LINGUIST List 14.1845

Wed Jul 2 2003

Diss: Lang Acquisition: Shi: 'Second Language...'

Editor for this issue: Naomi Fox <foxlinguistlist.org>

Directory

1. enchao.shi, Second Language Grammar and Secondary Predication

Institution: University of Arizona
Program: Second Language Acquisition and Teaching Program
Dissertation Status: Completed
Degree Date: 2003

Author: Enchao Shi 

Dissertation Title: Second Language Grammar and Secondary Predication

Linguistic Field: Language Acquisition 

Subject Language: English (code: ENG)

Dissertation Director 1: Andrew Barss
Dissertation Director 2: Douglas Adamson

Dissertation Abstract: 

This thesis presents empirical evidence for second language grammar or
the final L2 state. Following the minimalist spirit and the Continuity
Hypothesis, we propose that L2 grammar or G2 is essentially
non-distinct from the native final state (G1) (cf. Flynn and Lust
2002). Several assumptions are critical to the study, one of which is
that grammatical theory is not deducible from analysis of
developmental data (cf. Chomsky 1985, 1999). A second assumption is
that the characterization of the initial L2 state is a matter of
logic, not merely a developmental issue. Another assumption is that
competence is inaccessible to analysis (cf. Chomsky 1988, Moravcsik
1998); extra care is warranted to avoid research artifacts

 We propose that G2 constitutes two components: a Computational
system and a lexicon, and G1 is external to the formation of G2. As
such, we argue that UG constitutes the initial L2 state and the
development of an L2 lexicon proceeds primarily independent of G1, a
hypothesis we called the CHL2 Uniformity Hypothesis (CUH). However, we
propose that L2 development is sometime mediated by the Interpretation
Device (ID), which translates L2 sensory data into an L1 for
comprehension. The ID is a processing mechanism and its use is
constrained by the Relativized Transfer Condition (RTC), an
acquisition procedure. Given that the ID is largely not amendable to
analysis, we put the RTC to test. We also wanted to find out whether
adult L2 learners are conservative.

 
 The experiments focused on the English resultatives (Mary
painted the house red) and depictives (John ate the meat raw) - the
former but not the latter is available in Mandarin. Nineteen Mandarin
speakers of English and nineteen native speakers of English
participated. The L2 subjects had lived in the United States for an
average of 10 years and 5 months. Four tests were administered: the
Guided Production (GP) test, the Clause-combining (CC) test, the
Grammaticality Judgment (GJ) test, and the Interpretation (IT)
test. Results, derived via application of t-test, one-way ANOVA, and
factorial ANOVA, are of significance. L2 subjects showed knowledge of
resultatives and depictives on a par with the controls on most
tests. L2 subjects nonetheless produced fewer resultatives and
depictives than the controls on the GP test, but no difference has
been found with the canonical constructions. We attribute such
irregularities to the modalities of measurements and functions of RTC,
because between-group differences were drastically reduced to
insignificance on the other tests (CC, GJ, and IT tests), where
choices for constructions were directed to resultatives and
depictives. This result indicates that the final L2 state coincides
with that of the native speakers; hence evidence for the CUH. This
result also led us to a logical conjecture that the linguistic and
learning mechanisms that lead to the final L2 state must be the same
as those employed by the native speakers, a result compatible with the
Continuity Hypothesis. Furthermore, the study shows that the RTC
constrains adult L2 development, since L2 knowledge was indeed
affected by the linguistic canonicality. Finally, we conclude that
adult L2 learners are relatively conservative.