Papers 1-9 Friday Morning
Papers and Posters Presented at the 32nd Annual Meeting of the Psychonomic Society The Hyatt Regency San Francisco, San Francisco, California November 22-24, 1991
A1TENTION I ASSOCIA11VE LEARNING: ANIMAL I Embarcadero AB, Friday Morning, 8:00-9:40 Embarcadero CD, Friday MoroiJIK, 8:00-9:50
Chaired by lanu!s C. lohnston. NASA-Ames Research Center Chaired by Lorraine G. Allan, McMaster University
8:00-8:20 (1) 8:00-8:15 (6) Selective Attention to an Item Is Stored as a Feature 01 the Item. Iocentive Shifts 10 Pavloviao Fear CoadItIoaIoa: A US Modula GEORGE SPERLING, New York Uniwmty, & STEPHEN A. WURSf, tion View. MICHAEL S. FANSELOW & STACEY L. YOUNG, SUNY at Oswego-Subjects must detect a repetition in a stream of 30 UCU-CSs associated with painful USs produce a naloxone reversi characters flashed at to per second. Items alternate in either color ble analgesia which acts as negative feedback, reducing the impact of (black/white). size, orientation, or spetial frequency . Selectively attending predicted USs. Accordingly, naloxone "disregulated" conditioning; a feature (e.g., black) never improves detection of repeated attended learning curves went to the same asymptote regardless of US intensity . (black) versus unattended (white) items. Many counterintuitive results Shifting drug treatment during acquisition had effects that paralleled US are explained by assuming (I) all items are stored in short -term mem intensity shifts. Naloxone blocks the retarded acquisition found when ory (there is no perceptual flItering) and (2) attention to an item is itself strong shock experience follows mild shock. These data indicate that stored as a feature of that item. US intensity effects can be explained by conditional analgesic processes. 8:25-8:40 (2) 8:20-8:35 (7) Wby is There a Set-8ize Effect 10 Visual Search? JOHN PALMER Behavioral, Vocal, and Hormonal ReiatioDlhipalo Five Captive & CYNTHIA T. AMES, University of Washington-Set-size effects in AIrk:ao Elephants. MELISSA R. SHYAN, ROBERT H. I. DALE, visual search may be due to attentional effects on perception or to ef Butler University, JOHN K. CRITSER, ESTHER NOILES, Metlwdist fects of sensation, decision, or memory . In simple search tasks such Hospilal ofIndiana , Inc. , & DEBORAH J. OLSON, Indionopolis Zoo-- as detecting a long line among short lines, we have previously found Five captive African elephants were studied to compare behaviors and the modest set-size effect predicted by an attentional effect on only de vocal activities to their estrous cycles. Their behaviors were observed cision and not perception. To pursue this result, we investigated several over a two-year period. Vocalizations were recorded and analyzed to more complex search tasks that have larger set-size effects. determine whether an estrogen-related voca1ization used to attract mates 8:45-8:55 (3) (the "estrous song") was being produced. These data are being ana The Stroop Effect: Iocorporatiog Noocolor Words Ioto the lyzed to determine whether these vocalizations and any associated be RespIMR Set. COLIN M. MACLEOD & STUART A. GRANT, Uniwr haviors could be used as noninvasive behavioral markers for peak fer sity of Toronto. Scarborough Campw-In the classic Stroop effect, in tility during the estrous cycles. compatible color words interfere with naming the ink colors in which 8:40-8:55 (8) they are printed (e.g., the word "red" in green ink. say "green"). Or Modulation 01 Ethanol Relolorcement by Pavloviao Coodldoaed dinarily, noncolor words (e.g .• "horse") interfere virtually not at all Tolerance. CHRISTOPHER L. CUNNINGHAM, Oregon Health with color naming because they are not possible responses. We describe Sciences University-Previous studies show that stimuli correlated with a situation wherein noncolor words interfere as much as color words, ethanol injections come to elicit a hyperthermic response which inter we investigate its parameters, and we consider its implications for in feres with the drug's unconditioned hypothermic effect, thereby produc terpretations of the Stroop effect. ing tolerance. The present study shows development of a similar con 9:00-9:10 (4) ditioned response to a stimulus that signals aVailability of orally Beyond Reaction Time: A Forced-Cholce Procedure lor Study self-administered ethanol in a discrete-tria1 operant task. Responding ing Visual Search. JAMES L. ZACKS & ROSE T. ZACKS, Michi for ethanol was greater on signaled than on unsignaled tria1s, suggest gan State University-Investigation of visual search processes has relied ing conditioned tolerance to ethanol's thermal effect altered its reinforcing heavily on RT procedures. We will describe a forced-choice procedure efficacy. with limited duration displays that can be used to provide converging 9:00-9: 15 (9) evidence of a serial search process, and estimates of the parameters that Flavor-Drug As8cM:latioDS Produced by PosItively ReloIon:iDa characterize the search process. The estimates are based on both the Drup: A Dose-RespoDSe Analysis. LINDA A. PARKER, Wilfrid changes in the threshold duration as a function of array size and on the Laurier University-Drugs which are positively reinforcing in a place' shape of the psychometric functions relating accuracy to presentation conditioning or self-administration paradigm have also been shown to duration. have aversive properties at equiva1ent doses in a conditioned taste aversion 9:15-9:35 (5) (CTA) paradigm. A series of experiments with Sprague-Dawley rats Asymmetries 10 Visual Search lor Conjunctively Deftned Targets. investigated this paradox using the taste reactivity and CT A paradigms. ASHER COHEN, Indiana University-This study demonstrates asym Taste reactivity responses elicited by sucrose, which was paired on five metry in visual search for conjunctively defmed targets despite sym occasions with various doses of o-amphetamine (0, 2, 3, S, 10 mglltg, metry in search between the simple features that define these targets. i.p.), nicotine (0, .4, .8. 1.2, 2.0 mgl1tg, s.c.), or morphine (0, 2, 8, For instance, search for a red X among red Os and blue Xs is faster 20,80 mgIlr;p. i.p.), were measured. The only doses of these drugs which under some circumstances than search for a red 0 among red Xs effectively conditioned rejection taste reactivity responses were those and blue Os, while there is symmetry in search between Xs and Os. that are higher than those capable of producing a place preference. How An explanation of this phenomenon is proposed and supported by several ever, lower doses, which were within the range capable of producing experiments. a place preference, produced aCTA.
473 Friday Morning Papers 10-21
9:20-9:30 (10) tion hypotheses, but fmdings are in line with the T -trial programming Spedficlfy or Pavlovian Drug Conditioning Effects 00 Drug-Taking hypothesis. Bebavior. MARVIN D. KRANK, Mount Allison University-Pavlov 9:30-9:45 (16) ian theories of addiction postulate that conditional stimuli (CSs) for drug Pavlovian Principles in Human Parallel Distributed Processing. effects exert motivational control over drug-taking behavior. I tested ROBERT FRANK WEISS, NICHOLAS B. McDONALD, CHRISTINE the nature and specificity of drug CS effects on operant behavior for LITTLE, & ROBERT D. SHULL, University ofOklahoma-The least ethanol, using the transfer of training design. The results indicate that mean-squares solution for parallel distributed processing is mathematically a discrete cue light CS trained with response-independent ethanol has identical to the Rescorla-Wagner equation for Pavlovian conditioning. an incentive effect on ethanol-reinforced behavior. The incentive ef This unexpected confluence of cognition and conditioning is sparking fects influence both general appetitive behavior (e.g., approach) and interest among cognitive psychologists in the strange new laws of Pav reinforcer specific choice behavior. lovian conditioning successfully predicted by Rescorla-Wagner theory. 9:35-9:45 (11) In experiments that make full use of the techniques of modem condi The Comparator Ratio: Stimulus Control or the Numerator. tioning research, we demonstrate a human analog of Pavlovian blocking. T. JAMES MATTHEWS, ORN BRAGASON, & BENETTA YEE, 9:50-10:05 (17) New York University-Pigeons were tested on two random time sched Memory ror Fractal Textures. KEITH CLAYTON & DAVID L. ules of food reinforcement in which a response key was transilluminated GILDEN, Vanderbilt University-Ecological surveys ofnatura1 textures throughout the trial by yellow light which increased at either a fixed that were measured on the fractal dimension reveal a greater frequency rate or at a variable rate that was proportional to the trial duration. On of textures with beta near 2. Recent psychophysical research shows max the fixed-rate procedure, subjects initiated autoshaped pecking at a fixed imal sensitivity to fractal contours with beta near 2. In the present study, time before the reinforcer. On the longer trials of the variable-rate proce images varying in fractal dimension (five levels of beta from .8 to 4) dure, subjects responded at the midpoint of the trial. were presented to subjects followed by a forced-choice recognition test. Memory was poorest for images with beta near 2. HUMAN LEARNINGIMEMORY I 10:10-10:20 (18) PaCUlC West DE, Friday Morning, 8:00-10:50 Prospective Memory: The Effects or Cue Familiarity and Cue Dis tinctiveness. MARK A. McDANIEL, Purdue University, & GILLES O. Chaired by Denise C. Park, University of Georgia EINSTEIN, Furman University-Apparent similarities between cued retrospective memory tasks and event-based prospective memory tasks 8:00-8:15 (12) prompted the hypothesis that the familiarity and the distinctiveness of The Effect or Cue and Retrieval Strategy Compatibility in Part-List the target event (cue) would influence prospective memory performance. Cuing Inhibition. DAVID R. BASDEN & BARBARA H. BASDEN, The data confIrmed that unfamiliar target events and target events that California State University, Fresno (read by Barbara H. Basden) were distinctive relative to the local context benefited prospective mem Providing some of the study-list members at recall is inhibitory. Perhaps ory. Also, prospective memory performance correlated with performance part-list cues alter memory traces in a way that favors retrieval of cue on an indirect retrospective task but did not generally correlate with over noncue words. Alternatively, the presence of cues may encourage recall and recognition performance. Several frameworks of retrospec the use of a less effective retrieval strategy. Several experiments are tive memory are applied to the present effects. described in which inhibition was reduced or eliminated when subjects 10:25-10:45 (19) organized the study list in a way that was compatible with the presence Learned Covariation: COII8Cious or Unconscious Representation? of cues at recall. DON E. DULANY & RUSSELL POLDRACK, University ofRlinois 8:20-8:40 (13) After viewing pictures of persons with physical traits revealed and psy Adaptation-Level Coding and Isolation Effects in Serial Learn chological traits described, subjects judged whether other persons with ing. G. J. JOHNSON, University of British Columbia (sponsored by the same physical traits would have the associated psychological traits. F. P. Valle)-The distinctiveness model of serial learning is extended We were able to reproduce Lewicki's (JEP: LMC, 1986) two positive to the von Restorff effect. Predictions are generated with respect to rela effects only under these conditions: on latency of judgment by introducing tive performance on an isolated item, presence or absence of a spread a sequence bias consistent with reported sequence constraints and on of effect, number of isolates, and serial position of an isolate. Other correctness of judgment by introducing awareness of covariation with derivations concern the frequency with which the serial position of an instructional hints. isolated item is involved in intra1ist generalization errors. 8:45-9:05 (14) RECOGNTInONffiECALLI Memory ror Bizarre Imagery: A Multinomial Modeling Analysis. Pacific West FG, Friday Morning, 8:00-10:35 DAVID M. RIEFER, California State University, San Bernardino, & JEFFREY N. ROUDER, University ofCalifornia, Irvine-A new multi Chaired by Richard Schweicken, Purdue University nomial model is presented that separately measures storage and retrieval factors in memory by examining free followed by cued recall of paired 8:00-8:20 (20) associates. The model is used to explore the storage-retrieval basis be The List-Strength Effect (LSE) and Global Matching Models. hind the bizarreness effect, the finding that bizarre imagery is remem BENNET B. MURDOCK, University of Toronto-It has been claimed bered better than common imagery. The model suggests that bizarre that the LSE is counter to all the current global matching models (SAM, images are better retrieved but not better stored than common images, MINERVA, CHARM, TODAM, and matrix models). Whether or not a fmding that is consistent with current theories. this is true depends on the underlying assumptions. If memory is con 9:10-9:25 (15) tinuous, then most, if not all, of the current global matching models The Hidden Power or Tests: Control or the Ultimate Acquisition would seem to predict not only the absence of an LSE, but also the Level. CHIZUKO IZAWA, Tulane University-Two experiments, utiliz presence of a list-length effect in recognition memory. ing constant study time, unveiled previously unknown facilitating ef 8:25-8:45 (21) fects of unreinforced test (T) trials administered successively with the Interference and Forgetting in Recognition Memory. RICHARD M. traditional list-repetition (as opposed to the novel item-repetition) tech SHIFFR1N & KIM MARINELLI, Indiana University-The causes of nique in cued recall; total T -time was held constant. Rapid successive interference and forgetting are examined in a study presenting a mixed Ts uncovered long-term memory retrieval facilitation effects. Data did list of exemplars (but not prototypes) from 16 semantic categories (e. g. , not support the total T -time, the T -frequency, or the T -exposure dura- words related to "butterfly") and 8 orthographic/phonemic categories
474 Papers 22-33 Friday Morning
(e.g., words sharing the vowel sound, number of letters, and one con varied the correlational structure organizing possible events and the par sonant with "cat"). Recognition tests were carried out immediately and ticular attributes that correlated. Subjects were tested for knowledge of after one week for exemplars, prototypes, and other types of distrac individual correlations among event attributes. Individual correlations tors. The results rule out a wide class of models. were learned more easily when part of a coherent system of correlations. 8:50-9:00 (22) 8:20-8:35 (28) Reorganized Memory Lists, Expanded Memory Lists, and Retrieval. The Role of Attcntional Resources In Suggestibility and Source DAVID BURROWS, Skidmore College-Subjects determined whether Monitoring. MARIA S. ZARAGOZA & SEAN LANE, Kent State test words belong to expanded memory lists. Test cycles alternated with University-Subjects who are exposed to misleading suggestions about presentation cycles. In the standard condition, presentation cycles con an event they have witnessed sometimes come to believe they remem sisted of 12 words to be added to the list. In the reorganization condi ber seeing items that were merely suggested to them, an error we refer tion, presentation cycles consisted of 12 words, 9 to be added to the to as source misattribution errors. An experiment was conducted to test list, and 3 to be deleted. Accuracy declined gradually with list length, the hypothesis that source misattribution errors are a function of the and was equivalent for the two conditions. Latencies were 670 msec attentional resources available at test. The results show that retrieval greater for reorganized lists. conditions play an important role in the prevalence of source mis 9:05-9:15 (23) attribution errors. A Context-Dependent Stimulus SuffIX Effect. AIMEE M. SUR 8:40-8:55 (29) PRENANT, IAN NEATH, & ROBERT G. CROWDER, Yale Univer Topography Facilitates Symptom-FaUlt Learning during sity (read by Robert G. Crowder)-Precategorical acoustic store (PAS) Troubleshooting by Restricting Search. LAWRENCE W. BARSALOU, explains stimulus suffix effects by assuming that the suffix enters into University of Chicago, & CHRISTOPHER R. HALE, Georgia Insti an acoustic store on physical criteria and interferes with earlier list items tute of Technology-Knowing a system's topography during trouble contained therein. One line of evidence incompatible with this explana shooting facilitates identifying the faults that produce symptoms. Of in tion, a context-dependent suffix effect, has been demonstrated only once terest is whether such effects reflect intelligent guessing versus learning. (Ayres, Jonides, Reitman, Egan, & Howard, 1979). The two experi We report results from four experiments showing that topography facili ments reported here replicate and extend the finding that a clearly tates both. Unexpectedly, topography does not facilitate learning through nonacoustic factor can determine whether the suffix effect obtains. mnemonic elaboration but through restricting search. When subjects who 9:20.9:40 (24) do not know topography can restrict search similarly, they learn at the Understanding Human Memory. MICHAEL S. HUMPHREYS, same rate. The max rule, the Luce rule, and search probability consti JANET WILES, & SIMON DENNIS, University ofQueensland-Marr's tute possible mechanisms. (1980) framework for understanding an information processing device 9:00-9: 15 (30) is applied to cued recall with an extralist associate. Our computational Primary and Secondary GeneraHzatioo in Categorization and Typi theory specifies that pairwise information is used and that the goal is cality of WeD-Defined Concepts. ANDRE V ANDIERENDONCK, to find the intersection of the items related to the cue and the items in University of Ghent-According to exemplar models of categorization, the list. We then show how this characterization of the computation con primary generalization is the only principle needed to explain categori strains theories at the next level, which may also be constrained by the zation performance and typicality. Other views support the idea that hardware implementation level. in addition to primary generalization, subjects use some form of sec 9:45-10:05 (25) ondary generalization. Experimental results are discussed which test the Implicit Memory: Effects of Associative Set Size, Connectivity, hypothesis that a decrease in interexemplar similarity results in increased and Network Intersection. DOUGLAS L. NELSON, NANCY GEE, opportunities for secondary generalization changing the form of the & THOMAS SCHREIBER, University of South Florida-By collect categorization and typicality gradients. ing discrete word associations for a very large sample of words, n x n 9:20.9:35 (31) associative matrices were constructed for over 1,500 items. Associa Inferring Sex and Age from Handwriting. WILLIAM N. HAYES, tive structures vary naturally in size and connectivity, and the networks Albion College-Handwriting samples were collected from 30 people for any two words intersect to different degrees. In several experiments, aged 60 + (15 female and 15 male) and 30 college students (15 female subjects studied words whose associates varied in set size and connec and 15 male). Another group of students, haIffemale, half male, rated tivity, with recall cued by related words sharing differing numbers of the samples on two categories, sex and age. Accuracy for the sex rat intersecting associates. Each variable independently affected recall. ings was well above chance, and for the age ratings was slightly above 10:10.10:30 (26) chance. Handwriting apparently conveys information about sex and, to An Acquisition, Loss, and Decision Model of Retrieval from SfM. a lesser extent, age. BARBARA ANNE DOSHER, Columbia University-A forgetting model 9:40-10:00 (32) of STM, including acquisition, loss, and decision components, was ap ALCOVE: An Exemplar-Based Connectionist Model of Category plied to reaction time and error data in speeded item recognition. Each Learning. JOHN K. KRUSCHKE & ROBERT M. NOSOFSKY, In serial position of STM lists (lengths 2-5) of digits, discriminable and diana University (read by Robert M. Nosofsky)-Kruscbke's (1990) confusable letters, trigram syllables, and novel pictographs were tested. ALCOVE model implements key features of the exemplar-based gener The model is applied to differences in recency proflies across materials, alized context model (GCM) (Nosofsky, 1986) within a multilayered allowing the estimation of acquisition, interference-decay loss, immediate connectionist network. The model provides excellent quantitative fits match, and encoding contributions. to category learning data and accounts for fundamental phenomena such as prototype effects, effects of specific exemplars, sensitivity to cor CATEGORIZA TION related dimensions, attentionallearning, and base-rate neglect. Unlike Golden Gate A, Friday Morning, 8:00-10:05 standard backpropagation models, ALCOVE does not suffer catastrophic interference, and learns as people do. Chaired by Roger Chaffin, Trenton State College INFORMATION PROCESSING I 8:00-8:15 (27) Golden Gate B, Friday Morning, 8:00-9:40 Learning Event Categories: Effects of Correlational Structure. DORRIT BILLMAN & ALAN KERSTEN, Georgia Institute ofTech Chaired by Alinda Friedman, University of Albena nology-Event categories are central to cognition and language but have been little studied relative to object categories. We analyzed verb and 8:00-8:10 (33) syntactic features to design the event attributes for our experiments. Sub Visual Stimulus Intemity Does Not InJIueDce Respoo!iI! Force. JEFF jects watched animated events showing two characters interacting. We MILLER, University of California, San Diego, ROLF ULRICH, Univer-
475 Friday Morning Papers 34-44
sity of Tubingen, & KERRY PFAFF, University of California, San increments favored accompaniment. With longer increments, fast Diego-The forcefulness of a keypress response did not depend on stimu rise/falls favored replacement; slower ones cued transformation. lus intensity in experiments using simple, go/oo-go, and choice RT tasks 10:05-10:20 (39) with visual stimuli. Subjects responded with index finger flexions, and Prosodic Organization in Song Composition and Performance. force was measured. Task-irrelevant increases in stimulus intensity MICHAEL H. KELLY, University of Pennsylvania, & CAROLINE reduced RT by approximately 100 msec but did not affect response force. PALMER, Ohio Swe University (read by Caroline Palmer)-We exam This rmding supports models with discrete stimulus coding and is con ined the alignment of phrasal stress patterns with musical meter in song. sistent with the "constant stage output" assumption of the Additive Factor Analyses of song compositions showed adjective-noun phrases aligned Method. with musical meter, as predicted by clitic rules of stress assignment. 8: 15-8:30 (34) VocaIists' performances of songs containing adjective-noun phrases in Mental Chronometry: Further Beyond Reaction TIme. DAVID A. dicated that sung durations of adjectives and nouns varied according BALOTA & RICHARD A. ABRAMS, Washington University-Last to principles of clitic group formation. Prosodic stress and musical meter year, we presented evidence that factors (word frequency and memory had independent effects on sung durations, suggesting separate but similar set size) that affect onset latencies in two classic response-latency para organizational principles. digms (lexical decision and memory scanning) also influence response 10:25-10:40 (40) dynamics after response initiation. This year, we provide evidence General Principles of Musical Expectancy in the Implication regarding word-frequency effects in a manual and vocal lexical deci Realization Model. CAROL L. KRUMHANSL & E. GLENN sion task. These data eliminate a response-completion account of our SCHELLENBERG, Carnell University-The implication-realization previous rmdings and extend these findings to speech production. model (Narmour, The analysis and cognition of basic melodic struc 8:35-8:50 (35) tures) specifies basic perceptual principles governing expectancies in Effects of Response Uncertainty on Semantic Retrieval. JAMES M. music. The first test of the model found strong support for the model CLARK, University of Winnipeg-Selected RT effects of response un using Western folk tunes (psychonomic meeting, 1990). Recent experi certainty (Le., number of alternative responses) are reported for such ments have found additional support using atonal music and Chinese semantic retrieval tasks as picture naming, free association, synonym folk tunes; effects of musical training and acculturation are minimal. production, and generation of translation equivalents. Reported rmd Affective responses to melodic patterns have also been explored within ings include: increased RTs with response uncertainty, independence the context of the model. from familiarity and other item attributes, and unique uncertainty ef 10:45-11:00 (41) fects for different operations (e.g. , generation of instance vs. super Neural Seif-Organization of Auditory Categories. JAMSHED J. ordinate names for same pictures). Possible explanations are described, BHARUCHA & W. EINAR MENCL, Dartmouth College-Auditory including mutual inhibition among competing alternatives. categories can be learned passively by neural self-organizing mecha 8:55-9:10 (36) nisms. A network presented with a tonotopic representation of tone spec Independence of Form and Color in Perception and Memory. tra discovers similarities between tones and learns to classify them by JOHN CERASO, Rutgers at Newark-Whether presented as units or forming specialized detectors. The learned categories include octaves as separate entities, forms and colors tend to be independently perceived and timbres. Following learning, the network generalizes its categori and independently recalled. Yet, on other measures, there are differ zation to degraded spectra that are missing the very cues (e.g. , octave ences between unitary and separate stimuli: Units show better associa harmonics) that led to the learning of those categories. ti ve recall; perceptual identification of unitary form and color is not dis 11:05-11:25 (42) rupted by divided attention, whereas separate form and color is disrupted. Perceiving Implied Harmony. MARl RIESS JONES, SUSAN These rmdings suggest that the independence measure is not diagnostic HOLLERAN, & DAVID BUTLER, Ohio State University-Listeners of unitariness. (musically trained and untrained) heard standard-comparison pairs of 9:15-9:35 (37) novel tonal melodies. Melodies were accompanied by: (1) no chords, A Backwardly Muked Stimulus Is Not Subliminal. JOHN THEIOS (2) key-consistent chords, or (3) key-inconsistent chords in certain mea & GLORIA MARMOLEJO, University of WISconsin-Madison -Critics sures. A target tone occurred in one unaccompanied measure of the com of unconscious priming have erroneously assumed that stimuli back parison; its pitch could change (relative to the standard) to become con wardly masked by a pattern are subliminal. In fact, masked stimuli are sistent or inconsistent with the chord implied locally. All listeners were superthreshold and activate the visual system all the way up to the seman less likely to notice target changes consistent with local implied harmony. tic level. The mask has been chosen so that it captures the observer's 11:30-11:45 (43) attention, diverting it away from the stimulus. Thus, the observer has Melody Is Separable from Rhythm in Music Discrimination: Evi no conscious experience of the stimulus, but it does automatically acti dence from Neuropsychology. ISABELLE PERETZ, University of vate the perceptual and semantic systems enough to associatively prime Montreal, & IffiGINE KOLINSKY, Free University of Brussels-The related stimuli. study of a brain-damaged patient who suffered from a severe distur bance in the processing of pitch patterns without accompanying disorder in the processing of temporal patterns will be presented. This selective AUDITORY PERCEPTION deficit was (a) replicated across different sets of material, (b) supported Embarcadero AB, Friday Morning, 9:50-U:15 by a reversed association, and (c) maintained in conditions that promote integration in a normal brain. The results argue against the view that Chaired by Cynthia H. Null, NASA-Ames Research Center melody and rhythm are treated as a unique dimension. 11:50-12:10 (44) 9:50-10:00 (38) Selective Attention in Auditory Detection. ERVIN HAFTER, Auditory Intensity Changes Can Cue Perception of Transforma KOUROSH SABERI, LYNE PLAMONDON, & ERIC JENSEN, tion, Accompaniment, or Replacement. ALBERT S. BREGMAN, University of California, Berkeley-In a masking experiment, listeners McGill University, & LUC ROUSSEAU, Laval University-Amplitude detected tones, sets of tones, or harmonic complexes, whose frequen change can cue the number of sounds in a mixture. When part of a spec cies varied randomly from trial to trial. Uncertainty was reduced by trum increases in intensity, listeners hear either that a first sound has various stimulus cues, each directed at a different level of auditory "transformed" in intensity or that a second sound has begun, and either processing, levels that ranged from auditory fIlters to pitch to musical "accompanies" or "replaces" it. We presented a pure tone which under relations to auditory memory. Results show that auditory attention could went an intensity increment, then returned to the lower value. Shorter be focused downward in the processing chain but not upward.
