Abstracts 2005 - The Psychonomic Society

Papers 228–234 Saturday Afternoon
tive to the multidimensional structure of the family tree. I explore this
finding further in three experiments. The first tests whether these effects
occur with complex structures for which a familiar prototype
such as a family tree is not available. The second looks at comparisons
involving a more complex family tree in which up to four degrees of
lineality are possible. The third demonstrates that the results depend
on availability of item-specific information.
2:30–2:45 (228)
The Effects of Working Memory Capacity and Cognitive Load on
Instructional Design: Subgoals Still a Good Idea. RICHARD
CATRAMBONE, Georgia Institute of Technology—Learners studied
example physics mechanics problems that either did or did not highlight
relevant subgoals. They then solved isomorphs and far-transfer
problems. During training and transfer, participants provided cognitive
load ratings. Transfer performance was superior for participants
who studied examples that highlighted subgoals. Cognitive load ratings
were lower for the subgoal condition during both the training and
problem solving phases. Participants’ working memory capacity affected
performance in the no-subgoal condition but not in the subgoal
condition. This result suggests that examples designed to convey subgoals
can place reduced loads on working memory, so that learners
can more effectively acquire useful problem solving knowledge even
with lower working memory capacity; less effective instructional materials
might require learners to have greater working memory capacity
in order to learn from them.
2:50–3:05 (229)
Symbolic and Perceptual Comparisons With Remembered Instructions.
WILLIAM M. PETRUSIC, Carleton University, SAMUEL
SHAKI, College of Judea and Samaria, & CRAIG LETH-STEENSEN,
Carleton University—Semantic congruity effects (SCEs) were obtained
in each of two experiments, one with symbolic comparisons and the
other with comparisons of visual extents. SCEs were reliably larger
when the instructions indicating the direction of the comparison were
represented by consonant–vowel–consonant (CVC) nonsense syllables
that had been associated with the usual instructions in a preliminary
learning phase of the experiment. The enhanced SCE with the
CVC instructions is not readily explained by any of the non–evidence
accrual theories of the SCE (e.g., expectancy, semantic coding, and
reference point). On the other hand, the general class of evidence accrual
views of SCEs, such as those developed in Leth-Steensen and
Marley (2000) and Petrusic (1992), receive considerable empirical
support.
Spelling and Print-to-Sound
Conference Rooms B&C, Saturday Afternoon, 1:30–3:30
Chaired by Rebecca Treiman, Washington University
1:30–1:45 (230)
Long-Distance Anticipatory Effects in Spelling-to-Sound Translation.
REBECCA TREIMAN & BRETT KESSLER, Washington University—Phonological
decoding of written words is often viewed as a
serial process. To test this view, we examined whether college students’
productions reflect sensitivity to English spelling patterns that
require long-distance nonserial processing. Word-initial g is usually
velar (hard), but before e, i, or y, it is coronal (soft) if the root word
is classical or Romance (e.g., giant vs. Germanic give). Students were
more likely to use the coronal pronunciation for nonwords with classical
or Romance spelling patterns (e.g., gepous, gileph) than for those
with native Germanic patterns (e.g., gepness, gileff ). This tendency
was found also for rare words, and to some extent for initial c. Students
who had studied a Romance language produced more coronal
pronunciations of initial g and c than did those who had not. Thus,
when decoding word-initial letters, readers consider complicated and
largely untaught patterns that extend to the ends of words.
36
1:50–2:05 (231)
A Challenge to Chaining Theories of Serial Order: An SRN Investigation.
ARIEL M. GOLDBERG, ADAM B. BUCHWALD, & BRENDA
C. RAPP, Johns Hopkins University (read by Brenda C. Rapp)—Serial
behavior is the production of a set of items in the proper sequence. For
example, spelling is an everyday task requiring serial behavior. Prominent
among theories of serial order are chaining theories, which hold
that serial order is achieved through interitem associations. We used
simple recurrent networks (SRNs), which are generally assumed to instantiate
complex chaining, to test the chaining hypothesis of serial
order in spelling. The performance of SRNs trained to spell a large
corpus of words was compared with that of two dysgraphic individuals
suffering from damage to the serial order component of the spelling
process. Although SRNs with simulated damage produced length effects
matching those of the dysgraphics, very significant differences were
found for multiple features, including serial position effects. These
findings constitute an important challenge to chaining theories of serial
order in spelling and, by extension, to chaining theories of serial order
more generally.
2:10–2:25 (232)
Nature of Orthographic Representations: Disrupting the Graphemic
Buffer in Unimpaired Spellers. ANGELA C. JONES & JOCELYN R.
FOLK, Kent State University (read by Jocelyn R. Folk)—We investigated
the hypothesis that the orthographic representations involved in
spelling are multidimensional; consonant/vowel status, doubling information,
and syllabic structure are part of orthographic representations
(see, e.g., Caramazza & Miceli, 1990; McCloskey et al., 1994).
The evidence for this hypothesis has come exclusively from studies
of dysgraphic patients with damage to the graphemic buffer, the processing
stage responsible for maintaining the activation of orthographic
representations as serial output occurs. In the present study,
we sought to examine the nature of orthographic representations in a
neurologically intact sample by inducing cognitive load in unimpaired
spellers, hindering the operation of the graphemic buffer. We examined
the nature of the spelling errors produced under these conditions.
We found convergent evidence for the multidimensional nature of orthographic
representations. In addition, we found evidence of texture
across and within orthographic representations; some words or letters
within a word may be more strongly represented than others.
2:30–2:45 (233)
Even Preschool Children Show Orthographic Influences on Phonological
Awareness Tasks. ANNE CASTLES, JOANNE WEDGWOOD,
& VIRGINIA HOLMES, University of Melbourne—It is well established
that experienced readers show influences of orthographic
knowledge on their performance on tasks ostensibly tapping pure
phonological awareness. For example, people tend to report that there
are more phonemes in the word pitch than the word rich, because of
the extra letter in the former word. Here we demonstrate that even
preschool children are influenced by their emerging orthographic abilities
when performing such tasks. Three- and 4-year-old children were
required to identify the initial phonemes of words describing pictured
objects, and their familiarity with the written form of the letters corresponding
to those phonemes was assessed. An item-specific effect
of letter knowledge was found: The children performed significantly
better on items in the phonemic awareness test if they could identify
the corresponding letter than if they could not. These results have important
implications for theories of the role of phonological awareness
in reading acquisition.
2:50–3:05 (234)
The Effect of Age of Acquisition on Bilingual Word Translation.
J. MICHAEL BOWERS & SHELIA M. KENNISON, Oklahoma State
University (read by Shelia M. Kennison)—An experiment with Spanish–
English bilinguals tested the hypothesis that connections between
conceptual representations and some first language (L1) words are