I. Introduction
In her seminal article, Lardiere (2005)
challenged the prevailing notion of parameter
resetting, arguing that the formal task facing a
learner is much more complex than the simple
parametric selecting of a new feature. She
attributed the existing cross-linguistic variations
to permutations in the configuration of morpho-syntactic features which make up grammatical
categories. Difficulties in learning a second
language, in the framework of parameter setting
and resetting, are explained through the
availability of features. If a feature is selected in
the L1, it would be easy for the learner to acquire
that feature in the L2. On the other hand, features
that are morphologically selected in the L2 but
not in the L1 would be no longer available and,
therefore, unacquirable, resulting in a fossilized
grammar. In the later proposed framework of
feature reassembly (Lardiere, 2005, 2007, 2009),
the difficulties are attributed to the massive task
of determining how to reconfigure features from
the way they are mapped in the L1 into new
configurations which may be realized on quite
different lexical items in the L2.
However, as White (2009) suggests, there are
two parts to this proposal. If a feature exists in
the first language, sometimes the learners should
associate that feature with different lexical items
in the second language. An example would be
the acquisition of articles in a second language
like English or Greek by learners whose L1 is
article-less such as Mandarin Chinese which has
the feature definiteness but this feature is
realized on other determiners such as
demonstratives. So, Chinese learners’ task would
be to discover that the feature is realized on
different lexical items in English: the and a. In
the second place, if feature combinations in L1
and L2 differ, the learner will have to acquire a
new language-specific configuration, assembling
features into different bundles in the L2 from the
L1. An example would be the way the four
features of plural, definite, human, and animate
are assembled in English and Korean resulting in
different realizations of plural in each language
(Choi, 2009). There are other cases which
involve different combinations of features across
two languages and the learning of such
grammatical points will depend on successful
reassembly on the part of learners (see for
example Renaud, 2009; Yuan & Zhao, 2009). It
is this second form of reassembly which is more
complicated and more likely to pose serious
problems to learners (Lardiere, 2005).
One area of grammar that has been studied in
research on L2 acquisition is the article system.
It is of particular interest in testing hypotheses
about L2 acquisition because of its complexity
and the number of features that play a role in
determining the correct article to be used. In case
of English, those features include definiteness,
specificity, countability, genericity and number.
So, much information is encoded in the two little
words the and a. Wakabayashi (2009) noted this
intricacy of marking definiteness in English and
believed that it has been largely ignored in the
existing SLA research which has focused on
only one or two features that determine article
use. Examples of such studies are Ionin,
Zubizarreta and Maldonado (2008), Ko, Ionin
and Wexler (2009) and Snape, Leung and Ting
(2006). This gave this study the impetus to
present a comprehensive account of how the
system works in English with regard to the
above mentioned features and to find out if those
features also operate in Persian. Also, the
acquisition of the article system in English by
Persian speakers seems to be particularly
suitable for testing the theoretical predictions of
the feature reassembly hypothesis (FRH
henceforth) because many features seem to be
bundled into the articles (lexical items) in both
languages. The discrepancy in feature
combination may explain Persian speakers’
reported failure to achieve native-like fluency in
the use of English articles (Afzali, 2008;
Rahmani, 2010; Momenzade & Youhanaee,
2013). The aim, then, was to seek explanation
for the observed pattern of article use by Persian
learners. If it can be shown that the
configurations of the system differ between the
two languages but the features are present at
both scenes, then the reassembly hypothesis
might explain the process of acquisition as it
really is.
II. Properties of the article system in English
and Persian
In the following, the article system in English
and Persian is elaborated on in order to clarify
how the features of definiteness, specificity,
genericity, number and countability actually
affect article use. Each point is presented with an
example. Similarities and differences between
the two languages are highlighted and
predictions are made as to the performance of
Persian learners of English on each point.
a. Definiteness and specificity
The English article system marks definiteness,
not specificity (Ionin, 2006). Therefore, definite
determiner phrases (henceforth DPs) in English
are marked by the and indefinite DPs by a. As
for specificity, DPs in English may come with
the or a in both [±specific] situations (see Ionin,
Ko & Wexler, 2007). Persian, on the other hand,
does not have a definite article in its formal
written form- though there is an enclitic used to
mark definiteness in the colloquial language
(Ghomeshi, 2003)- but marks indefinite NPs in
three ways (using an enclitic -i noun finally, the
word yek before the noun, or both). Indefinite
NPs in Persian can bear specificity as well (see
Karimi, 1999). Table 1 presents examples for the
two types of NPs in the two corresponding
languages:
It is, then, reasonable to predict that Persian
learners should have no difficulty with the
indefinite article but, since an overt article is
lacking in Persian, they may have problems in
learning how to use the. However, any difficulty
should be of short duration on the grounds that
the concepts exist and are implemented in
Persian and that there is an informal definite
enclitic in that language.
