isabelli-multicompetence
TRANSCRIPT
Title: Multicompetence and its Role in Second Language Acquisition Research Author: Casilde A. Isabelli Address: Dept. of Foreign Languages & Literatures / MS 100 University of Nevada at Reno 1664 N. Virginia Reno, NV 89557 Contact Information: Office number: 775-784-6055x300 Home number: 775-322-3591 Fax number: 775-784-4197 (Attn: Cassie) Email address: [email protected]
Full name of Institution: Department of Foreign Languages & Literatures University of Nevada at Reno
Abstract: Cook (1992) suggests that when testing the acquisition of subtle UG properties, the linguistic competence of the advanced L2 learner should be measured against the competence of an L1/L2 fluent bilingual and not that of an L2 monolingual because of differing cognitive processes between bilinguals and monolinguals. The goal of this empirical study is to compare the use of that-trace effects in Spanish by three groups: monolingual Spanish speakers; fluent bilingual English/Spanish speakers; and advanced English L2 learners of Spanish. Using reaction time grammaticality judgment tests and a follow-up interview on the subjects’ judgments, the results show differences between the bilingual and monolingual groups, suggesting that the L2 learners are performing like bilinguals. The implications of these results might suggest an additional approach to testing advanced L2 learners on acquisition of UG properties and help clarify the question of UG access during SLA.
1
Multicompetence and its Role in Second Language Acquisition Research
Abstract
Cook (1992) suggests that when testing the acquisition of subtle UG properties, the
linguistic competence of the advanced L2 learner should be measured against the
competence of an L1/L2 fluent bilingual and not that of an L2 monolingual because of
differing cognitive processes between bilinguals and monolinguals. The goal of this
empirical study is to compare the use of that-trace effects in Spanish by three groups:
monolingual Spanish speakers; fluent bilingual English/Spanish speakers; and advanced
English L2 learners of Spanish. Using reaction time grammaticality judgment tests and a
follow-up interview on the subjects’ judgments, the results show differences between the
bilingual and monolingual groups, suggesting that the L2 learners are performing like
bilinguals. The implications of these results might suggest an additional approach to
testing advanced L2 learners on acquisition of UG properties and help clarify the
question of UG access during SLA.
1 Introduction
In generative second language acquisition (SLA) research, it is the norm to measure
acquisition by the approximation of the second language (L2) learner’s use of the target
language to a monolingual native speaker’s use of that target language. However, this
norm suggests a monolingual bias in research when comparing the L2 learner to a
monolingual speaker, since a ‘multicompetent’ speaker (the L2 learner) is compared to a
‘mono-competent’ speaker. The question here is whether SLA research needs to take
into account who they are using as a standard when determining acquisition. In addition
2
to comparing the L2 learner with a monolingual speaker, should we be comparing the L2
learner to a multicompetent L2/L1 fluent bilingual as well? Cook’s (1996) theory of
multicompetence states that the internalized grammars of advanced L2 language learners
differ from those of native speakers of the target language, despite the formers’ near-
native performance. This difference is hypothesized to be due to the compound state of a
mind containing two grammars in contrast to a model in which the L1 and L2
competencies appear to be separate. Within the generative framework of SLA, the
present study attempts to determine whether the linguistic competencies of
multicompetent bilingual speakers (Spanish/English) differ from monolingual speakers of
Spanish. In addition, this study will compare bilingual and monolingual speakers with the
advanced Spanish L2 learner.
2 Universal Grammar in SLA
Chomsky (1981) proposed that an innate Universal Grammar (UG) bridges the gap in
L1 acquisition between available experience (e.g., underdetermined input, degenerate
input, and lack of negative evidence) and attained competence (i.e., a complex adult
grammar). This gap is often referred to as the learnability problem, or the logical problem
of language acquisition, because of the “mismatch between primary linguistic input and
the system actually attained” by language learners (White, 1989, p. 5) due to poverty of
the stimulus. Poverty of the stimulus is characterized by the following: (1) a restricted
sample of the infinite number of sentences a grammar can generate; (2) input that is a
poor sample of the language to be learned; and (3) a lack of information about which
sentences are not grammatical (which would allow the language learner to modify
3
hypotheses that prove to be incorrect). The learnability problem found in L1 acquisition
can be extended to incorporate L2 acquisition due to the same circumstances the learner
faces. In the L2 learning situation the L2 learner attempts to attain the target language,
but unlike L1 learners, the majority of L2 learners fail to acquire full competence in the
L2. However, despite poverty of the stimulus, some L2 learners are able to project
beyond L2 input and become competent and proficient in the L2 just as L1 learners do
because of access to UG (White, 1987, 1988, 1990).
Over the past several decades, generative research has examined the central question
of whether SLA is similar to L1 acquisition by determining the availability of UG to the
L2 learner. One way of measuring availability of UG during SLA is to determine
whether the L2 learner can “reset” a specified UG parameter from the L1 setting to the
L2 setting.
