| |
Abstract:
Agrammatic aphasia is associated with omissions and
substitutions of inflections and "function words" (1). We propose
that a single deficit contributes to all of these problems. The
deficit is hypothesized to impair syntactic structure building, and
to specifically affect concatenation. It likely results from damage
to left frontal/basal-ganglia circuits that also underlie motor
sequencing. Neuropsychology. Non-fluent aphasia with agrammatism is
associated with damage to left frontal/basal-ganglia structures.
Fluent aphasia with anomia is associated with damage to left
temporal/temporo-parietal regions and the possible sparing of
(aspects of) grammar. This dissociation suggests that (at least
some aspects of) grammar may be computationally and
neuroanatomically distinct from other linguistic faculties, and may
be particularly dependent upon left frontal cortex (1, 2).
Moreover, grammar may be subserved by a frontal/basal-ganglia
procedural system implicated in motor sequencing (3).
Linguistic Theory. According to the Principles and Parameters
framework (e.g., 4), syntactic computations depend on the
construction of hierarchies of lexical and functional categories
through sequential operations of concatenation, from subordinate
(lower) to super-ordinate (higher) categories. Functional
categories correspond to function words and license inflections.
The nature of functional categories and their place in the
syntactic hierarchy predict that an impairment of syntactic
structure building should disturb inflection and function words
together, and moreover, should do so in a graded manner, with forms
dependent upon higher projections being more prone to errors.
The Study. We examined the inflectional errors of
English-speaking agrammatic non-fluent aphasics with left
frontal/basal-ganglia lesions, and anomic fluent aphasics with left
temporal/temporo-parietal lesions in (i) the elicited past-tense
production of 20 regular and 16 irregular verbs in sentence
contexts, and (ii) the isolated-word reading of 17 irregular and 17
regular past-tense forms. Two non-fluent and 6 fluent aphasics
completed the production task. Nine non-fluent and 5 fluent
aphasics completed the reading task. Age- and education-matched
normal subjects served as controls.
Individually and as a group, the non-fluent but not the fluent
aphasics were more impaired than controls, in both production and
reading, at both regulars and irregulars. Individually and as a
group, the non-fluent aphasics, but not the fluent aphasics or
controls, showed a graded impairment, producing many unmarked
errors, some -ing-substitutions, and few -s-substitutions,
respecting the bottom-up order of syntactic projections from the
lexical category of Verb, to the functional categories of
Participle, Tense, and Subject-Verb Agreement (5).
A third of the non-fluent aphasics' unmarked errors in both
tasks were on irregular verbs, precluding an impairment of
morphological affixation as the sole factor yielding unmarked
forms, and arguing against a purely phonological explanation of the
deficit (6). The impairment was graded within subjects, a pattern
that is not consistent with a categorical deficit of functional
projections (7, 8). The fact that the non-fluent but not the fluent
aphasics also had greater difficulty at regular than irregular
past-tenses suggests an impairment of concatenation, both in
morphophonology and in syntax. The distinct response patterns of
the fluent aphasics suggests a relative sparing of structure
building. Patterns of inflectional errors found in the isolated
past-tense reading task support the view that word-formation
requires syntactic structure building (9), and challenge theories
of syntax-independent word-formation (4, 10).
|