Besides definiteness and specificity, there are
still a number of other features that affect how
articles are used in English. They include
genericity, number, and the count/mass
distinction. The same features also exist in
Persian but the two languages do not behave
similarly (as presented in the following sections).
How those features influence article use in
English plus the difference there is with Persian
is explained in the following.
b. Genericity
In English, the concept of genericity can be
expressed using the, a/an, or Ø. The three forms
are not always interchangeable, a point that is
not the concern of the present paper (for more
information see Quirk, Greenbaum, Leech &
Svartvik, 1972). In Persian, however, only bare
singular or bare plural nouns can be interpreted
as generic (Moin, 1990). To identify the
acquisition problem for Persian learners, one
should bear in mind that English uses three ways
to denote genericity but Persian uses two.
Examples are:
It has to be mentioned that bare plurals in
Persian depend on context for their
interpretation. They may also be interpreted as
definite. So, example (13) above can also be
interpreted as ‘The flowers need light.’ when the
speaker is referring to some flowers in his office
for example.
Although Persian has an overt indefinite article,
it cannot be used in a generic NP which is a
point of divergence from English. It is also of the
same difficulty to Persian learners to figure out
that the definite article can be used in a generic
NP. It seems, then, reasonable for Persian-speakers to be able to best understand and
produce the first type of generic sentences in
English (bare plurals) since the same structure
exists in Persian as well. For the other two types,
difficulties are expected at low levels of
proficiency simply because the other two types
involve article use that makes them different
from their L1 structure. At higher levels,
however, learners are expected to comprehend
and produce the other two forms as they master
the articles.
c. Number
As to number, both languages select the feature
but the realization is different again. In English,
singular nouns can be accompanied by definite,
indefinite, or zero articles (in case of mass
nouns). Plural nouns also may occur with
definite or zero articles (in case they are
generic). In Persian, however, the indefinite
article may precede plural nouns as well as
singular nouns. If nouns (singular, plural, or
mass) are definite or generic, then they are
accompanied by a zero article. Another point to
mention is that in informal spoken Persian, the
definite enclitic is only used with singular nouns
(Ghomeshi, 2003). So, whereas in English the
definite article is used with both singular and
plural nouns, in Persian it is used only with
singular nouns. Similarly, whereas in English the
indefinite article can only be used with singular
nouns, it is used with both singular and plural
nouns in Persian. Examples are displayed in
Table 3
To make things more complicated, bare plural
NPs in English are either indefinite or generic
whereas the same construction can be generic or
definite. As a result, the learning task for Persian
learners of English would be to use the definite
article for singular and plural nouns and to use
the indefinite article just for singular nouns. It is,
then, expected that Persian speakers should have
no serious or long-lasting problems since both
features (definiteness and number) exist in their
L1.
d. Countability
The last feature to be discussed is the count/mass
distinction on nouns. In English, mass nouns can
be definite and, thus, preceded by the. A zero
article can also come before a mass noun in a
generic or indefinite sense though in the
indefinite reading words like some are preferred.
In Persian, on the other hand, mass nouns cannot
take any article other than zero in which case
they can be interpreted as definite, indefinite, or
generic. Although on occasions people use the
indefinite article with a mass noun in the
informal spoken language (Moin, 1990), it is not
acceptable in the formal written variety we are
discussing. There is, however, a very important
difference between English and Persian in that
mass nouns can be pluralized in Persian, a
construction which is impossible in English.
(Table 4)
Again, it seems that differences between the two
languages are not due to absence of some
features but to how those features are realized on
lexical morphemes cross-linguistically. If we
take the view that absence of features is the
source of difficulty, Persian learners should have
no serious problems. If they face difficulties,
then, the responsibility falls on the different
assembly of features in the two languages.
III. Review of literature
As a difficult linguistic element to be acquired
by learners, studies abound on the acquisition of
the English article system and on the difficulties
learners face. These pieces of research can be
generally divided into two groups based on
learners’ L1 background: if they speak a
[+article] or a [–article] first language. Two
notions dominate such studies: transfer and
fluctuation. That is, either learners have access to
UG settings and, therefore, fluctuate between the
two settings of using articles for specificity or
for definiteness, or they don’t have such access
and transfer their L1 settings into their pattern of
L2 article use. Studies on learners with article-based L1s have shown that the second alternative
works and the transfer option is dominant
(Snape, Leung & Ting, 2006; Ionin, Zubizarreta
& Maldonado, 2008; Jaensch & Sarko, 2009;
Garcia Mayo, 2009). Learners from article-less
L1s, however, have demonstrated that, for them,
fluctuation is the prevailing factor and that they
use English articles sometimes to mark
specificity and other times definiteness (Guella ,
Déprez & Sleeman, 2008; Zdorenko & Paradis,
2008).The few studies conducted on the use of
articles for generic reference have given support
for the strong role L1 plays in L2 acquisition of
articles as well (Perez-Leroux, Munn, Schmitt &
DeIrish, 2004; Ionin & Montrul, 2009; Snape,
Garcia Mayo & Gürel, 2009).
