Gender in Language Contact

94

ÚrtakGreinin snýr seg um kynstillutan í tveimum tvímæltumhøpum. Í tí fyrra partinum av greinini verður hugt atføroyskum, og vit fara at vísa á, hvussu donsk lániorðhjálpa at gera av, hvat kyn er ómarkað, tí fleiri donskorð, sum eru felagskyn skifta k yn til hvørkikyn í før-oyskum enn øvugt. Seinni parturin er um kynstillutan íkatalanskum arvorðum borið saman við spanskt, harmálini bæði ikki hava sama kyn. Vit síggja, at einbrúkstendensur er í trimum aldursbókum (3-5, 19-23and 32-40). Munurin er mest at síggja, har munur er á,hvussu nógv katalanskt verður lært í skúlunum.

AbstractThis article deals with gender assignment in twodifferent bilingual contexts. In the first part, we look atFaroese, and will show how Danish loanwords help usestablishing default gender, as more Danish commongender nouns end up as neuter than neuter as commongender. The second part focuses on gender assignmentin Catalan words that have lexical cognates in Spanishand which differ in gender in the two languages. Itobserves usage trends among three age groups (3-5, 19-23 and 32-40 years old) differing mainly in the degreeof Catalan received at school. The speakers are fromtwo districts of Barcelona (Gràcia and Nou Barris)differing in the presence and usage of Spanish.

1. IntroductionThe goal of our investigation is twofold: (i)to show the status of borrowings whenestablishing the notion of default gender ina language and (ii) to look for possibletrends in the use of lexical cognates withdiffering gender in the two languages invol-ved in a language-contact situation. Thus,the issue of gender assignment is dealt within both a theoretical – in the case of (i) –and a usage-based approach – in the case of(ii). We use data from Faroese-Danish lan-guage contact in order to find evidence for(i), and Catalan-Spanish language data inorder to find evidence for (ii).

1.1. Arbitrary or rule-based genderassignment

Leonard Bloomfield wrote the following ongender in German, French and Latin:

There seems to be no practical criterion bywhich the gender of a noun in German, French,or Latin could be determined.

(Bloomfield, 1933:280).

Fróðskaparrit 55. bók 2007: 94-114

Gender in Language Contact: Evidencefrom Faroese-Danish and Catalan-SpanishKyn og málsamband – dømi í føroyskum-donskum og katalanskum-sponskum

Ariadna Benet1, Susana Cortés2 and Hjalmar P. Petersen3

SFB 538 Mehrsprachigkeit, Max-Brauer-Allee 60, 22765 Hamburg, Týskland.Emails: 1) [email protected], 2) [email protected]) [email protected]

Gender assignment is fully arbitrary, accor-ding to Bloomfield, and this was a viewlong upheld in the linguistic literature. Morerecent research on the matter in Indo-European, and also in other languages, hasshown that gender assignment rules indifferent languages are rule based in thatlanguages have a predominant either se-mantic, morphological or phonological as-signment system; or there can be also amixture between these principles. This hasbeen incorporated into different linguistictheories, see for example Steinmetz (1985,1986, 1997), Köpcke & Zubin (1984),Bittner (2000), Doleschal (2000), Leiss(2000), Trosterud (2001, 2006), Froschauer(2003), Nesset (2006), Rice (2006), amongothers.

With regard to French, as it is mentionedby Bloomfield above, and as gender assign-ment usually is taken to be totally arbitraryin that language, Mel´čuk (1958), (cf.Bidot, 1925), has shown that there are pho-nological assignment rules at work for “notless that 85 per cent of the frequently occur-ring nouns” (Corbett, 1999:57).

As for gender assignment in German,we refer to the work by scholars such asKöpcke (1982), Köpcke & Zubin (1984),Steinmetz (1985, 1986, 1997), Doleschal(2000). All these studies show that genderassignment is rule based, and the main as-sumption is that there is a hierarchy ingender assignment in that semantic assign-ment rules (SAR) take precedence overmorphological assignment rules (MAR),which again take precedence over phono-logical assignment rules (PAR) (Corbett,1999; Corbett & Fraser, 2000a, b). Al-

though such a formulation (SAR >> MAR>> PAR) might be too strong, given that forexample Faroese tjaldur ‘oystercatcher’ isneuter even though names of birds aremasculine or feminine, there is a certaincore, which follows this principle:

The Core Semantic Override PrincipleRules referring to biological sex take prece-dence in gender assignment

(Nesset, 2006:18).

According to The Core Semantic OverridePrinciple, gender follows sex for examplein German: der Mann-m., die Frau-f. anddas Kind-n. The same pattern is observed,for example, in Faroese, Icelandic and OldNorse: mað(u)r-m. ‘man’, kona-f.‘woman’, barn-n. ‘child’; and Latin: vir-m.,femina-f., but liberi-m.pl. (becausemasculine is default in Latin = heredenoting ‘human beings of both sexes’).

Evidence in favor of a rule based ap-proach to gender are that:

(1) Native speakers typically make few mis-takes,

(2) borrowed words acquire a gender, whichshows that there is a mechanism for assig-ning and not just remembering gender, and

(3) when presented with invented words, spea-kers give them a gender and they do so witha high degree of consistency.

(Corbett, 1999: 7).

In addition to this, we note that there arevery few changes of gender in languagechange, unless the whole system ‘breaksdown’ (cf. English), with consequences forthe whole morphological system of thelanguage. A fourth evidence in favor ofgender as rule based is the fact that children

KYN OG MÁLSAMBAND – DØMI Í FØROYSKUM-DONSKUMOG KATALANSKUM-SPONSKUM

95

acquire a complex system like genderpretty effortlessly.

As gender is so crucial for the morpho-logy of a language – requiring agreement –the speaker and the child acquiring thegender system apply one main principle:Preserve gender, which would explain thevery few gender changes between OldNorse and Faroese or Icelandic, and alsowhy the gender of a borrowed word usuallyis preserved - if the speaker is more or lessa balanced bilingual1. An interesting exam-ple is that where the speaker, a woman fromthe elder generation (70+), first uses com-mon gender, corresponding to the Faroesemasculine noun sandur (‘sand, beach’);then she immediately corrects herself anduses the correct Danish gender, which isneuter:

(1) og vi endte nede på sanden, sandet dér…and we ended up down on the beach-m./beach-n. there‘And we ended up down on the beach overthere’

(WO2V).

The rule with regard to Danish borrowingsin Faroese is that Danish neuter nouns areneuter in Faroese, common gender nouns inDanish end up as either masculine or femi-nine, cf. also opus, which is neuter in Polish(and Latin), even though a word with a finalconsonant should be masculine (Kryk-Kastovsky, 2000: 734). Example (1) showsthat the gender systems of a bilingualspeaker are autonomous (Costa et al.,2003); this would indeed be expected inasymmetrical gender systems like Faroeseand Danish.

The function of gender as a referencepoint for declensional classes is decisive forhaving two independent systems, and whenthe systems change they do so in certainsteps, for example in Norwegian, where thechange first was from neuter plural collec-tive nouns to feminine singular collective,then masculine became used for nomenagentis, neuter as nomen actiones; mascu-line became also used to denote abstracta,neuter to denote concreta, ultimately endingup with common gender and neuter, Beito(1986).

