Null-prep
.....5) *Who did John give the book?
.....6) *The man who John gave the book
was Kevin.
By contrast, Klein (1993a) reports, L1 studies of the development of
piping and stranding constructions have rarely detected the null-prep
phenomenon.
On the other hand, null-prep constructions are rare in natural languages.
As Klein (1995a) explicates, null-prep constructions are highly constrained
by UG-that is, they are merely licensed in relative clauses of particular
languages and are generally prohibited in interrogatives due to identification,
licensing and recoverability issues. Nevertheless, the results of a
number of studies of the interlanguage grammar of child and adult L2
learners explicitly exhibit instances of null-prep phenomenon in interrogative
constructions as well as relative clauses (Mazurkewich, 1985; Bardovi-Harlig,
1987; Klein 1993a, 1993b, 1995a, 1995b; Klein and Casco, 1999; Kao,
2001).
Making a comparison between L1 and L2 research, MacLaughlin (1993) highlights
a critical difference between the two and argues against the difference.
He notes that in some L2 research it is presumed that adult L2 learners,
just like L1 learners, have access to UG and it accounts for their interlanguage
as well. However, the results of L2 research are appraised differently
from those of L1 research. He observes that L1 acquisitional data inspire
changes in the linguistic theory whereas L2 acquisitional data do not.
By this is meant that if predictions made based on the linguistic theory
adopted are not supported, the results are interpreted in terms of differences
between L2 and L1 acquisition. As a matter of fact, in contrast with
L1 acquisitional data, "L2 results are rarely, if ever, used to
influence the linguistic theory" (MacLaughlin, 1993, p. 3). He
criticizes this unidirectional view of L2 acquisition and suggests that
L2 data be used as "a viable input to the formulation of linguistic
theory" (MacLaughlin, 1993 p. 9).
Klein and Casco (1999) seem to have taken a step in the right path by
attempting to resolve the controversies originating from the results
of research on preposition pied-piping and preposition stranding and
on the null-prep phenomenon (specifically, Klein, 1995a and Dekydspotter,
Sprouse and Anderson, 1998) by revising the linguistic theory. Still
their alternative proposal also contributes to the illumination of the
differences between L1 and L2 acquisition.
On the other hand, Kaplan and Selinker (1997) argue that exploration
of the nature of the null elements observed in the interlanguage and
their relevance to UG merit further investigation. They report that
"Apparently, there is little serious empirical work investigating
these phenomena, and NO work of any kind…linking them in a clear
theoretical framework" (p. 173).
Accordingly, the nature of interlanguage grammars, in general, and the
null elements, in particular, warrants further empirical investigation.
Thus, the development of pied-piping, preposition stranding and null-prep
constructions lends itself well to such an investigation. Studies done
on the development of pied-piping and stranding constructions in the
interlanguage of L2 learners may shed more light on the nature of the
interlanguage grammars and, as White (1998) posits, would deal with
the role of UG in SLA from a new perspective.
This study tends to probe into the issue of the development of preposition
pied-piping and preposition stranding constructions in the Iranian EFL
learners' interlanguage. It is attempted to see how the Iranian EFL
learners' interlanguage grammar would compare with that of other L2
learners discussed in the literature. In other words, the study is conducted
to specifically find out whether the null-prep phenomenon is significantly
observed in the Iranian EFL learners' interlanguage grammar and whether
the development of preposition pied-piping and preposition stranding
in relative clauses and interrogatives would differ. To study developmental
changes, one could do either a longitudinal or a cross-sectional study.
The former investigates changes over time whereas the latter focuses
upon a cross-section of a population at a single point in time (McBurney,
1998). However, due to practicality considerations, this study is intended
to be a cross-sectional one. Hence, performances of various groups of
learners are also compared to uncover differences in terms of proficiency
levels. Further investigation of the results might furnish sufficient
evidence to assess the competing proposals and may hopefully help to
resolve the controversies.
Studies
on Preposition Pied-piping and Preposition Stranding
Mazurkewich (1985), following Markedness Theory, examines the acquisition
of dative WH-questions in English by second language learners with two
different native languages, namely French and Inuktitut. She notices
a stage when the subjects show evidence of 'no-prep' constructions.
However, she actually dismisses the stage on the ground that the L2
learners at this stage lack subcategorization knowledge of verbs for
their prepositional complements. She asserts that the results of her
study support the theory of markedness for the reason that the two groups
of subjects under study prove to have acquired the dative preposition
pied-piping, the presumably unmarked structure, before preposition stranding,
the marked counterpart. Moreover, a constant increment is observed in
the use of the two structures as the level of competence in English
increases, though instances of pied-piping always outnumber instances
of preposition stranding.
