1、Age of Acquisition and Proficiency as Factors in Language Production: Agreement in Bilinguals,Rebecca Foote February 21, 2007 Univ. of Illinois at Urbana-Champaign,2,Central question in SLA,Why is the outcome of second language (L2) acquisition different from that of first language (L1) acquisition?
2、 Specifically, why is success inevitable in L1 acquisition but failure to become nativelike the standard for L2 learners (late bilinguals)?,3,Age effects in language acquisition,Critical period hypothesis Critical / sensitive period for language learning: “a limited developmental period during which
3、 it is possible to acquire a language, be it L1 or L2, to normal, nativelike levels” (Birdsong, 1999) But does this apply to all aspects of language acquisition? Are there particular linguistic structures or psycholinguistic processes that are not susceptible to age effects?,4,What is it thats not n
4、ativelike about L2 learners (late bilinguals)?,Variability in late bilingual production of agreement morphology: Example: Jim runs to the store. *Jim run to the store. Two possible sources for the problem: Deficit in linguistic knowledge (competence) Beck (1998), Eubank, (1994), Hawkins & Chan, (199
5、7), Meisel (1990, 1997), Schachter (1989, 1990) Psycholinguistic processes of language production (performance) Lardiere (1998a, 1998b), Prvost & White (2000), Fernndez (1999),5,What about early bilinguals?,Dont seem to show same morphological variability as late bilinguals Dont always pattern with
6、monolinguals Kohnert, Bates, & Hernndez (1999), Montrul (2002) Are they more accurate in language processing and production than late bilinguals? What if they are not of “nativelike” proficiency?,Experimental study: Subject-verb number agreement production in English-Spanish bilinguals,7,Subject-ver
7、b number agreement production in Spanish and English,Verb number controlled by the number of the subject noun phrase The cats are black. Los gatos son negros. The lamp is beautiful. La lmpara es bella. But how do we account for subject-verb number agreement errors (The occurrence of these errors are
8、 underestimated)? Agreement potentially influenced by: Grammatical number of the subject noun phrase “Notional” or “conceptual” number of the subject noun phrase Attraction Conceptual number and attraction working together,8,Grammatical vs. conceptual number,Grammatical number Lexically specified nu
9、mber of head noun Conceptual number Number of referent in the speakers mental model one thing or more than one thing? Singular or plural? Jeans Faculty Family Clothes Clothing,9,Attraction,*The tree behind the houses ARE large. Error common to native speakers of both Spanish and English (Bock Viglio
10、cco, Butterworth, & Garrett, 1996),10,Conceptual number and attraction working together,Single referent complex NP: The bus to the schools Distributive referent complex NP: The label on the bottles Distributivity effects in subject-verb number agreement The label on the bottles are vs. The bus to th
11、e schools are,11,Subject-verb number agreement production in Spanish and English,Psycholinguistic research on agreement production Speech error research determine which factors play a role Grammatical number only? Conceptual number as well? Bock & Miller (1991) Sentence-fragment completion methodolo
12、gy “The label on the bottles”,12,Cross-linguistic differences in subject-verb number agreement production,English: Grammatical number (no distributivity effects) Bock & Miller (1991) The label on the bottles are = The bus to the schools are Spanish: Conceptual number as well as grammatical number (d
13、istributivity effects) Vigliocco, Butterworth, & Garrett (1996) The label on the bottles are The bus to the schools are,13,Maximalist model of agreement production (Vigliocco, Butterworth, & Garrett, 1996),Interactive model that assumes a language production system with various levels or steps (Leve
14、lt, 1989; 1999) Message conceptualization Grammatical encoding Morphophonological encoding Phonetic encoding Articulation Accuracy of agreement production ensured by systems use of the maximal amount of information available to it; information across levels or steps is shared Accounts for cross-ling
15、uistic differences in the influence of conceptual number on agreement production by proposing different mechanisms of agreement for different languages English: Feature copying Spanish: Feature unification,14,Maximalist model of agreement production feature copying (from Vigliocco, Butterworth, & Ga
16、rrett, 1996, p. 270),15,Maximalist model of agreement production feature unification (from Vigliocco, Butterworth, & Garrett, 1996, p. 271),16,Maximalist model of agreement production cross-linguistic differences,Languages with richer inflectional morphology, flexible word order and the possibility
17、of null subjects are more likely to produce agreement by means of feature unification and therefore to show conceptual effects in agreement production Languages with poorer inflectional morphology, less flexible word order and little possibility of null subjects are more likely to produce agreement
18、by means of feature copying, and therefore to show no conceptual effects in agreement production,17,Agreement production in bilinguals,How do cross-linguistic differences play out in bilingual agreement production? Cause of variability in late bilingual production of agreement morphology? Proficient
19、 bilinguals appear to process and produce agreement in the same way in both of their languages Nicol, Teller, & Greth (2001) Nicol & Greth (2003),18,Experimental study,Experiment 1 Agreement production in English and Spanish monolinguals Experiment 2 Agreement production in English-Spanish bilingual
20、s,Experiment 1: Monolinguals,20,Experiment 1: Research Question,Are there differences in the mechanics of subject-verb number agreement production in Mexican Spanish, Dominican Spanish, and English? Previous research indicates cross-linguistic differences between Spanish and English (Vigliocco, Butt
