S1 (FriAM 1-65) - The Psychonomic Society

Friday Morning Papers 62–65
logical congruency effect in object, but not in color naming. Differences
in the typical order of object nouns and color adjectives in
French and English provide a plausible account for the cross-linguistic
contrast in phonological activation.
10:40–10:55 (62)
Word Preparation Addresses Syllables but Not Segments in Mandarin
Chinese. JENN-YEU CHEN, National Cheng Kung University,
PADRAIG G. O’SEAGHDHA, Lehigh University, & KUAN-HUNG
LIU, National Cheng Kung University (read by Padraig G.
O’Seaghdha)—In Mandarin, speakers benefit from foreknowledge of
what the first syllable of a disyllable will be, but not what the onset
is (Chen, Chen, & Dell, 2002). We replicated the absence of an onset
preparation benefit for disyllables (Experiment 1). Then we provided
stricter tests using monosyllables. We found no preparation effect
whether targets were cued by other monosyllables that could form a
compound (Experiment 2) or that could not (Experiment 3), or by a
nonlexical symbol (Experiment 4). These results are in marked contrast
to implicit priming results with European languages. We outline
a theory in which advance planning addresses the primary level of
phonological encoding below the word—syllables in Mandarin, segments
in English and other Indo-European languages. In Mandarin,
segments may be activated but are not prepared until after syllable
production begins. Thus, the control structure for word encoding is
fundamentally different in these languages.
11:00–11:15 (63)
Interactive Effects in Speech Production: Feedback or Perceptual
Monitoring? MELISSA M. BAESE & MATTHEW A. GOLDRICK,
Northwestern University (read by Matthew A. Goldrick)—Interactive
effects are well documented at lexical and phonological levels of the
production system (e.g., the mixed error effect; Dell, 1986). Acoustic
analysis reveals such effects extend to postphonological processes:
Lexical properties of words modulate their phonetic properties.
Specifically, voice onset time (VOT) varies according to lexical
neighborhood structure. For example, the /k/ in “cod” has a longer
VOT than the /k/ in “cop.” This reflects contrasting neighborhood
structure: “cod” has a highly similar neighbor “god,” whereas “cop”
does not (“gop” is not an English word). We then examine the source
of this interactive effect. It could reflect enhancement of articulation
to help perception (listener modeling) or feedback-driven enhancement
of lexical and phonological representations. To contrast these accounts,
participants produced words in an interactive task where phonetic
enhancement was unnecessary for successful communication.
10
Consistent with a production-internal feedback mechanism, phonetic
enhancement was still present in speakers’ productions.
11:20–11:35 (64)
Amnesic H.M.’s Sentence Production: Parallel Deficits in Memory
and Language. DON G. MACKAY & CHRISTOPHER B. HADLEY,
UCLA, & LORI E. JAMES, University of Colorado, Colorado Springs—
To test conflicting hypotheses regarding amnesic H.M.’s language abilities,
this study examined H.M.’s sentence production on the Language
Competence Test. The task for H.M. and 8 education-, age-, and IQmatched
controls was to describe pictures using a single grammatical
sentence containing 2 or 3 prespecified target words. The results indicated
selective deficits in H.M.’s picture descriptions: H.M. produced
fewer single grammatical sentences, included fewer target words, and
described the pictures less completely and accurately than did the controls.
However, H.M.’s deficits diminished with repeated processing of
unfamiliar stimuli, and virtually disappeared for familiar stimuli, selective
effects that help explain why some researchers claim that H.M.’s
language production is intact. Besides resolving the conflicting hypotheses,
this study demonstrates parallel deficits and sparing in H.M.’s
language and memory, and replicates other well-controlled sentence
production results. Present results comport in detail with binding theory
but pose problems for system theory accounts of H.M.’s condition.
11:40–11:55 (65)
Not Saying What’s on Your Mind: How Speakers Avoid Grounding
References in Privileged Information. LIANE WARDLOW LANE &
VICTOR S. FERREIRA, University of California, San Diego (read by
Victor S. Ferreira)—Speakers’ descriptions of shared objects sometimes
implicitly refer to private information (i.e., privileged objects). In
Experiment 1, speakers coped with a cognitive pressure to attend to
privileged objects and a communicative pressure to avoid referring to
privileged objects. The results show that when cognitive and communicative
pressures compete, cognitive pressures win: Speakers implicitly
referred to privileged objects more, even at the expense of communicative
success. Experiment 2 showed that this was not due to task
confusion. Experiment 3 directly tested whether increased attention
causes increased references to privileged objects. The results showed
more references to privileged information when privileged objects were
made relatively more salient. Experiment 4 combined the attention manipulations
of Experiments 1 and 3, revealing that they were not independent,
suggesting that the separate effects were caused by similar or
related mechanisms. Overall, speakers are compelled to cope with cognitive
pressures, even at the expense of communicative efficacy.