Bardovi-Harlig (1987) criticizes Mazurkewich's (1985) analysis on the
ground that she overlooks a piece of evidence which does not support
her claim: the Inuktitut speakers have actually produced more instances
of preposition stranding in the early stages. Bardovi-Harlig (1987)
also maintains that the early acquisition of pied-piping by French learners
can be due to L1 transfer because the structure meets all the conditions
necessary for transfer to occur, i.e. it is considered to be unmarked,
regular, productive and common. She raises a significant question as
to the role of English itself in the acquisitional sequence. She notes
that there is such a growing tendency in English towards using preposition
stranding instead of preposition pied-piping that many linguists have
considered stranding as the norm and the cases in which stranding is
impossible as exceptions. As such, she comments that salience might
influence the acquisitional sequence of the learners of English.
In an attempt to compensate for the shortcomings of Mazurkewich's study
(1985), Bardovi-Harlig (1987) replicates her test for questions and
expands it to include relative clauses as well so that she can examine
the acquisition of preposition pied-piping and preposition stranding
across two structures, one linguistically simple (dative WH-questions)
and the other linguistically more complex (relative clauses). She administers
the test to a group of subjects who have a variety of native languages
and are at different levels of proficiency in English.
The findings of the study support the stepwise progression from the
simple to the more complex structure, i.e. both constructions first
appear in WH-questions and then in relative clauses. The so-called null-prep
constructions are also employed in both questions and relatives, especially
in the early levels, before pied-piping and stranding are attempted.
However, instances of null-prep diminish with a rise in the proficiency
level. She attributes the emergence of null-prep constructions to lack
of subcategorizational knowledge, just as Mazurkewich (1985) does. The
results of her study further show that preposition stranding, which
is assumed to be the marked structure, is acquired before preposition
pied-piping, the presumably unmarked counterpart. This, she concludes,
suggests that salience is also a determining factor in the acquisition
order.
Van Buren and Sharwood Smith (1985), carrying out a project which is
aimed at exploring the acquisition of preposition stranding in English
and Dutch as a second language, discuss the application of Government
and Binding Theory to second language acquisition. The project is also
intended to investigate the role that the marked and unmarked properties
of the learner's native language and the target language play in the
acquisition process. The results of the studies show evidence of both
pied-piping and stranding as alternatives but they report a general
tendency in favor of pied-piping. In other words, the subjects' responses
to a variety of tests employed do not emulate the widespread presence
of stranding in the input. They argue that although Government Binding
Theory has undeniable potentialities for generating sophisticated linguistic
research questions, the acquisitional aspects of the theory should be
elaborated and adapted so that it can handle the special features of
second language acquisition as well. They suggest that incorporation
of the concept of markedness in Universal Grammar will enhance the precision
of the theories of second language acquisition formulated within the
UG framework.
Klein (1995a) argues that to attest strong UG claims in L2 acquisition,
the learners' interlanguage grammars should be shown to be constrained
by UG, i.e. their grammars should present no evidence of constructions
forbidden in the world's languages. An illicit or 'wild' grammar would
indeed downgrade the role of UG in second language acquisition.
Building upon earlier studies of preposition pied-piping and preposition
stranding in English, Klein (1995a) examines the interlanguage of child
and adult learners acquiring the structures in English and finds considerable
instances of null-prep phenomenon. She notes that the null-prep construction
is severely constrained by UG as it is merely permissible in non-movement
relative clauses while it is absolutely forbidden in interrogative constructions.
Thus, the occurrence of null-prep in L2 questions, as it is observed
in the interlanguage of child and adult learners of English as a second
language, appears to be an unequivocal instance of wild grammar. This
observation runs against the conception that the learner's interim grammars
conform to a natural language system.
Finally, she raises several questions as to the nature of null-prep.
The first deals with the reason why null-prep phenomenon shows up. She
speculates that the construction emerges because learners may not have
learned the licensing and identification restrictions on the constructions.
Positing that null-prep only occurs in combination with Pro while questions
normally require a trace, she attempts to account for the illegitimacy
of null-prep instances in wh-questions. The implication, she speculates,
is that the "L2 learners have (incorrectly) set the licensing parameters
for these null-prep elements but have not yet considered the necessary
identification requirements" (Klein, 1995a, p. 107).