476 Papers 45-56 Friday Morning
3-DIMOVEMENT PERCEPTION I lish the appropriate optical cues specifying time-to-passage for non Embarcadero CD, Friday Morning, 10:00-12:35 collision cases and present data on observers' relative and absolute pas sage judgments. Chaired by Philip J. Kellman, Swarthmore College 12:00-12:10 (51) Scene Perception UDder Dynamic Occlusion. PETER M. VISHTON, 10:00-10:15 (45) PHILIP J. KELLMAN, & THOMAS F. SHIPLEY, Swarthmore Col Perception of Relative Depth from Translation and Rotation. lege (read by Philip J. Kellman)-Scene perception under dynamic oc MYRON L. BRAUNSTEIN, JEFFREY C. LITER, & JAMES S. TIT clusion was studied in two experiments. Subjects viewed real scenes TLE, University of California, Irvine-Random-
477 Friday Morning Papers 57-67
weighting the most recent events more heavily. Contingencies of rein to be judged. Group differences emerged with regard to the shapes of forcement should be maximally effective if they match the animal's mem the serial position functions, with these curves flattening less for the ory. I measure the effectiveness of contingencies with differing win ordinal condition as distance level increased. Implications for the posi dows, and [md optimal control for windows with half-lives of three tional discrirninability model will be discussed. responses. This resonance identifies the animal's memory window, and 10:30-10:40 (62) tells us how to optimize reinforcement. Speech-Controlled Motor Behavior and Problem Solutions: What 12:10-12:20 (57) Is Relevant? ELIZABETH F. SHIPLEY, University ofPennsylvania Reality Versus the Statistical Rat: Empirical Modeling of Object In a follow-the-leader game, 3-year-olds produced either one of two Investigation. MICHAEL J. RENNER & CHARLES P. SELTZER, incompatible solutions (match color or block number) or repetitive motor Memphis Swe University-Over four nights, 12 male rats were individu behavior (play all blocks). Models' verbalizations relevant to the solu ally videotaped in a large arena containing objects; component behaviors tions ("blue, blue," "one, two") decreased repetitive motor behavior of object investigation 0,125 bouts) were coded. Each rat's behavior but did not determine specific solutions. Neither nonsense nor irrele sequences were condensed to a behavioral "grammar," a template vant familiar words facilitated solutions. These results refine prior [md describing bouts via individual transitions between contiguous acts. Many ings on the irrelevance of semantic factors in speech controlled motor bouts allowed by the grammars, however, did not actua1ly occur. Each behavior. real rat's behavior was compared to its grammar-generated "statistical 10:45-11:00 (63) rat. " Compliance and omission patterns illuminate the organization of The Conceptual Organization of Mental Verbs. PAULA J. investigatory behavior. SCHWANENFLUGEL, University of Georgia, WILLIAM V. FABRICIUS, Arizona State University, KELLY BIGLER, University LETIERIWORD PROCESSING I ofMichigan, & JOYCE ALEXANDER, University of Georgia-Major Pacific West FG, Friday Morning, 10:45-12:10 organizational features underlying naive adults' concepts of mental verbs were examined. The 30 mental verbs used represented early acquired Chaired by Nancy B. Marshall, University of Alabama at Birmingham verbs designated by experts as ways to know or come to know some thing. Subjects rated pairs of verbs in tenns of the similarity of the mental 10:45-11:05 (58) processes involved. Multidimensional scaling, additive similarity trees, Interpretations of the Frequency Effect in Word Recognition. and Pathfinder network techniques yielded three organizing dimensions, KENNETH I. FORSTER, University ofArizona-Word-frequency ef three major verb clusters, and three central mental verbs ("think," "ex fects in tasks such as lexical decision place powerful constraints on models amine," and "observe"). of word recognition, provided they operate during lexical access and 11:05-11:20 (64) not at some postaccess stage. Recent work has raised doubts about this Individual Differences in Working Memory. RANDALL W. ENGLE, assumption. This work is briefly reviewed, and new studies are described JUDY CANTOR, & JULIE CARULLO, University ofSouth Cmolina dealing with familiarity, repetition effects, and tasks that show lexical Does the correlation between tests of working memory and compre effects without frequency effects. hension occur because of differences in processing, not storage? A 11:10-11:30 (59) moving-window version of the reading span task was used to measure Naming Reaction Times to Braille Words: Effects of Lexicality, the time to read the words in the sentence and the time spent processing Frequency, and Uniqueness Point. PAUL BERTELSON, PHILIPPE the to-be-remembered word. When the time to process the sentence and MOUSTY, RONALD PEEREMAN, & MONIQUE RADEAU, Univer the span words was partialed out, the correlation between reading span site Libre de Bruxelles-The latencies of naming responses to high- and and comprehension was undiminished. low-frequency words and to pseudowords presented in braille were mea 11:25-11:40 (65) sured in blind readers. The effects of lexicality and of frequency showed Learning and Performing F1ueut Cognitive Sequences. RICHARD A. the pattern generally obtained with printed material. The location of the CARLSON, Pennsylvania Swe University-Fluent performance ofrou uniqueness point influenced RTs to high-frequency words, not to low tine cognitive activities such as skilled problem solving requires rapid frequency ones. The results are consistent with a two-process interpre selection and integration of procedural subsequences. Recent research tation of the pronunciation of braille words. has examined roles of acquisition context, consistency, goal structures, 11:35-12:05 (60) and metacognition (awareness of repeated sequences) in the develop Lexical Ambiguity and Eye Fixations: Frequency and Familiar ment of fluent cognitive sequences. Learning data from problem solving, ity Effects. KEITH RAYNER, JEREMY M. PACHT, SARA C. attention switching, and motor sequence paradigms are reviewed. Two SERENO, University of Massachusetts, & SUSAN A. DUFFY, Am learning processes-learning by restructuring, and explicit sequence learn herst College-Readers look longer at biased ambiguous words when ing-are proposed to integrate findings from these diverse paradigms. context instantiates the subordinate meaning compared to unambiguous 11:45-12:00 (66) control words. In the present studies, we examined fixation times on Aesthetics, Laterality and Visual Scanning. JOHN P. McLAUGH ambiguous words and control words to determine if the effect is due LIN & ANDREW M. MEAD, University of Delaware-Lateral dif to subjects 0) treating the subordinate sense of a biased word as a low ferences contribute to evaluations of art. Paintings preferred by dex frequency word or (2) lacking familiarity with the subordinate sense. tra1s elicited fixations farther to the right in the picture space, indicating Implications for models of lexical ambiguity resolution will be discussed. that the amount of a picture in the right visual field determines prefer ence. For paintings producing the strongest effect, mirror reversals pro COGNITION duced differences in the frequency, distance, and direction of saccades. Golden Gate A, Friday Morning, 10:15-12:30 This sensitivity to compositional change reveals a type of intersubject consistency in scanning patterns. Sinistrals, however, showed these ef Chaired by Margaret Jean Intons-Peterson, Indiana University fects less consistently. 12:05-12:25 (67) 10:15-10:25 (61) Incubation Effects in a Multisolution Anagram Task. CATHERINE G. Serial Position Effects in Numerical Comparisons: Magnitude PENNEY, Memorial University of Newfoundland-Four experiments Versus Order Judgments. DANIEL B. BERCH & ANGELA investigated incubation effects in a multisolution anagram task in which BIRKHEAD-FLIGHT, University of Cincinnati-Subjects were given subjects produced five-letter words from the letters in a ten-letter starter either a standard, numerical magnitude comparison task or an "ordi word. Subjects worked as long as they were being productive before nal" numerical comparison task involving the same digits pairs, but for taking a break. After the break, subjects produced new solutions not which the correct numerical (left-to-right) order of the two digits had generated before the break. The number of new solutions depended on
478 Papers 68-79 Friday Morning/Friday Afternoon
the length of the break and whether subjects were exposed to solution actions. However, children frequently disagreed with their parents as words during the break. to which emotion they had experienced, especially when parents reported anger and fear reactions. Parents' skill in predicting accurately their COMPREHENSION/REASONING children's emotional response was directly related to the amount of con Golden Gate B, Friday Morning, 9:50-12:20 tent overlap found in the two accounts of the same event. 11:55-12:15 (74) Chaired by Anhur C. Graesser, Memphis State University Concepts in Situated Reasoning. JAMES G. GREENO & JOYCE L. MOORE, Stanford University and Institute for Research on Leaming 9:50-10:05 (68) We view reasoning as interactive activities of agents in situations that Hypothesis-Testing Goals and Strategies: Two Rules Are Better generate information. Concepts, in this view, correspond to bundles of than One. CHARLES M. WHARTON, THOMAS D. WICKENS, & constraints on physical and social activities. Reasoning in accord with PATRICIA W. CHENG, UCLA (read by Thomas D. Wickens)-An a concept can be analyzed to show which constituent constraints result effective strategy to recover from an erroneous inference in hypothesis from affordances of situations and which result from agent(s)' abilities. testing is to examine both what is the correct hypothesis and what is These analyses support attributions of understanding to the agent(s). We not. The efficiency of dual hypotheses relative to a single hypothesis present an analysis and data involving the concept linear function. has been attributed both to task difficulty and to problem representa tion. Subjects in three experiments performed isomorphs of Wason's HUMAN PERFORMANCE (1960) 2-4-6 task with instructions for single-rule or dual-rule testing. Embarcadero AB, Friday Afternoon, 1:00-4:00 Results and models support problem-representation theories. 10:10-10:25 (69) Chaired by Stephen R. Ellis, NASA-Ames Research Center H Saddam Is Hitler Then Who Is George Bush? Coherence in Ana logical Mapping. BARBARA A. SPELLMAN & KEITH J. 1:00-1:10 (75) HOLYOAK, UCLA (read by Keith J. Holyoak)-Naturalistic analogi No Influence of Music on Maximal Performance and Fatigue. cal mapping between the Persian Gulf War and World WarD revealed JOHAN HUETING, ERIC SOETENS, & ILSE CAUWENBERGHS, that the analogy is bistable-an' 'analogical Necker cube." When Sad University of Brussels-The question of whether music enhances per dam was presumed to map to Hitler, subjects mapped Bush and the United formance, mood, sales, and so forth has long been a much-debated topic. States either to Churchill and England or to Roosevelt and the United Straightforward experiments with clear data are scarce, however. The States of World War D; in addition, Saudi Arabia mapped to England influence of music was investigated during an exhaustive task on a bi only if the United States did not. These pattems, simulated using Holyoak cycle ergometer. Music was presented with a Walkman, and effects were and Thagard's (1989) ACME model, reflect multiple constraints that compared to a control session. No differences were found, neither be determine analogical coherence. tween maximal performances nor between continuously self-rated feel 10:30-10:45 (70) ings of fatigue and recovery. Apart from the music, striking differences Validation of Motive Bridging Inferences in Brief Texts. MUR were found between the course of fatigue and of lactic acid curves. RAY SINGER, University of Manitoba-The validation model states 1:15-1:30 (76) that the motive bridging inference that links the sentences Joan wanted Verbal Protocols and Complicated Performance. PAT-ANTHONY to rent a modem house, She bought the newspaper, must be validated FEDERICO, Navy Personnel Research and Development Center-In against pertinent knowledge. Subjects read motive and control temporal order to determine the impact of different verbal protocol procedures sequences. As predicted, answer times for subsequent questions about on complex task performance (i.e., tactical decision making), subjects pertinent knowledge, such as Do newspapers advenise houses?, were were randomly assigned to one of three conditions: experimental con faster in the motive condition. This outcome obtained both for adjacent current or retrospective protocols or control no protocol, and administered and separated goal and action sentences. 256 simulated scenarios. Obtaining verbal protocols from the subjects 10:50-11:05 (71) did not have any signifIcant impact upon their complex task performance. The Role of Pretests in PredictioDi ofPedonnance on Text. RUTH H. 1:35-1:50 (77) MAKI, Nonh Dakota State University, & MATT SERRA, Purdue Uni Using Attention Operating Characteristics to Calibrate Inferences versity-Two experiments were conducted in which subjects answered about Performance Operating Characteristics. BARRY H. KANTO multiple-choice pretest questions before predicting performance on W1TZ, Banelle Human Affairs Research Centers, GREG C. ELVERS, multiple-choice tests. Although subjects could predict their performance University ofDayton, & JOHN PALMER, University of Washington with greater -than-chance accuracy, subjects who had pretest questions Performance on dual tasks can be represented as attention operating that were either identical to or similar to final test questions did not predict characteristics (AOCs: same tasks on both axes) or performance oper more accurately than subjects who saw no pretest questions. Apparently, ating characteristics (POCs: different tasks on each axis). There are seri pretests without feedback do not help subjects to predict test performance ous problems in interpreting POCs arising from combining two differ more accurately. ent dependent variables. A new methodology is presented to mitigate 11:10-11:25 (72) such difficulties by testing for consistency between AOCs and POCs. Early Learning ofMatbematical Concepts. PATRICIA BAGGETT, Initial results for two perceptual tasks showed predicted consistency. University of Michigan, & ANDRZEJ EHRENFEUCHT, University 1:55-2:15 (78) of Colorado-Arithmetic is currently taught in an inconsistent fashion. The Ubiquitous Congruity Effects. DANIEL HOLENDER, Free We are developing and testing a curriculum based on the arithmetic of University of Brussels-A general theory based on the concept of con real numbers (rather than positive integers). This approach stresses a gruity is proposed of the effects observed in Stroop-like and priming verbal component of mathematics and the use of manipulatives and cal tasks, and in some serial tasks. Congruity effects arise from the fact culators. It introduces all four arithmetical operations in the fIrst grade. that irrelevant aspects of the situation may be related to the responses Some lessons have been tested in kindergarten, fIrst, and second grade according to the same descriptors as those used for mapping the rele classrooms. vant aspects into the responses. Hence, the locus of the effects is in late 11:30-11:50 (73) decisional processes dealing with conscious mental contents. A Comparison Between Cblldren's and Parents'Memory for Em0- 2:20-2:40 (79) tional Events. NANCY L. STEIN & MARIA LIWAG, University of Dimensional Overlap and Population Stereotype as Joint Predic Chicago-Our study compared the correspondence between parent-child tors of Stimulus-Response Compatibility. LINDSAY M. OLIVER memory for emotional events experienced by children. Parents accurately & SYLVAN KORNBLUM, University of Michigan (read by Sylvan identified events that caused their children to experience emotional re- Kornblum)-Stimulus-response compatibility has recently been attributed
479 Friday Afternoon Papers 80-90
to two separate factors: one factor, at the set level, is based on the similar the distinction that can be made between "perceived" size and "reti ity, or dimensional overlap, between the stimulus and response sets; nal" size. The results suggest that the size-congruency effect in mem the other, at the element level, is based on the pairing, or mapping, ory for visual shape occurs as a result of changes in the perceived size between the individual stimulus and response elements. Independent mea of shapes between the encoding and the testing phases, with little or sures of these two factors, and data showing their joint effects on RT no contribution of retina1 size. in stimulus-response compatibility tasks will be presented. 2:00-2:15 (86) 2:45-3:05 (80) Visual Orientation and Symmetry Detection under Aftine Trans A Cognitive Task Analysis of Intelligent VebiclelHighway Navi formations. WALTER GERBINO, University of Trieste, & LING gation Systems. RALPH NORMAN HABER & L YN R. HABER, ZHANG,ISAS, Trieste-In normal mirror symmetry, the axis is verti University of illinois at Chicago-Intelligent vehiclelhighway systems cal and the virtual lines connecting corresponding points are horizon are becoming available for automobiles. Display of navigational infor tal. We studied affine transformed dot patterns to separate the contri mation is standard, while dynamic systems provide on-line updated route bution of these two properties to symmetry detection. Results indicate guidance alternatives based on traffic density, accidents, and construction, that matching along the horizontal dimension is more important than using information from roadway sensors and probe and other vehicles, the vertical axis. A 3-D frame, either preceding or simultaneous, was processed by a traffic control center and transmitted back to vehicles. presented. A simultaneous frame facilitates symmetry detection, but does A task analysis of the use of such standard and dynamic systems by not eliminate the orientation bias intrinsic to the matching process. drivers is presented, focusing on attentional demands, safety, and driver 2:20-2:40 (87) ability to benefit from the information presented. Prevailing Lightness and Texture Segregation. JACOB BECK & 3:10-3:30 (81) WILLIAM GOODWIN, University of Oregon-Perceived segregation Components of Orientation Ability in Blind Travelers. L YN R. was investigated in displays in which the elements in one quadrant HABER & RALPH NORMAN HABER, University of Illinois at differed from those in other quadrants by interchanging light and dark Chicago-Research publications reveal a variety of definitions of orien areas. The light and dark areas composing an element were equal. Inter tation ability underlying travel by the blind. We present a systematic changing figure-ground lightnesses failed to yield strong texture segre analysis of these components, which include: knowing the locations of gation. This is consistent with previous fmdings that shape differences objects, knowing one's location, knowing how these change with self fail to predict perceived segregation. Perceived segregation appears to motion, knowing stimuli that provide orientation information, know occur in terms of differences in the prevailing lightnesses of the texture ing how to make oriented psychomotor responses, and knowing the se elements. The experimental results will be compared to predictions from quence of route segments and choice points. The analysis is applied to computational models. current research with blind travelers, and extensions to navigation, map 2:45-3:00 (88) reading, and visual guidance by the sighted are examined. An Attribute Intfgndion Model of Image ~ WILLIAM R. 3:35-3:55 (82) UTIAL, ROBB LOVELL, SRlRAM DAYANAND, & TOM SHEP Why Humans Get Hot. DAVID L. GILDEN & KERSTIN MAC HERD, Arizona State University-We present a computational model DONALD, Vanderbilt University-In a variety of tasks ranging from of image analysis which results in the segmentation of objects in a scene. putting golf balls to signal detection, we have shown that performance The significant aspect of this model of this important aspect of human is streaky; success tends to segregate from failure. Why is this? We visual perception is that it is based upon the integration of a number consider three classes of models to account for our data: stochastic fluc of weak and fallible analyzers, each sensitive to a different attribute (e.g., tuations in effort, effort contingent upon prior outcome, and wave modu texture, brightness, color, and depth) rather than on the performance lation of ability. The wave modulation model not only is able to gener of a single algorithm especially tuned to respond to a single attribute. ate streaky sequences, it reproduces the exact statistical structure of By a process of spatial averaging, the variable and inaccurate outputs observed sequences. of the several weak analyzers are pooled to segment objects in a powerful and accurate manner. This procedure is consistent with contemporary PATTERN RECOGNITION views of both neuroanatomy and psychophysical function. Embarcadero CD, Friday Afternoon, 1:00-3:30 3:05-3:25 (89) Multiple Mechanisms for Detection of Renection Symmetry in Chaired by Mario Zanforlin, University of Padova Brief Exposures. C. W. TYLER, L. HARDAGE, & R. MILLER, Smith-Kettlewell Institute-Reflection symmetry (a "pop-out" feature 1:00-1:15 (83) which receives preferential visual processing compared with repetition Infant Perception of Illusory Contours in AppIIftIIt Motioo Displays. or rotation symmetry) was imposed on a field of either static or dynamic ALBERT YONAS, DOUG A. GENTILE, & KIRSTEN CONDRY, noise, with same or opposite luminance contrast in the two half fields. University of Minnesota-In a series of experiments, we have inves Profound differences in symmetry detection with duration, eccentric tigated 4- and 7-month-old infants' abilities to perceive illusory con ity, and the width of a masking strip around the symmetry axis implied tours using kinetic displays. Seven-month-old infants do pay attention the existence of three separate mechanisms mediating static, dynamic, to illusory contours. The data on the 4-rnonth-old infants are not as clear. and opposite contrast symmetry perception. Methodological issues regarding luminance, spatial frequency, and dMAX will be discussed. PSYCHOLINGUISTICS I 1:20-1:35 (84) Pacific West DE, Friday Afternoon, 1:00-2:30 Effects of Hierarchical Structure and Spatial Frequency on GlobalILocal Analysis. MARVIN R. LAMB & E. WILLIAM YUND, Chaired by Marilyn Shatz, University of Michigan VAMC, Martinez-Recent studies have shown that reaction times to lo cal and global targets depend on the activity of several dissociable mech 1:00-1:15 (90) anisms. A series of experiments will be presented examining the extent Action Plaus for the Production of Word Sequences. CHRISTINE A. to which the operation of these different mechanisms depends on the SEVALD & GARY S. DELL, University of lllinois (read by Gary s. hierarchical structure of the stimulus or on its spatial frequency content. Dell)-What units of speech must be mentally represented to correctly 1:40-1:55 (85) specify a sequence of syllables for production? We used a parameter Size Effects in Recognition Memory Are Based on Pen:eived Size. remapping rationale to determine what units are present in the speech BRUCE MILLIKEN & PIERRE JOLICOEUR, University of Water plan. We found that units corresponding to syllables and rimes can be loo (read by Pierre Jolicoeur)-The present research considers the size "reused" during the execution of a sequence, resulting in a repetition congruency effect in recognition memory (Jolicoeur, 1987) in light of benefit. This was not true for either onset or CV units.