21、erworth, & Garrett, 1996) with respect to the factors that influence singular or plural agreement Grammatical number Conceptual number,21,Experiment 1: Participants,English monolinguals (n = 18) Mexican Spanish monolinguals (n = 32) Dominican Spanish monolinguals (n = 28),22,Experiment 1: Materials,
22、Language history questionnaire Experimental task Modified sentence-fragment completion task,23,Experiment 1: Stimuli,Auditory stimuli 2 sets (1 English, 1 Spanish), 32 complex subject noun phrases 16 with distributive referents as head nouns The label on the bottles 16 with single referents as head
23、nouns The bus to the schools 64 filler subject noun phrases The chairs Pictorial stimuli Each fragment associated with a line drawing of its referent Example: The stamp on the envelopes,24,Experiment 1: Procedure,Participants see a picture in the center of the computer screen, followed by a sentence
24、 fragment played over the speakers Repeat and complete fragment as quickly and accurately as possible, naming the color of the head nouns referent,25,Experiment 1: Example trial,26,Experiment 1: Results,Participants responses digitally recorded, transcribed, and coded for errors Predictions If parti
25、cipants are sensitive only to grammatical number, they will make approximately the same number of errors with single and distributive referent fragments. The label on the bottles are = The bus to the schools are If participants are sensitive to conceptual number, they will make more errors with dist
26、ributive referent fragments than with single referent fragments. The label on the bottles are The bus to the schools are,27,Experiment 1: Results Are participants sensitive only to grammatical number, or to conceptual number also?,The label on the bottles are The bus to the schools are,28,Experiment
27、 1: Research Question,Are there differences in the mechanics of subject-verb number agreement production in Mexican Spanish, Dominican Spanish, and English? NOAll three languages/dialects evidence conceptual effects in agreement production,29,Experiment 1: Discussion,English, Dominican Spanish, and
28、Mexican Spanish all show sensitivity to conceptual number in agreement production In terms of the maximalist model (Vigliocco, Butterworth, & Garrett, 1996), all three effect agreement by means of feature unification,Experiment 2: Bilinguals,31,Experiment 2: Research questions,Do English-Spanish bil
29、inguals produce subject-verb number agreement in the same manner in both languages? What role do age of acquisition and proficiency play in bilingual agreement production? Manner Accuracy,32,Experiment 2: Participants,Early English-Spanish bilinguals (age of acquisition of both languages 11) Interme
30、diate proficiency 52 Advanced proficiency 18,33,Experiment 2: Materials,Language history questionnaire Proficiency test(s) Experimental task Modified sentence-fragment completion task,34,Experiment 2: Stimuli & Procedure,Same as Experiment 1 Completed in both English and Spanish different list each
31、time, sessions at least 1 week apart,35,Experiment 2 Results: Manner Are English-Spanish bilinguals sensitive only to grammatical number when producing English?,The label on the bottles are The bus to the schools are,36,Experiment 2 Results: Manner Are English-Spanish bilinguals sensitive to concept
32、ual number when producing Spanish?,The label on the bottles are The bus to the schools are,37,Experiment 2 Research questions: Manner,Do English-Spanish bilinguals produce subject-verb number agreement in the same manner in both languages? YES What role do age of acquisition and proficiency play in
33、bilingual agreement production? Manner NONEall groups are sensitive to conceptual number in both Spanish and English,38,Experiment 2 Discussion: Manner,Results support NO cross-linguistic differences in how bilinguals produce agreementconceptual number affects agreement processes in both English and
34、 Spanish regardless of age of acquisition or proficiency level. In terms of the maximalist model (Vigliocco, Butterworth, & Garrett, 1996), bilinguals effect agreement by means of feature unification in both of their languages, no matter the AoA or proficiency level,39,Experiment 2 Results: Accuracy
35、 AoA Are early bilinguals more accurate in agreement production than late bilinguals?,40,Experiment 2 Results: Accuracy Proficiency Are advanced bilinguals more accurate in agreement production than intermediate bilinguals?,41,Experiment 2 Research questions: Accuracy,What role do age of acquisition
36、 and proficiency play in agreement production in English-Spanish bilinguals two languages? Accuracy More proficient bilinguals are more accurate, no matter whether late or early Early bilinguals less accurate than late bilinguals,42,Experiment 2 Discussion: Accuracy,Advanced proficiency bilinguals m
37、ore accurate than intermediate proficiency bilinguals Early bilinguals NOT more accurate than late bilingualsother way around. Contrary to an across-the-board interpretation of the CPH,43,Central question in SLA,Why is the outcome of second language (L2) acquisition different from that of first lang
38、uage (L1) acquisition? Variability in late bilingual production of agreement morphology Psycholinguistic processes of language production (performance) as a source of the variability The aspects of the agreement production system investigated are not responsible for the variability at the level of i
39、nflectional morphology present in bilinguals Is there really a problem with morphological variability in late bilingual agreement production? Subject-verb number agreement vs. adjective agreement Adjective agreement in Spanish: La etiqueta en las botellas es roja.,44,Subject-verb number agreement vs
40、. adjective agreement errors,45,Conclusion,Age effects in second language acquisition should be investigated in terms of specific linguistic structures or psycholinguistic processes Younger may not always be better!,46,Acknowledgements,Silvina Montrul J. Kathryn Bock Ann Abbott Melissa Bowles Jonathan Foote Justin Davidson Dept. of Spanish, Italian, and Portuguese, University of Illinois at Urbana-Champaign University of Illinois Graduate College,47,Thank you!,Questions? Comments?,