Another significant issue raised by Klein (1995a) is whether null-prep
constructions appear as a general language acquisition process or they
are merely the result of language-specific input factors. In case the
phenomenon is systematically observed in L2 acquisition data, a reformulation
of UG principles seems warranted so that the data may be coped with
more efficiently. Alternatively, she offers, the relationship between
UG and acquisition may need to be revised in the sense that the mandate
for conformity of intermittent grammars with UG is relaxed, thus allowing
for non-conformist grammars as transitional stages. In fact, she favors
the latter suggestion inasmuch as it endorses the role of UG in the
process of L2 acquisition, yet it casts off the strong constraints laid
by UG claims on the interlanguage grammars.
In her study of second language acquisition as opposed to third language
acquisition, Klein (1995b) intends to explore differences between unilinguals
and bilinguals while acquiring English as a second language and third
language, respectively. She compares the acquisition of subcategorisation
knowledge of unilingual learners of English with that of multilingual
learners (lexical learning) as well as their acquisition of preposition
stranding (syntactic learning). She specifically sets out to test the
following hypotheses: 1) Multilinguals learning English will significantly
surpass their fellow unilinguals in subcategorization knowledge. 2)
Both groups will evidence null-prep phenomenon.
It is surprising to note that the data show no instances of pied-piping
in spite of the fact that some of the learners' prior languages demonstrate
this construction and the match is assumed to augment the rate of acquisition.
This observation, indeed, contradicts language transfer. The results
support the first hypothesis thereby attesting the advantage of multilinguals
over monolinguals in lexical learning and parameter setting. The second
hypothesis is similarly supported inasmuch as both groups demonstrate
null-prep, albeit to different degrees.
In their study of English-French interlanguage, Dekydtspotter, Sprouse,
and Anderson (1998) report to have found compelling evidence that rejects
the claims as to the divergence of null-prep grammars from UG constraints.
First, they present some data suggesting that "construals between
interrogative operators and covariant Pros are not banned by UG"
(p. 351). This is supposed to illustrate that null-prep phenomenon falls
within the UG-governed binding options, contra Klein's (1995a) assertion.
They also provide instances of nonmatching PP-gaps in popular French
and Yoruba interrogatives which they interpret as A-bar binding. Moreover,
it is suggested that null-prep manifests familiar argument-adjunct asymmetry.
They suggest that the asymmetry may be accounted for by positing the
rule of preposition incorporation. The rule implies that in null-prep
constructions the category status of the complement of the verb undergoes
a change. In other words, the VP complement becomes an NP rather than
a PP in such constructions. Accordingly, they conclude that null-prep
in L2 development displays a preliminary stage when learners form WH-questions
by binding construals rather than WH-movement.
Klein and Casco (1999) scrutinize L2 learners' grammar with a focus
on the differences between L1 and L2 acquisition and present some evidence
which is suggested to weaken Dekydspotter, Sprouse and Anderson's (1998)
analysis. The results reveal that L2 learners optionally select null-prep
in questions well up to the advanced level. Of course, the number of
correct instances of piping and stranding rises as general proficiency
increases. Robust evidence of optional null-prep exhibited in both adjuncts
and arguments is claimed to provide counterevidence to Dekydspotter,
Sprouse and Anderson's (1998) analysis in that they offer the rule of
preposition incorporation to propel null-prep constructions into UG-constrained
hypothesis space.
However, Klein and Casco (1999) argue that the rule does not hold in
the case of adjuncts inasmuch as incorporation of a preposition into
a verb is banned across a CP as is the case in adjuncts. Refuting Dekydspotter,
Sprouse and Anderson's (1998) analysis of preposition incorporation,
Klein and Casco rationalize that as stranding is quite a rare phenomenon
in most languages of the world, L2 learners' prior exposure to L1 usually
leads to a bias against stranding. Instead, L2 learners attempt null-operator
movement that is more economical as it merely moves the features rather
than an overt operator. This analysis readily permits null-prep as an
interim strategy for adjuncts as well as arguments. Hence, it manages
to place L2 learners' null-prep grammar within the limits permitted
by UG. Yet, they call for further research to help support their analysis.
In her study of the acquisition of English prepositional verbs by Japanese
EFL learners, Kao (2001) further examines Klein's (1995b) report that
preposition stranding is the generally preferred form and her account
that it may be due to "the frequency of stranding in informal language
and the potentially marked nature of piping in English." (p. 198)
Kao warns that Klein's findings should be interpreted with caution since
her targeted sentences raise the possibility of stranding. Her relative
clauses all include the complementizer that instead of the WH-pronouns.