In a parallel way to (1), bilingual spea-kers in Catalan and Spanish show confusionnot with borrowings, but with cognates thatdiffer in gender in one language and theother. For example, the Catalan word forroad sign is senyal-m. whereas it is señal-f.in Spanish. It has been observed that bilin-gual speakers display confusion in genderassignment when they speak Catalan, andsome display the confusion in their Spanishproduction too. The study presented in sec-tion 4 focuses on the Catalan production ofcognates displaying such a mismatch.

In this study we will observe the choicesas regards to gender when Catalan speakersuse a set of cognate words that differ in gen-der in Catalan and Spanish. In this sense,Costa et al. (2003)’s study on the gendersystems of bilingual speakers is relevant.They try to find evidence for or against theautonomy of the systems of both languages;the pairs of languages considered havestructurally identical gender systems (i.e.,Catalan-Spanish and French-Italian, withtwo genders in each of the languages of thelanguage pair) and another pair is asym-

GENDER IN LANGUAGE CONTACT: EVIDENCEFROM FAROESE-DANISH AND CATALAN-SPANISH

96

metrical in this respect (i.e., Croatian-Italian, Croatian having three genders andItalian only two). They compare the naminglatencies of pictures whose names differ ingender with those whose names have thesame gender in both languages. The resultsshow that the naming latencies do not varydepending on the fact that the word pairshave different or the same gender in bothlanguages, no matter whether both langua-ges are structurally asymmetrical or identi-cal. This leads to the conclusion that thegender systems of a bilingual speaker areautonomous, even when both languages aresimilar, as it is the case of Catalan andSpanish.

2. Gender in language contactThe aim of this paper is to observe howgender manifests itself in language contact.We will draw attention to two differentlanguage contact situations: one involvesFaroese and Danish, and the other Catalanand Spanish. Specifically, for the first casewe will take borrowings from Danish intoFaroese into consideration. For the secondlanguage contact situation, we will observethe gender that Catalan speakers tend to usefor cognate words that have different gram-matical gender in Spanish and in Catalan.

2.1. Faroese: how borrowings can helpestablishing default gender

In entering a room, an Icelander would askif allir-m.pl. have arrived, that is, masculineplural when referring to both sexes. AFaroese would use neuter plural øll.

Another evidence in favor of neuter asdefault in Faroese is the use of past par-

ticiple neuter when referring to both sexesopposite to masculine plural in Icelandic:(Far.) brekað-n.pl., ~ (Icl.) fatlaðir-m.pl.‘handicapped’.

Faroese has three genders: masculine,feminine and neuter, whereas Danish hastwo: common gender and neuter. Commongender nouns in Danish are a merger of OldDanish masculine and feminine. One im-portant principle noticeable in the Faroese-Danish language contact is that gender ispreserved, so that Danish neuter nouns, notonly cognates, are neuter in Faroese; andDanish common gender nouns are eithermasculine or feminine in Faroese.

We will use the data from Faroese-Da-nish to find further evidence in favor of a de-fault neuter, assuming that cyclic deriva-tions with one default gender are crucial inlanguage processing and derivation (cf. Kip-arsky’s (1982) and Anderson’s (1992) Else-where Principle or Elsewhere Condition,which has also been labelled Panini’s The-orem after the famous Indian grammarian).

The way we expect default to be repre-sented in the borrowings is that (1) there aremore changes from Danish common gendernouns to Faroese neuter or from Danishcommon gender and neuter nouns to neuterin Faroese, than (2) to the opposite gender:that is, changes from Danish neuter to mas-culine or feminine in Faroese. If we findmore evidence in favor of (1) and there arenot semantic, morphological nor phono-logical motivations behind the change, theonly explanation is neuter as default.

2.2. Catalan: trends in gender usageBoth Catalan and Spanish have two genders


97

for substantives and adjectives: masculineand feminine (when talking about pro-nouns, also neuter should be considered).The usual gender markers for substantivesin each language are the following ones(Lloret & Viaplana, 1998; Clua, 2003):

(1) Catalan:Masculine: -[Ø] [‘gat-Ø] catFeminine: -[ə] [‘kaz-ə] house

Spanish:Masculine: -[o] [‘gat-o] catFeminine: -[a] [‘kas-a] house

There is also a set of other gender markersfor substantives which are not as usual asthe previous ones:

(2)Catalan:Masculine:-[u] ([‘amu], owner), -[ə] ([‘diə], day),-[us] ([‘tipus], type), -[i] ([bi’ƒçti], moustache),-[is] ([‘bRindis], toast), -[s] ([‘fons], bottom)Feminine:-[u] ([‘fotu], photo), -[us] ([‘bEnus], Venus),-[i] ([‘dçzi], dose), -[is] ([‘glçtis], glottis),-[s] ([‘Eliks], helix), -[Ø] ([‘kçl], cabbage),

Spanish:Masculine:-[a] ([‘dia], day), -[e] ([o’βoe], oboe),-[i] ([‘taksi], taxi), -[is] ([‘bRindis], toast),-[s] ([‘toRaks], thorax), -[Ø] ([li’mon], lemon)Feminine:-[o] ([so’pRano], soprano), -[e] ([‘pRole], progeny),-[i] ([me’tRopoli], metropolis), -[is] ([‘dosis], dose),-[Ø] ([‘kRuθ], cross)

In general, cognate words have the samegender in Catalan and Spanish: casa-casa,f. ‘house’, paret-pared, f. ‘wall’, boca-boca, f. ‘mouth’, armari-armario, m. ‘cup-board’, peix-pez, m. ‘fish’, llapis-lápiz, m.‘pencil’, etc. However, in some cases, Cata-lan and Spanish cognates show differentgenders, for instance: dent (Cat.), m. - dien-te (Sp.), f. ‘tooth’. These cases normallycorrespond to words that in one and/or theother language take one of the more alter-native gender markers presented in (2). Dueto this fact and also to the contact languagesituation, when bilingual speakers face thisdifferent behavior in both languages mayget influence from one or the other lan-guage and thus they may use the wronggender, which is made apparent through theagreement with articles or adjectives.

The speakers of the present study belongto three different age groups (children,young adults and adults) in two differentdistricts in Barcelona, differing in theamount of Spanish usage. Consideringthese six groups, we will see whether anykind of relationship can be establishedbetween the characteristics of the group ofspeakers and their behavior as regards togender assignment in the case of suchcognate words. And in a more general view,we will try and account for the speakers’decisions considering semantic, morpholo-gical or phonological criteria.

2.3. Meeting point in the study of genderin Faroese and Catalan

Both studies presented in this article contri-bute differently to an explanation of genderassignment in a language contact situation.


98

While through the Faroese-Danish case weare trying to find out a more theoreticalaspect of Faroese (i.e., to account for thedefault gender of the language throughloanwords), the case of Catalan-Spanishlooks for trends of usage among differentgroups of speakers and their correspondinglinguistic explanation. The common pointof both approaches is that they look forpossible semantic, morphological or phono-logical patterns in gender assignment thatcould explain both the speakers’ usage (forthe case of Catalan) and the establishmentof grammatical decisions (for the case ofFaroese).

3. The Faroese studyThe Faroese data builds on simplex nounsin the following two dictionaries: Donsk-føroysk orðabók (1995) (Danish-Faroesedictionary) and Føroysk orðabók (1998)(Faroese dictionary). The data from Før-oysk orðabók was used to look into seman-tic, morphological and some phonologicalassignment rules in Faroese. This work wasneeded in order to understand the mechan-ism behind assignment in borrowings,which we found in the Danish-Faroesedictionary. Where there were doubts, weasked different people by sending themsentence examples via e-mail, but also bylooking at the word in Google by simplywringing the ‘WORD site:fo’. In somecases we used the collection of words inFøroyamálsdeildin (Department of Faroese,Fróðskaparsetur Føroya).