Most studies that have investigated learning of
the articles by Persian learners have approached
the problem from a pedagogical perspective and
so, do not contribute to this research (for
instance, Afzali, 2008; Dabaghi & Tavakoli,
2009; Farrokhi & Sattarpour, 2012; Geranpaye,
1995; Mobini, 2006; Pashazade & Marefat,
2010; Sarani, 2009).Two studies, however, have
tried to provide a theoretical account for article
acquisition using the UG framework. Rezai and
Jabbari (2010) were able to show that the
learners in their study did acquire the English
definite article though some fluctuation was
observed in their teasing apart the definiteness
and specificity features in singular contexts.
They believed that their study provided another
piece of support for the interpretability
hypothesis (Tsimpli & Dimitrakopoulou, 2007)
which claims that interpretable features are
accessible to the L2 learner while uninterpretable
features are difficult to identify and analyze in
the L2 input due to persistent L1 effects on adult
L2 grammars. The definiteness feature is
interpretable, they argue, at both conceptual
level (LF) and phonetic level (PF) in English.
The same feature is, however, interpretable only
at the LF in Persian. Nevertheless, the fact that
definiteness is an interpretable feature in both
languages poses no serious problems to learners
and the study supported this claim. Elsewhere,
Rezai (2012) tested 50 learners of English on the
use of articles for mass nouns and, again, the
results substantiated the interpretability
hypothesis. He also concluded that article
suppliance created more learnability problems in
the plural and indefinite mass contexts compared
to the count singular ones.
Considering research on feature reassembly
hypothesis, on the other hand, numerous studies
exist all of which point to the explanatory power
of the FRH for the acquisition failure of learners.
In fact, Montrul and Yoon (2009, p. 291) called
selecting and reassembling of features ‘the
logical problem of L2 acquisition’. Dominguez,
Archi and Myles, (2011) studied aspect in
Spanish where English learners had to overcome
a morphological contrast between the two
languages and reported that cases where a
meaning-requiring reassembly was necessary,
students failed even at the advanced level. In
their study of English learners’ acquisition of
Chinese resultative constructions where there
was an asymmetry between reconfiguration of
syntactic and thematic features, Yuan and Zhao
(2008) concluded that the learners were able to
reassemble syntactic structures but not the
thematic ones. The writers, then, suggested that
reconfiguration of syntactic and semantic
features did not develop in a uniform fashion in
L2 acquisition. Choi (2009) also added to the
literature by studying the interpretation of wh-in-situ expressions in L2 Korean by adult native
speakers of English, a structure that posed
difficulty to English learners of Korean. She
adequately showed that the difficulty for such
learners did not reside in parametric selection
because both L1 and L2 selected the relevant
features. Instead, the learners had difficulty in
reconfiguring such features into a different L2
configuration. Other areas of grammar which
have been considered ideal for testing the
predictions of the FRH include the existential
quantifier (Gil & Marsden, 2013), wh-questions
(Muroya, 2013a), accusative clitics
(Shimanskaya & Slabakova, 2014), VP adverbs
(Muroya, 2013b), past participle agreement
(Renaud, 2009), inflectional morphology
(Muroya, 2012), and spatial predicates (Stringer,
2012) all of which have lent support to the
hypothesis.
Concerning studies on English articles, Lardiere
(2005) presented data from a Chinese learner of
English (Patty) who had apparently been able to
use the article system to mark definiteness and
number. These two features are lacking in
Chinese and the researcher argued that ‘the
acquisition of definiteness and plural marking is
not a matter of mere parameter resetting from a
minus value in Chinese to a plus value in
English. Rather, it involved ‘a more painstaking
process of reassembling the relevant features
from the way they were conditioned and realized
in the L1 to that of the L2’ (p. 183). In another
study, Kaku (2006) provided evidence by testing
a group of L1 Japanese learners of English.
Japanese does not have an article system but the
concept of in/definiteness is expressed through
other means (demonstratives, pronouns, etc). In
other words, the two features required for the
English articles (definiteness and referentiality)
are expressed by case markers and
demonstratives. Using an elicitation task and a
translation task, she showed that, overall, a good
proportion of articles had been supplied correctly
especially by the advanced learners. The
researcher, then, suggested that it was plausible
that the features of the English articles could be
acquired (based on the performance of the
advanced students) and that the variability in
article choice among the intermediate learners
could be because of unsuccessful reassembly of
features in the L2. However, this study was
based on data from only five participants (three
advanced and two intermediate) so it is difficult
to make generalizations based on it.