480 Papers 91-102 Friday Afternoon
(91) 1:35-1:50 (97) Withdrawn A Simple Dynamic Model for Recurrent Choice. DERICK G. S. 1:2.... 1:35 (92) DAVIS & J. E. R. STADDON, Duke University (read by J. E. R. Plausibility and Argument Structure in Sentence Comprehension. Staddon)-The dominant process theories of recurrent choice are all based SHARI R. SPEER, Nonheastern University, & CHARLES E. on a simplified version of the law of effect which assumes that prefer CLIFTON, JR., University ofMassachusetts at Amherst (read by Charles E. ence now is determined solely by reinforcement conditions in the re Clifton, Jr.)-Subjects read sentences of the form NP V NP PP. Plau cent past. We term these Markovian models. Data from experiments sibility (determined in a previous norming experiment) and argument on successive discrimination reversal show that Markovian models cannot versus adjunct status of the PP were manipulated factorially. Thus, sen be correct. We propose a simple non-Markovian alternative that can tences were either highly plausible or implausible, and PPs were either account for these and other transitional effects as weU as standard match verb arguments or VP adjuncts. Reading times showed an early, short ing relations. lived advantage for arguments over adjuncts and an independent, longer 1:55-2:10 (98) lived processing advantage for plausible over implausible conditions. Dynamics or Temporal Learning. JENNIFER J. HIGA, Duke 1:46-2:00 (93) University-Temporal learning is generally thought to be a slow process A Reaction Time Study of Sentence Processing in Chinese. PING (e.g., standard temporal schedules such as fIXed-interval usually require LI, University of California, San Diego, BRIAN MAcWHINNEY, hundreds of cycles for complete learning). Yet under some conditions Carnegie-Mellon University, & ELIZABETH BATES, University of it is rapid, and dynamic properties can be studied. I will present results California, San Diego (read by Brian MacWhinney)-Chinese is often from experiments investigating such conditions, showing that pigeons described as a language that uses word order and no inflectional devices can quickly learn a series of short interfood intervals (lFIs) and how for marking granunatical relations. However, research by Miao has this ability depends on the number and separation of these IFIs. shown that Chinese speakers do not make consistent use of word-order 2:15-2:30 (99) cues. But then how can Chinese speakers understand sentences if they Inbibitory Learning in the I.eMer Octopus (EWorwdrrllOllrl). W. F. have no cues to rely on? This study measured reaction times in a com ANGERMEIER, University of Cologne-Octopi (Eledone cirrhosa) were puterized picture-choice task to study the utilization of a wider variety trained to approach a model of a crab attached to the outside wall of of cues. Evidence was found for a more subtle set of integrations of their living tanks. After this response was inhibited, a live crab was word~rder patterns with markings for indefiniteness, preverbal objects, presented to the animals to see whether this inhibition would transfer and passivized subjects. The results are discussed within the framework to the natural feeding situation. The results indicate transfer of inhibi of the competition model. tion and long-term memory of the learned inhibition. Findings are dis 2:05-2:25 (94) cussed in terms of evolutionary significance. Verb-Based Lexical Information and Parsing. MICHAEL K. 2:35-2:45 (100) TANENHAUS, JOHN C. TRUESWELL, & CORNELL JULIANO, Dissociation of Contingency and Blocking Effects. BEN A. WIL University ofRochester-We report a series of experiments using several LIAMS, University of California, San Diego-Contingency and block response measures, including self-paced reading, cross-modal naming ing effects were dissociated by changing the identity of the reinforcer. and eye-tracking that support two conclusions: (I) Contrary to recent In the contingency procedure, freely presented reinforcers suppressed claims in the literature, subcategorization information is used immedi operant behavior more when the free reinforcers were the same as the ately in parsing and (2) a complexity difference between two structures contingent reinforcer than when they were different. In the blocking can easily be mistaken for a temporary syntactic garden-path, particularly procedure, changes in the identity of the reinforcer from single-element when eye-movement data are parsed into first and second pass fixations. pretraining to compound-element training did not attenuate the degree of blocking. ANIMAL LEARNINGIBEHAVIOR I PacifIC west FG, Friday Afternoon, 1:06-2:50 REPETITIONfPRIMING EFFECTS I Golden Gate A, Friday Afternoon, 1:06-2:20 Chaired by linda A. Parker, Wilfrid Laurier University Chaired by J. W. Whitlow, Jr., IWtgers University 1:06-1:10 (95) The Role ofVasopressm in Stress-Induced Analgesia. HELEN M. 1:06-1:15 (101) MURPHY & CYRILLA H . WIDEMAN, John Ca"oll University A Computer Simulation of PrimIna Effects In Simple Multiplica Tail-flick latencies were measured in vasopressin-containing (LE) and tion. JAMIE CAMPBELL & PAUL MEAGHER, University of vasopressin-deficient (01) rats under ad-lib and food-restricted condi Saskatchewan-We present a computer model of a network-interference tions. On the first day of the ad-lib phase, LE rats showed a longer latency theory of memory for simple multiplication and addition facts. In the than did DI rats. On all of the other ad-lib days, no significant differ model, representations of numerical facts that are related to the pre ences were noted between the two groups. With the introduction of food sented problem receive continuous activation and compete by way of restriction stress, significant differences were again noted, with LE mutual inhibition. Numerical primes preactivate components of the net animals exhibiting a longer latency. It appears that vasopressin plays work, biasing the tradeoff between excitation and inhibition. Different a modulatory role in the stress phenomenon. degrees of interference or facilitation are observed as a function of prime 1:15-1:30 (96) type and problem difficulty. Effects or Repeated Sessions of Intruder Defeat on HypoaIgesia 1:2.... 1:35 (102) and Freezing. JON L. WILLIAMS, PAUL D. WORLAND, & JONA Controlled and Automatic Retrieval: Influence or Relational and THAN M. JUST, Kenyon College-Rats exposed to one session of defeat DistlDctive Processing. R. REED HUNT, University of Nonh Caro as colony intruders by dominant (alpha) male conspecifics were found lina at Greensboro--Organizatioo and distinctiveness independently affect to show partially opioid-mediated hypoa1gesia and freezing when for cued recall. These influences were analyzed in terms of controUed and malin tested 24 h later with alpha-colony odors present. Recently, we automatic retrieval processes using Jacoby's opposition logic. Subjects found that four sessions of intruder defeat, given on two consecutive sorted or rated the pleasantness of words from natural or ad hoc cate days, resulted in naltrexone-reversible hypoaIgesia during later tests when gories. Tests were either direct cued recall, indirect category produc alpha-colony odors were either present or absent, but freezing only oc tion, or category production with direct instructions to exclude list words. curred when these odors were present. Distinctive prncessing had little effect on automatic generation but sub-
481 Friday Afternoon Papers 103-115
stantially influenced conscious recollection. Relational processing primar takes on a variety of forms not present in modal box models of the 60s ily affected automatic generation. and 70s. The forms include data and control memory in feed-forward (103) and recurrent networks with multiple time constraints and learning rates. (withdrawn) This variety provides a far more robust processing system than box 1:40-1:55 (104) models, and can interpret such memory phenomena as episodic and Perceptual Fluency and Recognition Memory - What's the Con procedural memory, the role of context storage, buffering and work nection? JOAN OAY SNODGRASS, New York University-Two load effects, and stages of skill acquisition. processes-fluency and explicit retrieval-are often implicated in recog 2:40-3:00 (110) nition memory judgments. Because fluency also supports priming in frag Very Short-Term Conceptual Memory. MARY C. POTTER, ment completion, we asked whether level of fragmentation at study ac MIT-Is short-term memory (STM) for conceptual information miss counted for performance on both fragment completion and speeded ing from current models of short-term memory? Several phenomena recognition. Across four experiments, there was no connection between will be discussed that give evidence for fleeting conceptual representations. successful fragment completion and successful recognition memory. I argue that these representations are fundamental to cognitive processing Moderately fragmented pictures at study were best for fragment com and the form that long-term memory takes, but they do not surface readily pletion while complete pictures at study were best for recognition with many of the standard methods for studying and testing STM. memory. 3:05-3:25 (111) 2:00-2:15 (105) Discussion. RICHARD M. SHIFFRIN, Indiana University-In one Mood Congruent Memories: Explic:itIy Remembered, ImpUcitly guise or another, the concept of short-term memory has survived numer Forgotten. SUSAN T. BLAZ, TARA C. RANKIN, & DAVID B. ous empirical and theoretical attacks since the 19608, perhaps for utilitar MITCHELL, Southern Methodist University (read by David B. Mitchell) ian rather than fundamental reasons. We look at the key issues from We investigated the influence of experimentally induced positive mood the perspective of 25 years of research. states on memory for self-referent personality trait adjectives. Across three experiments, we systematically varied the amount of perceptual SPEECH PERCEPTION information available at test. The implicit tests were free association, Embarcadero AB, Friday Afternoon, 4:10-5:30 word-stem completion, and "know" judgments; free recall, cued recall, and recognition served as parallel episodic tests. Both data-driven and CluJired by Beatrice de Gelder, Tilburg University conceptually driven tests revealed mood congruency in episodic mem ory but not in implicit memory. 4:104:25 (112) Segmenting Speech by Recognizing Words. ANNE HENLY & SYMPOSIUM I: HOWARD NUSBAUM, University of Chicago (read by Howard Nus SHORT-TERM MEMORY: WHERE DO WE STAND? baum)-According to most theories, listeners recognize words in the Golden Gate B, Friday Afternoon, 1:00-3:30 order by which they are spoken. Recognizing one word locates auto matically the beginning of the next, eliminating an explicit segmenta CluJired by Robert G. Crowder, Yale University tion mechanism. To test the predictions of segmentation by recogni tion, subjects identified target words in the context of preceding and 1:00-1:20 (106) succeeding monosyllabic words and nonwords. The results indicate that Introductory Remarks. ROBERT O. CROWDER, Yale University listeners can use lexical status in recognition, but it is not done the way The identification of a separate short-term memory system was among most theories would predict. the proudest accomplishments of early work in cognitive psychology. 4:30-4:45 (113) Now, a quarter~ntury afterwards, the evidential basis for such a sys Migration of Speech Units In an musory Word Detection Task. tem is in dispute. New experimental settings and new theoretical con JOSE MORAIS & REoINE KOLINSKY, Universite libre de Bruxelles texts are being suggested for this concept. We should ask: To what ex Evidence concerning the units of speech processing was obtained by tent do these suggestions serve either (1) James's intuitive definition using tasks that do not require attention to these units. We created con or (2) Hebb's reasoning about memory? ditions such that illusory perceptions of a word or a pseudoword target 1:25-1:45 (107) could occur through the extraction and subsequent recombination of parts Activation, Attention, and Short-Tenn Memory. NELSON COWAN, of two dichotically presented items. The aptitude of phonetic features, University of Missouri-I argue that the term "short-term memory" phonemes (consonants and vowels) and syllables to cross from one ear (STM) is vague because it is used to refer to either (1) the set of represen to the other and elicit such illusions was compared. tations from long-term memory currently in a state of heightened acti 4:50-5:05 (114) vation or (2) the focus of attention or content of awareness. A more Making the (Auditory) Scene with Speech. ROBERT E. REMEZ, coherent representation of STM depicts (2) as a subset of (I). Research JENNIFER S. NUTTER, Barnard College, & PHILIP E. RUBIN, to be discussed helps to isolate and characterize the activation compo Haskins Laboratories-Our studies of sine-wave sentences frustrate the nent and also to assess the modifying effects of attention. attempt to characterize perceptual organization in a few rules derived 1:50-2:10 (108) from Gestalt principles. In such sentences, phonetic perception survives Short-Term Memory and Sentence Processing: Evidence from multiple violations of grouping principles and concurrent impressions Neuropsychology. RANDI C. MARTIN, Rice University-Theories of sinusoidal pitch and timbre. Does this mean that phonetic perception of sentence comprehension typically assume a working memory sys is independent of auditory scene analysis? We report experiments to tem used to hold the results of phonological, syntactic, and semantic assess the extent to which the perceiver experiences divergent percep analyses. Several proposals are considered about the possible overlap tual effects of the phonetic and auditory aspects of sine-wave sentences. between the short-term memory revealed by traditional span tasks and 5:10-5:25 (115) the working memory involved in sentence processing. Evidence is Perceptual Restoration, Perceptual Bias, Priming, and Pseudo brought to bear from studies of the sentence-processing abilities of brain words: InsIghts from a Newer Methodology. ARTHUR O. SAMUEL, damaged patients who have severely restricted memory spans. SUNY at Stony Brook-Samuel (JEP: General, 1981) reported that per 2:15-2:35 (109) ceptua1 restoration of missing phonemes was stronger in real words than Varieties of Short-Term Memory In a ConnectionBIControi Archi in pseudowords and stronger in primed words than in unprimed. These tecture for Working Memory. WALTER SCHNEIDER, University effects did not appear in Samuel (J. Merrwry & Language, 1987). A of Pittsburgh, & MARK DETWEILER, Pennsylvania Stale University- new method is reported that clarifies when such lexical effects will and The nature of working memory in a connectionist control architecture will not appear. The method and results will be used to address the idea
482 Papers 116-126 Friday Afternoon
of "perceptual bias," a concept that has not received adequate atten depends on whether subjects attend to the context circles. Subjects saw tion in recent years. a briefly (150 msec) presented test circle surrounded by small and large context circles of different colors. After judging test circle size, sub ATTENTION n jects judged the color of either the small or large context circles. Per Embarcadero CD, Friday Afternoon, 3:40-5:45 ceived test size depends on which context circles were task relevant.
Chaired by Raymond M. Klein, Dalhousie University 3-DIMOVEMENT PERCEPTION n Pacific West DE, Friday Afternoon, 2:40-4:50 3:40-4:00 (116) Brightness Perception: Automatic Detection Versus ControUed Chaired by Maggie Shiffrar, NASA-Ames Research Center Identification in Concurrent Tasks. ANNE-MARIE BONNEL & PIERRE BERTUCCI, CNRS-UVF, Marseille (sponsored by Ervin 2:40-2:55 (Ill) Hafter)-Is there a qualitative difference between "automatic" or Orientation Affects Both Structural and Episodic Representations "capacity-free" detection and "controlled" or "capacity-limited" iden 01 3-D Objects. LYNN A. COOPER, OJlumbia University, DANIEL L. tification? A series of dual-task experiments on brightness perception SCHACTER, Harvard University, & CASSANDRA MOORE, Colum showed that different kinds of information are required for detection bia University-Previous work has demonstrated functional dissocia and for identification. It may be that transient channels mediate global tion between structural and episodic representations of unfamiliar, 3-D detection and behave essentially as "change detectors," which detect objects, and invariance of structural representations over changes in object signal onsets. In contrast, sustained channels mediate focal identifica size and parity. The present experiments examine the effects of manipu tion of the direction of change and behave as "integrative detectors. " lation of an object's orientation in the picture plane on priming and recog 4:05-4:25 (117) nition. Unlike reflection, rotation in the plane significantly reduces the Dimensions 01 Experience: What's Important lor Perception? magnitude of object~ision priming, suggesting that structural represen ROBERT D. MELARA, Purdue University-Cognitive psychologists tations of objects are axis and reference-frame based. distinguish between integral and separable experiences. "Integral" is 3:00-3: 15 (123) often meant to convey that perceivers have limited access to the consti Lateral AsynunetrWs in Representational Momentum. ANDREA R. tuents of the experience. For example, two dimensions of color HALPERN, Bucknell University, & MICHAEL H. KELLY, Univer saturation and brightness-are called integral because color stimuli are sity of Pennsylvania-Is representational momentum (RM) solely an in thought to be coded initially as unanalyzed points in color space. In this ternalization of environmental properties? Five experiments demonstrate talk, I discuss the difficulties-both logically and empirically-in main one aspect of RM that cannot have been learned from the environment. taining the traditional view of integral experiences. I argue instead that For our sample of right-handed subjects, objects apparently moving to certain dimensions are primary in experience. I report supporting evi the right engender a larger RM effect than those moving to the left. dence from color, sound, and touch perception. This effect is insensitive to changes in apparent velocity, type of ob 4:304:45 (118) ject, and retention interval, and it may be confined to objects in the left Is There a Color Advantage in ViwaI Search? DAVID W. MARTIN half of visual space. & ALAN E. BENSON, New Mexico State University-Ionides and 3:20-3:35 (114) Yantis (1988) reported that, in contrast to abrupt-onset targets, detec Stereokinetic Phenomena with Rotating Ellipses FoUow a Mini tion of uniquely colored targets in visual search was not enhanced when mum Principle. MARIO ZANFORLIN, University ofPadova-A rotat the cue was randomly predictive of target identity. However, we have ing ellipse gives rise to three different percepts: an elastic figure, a rigid consistently found a color-advantage effect with 5-character displays tilting disc, and a solid ellipsoid. Minimizing relative velocity differ when subjects are required to identify rather than detect a target. in an ences among all the points of the configuration yields three mathemati identification experiment, display size was manipulated and these results cally possible minimal solutions corresponding to the three percepts. were replicated. Explanations are offered for the detection/identification The hypothesis predicts: (I) the precise size of the perceived "objects" difference. in depth and (2) what stimuli should be added to obtain a unique minimal 4:50-5:05 (119) solution and a unique percept. Theoretical predictions are supported by Attentional Capture by Equiluminant Objects. STEVEN YANTIS experimental results. & ANNE P. HILLSTROM, Johns Hopkins University-Abrupt visual 3:40-3:55 (l2S) onsets capture attention. Possible mechanisms include (a) the activa Do Moving Observers Decompose Retinal Flow? lAMES E. CUT tion of low-level visual channels that are sensitive to luminance incre TING, Cornell University-Moving observers find their way through ments and (b) a higher level mechanism that orients attention to newly cluttered environments with ease and without injury. Most approaches presented obP:ts even without a luminance increment. Observers visually to wayfinding consider this feat to be done on the basis of information searched for a prespecified target in an array of objects defined by dis in optical flow. Thus, they insist that retinal flow be decomposed into continuities in luminance, motion, texture, or binocular disparity. The its component flow fields, one of which is the familiar radial expansion abrupt appearance of an equiluminant object tended to capture atten pattern of optical flow. I will present several lines of evidence, how tion, supporting mechanism (b). ever, that suggest decomposition is not performed by human observers. 5:10-5:25 (120) 4:00-4:15 (ll6) Novel Visual Shapes in Negative Priming. BRETT DE SCHEPPER Perceptual-Motor Organization and the Calibration of BiomecbaD & ANNE TREISMAN, University of California, Berkeley (read by Anne ical Information for Locomotion. JOHN I. RIESER, DANIEL A. Treisman)-We report a series of experiments on priming by novel visual ASHMEAD, Vanderbilt University, HERBERT L. PICK, IR., Univer shapes, exploring the memory traces left by a single exposure, either sity of Minnesota, & ANNE GARING, Vanderbilt University-People attended or unattended, and how they change over 30 successive presen coordinate the force and direction of locomotion with the distance and tations. Whereas no explicit memory seems to be available for the un direction of surrounding objects. To test coordination, people view a attended shapes in a superimposed pair (Rock & Gutman, 1981), we target, close their eyes, and attempt to walk to its location (or tum to do find negative priming when the novel unattended shape on one trial face it). Precise walking depends on the calibration of vision and ac becomes the attended shape on the next. tion, that is, on knowing how far to walk or tum. We induce recalibra 5:30-5:40 (121) tion when people walk at one rate while seeing a different rate of opti Attentional Modulation of Size Contrast. GORDON L. SHULMAN, cal flow. Transfer of the recalibration effects, at the level of specific Washington University Medical School-The perceived size of an ob muscles or limbs or directions, is used to probe the organization of the ject depends on the size of surrounding context objects. This illusion underlying system.
