| |
Abstract:
Semantically related word substitution errors of the sort
"Put it on my FOOT" when "Put it on my HAND" was intended are the
most common type of lexical substitution errors. In models of
language production (e.g., Garrett, 1980; Levelt, 1989) they are
accounted for in terms of failure in the mapping between conceptual
structures and abstract lexical representations, also referred to
as "lemmas". It is a well known fact that these errors are
constrained by the grammatical category of the target, such that
nouns substitute for other nouns, verbs for verbs, etc. This
grammatical category constraint, however, may be accounted for in
different manners. On one account, it may be due to the fact that
words belonging to the same grammatical category are also
semantically more similar (semantic account); alternatively it may
arise as a consequence of phrasal mechanisms that constrain the set
of potential intruding words (syntactic account).
In a series of two experiments with German speakers we assessed
these two hypotheses by investigating whether a syntactic variable
that does not have a strong semantic correlate (grammatical gender)
constrains potential target-intruder pairs. In both experiments,
pictures of animals and body parts were presented consecutively on
a computer screen and participants were instructed to simply name
the picture (bare noun production, Experiment 1), or to produce a
det+N phrase (phrase production, Experiment 2). Lexical errors in
both experiments were predominantly semantically related word
substitution errors in which the target and the intruder were both
members of the response set (e.g., "Esel" [donkey] => "Pferd"
[horse]). In Experiment 1 (bare noun production), no gender
constraint was observed: the likelihood that the target and the
intruder shared the same gender was at chance level; while a gender
constraint was found when speakers produced phrases (Experiment 2).
This latter finding is in line with observations from spontaneously
occurring semantically related errors in German (Marx, 1999). Since
grammatical gender has little semantic force, the present results
falsify a semantic account. This conclusion is further supported by
the fact that no such constraint was found in single word
production. The finding of a gender constraint in phrasal
production also indicates that the mechanisms of lexical selection
at the lemma level is sensitive to the grammatical properties of
the target lemma, therefore, suggesting that such properties may be
active and used before selection occurs.
Garrett, M.F. (1980). Levels of processing in sentence
production. In B. Butterworth (Ed.). Language production: Volume I:
Speech and talk. London: Academic Press.
Levelt, W.J.M. (1989). Speaking: From intention to articulation.
Cambridge, MA: MIT Press.
Marx. E. (1999). Gender processing in speech production: Evidence
from German speech errors. Journal of Psycholinguistic Research,
28, 601-621.
|
