Interlanguage Variation of Chinese Learners of English

1
Interlanguage Variation of Chinese Learners of
English

• D. Victoria Rau
• Providence University

2
Definition of Interlanguage

• Interlanguage is an L2 learners multi-competence
  (Cook 1992, 1996) which has its own grammar
  (Selinker 1972) and is subject to systematic
  variation, as any natural language (Labov 1969).
• The L2 target is the standard variety of the
  language that the learning community adopts.

3
Research Designs for the Study of Interlanguage
Variation
4
Chinese English interlanguage,VARBRUL program
(Table 1)

• Bayley (1991, 1994, 1996)
• 20 Chinese adults
• Linguistic variable past tense marking
• Independent variables verb type, preceding
  segment, following segment, aspect, proficiency,
  social network, interview type

• Young (1989a, 1991)
• 12 Chinese adults
• Linguistic variable plural inflectional morpheme
  -s
• Independent variables developmental, semantic,
  syntactic, phonological

5
Current studies on Chinese English Interlanguage

• Phonology (r) (S. Chen 2001), (th) (Chang 2002,
  Rau in progress)
• Morphology past tense marking (P. Chen 2002),
  article (H. Chen 1998)
• Syntax relative pronoun (H. Chen Tai 2003)

6
Current studies in Taiwan

• P. Vicky Chen (2002)
• 37 Chinese university students
• Linguistic variable past tense marking
• Independent variables verb type, lexical aspect,
  type of writing

• S. Adeline Chen (2001)
• 9 Chinese university students
• Linguistic variable consonant (r)
• Independent variables phonological environment,
  speech style, proficiency, L1 transfer

7
Effect on English (r) Internal and external
factors

• (1) word position (e.g., red, very, jar, bring,
  short)
• (2) vowels following word-initial /r/ (e.g.,
  river, room, rock)
• (3) vowels preceding word-final /r/ (e.g., dear,
  there, your, door, jar)
• (4) consonants preceding post-consonantal /r/
  (e.g., bring, from, tree, three, crimson)
• (5) vowels following post-consonantal /r/ (e.g.,
  free, group, promise)
• (6) vowels preceding pre-consonantal /r/ (e.g.,
  fearful, short, yard)

• (1) speech style (word list reading and story
  retelling)
• (2) English oral proficiency level (high, mid,
  and low)
• (3) how participants pronounce Mandarin (r)
  (r/?, z, and l).

8
Major findings of Chinese production of consonant
(r)

• (1) The VARBRUL weights can be transformed into
  an environment continuum to indicate an
  acquisition hierarchy (Table 3).
• (2) Speech style variation (i.e., attention paid
  to form) is stable across groups regardless of
  proficiency (Table 2). Word list reading
  promotes accurate production of /r/ while story
  retelling inhibits it.
• (3) Mid to low proficiency and L1 production of
  Mandarin /r/ has a strong effect on the
  inaccurate production of /r/ in L2 (Table 2).

9
Environment continuum

• Acquisition hierarchy of /r/ (Table 3)
• Sounds occurring in heavier environments are
  mastered earlier than those in lighter
  environments (Bailey, 1973).

10
Acquisition Hierarchy of (r)
11
(No Transcript)
12
(No Transcript)
13
(No Transcript)
14
(No Transcript)
15
English proficiency and L1 production of /r/
combined

• Group A High
• High /r/ (.98)
• High /z/ (.98)
• High /l/ (.92)
• Mid /r/ (.97)
• Mid /z/ (.90)

• Group B Low
• Mid /l/ (.00)
• Low /r/ (.10)
• Low /z/ (.06)
• Low /l/ (.00)

16
Comparisons Between Two Proficiency Groups (Table
4)
17
Comparisons Between Two Proficiency Groups (Table
4)
18
Markedness Differential Hypothesis (MDH)

• Unmarked phenomena are acquired before marked
  phenomena.
• Front vowels following onset cluster with /r/
  (e.g., free) promote accurate production of
  English /r/ whereas round (e.g., group) and low
  vowels (e.g., promise) inhibit it regardless of
  proficiency levels.
• Environments themselves can be in markedness
  relationships (Carlisle 1994).