The following nouns are common gen-der in Danish, neuter in Faroese:

afasi ‘aphasia’, alkohol ‘alcohol’, aktiv ‘active’,aorist ‘aorist’, antabus ‘antabuse’, baby ‘baby’,blits ‘flash’, boykott ‘boycott’, bonus ‘bonus’,brint ‘hydrogen’, buffe ‘sideboard’, butan‘butane’, dativ ‘dative’, deodorant ‘deodorant’,ebonitt ‘ebonite’, fallitt ‘bankruptcy’, forbannilsi‘curse’, forlovilsi ‘engagement’, forbrydilsi‘crime’, fornermilsi ‘insult’, fornýggilsi‘renewal’, forsinkilsi ‘delay’, galopp ‘gallop’,garanti ‘guarantee’, ginn ‘gin’, grafitti>‘graphite’, granitt ‘granite’, hekti >’hook’,herpes ‘herpes’, herts ‘hertz’, infinitiv‘infinitive’, hobby (hobby-f.)’hobby’, judo ‘judo’,indikativ ‘indicative’, kaffi ‘coffee’, karat ‘carat’,kompliment ‘compliment’, krem ‘cream’, jogurt‘yogurt’, luksus ‘luxury’, lockout ‘lockout’,metadon ‘methadone’, manna ‘manna’, moskus‘musk’, melis ‘refined sugar’, neksus ‘nexus’,modell2 ‘model’, nitroglyserin ‘nitroglycerine’,narkomani ‘drug addiction’, nugga ‘nougat’,notits (notits-f.) ‘note’, onani ‘masturbation’,paprika ‘paprika’, papyrus ‘papyrus’, pedofili‘pedophilia’, pláss ‘place’, passiv ‘passive’,perno ‘pernod’, porno ‘porn’, poesi (poesiur-m.,poesi-f.) ‘poetry’, reggae ‘reggae’, pláss ‘place,’pynt ‘decoration’, sex ‘sex’, rísalamang ‘ris a lamande’, silicon ‘silicone’, sement ‘concrete’,sjalusi ‘jealousy’, sirkumfleks ‘circumflex’,spiritus ‘spirits’, skepsis ‘skepticism’, suvenir‘souvenir’, stripptis ‘striptease’, te ‘tea’, teyp‘tape’, telefaks ‘telefax’, tekk ‘teak’, tennis‘tennis’, teleks ‘telex’, terror ‘terrorism’,terpentin ‘white spirit’, ting ‘thing’, trend(trendur-m., trend-f.) ‘trend’, tyfus ‘typhoidfever’, tubbak ‘tobacco’, upplivilsi ‘experience’,vígilsi ‘wedding’, urin ‘urine’.

Some of these nouns change gender due tosemantic or morphological assignment ru-les. More specifically, one of these rulestates that alcoholic drinks are neuter: alko-hol ‘alcohol’, ginn ‘gin’, perno ‘pernod’,spiritus ‘spirits’; chemical elements andmetals are neuter: brint ‘hydrogen’, butan


99

‘butane’, ebonitt ‘ebonite’, nitroglyserin‘nitroglycerine’ and terpentin ‘white spir-its’. One important semantic rule is thatbeings of both sexes are neuter, cf. barn-n.‘child’, hence the change in baby fromcommon gender to neuter. There is onemore semantic rule observable in the dataabove: Latin grammatical terms are neuter:aorist ‘aorist’, dativ ‘dative’, infinitiv‘infinitive’, indikativ ‘indicative’, passiv‘passive’, sirkumfleks ‘circumflex’. Thesenouns do then have a semantically motiva-ted change from Danish common gender toFaroese neuter.

In other cases it is possible to find mor-phologically motivated clues behind thegender change: forbannilsi ‘curse’, forlov-ilsi ‘engagement’, forbrydilsi ‘crime’, for-nermilsi ‘insult’, fornýggilsi ‘renewal’,forsinkilsi ‘delay’, upplivilsi ‘experience’,and vígilsi ‘wedding’, as the suffix -ilsi isassociated with neuter.

In the Donsk-føroysk Orðabók (1995)there are 51 simplex Danish words withword final -ment, which are neuter, and onlycement ‘concrete’, kompliment ‘compli-ment’, and konsument ‘consumer’ are com-mon gender in Danish. The reason for thechange of gender in Dan. cement-cg. > Far.sement-n. ‘concrete’, and Dan. kompliment-cg. > Far. kompliment-n. ‘compliment’ isobvious. The opposite is also obvious, thatis, where common gender is preserved asmasculine – more specifically, in Danishkonsument-cg. to konsumentur-m. ‘consu-mer’ – there is a semantic rule: persons aremasculine in Faroese (cf. other borrowingssuch as kommunistur-m. ‘communist’, pro-fessari ‘professor’, and native nouns such

as einstaklingur ‘person,’ and the historicalchange of Old Norse persóna-f. to persón-ur-m. ‘person’). Note also sjúkrarøktar-frøðingur-m. Lit.: sick-nursing-researcher= ‘nurse’.

The above changes are obvious and in-disputable. Then we observe nugga-n.(Dan. nougat-cg.). Words with word final-a are, as a rule, feminine. There are 1,262simplex Faroese feminine nouns with thissuffix, 32 neuter3 in Føroysk Orðabók(Faroese Dictionary) from 1997. Thechange of gender in nugga is suspect. If welook at simplex nouns with word final -ggwith different nuclei vowels, the followingpicture emerges:

Table 1Three words with word final -egg are masculine(veggur ‘wall’, leggur ‘calf’, seggur ‘man’). Thereare four neuter nouns with word final -egg (egg ‘egg’,dregg ‘grapnel’, legg ‘pleat’, skegg ‘beard’). There isone feminine word with -ágg, (gágga ‘whelk’), butthere is no actual rule at play here, as-ágg and -ogg >are phonologically the same [çg:].///// indicates that it is not possible to find any rule.

This could perhaps explain the change ofgender in nugga ‘nougat,’ although we areinclined to take the change to be simply thatneuter is default, especially since fourwords are too few to establish a rule, andthey do not have -u as the nuclei vowel.

The ending -o in Danish common gen-der nouns is as a rule associated with femi-nine in Faroese, e.g. konto-f. ‘account’.


100

i/y e ø u o a/æ á ú ó ei ey oy í/ý

gg //// 3,4 /// /// /// /// 1 /// /// /// /// /// ///

Gend. m,n

Strangely enough three words end up asneuter: judo ‘judo’, porno ‘porn’ and radio‘radio’. Neuter in porno and judo can bestbe explained as a result of default gender,but it is possible that the semantic field judobelongs to has cased neuter; compare forexample borðtennis ‘table-tennis’ and bad-minton. Another case of semantic analogyis perhaps in buffe-n. ‘side board’ (< Dan.buffet-cg. ‘side board’) in analogy to nativeborð-n. ‘table.’ A pretty clear case of se-mantic analogy is in pláss-n. ‘place,’ (Dan.plads-cg.) corresponding to the native nounstað-n. ‘place’.4 The noun urin might havechanged from common gender to neuterbecause of conceptual association with thenative piss-n. ‘pee’.

