The influence of L1 syntax on L2 processing
L1
Alice Foucart
ESRC Centre for Research on Bilingualism, Bangor University – 7th March 2011
L2
Outline:
Part 1

Gender processing in L2: review

Experiments gender processing in L2 (ERPs and eye-tracking)
Part 2

Agreement processing in L2 production: conceptual gender in
possessive structures
Conclusions
Grammatical gender
English:
the car
French:
lafem voiturefem
Spanish:
elmasc cochemasc
German:
dasneu Autoneu
Research questions:

Is gender represented and processed in a similar manner in
native and non-native speakers?

Is gender processing in L2 affected by the L1?
Gender processing in L1
Example:
Lafem / *lemasc cleffem était dans la serrure
The key was in the keyhole
 P600 effect (syntactic integration)
(Barber & Carreiras, 2005; Foucart & Frenck-Mestre, 2004; Gunter et al., 2000; Hagoort & Brown
1999)
 Sometimes preceded by a LAN (morpho-syntactic violations)
(Barber & Carreiras, 2005; Gunter et al., 2000)
=
Gender is syntactically processed
Comprehension in L2: ERPs studies
Difficulties in lexical-semantic integration

Similar but with reduced amplitude and later peak latency
(Ardal et al. 1990; Hahne & Friederici, 2001; Weber-Fox & Neville, 1996)

L2 learning (Osterhout et al., 2006)
Word category violations

Sometimes found, and often with different topography
(Hahne, Müller & Clahsen, 2006; Rossi et al., 2006; Weber-Fox & Neville, 1996)
Morpho-syntactic violations
Syntactic integration