483 Friday Afternoon Papers 127-139
4:20-4:30 (127) 4:15-4:25 (133) Increases in IPD Reduces Perceived Depth in Stereograms. Length of Exposure to SucroIe and Negative Consummatory Con ROBERT FOX & DEBORAH L. MAUK, Vanderbilt University trast Effects in Rats. FRED VALLE, University ofBritish Columbia Increasing interpupillary distance (lPD) increases retinal disparity when Rats given alternating daily 2-h exposure to 16% and 4% sucrose solu 3-D space is viewed, but increases in IPD do not change the simulated tions prior to their daily meal of Purina Chow show negative consum disparity built into stereograms. If depth constancy is operative, inverse matory contrast effects to both the sucrose solution and the Purina Chow linear relationships will obtain between IPD and perceived depth inter on 4% days (Valle & Valle, AlAB, 1991). The present experiments vals even when disparity remains constant. Results from a factorial ex investigated the length of daily sucrose exposure required to engender periment (3 IPDs, three viewing distances) confonn to such relation the secondary contrast effect to Purina Chow: neither 15- nor 6O-min ships, albeit with a small systematic underestimation of perceived depth exposure is sufficient. (i. e., constancy prevails). 4:304:40 (134) 4:35-4:45 (128) Selective Breeding ror Negative Contrast. CHARLES FLAHERTY Misdirected Visual Motion: Mae and Pbi. ERIC J. HIRIS, Vander & KATHLEEN KRAUSS , Rutgers University-Degree of decrement bilt University, ROBERT H. CORMACK, New Mexico Tech , & RAN in lick frequency following a shift from 32% to 4% sucrose was used DOLPH BLAKE, Vanderbilt University (read by Robert H. Connack) as the criterion for a selective breeding program using Sprague-Dawley A peripherally viewed rectangle moving across a grating can appear derived rats. Reliable differences between the large and small contrast to deviate 90° from its true path. A model providing the requisite stimulus lines appeared by the fourth generation and have been maintained through energy underlying this illusion makes predictions relative to motion after the seventh generation. Behavior of the two lines in the open field, in effects and apparent motion. Aftereffects should be opposite the illu other contrast procedures, and in response to selected drugs will be sory direction of motion; they are. Two-point phi-motion should fail reported. to elicit the illusion under relevant conditions; it so fails . Why there 4:45-4:55 (135) is no illusion with foveal viewing remains unknown. Simultaneous ConcIitionIng in Honeybees. JOHN D. BATSON, Fur man Univemty, JAMES S. HOBAN, & M. E. BITTERMAN, University ANIMAL LEARNlNGIBEHAVIOR II ofHawaii (read by M. E. Bittennan) (sponsored by M. E. Bittennan) Pacific West FG, Friday Afternoon, 3:00-5:00 Proboscis extension was classically conditioned in two experiments designed to explicate the role of amount of reward in the development Chaired by Helen M. Murphy, John Carroll University of odor preferences by honeybees. The results contradict the long standing assumption that honeybees learn nothing about odors present 3:00-3:15 (129) during feeding. Learned Industriousness. ROBERT EISENBERGER, FRED A. MASTERSON, & FRANCES WEIER, University of Delaware RECOGNlTIONIRECALL II Extensive research with lower animals and humans indicates that re Golden Gate A, Friday Afternoon, 2:55-5:30 warded effort contributes to durable individual differences in industri ousness. It is proposed that reinforcement for increased physical or cog Chaired by Robert A. Bjork, University of California, Los Angeles nitive performance, or for the toleration of aversive stimulation, conditions reward value to the sensation of high effort and thereby reduces 2:55-3:15 (136) effort's aversiveness. The conditioning of secondary reward value to The Regularities or Recognition Memory. MURRAY GLANZER, the sensation of effort provides a dynamic mechanism by which rein New York University, JOHN K. ADAMS, Montclair State College, & forced high perfonnance generalizes across behaviors. KISOK KIM, New York University-Three regularities in recognition 3:»3:35 (130) memory are presented: (1) the mirror effect, (2) the order of receiver Context Specificity or Operant Discrimination Performance in operating characteristic slopes, and (3) the symmetry of movement of Pigeons. DAVID R. THOMAS, University of Colorado at Boulder underlying distributions. Sets of data are presented that demonstrate these Pigeons learned a key color discrimination in two similar chambers while regularities. These regularities are related to attentionllikelihood the a particular odor was present. Then they learned the reversal in one ory and to signal detection theory. of these chambers with a different odor present. The reversal did not 3:»3:40 (137) transfer to the alternative chamber. Subsequent experiments revealed Do Deliberate, Metamemorial Judgments Underlie the MIrror Er that visual contextual cues played an important role and that the use fect? DOUGLAS L. HINTZMAN, DAVID CAULTON, & TIMOTHY of a discrimination and its reversal (rather than single-stimulus train CURRAN, University of Oregon-In the mirror effect, categorization ing) was essential. Furthennore, the use of different odors potentiated of both old and new recognition items is better for one stimulus type the acquisition of control by visual contextual cues. (e.g., concrete words) than for another (abstract words). One hypothe 3:40-3:50 (131) sis is that this reflects a postretrieval judgment of the memorability of Opioid Versus Nonopioid Proactive Effects or Shocks on Gastric the test item rather than a fundamental aspect of the recognition process Ukers. J. BRUCE OVERMIER, University ofMinnesota, & ROBERT itself. To test this hypothesis, we examined recognition judgments using MURISON, University of Bergen-Rats exposed to either intennittent two paradigms: divided versus undivided attention and the response 5-sec shocks or to a single long shock show dramatically increased ulcer signal method of speed-accuracy tradeoff. ation induced 72 h later by exposure to restraint-in-water stress (at 19° C). 3:45-4:00 (138) A subcutaneous injection of naltrexone 20 min prior to the shock ses Recognition, Recall, and Episodic InfOl'lDlltion. DAVID L. HORTON sion blocked the proactive effects of intennittent shock but not the long & TIMOTHY J. PAVLICK, University ofMaryland-Further evidence shock. This suggests there are two types of proactively induced vulner will be presented for two kinds of recognition memory that depend on ability to ulceration; one is opioid mediated and the other is not. qualitatively different underlying processes. It will be shown that im 3:55-4:10 (132) provements in performance in these two kinds of recognition memory Body Weight Set-Point and Muscular Exercise in Rats. MICHEL are often inversely related to one another. Some implications of this CABANAC & JULIE MORRISSETTE, Laval University-Rats' body interpretation will be discussed, particularly as it pertains to the dis weight set-point was measured with the hoarding method to be 497 ±IS g. tinction between episodic and semantic memory. Then the rats were trained to run I h every day. Mean body weight 4:05-4:20 (139) set-point was 433 ±9 g (p < .05) when training took place just before On the Reality of the TulvIng-Wi1eman Function. ENDEL TULV hoarding session, and 504±21 g (n.s.) when training took place after lNG, University of Toronto, & ARTHUR J. FLEXSER, Florida Inter hoarding session. This suggests that body weight set-point is acutely national University-Hintzman's (1991) recent claim that the Tulving lowered by muscular exercise but not by chronic training. Wiseman (1975) function is a mathematical artifact is wrong. Data from
484 Papers 140-148 Friday Afternoon
a large number of experiments and experimental conditions in which lexical meaning and functional syntactic role involve different process recognition hit rate is equal to or exceeds probability of cued recall ing mechanisms. and therefore not subject to any mathematical constraints-adbere closely 4:00-4:20 (144) to the Tulving-Wiseman function. These empirical facts attest to the Prosodic Information for Syntactic Structure in Parental Speech. reality of the function. ANNE LEDERER & MICHAEL H. KELLY, University of Pennsyl 4:25-4:45 (140) vania (read by Michael H. Kelly)-Thirty-six mothers read one of six Recognition Failure in a Composite Holographic Associative Recall phrase structurally ambiguous sentences to their infants (mean age = Model (CHARM). JANET METCALFE, Dartmouth College-The re \0 months), with a prior story context indicating the intended interpre lation between recognition and the recognizability of recallable words tation. Analysis of the two interpretations revealed that syllable dura in human memory is an orderly function, but one that has resisted theo tions and FO at prosodic boundaries are strongly correlated with the num retical interpretation. A composite, holographic, associative recall and ber of syntactic boundaries crossed. This is the first demonstration that recognition model-CHARM-is applied to this situation. It is shown the hierarchical syntactic structure of an utterance is recoverable from that because of the interactive nature of the composite memory trace the prosodic structure of parental speech. in the model, a relation like that found in the data provided by human 4:25-4:40 (145) subjects is automatically produced. The TIp-of-the-Toogue (T01) Experience: The Effeets of Reaney, 4:50-5:05 (141) Frequeucy, and Aging. DEBORAH M. BURKE, KATHLEEN RAS'ILE, Knowing Wbo's New: The Pbenomenon of Jamais Vu. J. DON Pomona College, & CARRIE MARINER, University of Virginia-We READ, JOHN R. VOKEY, & MATTHEW DAVIDSON, University induced TOT states in young and older adults and found that probabil of Lethbridge-Subjects sometimes less frequently recognize a person ity of a TOT increased with subject age but decreased by about one half that they have seen than they falsely recognize a person not previously for target words presented earlier in a pronunciation task. Recency ef seen. Such "negative recognition" or "jarnais vu" occurs when a per fects were comparable across age, suggesting that age-related increases son's appearance in photos and video clips has changed from study to in spontaneous TOTs are not because of age differences in word recency. test and when episodic familiarity for the person is perhaps discounted A model is presented in which TOT probability is affected by word because specific retrieval is lacking. recency and frequency and by subject age. 5:10-5:25 (142) 4:45-5:00 (146) Cognitive Resources and Retrieval Interference Effects in Nor Toddlers' Understanding of the Normative Concepts of Color and mal People: The Role of the Frontal Lobes and Hippocampus. Number. MARILYN SHATZ & ANDREA BACKSCHEIDER, Univer MORRIS MOSCOVITCH, University of Toronto, Erindale Campus sity of Michigan-We propose that children's earliest understandings Four experiments are presented that show that a secondary concurrent about the normative concepts of color and number come from discourse interfering task (sequential finger tapping) at study and test impairs per with parents and that these understandings are top-down, concerning forrnance in normal people only on memory tests sensitive to frontal lexical categories, not word-to-world mappings. Our study examined lobe damage (memory for categorized lists, memory for temporal order, how toddlers under the age of two use color and number terms. Virtu release from PI, and phonemic/fluency) but not on tests sensitive to hip ally none could consistently map color or number terms to the world pocampal damage (memory for random lists, cumulative learning, mem correctly. However, even toddlers unable to use them appropriately knew ory on prerelease trials in PI test, and category fluency). These studies that color and number terms belonged to separate lexical categories. support the more general hypothesis that the hippocampus acts like a 5:05-5:15 (147) module in which cues elicit memories mandatorily, effortlessly, and au Repair Maneuvers. SAMUEL FILLENBAUM, University of North tomatically and the frontal lobes act like central systems that require Carolina at Chapel Hill-If a cooperative speaker says what should not cognitive resources for their operation. need saying, then maybe it really did need saying, and normal assump tions about reality may be revised. But what if the speaker asserts what has been presupposed immediately before? People recognize that some PSYCHOLINGUISTICS n thing is amiss and will reverse the order of presupposed and asserted Golden Gate B, Friday Afternoon, 3:40-5:45 materials or delete the latter, responding very differently than when the assertion conflicts with previously presupposed material or when nor Chaired by Laurie B. Feldman, SUNY at Albany mal assertion-presupposition order obtains. 5:20-5:40 (148) 3:40-3:55 (143) Spelling Ability and Word Recognition Strategies. VIRGINIA M. Word Recognition and Ambiguity in Meaning and Syntactic Func HOLMES, University ofMelbourne (sponsored by Edward J. Shoben) tion. LEONARD KATZ, Haskins Laboratories, MIRA PETER, & Knowledge of how to spell long words with idiosyncratic spelling pat KARL REXER, University of Connecticut-The effects of polysemy terns most strongly discriminates poor from good adult spellers. Thus, and functional ambiguity on lexical decision were examined, factori poor spellers make more errors than good spellers in resolving the lexi ally, in Serbo-Croatian and English. A polysemous word was recog cal status of long, irregular words. Poor spellers also have greater nized faster than a semantically unambiguous word. In contrast, am difficulty classifying nonwords based on long, regular words with pairs biguity in a word's case role (its functional role in its phrase) had the of medial letters transposed. Further, the idea that poor spellers process opposite effect: recognition was slowed. Apparently, the resolutions of words in a rapid, cursory fashion is not substantiated.
485 Friday Evening Posters 149-160
POSTER SESSION I (155) Pacif'1C Concourse East, Friday Evening, 5:45-7:15 Complex Conditional Relationships Learned by a Language Trained Sea Lion. ROBERT C. GISINER, Naval Ocean Systems (149) Center, Hawaii Laboratory, & RONALD J. SCHUSTERMAN, Califor PrerecognItioo Pattern Processing by Left and RJgbt Hemispheres. nia State University, Hayward (sponsored by Paul E. Nachtigall)-A LLOYD L. AVANT, AllCE A. TIDEMAN, & MICHAEL B. TEPIN, sea lion was trained to respond appropriately to instructions conveyed Iowa SUJte University-Good and poor gestalt dot patterns were pre by two to seven signs. She was then presented with probe trials con sented in paired pre- and postmasked 100msec presentations to the taining missing, added, or disordered signs. The sea lion's responses RVFILH or LVFIRH, and patterns were not recognized. Subjects judged showed that she had learned more than the specifically trained paired which of the paired flashes appeared to last longer. Duration judgments associate relationships between signs and referents. Signs were also showed differences between LH and RH prerecognition processing of treated as cues about the next sign in sequence (sequential conditional these patterns. Implications for earliest hemispheric visual processing relationships) and as cues for response organization (hierarchical con are discussed. ditional relationships). (1SO) (156) Effects or BIuITing OIl Hemkpberic Asymmetry for Processing Spa Retroactive Interference of Delayed "Symbolic" Matching to Sam tial Information. ELIZABETH L. COWIN & JOSEPH B. HELUGE, ple in Callfornla Sea Lions. RONALD J. SCHUSTERMAN, BRIGIT University of Southern California (presented by Joseph B. Hellige) GRIMM, University of California, SanUJ Cruz, ROBERT C. GISINER, Subjects indicated whether a dot was presented above or below a line Naval Ocean Systems Center, Hawaii Laboratory, & EVELYN or whether the dot was within 3 mm of the line. Stimuli were projected HANGGI, University of California, Santa Cruz-Two female Califor to the LVF or RVF on each trial and were presented clearly or blurred. nia sea lions were trained on two-choice delayed conditional discrimi Clear stimuli produced a task x visual field interaction that changed nation tasks. Delays were gradually extended for both sea lions, and with practice. Blurring disrupted performance and eliminated the task although Rocky showed little or no forgetting over delay intervals of x visual field interaction found with clear stimuli at early stages of from 1 sec to 2 min, Rio showed some forgetting over delay intervals practice. of from 1 sec to 45 sec. In retroactive interference experiments, both (151) sea lions showed nearly complete forgetting when irrelevant compari Perception of Conspecific Faces by a Bird. SUSAN D. BROWN son stimuli were shown during the delay intervals. & ROBERT J. IXlOLlNG, University ofMaryland (presented by Robert (157) J. Dooling)~The perception of conspecific faces by budgerigars was A Cognitive Perspective on Parent -Young Recognition in Kitti assessed by analyzing reaction times to discriminate between faces. Bud wakes (RJaG trldtJctyID). RITA E. ANDERSON, Memorial University gerigars discriminate among other budgerigar faces more efficiently than ofNelifoundland, JULIE M. PORTER, Fisheries and Oceans Canada, among the faces of other avian species. The salient features of conspecific & ANNE E. STOREY, Memorial University ofNewfoundland-A cog faces were confirmed using computer-generated synthetic models of nitive analysis of memory-related behaviors challenges the standard be natural faces. These models were further modified to demonstrate that havioral ecology claim that black-legged kittiwake .(Rissa tridactyla) efficient processing (discrimination) of faces in this species depends on parents do not recognize their chicks. Results of four experiments sug a normal configuration of features. gest that the lack of recognition reflects both lowered chick vocal iden (152) tifiability and reduced differential responding by chicks and parents to Palatability Differentially Affects Aversive Conditioning of Gus each other, relative to ground-nesting gulls. Kittiwake memory-related tatory and Nongustatory Food Cues. STUART R. ELLINS, behavior is discussed in terms of decision rules versus adaptation of mem CHRISTINE L. ROBERTS, & DAVID K. KENNEDY, California Swe ory systems. University, San Bernardino-Varying contributions of saccharin produce (158) tastes ranging from sweet (palatable) to bitter (unpalatable). In a series Context Value Modulates Levels of Response Recovery after Ran of experiments, we found that although unpalatable tastes followed by dom Control Training. LYNN ARONSON, Columbia University, LiCI-induced illness produce stronger conditioned taste aversions, palat PETER D. BALSAM, Barnard Callege, Calumbia University, & JOHN able tastes potentiate stronger conditioned aversions for auditory cues GmBON, New York State Psychiatric Institute and Columbia Univer that are spatially contiguous with ingesta. The results are discussed in sity (presented by John Gibbon)-Pigeons exposed to autoshaping proce terms of attentional differences in the aversive conditioning of gusta dures followed by response elimination using random control show tory and nongustatory food cues. response recovery in extinction (Lindblom & Jenkins, 1981). In Ex (153) periment 1, subjects receiving intervening context extinction before test Shock Avoidance Via Learned Orientation by Individual Drosoph ing showed stronger recovery. In Experiment 2, subjects tested in a con ilII~. KATHARINE L. LOFDAHL & JERRY HIRSCH, text not present during random training also showed stronger recovery. University of Illinois at Urbana-Cluunpaign (presented by Jerry Levels of recovery may depend on a comparator-like performance rule, Hirsch)-We report rapid shock avoidance learning (and more rapid sensitive to context value (Le., Gibbon & Balsam, 1981). relearning) of an orientational response accompanied by inhibition of (159) locomotion by individual Drosophila using a direct current voltage source Conditioning Flavor-Flavor Associatiom Via Opponent Processes to establish an optimal shock intensity. of SiclmesslWellness. LEWIS M. BARKER, MARK BOWMAN, & (154) LAURA HERBERT, Baylor University-Aversions to flavor A and The Effects of Medial Frontal and Orbitofrontal Lesions on Tim preferences for flavor B can be conditioned concurrently in rats when ing Performance in tbe Rat. DAVID L. FREDERICK & JOSEPH D. A precedes and B follows lithium injections. A also becomes associated ALLEN, University ofGeorgia (presented by Joseph D. Allen)-Timing with B. When, following conditioning, A is extinguished, reconditioned, shifts produced by medial frontal or orbitofrontal cortex aspirations were and reextinguished, preference for B tracks the aversion to A. If A is examined in 18 8 486 Posters 161-171 Friday Evening L. SPANGLER, BREIT HELLER, & PAOLO GAROFALO, National preserve space constancy, the CNS performs complex transformations Institule on Aging (sponsored by Donald K. Ingram)-Heptophysostigmine of continually changing retinal signals. Four subjects were tilted to 62 (1.0, 1.5, 2.0 mg/kg i.p.), a long-acting form of the indirect choliner different body positions (pitch and roll combined) and set visual targets gic agonist physostigmine, was administered prior to injection of the to their apparent vertical and horizontal. No visual direction cues were muscarinic cholinergic antagonist scopolamine (0.75 mg/kg i.p.) in 3- provided. Angular distortions as large as 40° were found when the body month-old F-344 rats. All rats were subsequently tested in a complex, was tilted more than 90°; a mathematical description of this nonorthogo shock-motivated 14-unit T-maze, known to be sensitive to cholinergic nality is presented. antagonism and demonstrating robust age-related effects in rats. Hepto (167) physostigmine attenuated the effects of scopolamine in a dose-related Tak Dependence In the Global Effect for Saccadic Eye Move manner. ments. CLARA CASCO & GABRIELLA LUPI, University of Padua (161) (sponsored by Lucia Colombo)-We recorded saccade amplitude while Stimulus Control of Pavlovian Inhibition. B. KENT PARKER, subjects performed either a " physical" or a "name" letter-matching SHERRY L. SERDIKOFF, & THOMAS J. SPENCER, West Virginia task on peripherally presented letter pairs. Saccades landed closer to University-The present study examined stimulus control of Pavlovian one of the two letters in the name-letter-matching task. In the physical inhibition. Training consisted of the presentation of a B stimulus in a letter matching, however, saccades fell to an intermediate position be feature-negative discrimination in which the A stimulus was followed tween the two targets. This is called global effect, and it is stronger by grain presentation only if not preceded by B (B-+A-IA+). Control with physically identical than with physically different letters. Our results by the B stimulus was assessed using generalization training. The results indicate that the activation of an automatic subsystem with low resolu are compared with those obtained in studies examining (a) stimulus con tion, underlying the global effect, is task dependent and is more likely trol of Pavlovian facilitation and (b) operant feature-negative discrimi to occur with tasks involving sensory codes. nation training. (168) (162) The Detection of Texture Density: Evidence from Adaptation. Landmark Learning in Honeybees. P . A. COUVILLON & M. E. FRANK H. DURGIN & DENNIS R. PROFFITT, University of Vir BITTERMAN, University of Hawaii-The performance of foraging ginia (presented by Dennis R. Proffitt)-The perception of texture density honeybees trained to discriminate between two feeding places, one of is susceptible to powerful adaptation effects amounting to a retinally which was distinguished by a pair of landmarks 30 cm to either side localized response suppression of about 50%, measured by a staircase of it, was impaired by removal of the landmarks after feeding began. method. The visual registration of texture density appears to be accom The results are compatible with accumulating evidence of excitatory plished by a detection system specifically sensitive to the density or "backward" conditioning in vertebrates. clustering of elements. This detection system may be locally distorted (163) by exposure to high-density stimuli. The effects cannot be explained Differences in AfIiliative Behavior among Voles (Mlcrotru). AL by luminance, contrast, spatial frequency, or size information. LEN L. SALO, University ofFlorida, LAWRENCE E. SHAPIRO, Na (169) tionailnstitUle ofMenJal Health, & DONALD A. DEWSBURY, Univer Rapid DIscrimination of Textures Comprising IM.pectrum sity of Florida (presented by Donald A. Dewsbury)-We are working Micropatterns. RICK GURNSEY & PAULINE PEARSON, Univer to characterize traits that lead prairie and pine voles to be generally sity of Western Ontario (sponsored by Mel GoodaIe)-Several recent monogamous and meadow and montane voles to be generally polygamous. theories assume that texture discrimination involves localizing energy We have previously shown that prairie voles are more contact prone, gradients within spatially restricted regions of "neural images," each that is, huddle more when in male-female pairs in I-h tests, than are of which represents the convolution of the input image with an orienta montane voles. We now show that pine voles huddle more than meadow tion and scale selective filter. This assumption was tested using textures voles. These data suggest that contact proneness may be a factor en comprising L- and X -type micropatterns that had identical amplitude gendering monogamy. spectra but different phase spectra. The resultant textures were easily (164) discriminable at brief exposures and discrimination showed an interesting Attenuation of the Habituation of Neophobia by Swim Stress. T. R. dependence on eccentricity of presentation. SCHACHTMAN, B. R. BUTCHART, J. L. CALTON, University of (170) Missouri-Columbia, & S. REILLY, Pennsylvania State University Med MagnItude Estimations of Reflex-Eliciting Stimuli: (Non)Contri ical School-Previous reports found that external stressors during clas bution of the Reftex. CHARLES B. WOODS, University of Florida, sical conditioning of a taste aversion can produce poor conditioned & P. HAERICH, La Sierra University (presented by P. Haerich) (spon responding at test. The present experiments tested one theory of such sored by W . Keith Berg)-Magnitude estimations and e1ectrophysio effects: that external stressors inhibit learning about internal events as logical measurements of the reflex blink were collected from subjects claimed by the gut-defense/skin-defense systems theory. Supporting this presented with acoustic (white noise) and tactile (periocular air puff), view, it was found that swim stress disrupted processing during simple reflex-eliciting stimuli of various intensities. Results indicated that flavor preexposure and maintained a neophobic response at test rela (I) perceptions of intensity changes were consistent across subjects (with tive to a nonstressed control condition. the power function for air puffs being greater than that for noise), (165) (2) reflex probability (but neither amplitude nor latency) was signifi Intrasession Dynamics of Waiting in Pigeons. CLIVE D. L. cantly associated with increasing intensity, and (3) magnitude estima WYNNE&J. E. R. STADDON, Duke University (sponsored by J. E. R. tions were independent of the elicited reflex. Staddon)-Pigeons were exposed to series of interfood intervals that (171) varied within a session. In different conditions, the frequency (number Color DI8crImination and CoIor.