19
Post-Vocalic /r/ (3. Vowels preceding word-final
/r/)
20
Post-Vocalic /r/ (6. Vowels preceding
pre-consonantal /r/)
21
Chronological Corollary of the Ontology
Phylogeny Model (OPM)

• As IL develops chronologically, L2 increases, L1
  decreases, and U (UG) increases and then
  decreases.
• In the cases of post-vocalic /r/, the low
  proficiency group demonstrates L1 transfer or
  canonical CV syllable by leaving out /r/
  completely, while the high proficiency group
  shows variation determined by the markedness of
  the environments.
• High or front vowels (e.g., dear, there, your,
  fearful) promote accurate production of /r/ while
  non-high or round vowels (e.g., door, jar, short,
  yard) inhibit it.

22
Low Proficiency Learners

• In the high proficiency group, L2 has reached a
  threshold that the position of /r/ does not
  affect its production accuracy, but in the low
  proficiency group, it still does.
• Intervocalic position (e.g., very) promotes
  accurate production of /r/ whereas word-initial
  (e.g., red) and post-vocalic /r/ (e.g., jar,
  short) inhibit it.

23
Effects of Word Position on /r/ of Low
Proficiency Learners
24
Constraint on Onset Clusters
25
Obligatory Contour Principle (OCP) and Markedness

• The presence of coronal sounds (e.g., tree,
  drink, three) inhibit accurate production of /r/
  in the low proficiency group.
• In the high proficiency group, the most marked
  interdental fricative still inhibits accurate
  production of /r/ in three.

26
Reordering of Constraints

• E. Broselow et al. (1998). The emergence of the
  unmarked in second language phonology, SSLA 20
  261-280.
• L. Lombardi. (2003). Second language data and
  constraints on manner Explaining substitutions
  for the English interdentals. SLR 19.3225-250.

27
Emergence of the UnmarkedDeletion of
Post-Vocalic /r/

• High proficiency group Faithfulness gtgt
  Markedness
• Low proficiency group Markedness gtgt Faithfulness

28
Stylistic Variation
29
Imbalanced or Informed Design?

• The source of variation is linguistic (Preston
  1996).
• The range of several linguistic factors is larger
  than the proficiency level range, and the
  proficiency level range is larger than the
  stylistic range (Bayley 1991, S. Chen 2001,
  Preston 2002).

30
Insights

• (1) The concept of L1 membership should be
  clearly defined. Negative transfer is apparent
  when /r/ is substituted by /l/ in rock by
  Mandarin speakers whose L1 /r/ is realized as
  l.
• (2) The concept of L2 target language is not
  stable.
• (3) Each linguistic variable has its own history,
  the meaning of which should be explained under
  its own sociocultural contexts.
• (4) Speech community L2 norm free, fink (HK) vs.
  sree, sink (Taiwan) for three and think
  (http//classweb.gmu.edu/accent/)

31
Reconsidering L1

• Substitutions for English interdentals tend to be
  consistent based on L1 t for speakers of
  Russian (high ranked markedness), s for
  speakers of Japanese (high ranked faithfulness)
  (Lombardi 2003).
• We should go a step beyond auditory salience and
  weight to ask why L1 substitution of (th) in EF
  is s while that in QF is t (Brannen 2002).

32
Future Directions

• A more balanced design in interlanguage variation
  is called for.
• Add a socio-psychological dimension to
  interlanguage variation.
• Longitudinal studies based on the variationist
  model on SLA are absolutely needed to account for
  how a learners multicompetence changes over
  time.

33
Attitude Questionnaire

• Please rank the following five sounds, s, f,
  t, ?, and ?, as in three and think, from 1
  (most acceptable) to 5 (least acceptable).
• ? 1
• f
• t
• ?
• s

34
A theory of interlanguage variation

• Empirically verifiable
• Account for the existence of all the factors of
  variation (linguistic context, psychological
  processing factors, interlocutor, topic and
  social norm, and function-form relationships)
• Address longitudinal process of SLA
• Consistent in maintaining key theoretical
  distinctions
• Psycholinguistically plausible and
  sociolinguistically solid

35
The End