In essence, UG (under Chomsky’s 1981 Principles-and-Parameters approach)
postulates a small number of universal principles that limit the types of grammars found
in natural languages. An example of a principle is the extended projection principle
(EPP), which states that all natural languages must contain a subject. Parameters are then
the range of possible but limited variations of a UG principle. An example of a
parameter is the Null Subject Parameter (NSP), which is a variation of the EPP. The NSP
has two settings: [+NSP], found in Italian, Spanish, Greek or Arabic, which allows the
subject position to be phonologically empty due to the recoverability of the subject
pronoun in the rich inflection of the NSP language; and [-NSP], found in English and
French, which does not allow the subject position to be phonologically empty as a result
of weak inflection. Furthermore, a parameter setting will underlie a range of surface
4
syntactic properties in a particular language, thereby creating a “cluster,” or group of
properties that are related to that parameter. Chomsky’s Minimalist Program (1995)
overlaps considerably with his Principles-and-Parameters approach but differs in that all
language variation is seen as morpho-lexical, and all parametric variation is thought to
arise from the lexicon.
As mentioned previously, a standard method of testing UG availability in SLA is to
contrast L1 and L2 principles with parameter settings of UG. Because UG is designed to
put constraints on a language system and to inform one as to what is not possible,
research on principles and parameters in SLA therefore needs to investigate aspects of
UG not exemplified in the L1. The L2 learner’s ability to “reset” the parameter setting
that is not operative in their L1 is evidence that UG is available. Parameter resetting
means that the L2 learner will not lose the L1 parameter setting, but rather will assume
the L1 setting initially and then “reset” it for the L2 while still maintaining the setting for
the L1 (White, 1989, p. 80).
For the purposes of the present study, three clustering properties (after Chomsky,
1981; Burzio, 1986; and Rizzi, 1982) of the [+NSP] will now be reviewed.
2.1 The Null Subject Parameter
The first clustered property of [+NSP] allows null subjects, pro, in tensed clauses:
(1) a. [pro] Salieron a las ocho.
b. * Left at eight.
c. They left at eight.
d. [pro] Está lloviendo.
5
e. * Is raining.
f. It is raining.
g. * Lo está lloviendo.
All empty categories must be properly governed, as stated by the empty category
principle (ECP). In the Spanish sentences (1a, d), pro is properly governed by the
inflection of the verb, because in null subject languages inflection licenses pro. But in
English, as seen in (1b, e), lexical subjects are required due to their inability to recover
from the weak inflection. In (1d), the absence of the subject pronoun of weather verbs in
Spanish is obligatory, whereas English always requires pleonastic pronouns it and there,
as seen in (1f). Expletive pleonastic it, which is not an argument and to which no theta
roles are assigned, acts as a ‘dummy’ pronoun; it is simply a placeholder for the unfilled
subject position.
The second property of the [+NSP] is free verb-subject (V/S) inversion, illustrated
below:
(2) a. [proi] Han llegado mis padresi.
b. * Have arrived my parents.
c. My parents have arrived.
d. Therei have arrived my parentsi.
Verb-subject inversion in Spanish functions such that co-indexing exists, as in (2a),
between a non-argument subject, pro, and a post-verb indefinite subject, mis padres.
Sentence (2a) is similar to the English sentence (2d), where if the subject position is not
occupied by a referential expression, then it must be filled with the expletive existential
there, also co-indexed with a post-verbal argument, my parents. But, unlike in English,
6
the non-argument co-indexed subject in the Spanish (2a) need not be filled with a
pronoun because null subject languages allow for the position to be null.
The third property of the [+NSP] is the apparent that-trace filter violations, illustrated
in the following:
(3) a. *Whoi does the FBI think [CP ti that [IP ti killed the president]]?
b. Whoi does the FBI think [CP ti [IP ti killed the president]]?
c. ¿Quiéni cree el FBI [CP ti que [IP proi mató tI al presidente]]?
d. *¿Quiéni cree el FBI [CP ti [IP proi mató ti al presidente]]?
The English sentence (3a) is ungrammatical due to the violation of ECP. Proper
government of the wh-trace, t—left in subject position—is blocked because of the
presence of the complementizer that in COMP. But once the complementizer that is
deleted in (3b) the trace in CP properly governs the original one. In the Spanish (3c), pro
is extracted from post-verbal position, allowing the trace to be properly governed by the
lexical verb matar.
3 What is ‘multicompetence’?
The term ‘competence,’ as described by Chomsky (1965), denotes the mental
representation of linguistic rules that constitute a person’s internal grammar and includes
implicit knowledge of what constitutes grammatical sentences. Cook (1991) builds from
the term ‘competence’ to derive the term ‘multicompetence’—“the compound state of a
mind with two grammars” (p. 112)—, which is different from the state of mind with one
grammar, monocompetence. In 1992, Cook further states that the start of
multicompetence within the UG model would be when the L2 learner’s linguistic
7
competence starts to differentiate between the L1 and L2 parameter settings (1992, p.
581).
The term multicompetence stems from Cook’s research in language acquisition,
specifically the poverty-of-the stimulus argument (1991), in which he argued that it “is
not so much how the child learns a grammar with a setting for each parameter as how the
child learns one or more grammars with one or more settings for each parameter—
multicompetence” (1992, p. 558). One of the questions Cook addresses with respect to
multicompetence is whether bilinguals’ and/or L2 learners’ language systems are
independent of each other, or not in the same “mind.” In his review of psycholinguistic
research with multilingual people, he provides evidence that lexicon is, in fact, a single
system (Beauvillain & Grainger, 1987; Caramazza & Brones, 1979); that L1 competence
is not cut off during L2 processing (Altenberg & Cairns, 1983; Blair & Harris, 1981); and
that the multilingual’s brain appears to store both languages in the same areas (Paradis,
1989).