As the subjects have only been required to correct the sentences, they
have been free to change one place (PS) or two (PiP). It is naturally
expected that the subjects favor a minimal correction (inserting a preposition
leading to PS) over a two-step process (inserting a preposition and
changing that to a relative pronoun leading to PiP).
Accordingly, Kao (2001) intends to replicate Klein's study to verify
her claims. Of course, she avoids the pitfall in Klein's study by including
relative pronouns rather than that in her relative clauses. The
results of her study evidence instances of null-prep constructions though
to a far lesser degree than those of Klein's studies. Moreover, stranding
turns out to be the preferred option and the subjects show the preference
by rejecting even correct piping sentences in favor of stranding.
Kao examines six grammar books used by Japanese students and finds that
although both preposition pied-piping and preposition stranding are
introduced, five out of six references recommend students to leave the
preposition at the end of the sentence when the relativized NP is the
object of a preposition. Or in the case of interrogatives, the references
generally state that the WH-word/phrase should be placed at the beginning
of the sentences. Kao speculates that some learners may avoid piping
because piping may seem to violate the grammar rule taught in the books.
Her idea is indeed supported by the results of an ad hoc follow-up test
given to the subjects different from the ones taking part in the main
experiment.
Kao also refers to Sharwood-Smith's (1993, cited in Kao, 2001) distinction
between consciousness-raising and input enhancement and comments that
as the results of her study show, although formal instruction is intended
to help the learners acquire a rule (e.g., prepositional verbs and their
co-occurrence restrictions or wh-question formation), the product the
learner carries away is the presumably preferred structure: preposition
stranding. Accordingly, Kao highlights the effect of instruction on
the learner's performance in the second language.
As it was discussed above, extensive investigations done on the acquisition
of preposition pied-piping and preposition stranding have generally
produced mixed results. Specifically, the two most recent investigations,
Dekydspotter, Sprouse and Anderson (1998) and Klein and Casco (1999),
have made rather contradictory claims about the interim stage in the
development of preposition pied-piping and preposition stranding at
which some L2 learners evidence null-prep constructions.
In an attempt to resolve the controversies discussed above, this study
was intended to further investigate the issue. It specifically sought
to answer the following questions:
1. a.
Do Iranian EFL learners robustly evidence null-prep in relative clauses
and/or questions?
b. How do the interlanguage grammars of Iranian EFL learners at various
levels of proficiency differ in terms of null-prep constructions?
2. a. Are there any differences between questions and relative clauses
in terms of the emergence of preposition pied-piping and/or preposition
stranding and null-prep constructions?
b. How do the interlanguage grammars of Iranian EFL learners at various
levels of proficiency differ in terms of the emergence of preposition
pied-piping and/or preposition stranding and null-prep constructions?
3. Do Iranian EFL learners evidence null-prep in adjuncts and/or arguments?
Method
Participants
A total of 80 adult Iranian male and female EFL learners studying at
Shiraz University Language Center constituted the participants in the
study. The main body of the students at this institute study in conversation
classes ranging from levels 1 to 12. Candidates for the courses at this
institute are initially interviewed. Their performances are then matched
against the contents of the textbooks they are going to study, i.e.,
New Headway English Course (Soars and Soars, 1998). The series
includes 6 volumes, beginner, elementary, lower intermediate, intermediate,
upper intermediate and advanced. On the basis of the volumes studied
at each level, levels 1 to 4 are considered to be beginner, levels 5
to 10 intermediate and levels 11 and 12 advanced.
As one of the goals of the study was to compare the performance of learners
at different proficiency levels, namely, low, mid and high groups, sampling
was carried out in the following way. The participants of the study
were selected from levels 4 (N=30), 8 (N=30) and 12 (N=20) which best
represented the afore-mentioned levels of proficiency, respectively.
It is worth mentioning that the low group was selected from level 4
to make sure that the participants had already been exposed to the constructions
under study, i.e., preposition pied-piping and preposition stranding.
From among all classes at levels 4, 8 and 12, two classes were randomly
chosen, one from the boys' section and the other from the girls' section
in each level, and all the students in each class took the instrument
utilized in the study.
Materials
and Procedure
The instrument for data collection comprised a grammaticality judgment
and correction task. A total of 8 verbs were tested in the task. The
items included an equal number of instances of adjuncts and arguments,
i.e., four verbs with arguments and four with adjuncts. Each verb was
initially presented in a declarative sentence to test the subcategorization
knowledge of the participants. In addition, for each verb two sets of
sentences were developed, one interrogative and the other with a relative
clause, to check the differences between the two structures. Thus there
were a total of 24 targeted sentences. In all the targeted sentences
the preposition was omitted. The task also included 24 incorrect filler
sentences.
