes just before the and usually are doubled by the corresponding clitic: Lo verb, i.e. in the identical position to spanish proclitic lo. ln conozcoa Juan'I know John'. Contemporarysyntactic eue- analyses chua, the case marker -ta has other functions, including adverbial of the extensiveclitic doubling in Andean Spiniih have usuallydeparted and locative uses. It is also used to signal direct objects in certain from the notion that the Andeanclitics, including invariant /o, double-object constructions involving verbs of helping and teach- occupythe samestructural position and carry out the samefunction ing. In nearly all instances,however, -ta does not appear in immedi- as in other varietiesof Spanish(e.g. Barrenechea ate preverbal position, nor in any other single canonical position and orecchia ry77, Lujin ry87, Sufier 1988).The few theoretical studies that might cause -ta to be calqued by an object clitic in Andean of Nahuatl-influencedstructures in Mexican interlanguage 88 BEFoRE AND AFTER SPAIN _ THE NATIVE AMERICAN CoNTRIBUTIoN SUMMARY OF NECESSARYCONDITIONS 89 Spanish(e.g. Hill r987) have postulateda different syntactic struc- ture for direct object nouns, but regard invariant lo as carrying out combinedwith intransitiveor existentialverbs in Nahuatl-influ- the functions of a normal Spanishdirect object clitic. The theoreti- enced Spanish,a Spanishfunction is never assumed,and this cal arguments entail the conditions under which a direct object elementcannot be analysedas a Spanishsyntactic pattern. noun and its respectiveclitic are allowed to co-occur in the same The caseof clitic doublingin indigenousinterlanguages illustrates clause. It is possible that Spanish speakers in Mexico or the that substrateinfluence need not alwaystake the form of semanti- Andean region who have natively acquired clitic doubling with cally and syntacticallytransparent borrowings, but may involve invariant Io are treating this element on a par with object clitics in transferat a more abstractlevel. Since no one-to-onecorrespond- other Spanishdialects, but for interlanguagespeakers this is doubt- ence betweenSpanish and the indigenousconstructions can be ful. Moreover, the range of environmentspermitted by invariant /o postulated,the casefor substrateinfluence will alwaysbe circum- in these indigenous-influenceddialects, including questions and stantial,bolstered as much by the lack of comparableconstructions with indefinite direct objects, departs so radically from usual Span- in Spanishdialects without a demonstrableindigenous substratum 'parameter ish patterns to make a simple resetting' interpretation asby positivecorrespondences with a givensubstrate language. appear unlikely. Closer examination of the use of lo in Andean and Nahuatl-influenced Spanish shows differencesas well as similarities, Summary of necessaryconditions and the differences highlight the fact that under situations of bilingual contact, both interlanguageshave appropiated a Spanish Mere geographicalproximity of Spanishand Native American clitic and its accompanyingposition vis-d-visthe Spanishverb, but languagesis not sufficientto postulate substratalinfluences on for different reasons in each case. The use of invariant /o in both regional Spanishdialects, nor is a demographicpredominance of bilingual contact situationscan be consideredas an intrusion of the indigenouspopulations. Lexical borrowing can occur under the indigenous substratein a very real sense,but not in a transparent most superficialcontact conditions, but transferof phonologicalor word-by-word fashion.The wide rangeof syntacticand morphologi- syntacticpatterns requires a specialmix of demographic,sociolin- cal functions represented by Spanish /o, including direct object guisticand historicalconditions. Two basicmechanisms allow for clitic, neuter article, component of relative pronouns, etc., creates indigenousphonological or syntacticinfluences on regionalSpanish the potential for this same item to be used as a portmanteau dialects.The first is the mixed-ethnicbilingual household,with a particle in the interlanguage. Originally, this element was inter- EuropeanSpanish-speaking father and an indigenousmother. In preted by indigenous speakers as a calque of an invariant mor- sucha household,if both parentsbecome bilingual, code-switchifrg pheme in their respective languages, which fortuitously occupied iathei thbn siructuralinteiference will dominatethe linguisticenvil the same position in the linear string as Spanish preverbal /o, was ronment, and children will learn fluent and minimally modified monosyllabic, and sometimes bore a phonetic resemblance to lo. versionsof eachlanguage. Transfer from the indigenouslanguage As these speakersbecame more fluent in Spanish, /o was expanded to Spanishwill be limited to intonation,rhythm and