A second question Cook (1992) addresses is how multicompetence differs from
monocompetence in terms of a “state of mind.” He approaches this question in four ways.
First, to examine the hypothesis that L2 users differ from monolinguals in knowledge of
the L1, he reviews the phonological work of Flege (1987) (L2 French, L1 English),
Nathan (1987) (L2 Spanish/English), and Obler (1982) (Hebrew/English bilinguals)
showing that L2 learners’ linguistic competence might have an effect on that speakers’
L1 competence because L1/L2 bilingual adults produce and perceive differently than L1
and L2 monolinguals on the same tasks. Cook also reviews the studies of Caskey-
Sirmons and Hickson (1977) (L1 Korean, L2 English) and Magiste (1979) (L1 German,
8
L2 Swedish), which respectively demonstrate that bilinguals’ L1 vocabulary is affected
by their second language, and that bilinguals’ rate on production and comprehension
tasks is much slower than that of monolinguals. Cook concludes that “multicompetence
in some respects contains a different state of L1 knowledge” (p. 561).
A second way in which Cook (1992) addresses the question of whether
multicompetence is a different state of mind than monocompetence is by showing that
people who know an L2 have a different metalinguistic awareness from people who
know only an L1. A brief review of the bilingual studies of Ianco-Worrall (1972)
(Afrikaans/English children), Ben-Zeev (1977) (Hebrew/Spanish), Galambos and Hakuta
(1988) provide evidence that bilinguals have an enhanced metalinguistic awareness over
monolinguals. Cook concludes that L2 users are more consciously aware of language in
comparison to monolinguals.
A third way in which Cook (1992) addresses the question of difference is by outlining
the differences in the cognitive processes of multicompetent and monocompetent
individuals. In a review of the limited extant research in this area, Cook highlights work
by Landry (1974), who found that L2 learners, after being in a five year FLES program,
scored higher on standard tests of divergent thinking (which value flexibility, originality,
and fluency) than monolinguals. Lambert, Tucker, and d’Anglejan (1973) provide
evidence to suggest the existence of higher creativity of “unusual uses” in children who
were enrolled in immersion schools as compared to monolingual children. Peal and
Lambert (1962) found that on verbal and nonverbal IQ tests, bilingual children
outperformed monolingual children. Cook suggests that these results found among
advanced L2 learners are indicative of their different cognitive processes.
9
Finally, Cook (1992) shows that advanced L2 users differ from monolinguals in their
knowledge of the L2. He examines work by Coppieters (1987), who administered French
grammaticality judgment tests to L1 and near-native L2 French bilinguals (individuals
who started learning French as adults but were accepted by native French speakers as
being native). Her results showed that the near-natives’ grammaticality judgment tests
scores were significantly different from the L2 monolinguals’ scores, especially on basic
grammatical contrasts such as the two past tenses, the third person pronouns, and the
placement of adjectives before or after a noun. The results thus suggest the independence
of language use versus linguistic competence.
The studies by Balcom (1995) and (1998), found that English/French bilinguals
performed significantly differently than native English speakers with little exposure to
French on their acceptance of inappropriate passive morphology. Her results support
Cook’s hypothesis in that on the production tasks, the advanced English learners’
performance was indistinguishable from the monolingual English speakers, but the
grammaticality judgment tests showed that the learners accept significantly more
ungrammatical passives with unaccusative verbs and with other verbs having a theme in
subject position.
Although empirical research in this area is lacking, Cook believes that “the grammar
of the L2 in a multicompetent speaker is not the same as the apparently equivalent
grammar in a monolingual” (p. 562). He argues that if people who speak two languages
are competently different than monolinguals, then this contrast should be introduced into
the field of SLA research. Presently there is a monolingual bias in research; L2 learner
grammaticality judgment test scores are often compared to those of L2 monocompetents,
10
rather than to those of multicompetent L2/L1 fluent bilinguals. In the next section, a
review of SLA research on the Null Subject Parameter using monolingual L2 as the
control group will be re-analyzed within Cook’s multicompetence framework.
3.1 Re-examining studies on the resetting of the Null Subject Parameter
The studies of Liceras (1989), Phinney (1987), White (1985, 1986), Muñoz (1997),
and Isabelli (2004) have addressed the question of whether L2 adult learners of Spanish
reset the NSP from the [-] value in English, to the [+] value in Spanish, or vice versa.1
These studies have produced conflicting evidence as to whether or not the NSP was reset;
however, it is proposed that the results can be clarified with reference to Cook’s (1996)
multicompetence theory.
Liceras (1989) investigates, cross-linguistically, the resetting of the NSP by 30
English and 32 French speakers enrolled in four consecutive Spanish level university
courses in Canada. Five Spanish speakers acted as controls. Liceras administered to all
the subjects a grammaticality judgment test with 17 items that tested the three properties
of the NSP. Her results showed that the first property of the [+NSP], overt null subjects,
was acquired, but that the second and third properties were not. She concludes that these
last two properties might not form part of the [+NSP].