The participants were asked to judge the acceptability of the items.
In case a sentence was judged unacceptable, the participants were required
to correct it, for instance, inserting a preposition in the target sentences.
Acceptance of the declarative sentence with no preposition suggested
that the participant did not have the subcategorization knowledge for
the verb and hence his/her responses to the interrogative and the relative
constructions did not count. However, if a participant rejected the
declarative sentence and inserted the missing preposition, then his/her
responses to the interrogative and the relative constructions were further
analyzed as he/she appeared to possess the required subcategorization
knowledge for the verb. In the case of relative and interrogative constructions,
acceptance of the sentence would be considered as a null-prep response.
Alternatively, the result of rejecting the sentence in the test and
inserting the preposition in the initial position of the clause was
considered as piping and rejecting the sentence in the test and inserting
the preposition in the final position was considered as stranding.
Results
and Discussion
The results delineated in Table 1 clearly show that although the subcategorization
knowledge of the Iranian EFL learners increases with their proficiency,
the interlanguage grammar of the participants at various levels of proficiency
who possess the relevant subcategorization knowledge evidence null-prep.
The low group accurately subcategorized 72 declarative sentences. However,
they accepted 66.7 and 76 percent of the corresponding interrogatives
and relatives, respectively, without the preposition. The mid group
accurately subcategorized 113 declaratives. Yet, they accepted 54 and
61.9 percent of the interrogatives and relatives without the preposition.
The high group who had accurately subcategorized 101 declaratives accepted
37 percent of the interrogatives and 44.6 percent of the relatives without
the preposition.
Figure 1 as well as the percentages in Table 1 apparently indicates
that the incidence of null-prep in relative clauses (53.6%) is greater
as compared with interrogatives (46.4%). However, the results of a chi-square
test showed no significant differences between the two structures, i.e.,
interrogatives and relative clauses (X2= 1.66 df=1 p>0.05). As such,
null-prep can be similarly observed in both relatives and interrogatives,
hence providing a positive answer to question 1a posed earlier. As for
question 1b, although the incidence of null-prep seems to decline as
the proficiency level increases (interrogative: Low=66.7% Mid=54%, and
High=37%; relatives: Low=76%, Mid=61.9% and High=44.6%), a pair of chi-square
tests did not show significant differences among proficiency levels
in terms of the use of null-prep in each of the two structures (interrogatives:
X2=5.42 df=2 p>0.05; relatives: X2=5.58 df=2 p>0.05).
Table 1: Null-prep Results by Proficiency Level
| Prof.
Level |
No.
of Part. |
No.
of Subcat. |
Null-prep
Inter. No. (%) |
Null-prep
Relat. No. (%) |
| Low |
30 |
72/240 |
48
(66.7) |
55
(76.4) |
|
Mid |
30 |
113/240 |
61
(54) |
70
(61.9) |
| High |
20 |
101/160 |
38
(37) |
45
(44.6) |
| Total |
80 |
286/460 |
147
(46.4) |
170
(53.6) |
Note: Prof.=proficiency
; Part.=participants; Subcat.=correctly subcategorized
declaratives; Inter.=interrogatives; Relat.=relative clauses
Figure
1: Null-prep in Interrogatives vs. Relatives
Further
examination of the results tabulated in Table 2 shows robust evidence
of null-prep in both questions and relatives (51.4% and 59.4%, respectively)
despite correct subcategorization knowledge indicated in the corresponding
declaratives. The results also suggest that in both structures L2 learners
are more prone to use preposition stranding (interrogatives: 28.3% and
relative clauses: 22.8%) than preposition pied-piping (20.3% in interrogatives
and 17.8% in relative clauses). The results of a chi-square test showed
significant differences between the two structures in terms of the emergence
of the three types of constructions (X2=323.04 df= 4 p<0.01), hence
providing a positive answer to question 2a.
