Acquisition of Grammatical Gender - PowerPoint Presentation

Slide1

Acquisition of Grammatical Gender in Latvian-Russian Bilinguals

Olga Urek, Marit Westergaard & Agrita Tauriņa

Project Nr.

NFI/R/2014/053

L

atvian

language in monolingual and bilingual acquisition: tools, theories and

applicationsSlide2

IntroductionLAMBA project – EEA Norway grants WP4:

Experimental study of grammatical and (morpho)phonological acquisition by monolingual and bilingual learners of LatvianExperiment 3: Acquisition of morphosyntactic properties: adjectival gender agreementThe experiment focuses on the acquisition of adjectival gender agreement by monolingual Latvian-speaking children and bilingual learners of Latvian and Russian.In this talk: only Russian data (monolinguals and bilinguals)Slide3

Gender in RussianThree-gender system, nouns of different genders unequally distributed (Corbett 1991, Polinsky 2008):

Masculine – 46%Feminine – 41%Neuter – 13% Gender agreement as a suffix on adjectives, demonstratives, possessive pronouns, past tense verbs, etc. Mo-ya bel-aya kurtk-a My-FEM white-FEM jacket ‘My white jacket’ Mo-y bel-iy stol-Ø My-MASC white-MASC table‘My white table’Slide4

Gender in RussianGender assignment: mostly transparentRule of thumb:

-C in Nom. sg = masculine (e.g. dom ‘house’)-á in Nom. sg = feminine (e.g. ruka ‘hand’)-ó in Nom. sg = neuter (e.g. vedro ‘bucket’)In some instances, gender assignment is opaque:-C´ in Nom. sg = masculine (e.g. gus´ ‘goose’) = feminine (e.g. rys´ ‘lynx’)-V in Nom. sg = feminine (e.g. partʌ ‘desk’) = neuter (e.g. sitʌ ‘sieve’) Slide5

Gender in Russian: AcquisitionIn monolinguals, gender assignment in transparent nouns is largely acquired by the age of 2;5 – 2;7 (Gvozdev

1961, Polinsky 2008)Errors with opaque nouns persist up to the age of 7 (ibid):-C´ in Nom. sg = masculine (e.g. gus´ ‘goose’) = feminine masculine (e.g. rys´ ‘lynx’)-V in Nom. sg = feminine (e.g. partʌ ‘desk’) = neuter feminine (e.g. sitʌ ‘sieve’) masculine (e.g. jablok vs. jablokʌ ‘apple’)Are bilingual speakers qualitatively and quantitatively similar?Slide6

Previous StudiesPolinsky (2008):Gender assignment in lower-proficiency heritage Russian speakers in the US;

Results: gender system undergoes re-analysisAll -C´ nouns reinterpreted as masculine;All -V nouns are reinterpreted as feminine;for some speakers, also suffix-stressed neuters are reinterpreted as feminine (correlated with proficiency (speech rate) in Russian);Slide7

Previous StudiesRodina & Westergaard (2016):

Gender assignment in Norwegian-Russian bilingual children in Norway;Two groups of bilinguals: RR: only Russian at home; NR: both Russian and Norwegian at home;Results: RR & L1 children: -C´ nouns reanalyzed as MASC -V nouns reanalyzed as FEMNR children: all FEM and NEUT reanalyzed as MASC (cf. Polinsky 2008)Possible reduction of gender system linked to the amount of exposure;Slide8

This study: research questionsWill differences between mono- and bilingual children be mainly quantitative?