The article system, therefore, is an ideal
grammatical category for comprehensive testing
of the FRH since it involves lexical encoding of
more than one feature. We believe that
investigating the acquisition of articles by
Persian speakers is a fruitful case because of the
variety of features involved, the complexities
that exist in the Persian article system, and the
cross-linguistic comparability of the two
systems.
IV. The focus of the present study
Previous studies on Persian L2 learners of
English have documented the article system as a
major barrier to native-like proficiency even at
the advanced level (Mobini, 2006; Momenzade
& Youhanaee, 2013) and have sought to
alleviate the problem by providing pedagogical
suggestions (Dabaghi & Tavakoli, 2009;
Pashazade & Marefat, 2010). Nevertheless,
Persian is an article-based language which marks
definiteness and which bundles some other
features into its article system besides
definiteness. The main aim of this study was,
then, finding out whether Persian learners of
English are able to reconfigure the assembly of
features from their L1 to the L2 and if so, at
what level of grammatical proficiency. Besides,
we were interested in finding out if Persian gives
clues as to what article should be used with
nouns in English. To that purpose, the following
questions were formed:
Are Persian EFL learners eventually able
to reassemble definiteness, specificity,
genericity, number, and countability on
English articles?
Is there any order of acquisition with
regard to the features related to article
use?
Does the function of nouns as
subject/object help correct use of articles
in English?
V. Method
Participants
Forty three Persian-speaking learners of English
at Sheikhbahaee University took part in the
present study. Initially, they took the Oxford
Placement Test (2001, version 1) based on which
their level of grammatical knowledge was
determined and three groups of grammatical
proficiency were identified. In the highly
advanced group, learners who scored 55-60 on
the OPT were placed. They had an age range of
25-40. This group included 11 people. The
intermediate group consisted of those
participants who could score 37-47 on the OPT.
They were 14 in number and 19-27 years old.
The elementary group included 18 participants
who scored 18-27 on the placement test. They
had the same age range as the intermediate
group. It should be mentioned that wider gaps
were considered between the groups in order to
make sure that they belonged to different levels
of grammatical knowledge. Fifteen native
speakers were also included in the study as the
control group. They were 9 males and 6 females
and aged 8-34, from school children to adults.
Actually, they were the only available natives to
the researchers. They all were originally
Americans having lived in their home county all
their lives and volunteered to take our English
test. Table 5 summarizes the specifications of the
participants.
Instruments
Data elicitation in the present study was based
on an OPT test, a grammaticality judgment test
and a translation test, the description of each is
presented in the following.
a. The Oxford Quick Placement Test (OPT)
The quick placement test used in this study is a
flexible test of English developed by Oxford
University Press and Cambridge ESOL (2001) to
give teachers a reliable and time-saving method
of finding a student’s level of English. The paper
and pencil version of the test used here includes
60 multiple-choice items to be answered within
30 minutes. According to the publishers,
elementary students are the ones who score 18-29 on the test. Intermediate learners score 30-47
and those who score within the range of 48-54
are considered as advanced. Any learner who
scores above 55 (55-60) is a very advanced one.
The test also has a writing task with a different
scoring method which was not used in the
present study.
b. The grammaticality judgment (GJ) test
To tap the participants’ understanding of
grammaticality/ungrammaticality of the forms
associated with the English article system, a
grammaticality judgment (GJ) test was
developed and used for the purposes of this
study. The test included 100 items. There were
eleven categories on the test which measured the
five features that have an effect on article use.
For each category, both grammatical and
ungrammatical items were included. The reason
was that acquisition implies knowledge of both
grammatical and ungrammatical forms. So, only
when learners know the correct and incorrect
form of a grammatical item, they can be said to
have acquired it. To distract the participants’
attention from the point that was tested, some
filler items were also included. The categories on
the test and the number of items in each category
are presented in Table 6 below.
Two points have to be mentioned: first, an effort
was put into considering all types of nouns in
English with which an article should be used, so
we came up with eleven categories. Second, the
number of ungrammatical items across
categories was not equal because the cases were
different. For example, considering plural nouns,
whether definite or indefinite, the use of a was
completely out of the question. So, the number
of ungrammatical items in such groups was 4
(zero article +N for definite plurals and the +N
for indefinite ones) whereas for singular nouns,
more ungrammatical items could be included. To
further clarify the point, consider the sixth
No Categories Grammatical Ungrammatical Total
category (indefinite, singular, non-specific
nouns). Ungrammatical items included four
sentences containing the+N and four including
zero article+N. Finally, the ninth category
included no ungrammatical sentences simply
because the corresponding ungrammatical items
(zero article +N) were included in the 8th
category. So, there was no need to make the test
lengthier than it was.