possiblysome to include the entire range of Spanish direct object clitics, particu- segmeltaltraits. Paraguaycomes closest to embodyingthis configu- larly when doubling a direct object noun, but except for Spanish- ration.' In bilingual householdswhere children ipend most time dominant bilinguals, it is unlikely that these clitics occupy either of with primary care-giverswho speak only indigenouslanguages the canonical argument positions. Additional fluency in Spanish andlor nonfluentSpanish, acquisition of Spanishby childrenmay brought further expansion of Io in the interlanguage, with the not coincidewith the Europeanstandard. The other method of adoption of Such Spanish-like patterns as enclitic adjunction to transferis the graduallanguage shift of a predominantlyindigenous 'climbing' infinitives, clitic with restructuring verbs (e.g. quiero population, resulting in the stabilization and nativization of a 'I hacerlo ---+Io quiero hacer want to do it'), etc. The use of double Spanishinterlanguage. Whether or not suchnativized interlanguage clitics sometimes found in Andean Spanish (e.g. lo quiero hacerlo) becomesaccepted as a regionalor national standarddepends upon provides indirect proof that a non-spanish syntactic substratum is sociopoliticalconsiderations. It is possiblefor the two to coexist for long periods at work: /o is still carried over as a representation of the Quechua of time without interlanguagefeatures entering accusative marker, but is optionally attached to both possible prestigenorms. This hasoccurred, for example,in much of Bolivia, positions in the corresponding Spanish sentence.In the caseof lo Ecuador,the Yucatan,and in the Amazon Basin.With the gradual upwardsocial mobility of indigenousand mestizogroups, the door - 90 BEFORE AND AFTER SPAIN THE NATIVE AMERICAN CONTRIBUTION 9r 'en is opened for upward transfer of certain featuresinto regional el aspecto morfologico, los nahuas trasladaron al espafiol191 norms. patronei, formas y procedimientos propios del polisintetismo. Unieron dor o mis palabras, suprimiendo fonemas, para formar nuevaspalabras, surgiendo isi en el habla mestiza formas como vapui (vaya pues)-,puesi (puis si), vud (voy a) . . . idiay (y de ahi), aloshte (ya lo oiste), onde Notes (donde) . . . enque(aunque), ende(desde\,endenantes(desde antes). . . En otros casos,nos conformamos con suprimir fonemas put (pues\, ay of the Gilegiienceare studied r. The question of the language and origins (ahi), ma (toma), va (vaya), ante (delante), baio (debajo), tas (est6s).'(as (1883), edition of the work, by Elliot by Brinton in the first critical ior morphology, the Nahua translated their polisynthetic forms and (I938: Mintica (t989), Arellano (I984), (r884), Henriquez Urefia 325-7), patternJ to Spanish. By joining two or more words together, while (t884) and Arellano (1984) feel that the and many others. Elliot Lliminating phonemes,to form new words, the mestizo speechdeveloped is work of two authors, at least one of whom was a Giiegiiense the forms like vapud (vaya pues'),pues{ (pues si), vud (voy a) . . . idiay (y de native Spanish speaker. ahi), aloshte(yalo oiste), onde (donde) . . . enque(aunque), ende(desde), (1978: I7) also claims Nahuatl influence for Spanish z. Geoffroy Rivas endenantes(desde antes) . . . in other cases,we simply eliminate phonemes: lyl: 'la ll se yefsmo (neutralization of lyl and l['l) and of epenthetic pud (ptes), ay (ahi), ma (toma), va (vaya), ante (delante), baio (debajo)' por que no s6lo substituy6 a aquel fonema cambi6 una.y muy marcada, ras (estis).) sino que se introdujo donde no existe. No s6lo decimos cabayo, estreya, etc., sino que la pronunciamos separando el diptongo ia y decimos diya, The reader familiar with Spanish of other regions will immediately notice habiya, teniya, etc.' However, loss of the palatal lateral phoneme has that most if not all of these examples are widely attested elsewhere, occurred throughout much of Spain and Latin America, and dialect representing either the natural fusion of words in connected speechQtuesi, levelling resulted in extensive neutralization of lK/ and lyl in Spanish vua), resulti of popular morphology (ua), frequent loss of initial or final America. Neither Pipil nor Lenca contain a palatal lateral phoneme, syllables (tQ, ol irchaic but authentic Spanish carryovers from earlier and nothing suggests that such a phoneme existed in previous centuries. clnturies (onde, endenantes,enque, etc.). None of the examples is qualita- Thus, incipient yeismo in the Spanish arriving in Central America may tively different than connected speechfound in other Spanish dialects, and have been reinforced through contact with bilingual speakers whose theri is no justification for positing a subqtrate influence for this admixture Spanish continued to be influenced by