A closer examination of the methodology of this study, however, reveals several
important limitations. First, the five control subjects against whom the L2 learner subjects
were compared were Spanish speaking graduate students at a Canadian university. Were
these fluent English/Spanish bilinguals? French/Spanish bilinguals? Monolingual
Spanish speakers? Liceras does not provide a detailed description of the controls’
11
language backgrounds, which would allow for a clearer understanding of the results
obtained.
Liceras’s results also show that for both the French and English subjects, the
percentage of correct acceptances of grammaticality judgment items on V/S inversion
and that-trace increased with the level of the Spanish course, suggesting that the Spanish
L2 learners entertained the two parameter settings of the NSP. But the more advanced
Spanish learners showed signs of multicompetence when their linguistic competence
started to differentiate between the L1 and L2 parameter settings (when they correctly
accepted 60% or more of the items). Would Liceras’ conclusions be different if the
advanced learners were compared to multicompetent speakers?
Phinney (1987) collected production data in the form of free compositions by Puerto
Rican university students of English and American university students learning Spanish.
Phinney’s results provide no conclusive evidence in either direction of the NSP being
reset, but suggest that it is not difficult for L2 learners to reset the parameter. These
inconclusive results might be due to a combination of methodological errors. First, there
was no control over the topics subjects wrote about, and therefore they could not be
specifically directed to produce the three properties of the NSP. In addition, the study
assumes that free production is representative of linguistic competence, whereas there is a
great deal of research to discount this notion. Finally, and most importantly, there was no
control group with which to compare and contrast results. Differences between
multicompetence and monocompetence would have likely been observed had there been
a monolingual Spanish and English control group, as well as a bilingual Spanish/English
control group.
12
White’s (1985, 1986) subjects consisted of Spanish speaking adults learning English
in the classroom setting, and a control group consisting of native French speakers
learning English. The subjects were administered a grammaticality judgment test that
tested the three properties of the NSP, and a question-formation task that focused on the
that-trace property. Results showed that Spanish speakers had a higher error rate when
judging the grammaticality of English sentences with missing subjects and that-trace
violations than did the French speakers learning English. The results suggest that initially
the L2 learner opts for the L1 setting of the pro-drop parameter, but as proficiency
increases they switch to the L2 setting, at least in regard to the first property. White is
uncertain if the V/S inversion and that-trace effects are properties of the NSP.
Muñoz (1997) administered a grammaticality judgment test testing the three
properties of the [+NSP] to six levels of 124 university Spaniards learning English as an
L2. Her control group consisted of White’s (1986) control subjects, French learners of
English. Muñoz , like White, found that only the first property was acquired, and suggests
that V/S inversion and that-trace do not pertain to the NSP.
Neither White (1985, 1986) nor Muñoz (1997) used monolinguals of the target
language or bilinguals of the L1 and L2 to compare the subjects against. A better
analysis of parameter resetting would have been to compare the multicompetent and
mono-competent speakers to determine, first, whether they hold different competencies,
and second, whether there is a cognitive difference between the two. Such a comparison
would have likely resulted in different conclusions about the NSP in SLA.
Isabelli (2004) measured Spanish syntactic development and acquisition of three
syntactic properties proposed to be related to the NSP in L2 learners of Spanish in a
13
study-abroad context. Judgment tests and oral narrative measures were administered to 29
advanced learners before and after a one-year study-abroad stay in Barcelona. Statistical
results showed significant improvement on all properties. The results also showed that
the participants performed as native-speakers on subject pronoun omission and subject-
verb inversion but did not perform as native-speakers on the third property (that-trace
effect), suggesting that the property was not acquired and thus the parameter was not
reset. Although the Minimalist program is constructionist2 (that is, properties within a
parameter are not acquired all at once but gradually), the lack of acquisition of the last
property even after nine months of immersion is surprising. Do the learners need more
exposure to the input to acquire this last abstract property? Or will the learner ever match
the competence of the monolingual speaker, to whom the learner was compared in this
study?
In summary, Liceras (1989), White (1985, 1986), and Muñoz (1997) found that only
one property of the NSP was acquired and thus conclude that the other two properties are
not part of the NSP, or that the NSP is not reset. Isabelli’s (2004) data suggest that only
two of the properties were acquired and that the failure to acquire the that-trace property
did not allow the NSP to be reset. Phinney’s (1987) results neither support nor contradict
parameter resetting. It is proposed that the conflicting evidence provided by these studies
can be understood as a result of monolingual bias in the studies’ methodological
procedures. The failure of these studies to account for acquisition of the third property
may be due to their failure to compare L2 learners with multicompetent L2/L1 fluent
bilinguals. According to Cook (1996), when testing the acquisition of subtle UG
properties (such as the NSP), the linguistic competence of the advanced L2 learner should
14
be measured against the linguistic competence of an L1/L2 fluent bilingual and not that
of an L2 monolingual because of differing cognitive processes between bilinguals and
monolinguals. The present study is an attempt to re-examine the issue of parameter
resetting in L2 learners through the lens of Cook’s multicompetence theory.