Each item on the test included two sentences.
Following the stem, the test takers had three
choices. They were supposed to judge the
grammaticality or ungrammaticality of the
second sentence in each pair based on the first
one. They were asked to choose √ if the second
sentence was, in their opinion, grammatically
correct and * if incorrect. They were also given
the option of choosing ? if they didn’t know or
were not sure of the answer. Examples of each
category on the test are provided below. They
are all selected from the grammatical items:
1. I saw an accident on my way back home. A
man was injured in the accident.
2. I’m going to the boss today to ask him for a
raise. The man who comes with me will not
regret it.
3. Mr. Peterson owns two houses. He tries to
keep the houses in good condition.
4. I returned from the shopping center to find
that my car had been punctured. I must find
the wrongdoers whoever they are.
5. Alice is so angry. A pig has stepped on her
front garden.
6. My pen ran out of ink. Will you give me a
pen, please?
7. Whenever I can’t find something, my dog
does it for me. Dogs are very intelligent.
8. When you are at a zoo, you should stay away
from lions’ cages. The lion is a dangerous
animal.
9. Large animals usually give birth to their
children. But an ostrich, which is the biggest
bird of all, lays eggs.
10. What would you like best for desert? Oh, I
prefer ice cream to the rest.
11. We’ll never go to that restaurant again. The
food was awful.
c. The translation test
The test was made up of 72 items and included
categories some of which were similar to the
ones in the GJ test. There were still some
categories in the translation test that were not
found in the other test. The reason was that this
test was constructed based on the concept of
definiteness and how it is assembled with other
features in Persian. Therefore, there were cases
which existed in Persian but not in English (see
Tables 1-4). The purpose of developing the
translation test was to detect possible L1 transfer
effects on how Persian-speaking EFL learners
used the article system in English. The following
table summarizes the details of the test:
The definite, indefinite, and generic items on the
test were evenly distributed for number,
specificity, countability, and function
(subject/object). Considering the nouns in object
position, the test included both direct objects
(with the object marker ra) and indirect objects
(without the object marker ra) so that the
influence of the object marker in correct article
use could be detected. A sample item from the
test is:
وقتی از اتاق بیرون میرفتم بچه ها گریه کردند.
The underlined word is plural, definite, specific
and in the subject position.
Data Collection, scoring, and analysis
Prior to data collection, a pilot testing was done
with 15 language learners. Few inconsistencies
that existed in the tests were modified and an
approximate time limit was set for each one. The
language of instruction was decided to be
Persian. Also, two experts in the field of
language acquisition scrutinized the tests and
validated the content of each one. After data
collection, the internal consistency was
measured for each test, using a Cronbach alpha
coefficient. For the GJ test, the alpha coefficient
was .96. It was calculated to be .87 for the
translation test. Therefore, both tests could be
considered reliable with the specified sample of
the study.
The L2 learners voluntarily participated in the
study. Initially ninety-five learners took the OPT
based on which forty-three were selected and put
into three quite distinct levels. The participants
took the GJ test in the following week. On the
third week, they were asked to do the translation
test. Because the tests were rather long and they
had to be done at times other than class hours, a
one week interval was observed between the
tests. So, the participants met once a week for
three weeks. The second test was the GJ test.
Each correct answer was worth one score and
each incorrect one was given a zero score. Also,
for each incorrect item on the test, only if the
participant had circled * and supplied the right
answer, s/he was scored 1. To put it differently,
just circling *was not enough to indicate the
participant necessarily knew which part of the
sentence made it ungrammatical. For the correct
answers as well, the participants had to mark √
to merit 1. They weren’t granted any scores if
they had left the sentence intact or if they had
chosen ?.On the translation test, the concern was
with the noun phrase in question. It didn’t matter
if the whole sentence was translated correctly.
Each correct answer was granted 1and the wrong
answers were given a zero.
The 1/0 coded data was submitted to the
Statistical Packages in Social Sciences (SPSS,
16) software for the purpose of analysis. For
each category on the tests, the mean percentage
for each individual participant and later for each
proficiency group was calculated. Since there
were four groups of participants and one
independent variable on each category, one-way
between groups ANOVA was performed as the
proper statistical procedure to see if inter-group
differences existed with regard to the features in
question. Wherever necessary, the paired
samples t-test was conducted to detect
significant intra-group differences in
performance. Also, in cases where comparisons
needed to be made between two groups, the
independent samples t-test was used.
VI. Results
The first research question asked if the Persian
learners could eventually reassemble
definiteness, specificity, genericity, number, and
countability on English articles. The results are
presented separately for each feature so that
group performance can be compared. The first
feature to examine is definiteness. Table 8
presents results on the GJ test.
