申論題

Morphological congruency and the acquisition of L2 morpheme

    The acquisition of grammatical morphemes by second-language (L2) Iearners
has received a great deal of attention in SLA research. There is a consensus that
grammatical morphemes, such as plural, tense, and gender markers, are difficult
structures for adult leamers to acquire, but researchers disagree as to whether they are
ultimately acquirable.
    Much research evidence has accu cumulated over the past two decades that is
directly related to the issue. However, the evidence has been mixed, allowing each
side to find support for their position. For example, a number of longitudinal studies
have shown that their participants had difficulty in accurately using grammatical
morphemes in spontaneous L2 production and that they exhibited little progress over
an extended period of time, even though they were considered highly proficient L2
users othervise (e.g,Jia, 2003; Lardiere, 1998, 2003; Long, 2003; Schmidt, 1983).
On the other hand, native like accuracy in morpheme use has been observed in adult
L2 learners in other studies (e.g., Flege et al., 1999; Goad & White, 2006;
Herschensohn, 2001; White, Valenzuela, Kozlowska-MacGregor, & Leung, 2004).
    A number of factors may be responsible for such conflicting findings. One can
expect L2 leamers' proficieney, for example, to play an important role, and the
participants' L2 proficiency has varied a great deal across different studies. Some
.'studies examined beginning to intermediate learners who have limited experience in
the L.2 (c.g, Herschensohn, 2001); other studies have tested advanced learners who,
with many years of residence in the target language environment, had probably
reached a steady state in their L2 morphosyntax (e.g, Lardicre, 1998).
    However, L2 proficiency alone does not provide an adcquate explanation of the
conflicting results. Another important factor is research method. A variety of methods
have been used in studying morphological development. Data collection in some
studies involved only a single test session lasting no more than an hour, other studics
have been longitudinal, tracking the same L2 learners' morphological performance
over a period of scveral years. There is much diversity in the specific tasks used for
data collection as well, such as grammaticality judgment, sentence completion,
picture- or video-based rctelling, interviewing, and self-paced reading. Particularly
significant in this regard is the fact that these tasks differ in the extent to which they
allow participants to rely on their explicit knowledge. Some tasks draw the
participants' attention to grammatical accuracy and give them plenty of time and
motivation to apply explicit knowledge. Consequenty, the resuls are informative
about the participants' possession of explicit knowledge of a target structure more
than the development of an ability to use the structure spontancously in
communication. Some other tasks are designed or adopted to minimize the
invoivement of explicit knowledge. Such methodological differences could have also
contributed to the inconsistency in the findings.