Do we find changes/reductions in the gender system of bilinguals?How does the amount of cumulative exposure affect the acquisition of gender in transparent gender systems?Slide9

This study: predictionsQuantitative differences between monolinguals and bilingualsQuantitative differences between bilinguals depending on the cumulative amount of exposure to a given language

;Qualitatively similar gender acquisition in bilinguals and monolingual learners;More problems with opaque than transparent nouns;Masculine agreement overused with feminine opaque nouns;Feminine agreement overused with neuter opaque nouns;Slide10

This study: stimuliBased on Rodina & Westergaard (2016)

5 Transparent & opaque picturable nouns of each genderTransparentOpaqueMasculineglaz ‘eye’rul´ ‘steering wheel’

Feminine

grusha

‘pear’

ten´

‘shadow’

Neuter

pero

‘feather’

uho

‘ear’Slide11

This study: procedureFollowing Rodina & Westergaard (2016):

Experimenter: “That´s what we call ‘an apple’. What color are they?”Child: “A blue-NEUT apple and a red-NEUT apple”Experimenter: “And now, what has disappeared?”Child: “The blue-NEUT apple”Slide12

This study: procedureFollowing Rodina & Westergaard (2016):

Experimenter: “That´s what we call ‘an apple’. What color are they?”Child: “A blue-NEUT apple and a red-NEUT apple”Experimenter: “And now, what has disappeared?”Child: “The blue-NEUT apple”Slide13

Participants42 monolingual Russian-speaking childrenmean age 4.1 years

SD = 0.5329 monolingual Latvian-speaking childrenmean age 4.0 yearsSD = 0.4218 bilingual Latvian-Russian speaking childrenmean age 5.2 yearsSD = 0.84Slide14

ParticipantsRecruited in kindergartens in Riga area, Latvia:Latvian monolinguals both parents speak Latvian natively;

Latvian is the only language of instruction in the kindergarten;Russian monolingualsboth parents speak Russian natively;Russian is the only language of instruction in the kindergarten;caveat: Russian-speaking children are exposed to Latvian via L2 teaching + through a community Bilingualsmixed families;Latvian is the language of instruction in the kindergarten;UBiLEC to measure cumulative exposure to both languagesSlide15

Composition of population by ethnicityCentral Statistical Bureau of Latvia, 2013Slide16

Share of Ethnic Latvians by Region (2011 population census)Slide17

Language mainly used in the household2011 Population CensusSlide18

Latvian and Russian language skills2000 Population Census:

of 1 383 105 native-speakers of Latvian, 75% (1 038 723) reported some knowledge of Russian;of 891 451 native-speakers of Russian, 56% (495 619) reported some knowledge of Latvian;395000-strong de-facto monolingual Russian community (≈17% of total population);≈ 1 530 000 individuals with varying degrees of command of both Latvian and Russian ( ≈ 78% of total population) Slide19

Accuracy rates by condition (Russian)Adjective-noun agreement in Russian (% of correct responses)

TransparentOpaqueMF

N

M

F

N

RUS-MON

99.5%

(378/380)

94.3%

(446/473)

83.0%

(234/282)

95.3%

(406/426)

75.4%

(239/317)

84.2%

(251/298)

RUS-BIL

99.2%

(249/251)

94.5%

(290/307)

71.1%

(133/187)

89.0%

(227/255)

74.8%

(178/238)

77.8%

(147/189)Slide20

PredictionsMore problems with opaque than transparent nouns;

Masculine agreement overused with feminine opaque nouns;Feminine agreement overused with neuter opaque nouns;Slide21

Monolinguals: RussianSlide22

Monolinguals: RussianEffect of OpacitySlide23

PredictionsMore problems with opaque than transparent nouns;

YES – lower accuracy rates for opaque MASC and FEM nounsMasculine agreement overused with feminine opaque nouns;Feminine agreement overused with neuter opaque nouns;Slide24

Monolinguals: RussianErrors

***Slide25

Monolinguals: RussianErrors

*Slide26

Monolinguals: RussianErrorsSlide27

PredictionsMore problems with opaque than transparent nouns;

YES – lower accuracy rates for opaque MASC and FEM nounsMasculine agreement overused with feminine opaque nouns;YES Feminine agreement overused with neuter opaque nouns;NO – no significant difference in the erroneous use of feminine agreement between NEUT-TRANS and NEUT-OP Slide28

Monolinguals: RussianEffect of Gender

**Slide29

Monolinguals: RussianEffect of GenderSlide30

Bilinguals: RussianSlide31

Bilinguals: RussianEffect of Opacity

*****Slide32

PredictionsMore problems with opaque than transparent nouns;