For the sake of comparison, a one-way between
groups ANOVA was used. It showed that the
groups’ performance on this test was
significantly different from each other (Table 9).
At the p< .05 level, the four groups’ performance
on the GJ test was significantly different from
each other both on the definite items (F =
160.49, p = 0.00, r = 0.8) and on the indefinite
items (F = 75.54, p = 0.00, r = 0.8). A Sheffe
post-hoc test was also conducted which revealed
that the learner groups were significantly
different from each other and from the native
group. On the indefinite items of the test,
however, the advanced group could catch up
with the natives as the difference between the
two groups was not significant (p= 0.08).
Elementary and intermediate learners were
significantly different from both advanced
learners and native speakers.
The second feature involved in article selection
was specificity. Table 10 summarizes
performance on [± specific] items.
Table 10
Specificity (%) on the GJ Test
Performance in the [definite non-specific]
context was not as good as that in [definite
specific] for all proficiency groups which shows
that our participants had more doubts about
using the for non-specific definite nouns and
were clearly non-native-like even at the
advanced level. Considering the indefinite
context, using a was unaffected by specificity for
the elementary group as no difference could be
detected. Evidence for these conclusions came
from a one-way between groups ANOVA the
results of which are tabulated below.
At p< .05, there was a statistically significant
difference among the groups in their
performance in the four contexts (p= 0.00). The
effect size (r) was large for all contexts. Post-hoc
comparisons using the Scheffe test indicated
that, in all four contexts, the elementary and
intermediate groups performed significantly
different from the native speakers (p = 0.00 for
the elementary group and p = 0.00 for the
intermediate group). The advanced participants
performed similar to the natives in [definite,
specific] and [indefinite, non-specific]
(p = 0.28 and p = 0.75 respectively). In the two
suspected contexts of [definite, non-specific] and
[indefinite, specific], in which learners are prone
to fluctuation, even the advanced learners
performed significantly different from the
natives (p = 0.00 and p = 0.00 respectively).
The next feature to be checked is genericity.
Performance on singular and plural generic NPs
is displayed in Table 12.
In order to see if the observed differences in
performance on singular and plural generics was
meaningful, a one-way between groups ANOVA
was calculated. The results (Table13) revealed
significant differences between the groups. The
effect size, using eta squared, was large for both
singular and plural generic nouns.
Further post-hoc calculation using a Scheffe test
revealed that, considering singular generic NPs,
none of the learner groups conformed to the
native speaker norm as there was a significant
difference detected (p = 0.00 for all learner
groups). For the plural generic items, however,
the advanced group performed similarly to the
native speakers (p = 0.00 for the elementary and
intermediate groups but p = 0.38 for the
advanced group).
The next feature to be examined is number.
Group performance is presented in Table 14.
Again, differences in performance among the
four groups were tested by a one-way between
groups ANOVA which pointed to significant
gaps (Table 15).
In other words, at p< .05, the participant groups
were significantly different from each other in
using the correct article for both singular (F =
93.35, p = 0.00, r = 0.8) and plural nouns (F =
79.40, p = 0.00, r = 0.8). In both cases there was
a large effect size: using eta squared, 80% of the
total variance could be accounted for by the
groups’ knowledge state. The results of a post-hoc Scheffe test also revealed that the learner
groups differed from each other and from the
native controls in performance on both singular
and plural nouns (p = 0.00 on both singular and
plural nouns for all groups).
The last feature to put under scrutiny is
countability. Only the participants’ performance
on definite items was analyzed mainly because
the tests did not include indefinite mass nouns.
Table 16 summarizes group performance on the
two types of nouns:
At p< .05, the groups were significantly different
from each other in using the correct article for
both count (F = 123.93, p = 0.00, r = 0.8) and
non-count nouns (F = 67.59, p = 0.00, r = 0.7)
based on a one-way between groups ANOVA.
Furthermore, Persian learners were significantly
different from the native control group in
performance on both count and mass NPs. This
was revealed using a Scheffe post-hoc test (p =
0.00).
Based on the results presented so far, it is clear
that the participants in this study could not be
claimed to have acquired the features in question
at none of the three proficiency levels, so the
answer to the first research question is negative.
To arrive at an answer to the second research
question concerning the order of acquisition of
the features, performance of the participants in
the advanced group was further analyzed to
compare their degree of mastery over the five
features. Mastery was defined as correct article
use over 90% for each feature in question in
which case no significant difference could be
detected between natives and non-natives. The
findings on the GJ test are presented in Table 18.