In the present study, multicompetence will be examined by studying the acquisition
of the third property of the NSP, that-trace effects, by the following speakers: advanced
L2 Spanish learners, monolingual Spanish speakers, and bilingual Spanish/English
Heritage speakers. The third property of the [+NSP] was chosen as a focus since, as
discussed with reference to the extant research, it has been shown to be the property most
difficult to acquire (if at all).
4 The Study
The purpose of the present study is twofold: 1) to determine whether mono-competent
L1 and multicompetent L1/L2 Heritage Spanish speakers (HSS) have the same linguistic
competence; and 2) to determine whether the advanced Spanish L2 learner’s linguistic
competence is more similar to the Spanish mono-competent L1 speaker or to the
multicompetent Spanish/English bilingual L1/L2 speaker. Specifically, the study
examines the treatment of Spanish that-trace effects by three groups: monolingual
Spanish speakers; fluent bilingual English/Spanish Heritage speakers; and advanced L2
learners of Spanish. The questions addressed are whether HSS treat that-trace items in
the same manner as monolingual Spanish speakers, and whether advanced L2 learners of
Spanish treat that-trace items in the same manner as Heritage Spanish speakers. The
research questions are as follows: Is there a significant difference on grammaticality
15
judgment test scores on the that-trace effect of the [+NSP] between: (a) monocompetent
monolingual Spanish speakers and multicompetent HSS?; (b) advanced L2 Spanish
learners and monolingual Spanish speakers?; and (c) advanced L2 Spanish learners and
the multicompetent HSS?
4.1 Participants
In order to investigate the use of the that-trace filter in Spanish, three groups of
participants were tested. Group 1 consisted of 10 advanced L2 learners of Spanish whose
first language is English and who were majors and/or graduate students studying Spanish.
These were adult learners who started learning the L2 after puberty. Group 2 consisted of
10 HSS. These HSS are Latino adults who were born in the United States to at least one
parent born in a Spanish-speaking country. As children, these HSS were exposed to two
languages from birth—the family language (Spanish) and the community language
(English)—and who may be more dominant in the community language. Group 3
consisted of five monolingual Spanish speakers from Spain and Mexico, although they
had some speaking skills in English. Groups 2 and 3 served as the control groups.
4.2 Materials
Since UG is designed to put constraints on a system and to “tip off” the language
learner as to what is not grammatical, the most efficient method of determining whether
the target language parameter setting was being reset was to administer a grammaticality
judgment test (GJT). The three subject groups were tested on their knowledge of that-
trace effects in Spanish using a computerized GJT.
16
4.3 The test
The GJT tested for the linguistic competence of the that-trace property of the [+NSP].
The GJT used high frequency Spanish lexicon. Since the hypothesis being tested is
whether the advanced L2 learners’ linguistic competence of that-trace use resembles that
of a monolingual L1 Spanish speaker or of an L1/L2 bilingual speaker, the target items
involved eight grammatical Spanish sentences (see Table 1), which, in comparable
English, would violate the [-NSP] and be ungrammatical.
Table 1 That-trace effects grammatical target sentences
1. ¿Quién cree la policía que mató al joyero?
* ‘Who does the police believe that killed the jeweler?’
2. ¿Quién piensan los padres que raptó a su hija?
* ‘Who do the parents think that kidnapped their daughter?’
3. ¿Quién crees tú que habla japonés aquí?
* ‘Who do you believe that speaks Japanese here?’
4. ¿Quién cree el FBI que asesinó al presidente?
* ‘Who does the FBI think that assassinated the president?’
5. ¿Qué creen Uds. que hizo el ruido?
* ‘What do they believe that made the noise?’
6. ¿Qué pensamos nosotros que causó el derrumbe?
* ‘What do we think that caused the cave-in?’
7. ¿Qué piensan Uds. que rompió en el coche?
17
* ‘What do you (pl) think that broke the car?’
8. ¿Qué crees tú que causó la confusión?
* ‘What do you think that caused the confusion?’
The next set of items was comprised of eight ungrammatical Spanish sentences that
violated the [+NSP] (see Table 2), and that, in comparable grammatical English
sentences, would not violate the [-NSP] setting.
Table 2 That-trace effects: ungrammatical target sentences
1. *¿Quién pensamos nosotros robó el dinero?
‘Who do we think stole the money?’
2. *¿Quién creen Uds. pegó al hombre?
‘Who do you (pl.) believe hit the man?’
3. *¿Quién piensan Uds. regaló las flores a Tina?
‘Who do you (pl.) think gave the flowers to Tina?’
4. *¿Quién crees tú va a ganar mañana?
‘Who do you believe will win tomorrow?’
5. *¿Qué pensamos nosotros hizo el ruido?
‘What do we think made the noise?’
6. *¿Qué creen Uds. causó el fuego?
‘What do you (pl.) believe caused the fire?’
7. *¿Qué crees tú causó el accidente?
‘What do you believe caused the accident?’
18
8. *¿Qué piensan Uds. rompió en el otro cuarto?
‘What do you (pl.) think broke in the other room?’