Native -us words are associated withmasculine: gánus/glanus/kánus ‘person thatstares’, nemus ‘person that has to toucheverything’. -us in borrowings is, for in-stance, found in other nouns such as bam-bus/bambusur ‘bamboo’, and syklus/syk-lusur-m. ‘cycle’. In Danish there are neuternouns ending in -us such as genus ‘gender’,

korpus ‘corpus’, and kursus ‘course’. Butthe change of gender in antabus ‘antabuse’,bonus ‘bonus’, luksus ‘luxury’, moskus‘musk’, papyrus ‘papyrus’ and tyfus‘typhoid fever’ is still strange. Spiritus-n.has changed gender due to the semanticassignment rule alcohol is neuter, and pa-pyrus has changed gender due to a semanticanalogy with pappír-n. ‘paper’.

There are 1,314 simplex masculinenouns with word final -i in Faroese in thenominative singular. Out of these, 53 aremasculine as a result of the semanticassignment rule: persons ⊃ m., like granni‘neighbor’. Inherited neuter nouns withword final -i (as kvæði ‘ballad’) are close to500, so one should not expect commongender nouns with word final -i to change toneuter based on these numbers alone. Stillwe find the following changes from com-mon gender in Danish to neuter in Faroese:afasi ‘aphasia’, garanti ‘guarantee’, grafitti‘graffiti’, hekti ‘hook’, hobby ‘hobby’5, kaffi‘coffee’, narkomani ‘drug addiction’, onani‘masturbation’, pedofili ‘pedophilia’, poesi


101

Figure 1. (1) Neuter and common gender in Dan.,neuter in Far. (2) Neuter and common gender in Dan.,masc. in Far. (3) Neuter and common gender in Dan.,feminine in Far. (4) Neuter and common gender inDan., all three genders in Far.

The figure shows, (1) how many nouns are neuter andcommon gender in Danish, and what gender the samewords have in Faroese, that is neuter. Then there area few nouns that are neuter and common gender inDanish, but end up as masculine in Faroese. (3) illu-strates how many nouns are neuter and common gen-der in Danish and end up as feminine in Faroese. (4)shows the nouns that are neuter and common genderin Danish, and might have all three genders in Faro-ese. That is, they are in a transition period.

60

50

40

30

20

10

01 2 3 4

‘poetry’ (might be masculine poesiur andfeminine also poesi), reggae ‘reggae’sjalusi ‘jealousy’. There are neither seman-tic nor any morphological motivations be-hind the change.

3.1. Common or neuter in Danish, neuterin Faroese

Some nouns in Danish might have bothcommon gender and neuter. Most of themend up as neuter in Faroese (Figure 1).

Note here the semantic rule chemicalelements and metals ⊃ n., which will ac-count for the gender of some of the nounsabove. For example anilin ‘aniline’, sellu-loid ‘celluloid’, karbid ‘carbide’, klor ‘chlo-rine’, naftalin ‘naphtalene’, natron ‘soda’,salpetur ‘saltpeter’, trotyl ‘TNT’, uran ‘ura-nium’, but the other end up as neuter as aresult of default.

If we look at the opposite situation, thatis, how many Danish neuter nouns changeto masculine or feminine in Faroese, wefind the following nouns:

The change of centner to masculine sentn-ari ‘hundredweight’, gelænder to gelend-ari-m. ‘banisters’ and register to registari-m. is regular. The Danish ending is associa-ted with the nomen agentis ending -er as

løber-m. ‘runner’ ~ Faroese renn-ar-i-m.‘runner’. Also -el, in kápil ‘cable’, has amorphological explanation, cf. nouns in -ilor -ul in Faroese as stuðul-m ‘support’.

As for gardin ‘curtain’, there are onlytwo words ending in -din in Petersen’sdatabase and both are feminine, gardin-f.and sardin-f. ‘sardine’.

The change of spinet-n. to femininespinett-f. ‘spinet’ is in accordance with thesemantic rule: musical instruments ⊃ f.,which belongs to the Semantic AssignmentRules.

It is not surprising that smykke-n. chan-ges to smúkka-f., if we look at the generaltendency for nouns ending in -e in Danishto become feminine in Faroese.

Neuter nouns ending in -e change theirending to -i: bælte > belti ‘belt’, hæfte >hefti ‘pamphlet’, orgie > orgi ‘orgy’, rygte> rykti ‘rumor’, studie > studii ‘study’; to -a: dogme > dogma ‘dogma’, medie > media‘media’, or they preserve the Danish (orEnglish) ending: hardware, software, pate,

resume.Given the data

above, especiallycommon gender >neuter and commongender/neuter >neuter, we concludethat there is eviden-ce in borrowings forneuter as being the

default gender in Faroese.

4. The Catalan studyAs described in 1.1., the Catalan study foc-uses on the production of Catalan words,


102

Table 2. The following Danish neuter nouns change gender, when they areborrowed into Faroese

centner > sentnari-m. ‘hundredweight’ credo > kredda-f. ‘creed’fýr > fýrur-m. ‘fireplace’ gardin > gardina-f. ‘curtain’gelænder > gelendari-m. ‘banisters’ kabel > kápil-m. ‘cable’register > registari-m. ‘register’ smykke > smúkka-f. ‘ornament’spinet > spinett-f. ‘spinet’

which happen to be cognates with Spanishwords but differ in gender in the two lan-guages.

4.1. Methodology4.1.1. Target words

A list of Catalan target words was devised,with cognates that differ in gender inSpanish and Catalan. The target words inthe present test are the following:

Table 3. Target Catalan words and their Spanishcognates.

4.1.2. SubjectsThe subjects for this study belonged to 3different generations in 2 different districtsin Barcelona, namely Gràcia and Nou Bar-ris:• G1 (age: 3-5): 10 children in Gràcia (8

girls and 2 boys; mean age = 3.9) andNou Barris (5 girls and 5 boys; mean age= 4.5).

• G2 (age: 19-23): 10 young adults inGràcia (5 male and 5 female; mean age =21) and 10 in Nou Barris (4 male and 6female; mean age = 20.8).

• G3 (age: 32-40): The group of adultscomprised one of the members of eachparental unit of the children interviewed:10 in Gràcia (9 female and 1 male; meanage = 36.4) and 10 in Nou Barris (6female and 4 male; mean age = 37.1).

In the case of some words which do not be-long to the child lexicon, we did not includeG1 in the test, because the chances to obtainthe word from them were too scarce. There-fore, the cases in which this was the casewill be indicated in the results section.


103

Figure 4. Figure 4 shows (1) nouns that are commongender in Danish, but are borrowed into Faroese asneuter. In this particular case there are no specificrules, semantic, morphological or phonological thatmight explain the change. The only explanation is thatneuter is default. The numbers show how borrowingssupport other findings in the language. In (2) there isa semantic explanation of why the Danish commongender nouns end up, when borrowed, as neuter inFaroese. In (3) the motivation behind the change issemantic analogy and in (4) formal rules are respon-sible for the change of Danish common gender toneuter in Faroese.

60

50

40

30

20

10

01 2 3 4

Catalan Spanish Glossfront - m. frente - f. foreheaddent - f. diente - m. tootholor - f. olor - m. smelldubte - m. duda - f. doubtllum - m. luz - f. lampfull - m. hoja - f. sheet (of paper)sida - f. sida - m. AIDSanàlisi - f. análisis - m. analysispostres - f. postre - m. dessertsenyal - m. señal - f. sign

4.1.3. EquipmentThe data were recorded with a Sony ECM-CS10 unidirectional lapel microphone,plugged into a portable Mini-Disc Hi-MDWalkman MZ-RH10 Sony recorder.