The + mark shows that learner performance was
accurate over 90%
The observation we made was that definiteness
was the only feature in which five of the
advanced participants were accurate. Three
participants could satisfy the accuracy criterion
in article use for the features of specificity and
number. Three participants were accurate in
article use for generic expressions. And only two
of the participants could achieve the accuracy
criteria for the feature of countability. So, the
order of definiteness > specificity/ number >
genericity > countability was found as the
answer to the second research question. It has to
be mentioned that the learner participants in this
study, as will be discussed in the following
section, could not be considered to have acquired
those features so the order that is found is, at
best, tentative.
The last research question was formed to see if
the grammatical function of the noun as subject
or object in Persian would be of any help to the
learners in appropriate article use in English.
Results on the translation test can help arrive at
an answer to that question. (Figure 1)
As the figure depicts, the participants were more
successful in indefinite article use either for
subject or object NPs. This is not surprising
because it was previously concluded that the
Persian learners had fewer problems using the
indefinite article and were actually native-like at
the advanced level (Table 8).However,
performance of the learners showed that they
were more successful in using the definite article
for NPs as objects. Although the results of a
series of paired samples t-tests showed that there
was no significant difference in definite article
use for subject/object NPs for the elementary
group (t(17)= 1.25, p
other two groups had a better performance on
object nouns (t(13) = 2.58, p
for the intermediate group and t(10) = 3.12,
p
answer to the third research question is, then
partially positive. NP function helped the
participants in using the correct definite article
for subject NPs at higher levels of grammatical
knowledge.
VII. Discussion
The first question in this study concerned the
acquisition of English article system.
Comparisons between our learner groups and the
native group revealed that Persian learners were
significantly different form the natives in their
article use, a difference which was observed for
all the features that affect article use. In other
words, the study revealed that Persian learners,
regardless of their L2 state of grammatical
knowledge, were non-native-like in their article
use for ±definite, ±specific, ±singular, ±generic,
and ±count nouns. These results were not in line
with the ones reported by Rezai and Jabbari
(2010). They reported that their Persian subjects
were able to acquire the definite article and,
hence, could provide support for the
interpretability hypothesis. According to them,
definiteness is interpretable in both English (at
both LF and PF) and Persian (just at LF) and so,
it should not pose serious problems to the
learners. More than that, following Ghomeshi
(2003), we assume that there is a definite enclitic
in the spoken form of Persian. So, definiteness
would be considered as interpretable at both LF
and PF (quite like English) and it should not be
an obstacle to learning anyway. Results from the
present study, however, do not seem compatible
with Rezai and Jabbari (2010), henceforth R & J.
We suggest that this is due to the scope and
methodology of their study and the present one.
First, R & J considered only two features of
definiteness and specificity whereas we looked
at the combination of three more grammatical
features and more importantly their assembly in
English and Persian articles. Second, the main
source of their data was an elicitation task. In
fact, in the present study we also used an
elicitation task the results of which showed no
difference between native and non-native
speakers, which is compatible with R & J results.
However, as we found significant differences
between the performances of the L2 learners on
the other two tasks, we found that the elicitation
task was easy and its results did not perfectly
represent the status of knowledge of English
articles. This is not surprising given the fact that
the context of English L2 learning in Iran is an
instructed setting, where English is taught
explicitly in most schools and hence all Persian
learners of English have a good command of
explicit grammatical knowledge. Therefore, we
concluded that this pen and paper task tapped
only the learned linguistic knowledge of our
learners and decided to exclude its results. Third,
R & J pooled the data of their intermediate and
advanced learners, which actually blurs the
picture additionally. Finally, the lack of a group
of native speakers was a major hindrance to
make detailed comparisons and carry out in-depth analyses. In fact, considering these
differences, we construe that R & J’s study does
not provide counter evidence to our study.
The results of the present study then indicated
divergence between native speakers and non-native L2 learners due to morphological
variability in the performance of L2 learners
which seems not to disappear even at advanced
levels hence a persistent problem. We suggest
that the source of this problem could be
morphological feature combination and
conditioning as proposed by Lardiere 2005. She
assumes that L2 learners ‘initially seek the
morphological equivalents from already
assembled lexical items in the L1 to analyze L2
conditioning environment’ (2009, p. 213). Proof
of this assertion was found in our study where
the participants had more difficulty in using the
definite article than the indefinite one, where
they could not adequately mark [± definite] on
count/non-count as well as singular/plural NPs,
and where they could not use the correct article
for singular generic NPs. We would like to
argue, then, that using English articles to mark
in/definiteness by Persian learners reflects L1
transfer effects. This effect is not syntactic in
nature, though, because the DP structure and
definiteness as a feature exist in both languages.
There are definite, indefinite, and zero articles in
both English and Persian. And article use is
affected by other features as specificity, number,
countability, and genericity in both languages.
Despite all these cross-linguistic similarities, the
developmental patterns of the participants’
performance did not bear significant
resemblance to the native speaker group. They
obviously had problems in using English articles.