indigenous languages. To claim of popular elements. indigenous influences as the sole cause behind neutralization of lAl and which ly lin El Salvador is to completely ignore the wider comparative perspec- 3. For example, Cassano (t97za) challenges earlier descriptions Spanish. tive, which situates this sound change in the context of the evolution of claimed that ltl and ld/ are usually alveolar in Paraguayan affricate ly I Peninsular Spanish during the I6th and rTth centuries. In Pipil, the Granda (r98o, r98za) disputes the existenceof a uniformly palatal lateral is absent, while there is a palatal fricative /y/. This throughout Paraguay. predominance of settlers element occurs freely before nonfront vowels (lal and /u/), but does not 4. Granda (tglil has pointed to a demographic contributing occur before the front vowel lil-lel . On the other hand, Pipil does have from northern Spain, where /,(/is still viable, as another diphthongs and hiatus combinations whose first element is [i]. Available factor. Po.zzi-. descriptions of Salvadorian Pipil give no hint of epenthetic intervocalic 5. Not all agree with this line of reasoning, however. For example, in Andean from which we are led to conclude that the epenthetic sound in Escot Ogli has suggestedthat clitic-doubled direct objects [y], periods modern Salvadorian Spanish cannot be a direct transfer from Pipil. Loss Spanish may be an archaism, giving some examples from earlier normal clitic- of lyl in contact with front vowels, however, could have been nudged to ol Spanish. Lozano (rgZS) believesthat analogy with the Juan'I gave completion by the lack of corresponding combinations in Pipil. Lenca, doubling of indirect objects in Spanish (e.g. le di el dinero a Aymara like Pipil, contains /y/ and lacks a palatal lateral. It freely permits the money to John') is at work, rather than Quechua or intervocalic lyl incontact with nonfront vowels, while excluding intervo- influence. of calic ly I in cbntact with front vowelsi some instances of word-initial ly I 6. Also found in Mesoamerican interlanguage is the combination (esa estos mis verbos, before front vowels are found. Lenea also contains instances of diph- article/demonstrative and possessive tus naguas, juino, were thongs and hiatus combinations involving lil, with no indication of . un tu desimulo, este tu munguieca etc.). Such configurations in old Spanish, and continue to exist in Portuguese. epenthetic [y] (Mendoza rg5g\. Geoffroy Rivas (1978: 17) offers yet not infrequent Spanish another broad generalization on possible Nahuatl influence on Salvado- However, they are not frequently found in Latin American attempts rian Spanish pronunciation: except for Guatemala and El Salvador. (Martin 1978, 1985) to trace this construction in Guatemalan Spanish to Mayan influence, but the interlanguage data from Pipil and Lenca speakers suggest 92 BEFOREAND AFTERSPAIN - THE NATIVEAMERICAN CONTRIBUTION

broader areal characteristics.The noteworthy absenceof similar construc- tions in most of Mexico in turn may indicate regional variation of Chapter 4 Nahuatl dialects. TheAfrican connection

Introduction Latin AmericanSpanish embodies linguistic and cultural contribu- tionsfrom four continents.In additionto the patrimonialEuropean heritageand the resultsof contactwith indigenouspopulations in the two American continents,Spanish came into contact with African languages,spoken by the tens of thousandsof Africans who formed the slave labour force of the developingcolonies. During the Spanishcolonial period, it is estimatedthat more than r.5 million African slaveswere imported into SpanishAmerica (Curtin t96g),and in manycolonies the African populationoutnum- beredresidents of Europeandescent for nearlythe entiretime. The possible African contributions to Latin American Spanish are closelybound up with the tragichistory of slavery,with racismand marginality, and with emergingnations' searchfor self-identity, often postulated on a European-Americanaxis which excludes Africa altogether.During most of the history of Latin America, Africans and their descendantshave occupiedthe lowestrungs of society,ranging from slavesto peonsand subsistence-levelfarmers and fishermen.Their lives and activitieswere not documented, exceptto criticizeor when they had run afoul of law or tradition, while few Africans were equipped to write their own history. Postcolonialtrends towards 'whitening', both demographicallyand through historical revisionism,further impede assessingthe full African linguisticimpact in Latin America.Without this key infor- mation, the reconstruction of Spanish in the Americas is incomplete. The history of Africans in Latin America beginswith the first European voyagesof exploration, some of which carried free African sailors.The African presencein the Iberian Peninsulahad