4.1.4. ProcedureAs the data reported here forms part of alarger study, data collection sessions lasted45 minutes each, approximately, in whichsubjects were asked to perform differenttasks. For the specific purpose of the currentstudy, subjects were asked questions aimedat eliciting some of the cognates in the listabove. Apart from being asked specificquestions, subjects were also shownpictures of objects, which they were meantto describe, in which they could see objectsin the list of target words.

4.2. Results4.2.1. First group: front - m., dent - f.,

olor - m., dubte - f.The words under study in this first grouphave in common that they are everydaywords. They are, in principle, high frequen-cy words as they include parts of the bodyand a word that has to do with the senses(i.e., olor). As for dubte, it is a more abstractnoun but is still often used in everydayspeech.

The original gender of these words inLatin is always kept in Spanish, but changesto the opposite gender in Catalan. Thus,Spanish follows the Preserve gender prin-ciple, but Catalan does not.

4.2.1.1. front - m., Spanish: frente- f.; ‘forehead’

The Catalan word for ‘forehead’ comesfrom Latin FRŌNS, FRŎNTIS, f., and theCatalan masculine gender is consistentlyused in most Catalan dialects. It is only keptas a feminine noun at the Northern andWestern part of the Western dialect (Coro-mines, 1980).

Table 4. front

The results show that Catalan speakers donot have a problem with the gender of thesewords overall. The only noticeable differen-ces are those between G1 in Gràcia andNou Barris, as well as in G3 in both dis-tricts.

The strong presence of Spanish in NouBarris is made obvious in speech produc-tion patterns in the language of children andadults. Nevertheless, G2 shows ceilingeffects in both districts. This could be due tothe fact that these subjects have receivedCatalan education from the beginning.Hence, they have learnt the gender distinc-tion from very early on, and that is whytheir production sticks to the standard re-gardless of the district they live in.


104

Group % of Catalangender usage

G1 Gràcia 87.5G1 Nou Barris 44.44G2 Gràcia 100G2 Nou Barris 90G3 Gràcia 90.9G3 Nou Barris 60

4.2.1.2. dent - f., Spanish: diente- m.; ‘tooth’

Catalan dent comes from Latin DENS,DĔNTIS, m. Although the etymological wordis masculine, it appears as feminine inCatalan from the very first occurrences intexts, in the 13th c. (Coromines, 1980).

Table 5. dent

The production reaches ceiling effectsacross the board. The only group that doesnot stick to the standard is the group ofchildren in Nou Barris. It looks as thoughthe influence of the language of the en-vironment plays a crucial role for this agegroup, which displays a major usage of theSpanish gender when they produce thisCatalan noun.

4.2.1.3. olor - f., Spanish: olor - m.;‘smell’

Catalan olor comes from vulgar Latin OLOR,OLORIS, m. (from classical Latin, ODOR,ODŌRIS). It often appears as ‘odor’, in thelearned version of the word, in the middleages but it disappears after the 15th c. InRamon Llull’s works olor occurs already.The masculine gender is generally associa-ted to Spanish influence and is rejected forthis lexical item (Coromines, 1980).

Table 6. olor

Again, the only group that shows a produc-tion with a large association of the Spanishgender when speaking Catalan is G1 in NouBarris.

4.2.1.4. dubte - m., Spanish: duda- f.; ‘doubt’

This word was not elicited from G1 infor-mants due to its abstract nature and the un-likelihood for it to be part of the activelexicon of children. Therefore, only resultsof this word’s production by G2 and G3 arereported here.

This is a deverbal noun, which comesfrom the Latin verb DŬBǏTARE. The closestnoun in Latin is DUBITĀTIO, DUBITATIŌNIS, f.

Table 7. dubte

Results show ceiling effects across theboard, with the lowest percentage of Cata-lan gender usage by G3 in Nou Barris. The


105


G1 Gràcia100G1 Nou Barris 37.5G2 Gràcia 91.66G2 Nou Barris 90G3 Gràcia 100G3 Nou Barris 86.66


G1 Gràcia 100G1 Nou Barris 55.55G2 Gràcia 100G2 Nou Barris 100G3 Gràcia 100G3 Nou Barris 90


G2 Gràcia 100G2 Nou Barris 90G3 Gràcia 90.9G3 Nou Barris 72.72

influence of Spanish in the environmentmight be playing a role here.

4.2.1.4. SummaryThe words in the present group do notconstitute problems for Catalan speakers,with respect to their gender marking inCatalan. The children in Nou Barris are thespeakers who show more confusion withrespect to this. Perhaps they are at a deve-lopmental stage where they need more in-put in both languages to have a clearer ideaof which gender to use in each language.Cognates present a further complication inthe acquisition process, especially in dis-tricts where Spanish is highly active in theenvironment, because of their similar formin both languages. Hence, Spanish has adetrimental effect on the acquisition ofCatalan gender in these cognate words inthe Spanish-speaking district.

4.2.2. Second group: llum - m. and full- m.

The words in the present group have incommon that they are possible in bothgenders in Catalan although each gender isassociated with a different meaning. Hence,llum means ‘light’ in feminine but ‘lamp’ inmasculine. In turn, fulla (with the extrafeminine morphological marker -a) means‘leaf’ whereas the masculine form fullrefers to ‘sheet (of paper)’. It is importantto mention that these two words come fromneuter gender words in Latin.

4.2.2.1. llum - m., Spanish: luz -f.; ‘lamp’.

This word comes from Latin LŪMEN,

LŪMǏNIS, n. Although neuter Latin wordsgenerally took the masculine gender inCatalan, llum showed a strong tendency totake the feminine gender, possibly due tothe influence of LUCEM, f. However, thefeminine eventually refers to ‘light’ where-as the masculine refers to ‘lamp’ (Coromi-nes, 1980). Spanish only has one referent,which corresponds to the feminine Catalanform.

Table 8. llum

The results are overall lower than those inthe previous group. The specialization ofgender might play a role in such a behavior.Children seem to be at a loss because, ifcognates are difficult enough to acquire,when a word can be associated with any ofthe genders in a single language dependingon the referent, the task is even harder. Andin G1 Nou Barris, the influence of Spanishreinforcing the gender with a differentmeaning in Catalan does not help. Youngadults and adults also show the problematicstatus in production of this lexical item.

4.2.2.2. full - m., Spanish: hoja -f.; ‘sheet (of paper)’.

As mentioned above, the Latin word Cata-


106


G1 Gràcia 12.5G1 Nou Barris 0G2 Gràcia 66.66G2 Nou Barris 68.42G3 Gràcia 65.21G3 Nou Barris 79.16

lan full comes from its neuter: FŎLǏUM. Inturn, the feminine word fulla, which refersto ‘leaf’, comes from the plural form of thisword, i.e. FŎLǏA. It was originally used as acollective noun referring to all the leaves ofa tree/bush, for example. Then, it adoptedthe singular meaning too, but semanticallyspecialised to designate ‘leaf’ and not ‘sheet(of paper)’ (Coromines, 1980). This seman-tic association with specific genders paral-lels the case of Norwegian described inBeito (1986).