Found in some studies, but not in other
(Foucart & Frenck-Mestre, 2004; Hahne, 2001, but Hahne & Friederici, 2001;
Weber-Fox & Neville, 1996).
Semantic processing: similar
≠
Syntactic processing: Proficiency and AoA
Syntactic processing in L2
- Different L1 and L2 processing
Shallow structure hypothesis (Clahsen & Felser, 2006)
 The syntactic analysis processed by late L2 speakers during language
comprehension is not as ‘full’ as that processed by native speakers.
- Similar L1 and L2 processing
Three-phases model (Friederici, 2002; Rossi, Gugler, Friederici, Hahne, 2006 )
 With enough exposure, highly proficient L2 learners can reach native-like
processing levels even if they learned their L2 late in life.
Sabourin & Stowe’s (2008) study
Participants:
Dutch native speakers, German-Dutch and Romance-Dutch learners
Materials:
Determiner-noun gender agreement violation
e.g.
Hetneu/*Decom kleine kindcom probeerde voor het eerst te lopen.
The small child tried to walk for the first time
Results:
Dutch native speakers and German-Dutch learners
→ P600 effect
Romance-Dutch learners
→ No effect
Conclusions:
Automatic gender processing in L2 not only depends on the presence of a
grammatical gender system in the L1 but also requires overlapping of lexical
gender.
Sabourin, L., & Stowe, L A. (2008) Second language processing: When are L1 and L2 processed similarly. Second Language Research, 24 (3), 397-430.
Foucart & Frenck-Mestre’s (2011) study
Participants:
French native speakers and German-French advanced learners
Materials:
Determiner-noun gender agreement violation
e.g.,
Hier lafem/*lemasc chaisefem était dans le salon;
Yesterday the chair was in the living room
Results:
French native speakers and German-French advanced learners
→ P600 effect
Conclusions:
With enough exposure, proficient later learners can process gender in a nativelike manner.
Foucart, A. & Frenck-Mestre, C. (2011). Grammatical gender processing in L2: Electrophysiological evidence of the effect of L1 - L2 syntactic
similarity. Bilingualism: Language and Cognition, DOI:10.1017/S136672891000012X
Tokowicz & MacWhinney’s (2005) study
Participants: English-Spanish learners
Materials:
(a) tense-marking (similar in English and Spanish)
e.g., Su abuela *cocinando/cocina muy bien
His grandmother *cooking/cooks very well.
→ P600 effect
(b) nominal number agreement (different in English and Spanish)
e.g., *El/Los niños están jugando
The boys are playing.
→ No effect
(c) nominal gender agreement (unique to Spanish)
e.g., Ellos fueron a *unmasc/unafem fiestafem
They went to a party.
→ P600 effect
Conclusions:
- Online sensitivity to gender agreement in L2 learners despite the absence of
grammatical gender in their L1.
- Features that are not present in L1 should be acquired faster than those that
are in conflict with L2 features.
Tokowicz, N., & MacWhinney, B. (2005). Implicit and explicit measures of sensitivity to violations in second language grammar: An event related potential
investigation. Studies in Second Language Acquisition, 2, 173–204.
Gillon Dowens, Vergara, Barber, Carreiras’ (2009) study
Participants:
Spanish native speakers, English-Spanish late learners (+ 12 yrs exposure)
Materials:
Determiner-noun gender agreement violation
e.g.,
Elmasc /*lafem /*losplur suelomasc-sing está plano y bien acabado,
The floor is flat and well-finished
Noun-predicative adjective gender agreement violation
e.g.,
El suelo masc-sing está planomasc/*planafem/*planosplur y bien acabado
Results:
Determiner-noun gender agreement violation
Similar pattern for native and non-native speakers
→ LAN + P600 effects
Noun-predicative adjective gender agreement violation
Native speakers → LAN + P600 effects;
L2 speakers
→ P600 effect
Conclusions:
- With enough exposure, late L2 learners can reach native-like processing even
when agreement involves features that do not exist in the learner’s L1.
- Processing non-local agreement may be more demanding in L2 (working memory).
Gillon-Dowens, M., Vergara, M., Barber, H. A., Carreiras, M. (2010). Morpho-syntactic processing in late L2 learners. Journal of Cognitive
Neuroscience, 22 (8), 1870–1887
Keating’s (2009) study
Participants:
Spanish native speakers, English-Spanish beginner, intermediate and
advanced learners.
Materials:
(a) in the determiner phrase
Una casafem pequeñafem/* pequeñomasc [cuesta mucho en San Francisco.],
A small house costs a lot in San Francisco.
(b) in the verbal phrase
La casafem [es bastante pequeñafem /*pequeñomasc y necesita muchas reparaciones.]
The house is quite small and needs a lot of repairs.
(c) in a subordinate clause
Una casafem [VP cuesta menos [CP si [VP es pequeñafem/*pequeñomasc y necesita
reparaciones]].
A house costs less if it is small and needs repairs.
Keating, G. D. (2009). Sensitivity to violations of gender agreement in native and nonnative Spanish: An eye-movement investigation. Language
Learning, 59, 503-535.
Keating’s (2009) study
Results:
(a) in the determiner phrase
Beginner and Intermediate learners:
Native speakers and advanced learners
→ no effect.
→ longer fixation times for violations
(b) in the verbal phrase
L2 learners:
Native speakers
→ no effect.
→ longer fixation times for violations
(c) in a subordinate clause
L2 learners:
Native speakers
→ no effect.
→ longer fixation times for violations
Conclusions:
- High proficient late L2 learners can acquire new abstract grammatical
features such as gender in their L2.
- Processing non-local agreement may be more demanding in L2 because of
the working memory cost.
Keating, G. D. (2009). Sensitivity to violations of gender agreement in native and nonnative Spanish: An eye-movement investigation. Language
Learning, 59, 503-535.
L2 gender processing:
Evidence from ERPs and eye-tracking studies
- Gender processing in L2 can only reach native-like level if the syntactic structures
are similar in L1 and L2.
(Sabourin & Haverkort, 2003; Sabourin & Stowe, 2008)
- Features that are not present in L1 (e.g., gender for native English speakers)
should be acquired faster than those that are in conflict with L2 features.
(Gillon Dowens et al. 2010; Tokowicz & MacWhinney, 2005)
- With enough exposure, late L2 learners can reach native-like processing even
when agreement involves features that do not exist in the learner’s L1.
(Foucart & Frenck-Mestre, 2011; Gillon Dowens et al., 2009; Keating, 2009; Tokowicz & MacWhinney,
2005)
- L2 learners may not be able to process gender as ‘fully’ as native speakers in
non-local contexts (e.g., predicative adjectives)
(Gillon Dowens et al., 2010; Keating, 2009)
Experiment 1: Method Foucart, A. & Frenck-Mestre, C. (under revision), JML.
Participants (N=14):
French native speakers
English-French advanced learners
German-French advanced learners
Materials:
96 nouns, inanimate
Same/different gender (French, la cleffem; German, der Schlüsselmasc; the key)
Electrodes (21 scalp sites):
- Central, Fz, Cz, Pz
- Frontal, central and posterior areas
of the left hemisphere,
Fp1, F3, F7, Fc5, C3, Cp5, T5, P3, O1
- of the right hemisphere,
Fp2, F4, F8, Fc6, C4, Cp6, T6, P4, O2
Experiment 1: Conditions
Conditions:
Correct
Incorrect
En ce moment les pommesfem vertesfem sont de saison.
En ce moment les pommesfem vertsmasc * sont de saison.
At the moment green apples are in season
Predictions:
+
?
 For French native speakers and
the two L2 groups
Experiment 1: Results
French native speakers
English-French bilinguals
German-French bilinguals
French native speakers: P600 effect
 Gender syntactically processed
English-French bilinguals: P600 effect
 Gender syntactically processed
German-French bilinguals: No effect
? New syntactic structure
? Plural agreement rule in L1
Experiment 2: Method and Predictions
Participants (N=14):
Same as Experiment 1
Materials and procedure:
Same as Experiment 1
Conditions
Correct
Incorrect
Souvent les / petitesfem/ pommesfem/ sont/ bien/ sucrées.
Souvent les / petitsmasc*/ pommesfem/ sont/ bien/ sucrées.
Usually small apples are very sweet
Predictions:
 For French native speakers and EnglishFrench bilinguals
 ? German-bilinguals
Experiment 2: Results
French native speakers
English-French bilinguals
German-French bilinguals
French native speakers: P600 effect
 Gender syntactically processed
English-French bilinguals: N400 effect
 Still acquiring gender agreement with
pre-posed adjective
German-French bilinguals: No effect
X New syntactic structure
 Plural agreement in L1
Experiments 3: Method and Predictions
Participants (N=14):
Same as Experiments 1 and 2
Materials and procedure:
Same as Experiments 1 and 2
Conditions
Correct
Incorrect
En automne les /pommesfem/ sont/ vertesfem/ sur cet/ arbre.
En automne les /pommesfem/ sont/ vertsmasc*/ sur cet/ arbre.
In autumn apples are green on this tree.
Predictions:
 For French native speakers and EnglishFrench bilinguals
 BUT no effect for German-bilinguals
Experiment 3: Results
French native speakers
English-French bilinguals
German-French bilinguals
French native speakers: P600 effect
 Gender syntactically processed
English-French bilinguals: No effect
? Gender agreement more complex when
not in a local context
German-French bilinguals: No effect
Experiments 4: Method and Predictions
Participants (N=14):
Same as Experiments 3
Materials and procedure:
Same as Experiments 3
Conditions
Correct
Incorrect
Predictions:
R1
R2
R3
R4
R5
R6
En automne les /pommesfem/ sont/ vertesfem/ sur cet/ arbre.
En automne les /pommesfem/ sont/ vertsmasc*/ sur cet/ arbre.
In autumn apples are green on this tree.
 Longer fixation times for
French native speakers
 ? English-French bilinguals
 BUT no effect for Germanbilinguals
Experiment 4: Results
R1
R2
R3
R4
R5
R6
En automne les /pommesfem/ sont/ vertesfem-vertsmasc*/ sur cet/ arbre.
In autumn apples are green on this tree.
 French native speakers: Longer reading times for incorrect agreement
 English-French bilinguals: Same pattern as the native speakers
 German-French bilinguals: No effect
Experiments 1-4: Conclusions