Also included in the GJT were 16 distractor items (Appendix A) that consisted of eight
grammatical simple structures and eight ungrammatical structures with a scrambled word
order. The GJT was thus composed of a total of 32 items (target and distractor), 16
grammatical and 16 ungrammatical.
4.4 Data Collection Procedure & Scoring
The data was collected one participant at a time on the researcher’s office computer.
The test was administered via a program that presented each sentence individually on the
screen. The participant had eight seconds to read and judge whether the Spanish sentence
was ‘possible’ or ‘impossible’ by hitting the corresponding letter on the keyboard. Before
starting the test, the participants took a warm-up test with five practice sentences. The test
items were administered in a randomized order. After completion of the test, each
participant completed a background questionnaire, during which time the researcher
printed out the sentences judged as ‘impossible.’ After completing the questionnaire, the
participant was given the sentences marked ‘impossible’ and was then instructed to make
the ‘impossible’ sentences ‘possible.’
Each item of the GJT was given a score of “1” if the participant correctly judged the
that-trace present sentences as ‘possible’ and correctly judged the missing that-trace
sentences as ‘impossible.’ The follow-up test allowing the participants to make the
‘impossible’ sentences ‘possible’ ensured that the sentences judged as ‘impossible’ was
19
due to the missing that-trace. If they corrected the sentence by inserting the
complementizer ‘que,’ then they received one point.
The participant received a score of “0” if they incorrectly judged the that-trace
present sentences as ‘impossible’ and incorrectly judged the missing that-trace sentences
as ‘possible.’ If on the follow-up test the participant corrected the ‘impossible’ sentence
by not inserting the complementizer ‘que,’ then they received 0 points.
The GJT target items were scored and the total scores were then tallied for each
participant. A score of “8” was the highest possible.
5 Results
The Heritage Spanish speakers’ GJT scores and the monolingual speakers’ test scores
are summarized in Table 3. On grammatical test items with the that-trace present, the
monolingual Spanish speakers performed better (100%) than the Heritage Spanish
speakers (81.3%) and the advanced L2 learners (63.8%). On the ungrammatical test
items with the that-trace absent, the monolingual Spanish speakers again performed
better (92.5%) than the Heritage Spanish speakers (71.2%) and the advanced L2 learners
(45%). In addition, the data suggest that the advanced L2 learners know aspects of
Spanish language (that-trace effects) that they could not have learned from the input or
through formal instruction; this knowledge is indicated by the fact that they scored more
than “0” on sentences that are grammatical in Spanish but are ungrammatical in English.
Table 3 Grammaticality Judgment Test Mean Scores
20
Item Range Mean % (Standard Deviation)
Monolingual Heritage Advanced
(n = 5) (n = 10) (n = 10)
That-trace present 0-8 100% (.00) 81.3% (2.42) 63.8% (3.31)
That-trace absent 0-8 92.5% (.89) 71.2% (2.26) 45% (3.27)
Table 4 presents the independent-samples t-tests comparing the monolingual Spanish
speakers’ and the Spanish Heritage speakers’ scores to the advanced L2 learners. For the
target sentences with the that-trace present and absent, the data yielded a non-significant
difference in mean scores between the Monolingual speakers and the Spanish Heritage
speakers (p = .095, df = 13 and p=.213, df=13 respectively) at the p < .05 level.
However, the data yielded a significant difference in mean scores between the
Monolingual speakers and the advanced L2 learners (p = .004, df = 13 and p = .002, df =
13 respectively) at the p < .05 level.
When comparing the Spanish Heritage speakers’ scores to the advanced L2 learner
group, the data yielded a non-significant difference in mean scores (p = .146, df = 18 and
p= .053, df = 18 respectively) at the p < .05 level.
Table 4 Summary Table of Independent-Samples T Test
Pair item t df p
Monolingual Spanish that-trace 1.363 13 .095
21
Heritage Speaker that-trace absent 1.594 13 .213
Monolingual Spanish that-trace 1.920 13 .004**
Advanced L2 learner that-trace absent 2.506 13 .002**
Heritage Speaker that-trace 1.079 18 .146
Advanced L2 learner that-trace absent 1.669 18 .053
** p < .05 (denotes a significant difference in mean scores)
6 Discussion and conclusions
The results indicate that, in answer to the first research question, there is not a
significant difference on grammaticality judgment test scores on the that-trace effect of
the [+NSP] between monocompetent monolingual Spanish speakers and multicompetent
Heritage Spanish speakers. Similarly, no significant difference in GJT scores was found
between advanced L2 Spanish learners and multicompetent HSS on the that-trace effect
of the [+NSP] (the third research question). However, a significance difference was found
between mono-competent Spanish speakers and multicompetent advanced L2 learners on
the that-trace effect of the [+NSP] (the second research question). Taken together, these
data suggest that the advanced L2 learners are performing more like the Heritage Spanish
speaking group and provide support for Cook’s ‘multicompetence’ hypothesis.
This study also confirms Balcom’s (1995) and (1998) findings that English/French
bilinguals performed significantly differently than native English speakers with little
exposure to French on their acceptance of inappropriate passive morphology. The
22
present study also supports Coppetiers’s (1987) findings that near-native speakers diverge
from native speakers in formal features and functional aspects of grammar.