Table 9. full

Here we see that most groups have pro-blems with using the appropriate genderwhen designating a sheet of paper in Cata-lan. Strikingly, the children in Gràcia arethose with more difficulties in producingfull instead of fulla. This group is followedby G2 in Nou Barris, and then closelyfollowed by G3 in Nou Barris and then G1in Nou Barris. The two groups withproduction more in accordance with thestandard are G2 and G3 in Gràcia. Thus, itlooks as though the influence of Spanish isimportant because most groups that displayproblems in producing this Catalan wordwith the Catalan gender are those living in

the district with a higher usage of Spanish.The children in Gràcia also show difficul-ties in producing ‘llum’ with its appropriategender when designating a lamp due to thepolysemy of the word. Therefore, the pro-blem might be developmental in the case ofthis group rather than caused by externalfactors.

4.2.2.3. SummaryThis second group seems to be more pro-blematic than the first one due to the poly-semy of the target words, be it in the sameform or with slight modifications. It is clearthat the acquisition of these polysemic cog-nates means a further complication forchildren. It constitutes a real handicap forbilingual children. However, it looks like itis a developmental stage they go throughbecause older groups show better produc-tion levels, even though hardly any of themshows ceiling effects.

4.2.3. Third group: sida - f. and anàlisi –f.

In this third case, we have grouped twowords belonging to the semantic field ofscience together: ‘syndrome’ and ‘analysis’.This was one of the cases in which we didnot try to elicit the words from the youngestgroup (G1), since the words are not part ofthe children’s lexicon. In fact, these arewords not commonly used in everydaysituations and often learned late at school.This means that adult speakers who havereceived education only in Spanish willhave learned these scientific words inSpanish and, as they are cognate words,they will probably keep on using the Span-


107


G1 Gràcia 30.76G1 Nou Barris 62.5G2 Gràcia 88.88G2 Nou Barris 58.33G3 Gràcia 70G3 Nou Barris 61.53

ish gender even when they use the Catalanword. In the groups under study, the onewhich has not received education in Catalan– at least not since the beginning of theirschool years – is the third generation (G3).We predict thus that this will be the groupwhich will use the Spanish gender the most.

4.2.3.1. sida - f., Spanish: sida –m; ‘AIDS’

This acronym stands for síndrome d’im-munodeficiència adquirida. Thus, the gen-der of the word is given by the substantivesíndrome. In Catalan it is feminine, whereasit is masculine in Spanish. The Catalanword preserves the original gender, since itwas feminine in Greek: ἠ συνδροµή, whichmeans ‘meeting, different things gatheringtogether’. But in Catalan, as well as inSpanish, there are other words with phono-logically similar suffixes: -asme, -isme,-itme: sarcasme (‘sarcasm’), periodisme(‘journalism’), algoritme (‘algorithm’).These words are always and in both lan-guages masculine. Therefore, it seems thatthe case of the feminine gender for ‘syn-drome’, and thus for ‘AIDS’, follows a non-expected direction for the Catalan speaker,leaving etymology aside. This may lead thespeaker to confusion, which will be moreapparent, the higher his/her usage of Span-ish is.

Besides, the Catalan word for ‘AIDS’ iswidely used in mass media, which in Cata-lonia are Spanish- or Catalan-speaking. Infact, Catalan-speaking media are relativelynew, since Catalan was prohibited in publicdomains until the mid 1970s. From then on,Catalan normalization was led mainly

through media and education. And the ef-fect of it falls for the most part on youngerspeakers and on the traditionally Catalanones. Therefore, it is expectable that theyoungest groups and/or those with a higherusage of Catalan in their district show thehighest number of productions with theCatalan gender.

Table 10. sida

The most noticeable result is the low usageof the Catalan gender by G3 in Nou Barris,which correlates well with the predictionstated above. This is the group with lessinfluence from the Catalan normalization,as they live in a mostly Spanish-speakingarea, where mass media in Catalan havelow audience indexes, and they have notreceived education in Catalan.

As for the effect of the education only inSpanish, which should influence the wholeof G3 as predicted at the beginning of thissection, here it seems that this effect isoverridden by the effect of mass media. G3in Gràcia is actually the group with thehighest percentage of usage of the Catalangender. They have not received formaleducation in Catalan, but nowadays theymay be acquiring at least part of what theydid not have the chance to learn at schoolthrough the Catalan-speaking mass media.


108


G2 Gràcia 62.50G2 Nou Barris 72.72G3 Gràcia 81.81G3 Nou Barris 36.36

4.2.3.2. anàlisi - f., Spanish:análisis - m.; ‘analysis’

Again, the Catalan word keeps the originalgender from Greek: ἠ ἀνάλυσις was femi-nine and means ‘dissolution of a whole inits parts’. Strikingly, the Spanish genderfollows a non-expected direction, since sub-stantives ending in -V+si in Catalan and in-V+sis in Spanish are feminine: paràlisi,diàlisi, hipòtesi, crisi. Thus, a morphologi-cal rule should not allow confusion to arise.

Here we can apply only the predictionabout the lack of schooling in Catalan, asthis is not a word especially used in themass media. Thus, the older generation willbe the most likely to use the gender inSpanish, i.e., the masculine.

Before analysing the results, it has to besaid that really clear productions weredifficult to find due to the coincidence ofthe phonetic form of the substantive prece-ded by the masculine article and that pre-ceded by the feminine one: un* anàlisi (m.)and una anàlisi (f.) are both pronounced as[u.nə.’na.li.zi]; the same would happen withthe definite article. A possible way to dis-ambiguate the gender usage in the inter-views was to try and elicit the word un/una,not as the indefinite article with the substan-tive, but alone as the numeral. Thus, forinstance, after having produced [u.nə.’na.li.zi], a second question was posed:“And how many analyses does the doctorwant to be made?”. However, the aim wasnot reached in all cases. So, here only theresults of the clear answers are presented.

Table 11. anàlisi

The percentage of productions using theCatalan gender is in general lower than withthe previous word síndrome. Thus, althoughthe Catalan gender follows a more logicaldirection – as similar suffixes have the samegender and the noun keeps the original gen-der from Greek – the influence of Spanishoverrides the Catalan morphological ruleassociated with the ending V +-si. And theeffect of the mass media is clear in the caseof sida, as referred to above, but not withanàlisi. The latter is a word not so common-ly used in mass media, but in interpersonalsituations, like between the doctor and thepatient, where linguistic correctness is notso controlled.

If we look at the results for groups, themost striking fact is that G2 in Gràcia is theone with the lowest rate, which do notcorrelate with the schooling in Catalan norwith the dominance of this language in thedistrict. On the contrary, G2 in Nou Barrishas the highest rate. Thus, it seems that inthe case of this word the effect of theeducation in Catalan is higher in the districtwith lower degree of Catalan usage.

4.2.4. postres - f. pl., Spanish: postre -m. sg.; ‘dessert’

This is a word taken from Spanish, from an


109


G2 Gràcia 33.33G2 Nou Barris 57.14G3 Gràcia 50.00G3 Nou Barris 50.00

expression like “fruta que se da a la postreen las comidas” (‘fruit which is given at theend of the meals’). Note that the gender inSpanish is also feminine, and sometimes itwas used also with the plural feminineform, exactly like the actual Catalan form.However, in other Spanish texts it appearsas masculine, already in 1490. The firstwritten data of the Spanish word used in aCatalan text is from the 18th c. (Coromines,1980). The Spanish substantive evolvedtowards the masculine singular form, whilethe Catalan one remained the same, i.e.feminine plural. There is reason to believethat during Catalan normativization, in thefirst half of the 20th c., some decision aboutkeeping the opposite gender from Spanishmight have been taken. In fact, during sometimes there had been attempts to find a moregenuine Catalan substantive (clou-dinar‘close-lunch’, llevant-de-taula ‘clearing-of-table’, darreries ‘at the end’), but no pro-posal had enough acceptance (Coromines,1980).