Is gender represented and processed in a similar manner in
native and non-native speakers?

Late bilinguals can attain native-like processing (P600, reduced amplitude)
HOWEVER:


Proficiency (N400, complex syntactic structure)

Non local processing seems to be more difficult (working memory).
Is gender processing in L2 affected by the L1?

Influence of the native language (gender vs. absence of gender)
Further research… in production
Is processing of conceptual gender agreement as difficult as processing of
grammatical gender agreement in L2 ?
Does L1 syntax affects the processing of L2-English gender agreement?

Possessives adjectives/pronouns processing in L2 English
Possessives adjective agreement rules
Possessive adjective agreement
Agree with the possessor (waitress):
English:
The waitress chases herFem / *hisMasc son
Greek:
I servitora kiniga ton gio tisFem / *touMasc

Agree with the possessee (son)
Spanish:
La camarera persigue a suMasc / *suFem hijo
French:
La serveuse poursuit sonMasc / *saFem fils
Experiment 5: Method (Santesteban, Foucart, Pickering, Branigan, 2010)
Participants:
12 English native speakers
24 Spanish-English bilinguals (AoA: 7.4; proficiency: 5.6 (self-rate: 1-low, 7-high))
24 French-English bilinguals (AoA: 10.6; proficiency: 5.5)
24 Greek-English bilinguals (AoA: 9.1; proficiency: 5.9)
Materials:
32 experimental pictures (8 per condition) + 72 fillers
Possessor characters: Masculine (e.g., boxer) vs. Feminine (e.g., waitress)
Possessee characters: Masculine (e.g., son) vs. Feminine (e.g., daughter)
Conditions:
2 (Possessor: masc vs. fem) x 2 (Possessee : masc vs. fem)
Gender match
The boxer chases his son (masc-masc)
The waitress chases her daughter (fem-fem)
Gender mismatch
The boxer chases his daughter (masc-fem)
The waitress chases her son (fem-masc)
Agreement errors
Similarly to number (Bock & Miller, 1991), the gender of a local noun can interfere with
agreement process and result in gender agreement errors called ‘attraction
errors’ (e.g., noun-adjective, Vigliocco & Franck, 1999; Franck et al., 2008; or antecedentpronoun, Meyer & Bock, 1999)
Number:
e.g.,
The picture of the postcards *are/is nice.
Gender:
e.g.,
El relojMasc de la ciudadFem es *viejaFem,/ viejomasc
The clock of the town is old
Hypotheses
Do L1 gender agreement rules affect conceptual gender agreement
production of L2 possessive pronouns?