{)ppooeot Strength in Normal and of sine-wave cycles per session) and noise content (amount of random Defective Color VWOO. CHARLES B. WOODS & KEITH D. WHITE, noise added into a sine wave) of variation were changed. With time to University of Florida (sponsored by Robert D. Sorkin)-Color-normal, first peck as dependent variable, the results showed that adaptation to dichromatic, and anomalous observers were tested at five different field temporal contingencies was very rapid, but not solely based on the im sizes with Rayleigh matches to measure color discrimination and bichro mediately preceding interfood interval. matic mixture thresholds to measure luminance additivity. These mea (166) sures were compared within and between observers to determine how Distortions in the Perception of Spatiai Axes. URS J. BUCHER, well the luminance additivity index of color opponent strength predicts University of a,rich and NASA-Ames Research Center, FRED W. color discrimination. Luminance additivity could predict color discrimi MAST, & NORBERT BISCHOF, University ofa,rich (sponsored by nation among normals (high and low discriminators) and among dichm· Malcolm M. Cohen)-To perceive a stable visual environment and mats but not among simple or extreme anomalous trichromats. 487 Friday Evening Posters 172-184 (172) dots in each frame are located at the midpoints between dots of the preced The Effect 01 Target Position in a Visual Searcb Task. MARISA ing frame. Although the perceived frame-to-frame direction of motion CARRASCO & DENISE EVERT, Wesleyan University (sponsored by could vary randomly, cooperativity is indicated by the emergence of R. A. Kincbla)-Two experiments, involving an orientation x color two temporally coherent motion patterns, one unidirectional (small dot conjunction visual search task, examined the effect of target position separations), the other oscillatory (larger dot separations), and a third under free and fixed viewing conditions (eye movements possible and perceptual state in which the dots appear to be stationary. impossible, respectively). Both experiments found an asymmetry: the (179) tilted target was processed faster and better than the vertical one. Re Early Vision Detects Differential Structure Prior to Stereopsis. action time and errors increased with increasing eccentricity. The ef JOSEPH S. LAPPIN & WARREN D. CRAFT, Vanderbilt University fect was more pronounced as more stimuli were displayed. The results Spatial derivatives of optical images constitute potentially visible infor seem to support the "zoom lens" hypothesis of covert attention. mation about the structure of surfaces in 3D. Is early vision sensitive (173) to differential structure prior to binocu1ar combination of the two monocu Localization illusion in Taste. L. BARTOSHUK, L. GREEN, lar haIf-images? Evidently so: Stereoacuity ( < \0 sec arc) remains in C. LEHMAN, S. MARINO, Yale University School of Medicine, & variant under dichoptic transformations that alter correspondences be J. TODRANK, University of Pennsylvania-When a taste solution was tween pairwise spatial relations in the two monocular haIf-images but "painted" from an area with many taste receptors into an area with do not change their second-order differential structure. few, the initial strong taste filled the tactile path and produced the illu (180) sion of strong taste from the receptor-poor area. Similarly, when touch 3-D Bilateral Symmetry Assumed in Judging Figural Identity and was spared but taste was locally damaged or anesthetized, painting a Orientation. DIANE J. SCHIANO, MICHAEL K. McBEATH, solution from the intact into the impaired area produced sensations NRC/NASA-Ames Research Center, & BARBARA TVERSKY, Stan "from" the impaired area. However, anesthetizing touch and taste caused ford University (sponsored by Mary Kaiser)-Subjects viewed symmetric this illusion to disintegrate. and asymmetric random polygons and described "what the figure looks (174) like" and its orientation. The findings demonstrate that viewers tend Generalization between Sensorimotor Maps. HELEN A. CUNNING to interpret indeterminate shapes as silhouettes of 3-D bilaterally sym HAM & ROBERT B. WELCH, NASA-Ames Research Center (spon metric objects. Symmetric shapes are interpreted as front or top views sored by Robert B. Welch)-In the first of two experiments, subjects and asymmetric shapes as proflle or oblique views. The assumption that alternately practiced (I) a visual tracking task involving 108 0 clockwise indeterminate stimuli depict symmetric, 3-D objects appears to be a fre rotation of hand relative to visual space and (2) the normal hand-eye quently used pattern-recognition heuristic. mapping. Eventually, subjects performed both tasks with little mutual (181) interference. Positive transfer was found to a 72 0 clockwise rotation Eft'ects of Temporal Recruitment on Perceived Direction of Mov and negative transfer to a 72 0 counterclockwise rotation. The second ing StImuli. BENNETT I. BERTENTHAL & TOM BANTON, Univer experiment extended these results to reflection transformations. These sity of Virginia-The perceived direction of a moving stimulus is bi results support the existence and generaIizability of "multiple adaptations. " ased by its past stimulation. In a series of experiments, we show that (175) this bias is mediated by the velocity, and not merely the direction of The Influence 01 Visual Pitch on Visually Perceived Eye Level Is the past stimulation. Moreover, this bias is offset by contradictory depth Spatiopic. WENXUN LI & LEONARD MATIN, Columbia Univer information but not by contradictory form information. These results sity (sponsored by Leonard Matin)-The elevation of visually perceived are most compatible with a global cooperative model (composed of eye level (VPEL) is systematically influenced by the pitch of a visual velocity-tuned local operators) that is modulated by higher level field consisting of a single vertical line in darkness. The effect is essen heuristics. tially spatiotopic with the VPEL-versus-pitch function nearly invariant (182) with variation of horizontal gaze direction. However, with gaze directed Motor Coactivation and Response Force in Bimodal Stimulation within the median plane, change in horizontal eccentricity of the line Paradigms. ROLF ULRICH & MARKUS GIRA Y, University of Tubin produced systematic change in the slope of the VPEL-versus-pitch gen (sponsored by Jeff Miller)-Response force and RT were measured function. in two experiments using visual, auditory, and bimodal stimuli. (176) Responses were faster to bimodal than unimodal stimuli, both when Pa-ception of Shrinking in Apparent Motion. MAURICE HERSHEN responses were made to any stimulus (Experiment I, redundant targets SON, Brandeis University-Apparent motion was produced by expos effect) and when responses were made only to the visual stimulus (Ex ing triangles of different sizes for 100 msec with 50 msec lSI and periment 2, intersensory facilitation effect). Response force was also 200 msec recycle interval. Triangles with one co\inear side appeared greater for bimodal than unimodal stimuli, suggesting that part of each to shrink and expand in a fixed depth plane. Concentric triangles were effect occurs in the motor system. perceived as constant-sized objects moving in depth. Type of pattern (183) (dots, outline, or filled), background (blank; horizontal, perspective, Mental Rotation and the Gravitational Upright. ALINDA FRIED or full texture), and orientation of base had no effect. MAN & DANIEL PILON, University of Alberta-We independently (177) varied the angular disparity between two 3-D elongated shapes and the Influence of Depth Cues on Velocity of Apparent Motion. angular distance between the top of one shape and the gravitational up MICHAEL K. McBEATH, NASA-Ames Research Center (sponsored right. Both factors accounted for a significant proportion of the vari by Roger N. Shepard)-A new objective technique was used to mea ance in response times for identity judgments. In general, the effect of sure experienced velocity of 3-D apparent motion with displays that varied angular disparity between th~ shapes increased as the top of the left consistency of depth cues. Subjects discriminated between tilted, max most stimulus was increasingly farther from the upright. Results are imally fast clockwise and counterclockwise stimulus sequences. Tem discussed in terms of their implications for shape recognition. poral variance was better accounted for by 3-D spatial separation than (184) by 2-D retinal separation, and the influence of each depth cue was ad Vigilance for Real and Subjective Contours. W. TODD NELSON, ditive. When all cues specified tilt in depth, velocity of apparent mo JOEL S. WARM, WILLIAM N. DEMBER, University of Cincinnati, tion remained relatively constant over separation in 3-D space. & RAJA PARASURAMAN, Catholic University ofAmerica (presented (178) by Joel S. Warm)-Signa1 detection for real and subjective contour tar Cooperative SeIf-Organization in the Perception of Apparent Mo gets in a vigilance task varied inversely with increments in event rate tion. HOWARD S. HOCK & GUNTHER W. BALZ, FloriiUl Atlantic and event synchrony, providing further evidence for functional equiva University-A row of dots is presented in a series of alternating frames; lence in the perception of real and subjective contours. Work-load rat- 488 Posters 185 -196 Friday Evening ings (NASA-TLX) varied directly with increments in task difficulty only blue? Our results suggest not. We found a consistent failure of inhibi in conjunction with the subjective targets, a result which supports the tion of return in several nonspatial domains (viz., color, orientation, notion that the perception of subjective contours drains more attentional and form). Implications of these findings are discussed. capacity than that of real contours. (191) (185) The Effect 01 RelatIve Probe Positioo on a Dual Target RSVP Task. Preattendve and Attentive Processes in Enumeration: Color Versus SANDRA DRAKE & KlMRON SHAPIRO, University of Calgary Brightness Contrast. LANA TRICK, University of British Columbia (presented by Kimron Shapiro)-Using rapid serial visual presentation (sponsored by Jim Enns)-Subitizing is the rapid (40-100 msec/item) (RSVP), subjects are required to identify a target letter from among apprehension of number in the 1-4 range. Counting is the slower a stream of nontarget letters and indicate if a subsequent probe occurs. (250-350 msec/item) process of enumeration when there are larger num When the probe is foUowed by other letters, identifying the target results bers of items. Trick and Pylyshyn (1991) found that subjects can subi in a probe-detection decrement of about 450 msec. The probe-detection tize target items in a background of distractor items, but only if the tar decrement is attenuated, however, when the probe is the last letter stimu gets and distractors differ by a feature. This suggests that there may lus, even when it is masked by a random-dot pattern. be a limited capacity preattentive individuation mechanism at work in (192) subitizing, because enumeration by its nature requires individuation to Spatial Distribution 01 Attention in Detection and ldentiftcadon prevent targets from being recounted or missed. The counting process Tasks. STEPHEN D. RADER & FRANCES J. FRIEDRICH, Univer seemed to require spatial attention. In addition, contrast between tar sity of UtDh (presented by Frances J. Fricdrich)-Evidence exists in sup gets and the background affect the subitizing slope but not the counting port of several models describing the allocation of spatial attention, in slope. The relative contributions of the color and brightness channels cluding spotlight, zoom lens, and gradient. However, the tasks used to this effect are examined. have varied widely across models. This series of experiments explores (186) how the distribution of attention differs for detection and identification Dividing Attention between Color and Shape: Extensions to the and as a function of cue size. The results suggest that the nature of the Interactive Race Model. J. TOBY MORDKOFF, University ofCalifor task places strong constraints on attention distribution and the effective nia, San Diego, & STEVEN YANTIS, Johns Hopkins University (spon ness of different cue types. sored by Steven Yantis)-Previous work has shown that the performance (193) of tasks requiring divided attention across spatial locations is weU described Movement Preprogramming: When RT Does Not Increase with by an interactive race model. Under this model, only one target on Movement Complexity. IAN M. FRANKS, University of British redundant-target trials directly activates the response. Three new exper Columbia-Reaction time increases as the movement becomes more com iments demonstrate that dividing attention between featural dimensions plex. Three experiments are described that investigated several situa color and shape-is best explained in terms of coactivation with both tions where RT did not increase across levels of complexity. Two sim target features activating the response on redundant-target trials. ple movements were compared. When subjects were asked to extend (187) and flex in a continuous movement, the RT to complete this movement Early Perception: Searching lor Textures and Single Elements. was significantly greater than the RT to complete the extension move TARA C. CALLAGHAN, St. Francis Xavier University (sponsored by ment. However, when subjects were asked to extend, pause, and flex, C. Ann Cameron)-This experiment bridges the gap between visual there was no comparative increase in RT. search for single elements and texture segregation for groups of ele (194) ments (orientation and luminance) using a single vsearch task. Number Adaptive Coordination and RecaUbration 01 the Eye-Hand System of target elements (I, 2, 6), variation in background (constant, vari Under 0ptica1 MIsalignment. GORDON M. REDDING, Rlinois State back), and similarity of elements within target (high, low) varied. Results University, & BENJAMIN WALLACE, Cleveland State University indicate lowering similarity of elements within target impairs ability to Direct -effect measures of terminal error in target pointing observed under fmd target but the particular effect depends on dimension varied and optical displacement with visual feedback showed complete adaptation whether one or both regions (target, background) contain variation. in about 30 trials of paced (3-sec rate) sagittal pointing and over· (188) compensation for the remaining 30 trials. Aftereffect measures observed An Interpretation 01 the Role 01 Attention in Location Cuing. M. L. with the optical misalignment removed and without visual feedback CHEAL & D. R. LYON, University of Dayton Research Institute-In showed slower adaptation, which was only about 40% complete after order to better understand the attention mechanisms that occur during 60 trials. These differences suggest a strategic coordination component a trial in the location-cuing task, a framework was developed within of direct-effect measures beyond the automatic recalibration reflected which a proposed temporal sequence of events is specified. This theo by after-effect measures. retical framework helps explain several puzzling aspects of location (195) cuing data, such as (I) some of the temporal differences between central Tendency or Pattern? The Nature 01 Veering IUDOIII Blind Pedes and peripheral preeuing and (2) the effect of invalid cuing on feature trians. DAVID GUTH, Western Michigan University (sponsored by discrimination. John Rieser)-The ability of blind (n= 18) and blindfolded-sighted (n=7) (189) pedestrians to establish and to maintain a 2S-m straight-line trajectory Attending to Objects or Their Locations. KEVIN A. BRIAND, in the absence of auditory and underfoot cues was assessed across 3 SUNY at Albany (sponsored by Robert Rosellini)-Covert visual atten to 14 IS-trial test sessions. In agreement with previous research, most tion has been postulated to occur through either attention to objects or subjects exhibited a clear .. veering tendency" across the 15 trials of the locations they occupy. The present experiment induced either an individual test sessions. Across sessions, however, the characteristics object-based or a location-based mode of attention and confirmed that of the tendency tended to vary. Implications of these fmdings for ex· making simultaneous judgments about two separated objects was difficult plaining and reducing veering will be presented. only when in a location-based mode. When presumably in an object (196) based mode, distance between objects had no effect on performance. Does Vision Guide the Hand or Update Knowledge 01 Target l0ca (190) tion? LINDA L. LAGASSE & DAVID A. ROSENBAUM, University Inbibition 01 Return to Locations and to Features. HO-WAN ofMassachusens (sponsored by David A. Rnsenbaum)-Traditionally, KWAK & HOWARD EGETH, Johns Hopkins University (presented vision has been thought to guide aiming movements by aUowing on by Howard Egeth)-lnhibition of return refers to a detrimental conse line comparisons of hand and target locations. An alternative view, pro quence of orienting attention to a specific spatial location. Would a similar posed here, suggests that vision merely helps update knowledge of tar effect occur in a nonspatial task? For example, would the response to get locations. We have tested this hypothesis by comparing aiming to a red stimulus be slower if the preceding stimulus were red rather than extemaIly held or subject-held targets in the light or in the dark. 489 Friday Evening Posters 197-208 (197) versus intentional manipulation. There was an age-related decline in Tapping Speed and Bimanual Coordination in Expert Versus memory, but only for subjects older than 65 years. Amateur Pianists: Talent or Practice? RALF TH. KRAMPE, Max (203) Planck Institute, Berlin, & K. ANDERS ERICSSON, University of Organizational Signals Affect Text Recall. ROBERT F. LORCH, Colcrado at Boulder (presented by K. Anders Ericsson)-Twelve expert JR., & ELIZABETH PUGZLES LORCH, University ofKentucky-This pianists displayed higher speed than 12 age-matched amateur pianists study tested four models concerning how signals influence text recaIl. for both simple finger tapping and series of fixed keystrokes with either Subjects read and recaIled a text in which all topics were signaled simul or both hands. The magnitude of the skill differences increased as a taneously by overviews, headings, and summaries (ALL); 1\0 topics were function of coordination complexity. Amount of past and current deliber signaled (NONE); or half of the topics were signaled (HALF). The most ate practice assessed from diaries and retrospective estimates for each important finding was that signaled topics were recalled better than un individual predicted the pattern of movement speed as well as asymmetries signaled topics in the HALF condition. The results support the hypothesis between left and right hands better than the expert-amateur distinction. that signals induce selective processing of text topics. (198) (204) Age Differences in Serial Mental Rotation: Replication and Ex Organizational Signals and Reading Instructions Affect Text tension. CHRISTOPHER HERTZOG & BART RYPMA, Georgia In Processing. ELIZABETH PUGZLES LORCH, ROBERT F. LORCH, stitute of Technology-Hertzog and Rypma's (1991) pattern of age dif JR., & JONATHAN M. GOLDING, University ofKentucky (sponsored ferences in a serial mental rotation paradigm is problematic for inferences by Robert F. Lorch, Jr.)-Readers' text-processing strategies are con of age-related slowing in speed of mental rotation. We found relatively strained by the manner in which a text is written. Subjects read a small differences in rotation RT slopes accompanied by larger differ computer-presented text under instructions to read for comprehension ences in decision RT slopes. The present experiment, now in progress, or to do a free recall. The text either did or did not contain organiza replicates and extends these results by imposing a response deadline, tional signals (overviews, headings, and summaries). When the text did minimizing opportunity for mental rotation to be deferred to the deci not contain signals, reading speed was unaffected by instructions; when sion phase of the task. the text did contain signals, reading speed was faster under compre (199) hension than under recaIl instructions. Automaticity of Frequency and Spatial Encoding: Concurrent (20S) Processing, Practice, and Aging. JAMES M. PUCKETT, Texas A&I The Effect of Purpose for Reading a Procedure on Performance University, CONSTANCE E. TOFFLE, & LESLEE K. POLLINA, Measures. VIRGINIA A. DIEHL, Western Illinois University, LIEN West Virginia University-In Experiment I, spatial and frequency en CHONG MOU, University ofMarylond, CAROL BERGFELD MILLS, coding tasks were given alone and in combination, and incidentally and Goucher College and American Institute for Research, & DEBORAH P. intentionally (the latter manipulation both within- and between-subject BIRKM1RE, U.S. Army HU11IIJlI Engineering Laboratory (sponsored by in order to assess practice effects). Practice and concurrent processing Carol Bergfeld Mills)-The effect of purpose for reading a procedure effects were obtained. In Experiment 2, young and old adults were given on task completion and text recaII was investigated. Forty-eight sub the same conditions as in Experiment 1 except that only incidental con jects read one of two texts; half read to do the task and half read to ditions were used. Again, a concurrent processing effect was obtained recaII the text. Condition did not affect task performance. However, but no age effect. Results suggest that these encoding tasks are non subjects who read more naturally (Read to Do) recaIled more active than optimally automatic (Sanders et aI., 1989). static text, while subjects who read to recall did not. Thus, purpose for (200) reading affected the type of information remembered. Mental Rotation Ability as a Function of Age, Sex, and Handed (206) ness. STEPHANIE M. CLANCY & CAROL A. BONTEMPO, Southern Loci of Visual Areas Involved in Reading Filtered Words That nlinois University (sponsored by Dennis L. Molfese)-Mental rotation Improve Reading Performance. TERI B. LAWTON, University of ability, as measured by the Vandenberg and Kuse Mental Rotations Test, Oregon (sponsored by Michael!. Posner)-Image-enhancernent fIlters was examined in a sample of young (18-22 years) and old (66-85 years) boost the amplitude of spatial-frequency components that are less visi adults as a function of sex and handedness. The mental rotation scores ble for the low-vision observer. Low-vision observers have experienced of males were higher than females, independent of handedness. Right a two- to fourfold increase in reading speed at up to 70% less magnifi banders, as assessed by the EdinWrgh Handedness InventoIy, had higher cation when reading filtered text. I also report reading-speed changes scores than left-banders. Young subjects exhibited superior performance for myopic viewers and discuss the neural systems involved in these relative to older adults. Several interactions involving age, sex, and improvements. handedness were also observed. The implications of the present find (207) ings for the study of spatial ability based on sex and handedness across Lexical Guidance in Sentence Parsing. BEVERL Y COLWELL the life span is considered. ADAMS, CHARLES E. CLIFTON, JR., University of Massachusens (201) at Amherst, & DON C. MITCHELL, University of Exeter (sponsored Aging and the Inbibition of Spatial Location. S. USA CONNELLY by Charles Clifton, Jr.)