Table
2: Null-prep, Preposition Pied-piping and Preposition Stranding Results
in Interrogatives vs. Relative Clauses
| Structure |
Null-prep
(%) |
PiP
(%) |
PS
(%) |
| Inter. |
51.4 |
20.3 |
28.3 |
| Relat. |
59.4 |
17.8 |
22.8 |
Table
3 depicts the differences among the three proficiency groups in terms
of the emergence of null-prep, preposition pied-piping and preposition
stranding. The table evidently shows that at the three proficiency levels
tested null-prep forms a stage of development prior to the acquisition
of preposition pied-piping and/or preposition stranding in both interrogatives
(Low: 66.7%; Mid: 54%; High: 37.6%) and relative clauses (Low: 76.4%;
Mid: 61.9%; High: 44.6%). However, null-prep gradually diminishes in
both structures as proficiency increases. The table also indicates that
apparently the rise in the use of both preposition pied-piping and preposition
stranding and the fall in the rate of null-prep generally occur in interrogatives
earlier than relative clauses, except for preposition pied-piping in
the mid group and preposition stranding in the high group where a reverse
sequence is observed.
Table
3: Results of Interrogatives vs. Relative Clauses by Proficiency Level
| Prof.
Level |
Null-prep
(%) |
PiP
(%) |
PS
(%) |
| |
Inter. |
Relat. |
Inter. |
Relat. |
Inter. |
Relat. |
| Low |
66.7 |
76.4 |
8.3 |
8.3 |
25 |
15.3 |
| Mid |
54 |
61.9 |
10.6 |
12.4 |
35.4 |
25.7 |
| High |
37.6 |
44.6 |
39.6 |
30.7 |
22.8 |
24.8 |
As for
question 2b, the results of a pair of chi-square tests showed significant
differences among the three proficiency groups in terms of the emergence
of preposition pied-piping and preposition stranding in each of the
structures (interrogatives: X2= 39.38 df=4 p<0.01; relatives: X2=
24.62 df=4 p<0.01). Figures 2 and 3 clearly depict the differences
among the three proficiency groups in interrogatives and relative clauses,
respectively. They present additional information that help to answer
question 2b.
Figure 2 shows that generally in each proficiency group as the occurrence
of null-prep in interrogatives decreases, instances of preposition pied-piping
and preposition stranding increase. Moreover, instances of preposition
stranding outnumber those of pied-piping in both the low and the mid
groups, the high group being an exception. Indeed, an abrupt increase
is observed in the use of preposition pied-piping by the high group
whereas the use of preposition stranding by the same group not only
does not show such a growth but also displays even a sudden decrease.
Figure
2: Null-Prep, Preposition Pied-piping and Preposition Stranding
in Interrogatives by Proficiency Level
Similarly,
Figure 3 reveals the gradual decline in the use of null-prep in relative
clauses and the increment in the use of preposition pied-piping as a
result of the growth in proficiency level. The mid group, as compared
with the low group, shows a similar increase in the tendency to make
use of preposition stranding; yet, the high group exhibits no such tendency.
Figure
3: Null-Prep, Preposition Pied-piping and Preposition Stranding
in Relative Clauses by Proficiency Level
Question
3 also receives a positive answer because further examination of the
data reveals evidence of null-prep in adjuncts (Adj) as well as arguments
(Arg). Table 4 shows that the three proficiency groups attempt null-prep
in both adjuncts and arguments rather similarly (Low: Arg=43% Adj= 57%;
Mid: Arg= 45% Adj=55%; High: Arg=41% Adj=59%). Figure 4 clearly depicts
the null-prep responses of the three proficiency groups in both types
of complement. The results of a chi-square test run to check the differences
among the three proficiency groups in terms of the use of null-prep
in arguments vs. adjuncts show no significant differences (X2=0.326
df=2 p>0.05).
Table 4: Results of Null-prep in Arguments vs. Adjuncts
by Proficiency Level
| Prof.
Level |
Arg
(%) |
Adj
(%) |
| Low |
43 |
57 |
| Mid |
45 |
55 |
| High |
41 |
59 |
Figure 4: Null-Prep in Arguments vs. Adjuncts by Proficiency Level
5.
Summary and Conclusion
Iranian EFL learners in this study evidently omitted the prepositions
in interrogatives and relative clauses in spite of the fact that they
had already demonstrated their knowledge of subcategorization requirements
of the verbs for the missing preposition. Moreover, they attempted null-prep
constructions even in the advanced level although the null-prep occurrence
decreased with the rise in proficiency. This is in line with the findings
of earlier research on preposition pied-piping, preposition stranding
and the null-prep phenomenon (cf. Mazurkewich, 1985; Bardovi-Harlig,
1987; Klein, 1995a, 1995b; Klein and Casco, 1999; Kao, 2001). The systematicity
of the occurrence of the phenomenon in L2 acquisition attests to null-prep
as an undeniable stage in L2 acquisition and thus shows irrelevance
of null-prep to "language-specific input factors" (cf. Klein,
1995a, pp. 107-108).
