| |
Abstract:
Although much research has been done looking at hemisphere
asymmetries in lexical semantics, relatively little research has
addressed the contributions of each cerebral hemisphere to
sentential processing. Recent studies have suggested that the left
hemisphere (LH) is more involved in semantic and syntactic
sentential processes than the right hemisphere (RH). For example,
some studies have found that the recognition of words presented to
the LH is influenced by the degree of constraint provided by a
sentence prime, but words presented to the RH are not, and that a
random rearrangement of word order disrupts processing more in the
LH (Faust, 1998). Previous research using semantically anomalous
sentences to look at message-level processing has suggested that
the RH may show priming, rather than interference, for semantically
anomalous sentences, but there were some methodological weaknesses
limiting this interpretation. The present study sought to provide a
stronger test of the idea that the RH uses only an intralexical
mechanism to facilitate word recognition in sentence contexts,
while the LH uses both intralexical and message-level mechanisms.
Anomalous sentences were created such that they were completely
normal up until the final target word. Anomalous and normal
sentences were investigated under two conditions; one where the
head noun phrase contained a word which was highly related to the
sentence-final target (R+ condition) and another where there were
no words in the sentence which were semantically related to the
target (R- condition). The neutral baseline condition was created
by changing the head noun and the verb to properly inflected
pseudowords, with the aim of leaving syntactic cues intact but
providing no semantic constraints (stimulus examples below).
Critical sentences were extensively normed using a cloze procedure,
a bizarreness rating task, and a relatedness judgement task.
Subjects viewed the centrally presented sentence fragments for 1050
(ms), then a centrally presented fixation marker for 350 ms. The
critical sentence-final target word was presented for 100 ms to
either the left visual field(LVF/RH) or the right visual
field(RVF/LH) for lexical decision. The results (see table below)
showed equivalent priming for normal sentences in both hemispheres
which was dependent upon the presence of a related word. Anomalous
sentences were bilaterally inhibited, but only in the absence of a
related word. The presence of a related word in the sentence
influenced priming in both hemispheres, indicating bilateral
sensitivity to lexical relations within sentential contexts. Since
the normal, anomalous, and neutral sentences differed only at the
semantic message level (all were syntactically well-formed), we
conclude from this study that the LH, as well as the RH, is
sensitive to some types of message level meaning. We do not claim
that message-level processing is equivalent across the hemispheres,
but suggest that RH semantic processes may support a type of
sentence level representation which is able to distinguish between
semantically normal and anomalous sentences.
Sample Stimuli:
R+ Normal: The weary campers set up the TENT
R+ Anomalous: The weary campers devoured the TENT
R+ Neutral: The weary deaves dod up the TENT
R- Normal: The weary husband set up the TENT
R- Anomalous: The weary husband devoured the TENT
R- Neutral: The weary struman brolled the TENT
Table: Normal Sentence Priming Anomalous
Sentence Priming
LVF/RH 50 14 -13 -32
RVF/LH 35 -5 -3 -49
|