-Eye movements were measured while sub & LYNN HASHER, Duke University (presented by Lynn Hasher) jects read sentences like "Although the audience yawned, the come Younger and older adults indicated the location of a target that appeared dian continued telling very bad jokes." The intransitive verb (e.g., on critical trials with a distractor. Across successive pairs of trials, both "yawned") was replaced half the time by an optionally transitive verb groups were slower to respond to the target when it appeared in a loca (e.g., "booed"), and each was followed half the time by a common. tion previously occupied by the distractor than when it appeared in a The pattern of flXation durations and regressive eye movements was new location. While older adults appear unable to inhibit identity, they interpreted in terms of how and when verb transitivity information is used. seem as able as younger adults to inhibit spatial locations. Interference (208) data are also presented. Word Class Effects in Intrasentential Code-Switcbing. SHOn (202) AZUMA, University of Utah, & RICHARD P. MEIER, University of Spatial Memory Changes in Adultbood. BOB UTTL & PETER Texas at Austin (sponsored by Donald J. Foss)-Bilinguals frequently GRAF, University ofBritish Columbia (presented by Peter Graf)-Age alternate, or "code-switch," between two languages within a single sen related changes in spatial memory were investigated as part of an ex tence. We report two sentence repetition experiments on SpanishlEngIish hibit on memory. The subjects were visitors to the exhibit. In an inten and Japanese/EngIish code-switching. Our results show that the switching tional condition, subjects were informed about an impending memory of closed-class morphemes causes significantly longer reaction times test before they entered the exhibit; in an incidental condition, they were and more errors than the switching of open-class morphemes. These not so informed. Memory performance was not affected by the incidental results are interpreted to support the claim that open- and closed-class 490 Posters 209-221 Friday Evening morphemes are involved in distinct levels of processing in speech sity of Colorado-Subjects read either of two 26-sentence versions of production. a small town. One description consisted of geographical relations (sur (209) vey) while the other presented the same locations sequentially encoun The Independence of Thematic Role Information. SARAH D. tered (route). Sentences were presented one at a time on 3 x 5 cards. BREEDIN & RANDI C. MARTIN, Rice University (presented by Feedback consisted of either (l) limited access to a map during read Randi C. Martin)-An aphasic patient is reported who shows disrupted ing, (2) limited referrals to previously presented sentences during read knowledge of thematic role assignment for verbs yet otherwise shows ing, or (3) access to the entire text upon completion. There were sig preserved knowledge of semantic and syntactic structures. His defec nificant differences in inferential reasoning but not propositional reca11. tive knowledge of thematic roles affects both production and compre (216) hension. The results suggest either an independent thematic role processor Semantic Interference and Orthograpbic Fadlitadoa ID DeftDidoo or lexical representations for thematic information that is separate from Naming. WIDO LA HEIJ, University ofLeilkn (sponsored by Judith F. other types of lexical representations. Kroll)-Semantic interference and orthographic facilitation are common (210) findings in Stroop-like naming tasks. We show that these context ef Tracking Sentence Topics in Speech and Writing. JEAN E. NEW fects also obtain when, instead of colors or pictures, definitions are used MAN, University ofNew Mexico (sponsored by Peder Johnson)-A dis as target stimuli. Furthermore, it is shown that the semantic interfer course continuation task investigated the influences of stimulus modality, ence effect (a> cannot be attributed to a strategic match/nonmatch deci word order, and focus on the production of sentence starting points, sion and (b) only obtains when the context word follows the definition. their surface forms, and underlying topic relations. Starting points for These findings will be related to context effects in studies on word speech depended on recency and focus, and repeated stimulus NPs. Writ retrieval and the tip-of-the-tongue phenomenon. ing used more pronomina1ization and was less influenced by focus and (217) word order; topic relations did not differ between modalities. This pattern Visual and Phonological Codes ID Repetition BIincIness. DAPHNE points to the importance of surface form in processing spoken discourse. BA VELIER & MARY C. POTTER, Mn'(presented by Mary C. ~) (211) Repetition blindness (RB) is the inability to detect the second occur Depth of Interpretation in the Processing of Literary Discourse. rence of a visual stimulus when repeated within a few hundred milli PETER DIXON, MARISA BORTOLUSSI, LESLIE C. TWILLEY, seconds. The role of visual versus phonological (name) similarity was & ALICE LEUNG, University ofAlbena-Is the processing of literature examined. RB occurs between visually similar items, but also between different from that of other kinds of discourse? We propose that literary phonologically identical items that are visually distinct. We propose that discourse is more likely to support depth of interpretation; that is, it RB results when the codes used for initial registration of two targets provides a greater opportunity for elaboration and inferencing that goes in STM are similar or identical. beyond the propositional content of the text. In our research, depth of (218) interpretation was measured by assessing how much readers' evaluations Repetition PrimIng Effects with Pseuclobomophones. LAREE A. changed after rereading literary and nonliterary texts. HUNTSMAN, San Jose Stale University, & SUSAN D. LIMA, Univer (212) sity ofWisconsin-MilwauJcee (sponsored by Susan D. Lima)-In a repe Working Memory Capacity Affects Use of Context in Parsing. tition priming experiment, orthographically similar word primes (e.g., JONATHAN KING, University of Califomi a, San Diego (sponsored DREAM), pseudohomophone primes (e.g., DREEM or DREME), and non by Patricia A. Carpenter)-The use of contextual information during word primes (e.g., DROAM) were all found to facilitate lexical decision parsing depends on available working memory capacity. Large differ response latencies to high-frequency word targets (e.g., DREAM). How ences between high- and low-apacity readers in their processing of object ever, only word and pseudohomophone primes reduced error rates on relative sentences in a "neutral" context collapse when the context pro high-frequency word targets. Low-frequency word targets benefited from vides thematic role assignment cues. But low-(unlike high-)capacity word primes only. The results suggest that both phonological and or readers do not use presuppositional information to guide their initial thographic codes exen an influence on lexical access. processing of temporarily ambiguous sentences, and thus always be (219) have in accordance with the memory-conserving principle of minimal Inhibitory Phonological PrimIng in Auditory Word Recopition. attachment. LOUISA M. SLOWIACZEK & MARY BETH HAMBURGER, SUNY (213) at Albany (sponsored by Donn Byrne)-Research in word recognition The Effect of Inferential Processes on Perceptual Identification. has suggested that lexical memory may be organized phonologically as REBECCA FINCHER-KIEFER, Gettysburg College (sponsored by well as semantically. In three experiments, subjects shadowed auditorily Paul R. 0' Agostino)-Perceptual identification responses were collected presented target items that were preceded by phonologically similar word for briefly presented target words that followed short texts. Results and nonword prime items. Results revealed a complex pattern of inhi demonstrated that targets that had been explicitly presented were iden bition and facilitation of target items preceded by word primes. How tified significantly better than targets not presented. Further, targets that ever, a significant effect was not found for targets preceded by phono had not been presented but were required for referential coherence (im logically related nonword primes. The results provide evidence for a plicit targets) or were predictive in nature (predicted targets) were iden phonologically organized network in lexical memory. tified as well as explicit targets. This pattern was only obtained when (220) the targets were related to the preceding text. LexIcal Representation of MorpboIogiaIIIy Complex Words. EnA (214) DREWS, Technische Universitat Braunschweig (sponsored by Karl Comprebensioo Strategies during Reading. JASON E. ALBRECHT Friedrich Wender)-In a series of priming experiments, we tried to dis & EDWARD J. O'BRIEN, University of New Hampshire (presented tinguish between the levels of lexical representation involved in process by EdwardJ. O'Brien)-Fourexperiments examined the effect of present ing of polymorphemic words. For this purpose, we used German and ing inconsistent location information concerning the location of the protag Dutch particle verbs, which can also differ in semantic complexity. Par onist, or information inconsistent with the perspective of the protagonist, ticle verbs sharing the same stem morpheme could be either semanti on the comprehension of short narratives. In Experiments I and 2, in cally transparent or opaque. The results show strong stem morpheme formation that violated the location of the protagonist resulted in increased priming effects. However, these effects are significantly influenced by reading times and poorer recall. In Experiments 3 and 4, subject read semantic complexity and experimental task. ing times were unaffected by violations of the protagonist's perspective (221) unless they were explicitly instructed to read from that perspective. Developing Episodic DIstinctiveness Via Retrieval Practice: Insu (215) lation from As80ciate Interfereace. TODD M. GROSS &. ROBERT A. Influence of Text and Feedback Variations on Discourse Process BJORK, UCLA (presented by Robert A. Bjork)-In an A-B, A-C paired ing. PHILIP LANGER, VERNE KEENAN, & KATE BLASH, Univer- associates paradigm, retrieval practice on some of the A-B and A-C 491 Friday Evening Posters 222-232 pairings was provided after the A-B and A-C study episodes, respec 1990) reported that repetition effects on a homograph (e.g., BAR) were tively. Practiced items did not suffer proactive or retroactive interference smaller if the semantic context accompanying each of the two presenta from competing responses on a later MMFR test, whereas unpracticed tions differed (steel vs. tavern) compared to repetition effects with con items did suffer such interference. When study and retrieval-practice texts identical or similar in meaning. However, in our experiments, the events were mixed in a single list, however, practiced as well as non repetition effects did not decrease as a function of the context differ practiced pairings exhibited associative interference. ences during the two presentations. The role of the overall experimen (222) tal conditions leading to such discrepancies is discussed. Test Awareness and the LonaevitY of AsIiodative Priming. JOHN M. (228) RYBASH, Hamilton College and MoluJwk Valley Community College, Mechanisms of Visual Word Recognition in Fluent Dyslexic & ADEEL HASAN, Hamilton College-Three groups of college stu Readers. JAMES HODGSON, Massachusetts General Hospital (spon dents performed an associative priming task either 10 min, 4 days, or sored by David Swinney)-A plausible model of visual word recogni 8 days after they generated sentences to pairs of semantically unrelated tion in dyslexia might propose that familiar words are recognized by words (e.g., flannel-broccoli; apricot-surgeon). The associative prim logographic processes, resulting in strong frequency but weak regular ing task consisted of word fragments that appeared in the same (e.g., ity effects, whereas unfamiliar words are identified through prelexical flannel-b_cc_i) or different (e.g., flannel-su_g_n) context relative recoding to phonology, implying strong regularity but weak frequency to items on the sentence-generation task. Subjects who recognized that effects. The predicted interaction of these two factors was sought in nam test items appeared on the sentence-generation task displayed associa ing experiments with 25 high school dyslexics. Strong regularity and tive priming at each time of testing. Subjects who failed to recognize frequency effects were found but no interaction. Possible explanations that test items appeared on the sentence-generation task exhibited non are explored. associative priming, but not associative priming, at each time of testing. (229) (223) Connectives Increase Word Activation and Clausal Integration. Does the Surprise ResponIIe Mechanism Play a Role in the Bizarre KEITH MILUS, Northern Illinois University, & MARCEL JUST, Imagery Effect? VIRGINIA A. LANG, Wichita State University (spon Cornegie-Mellon University (sponsored by Marcel Just)-Three experi sored by David G. Payne)-Surprise response has been proposed as an ments investigated the effect of connectives on the modulation of word underlying mechanism that can predict and explain the inconsistency activations during reading. Subjects read pairs of clauses that were either of the bizarre-imagery effect (Hirshman, Whelley, & Palu, 1989). This joined or not joined by a connective (e.g., because). Inunediately after set of experiments tests what, if any, role surprise response may have reading the sentence(s), subjects performed a word-recognition task. by manipulating the associate level of the kHle-remembered items. Mem The recognition time for the verb from the first clause (or sentence) ory performance is measured using both a free- and cued- (verb is the was facilitated when a connective was present. Overall, the data indi cue) recall test. Preliminary results indicate that the surprise response cate that connectives increase the activation of concepts and thus facili mechanism is not supported. tate interclause integration. (224) (230) Unconscious Memory on a Conceptual Indirect Test. COLLEEN DUfereoces in Cobesiveness IIDlODg DHrerent Types of Word-Initial KELLEY, Macalester College, & D. STEPHEN LINDSAY, Univer Consonant Clusten. REBECCA TREIMAN, Wayne State University, sity of Victoria-Performance on implicit memory tests can be medi & CAROL A. FOWLER, Dartmouth College and Hoskins Laborotories ated by explicit memory. To rule that out, we placed the effects of con In two experiments, subjects saw printed stimuli such as "gav" and scious recollection and unconscious memory in opposition and then tested "spem. " They substituted either the first phoneme or the first and sec for unconscious memory on a conceptually driven test. Subjects were ond phonemes of the second stimulus for the corresponding phonemes correctly informed that words on a list they had just studied were all of the first stimulus and pronounced the result ("sav" or "spav") as wrong (but related) answers to questions on a general-knowledge test. quickly as possible. The results suggest that word-initial consonant When the subjects' attention was divided during study, unconscious mem clusters are cohesive units. Consistent with certain linguistic theories, ory effects occurred at test. lsI clusters seem less cohesive than other clusters. (225) (231) Individual Differences in Inhibition from Unattended Stimuli. Does Perceptual Salience InftueDce "Pop-out" with Words and Non JOSEPH S. BROWN, THOMAS C. LORSBACH, & GREG B. SIMP words? SAL A. SORACI, JR., University of A.labama, JEFFREY J. SON, University ofNebraska at Omaha (presented by Thomas C. Lors FRANKS, MICHAEL T. CARLIN, THEODORE P. HOEHN, & bach)-We examine the extent to which inhibition by ignored stimuli JAMES K. HARDY, Vanderbilt University (sponsored by Keith Clay (ripper & Driver, 1988) is affected by spatial location. Leaming-ilisabled ton)-Previous research (Flowers & Lohr, 1985) has not found evidence (LD) and non-leaming-disabled (NLD) children named one of two let for the "pop-out" phenomenon (e.g., Treisman, 1988) with word and ters. On half the trials, a second pair containing the ignored letter from nonword letter strings. Flowers and Lohr utilized a visual search task the previous trial was presented. For LD children, inhibition occurred consisting of a circular arrangement of stimuli. This arrangement does only if the letter appeared in the same position as the previous trial. not optimize interstimulus contiguity, which is critical in facilitating For NLD subjects, there was significant inhibition at both positions. stimulus detection in certain contexts (Soraci, Carlin, Deckner, & Bau (226) meister, 1990). The present study examines "pop-out" with words and Implicit and Explicit Memory for Words and Voices. BARBARA nonwords on a same-different task, under conditions (i.e., high inter CHURCH & DANIEL L. SCHACTER, Harvard University (sponsored stimulus contiguity) designed to enhance perceptual salience. by Daniel Schacter)-We examined priming effects on auditory iden (232) tification and stem completion tests following semantic and nonseman Word Onset Venus Word Specification in Spoken and Visual tic encoding of target words. Speaker's voice was same or different at Word Recognition. SARAH C. WAYLAND, Northeastern Univer study and test. Overall level of priming was less affected by the encod sity, & ARTHUR WINGFIELD, Brandeis University (sponsored by ing manipulation than was explicit memory; voice effects were observed Arthur Wingfield)-We investigated the importance of word onsets in only under circumscribed conditions and were unrelated to study tasks. auditory and visual word recognition. SUbjects identified gated audi Results were interpreted in the context of neuropsychological research tory and visual stimuli that presented more and more of a word's be on auditory processing and perceptual representation systems. ginning. For comparison, responses were collected for entire words pre (227) sented with increasing SIN ratios and tachistoscopic durations. Analyses Context Cbange and Recognition of Repeated Words. AYOIN Y. of stress patterns, phonemic overlap, and word frequencies of the DURGUNOGLU & BARBARA J. HANCIN, University of nlinois responses indicated that the word-onset priority in recognition is due (sponsored by James H. Neely)-Masson and Freedman (JEP: /.Me, to increased efficacy in specifying the word. 492 Posters 233-244 Friday Evening (233) tests. Our research represents a case study of the challenges of estab The Effect of Word Frequency during Two Readings of a Text. lishing external validity of memory phenomena. GARY E. RANEY & KEITH RAYNER, University of Massachusetts (239) at Amherst (sponsored by Keith Rayner)-The effect of word frequency Translation Problems in Causal Reasoning: Graphical Represen on reading time was examined during two readings of short passages. tation and Inference Strategy. EMILY DffiBLE, University of The frequency of target synonym pairs was varied between the first and Washington, & HARRIET SHAKLEE, Seattle University (presented second reading. Each synonym pair contained one high-frequency and by Harriet Shaklee)-Previous work shows that people rely on cause one low-frequency word. Fixation times were shorter for high- than for present outcomes in making causal inferences. We manipulated data low-frequency words during each reading. During the second reading, presentation format to draw attention to cause-absent outcomes. Accuracy fixation times varied depending on whether a high- or low-frequency patterns across causal judgment problems suggest that incorporation of synonym was seen during the first reading. cause-absent outcomes was most likely with data in a pie chart, inter (234) mediate for bar charts, and least likely for tabled data. Results suggest ERP Evidence of Differences in the Processing of Concrete and that data representation format readily influences how people evaluate Abstract Words. PHILLIP J. HOLCOMB & JOHN KOUNlOS, Tufts the relevance of statistical information for causal inference. University-Event-related potentials were recorded to concrete and ab (240) stract words in a lexical decision and a concreteness judgment task. The Age Changes in Handedness Are Task Specific:. CLARE PORAC, N400 component was larger to concrete than to abstract words in both University of Victoria-Published reports have shown that the incidence tasks. Also, the N400 produced a greater repetition effect for concrete of left-handedness decreases among older individuals. In this study, than abstract words in the right hemisphere, but equal effects in the left respondents, ages 8 to 85 years (N = 654), described the degree of right hemisphere. These results support models that posit separate systems versus left-hand use using a 7-item inventory. Older individuals showed for imaginal and propositional representations of word information. a lower incidence of left-handedness, but the magnitude of this trend (235) differed for different handedness behaviors. These behavior-specific find Individual Differences in Lexical Access. JANET L. McDONALD, ings suggest that age-relatedness changes in handedness incidence in Louisiana State University-Subjects of differing verbal ability as mea volve an environmental adaptation component. sured by English ACT scores performed a lexical decision task on non (241) words consisting of prefixes and nonprefIxes combined with stems and Local Versus Globallncoberence: Some Evidence from ERP Data. nonstems. While subjects of average ability showed the reaction time MARIE ST. GEORGE, SUZANNE MANNES, & JAMES E. HOFF pattern predicted by a serial prefix stripping model, subjects of very MAN, University of Delaware (presented by Suzanne Mannes) (spon high or very low ability did not. In addition, very low-ability subjects sored by Frances Graham)-Event-related potential research has demon showed a larger prefix effect than did other subjects. Implications for strated that the N400 component is sensitive to local semantic expectancy, models of lexical access are discussed. with words less likely to occur within the context of a sentence being (236) followed by N400s of greater amplitude. In the present experiment, more Memory for Interrupted Problems: The Zeigarnik Effect Revisited. global, theme-related context effects were investigated by recording ERPs COLLEEN M. SEIFERT & ANDREA L. PATALANO, University of while subjects read titled and untitled paragraphs one word at a time. Michigan (sponsored by Edward E. Smith)-Zeigarnik (1927) found Data suggest that the N400 component is larger for untitled paragraphs, better free-recall memory for interrupted tasks than for completed ones, indicating sensitivity to global semantic incongruity. despite shorter solution attempts. Following this methodology, one study (242) showed completed tasks were better remembered; however, a second Quality of Motivation Questionnaire Predicts Codependency. experiment replicated Zeigarnik when subjects could abandon problems R. CHRIS MARTIN, BRAD RUTKOWSKI, & RAMA ESHELBREN they did not solve. Factors determining memory differences by solu NER, University of Missouri-Kansas City-The concept of codepen tion status include interruption method, time on problem, and differen dency has been criticized extensively for lack of anchoring. The Friel tial set size. Incomplete status can serve as a distinctive memory cue questionnaire is one of the few objective tests but does not identify areas that aids free recall of past problems. where behavior change is needed most. The Quality of Motivation Ques (237) tionnaire does show such information, and 25 couples given both tests Automaticity of Information Extraction from Tachistoscopic Dis showed that the two tests equally predict codependency and the scores plays. ROBERT L. STEPHENS, JACQUELYN Y. PEARSON, & were significantly correlated (p < .01). A model for anchoring the con MICHAEL D. SISSON, U. S. Amzy Aeromedical Research lAboratory cept is presented. (sponsored by Kenneth B. Melvin)-This experiment provides evidence (243) of differential levels of automaticity for processing information from Attributions of Self-Esteem as a Function of Duration of Eye Con tachistoscopic displays. Using three variations of the partia1 report proce tact. JOYLIN M. DRONEY & CHARLES I. BROOKS, King's Col dure, processing of both spatial and identity information in letter dis lege (presented by Charles I. Brooks)-Subjects viewed one of three plays was assessed simultaneously and then independently. Practice ef videotapes in which a model maintained eye contact with an interviewer fects on the resulting three tasks indicate that feature extraction is a highly for 5, 30, or 50 sec of a 6O-sec period. Subjects then completed the automated process, while spatial information extraction is a controlled Multidimensional Self-Esteem Inventory, answering as they thought the process. Implications for separate processing streams in the visual sys model in the tape would answer. Results showed that, for all self-esteem tem are discussed. scales, scores significantly increased as amount of eye contact increased. (238) Thus, increased eye contact produced attributions of high self-esteem. The Expanse of the Expanding Test Series Effect. JOHN J. (244) SHAUGHNESSY, Hope College, EUGENE B. ZECHMEISTER, WIL Developmental Correlates of Direct and Indirect Remembering. LIAM L. CULL, & HOLLY M. HART, Loyola University of LINDA J. ANOOSHIAN & PENNIE S. SEIBERT, Boise Stale Univer Chicago-Landauer and Bjork (1978) demonstrated that an expanding sity-Two testing sessions for 48 children at each of two age levels (4-5 test series (1-5-9) led to better cued recall than a uniform series (5-5-5) years and 8-9 years) and 96 parents provided measures of indirect mem or a massed series «()..()..{). Using expanding test series has been referred ory (picture identification), direct memory (recognition, recall), typi to as "a new and powerful mnemonic strategy." Our series of labora cality judgments, analytic and holistic processing (in classification), and tory experiments tested its generality over materials, pacing of study, theories of mind (4-5 year olds). Discussion of observed relations be modality, and the presence or absence of study opportunities following tween memory measures and other task performances will focus on the 493 Friday Evening Posters 245-251 significance of indirect memory to understanding developmental change sity College of London, & CARLO SEMENZA, University of Padua in other areas. (presented by Brian Butterworth) (sponsored by Patrizia Bisiacchi)-In (245) a series of experiments, speakers of Italian (whose morphology is richer Relationships between Weight, Body-Esteem, and Parental In than English, thus allowing more articulate experimental manipulations) fluence In AdoIe8ceats. MARY E. QUINLAN-HEINS & MARCIA Z. completed sentence fragments designed to elicit erroneous subject-verb LIPPMAN, Western Washington University (presented by Marcia Z. agreements. At variance with previous findings in English (Bock & Lippman)- The relationships between actua1 weight, body~steem, self Miller, 1991), broken agreement was shown to be affected reliably by esteem, and perceived parental evaluation of appearance were studied both semantic and morphological features of sentences subjects. This in a sample of 53 adolescent females. The Body-Esteem Scale, Piers result is discussed in light of current questions about the relative segre Harris Children's Self-Concept Scale, and a Believed-Parental-Evaluation gation of sentential constituents. Scale were administered to classroom groups. Perceived parental evalu (249) ation of appearance was found to be a better predictor of body~steem and self~steem than actual weight. Implications for intervention with Effects of Instruction on Normal Listeners' Perception of Error. adolescents were discussed . NANCY B. MARSHALL, LINDA WARREN DUKE, & AMANDA C. (246) WALLEY, University ofAlabama at Bimringham (sponsored by Linda W. Children's Inferences about the Health Consequences of Contact Duke)-The effects of pause time and task instructions on the ability with Artifacts Versus Natural Kinds. KEN SPRINGER, SoutMrn of college students to make judgments of appropriateness and correct Metlwdist University (sponsored by Curtis W. McIntyre)-This study ness for different sentence types were investigated. Sentence sets in examined children's ability to infer the health consequences of contact cluded affirmative active, affirmative negative, passive active, and passive with artifacts and natural kinds that have been altered in some way . By negative types. Each type sentence was presented containing no errors, age 4, children drew domain-specific inferences about physical contact, semantic errors, syntactic errors, and semantic and syntactic errors. The asserting, for instance, that licking a clean spoon that had lain in the results of statistical analyses will be discussed in terms of automatic versus sun for a week would not make one sick, whereas licking a clean piece effortful processing of language. of apple under the same conditions would cause illness. (250) (247) A Theory of Lemma Retrieval In Speaking. ARDI ROELOFS, Serial and Parallel Processing In Japanese Kanji Recognition. Nij~gen Institute for Cognition Research and Information Technology CHISATO AOKI, MASAOMI ODA, ATR Auditory and Visual Per (sponsored by Herbert H. Clark)-The paper reports a theory on con ception Research Laboratories, Japan, & TOSHIO INUI, Kyoto Univer ceptually driven word (i.e., lemma) retrieval. The mental lexicon is con sity (sponsored by Thomas S. Aiba)-We examined wbether Kanji Oogo ceived as a network with concepts, lemmas, and word forms. Lemmas graphic) characters consisting of left (hen) and right (tsukuri) radicals are retrieved via activation spreading from concepts towards lemmas, are processed holistically or by constituent in parallel or sequentially. with the highest activated lemma being selected. A computer model im Comparing Kanji-radical, radical-Kanji, and pseudo-Kanji-radical pairs, plementing the theory simulates well-known data on the time course of hen and tsukuri reaction times were identical in a visual tnatching task, object naming within the picture-word interference paradigm. Novel while tsukuri reaction time was faster than that of hen in a phonological predictions were experimentally tested and validated. task. The results suggest that radicals are parallel processed in the visual task and sequentially in the phonological task. (251) (248) Research Support from the National Science Foundation. FRED Computing Agreement In Sentence Production. GABRIELLA STOLLNITZ, NSF Program Director for Animal Behavior, & JOSEPH VIGUOCCO, University of Trieste, BRIAN BUTrERWORTH, Univer- YOUNG, NSF Program Director for Human Cognition and Perception. 