Attainment of Syntactic and Morphological Accuracy by Advanced Language Learners

1
Attainment of Syntactic and Morphological
Accuracy by Advanced Language Learners
Kathleen Bardovi-Harlig Theodora Bofman, 1989
Presented by Shuyuan Yang Eun-Young
Kwon Shirin Murphy
2
Motivation Methods
3
Motivation

• Previous studies of grammatical development in
  SLA are nonintersecting sets of studies
• The examination of formal features
• morphological development
• syntactic development
• The investigation of learners overall progress
• global L2 proficiency
• across various levels of learners

4
Morphological development

• Dulay and Burt (1973, 1974a, 1974b) for children
• Bailey, Madden and Krashen (1974) for adults
• Krashen, Houck, Giunchi, Bode, Birnbaum, and
  Strei (1977) for adults
• Kayfetz (1982) for adults

5
Syntactic development

• Relative clausesSchachter (1974) and Gass (1980)
• The pro-drop parameterWhite (1985) and Phinney
  (1987)
• Preposition stranding and pied pipingMazurkewich
  (1984) and Bardovi-Harlig (1987)

6

• Learners overall progress

• Larsen-Freeman (1973, 1983), Larsen-Freeman
• et. al (1977), Gaies (1976), and Monroe
  (1975)
• The widely used measure
• the average number of words per error-free
  T-unit

Syntactic complexity
Accuracy
7
The present study intends to

• examine the relationship between syntactic
  development, or complexity, and overall accuracy
  evidenced in the written English of advanced
  adult language learners.
• combine aspects of these previous studies in
  that it seeks to describe these three aspects,
  morphology, syntax, and error measurement in
  other words, overall accuracy with regard to the
  distribution of errors.

8
Participants

• 30 advanced EFL learners
• 5 native languages
• Learners were further divided into Pass/Non-pass
  subgroups

Table 1. Participants
9
Procedure

• Compositions written as part of the Indiana
  University Placement Exam
• Allotted time 45 minutes
• Topics nontechnical in nature

• The reason for using written texts instead of
  oral
• production

• more likely to be evaluated in detail
• comparable to the oral samples
• exhibiting similar evidence of development

10
Data Analysis

• The essays were scored along two parameters

• Syntactic complexity
• the average number of clauses per T-unit

• Surface errors
• syntactic
• morphological
• lexical-idiomatic

11
T-unit

• Devised by Hunt (1965)
• a minimal terminable unit
• one main clause with all the subordinate clauses
  attached to it

12
Lets practice together!

• one main clause with all the subordinate clauses
  
• attached to it

1. He likes to sleep and to catch frogs.
A simple sentence with a compound predicate ?
one T-unit
2. Hes Shirins cat whose name is Alyosha.
A complex sentence ? one T-unit
3. He likes to eat, but he hates fruit.
A compound sentence consisting of two main
clauses ? two T-units
13
Is this a syntactic, a morphological or a
lexical-idiomatic error?

I had an English teacher who had been travel a
lot of country.
I had an English teacher who had traveled to a
lot of countries. tense, preposition, plural
14
Morphological errors

• include errors in inflectional morphology (i.e.
  nominal morphology and verbal morphology), and
  errors in determiners, articles and prepositions
• the errors in derivational morphology were
  recorded separately

15
Is this a syntactic, a morphological or
a lexical-idiomatic error?
She loved me very much, so did I.
She loved me very much, and I loved
her. coordination
16
Syntactic errors

• including errors of word order, errors resulting
  from the absence of major and minor constituents,
  and errors in combining sentences

17
Is this a syntactic, a morphological or a
lexical-idiomatic error?
Since the date when oil have been explored, the
government of Saudi Arabia started to assess the
foundation for an industrial nation.
When oil began to be exploited, the government of
Saudi Arabia started to assess the foundation
for an industrial nation. tense,
lexical-idiomatic
18
Lexical-idiomatic errors

• include lexical-idiomatic or vocabulary errors

19
RESULTS
20
The complexity data and the analysis of errors
show that the interlanguage of these advanced
learners -- in both groups -- exhibits relatively
strong syntax but incomplete and variable
acquisition of grammatical morphemes.
21
Complexity
22
pass non-pass ? similar complexity scores

• Both groups same lower range (1.6
  clauses/T-unit)
• Non-pass group higher upper complexity scores
  (3.3 clauses/T unit)
• Pass group (2.5 clauses/T-unit)

23
Table 2. T-unit complexity scores for pass and
non-pass groups
24
Different L1 backgrounds ? same frequency of
syntactically complex T-units
Table 3. Frequency and complexity of combined
T-units by language
25
Analysis of Errors
26
Distinction between Pass Non-pass (by the
number of errors)
pass lt non-pass
27
Overall errors / clause

• Pass 456/751 lt Non-Pass 641/744

28
Errors in the error categories
Table 4. Distribution of errors per clause for
pass and non-pass groups
Pass lt Non-pass
29
Learners L1
? No significant effect on the distribution of
errors
Table 5. Errors per clause for all errors by
language group
30
Similarities between pass non-pass

• Distribution of errors by percentage of total
  number of errors ? same for both groups

Table 6. Distribution of errors I pass and
non-pass groups ( of total)
Syntactic lt Lexical idiomatic lt Morphological
31
Distribution of error typesa. Errors in
grammatical morphemes

• Nominal morphology
• Verbal morphology

Equal frequency in both groups
32
Table 7.
Distribution of errors in grammatical morphemes
by percentage of occurrence for pass and non-pass
groups
33
Distribution of error types b. Errors in syntax

• Word-order errors with
• Major constituents
• ex I apple eat
• Minor constituents adverb, intensifier
  placement
• ex I go always to a church

34
Table 8. Distributions of syntactic errors by
percentage of occurrence for pass and non-pass
groups
35
General findings in RESULTS

• The complexity data and the analysis of errors
  show that the interlanguage of these advanced
  learners -- in both groups -- exhibits relatively
  strong syntax but incomplete and variable
  acquisition of grammatical morphemes.