If L2 processing is affected by the ‘weaker’ L2 syntactic representations
of bilinguals, we expect no differences between bilingual groups.

If L1 agreement rules affect L2 processing, we expect larger gender
attraction effects for Spanish- and French-English bilinguals than for GreekEnglish bilinguals.
Experiment 5: Method (Santesteban, Foucart, Pickering, Branigan, 2010)
Experiment 5: Results (Santesteban, Foucart, Pickering, Branigan, 2010)
Errors
Match
Mismatch
10
5
0
SpanishEnglish
FrenchEnglish
GreekEnglish
English
natives
% of mismatch effects
(mismatch minus match)
% of agreement errors
15
Summary of attraction effects
15
10
5
0
Spanish- FrenchEnglish
English
GreekEnglish
 English monolinguals made no errors.
 Only bilinguals showed gender attraction effects
 Interestingly, similar attraction effects produced by all groups of
bilinguals (no significant mismatch x group interaction)
English
Experiment 5: Conclusions (Santesteban, Foucart, Pickering, Branigan, 2010)

L2 gender agreement errors during production of possessive
adjective seems to be due to ‘weaker’ syntactic representations,
not effects of L1 syntax
HOWEVER, there was no linear intervention of the possessee attractor.
Possessives pronoun agreement rules
Possessive pronoun agreement
Agree with the possessor (waitress):
English:
The waitress says that the son is hersFem / *hisMasc
Greek:
I servitora lei oti o gios ine dikos tisFem / *touMasc

Agree with the possessee (son)
Spanish:
La camarera dice que el hijo es suyoMasc / *suyaFem
French:
La serveuse dit que le fils est le sienMasc / *sienneFem
Experiment 6: Method (Santesteban, Foucart, Pickering, Branigan, 2010)
Participants:
21 English native speakers
21 Spanish-English bilinguals (AoA: 9.8; English proficiency: 5.5)
21 French-English bilinguals (AoA: 11.5; English proficiency: 5.3)
21 Greek-English bilinguals (AoA: 7.6; English proficiency: 5.8)
Materials, Conditions and Procedure:
Similar as Experiment 1
Gender match
The boxer says the son is his (masc-masc)
The waitress says the daughter is hers (fem-fem)
Gender mismatch
The boxer says the daughter is his (masc-fem)
The waitress says the son is hers (fem-masc)
Experiment 6: Method (Santesteban, Foucart, Pickering, Branigan, 2010)
Experiment 6: Results and Conclusions (Santesteban, Foucart, Pickering, Branigan, 2010)
% of agreement errors
15
Match
Mismatch
10
5
0
SpanishEnglish
FrenchEnglish
GreekEnglish
English
natives
% of mismatch effects
(mismatch minus match)
Errors
15
Summary of attraction effects
10
5
0
Spanish- FrenchEnglish
English
GreekEnglish
English

All groups showed gender attraction effects when the possessee attractor linearly
intervened between the antecedent and the pronoun.

Importantly, Spanish/French-English bilinguals produced more errors, and showed
larger gender attraction effects than Greek-English bilinguals and English
monolinguals.

Similar attraction effects for Greek-English bilinguals and English monolinguals.
 Production of L2 possessive pronouns is affected by L1 syntax.
Experiments 5 & 6: Conclusions (Santesteban, Foucart, Pickering, Branigan, 2010)
 Processing of possessive structures is affected by the gender of the
possessee
- Native speakers:
Gender attraction errors only occur with linear intervention of the
possessee (however, this might only be a sentence complexity effect)
- Bilingual speakers:
Gender attraction errors occur independently of the position of the
possessee in the sentence.
 At least part of bilinguals’ gender agreement error production is due to
‘weaker’ syntactic representations.
 However, L1 agreement rules that differ from L2 rules do also affect gender
agreement processing in L2.
General conclusions

Is L2 processing influenced by L1 syntax?
YES!
(Foucart & Frenck-Mestre, 2011; Sabourin, 2003; Santesteban et al., 2010)
HOWEVER:
- Proficiency
(Foucart, 2008; Osterhout et al., 2006; Santesteban et al., 2010)
- Complexity of the structure
(Gillon Dowens et al., 2010; Hahne, 2001; Keating, 2009)
THANK YOU!
English:
the end
French:
lafem finfem
Spanish:
elmasc finmasc
German:
dasneu Endeneu