494 Papers 252-262 Saturday Morning DEVELOPMENTAL PROCESSES I that a stimulus (S - ) that signals the nonreinforcement of an instrumen Embarcadero AB, Saturday Morning, 8:00-9:45 tal response provides information about the identity of the omitted out come. An S - trained with one outcome preferentially transferred to Chaired by Susan Kemper, University of Kansas a new response trained with the same outcome (Experiment I) and was slower to develop discriminative control over a new response reinforced 8:00-8:15 (252) with the same outcome (Experiment 2). In Experiment 3. an S - elevated Developmental Differences in Holistic Processing of Faces? SUSAN performance of a response trained with the same outcome when that CAREY, MIT-A composite photo of the top half of one person's face outcome had been devalued. and the bottom of another's fuses into a new person from which it is 8:20-8:35 (258) difficult to recover the identity of either half alone. This composite inter ReversibiHty of Wrthln-Incentive Selective A!IIIIlCiatioDi in AppetItive ference effect occurs only for upright faces, and is robust at ages 6, 10, and in Aversive Situations. LEIGH V. PANLILIO & STANLEY J. and adult. Nonetheless, adults are markedly more affected by inversion WEISS, American University, Washington. D.C. (read by Stanley J. overall than are children. Two senses of configural encoding of faces Weiss)-Barpressing in tone-plus-light (T + L) avoided shock for two are distinguished, only one of which undergoes developmental change. groups of rats. When tone-and-Iight absence (T+L) was shock free, 8:20-8:35 (253) tone was selectively attended to during a stimulus~lement test. How A Passover Effect in Cblldren's Word Learning. WILLIAM E. ever, when unpredictable unavoidable shocks and punishment was MERRIMAN, JOHN MARAmA, & LORNA H. JARVIS, Kent Stale programmed in T+ L. light controlled barpressing. Switching T+ L con University-Our goal was to determine whether young children who ob ditions over groups reversed modality control. When selective associa served a speaker pass over an object while introducing a new name for tions were produced solely with appetitive baselines, similar reversals similar objects would infer that the name did not apply to the passed occurred when contingencies were switched over groups. This within over object. Three-year-olds, but not 2-year-olds, tended to infer this. incentive reversibility strengthens the position that selective associations These results suggest that this inductive tendency is another source of are contingency rather than class-of-reinforcer determined. preschoolers' remarkable skill in word meaning acquisition. 8:40-9:00 (259) 8:40-8:50 (254) Contextual Conditioning and Performance in Appetitive Condi The Spacing Effect and Young Cblldren's ImpHcit and ExpHcit tioning. MARK E. BOUTON, CHARLES A. PECK, & DOUGLAS C. Memory. MELINDA Y. SMALL & ANTOINETI'E E. KAVANAUGH, BROOKS, University of Vemwnt-Contextual conditioning promoted Bowdoin College-Implicit (word production) and explicit (recall) mem performance to an extinguished CS but had less effect when the CS was ory were compared for preschoolers and first-graders. Recall improved merely paired or unpaired with food. When USs were added to the ms with age. Implicit memory occurred at both ages, but without an age of extinction, the effect was abolished. USs added to either condition effect. Spacing of repeated items had an effect on recall for both ages. ing or extinction could acquire the ability to cue the corresponding per There was no spacing effect on the implicit memory task. Comparisons formance. Contextual conditioning has complex effects that are simpli of incidental and intentional learning conditions indicated no effects on fied by recognizing that it, and the USs that produce it, can acquire memory. Theoretical implications are discussed for memory develop discriminative control over performance. ment and the spacing effect. 9:05-9:25 (260) 8:55-9:15 (255) Separate Effects of Aversion to Visual and Taste Cues on Drink Short-Term Memory for Item, Temporal, and Spatial Informa Ing Latencies: Bdongi"lJleso RecomIdered. JOSEPH J. FRANCHINA. tion in Young and Elderly. ROBERT E. TILL, University of Nonh P. ANDREW LEYNES. & LERA JOYCE JOHNSON, Virginia Poly Dakota, ALICE F. HEALY, University of Colorado, THOMAS F. technic Institute and State University-Presence of familiar visual cues CUNNINGHAM, St. Lawrence University, & LYLE E. BOURNE, JR., during prior aversion conditioning to a novel-tasting solution reliably University of Colorado-Short-term memory was examined for names, decreased the latency to start drinking, relative to that for cOnditioning temporal order, and spatial location of letters. Subjects saw three let to a familiar taste in the presence of novel visual cues. Latencies to com ters displayed one at a time, followed by digits. Each appeared arbitrarily plete nine subsequent drinking responses were similar for conditioning on the left, right, or center. Some subjects read aloud each letter; others to novel visual or taste cues and were reliably shorter than those for announced the letter's location. Age differences appeared in all combi conditioning to novel cues combined. Measures of intake and hedonic nations of encoding and test except one: no age effect occurred when expression showed reliable conditioning effects. subjects had named items at encoding and recalled only letter names 9:30-9:40 (261) at test. Failure to F1nd Keypnoas Codra!it for I\mt Reinforan. FRANCES K. 9:20-9:40 (256) McSWEENEY & CAM L. MELVILLE, Washington Stale University Fuzzy-Trace Tbeory and Developmental Stability of Reminiscence. No contrast occurred when rats pressed keys for milk reinforcers for CHARLES J. BRAINERD, University of Arizona-In long-term reten five component durations and two baseline schedules. The failure was tion designs, three variables may exhibit development trends: acquisi not caused by poor discrimination, a floor or ceiling, fatigue. satiation. tion rates, forgetting rates, and reminiscence rates. Developmental trends insensitivity to reinforcement, or a change in response fonn. These results in acquisition rates have long been well established and, more recently, imply that differences in response rates and in the part of the body used evidence of parallel trends in forgetting rates has become firm. As yet, did not produce differences in contrast among other responses, such as little is known about the development of reminiscence, though fuzzy keypecking and treadle-pressing. trace theory anticipates that age changes should be minima\ because reminiscence involves reconstruction from stored gist. We review evi REPETITION/PRIMING EFFECTS n dence from several experiments which suggests that between early child San Francisco A, Saturday Morning, 8:00-9:40 hood and adolescence, at least, reminiscence rates are remarkably stable. Chaired by Joan Gay Snodgrass, New York University ASSOCIATIVE LEARNING: ANIMAL n Embarcadero CD, Saturday Morning, 8:00-9:45 8:00-8:20 (262) Subliminal Semantic Activation between Objective and Subjective Chaired by T. James Matthews. New York University 'IbreshoIck. ANTHONY G. GREENWALD, UnivemtyofWashington Regression analysis was used to describe the relation between detect 8:00-8: 15 (257) ability of dichoptically masked priming stimuli and magnitude of their Associative Structure of Negative Discriminative Stimuli. RUTH M. priming effects. Data from two experiments indicated that this regres COLWILL, Brown University-Three experiments using rats showed sion function passes through the origin (i.e., no priming with no detec- 495 Saturday Morning Papers 263-274 tion). These fmdings are consistent with the subjective threshold thesis conjunctions, parenthetical remarks, and hesitation phenomena (filled of Cheesman and Merilde-that effective priming stimuli are detectable, pauses, repeats, and false starts). Other errors occurred with diminish even when subjects deny seeing them. ing frequency: additions, substitutions, and sequence reversals. Sources 8:25-8:40 (263) of these systematic transcriptional errors are discussed. Suppression oC Irrelevant Ambiguous Word Meanings in tbe 8:35-8:50 (269) Cerebral Hemispheres. CHRISTINE CHIARELLO, KIM CANNON, Causal Inferences in Comprebension: Does Syntax Playa Role? LORIE RICHARDS, & LISA MAXFIELD, Syracuse University-We MICHAEL E. YOUNG & CHARLES R. FLETCHER, University of investigated hemisphere-specific processes in the suppression of am Minnesota (read by Charles R. Aetcher)-Previous research suggests biguous word meanings. Homograph primes were shown twice using that causal inferences playa major role in story understanding. We lags of I or 6 intervening trials. Second presentation targets were unique, present two experiments that evaluate the role of world knowledge and being related to the same meaning as the first homograph-target pair, syntactic cues in readers' ability to make these inferences. Both experi an alternate meaning, or unrelated. Target pronunciations were slowed ments reveal that a reader's sensitivity to the causal structure of a text when a new meaning was instantiated, and this suppression varied over is diminished if that text is modified to remove the availability of either visual fields and lag. Implications for hemisphere differences in am syntactic cues or appropriate world knowledge. biguity resolution are discussed. 8:55-9:10 (270) 8:45-9:00 (264) The DefInite Article the FadHtates the Proces'l oCMapping. MORTON Variation oC Attention Predicts Repetition Blindness in Rapid Serial ANN GERNSBACHER & RACHEL R. W. ROBERTSON, Univer Visual Presentation. ELISABETH M. FINE & ADAM REEVES, sity of Oregon-The definite article the is one way to signal referential Northeastern University (read by Adam Reeves)-Random high coherence in English. For example, compare the sentences A man bought frequency words were shown in RSVP to study repetition blindness (RB) a dog and A boy was delighted with a gift with the sentences The man (Kanwisher). We explain RB using the "attention gating model" (AGM) bought the dog and The boy was delighted with the gift. Using a variety (Reeves & Sperling) of the perception of temporal order in RSVP. In of laboratory tasks, we demonstrated that referential coherence-as sig AGM, RB occurs when two presentations of the same item enter visual naled by the definite article the-facilitates the process we call "map short-tenn memory together. AGM predicted the amount of RB across ping" during discourse comprehension. the RSVP stream and also the effects on RB of an attention-grabbing cue. 9:15-9:35 (271) 9:05-9:15 (265) Context and the Resolution oC Ambiguous Idioms. LUCIA Orthographic Versus Phonological Similarity in Repetition Blind COLOMBO, University ofPadua-The influence of context on the reso ness. NANCY KANWISHER, UCLA-Repetition blindness is the lution of ambiguous idiomatic expression was investigated in two ex difficulty of reporting two items from a rapid display when they are periments with the self-paced reading paradigm. The context was neu similar in both orthography and phonology (Kanwisher, 1987) or just tral, or biased the literal or the figurative interpretation of the idiom. phonology (Bavelier & Potter, in press). Is all RB phonologically medi The idiom was followed by a disambiguating clause that always referred ated? In the present experiment, RB occurred to the same extent for to the literal interpretation of the idiomatic string. Reaction times to the word pairs with primarily phonological (certify-sir) or primarily ortho idiom and the post-idiom regions indicated whether either one, or both graphic (right-rig) similarity. Apparently, either phonological or ortho interpretations, had been activated and constructed. graphic overlap alone is sufficient to produce the full effect. 9:20-9:35 (266) ATTENTION m Effects oC Repeating the Same Relation on Relatedness Decision Golden Gate A, Saturday Morning, 8:00-9:50 Times. ROGER CHAFFIN & ROBERTA KELLY, Trenton State College-Subjects decided if noun:noun pairs were related or not. We Chaired by Colin M. MacLeod, University of Toronto varied relational similarity of related pairs to related pairs that immediately preceded them. Response time increased with similarity of the relation 8:00-8:20 (272) of preceding pairs. The decision criterion for relatedness reflected expec False Recognition, Attention, and Perception without Awareness. tations about the type of relation; decisions were faster when expectations STEVE JOORDENS & PHILIP M. MERIKLE, University of Water were more accurate. Subjects spontaneously preferred to look for specific loo (read by Philip M. Merilde)-Jacoby and Whitehouse (1989) found types of relation rather than for a general property of relatedness. that different patterns of false recognition occur following perception with and perception without awareness. We show that the "unaware" LANGUAGEIDISCOURSE PROCESSING n pattern of false recognition also occurs when attention is divided dur San Francisco B, Saturday Morning, 8:00-9:40 ing the recognition test and that the "aware" pattern of false recogni tion occurs when attention is focused during the recognition test on the Chaired by Sam Glucksberg, Princeton University critical stimuli. These results provide evidence for a link between at tention and awareness. 8:00-8:15 (267) 8:25-8:45 (273) Message Level Influences on Lexical Access. SUSAN A. DUFFY, The Fate oC Neglected Targets in the Callosotomized Brain: A Amherst College-In research on the effect of sentence context on lexi Chronometric Analysis. PATRICIA A. REUTER-LORENZ, Dart cal access, a critical issue is whether the message level representation mouth College and Medical School, GEORGE NOZAWA, Dartmouth influences access. This issue was addressed using a naming task in which College, & HOWARD C. HUGHES, Dartmouth College and Medical the target word was preceded by a sentence context. Within the con School (read by Howard C. Hughes)-The speaking left hemisphere of texts, critical lexical items were held constant while their role in the callosotomy patients can report the occurrence of a unilateral left or message was varied. Results will be discussed in terms of their impli right visual field light. L VF report accuracy drops by 34% with bilateral cations for modular and interactive models of access. presentations (simultaneous extinction). Paradoxically, simply unimanual 8:20-8:30 (268) or vocal reaction times to bilateral targets are significantly faster than Errors in Transcription. DANIEL C. O'CONNELL, Georgetown unilateral target RTs, demonstrating a robust redundancy gain. We dis University, SABINE KOWAL, Teclrnische UniversitIJI Berlin, ROBERT R. cuss the potential role of attention and response competition in mediat DORIN, Gonzaga University, & NATHALIE JAUMOUILLE, Tech ing redundant target effects in callosotomized as compared to normal nische UniversitIJI Berlin-Recent evideoce points to limitations in tran SUbjects. scribers' ability to accurately reproduce spoken discourse. Recordings 8:50-9:10 (274) of interviews of native speakers of German were transcribed (with un Dual-Task Interference and the Cerebral Hemispheres. HAROLD limited access) by native speakers of German (untrained transcribers). PASHLER & SHANNON O'BRIEN, University of Colifornia, San Nearly half of the errors were omissions of noninforrnative material: Diego-Previous studies have suggested that the two cerebral 496 Papers 275-285 Saturday Morning hemispheres constitute separate' 'resource pools," or that dual-task in 9: 15-9:35 (280) terference depends upon "cortical distance." However, these proposals Multiple Detector Models of Simple Reaction Time. PHILIP L. were based on dual-task studies that did not require two separate response SMITH, University of Otago-A class of models of visual simple reac selections-the operation often found to constitute a processing bottle tion time (SRT) is developed, in which the sensory processes are neck. We combined tasks requiring independent response selections, represented as diffusion processes with time-varying drifts. Within this with 1ateralized stimuli and responses. There was no evidence that direct framework, two current issues in SRT research are examined: namely, ing tasks to separate hemispheres allowed response selections to pro whether SRT is mediated by separate change and level detectors and ceed simultaneously. how information is combined across redundant targets. Model fits favor 9:15-9:30 (275) an account in which information combines coactively across stimuli for Failure to Find Generally Reduced Inhibition in Old Age. JAMES independent parallel change and level detectors. KIELEY, Pitzer College, & ALAN HARTLEY, Scripps College (read by Alan Hartley)-In testing the hypothesis that inhibitory functioning NEURAL NETWORK MODELING is generally impaired in old age, we found that inhibition of return and Embarcadero AB, Saturday Morning, 9:55-11:40 Stroop interference combined additiveIy in young adults but subaddi tively in older adults: Stroop targets had a smaller effect at an inhibited Chaired by Martha J. Farah, Carnegie-Mellon University than at an uninhibited location. At press time, we are not yet able to explain this, but we have ruled out eye-movement artifacts and the pos 9:55-10:10 (281) sibility that the inhibition particularly impacts semantic information. Connectionist Modeling of Multidimensional Generalization. 9:35-9:45 (276) ROGER N. SHEPARD, Stanford University, & JOSHUA TENEN "Effortless" Texture Segmentation and "ParaIlel" Visual Search BAUM, Yale University-We explore multidimensional extensions of Are Not the Same Thing. JEREMY M. WOLFE, MIT-If a texture the connectionist explorations of unidimensional generalization and clas defined region is effortlessly perceived, the texture segmentation is labeled sification begun by Shepard and Kannappan (Psychonomic meeting, "preattentive." If search RT for a target among distractors is indepen 1990) and based on the theory of generalization proposed by Shepard dent of set size, the search is labeled "parallel." Preattentive texture (Psychonomic meeting, 1984; Science, 1987). The connectionist architec segmentation and parallel search have been assumed to be equivalent tures we explore are chosen (a) to be neurophysiologically plausible and measures of the same parallel processing stage. I will demonstrate that (b) to account for the functional form of generalization and classifica this is not the case. Parallel search can occur with stimuli that do not tion in the contrasting cases of psychologically integral and separable support effortless texture segmentation and vice versa. stimulus dimensions. 10:15-10:30 (282) INFORMATION PROCESSING n Aperiodic Components Analysis of Evoked Potentials in Copitive Golden Gate B, Saturday Morning, 8:00-9:40 Tasks. BRUCE L. BROWN , DONOVAN E. FLEMING, MICHAEL J. ANDERSON, & ALLEN C. SCHOFIELD, Brigham Young University Chaired by Jerome Busemeyer, Purdue University Evoked potentials were obtained for four cognitive tasks (math, syno nyms, figure discrimination, and lateral tone localization) for eight sub 8:00-8:20 (277) jects. Raw waveforms were processed with aperiodic components anal Selective Attention under Speed Stress. ALLEN OSMAN & ysis. Three components accounted for 75% and 79% in two replications LlANGGANG LOU, University ofCalifornia , San Diego-We examined of the study. The resultant components were analyzed with a five-way the effects of speed stress on distractibility. Subjects made two-choice MANOVA (task type, hand of response, cerebral location, cerebral manual reactions in response to a target letter surrounded by irrelevant hemisphere, and individual subject). Significant differences were found flankers. Flankers were either identical to the target or letters signaling for task type, subject, and task-by-subject interaction. the alternative response. Instructions emphasized either speed or ac 10:35-10:50 (283) curacy. We measured reaction times, eIectromyographic activity, and Interference and Generalization in DIstributed Memory Models. event-related brain potentials. The results bear on the level to which STEPHAN LEWANDOWSKY, University of Oklahoma-An attrac unattended stimuli are processed and how speed is traded for accuracy. tive property of distributed memory models, or neural networks, is their 8:25-8:45 (278) ability to generalize to new inputs. However, that capability also con A Respome TIme Theory of Separability and Integrality in Speeded tributes to their tendency to show unreasonably fast unlearning, or "catas Classification. F. GREGORY ASHBY & W. TODD MADDOX, trophic interference ... The relationship between generalization and un University of California, Santa Barbara-A theoretical framework is learning is examined for a variety of network models, and it is shown developed for the speeded classification tests of perceptual separability that some solutions to the unlearning problem adversely affect gener and integrality, which were popularized by Gamer (1974). We show alization performance. that, if decisional separability holds, then the ftItering and redundancy 10:55-11:10 (284) tasks provide good tests of separability and integrality. However, if the A Neural Network Approach to Cognitive Mapping. NESTOR A. subject responds optimally, then redundancy gains and no interference SCHMAJUK & AARON THIEME, Northwestern University-We effects are likely, regardless of whether the dimensions are integral or present a real-time neural network that builds a topological cognitive separable. Experimental evidence supporting these predictions is map. The network combines recurrent and nonrecurrent properties that reported. allow the reading of the cognitive map without modifying it. The net 8:50-9:10 (279) work generates fast-time predictions of the remote future and real-time The Stop-Signal Paradigm: Old Models and New Data. HANS predictions of the present that are used to update the cognitive map. COLONlUS, Universitaet Oldenburg-Consider a simple reaction time Computer simulations portray latent learning and detour behavior in rats experiment where, in a certain percentage of the trials, the subject is and problem-solving tasks in humans. instructed (by presenting a stop signal after stimulus onset) to withhold 11:15-11:35 (285) the response. This stop-signal paradigm has been used to study basic MATHNET: A Distributed Model of Arithmetic Fact Retrieval. issues in automaticity and control. We compare several early and more MICHAEL McCLOSKEY & A. MARGRETHE LINDEMANN, Johns recent approaches to describe the underlying processing mechanisms Hopkins University-We present a connectionist model of the retrieval (Colonius, 1990; Logan & Cowan, 1984; Oilman, 1973) and present of basic arithmetic facts (e.g., 8x7=56), showing that the model gener some unpublished data from our lab. ates many of the phenomena observed in studies of normal and brain- 497 Saturday Morning Papers 286-296 damaged subjects (e.g., problem-size effects, error patterns, nonuniform CLUBB, ROBERT E. NIDA, University ofNonh Carolina at Chapel impairment). We also discuss some general issues arising in the context Hill, & LYNNE BAKER-WARD, Nonh Carolina State University (read of the model (e.g., to what extent do models of this sort interpret or by Peter A. Ornstein)-Three- and 6-year-olds' long-term recall of the explain, as opposed to merely reproducing, the phenomena of interest?). details of a physical examination was investigated. Children were in terviewed after 12 weeks, with subgroups receiving zero, one, or two ANIMAL COGNITION I additional assessments during this interval. Although immediate and Embarcadero CD, Saturday Morning, 9:55-11:35 delayed recall of the older children was superior, performance was im pressive at both age levels. The facilitative effects of repeated questioning Chaired by Ronald J. Schusterman, University of California, Santa Cruz and the contribution of prior knowledge about visits to the doctor are discussed. 9:55-10:10 (286) 10:15-10:30 (292) Knowledge of Ordinal Position by LIst-80phisticated Rhesus Mon Spilling the Beans on Cblldren's Comprehension and Production of keys. KARYL B. SWARTZ, Lehman College, CUNY. SHAOFU Idiolns. CRISTINA CACCIARl, University ofBologna , & M. CHIARA CHEN, & H. S. TERRACE, Columbia University-Rhesus monkeys LEVORATO, University of Padwl-The effects of context, frequency who had acquired at least 16 four-item lists (each consisting of novel of exposure, and semantic transparency on the comprehension and items) were trained on four lists consisting of items from previous lists production of idioms in children (ages 7 to II) are investigated in a set in which the ordinal position of each item, with respect to its original of experiments using different tasks (multiple choice, completion, and list, was maintained (2 lists) or violated (2 lists). Acquisition rates and paraphrase). The results challenge current hypotheses on idioms acqui error patterns supported the hypothesis that the original ordinal posi sition. A tentative model is proposed in order to account for the factors tion of each item influenced responding on derived lists. allowing children to overcome a literal-referential strategy and to develop 10:15-10:30 (287) a figurative competence. The Construction of Analogies by a Chimpanzee. DAVID L. 10:35-10:50 (293) ODEN, La Salle University, & ROGER K. R. THOMPSON, Franklin Aging and tbe Loss of Linguistic Complexity. SUSAN KEMPER, and Marshall College (read by Roger K. R. Thompson)-Analogical University ofKansas-Adults' language changes across the life span as reasoning by an adult chimpanzee (Pan troglodytes) was studied by per a result of limitations on their production of complex forms such as mul mitting her to construct analogies in whole rather than simply complet tiple sentence embeddings and cohesive devices. A five-year longitudi ing partial problems or judging those completed previously by an ex nal record of adults' language has now been analyzed using structural perimenter. She demonstrated her ability to produce analogies from a equation modeling with regards to identifying how age-related changes set of unarranged stimuli; the patterns of analogy construction and spon to verbal ability and working memory affect adults' language. taneous corrections of errors imply that the chimpanzee planned her anal 10:55-11:05 (294) ogies with foresight. Adult Age Differences in Prospective Memory. TIMO MANTYLA, 10:35-10:50 (288) University ofStocklwlm-Adult age differences in prospective memory Chimpanzee Communication 0: Dr. Whitmer and Peter. DOUG were examined in a laboratory paradigm, in which older and younger LAS CANDLAND, Bucknell University-Among Lightner Whitmer's adults generated associations to a large numher of words. Parallel with goals in founding the Psychological Clinic in Philadelphia at the turn this task, subjects performed an action when a semantically or phonem of the century was to test the ability of .. slum" children to determine ically defmed target word occurred. The results showed age-related dif the kind of education appropriate for them. While visiting Boston in ferences both in retrospective and prospective rememhering. The fmd 1908, he saw a performance by Peter, perhaps the fIrSt trained chim ings are discussed in terms of cue distinctiveness and self-initiated panzee to be seen in the U. S. Whitmer tested Peter's intelligence and retrieval processes. helped him learn the alphabet. This paper reviews the results of this 11:10-11:30 (295) early attempt at chimpanzee-human communication. Standardizing a Psychomotor Test Battery for Assessing Aging 10:55-11:10 (289) in Mice. DONALD K. INGRAM, JOHN M. HENGEMIHLE, Pigeons' Cognitive Maps Represent Relative Metric Space. JEFFERY LONG, & PAOLO GAROFALO, National Institute on DONALD F. KENDRICK, Middle Tennessee State University-Pigeons Aging-Our objective was to standardize a psychomotor battery that were trained to discriminate hetween two sets of pictures of a natural reflected aging in rodents. Using automated image analysis, we assessed location. One set (S + ) included scenes from a field of view from 254 0 motor performance across the life span (6-30 months) of male C57BLl6J to 272 0 • The second set included scenes from 290 0 to 308 0 (magnetic mice as follows: open-field locomotion in several apparatus; inclined north at 360 0). Generalization testing included pictures from 200 0 to screen ambulation; ability to remain suspended from a taut string; falls 360 o. Gradients showed the traditional inverted U -shaped functions, from a rotarod; and swim vigor. Varying in degree of linearity , signifi demonstrating that the pigeon's cognitive map accurately represents spa cant age-related decline in performance was observed in all tests ex tial extent and relative metric space. cept swim vigor. 