As for the predictions about the degreeof Catalan gender usage that we will find, itis clear that the word is mainly used in thecontext of a household, as well as in thedaily activities at school. Therefore, al-though schooling in Catalan might have aneffect, the most important influence willcome from the family and their linguisticbackground, at least in the speakers withCatalan dominance.

Table 12. postres

The clearest result comes from children: inboth districts they always used the Spanishgender (and in some cases the Spanish sing-ular form as well), although G3, theirparents, show much higher rates, at least inGràcia. This indicates that the correct gen-der in Catalan for this word is difficult to beacquired by children up to five years old.

As for the adults, G2 shows a non-ex-pected result, since the group in the districtwith a lowest usage of Spanish (i.e., Gràcia)is the one with the lowest percentage ofproductions with the Catalan gender. Thismay be attributed to education in Catalan,in the sense that it may have a strongereffect in the district where the word is usedthe least in Catalan: if it is not normallyused in Catalan, the young speakers havelearnt the Catalan gender at school; if it isused in Catalan in the familiar context, it isused with the Spanish gender and educationhas no effect.

However, the results are reversed in theoldest generation, with Gràcia showing thehighest rate. In this case the influence fromthe family language background and alsofrom mass media in Catalan may be thefactors at play.

This case, and especially the results of


110


G1 Gràcia 0G1 Nou Barris 0G2 Gràcia 22.22G2 Nou Barris 54.54G3 Gràcia 44.44G3 Nou Barris 9.09

G2, can be related to the fact that theCatalan word took the opposite gender fromthe original Spanish one. In general, thereare few productions of the correct gender inCatalan, which reflect the lack of natural-ness in the usage of the gender of les postresand therefore the idea that this attribution tothe opposite gender (and also number) mayhave been decided as a way to create a lin-guistic norm of Catalan contrary to theSpanish entry in order to keep forms as dif-ferent as possible. As we see from the re-sults, though, this attempt would not bereally working.

4.2.5. senyal - m., Spanish: señal - f.;‘sign’

The word comes from the Latin substantivi-zed adjective SĬGNĂLIS, which, in turn,comes from SIGNUM, neuter. In Spanish ittook the feminine but, in most of the Ro-mance languages form – in Portuguese,Italian, French – it evolved to the masculineform, and it is found with this gender evenin some old Spanish texts (Coromines,1980). However, the Catalan usage of theSpanish feminine gender is documented asa “non-accepted Spanish loanword from”both by Coromines (1980) and by Alcover& Moll (1930-1962).

In Catalan, substantives with the sameending (-al) are usually feminine: catedral(‘cathedral’), moral (‘morals’ or ‘morale’),canal (‘channel’). They could be used bythe speaker as analogy and would serve asinfluence, together with the Spanish genderof la señal, to produce the Catalan wordwith the feminine gender.

As a final remark on the grammatical

background of the word, it is worth pointingout that it is a quite highly polysemic word.It refers to ‘mark’, ‘road sign’, ‘sign of thecross’, ‘beep’ (like that of the telephone, forinstance) and others. This fact implies thatthe word is used in several semantic fields,some of them belonging to contexts inwhich Spanish is a more usual languagethan Catalan. It is the case for instance of‘sign of the cross’, which for many peoplewas learnt in Spanish, since during Franco’stimes schooling was only in Spanish andreligion was an essential part of education.In order to elicit the word, the picture of aroad sign was shown. This was anotherword not tested among G1 due to the highprobability of their early vocabulary notcomprising it.

The coincidence of the phonologicalendings and the polysemic characteristicsof the word may induce to the usage of theSpanish gender. The effect of education inCatalan, thus on G2, may have again animportant role, since the wrong genderusage of the word is quite strongly correc-ted by the Catalan grammar syllabuses atsecondary school.

Table 13. senyal

The highest percentage of productions withthe Catalan gender comes from the group


111


G2 Gràcia 62.5G2 Nou Barris 30G3 Gràcia 41.66G3 Nou Barris 28.57

of young adults (19-23 y.o.) in Gràcia, thedistrict with the highest usage of Catalan ofthe study. Education in Catalan and thelanguage of the environment do play a rolein this case. Actually, the second factorseems to be the most influent one, since itworks also with the oldest generation (G3).

The polysemic character of the word andthe several contexts in which the word isused can also be observed from the results.The relative low percentage of the Catalangender usage in both groups of G3 mayreflect the old usage in Spanish of thereferent ‘sign of the cross’. Another contextin which the word is mainly used in Spanishis that of traffic rules, since they are mainlycodified in Spanish and learnt also in thislanguage at the driving schools. This factmay cause the low results for G2, especiallyin Nou Barris, since many youngsters learnto drive at that age, and in this district withmore presence of Spanish the chances tohear the word with the Catalan gender arelower than in Gràcia.

4.3. Summary and discussionThe results have shown two different ten-dencies among the words studied. On theone hand, highly frequent and basic words(i.e., olor, dent, front) do not lead toconfusion, in general. Only the group ofchildren in Nou Barris shows percentagesof the Catalan gender production below oraround 50%. On the other hand, with therest of the words speakers do show uncer-tainty in gender assignment to severaldegrees. The causes are diverse dependingon each case. One factor we have seen thatmay lead to confusion is the polysemy of

the word, and specifically the fact of havingdifferent genders for different meanings(llum, full) within the same language.Another case that causes low percentagesof the Catalan gender production, at least insome groups, is when the cognate words aremore used or learned in Spanish than inCatalan (síndrome, anàlisi, senyal). Finally,with the word postres, we have seen that apossibly imposed Catalan grammaticalnorm might be responsible for the lowusage of the correct gender of this word inCatalan.

The conclusion from Costa et al. (2003)related to the autonomy of the gender sys-tems in the bilingual speakers can possiblyaccount for the relatively high degree ofcorrectness in the case of the first group ofwords. As for the rest of words, which dolead to confusion, other causes, mainly se-mantic and usage-related, make the speakeraccess the wrong gender system, in thiscase the Spanish instead of the Catalan one.

5. ConclusionWe have shown that we can use the lan-guage-contact data in Faroese as evidencein favor of a neuter default. The generalgender assignment process in which Danishcommon gender nouns go to either mascu-line or feminine in Faroese borrowingsshows exceptions. In the cases where com-mon gender nouns in Danish end up asneuter in Faroese, and there are no morpho-logical, phonological or semantic rules toaccount for them, the only possible expla-nation is to have neuter as the default,which is supported by other parts of gram-


112


113

mar, where neuter is also the default, as thechange of allir -m. pl. ‘all’ to øll - n.pl.

As for the language-contact study ofCatalan, basic words in the lexicon (i.e.,front, dent and olor) do not represent aproblem for Catalan speakers, althoughthese words have a differing gender inSpanish. When confusion in gender usagearises, it can be mainly attributed to two dif-ferent factors along the Spanish influence:semantic rules related to polysemy, as in thecases of llum, full and senyal; and the factthat the word is more often learnt in aSpanish-related context (e.g, sida, anàlisi).