The Morphological Congrueney Hypothesis

    The third factor, one that is the focus of the present study, is the type of
morphological markers under investigation and their morphological congruency
across languages. Languages difer in terms of which meaning is grimnaticalized and
thus morphologically marked. Some languages, such as English and Russian, mark
plurals, whereas plural marking is highly restricted and optional in other languages,
such as Chinese and Japanese.
    While leaming an L2, a learner may have to learn a grammatical morpheme that
does or does not have a counterpart in his or her first language (L1), Such
: crosslinguistic relationships in morphological marking,pan be described as involving
differing levels of morphological congruency. Two languages are morphologically
congruent when both grammaticalize and mark a meaning morphologically. They are
morphologically incongr ruent when a grammatical morpheme is present in one
language but not in the other. Morphological congruency always refers specifically to
a particular morpheme, rather than two languages in general. Thus, French and
English are morphologically congruent in plural marking but not in gender marking.
Chinese and English are morphologically congruent to some extent in aspect marking
but not in tense marking.
    Given the prevailing evidence for the influence of learners' L.1 in L2 leanning, it
is reasonable to suggest that morphological congruency may play an important role in
the learning of L.2 morphemes. In the present context, we are particularly interested in
how morphological congruency affects ultimate attainment in L2 morpheme
acquisition. We hypothesize that when L.2 learners reach an advanced or near-native
level of L2 proficiency, only congruent Iearners (i.e, those whose L1 has a
corresponding morpheme to the target L2 morpheme) are able to reach nativelike
proficiency in acquiring an L2 morpheme. Incongruent LI2 learners will find it
extremely difficult, if not impossible, to develop nativelike competence with respect
to the same L2 morpheme. We call this proposal the morphological congrueney
hypothesis.
    There are indications of such a morphological congruency effect in the literature.
One such indication is that when advanced L2 Iearners have been found to have great
difficulty in performning in a nativelike manner with respect to plural or past tense
marking, they have often turned out to be individuals whose L1 does not have an
obligatory grammaticalized plural or tense marker (Jia, 2003; Lardiere, 1998, 2003;
Long, 2003; Schmidt, 1983). These studies are similar in several important ways.
They are all longitudinal studies whose data collection spanned a period of several
years; their participants had lived in the target language environ ment for many years
and were likely to have reached a steady state in their morphosyntactic development
in English; their participants all had an L1 that does not mark plural or past tense
morphologically (i.c, Chinese and Japanese); finally, they all produced much lower
. accuracy rates in plural or past tense marking than native,speakers and showed litle
progress over the period of many years while being studied.
    A second indication comes from studies involving different groups of LI.2 learners
whose L1s vary with respect to whether they are morphologically congruent with the
target language. In such studies, learners with a congruent L1 almost always
outperform learners with an incongruent L1. In Hawkins and Liszka's (2003) study of
past tense marking in ESL speakers, for example, the accuracy rates for the congruent
group-a group of German leamers of English--were above 95%, but the
proficiency-matched incongruent group- the Chinese ESL speakers-showed much
lower accuracy rates on both regular and irregular verbs (62.5% and 84.2%,
respectively). Similarly, in a study of gender agreement in Dutch L2, Sabourin, Stowe,
and de Haan (2006) tested three proficiency-matched groups of participants whose
LIs either mark gender (German and Romance languages groups) or do not mark it
(the English group). English-spcaking participants' accuracy was significantly lower
than that of the other two groups (Experiment 2, English 51%, German 81%,
Romance 71%).
    A third line of evidence can be found in studies that bave tested the same group
of participants on a number of different morphemes (Bialystok, 1997; Franceschina,
2001; White, 2003). Some of these morphemes had a counterpart in the participants'
L1, but others did not. In line with the morphological congruency hypothesis, the
participants in all these studies did much better on the morphological structures that
were instantiated in their L1. For example, in a longitudinal case study of a Turkish
ESL speaker, White (2003) examined the participants' performance on four
morphemes (i.c., third-person -s, past tense, plural, and articles). Turkish is similar to
English in marking plural and past tense, but it does not have articles. Consistent with
the prediction of the morphological congruency hypothesis, the participant's accuracy
on articles was much lower than on plural and tense markers. Similar findings were
shown in another case study by Franceschina (2001), whose participant, Martin, was a
native speaker of English but was also a highly proficient French L.2 speaker. Among
the errors he made in French, 93% were related to gender, compared to only 7%
related to number marking. Again, morphological congruency scemed to dctermine
the learner's success in developing nativelike accuracy with individual morphemes.
 

A Psycholinguistic Approach to Testing the Hypothesis
    We put the morphological congruency hypothesis to the test by examining
morphological performance by two ESL groups in a self-paced reading task. The
target morpheme under investigation was the English plural marker. As Russian
marks plural and Japanese does not, at least not obligatorily, Russian and Japanese
ESL speakers were chosen to represent a morphologically congruent and a
morphologically incongruent group, respectively. The participants in both groups
were highly proficient ESL speakets who have had many years of residence in an
English-speaking country.
    The self-paced reading task was chosen for its advantages in minimizing the
involvement of explicit knowledge (see Jiang, 2004, 2007, for a discussion of its
advantages). In perforning the self-paced reading task, a participant reads sentences
word by word. Only one word appears on a computer monitor at any time. The
participant has to press a button to see the next word, which appears to the right side
of the preceding word to mimic the real-life lefit-to-right reading direction. The
word-by-word presentation of sentences allows one to record the participant's reading
time on each word as primary data. The task emphasizes both reading time and
comprehension. For the latter purpose, questions are given to ensure and check the
participant's reading comprehension. With this method, the acquisition of a structure
is determined by comparing individuals' reading times for a grammatical sentence
and its ungrammatical counterpart, as illustrated below:

la. They met several of the board members during their visit.
1b. *They met several of the board member during their visit.

    A delay in reading any underlined word of the ungrammatical sentence, as
compared to the same words in the grammatical version, would suggest that the
individual has developed an automatic sensitivity to errors involving a particular
structure--in this case, the plural morpheme. This sensitivity can then be taken as
evidence for the acquisition of the structure. By contrast, an absence of such
differences in reading time between the correct and incorrect versions is considered to
be evidence of a lack of integrated linguistic knowledge about the target structure.
END OF READING PASSAGE
Questions:
1. [COMPREHENSION CHECK QUESTIONS]