Syntactic lt Lexical idiomatic lt Morphological
36
DISCUSSION
37
Two possible interpretations of these results

• Communicative
• vs.
• Formal

38
Communicative

• Acquisition is driven by meaning and learner
  perception of salience
• Burt Kiparskys (1972) division of errors into
• Global overall sentence organization, including
  word order and missing, incorrect, or misplaced
  sentence connectors
• Local single elements or constituents, include
  nominal and verbal morphology, articles, and
  auxiliaries

39
How to apply to this study
?
Global
Syntactic
?
Local
Morphological
This interpretation is supported by Van Pattens
(1984, 1985) findings that learners process first
for meaning and then for form
40
Which group (pass/non-pass) is more
communicatively successful in the acquisition of
lexical and idiomatic items?
Table 9. Distribution of errors in pass and
non-pass groups ( of total)
The pass group!
41
A problem for the communicative interpretation
Errors in verb tense Subjects made more errors
in tense use (tense switching) than in tense form
(tense formation, subject-verb agreement)
Possible reason?
Transfer-of-training from instruction that
emphasizes formal accuracy over meaning and use
42
Formal

• (driven largely by findings from FLA)
• Long (1987) and Goldin-Meadow (1982)
• Learners select certain features of language
  from the input and acquire those features earlier
  and/or better

syntax
In the case of this study, _______seems to
precede __________.
morphology
43
Resilient Fragile properties
Goldin-Meadow (1982) divides language into
resilient and fragile properties Argues from a
critical period perspective
Resilient properties weeds
Fragile properties hothouse orchids
44
Resilient Fragile Properties comparison

• Based on data from deaf children who developed
  home sign and children who grew up in
  linguistically deprived conditions
• Genie (Curtiss, 1977)
• Newport, et al. (1977)

45
Resilient Properties

• Word-order production rules, constituent
  structure, recursion (the means for expressing
  more than one proposition in a single sentence)
• Can still be acquired after critical period or
  under conditions of deprivation

46
Fragile properties

• Movement rules, pro-forms, auxiliary structures,
  etc.)
• Can no longer be acquired after critical period
  or under conditions of deprivation

47
In this study..
?
Resilient
Syntactic
?
Fragile
Morphological

• Resilient (syntactic) features of language are
  not influenced by background variables such as
• L1 background, learning history, type of input,
  or pass/non-pass

• Fragile (morphological) features account for most
  of the errors across groups and are more
  susceptible to influence from background variables

48
Directions for future research
Four-way comparison between ESL and EFL child and
adult learners regarding fragile features
49
Questions motivating this four-way comparison

• Can fragile properties be learned after the
  critical period?

• Can the critical period (or sensitive period)
  hypotheses account for the lower morphological
  accuracy among advanced learners?

50
Question about this four-way comparison
If the answers to the previous questions are yes,
then do other variables such as ESL vs. EFL play
a role?
NO!
51
The biggest challenge to both the communicative
and the formal interpretations
Slobins (1982) hypothesis that individual
languages are balanced in terms of syntactic vs.
morphological difficulty(also based on FLA data)
Crosslinguistic study comparing early child
language acquisition among speakers of English,
Italian, Serbo-Croatian, Turkish, and Japanese
52
Motivation for Slobins study

• Naturalness hypothesis
• (Slobin calls it an Anglo-centrist or
  Indo-european-oriented position)
• (ie Bloom, 1973)
• claims that children learning English acquire
  syntax before morphology because word order is a
  natural reflection of the order of thought

53
But..

• Slobins study shows that children learning
  Turkish (a flexible, pragmatic word-order
  language in which grammatical functions are
  encoded in bound morphemes) acquire morphology
  before syntax and show overall
• faster and greater language acquisition in early
  stages than children learning strict word-order
  languages such as English and Italian

54
Reasons for Turkish childrens precocious
language acquisition

• (a selection of his reasons)
• In Turkish
• Morphology is highly regular
• Distinct morpheme for each grammatical function
• Morphology is consistent and obligatory, never
  optional
• Morphology is bound to the noun
• Morphemes are syllabic and stressed

55
Also

• Turkish children have the advantage of following
  local cues (decoding inflections of a single word
  regardless of position in sentence) rather than
  global cues that require learners of word-order
  languages to process the entire sentence to
  determine function of words

56
However.

• At later stage of L1 acquisition, Turkish
  children are slower than learners of
  Indo-European languages with regard to features
  such as relative clause constructions, which
  require complex syntax in Turkish

57

• Slobins findings support hypothesis of equal
  difficulty of acquisition across languages
• in some languages, morphological ease (or
  regularity) is counterbalanced by syntactic
  complexity (ie Turkish) whereas in other
  languages, syntactic ease (or regularity) is
  counterbalanced by morphological difficulty (ie
  English)

58
How does this pertain to the current study?

• The authors suggest that English may be easy
  __________ but difficult _____________ .

syntactically
morphologically
Communicative and formal interpretations are
learner-based, whereas Slobins position is
language-based.
59

• In order to test learner-based vs. language-based
  positions, the authors suggest a study comparing
  adult learners of Turkish (and other languages)
  from a variety of L1 backgrounds

60
Learner-based vs. language-based

• Support for learner-based position
• Learners have greater facility acquiring global
  or resilient features despite regularity of
  Turkish morphology
• Support for language-based position
• Learners acquire Turkish morphology early,
  accurately, and uniformly (as do Turkish children
  during FLA)