Why would such learners, regardless of their
language proficiency, have persisting problems
in article use? The answer relates to the fact that
the difference between the two languages is of a
morphological nature as the features that have a
role in article use are bundled differently and
into different lexical items in English and
Persian. The definite article (as a lexical item) in
Persian is not overtly realized in the formal
language. It is reasonable for Persian-speaking
learners to have persistent difficulty with the
English definite article. Other studies have also
pointed to the difficulty the learners face when
one or both of English articles are lacking in
learners’ L1s (Jaensch & Sarko, 2009; Kaku,
2006). As for the indefinite article, it
encompasses different feature combinations in
Persian as well. The present study is also backed
up by Lardiere (2005) where she argues that
knowledge of definiteness feature does not
correlate with Definite article use in the
performance of a Chinese L2 learner of English.
Similarly, Dominguez, Archi and Myles (2011)
claim that feature reassembly is a necessary
process during the course of acquiring a second
language. The feature reassembly hypothesis can
best account for such problems as the one in the
present study because it does not consider a
facilitative role for the L1 even in situations
where the point is seemingly identical across the
two languages. In fact, the situation for L1
Persian learners in this study is exactly like this:
the contexts seem to be identical across the two
languages. However, the participants’
performance showed that the identicality is
superficial. As they could display mastery in
none of the features realized in article use,
feature reassembly seems to provide a
reasonable explanation for such failure. The
reason for lack of mastery in article use could be,
therefore, the combination of the features which
differed across the L1 and L2 and the inability of
the learners to recombine those L1 features so
that they would conform to the L2 configuration.
The second research question, which asked about
the order of acquisition of the article related
features, cannot be definitely answered based on
the data from this study as it was revealed that
the learners at all three levels of grammatical
knowledge had not achieved full mastery over
article use as English native speakers. Being so,
no order of acquisition could be arrived at. It has
to be mentioned, however, that our participants
were more proficient in using the appropriate
article to mark definiteness and specificity on
English NPs and that they had the most difficulty
in dealing with the feature of countability.
Reasons for the observed pattern can be traced in
how the two languages realize each feature.
They both have means to show definiteness and
specificity on nouns (Table 1) whereas the
realization of definiteness on non-count nouns is
more complicated because of cross linguistic
differences (Table 4). Also, it was elaborated on
in Table 2 that the two languages stand in sharp
contrast as to their generic forms. Again, generic
NPs in Persian are bare while this is not the case
in English. In other words, difficulty in article
use occurred in situations where either the means
were different in the two languages or they were
absent in one language and present in the other.
Moreover, performance of these participants
showed their sensitivity to NP function within
the sentence which was the answer to the third
research question. They did much better in
contexts where the NP in question was in object
position. Performance on subject NPs was not at
all comparable to object NPs, showing that NP
function affected article use at least for the
participants in this study. The effect of noun
function on article use by learners from other
language backgrounds is yet unexplored since it
is not documented by any study so far. In case
that no such effect is found in other languages,
then, reassembly of features would be more
difficult for Persian learners because they have
to additionally be sensitive to noun function.
This implies that, for some learners with a
specific language background, reassembly of
features would be intertwined with other aspects
of grammar (for instance, NP function). So,
reassembly would be language specific, making
it more difficult for some learners but not others.
VIII. Conclusion
The central goal of this study was to find out if
Persian-speaking learners of English were able
to use the article system in a native-like fashion.
Analysis of the data, collected through two
different measures, showed that the participants
had not mastered appropriate article use even at
the very advanced level. Explanation was sought
from the feature reassembly hypothesis
according to which the source of such failure
was unsuccessful reconfiguration of the features
related to article use from the way they were
bundled in Persian into L2 English. Our data
were in line with Lardiere’s (2008, p. 114)
assertion that ‘persistent L2 variable
morphological production (omission and/or
faulty use) is attributable to differences in the
‘conditioning environment’ for the assembly of
features into lexical items between L1 and L2’.
The findings of this study indicate that the
problem Persian-speaking learners faced was not
acquiring new features but rearranging the
already existing ones into different lexical items
in their L2. Elsewhere, Lardiere (2009) suggests
that ‘if a feature contrast is detectable, it is
eventually acquirable’ (p. 214). Since the
advanced learners in this study were shown to be
highly proficient in other areas of English
grammar (scored above 55 on the OPT), the
question still remains as to how long they would
still need to figure out the contrasts and to be
able to perform like a native speaker. With this
goal in mind, attention should now be turned to
conducting studies that focus on the pedagogical
aspect and look for more fruitful methods of
instruction. The present study also paves the way
for future studies that focus on why the NP
function affects article use or on other possible
aspects of grammar that may play a role.