However, the far-from-perfect performance on the GJT by the Heritage speakers
(81% and 71%, respectively, with a large standard deviation) merits discussion of a topic
currently being investigated in the bilingual research: incomplete acquisition of Spanish
by the Heritage speakers. Research has shown (Silva Corvalán, 2003; Vihman &
McLaughlin, 1982) that early bilingual second and third generation immigrants for whom
the non-heritage, community language was stronger will have a weaker heritage language
since it was not acquired completely or because some aspects were lost. In addition,
recent generative bilingual research has shown that incomplete acquisition resembles a
particular stage of second language acquisition (Montrul, 2005a,b).
The results from these bilingual studies suggest that the issue of multicompetence
needs to be further investigated with Heritage speakers to determine whether incomplete
acquisition or multicompetence is in effect and to determine how to differentiate between
them during SLA research. The implications of these results might suggest an additional
approach to testing advanced L2 learners on acquisition of UG properties and help clarify
the question of UG access during SLA.
Endnotes:
1 Hilles’s 1986 study will not be considered here since she assumes a different version of
the clustered NSP properties.
2 Using the Minimalist Program for a theory of SLA of syntactic properties, a
constructionist approach is assumed. As Herschensohn (2000) eloquently summarizes,
23
“the ‘incompleteness’ of L2 acquisition follows from the fact that parameters are not
‘reset’ but rather are gradually acquired as the second-language learner’s lexical store
increases” (p. 117). However, one would assume that through this constructionist
approach the L2 learner will eventually arrive at the final state where the parameter is
reset (i.e., the acquisition of all the properties of the parameter, including the that-trace
property).
24
References
Altenberg, E., & Cairns, H. (1983). The effects of phonotactic constraints on lexical
processing in bilingual and monolingual subjects. Journal of Verbal Learning and
Verbal Behavior, 22, 174-188.
Balcom, P. (1995). Argument structure and multi-competence. Linguistica Atlantica 17, 1-17.
Balcom, P. (1998). These constructions don’t acquire easily: Middle constructions and
multicompetence. Canadian Journal of Applied Linguistics, 2, 5-20.
Beauvillain, C., & Grainger, J. (1987). Accessing interlexical homographs: Some
limitations of a language-selective access. Journal of Memory and Language, 26,
658-672.
Ben-Zeev, S. (1977). The influence of Bilingualism on cognitive strategy and cognitive
development. Children Develolpment, 48, 1009-1018.
Blair, D., & Harris, R. (1981). A test of interlingual interaction in comprehension by
bilinguals. Journal of Psycholinguistic Research, 10, 457-467.
Burzio, L. (1986). Italian syntax: A Government-Binding approach. Dordrecht: Reidel.
Caramazza, A., & Brones, I. (1979). Lexical access in bilinguals. Bulletin of the
Psychonomic Society, 13, 212-214.
Caskey-Sirmons, L., & Hickson, N. (1977). Semantic shift and bilingualism: Variation in
the color terms of five languages. Applied Linguistics, 19, 358-367.
Chomsky, N. (1965). Aspects of the theory of syntax. Cambridge, MA: MIT Press.
Chomsky, N. (1981). Lectures on government and binding. Dordrecht: Foris.
Chomsky, N. (1995). The Minimalist Program. Cambridge, MA: The MIT
Press.
25
Cook, V. (1991). The poverty of the stimulus argument and multicompetence. Second
Language Research, 7, 103-117.
Cook, V. (1992). Evidence for multicompetence. Language Learning, 42, 557-591.
Cook, V. (1996). Competence and multicompetence. In G. Brown, K. Malmkjaer, & J.
Williams (Eds.), Performance and competence in second language acquisition
(pp. 57-69). Cambridge: Cambridge University Press.
Coppieters, R. (1987). Competence differences between native and near-native speakers.
Language, 63, 544-573.
Flege, J. (1987b). Effects of equivalence classification on the production of foreign
language speech sounds. In A. James & J. Leather (Eds.), Sound patterns in
second language acquisition (pp. 9-39). Dordrecht: Foris.
Galambos, & Hakuta. (1988). Subject-specific and task-specific characteristics of
metalinguistic awareness in bilingual children. Applied Psycholinguistics, 9, 141-
162.
Herschensohn, J. (2000). The second time around: Minimalism in L2 acquisition.
Philadelphia, PA: John Benjamins Pub.
Hilles, S. (1986). Interlanguage and the pro-drop parameter. Second Language Research,
2, 33-52.
Ianco-Worrall, A. (1972). Bilingualism and cognitive development. Child Development,
43, 1390-1400.
Isabelli, C. (2004). The acquisition of the Null Subject Parameter properties in SLA:
Some effects of positive evidence in a natural learning setting. Hispania, 87, 150-
162.
26
Lambert, W., Tucker, C. R., & d’Anglejan, A. (1973). Cognitive and attitudinal
consequences of bilingual schooling. Journal of Educational Psychology, 85, 2,
141-159.
Landry, R. G. (1974). A comparison of second language learners and monolinguals on
divergent thinking tasks at the elementary school level. Modern Language
Journal, 58, 10-15.