YES – lower accuracy rates for opaque MASC and FEM nounsMasculine agreement overused with feminine opaque nouns;Feminine agreement overused with neuter opaque nouns;Slide33

Bilinguals: RussianErrors

***Slide34

Bilinguals: RussianErrors

**Slide35

Bilinguals: RussianErrorsSlide36

PredictionsMore problems with opaque than transparent nouns;

YES – lower accuracy rates for opaque MASC and FEM nounsMasculine agreement overused with feminine opaque nouns;YES Feminine agreement overused with neuter opaque nouns;NO – no significant difference in the erroneous use of feminine agreement between NEUT-TRANS and NEUT-OP Slide37

Bilinguals: RussianEffect of Gender

*******Slide38

Bilinguals: RussianEffect of Gender

**.Slide39

RU Monolinguals vs. Bilinguals: Transparent ItemsSlide40

PredictionsQuantitative differences between monolinguals and bilingualsQuantitative differences between bilinguals depending on the cumulative amount of exposure to a given language;

Qualitatively similar gender acquisition in bilinguals and monolingual learners;Slide41

RU Monolinguals vs. Bilinguals: Transparent Items

**Slide42

RU Monolinguals vs. Bilinguals:Opaque ItemsSlide43

RU Monolinguals vs. Bilinguals:Opaque Items

.Slide44

PredictionsQuantitative differences between monolinguals and bilingualsPARTIALLY – monolinguals are significantly better at NEUT-TRANS and MASC-OP items

Quantitative differences between bilinguals depending on the cumulative amount of exposure to a given language;Qualitatively similar gender acquisition in bilinguals and monolingual learners;Slide45

RU Bilinguals:Accuracy vs. Cumulative ExposurePearson´s product-moment correlation (t = 2.1462, df

= 17, p-value = 0.04658) Slide46

PredictionsQuantitative differences between monolinguals and bilingualsPARTIALLY – monolinguals are significantly better at NEUT-TRANS and MASC-OP items

Quantitative differences between bilinguals depending on the cumulative amount of exposure to a given language;PARTIALLY - there is a marginal correlation between accuracy and cumulative amount of exposure to RussianQualitatively similar gender acquisition in bilinguals and monolingual learners;Slide47

Error patterns in Russian:Monolinguals vs. BilingualsSlide48

Error patterns in Russian:Monolinguals vs. BilingualsSlide49

Error patterns in Russian:the Effect of Cumulative ExposureSlide50

Use of agreement in Russian: B2 vs. B1 childrenRodina & Westergaard (2016)

B2 childrenB1 childrenSlide51

PredictionsQuantitative differences between monolinguals and bilingualsPARTIALLY – monolinguals are significantly better at NEUT-TRANS and MASC-OP items

Quantitative differences between bilinguals depending on the cumulative amount of exposure to a given language;PARTIALLY - there is a marginal correlation between accuracy and cumulative amount of exposure to RussianQualitatively similar gender acquisition in bilinguals and monolingual learners;NO – bilinguals (especially those with lower exposure) are significantly more likely to overuse masculine agreement with NEUTcf. Polinsky (2008) & Rodina & Westergaard (2016)Slide52

Summary & ConclusionsOpaque nouns are more problematic for both monolinguals and bilinguals;Bilinguals fare worse than monolinguals on NEUT-TRANS and MASC-OP nouns;

Accuracy in bilinguals correlates with cumulative language exposure;Qualitative difference: Bilinguals (especially those with lower exposure) overuse MASCULINE agreement with neuter nouns (cf. Rodina & Westergaard 2016)Slide53

Thank you!Slide54

Accuracy rates by condition (Latvian)Adjective-noun agreement in Latvian (% of correct responses)

MasculineFeminineDecl. 1Decl. 2Decl. 4Decl. 5Decl. 6LAT-MON88.4% (425/481)88%(417/474)91.4%(497/544)88.3%(409/463)56.6%(275/486)LAT-BIL85.6%(285/333)85.6%(280/327)96.2%(381/396)98.8%(324/328)36.1%(120/332)TRANSPARENT

OPAQUE