The two studies in the present paper areconsistent with Costa et al. (2003)’s fin-dings, that is: the speaker uses the particulargender system of the language he/shespeaks when in that language mode. Theerrors found in structurally asymmetricgender systems such as Faroese and Danishare few, even in cognates, and when chan-ges are found in borrowings as those pre-sented above, they are motivated by theselected language (Faroese). The produc-tion of Catalan basic lexical items withoutincurring in gender confusion might pro-vide further evidence for the autonomy ofthe gender systems in bilinguals of struc-turally similar languages.

AcknowledgementsThis study has been carried out as part of the researchprojects "K8: Variance in multilingualism on the FaroeIslands" and "H6: Phonoprosodic development ofCatalan in its current bilingual context" at theCollaborative Research Center for Multilingualism. Weare indebted to the Center and its sponsors, theDeutsche Forschungsgemeinschaft (German ScienceFoundation) and the University of Hamburg for their

support. For the Catalan part of the study, we want toexpress our most sincere gratitude to M. Rosa Lloret,who gave us inspiring suggestions during her stay atour Center. We would also like to thank Laura Batalla,Laia Puig and Anna Schreibweis for their essential helpwith the transcription of the data. Finally, our mostsincere gratitude to the schools, children and parentswho kindly accepted to participate in the study, whichwithout their collaboration would not have beenpossible.

ReferencesAlcover, A. & Moll, F.B. 1930-1962. Diccionari català-

valencià-balear. 10 volumes. Palma de Mallorca:Editorial Moll.

Anderson, S. R. 1992. A-Morphous Morphology. Cam-bridge: Cambridge University Press.

Beito, O. T. 1986. Nynorsk grammatikk: lyd- og ordlæra.Oslo: Det Norske Samlaget.

Bittner, D. 2000. “Gender classification and the inflec-tional system of German noun”. In: Barbara Unter-beck and Matti Rissanen (eds.) Gender in Grammarand Cognition, pp. 1-25. Berlin, New York: Moutonde Gruyter.

Bloomfield, L. 1933. Language. New York: Holt, Rine-hart and Winston.

Clua, E. 2002. “Gènere i nombre en els noms i en elsadjectius”. In Joan Solà et al. (eds.) Gramàtica delcatalà contemporani. Volume 1, pp. 483-534. 3volumes. Barcelona: Editorial Empúries.

Corbett, G. 1999. Gender. Cambridge: Cambridge Uni-versity Press.

Corbett, G. & Fraser, N.M. 2000a. “Gender assignment:a typology and model”. In Senft, G. (ed.) Systems ofnominal classification, pp. 293-325. Cambridge:Cambridge University Press.

Corbett, G. & Fraser, N.M. 2000b. “Default genders”. InBarbara Unterbeck & Matti Rissanen (eds.) Genderin Grammar and Cognition. Part I “Apporaches toGender”, pp. 55-99. Berlin, New York: Mouton deGruyter.

Coromines, J. 1980. Diccionari etimològic i complemen-tari de la llengua catalana. 10 volumes. Barcelona:Curial Edicions Catalanes.

Costa, A., Kovacic, D., Franck, J. & Caramazza, A.2003. “On the autonomy of the grammatical gendersystems of the two languages of a bilingual”. Bilin-gualism: Language and Cognition, 6 (3), pp. 181-200.

Doleschal, U. 2000. “Gender assignment revised”. InBarbara Unterbeck & Matti Rissanen (eds.) Genderin Grammar and Cognition. I Approaches to Gen-der. pp. 117-167. Berlin, New York: Mouton deGruyter.

Froschauer, R. 2003. “Genus im Althochdeutschen: einefunktionale Analyse des Mehrfachgenus althoch-deutscher Substantive”. Heidelberg, Neckar: Winter.

Kiparsky, P. 1982. Explanation in Phonology. Dordrecht:Foris.

Köpcke, K. M. 1982. Untersuchungen zum Genussystemder deutschen Gegenwartssprache. Tübingen: MaxNiemeyer Verlag.

Köpcke, K. M. & Zubin D.A. 1983. “Die KognitiveOrganisation der Genuszuweisung zu den einsil-bigen Nomen der deutschen Gegenwartssprache”.Zeitschrift für deutschen Gegenwartssprache, 11,pp. 166-182.

Köpcke, K. M. & Zubin, D.A. 1984. “Sechs Prinzipienfür die Genuszuweisung im Deutchen: Ein Beitragzur naturlichen klassifikation”. Linguistische Be-richte. 93, pp. 26-51.

Kryk-Kastovsky, B. 2000. “Norm versus use: On genderin Polish”. In Barbara Unterbeck & Matti Rissanen(eds.) Gender in Grammar and Cognition. II Mani-festations of Gender, pp. 729-747 Berlin, New York:Mouton de Gruyter.

Leiss, E. 2000. “Gender in Old High German”. In:Gender in Grammar and Cognition. I Apporachesto Gender. Eds. Barbara Unterbeck & Matti Rissa-nen. Mouton de Gruyter. Berlin, New York.

Lloret, M.R. & Viaplana, J. 1998. “El binarismo delgénero gramatical en castellano y catalán”. Verba25, pp. 71-91.

Nesset. T. 2006. “Gender meets the Usage-Based Model:Four principles of rule interaction in gender assign-ment”. Lingua 1139, pp. 1-25.

Petersen, H. P. and Staksberg, M. 1995. Donsk-føroyskorðabók. Føroya Fróðskaparfelag.

Poulsen, J. H. W. P. et. al. 1998. Føroysk orðabók. Før-oya Fróðskaparfelag.

Rice, C. 2006. “Optimizing gender”. Lingua 116, pp.1394-1417.

Steinmetz, D. 1985. “Gender in German and Icelandic:Inanimate nouns”. Papers from a Symposium at theUniversity of Chicago, pp. 10-18.

Steinmetz, D. 1986. “Two principles and some rules forgender in German: Inanimate nouns”. Word 37, pp.189-217.

Steinmetz, D. 1997. “The great gender shift and theattrition of neuter nouns in West Germanic: The

example of German”. In I. Rauch and G. Carr (eds.)New insights in Germanic Linguistics II, pp. 201-224. New York: Peter Lang.

Trosterud, T. 2001. “Genustilordning i norsk er regel-styrt”. Norsk Lingvistisk Tidskrift 19, pp. 29-58.

Trosterud, T. 2006. “Gender Assignment in Old Norse”.Lingua Vol. 116/9, pp. 1441-1463.

Notes1 It is not a universal principle, however, as studies

have shown, that formal assignment may take prece-dence over the principle: Preserve gender (Corbett1999:70ff).

2 When referring to a human, the noun is feminine einmodell ‘a model’.

3 There are only three masculine nouns (two actually)with word final -a, babba, pabba ‘father’ and harra‘sir.’ Note that this speaks in favor of SAR >> MAR>> PAR as -a should be associated with the feminineor neuter, as in native kona-f. ‘woman’ and nativeeyga-n. ‘eye’.

4 An anonymous reviewer has pointed out that inhis/her language buffe is feminine. That is a pos-sibility. I think it is neuter, so the noun is, as seen, ina transition period.

5 This noun is in a transition period. All three gendersare found.


114

Gender in Language Contact

Documents

Transcript of Gender in Language Contact