Another finding of the study, namely the early emergence of preposition
stranding and greater tendency of the participants (in the low and the
mid groups) to use it, is in line with the findings of Van Buren and
Sharwood Smith (1985), Bardovi-Harlig (1987), Klein (1995b) and Kao
(2001). Given that Persian is a language in which preposition pied-piping,
but not preposition stranding, is permissible, one cannot account for
the observation by resorting to prior linguistic knowledge or L1 transfer.
In fact, the early emergence of preposition stranding and the learners'
preference to use it over pied-piping might be due to the salience of
the structure in English (cf. Bardovi-Harlig, 1987). The fact that the
sentences in the test included relative pronouns rather than "that"
as well as the observation of similar null-prep phenomenon in both relative
clauses and interrogatives refutes Kao's (2001) conjecture that learners
attempt it because they choose changing one place (resulting in preposition
stranding) rather than two (resulting in preposition pied-piping). However,
the results corroborate minimalists' view that derivations are directed
by the economy principles and that the most economical alternative always
wins out (cf. Klein and Casco, 1999). As for preposition pied-piping
and preposition stranding, there is a competition between the two because
they comprise the same set of lexical items. Since stranding involves
movement of less material (NP out of PP) than pied-piping (the entire
PP), the most economical derivation winning the competition is preposition
stranding.
On the other hand, a comparison between the number of piping and stranding
constructions attempted by the low and the mid groups showed a smooth
increase in both interrogatives and relative clauses. In addition, instances
of pied-piping outnumbered those of stranding. Nevertheless, the high
group evidenced an abrupt rise in the use of preposition pied-piping
but a sudden fall in the use of stranding. This unexpected change may
be reasonably attributed to the classroom instruction recommending the
prescriptivists' view of the supremacy of preposition pied-piping over
preposition stranding (cf. Kao, 2001).
Moreover, although the gradual reduction in the use of null-prep co-occurred
with a general rise in the use of preposition pied-piping and preposition
stranding in both interrogatives and relative clauses, null-prep was
retained longer in relative clauses. This is in line with Bardovi-Harlig's
(1987) finding that acquiring preposition pied-piping and preposition
stranding in relative clauses is more difficult than in interrogatives.
Finally, given that null-prep occurred rather similarly in both arguments
and adjuncts, no distinction can be made between the two structures
as far as the null-prep phenomenon is concerned. Thus, Dekydspotter,
Sprouse and Anderson's (1998) speculation of the rule of preposition
incorporation does not hold here inasmuch as the incorporation of a
preposition into a verb across a CP, as is the case in adjuncts, is
illicit in UG.
Alternatively, the study yields empirical evidence consistent with Klein
and Casco's (1999) proposal. Their analysis of null-prep incorporates
null-operator movement and hence is in keeping with the economy principles
because it presumes minimal movement, i.e., movement of wh-features.
Additionally, and more importantly, it makes no distinction between
arguments and adjuncts in terms of movement. Finally, Klein and Casco's
analysis seems more theoretically sound because, unlike Dekydspotter,
Sprouse and Anderson's (1998) proposal, it enjoys the privilege of reinforcing
the contrast between the L1 vs. the L2 learners' behavior. As far as
the evidence goes, L1 learners lack empty operators in their early stages
of language acquisition leading to the absence of the null-prep phenomenon
in their grammar whereas L2 learners have evidenced null-operator constructions
early on (cf. Klein and Casco, 1999).
As such, the results of this study furnish theoretically viable input
for a reformulation of the hypotheses (cf. MacLaughlin, 1993) as to
the nature of the null-prep phenomenon, the findings of the study being
fully compatible with Klein and Casco's (1999) analysis. The results
obtained also imply that the interlanguage grammar of Iranian EFL learners,
though at times diverging from that of L1 learners due to the absence
of obligatory prepositions, still falls within the space constrained
by UG.
Although this study primarily addressed theoretical issues independent
of their practical significance, it may be broadened and deepened to
help supply robust foundations for settling problems in second language
teaching practices. Since the results show that Iranian EFL learners
are obedient to the UG constraints on extraction even when they diverge
from L1 learners on their path to language learning, teachers do not
need to worry about the interim stage at which learners may omit the
prepositions in the prepositional constructions discussed earlier. As
numerous studies have attested to the role of input and instruction
in L2 acquisition (Long, 1985; Kellerman, 1985; Oxford and Scarcella,
1994; Sadighi and Tagharchi, 2001, to name just a few), as a result
of exposure to more positive input, learners would hopefully learn how
to correctly pied-pipe and/or strand the preposition.
