Is Syntax Separate or Shared Between Languages

1
Is Syntax Separate or Shared Between Languages?

• Cross-Linguistic Syntactic Priming in
  Spanish-English Bilinguals
• 
  - Robert J. Hartsuiker
• 
  Martin J. Pickering
• 
  Eline Veltkamp
• Eunah Kim
• 
  March 26, 2008
  

2
A Fundamental Question in Bilingualism

• Do bilinguals have separate stores for the two
  languages, or do they have a single store for at
  least some aspects of language?

3
Shared or Separate Lexicon?

• Most research into this question has considered
  the representations of words or concepts (e.g.,
  Kroll Stewart, 1994 McElree, Jia, Litvak,
  2000).
• Previous work on the bilingual lexicon
• At least some information is shared between
  languages (Colomé, 2001 Costa, Miozzo,
  Caramazza, 1999 Spivey Marian, 1999 Van
  Heuven, Dijkstra, Grainger, 1998).

4
Shared or Separate Syntax?

• However, whether bilinguals have shared or
  separate syntax is a different question. Lexical
  and syntactic information are usually assumed to
  be separated.
• ? A direct experimental investigation of
  bilingual syntax is needed.
• This study examined whether bilinguals have two
  entirely separate syntactic stores
  (separate-syntax account), or whether at least
  some syntactic information is shared between the
  languages (shared-syntax account).

5
Syntactic Priming in Bilinguals

• The study tested the bilingual shared vs.
  separate syntax accounts using the phenomenon of
  syntactic priming.
• Syntactic priming a speaker who processes a
  particular grammatical form displays a tendency
  to repeat that form. Many studies on monolinguals
  found the effects of syntactic priming (e.g.
  Bock, 1986).
• Shared-syntax account predicts cross-linguistic
  syntactic priming, but the separate-syntax
  account does not.
• Loebell and Bock (2003)
• Priming between alternative forms of English
  datives and their German equivalents
• No priming between English actives/passives and
  their German equivalents

6
The Experiment

• Language Spanish-to-English priming
• Construction active / passive
• Participants
• 24 native speakers of Spanish
• All spoke English as an L2
• All were resident in Edinburgh
• (length of residence 2 months - 7 years (avg.
  22 months)
• Moderate or high proficiency in English
• 1 female confederate (also a Spanish-English
  bilingual)

7
Confederate-scripting Technique
Sp. sentence
picture
picture
picture Eg. verb
8
Confederate-scripting Technique

• A modification of Branigan, Pickering, and
  Clelands (2002) method.
• Two bilingual interlocutors one confederate and
  one naive participant.
• The two interlocutors alternated in describing
  their cards to each other. After either
  participant finished speaking, the other took the
  topmost card from his or her selection box and
  determined whether it matched the description
  just heard or not.
• The naive participant described pictures to the
  confederate, and the confederate pretended to
  describe pictures to the naive participant, but
  in fact read aloud scripted sentences.
• The confederate spoke Spanish (prime sentences)
  and the native participant spoke English.

9
Materials

• Naive participants description cards
• a picture depicting an action an English verb.
• 128 cards - 32 target pictures 96 filler
  pictures
• Target pictures were designed to increase the
  likelihood of obtaining a reasonable number of
  passive responses in English.
• ? patient (left 16 animate/16 inanimate)
• agent (right inanimate)

hit
hit
approach
10
Materials

• Confederates description cards
• Spanish sentences.
• 128 cards in one session - 32 prime sentences
  96 filler sentences
• Prime sentences
• 4 conditions
• active transitive /passive transitive /
  intransitive/ OVS
• 32 sets
• 8 sentences per set (4 conditions X 2 sentences)
• Within each condition, half of the prime
  sentences had an agent and patient of the same
  animacy as the corresponding target picture, and
  half had an agent and patient with the opposite
  animacy.
• 8 counterbalanced lists by selecting a single
  prime sentence from each set.

11
Materials

• Order of prime sentences/ target pictures
• conferderate 3 fillers prime 3
  fillers prime ...
• Naive participant 3 fillers target 3
  fillers target ...
• The nouns and verbs in prime and target were
  never translation equivalents and were never
  related in meaning or form.

12
Scoring

• From the recording, the naive participants
  descriptions on experimental trials were
  orthographically transcribed.
• Scoring
• active
• S (agent of the picture) / V / O (patient of the
  picture)
• Alternative passive form should be possible.
• passive
• S (patient of the picture) / to be / participle /
  by-phrase (agent in the picture)
• Alternative active form should be possible.
• other
• sentences with passive morphology but without the
  by-phrase
• errors

13
Results

• 282 active (45)/ 221 passive (35)/ and 121
  other (19)
• Proportion of passives (out of actives and
  passives) for each prime condition
• Significant main effect of prime
• F1(3, 69) 5.20, p lt .01 F2(3,75)
  3.64, p lt .02
• Simple contrasts comparing the conditions with
  each other
• significant difference in

14
Discussion

• The results demonstrate cross-linguistic
  (Spanish-English) syntactic priming between
  production and comprehension in the context of
  interactive language use.
• Implication (At least for highly proficient
  Spanish-English bilinguals living in a culture in
  which English is dominant,) sentence form is
  shared between Spanish and English.

15
Discussion

• Pickering and Branigan (1998 monolingual
  lexical-syntactic representation)
• Combinatorial information such as the
  types of arguments a verb takes are represented
  at the lemma stratum. Lemma nodes are linked to
  combinatorial nodes (as well as other nodes), and
  those nodes are shared between lemmas.

16
Discussion

• Extension to bilingual lexical-syntactic
  representations Lemmas for English and Spanish
  verbs are connected to the same category nodes
  and to the same combinatorial nodes.

El camión es perseguido por el taxi (The truck
is chased by the taxi)
hit
17

• Thank you!