| |
Abstract:
In research of the "mental lexicon," an important issue has
been inflected forms: Are they stored as whole words, or generated
by rule when needed? Evidence from English has suggested that
irregular verb forms are stored, but regular forms are created on
demand. The present study claims that regularity is less important
than productivity for predicting the mental processing of
inflection.
The data used are from verb inflection by Japanese
native-speaker adults. The verb inflection system in Japanese,
though complex, is highly regular, but contrasts with English verbs
since regularity is not correlated with productivity. That is, when
new verbs enter the language, they almost always use one verb
structure, while other, equally regular, verb paradigms are
unproductive. Using nonce verbs of both productive and unproductive
type, we show that many native speakers cannot reliably generate
common inflected forms for the unproductive verb type.
Previous research was performed by Vance (1987, 1991) in
Japanese and Schnitzer (1993, 1996) in Spanish, who both presented
nonce verbs to adults. In comparison, the present experimental
design has several improvements. First, Vance used written
questionnaires, but our experiments are completely aural/oral.
Second, Schnitzer used oral presentation and response, but his
stimuli were single sentences, while we used longer, more natural
language samples. In our research, each subject listened to a
number of short (less than one minute) recorded dialogues. In each
one, two forms of a novel verb appeared (a total of four
instances). The subject was then asked to produce orally another
form of the same verb, one which was not heard in the recording.
The subject's productions were audio-recorded, and response times
were calculated. A within-subjects design compares performance on
nonce verbs of both productive and unproductive verb types.
This study shows that sheer algorithmic regularity in language
may be less relevant to the adult human user than strategies which
are tailored to the contemporary situation of use. Current work
extends the experimental paradigm to children and to
second-language learners. We hypothesize that both of these groups
may be able to inflect novel verbs better than native-speaker
adults, since they encounter previously unheard verbs of all types
fairly frequently, and thus for them all verbs are
"productive."
References
Schnitzer, M. L. (1993). Steady as a rock: Does the steady state
represent cognitive fossilization? Journal of Psycholinguistic
Research, 22(1), 1-20.
Schnitzer, M. L. (1996). Knowledge and acquisition of the Spanish
verbal paradigm in five communities. Hispania, 79, 830-844.
Vance, T. J. (1987). An Introduction to Japanese phonology.
Albany, NY: State University of New York Press.
Vance, T. J. (1991). A new experimental study of Japanese verb
morphology. Journal of Japanese Linguistics, 13, 145-166.
|