To add a word of caution, the researchers would like to emphasize that
the results of this study are by no means conclusive because of the
following reasons. First, participants in the study were selected from
a limited number of proficiency levels. Moreover, one elicitation task
was employed in this study which has most probably impinged upon the
results to a great extent. Therefore, further research including participants
with a variety of proficiency levels using different elicitation tasks
is needed if definite statements are to be made.
References
Bardovi-Harlig, K. (1987). Markedness and salience in second language
acquisition. Language Learning, 37, 385-407.
Clahsen, H. and Muysken, P. (1986). The accessibility of Universal Grammar
to adult and child learners: a study of the acquisition of German word
order. Second Language Research, 2, 93-119.
Dekydtspotter, L., Sprouse, R. A. and Anderson, B. (1998). Interlanguage
A-bar dependencies: binding construals, null prepositions and universal
grammar. Second Language Research, 14, 341-358.
Gass, S. (1996). Second language acquisition and linguistic theory:
The role of language transfer. In W.C. Ritchie and T.K. Bhatia (eds.)
Handbook of Second Language Acquisition. San Diego: Academic Press.
Haegamann, L. (1995). Introduction to Government and Binding Theory
(2nd ed.) Massachusetts: Blackwell Publishers Inc.
Hamilton, R. (1998). Underdetermined binding of reflexives by Japanese-speaking
learners of English. Second Language Research, 14, 292-320.
Hawkins, R. and Chan, Y. (1997). The partial availability of universal
grammar in second language acquisition: the 'failed functional feature
hypothesis'. Second Language Research, 13, 187-226.
Kellerman, E. (1985). If at first you do succeed…. In S.M. Gass
and C. G. Madden (eds.) Input in Second Language Acquisition. Cambridge:
Newbury House Publishers.
Kao, R. (2001). Where have the preposition gone? A study of English
prepositional verbs and input enhancement in instructed SLA.. International
Review of Applied Linguistics in Language Teaching, 39, 195-215.
Kaplan, T. and Selinker, L. (1997). Book review article: Toward second
language acquisition: A study of null-prep by E. C. Klein. Second Language
Research, 13,170-186.
Klein, E.C. (1993a). A problem for UG in L2 acquisition. Issues in Applied
Linguistics, 2, 33-56.
Klein, E.C. (1993b). Towards Second Language Acquisition: a Study of
Null Prep. Dordrecht: Kluwer.
Klein, E.C. (1995a). Evidence for a 'wild' L2 grammar: when PPs rear
their empty heads. Applied Linguistics, 16, 87-117.
Klein, E.C. (1995b). Second versus third language acquisition: Is there
a difference? Language Learning, 45, 419-465.
Klein, E. C. and Casco, M. (1999). Optionality in English non-native
grammars: Differences between L1 and L2. Proceedings of the Annual Boston
University Conference on Language Development, 23, 349-360.
Long, M. (1985). Input and second language acquisition theory. In S.M.
Gass and C.G. Madden (eds.) Input in Second Language Acquisition. Cambridge:
Newbury House Publishers.
MacLaughlin, D. (1993). The responsibility of linguistic theory to second
language acquisition data. A talk presented at the Workshop on Language
Acquisition and Linguistic Theory as part of the Boston University Applied
Linguistic Colloquium Series.
Mazurkewich, I. (1985). Syntactic markedness and language acquisition.
Studies in Second Language Acquisition, 7, 15-36.
McBurney, D.H. (1998). Research Methods (4th ed.). California: Brooks/Cool
Publishing Company.
Oxford, R.L and Scarcella, R.C. (1994). Second language vocabulary learning
among adults: state of the art in vocabulary instruction. System, 22,
231-243.
Sadighi, F. and Tagharchi, N. (2001). Learner-treatment role in acquiring
English phrasal verbs by Iranian EFL learners. Iranian Journal of Teaching
Languages, 2, 1-25.
Schwartz, B. and Sprouse, R. (1996). L2 cognitive states and the full
transfer/full access model. Second Language Research, 12, 40-72.
Soars, L. and Soars, J. (1998). New Headway English Course (eleventh
impression). Oxford: Oxford University Press.
Van Buren, P. and Sharwood Smith, M. (1985). The acquisition of preposition
stranding by second language learners and parametric variation. Second
Language Research, 1, 18-45.
White, L. (1998). Universal grammar in second language acquisition:
the nature of interlanguage representation. Presented at the Second
Language Research Forum at the University of Hawai`i.
To
view the rest of this article, either select the
PDF format, or
the MS Word format.