Liceras, J. (1989). On some properties of the “pro-drop” parameter: Looking for missing
subjects in non-native Spanish. In S. Gass & J. Schachter (Eds.), Linguistic
perspectives on second language acquisition (pp. 109-133). Cambridge: UP.
Magiste, E. (1979). The competing linguistic systems of the multilingual: A
developmental study of decoding and encoding process. Journal of Verbal
Learning and Verbal Behavior, 18, 79-89.
Montrul, S. (2005a). Second language acquisition and first language loss in adult early
bilinguals: Exploring some differences and similarities. Second Language
Research, 21, 199-249.
Montrul, S. (2005b, March). Interpreting mood distinctions in Spanish as a Heritage
language. Paper presented at the 20th Conference on Spanish in the United States,
University of Chicago, IL.
Muñoz, C. (1997). El parámetro pro-drop en la adquisición del inglés como lengua
extranjera. AESLA Conference proceedings.
Nathan, G. (1987). On second-language acquisition of voiced stops. Journal of
Phonetics, 15, 313-322.
27
Obler, L. (1982). The parsimonious bilingual. In L. Obler & L. Menn (Eds.), Exceptional
language and linguistics (pp. 339-346). New York: Academic Press.
Paradis, M. (1989). La lateralización cerebral diferencial en los bilingues: ¡Basta, por
favor! Investigaciones Psicologicas, 7, 95-105.
Peal, E., & Lambert, W. (1962). The relation of bilingualism to intelligence.
Psychological Monographs, 76, 1-23.
Phinney, M. (1987). The pro-drop parameter in second language acquisition. In T.
Roeper & E. Williams (Eds.), Parameter setting (pp. 221-238). Dordrecht: Reidel.
Rizzi, L. (1982). Issues in Italian syntax. Dordrecht: Foris.
Silva-Corvalán, Carmen. 2003. Linguistic Consequences of Reduced Input in Bilingual
First Language Acquisition. In S. Montrul & F. Ordónez (Eds.), Linguistic Theory
and Language Development in Hispanic Languages: Papers from the 5th
Hispanic Linguistics Symposium and the 4th Conference on the Acquisition of
Spanish and Portuguese (pp. 375-397). Somerville, MA: Cascadilla.
White, L. (1985). The pro-drop parameter in adult second language acquisition.
Language Learning, 35, 47-62.
Vihman, M. M., & McLaughlin, B. (1982). Bilingualism and second language acquisition
in preschool children. In C. J. Brainerd & M. Pressley (Eds.), Verbal processes in
children: Progress in cognitive development research (pp. 35-57). New York:
Springer Verlag.
White, L. (1986). Implications of parametric variation for adult second language
acquisition: An investigation of the pro-drop parameter. In V. Cook (Ed.),
28
Experimental approaches to second language acquisition. London: Pergamon
Press.
White, L. (1987). Against comprehensible input: The Input Hypothesis and the
development of L2 competence. Applied Linguistics, 8, 95-110.
White, L. (1988). Implications of learnability theories for second language learning and
teaching. MacGill Working Papers in Linguistics = Cahiers Linguistiques de
MacGill, 5, 148-162.
White, L. (1989a). Universal Grammar and second language acquisition (Rev. ed.).
Amsterdam: John Benjamins Publishing Company.
29
Appendix A
Grammaticality Judgment Test Distractor Sentences.
1. *¿Quién robó pensamos que nosotros el dinero?
* ‘Who robed do we think that we the money?’
2. *¿Quién dicen el pegó hombre Uds.?
* ‘Who they say the hit man they?’
3. *¿Quién flores regaló Uds. piensan?
* ‘Who flowers gave you (plural) think?’
4. *¿Quién Uds. su coche piensan su coche robó?
* ‘Who you (plural) his car think his car robbed?’
5. *¿Qué tú que causó crees la confusión?
* ‘What you that caused you believe the confusion?’
6. *¿Qué Uds. el fuego que causó creen?
* ‘What you (plural) the fire that caused you (plural) believe?’
7. *¿Qué el ruido pensamos hizo nosotros que?
* ‘What the noise we think made we that?’
8. *¿Qué causó que piensan en el coche Ud.?
* ‘What caused that you (plural) think in the car you (formal)?’
9. ¿Qué comiste tú ayer en el restaurante?
‘What did you eat yesterday in the restaurant?’
10. ¿Qué compró tu madre en el supermercado?
‘What did your mother buy at the supermarket?’
11. ¿Qué leyeron los profesores en clase?
30
‘What did your professors read in class?’
12. ¿Qué estudió el estudiante esta mañana?
‘What did the student study this morning?’
13. ¿Qué hicieron Uds. durante el verano?
‘What did you (plural) do during the summer?’
14. ¿Quién habló con tu amigo por la noche?
‘Who spoke with your friend last night?’
15. ¿Quién gritó a los niños pequeños ayer?
‘Who yelled at the little children yesterday?’
16. ¿Quién quiere ir conmigo a la fiesta?
‘Who wants to go with me to the party?’
17. ¿Quién quiere ver una película romántica?
‘Who wants to see a romantic movie?’
18. ¿Quién escribió la carta al presidente?
‘Who wrote the letter to the president?’