11:15-11:30 (290) Landmark Use in Pigeons: Determining Direction by Global and Local Landmarks. KEN CHENG, University of Toronto-In one ex TACTILE PERCEPTION periment, pigeons learned to go to one of four walls of a square arena San Francisco B, Saturday Morning, 9:50-11:20 for food. They continued to choose this location in the room after a num ber of manipulations, suggesting the use of global landmarks in deter Chaired by Paul Beneison, Universite fibre de Bruxelles mining which direction is which. In a second experiment, pigeons learned to find food at a constant location in a circular tray. Manipulations of 9:50-10:00 (296) large and small nearby objects showed that nearby landmarks influenced Stress-Induced Analgesia in Humans? GARY B. ROLLMAN, the calculation of direction as well. University of Western Ontario, & LISA M. CARSWELL, University of Victoria-Thirty subjects were exposed to one of three stress condi DEVEWPMENTAL PROCESSES 0 tions: physical (vigorous exercise), cognitive (mental arithmetic, digit San Francisco A, Saturday Morning, 9:50-11:35 symbol substitution), or emotional (scenes from an arousing film). Pain threshold and tolerance were determined with a constant pressure al Chaired by Leslie B. Cohen, University of Texas at Austin gometer. Results indicated significant stress effects for the three condi tions but no analgesic effect. These data suggest that cognitive variables 9:50-10:10 (291) implicated in animal studies of SIA may also apply to humans. In par Visiting the Pediatrician: Long-Term Retention and Forgetting. ticular, perceived controllability and self-efficacy may override stress BETTY N. GORDON, PETER A. ORNSTEIN, PATRICIA A. per se in engaging endogenous analgesic systems. 498 Papers 297-307 Saturday Morning 10:05-10:20 (297) but consistent with other research on perception of rotated objects, Haptic and Cross-Modal Recognition in Children. EMILY W. representations remained viewpoint specific at both SOAs. BUSHNELL, Tufts University-Five-year-old children were given eight 10:20-10:35 (302) objects in succession to explore haptically. These were then mixed with Implicit Processing of Faces by Prosopagnosic Patients. JUSTINE eight distractor objects and recognition was tested. Objects were familiar SERGENT, Montreal Neurological Institute-Prosopagnosia is an ac or unfamiliar, and on the recognition test, they were presented hapti quired inability to recognize known persons by their faces as a result cally or visually. Haptic and cross-modal recognition with familiar ob of a breakdown at different levels within the face-recognition system. jects were remarkably good, as was haptic recognition with unfamiliar Yet operations on faces are still perfonned in prosopagnosics' brains objects (all> 90%). Cross-modal recognition with unfamiliar objects whereby a perceived face activates pertinent memories, but either these was not as good as in the other conditions (about 80%). The results sug operations cannot be completed or their outcome fails to reach cons gest that haptic recognition may be mediated by vivid memory traces ciousness. The functional locus at which faces are implicidy processed analogous to visual images and that cross-modal recognition may in was examined in 3 prosopagnosics, and the results indicated a different volve different cognitive processes than does within-mode recognition. pattern of perfonnance depending on the nature of the prosopagnosic 10:25-10:35 (298) disturbance and on the processing requirements of the tasks. Remembering Braille: Cross-Modal and Intramodal Effects. 10:40-10:55 (303) SLATER E. NEWMAN, North Carolina State University, ANTHONY D. Effects of Inversion on Configura! Processing of Normal and Dis HALL,IBM Corporation, & SHARON M. PULLEN, North Carolina torted Faces. JAMES C. BARTLETT & JEAN SEARCY, Univt!rsity State University-Braille symbols were presented for haptic or visual of Texas at Dallas-If the mouth and eyes of a face are inverted, the examination on a single study trial along with auditory presentation of altered construction appears grotesque when upright but not when upside their letter names. All subjects were then tested haptically both immedi down. Several studies of this "Thatcher illusion" using three types of ately and 48 h later. Forgetting occurred in both conditions. Performance configurally distorted faces, along with grotesque-expression faces, in the cross-modal condition exceeded that in the intramodal condition revealed dissociations between simultaneous paired-comparisons and to about the same extent on each test. Some explanations for these results similarity -difference ratings, and between difference and similarity rat will be discussed. ings' suggesting separate modes of face processing that are differen 10:40-10:55 (299) tially sensitive to configural information, and differentially affected by Modeling Perception of Temperature Change Using the General stimulus inversion. ized Additive Model. JOHN COTTON & ABDUL RAHMAN 11:00-11:15 (304) OTHMAN, University of California, Santa Barbara-Refmetti (1989) Imagery in Scientists: Galton's Breakfast QuestioIlll8ire Revisited. modeled the estimation of magnitude of thermal sensation as a power WILLIAM F. BREWER, University ofIllinois at Urbana-Champaign, function E = kSB, where E is the sensory estimate of warmth and S & MARLENE SCHOMMER, Wichita State University-In 1880 Galton is the stimulus magnitude, using linear regression on 52 subjects. We gathered data on recollective autobiographical memory with the "break believed that intrasubject variabilities could perhaps partially be accounted fast questionnaire." He stated that scientists have "feeble powers of for by period and carryover from stimuli in previous periods. Data from visual representation." We gathered contemporary data from scientists 3 subjects in Refmetti's experiment were used. The generalized addi and nonscientists on Galton's task and found that both groups showed tive model was used to obtain a nonlinear equation of the fonn consistent imagery reports. A reanalysis of Galton's own data shows Y = B.+BIJ,.(XI) +B,J,.(X,) +B, J.(X,) , where Yis the estimate of warmth that his findings are consistent with our data in showing that scientists and fl (XI), J,.(X,), and fiX,) are nonlinear cubic spline functions of the do not have a deficit in imagery during recollective memory tasks. period, stimulus, and carryover effects, respectively. Computations of 11:20-11:40 (305) the estimates of the coefficients and residuals were done on the GAlM Evidence for Syntactic Analysis of Visual Patterns. RICHARD computing package. When modeled on individuals, period effects were CHECHILE & JANE ANDERSON, Tufts University-Subjects in three significant across all 3 subjects and carryovers were significant on 2 experiments learned abstract visual patterns which differed in syntactic subjects. complexity. After learning the patterns (i.e., at least three perfect, con 11:00-11:15 (300) secutive, paired-associate trials), the subjects were placed in a speeded Tactile Attention and Response Competition. PAUL M. EVANS, recognition test. In each study, there was a large (approximately Willamette University, & JAMES C. CRAlG,Indiana University (read 200 msec) effect associated with increased syntactic complexity for tar- by James C. Craig)-Previous studies have demonstrated that the iden gets and a larger (approximately 700 msec) effect for foils. These ef tification of a moving, tactile stimulus (target) can be interfered with fects are taken as an indication of preaccess syntactic analysis for visual by a second, equally intense tactile stimulus (distractor). The present patterns. study showed that target identification is interfered with only if the dis tractor has a different learned response from the target. These results HUMAN LEARNINGIMEMORY n provide evidence that tactile distractors are processed to the level of Golden Gate B, Saturday Morning, 9:50-11:30 incipient response activation even when subjects are instructed to at tend only to the target. Chaired by Michael P. Toglia, SUNY at Cortland 9:50-10:10 (306) PICTURE MEMORY/PROCESSING Implicit Learning: Within- and Cross-Modality Transfer of Tacit Golden Gate A, Saturday Morning, 10:00-11:45 Knowledge. LOUIS MANZA & ARTHUR S. REBER, Brooklyn Col lege and the Graduate Center, CUNY (read by Arthur S. Reber)-Four Chaired by Irving Biederman, University of Southern California transfer experiments were run using an artificial grammar learning task. The stimuli during training and testing differed in their surface forms 10:00-10:15 (301) but retained underlying structural commonalities. In Experiments 1 and Visual Object Representations: Holding on to an Old View. 2, the letters used to instantiate the grammar changed; in Experiment 3, STEPHEN J. LUPKER & G. KEITH HUMPHREY, University of the modality of presentation changed; in Experiment 4, both physical Western Ontario-Ellis and Allport (1986) proposed a model of object fonn and modality changed. Subjects were able to transfer tacit knowl perception wherein successively more abstract descriptions are gener edge in all cases, supporting the argument that tacit knowledge is likely ated as a function of processing time. To test the model, our subjects represented in an abstract fonn. decided whether successively presented pictures (identical, same name, 10:15-10:30 (307) or different) rotated in the frontoparallel plane had the same name. SOAs Effects of Perceptual Interference on Explicit Memory. ELLIOT were short (600 msec) or long (2,500 msec). Inconsistent with this model HIRSHMAN, University of Nonh Carolina at Chapel Hill-Nairne 499 Saturday Morning Papers 308-31 0 (1988) and Hirshman and Mulligan (1991) demonstrated that interfer BLANCHARD-FIELDS, LISA NORRIS, & LEWIS G. ROUSSEL, ing with perceptual processes at study improves later memory. A plau Louisiana State University-Following implicit/explicit training, sub sible explanation of this finding is that interfering with an item's per jects performed a string discrimination test. The letter set was changed ception increases the elaborative rehearsal the item receives. We tested midway through test preserving granunatical structure. Some abstract this explanation by varying presentation duration and, consequently, re knowledge was obtained when a fInite-state grammar was learned im hearsal time during study. The results of our experiments were not con plicitly or explicitly. Performance either stayed the same or improved sistent with the elaborative-rehearsal explanation. when letter set was changed using biconditional grammar. Feedback 10:35-10:50 (308) during string discrimination was not critical for obtaining transfer to Contextual Support ollmplicit and Explicit Memories in Older a new letter set, suggesting that an analogy mechanism was not neces Adults. DENISE C. PARK, University of Georgia, RAYMOND J. sary for transfer. SHAW, West Virginia University, & DAVID FRlESKE, University of Georgia-We examined the role of contextual support on implicit and 11:15-11:25 (310) explicit memory in young and old subjects. Despite a large N, no age Restoration 01 the Incongruity Effect. THOMAS J. AYRES, Failure effect occurred on the implicit task. The effect of contextual support Analysis Associates. Menlo Park-Words generated in response to a (number of letters in stem at retrieval) was equivalent for young and negative cue are generally recalled better than those with a positive old on implicit and explicit tasks. These findings suggest that data- 500 Posters 311-322 Saturday Noon POSTER SESSION n display. Perceived eye level was biased in the direction of box pitch; Pacific Concourse East, Saturday Noon, 12:00-1:30 the magnitude of the bias depended on the pattern of optical structure of the box. Results are discussed in terms of an oculomotor theory of (311) spatial stabilization. Perceptuallotegration of Shape and Texture Dimensions during (317) Haptic Processing. CATHERINE L. REED, Carnegie-Mellon Univer Inftuences of Closure and Occlusion on the Perception of Frag sity (sponsored by Roberta L. Klatzky)-Experiments investigated haptic mented Figures. JAMES M. BROWN & CHRISTOPHER KOCH, integration of shape and texture at perceptual levels of processing. The University of Georgia (sponsored by Milton H. Hodge)-Two experi haptic system combines tactile and kinesthetic information. Perceptual ments examined recognition of a repeated figure in stimuli where only separability and independence were assessed for planar and 3-D stimuli fragments of the figures were available. Fragments were either closed using a speeded classification paradigm and quantitative tests developed or open regions presented either with or without an occluder filling the by Ashby and Maddox (in press). Results indicated that shape and tex spaces between fragments. Recognition performance was superior for ture were perceptually separable yet highly dependent. Hand-movement open fragment stimuli, with or without an occluder. Experiment 2 com analyses showed optimized exploration for the simultaneous apprehen pared equiluminance and luminance open/occluded and open/unoccluded sion of both dimensions. stimuli. Recognition times were faster with an occluder and slower for (312) equiluminance stimuli. Surface Filling-in Is Not Directly Related to Contrast Mechanisms. (318) BlRGITIA DRESP, Universite Paris V-CNRS (sponsored by Claude Size Contrast as a Function of Conceptual Similarity. STANLEY Bonnet)-Some theories relate surface brightness to antagonistic con COREN & JAMES T. ENNS, University of British Columbia-Mag trast mechanisms, assuming that they directly generate a summative nitude of the Ebbinghaus illusion is known to vary as a function of the filling-in process. We present psychophysical evidence that such a visual similarity between the central (test) element and surround (in "filling-in" must occur at stages of processing which integrate more ducing) elements. Two experiments show that illusion magnitude also complex structural inputs than merely contrast. Increment thresholds, varies with conceptual similarity. Objects served as test stimuli, with generally assumed to reflect the output of antagonistic contrast mecha larger or smaller objects as inducers. Dlusion magnitude was greatest nisms, do not correlate with the intensity of surface brightness in the for visually identical inducers, next highest for same and similar con Kanizsa square measured with matching and rating procedures. ceptual class inducers, and least for inducers from a different concep (313) tual category. Lacuoarity: Fractal Dimension Alone Is Not Sufficient for Tex (319) ture Discrimination. FRANK M. MARCHAK, TASC (sponsored by Perceptua1 Set and Neural Fatigue Effects on Two Reversible Fig George L. Wolford)-Marchak (1989, 1990) showed that fractal dimen ures. GERALD M. LONG, GREGORY W. MONDIN, & THOMAS C. sionality accounted for subjects' ability to discriminate textures. Using TOPPINO, Villanova University-Observers reported the initial con higher resolution stimuli, it was found that fractal dimension alone is figuration and subsequent reversals of a reversible figure over a 3O-sec not sufficient to predict discrimination; it is possible to have textures testing interval following variable adaptation to an unambiguous ver with the same fractal dimension but different visual appearance. Lacunar sion of the figure. For a rotating 3-D Necker cube and a 2-D cube draw ity is a fractal parameter that describes the distribution of mass in a fractal ing, observers were more likely to report the reversible figure in the set. Fractal dimension and lacunarity together provide a more accurate same configuration as the adapted figure following brief adaptation texture descriptor. periods and in the opposite configuration following long adaptation (314) periods. Cognitive set and neural fatigue are implicated. Laterality Effects in tbe Perception of Relative Frequency in Au (320) dition. RICHARD IVRY & PAUL LEBBY, University of California, The Time Course of the Global Precedence and Consistency Ef· Berkeley-Laterality effects in auditory perception were explored. Sub fects. JAMES G. MAY, CAROLINA GUTIERREZ, & CHARLES A. jects were monaurally presented with two pure tones. The frequency HARSIN, University ofNew Orleans-The global precedence and con of one of the two tones was varied. For decrements in frequency, per sistency effects were examined with an RT paradigm which delayed the formance was better when the stimuli were presented in the left ear/right global information relative to the local information. With a delay of hemisphere. Conversely, a right earlleft hemisphere advantage was found 80 msec, the global precedence effect was obliterated, but the consistency when the frequency was increased. The results converge with recent effect was still significant. This suggests that the two effects may derive laterality findings in visual perception and account for disparate results from different neural loci. in the study of music perception and language. (321) (315) Incompatible Eye Movements DIsrupt Spatial EococIing. DALE S. Assimilation, Contrast and Figure/Ground. SUSAN PETRY & KLOPFER, Bowling Green State University (sponsored by Michael E. ALAN COOPER, Adelphi University-While the mechanisms causing Doherty)-Research indicates that eye position information is not needed brightness contrast are relatively well understood, its converse, bright for integration in spatial working memory to occur. Yet, with patterns ness assimilation, is rather poorly understood. We present experiments shown sequentially so that the first piece is to be remembered to the investigating the relationship of figure-ground to assirni1ation. Black/gray left of the second piece, and so forth, pattern memory is better when or white/ gray square-wave gratings were manipulated by changing spatial the pieces are presented from left to right than when presented randomly. parameters, surrounds, flickering, apparent motion, and eccentric view It appears that incompatible eye movements disrupt spatial encoding and ing to alter the region seen as figure. The results suggest that contrast that eye position information aids spatial integration. is a property of figure, assimilation of ground. Attentional and physio (322) logical models are discussed. The Relationship between Spatial Organization and the Perce~ (316) tion of Rotational Motion. JOHN R. PANI & SUSAN J. HESPOS, Optic Bias of Perceived Eye Level Depends on Structure of the Emory University (sponsored by Ulric Neisser)-We investigated per Optic Array. KENNETH NEMIRE & STEPHEN R. ELUS, NASA ception of 3-D geometric forms undergoing various rotations. Basic im Ames Research Center (sponsored by Arnold E. Stoper)-Head-relative plications are: (1) Comprehension of rotational motion can be conceived visual direction is determined by a combination of retinal and extra as the recognition of an implied spatial organization centered on the axis retinal information. We manipulated visual information by presenting of rotation; when that organization is aligned with the object-centered pitched virtual boxes, of different optical structures, in a head-mounted or the environmental reference systems, the structure of the rotation 501 Saturday Noon Posters 323-335 becomes apparent. (2) Type of rotation affects kinetic depth for spatial (329) organization. (3) Rotations aligned with object-centered reference sys A Line Facilitates the Perception of a Second Line. DONALD L. tems reinforce those systems. KING, HESTER HICKS, & PAMELA D. BROWN, Howard Univer (323) sity-The stimuli were a vertical line connected to a horizontal line (VH), Converging Evidence on Sex Differences in Spatial Ability. a vertical line alone (V), a horizontal line alone (H), and a blank tachisto THOMAS G. BEVER, University ofRochester-A series of new studies scope card. V was always 10 mm. Stimulus duration and then intensity confirms and expands the hypothesis that men and women are gener were decreased progressively. When H was 2 mm, VH was identified ally equal in the ability to learn to negotiate new areas. The studies also more accurately than H, but when H was 10 mm, VH and H were iden show that the previously reported differences are a function of greater tified equally accurately. Therefore, V facilitated the perception of the relative utilization by females of egocentric than geocentric cues: both 2-mm H. kinds of cues require configurational and recognition computations. The (330) spatial difference may underlie other kinds of cognitive sex differences The Face-Detection meet: Semory and CogDitive Factors. DEAN G. that will be discussed. PURCELL, Oakland University, ALAN L. STEWART, Stevens Insti (324) tute of Technology, & ERIC J. HIRIS, Vanderbilt University-Normal The Geometrical Bases of Reasoning about an Object's Orienta faces have lower detection thresholds than either upside-down or rear tion. LAWRENCE M. PARSONS & CHARLES H. SCOTT, Univer ranged faces . We show that both sensory and cognitive factors deter sity of Texas at Austin-The geometrical bases of people's imagined spa mine when this face-detection effect (FOE) will occur. The FOE is not tial transformations is unknown. Our results indicate that even high-spatial found if the stimuli contain dark blobs or have a high contrast facial ability individuals are generally quite poor at rotating an object about outline. Furthermore, the FDE is eliminated when thresholds are de an arbitrary axis and at perceiving the unique axis and angle of rotation termined for both faces and objects. Although the FDE is eliminated, separating two orientations of an object. We suggest how this and other the object-detection effect is not. recent evidence weighs against the idea that people mentally represent (331) minimum-length trajectories in the Cartesian space in which an object Effects of Early Visual Features on Pattern and Object Percep is oriented. tion. THOMAS SANOCKI, University ofSouth FloritkJ (sponsored by (325) Douglas L. Nelson)-The idea that recognition involves a contingency, A Test for Models of Motion Processing: Nonrigidly Perceived in which later processing is modified contingent on earlier analyses, Rotating FJHpses. MAGGIE SHIFFRAR & JEFFREY B. MULUGAN, was distinguished from bottom-up feature models. Common-feature shape NASA-Ames Research CenJer (sponsored by Jennifer J. Freyd)-We in primes were presented briefly, followed immediately by a target pat vestigated the perceived nonrigidity of rotating ellipses under a variety tern, and then a mask and forced-choice alternatives. Primes facilitated of conditions, following Wallach, Weisz, and Adams (1956). We mea choices between same-shape alternatives even though, according to fea sured the degree of perceived rigidity as a function of aspect ratio and ture models, the primes provided no discrimination-relevant informa found that a number of factors other than aspect ratio can profoundly tion. Experiments with common objects will also be reported. affect the percept. We compare the results with the predictions of ex (332) isting models of motion processing and examine how they might be ex Is Auditory Word Recognition Serial or Interactive? MARK A. tended to generate categorical judgments of rigidity . PITT, Ohio State University, & ARTHUR G. SAMUEL, SUNY at Stony (326) Brook (sponsored by Neal Johnson)-Previous research examined this The musions of tbe Veil. GLENN E. MEYER & JAMES question by having subjects monitor for phonemes at locations surround McMULLEN, Lewis and Clark College-If an intermediate gray tar ing a word's uniqueness point (phonetically unique measuring from word get moves along a ramp of light to dark, perceived lightness changes. onset). When compared with a nonword control condition, the results However, motion causes this change to be interpreted as a change in favored serial models. However, the beginnings of the non words formed a lighting source. As targets cross the isoluminant point, a veil of mist the beginnings of legal words, possibly diminishing any word advan is perceived which the target must penetrate. If targets, which are also tage. We controlled for this problem by creating nonwords with early ramped in brightness, are moved across ramps, subjects perceive shadows nonword points and found data in support of interactive models. and illumination changes sweeping across the targets' 3-D structure. (333) (327) Extensions of a Witbin-Trials Manipulation of Visual Quality as Dissociation and Self-Reports of Eating, Dieting, and Weight. a Means of Testing for Serial and Parallel Processing. DALE ELLEN F. ROSEN, College of William and Mary, & LINDA C. DAGENBACH, Wake Forest University, & HOWARD EGETH, Johns PETTY, Hampton University-Sanders (1986) defines dissociation as Hopkins University (sponsored by Howard Egeth)-The within-trials a personality trait which modifies links between affect, cognition, be visual quality diagnostic for parallel and serial processing described previ havioral control, and perception. She found that bingeing college stu ously by the authors is further developed to apply to varying display dents score high on the Perceptual Alteration Scale (PAS). However, sizes and to target-present trials. These extensions provide a potential PAS only measures affect and control (Fischer & Elnitsky, 1990). In means of distinguishing between serial and at least some limited-capacity this study, college women responded to both the PAS and the Dissocia parallel models of visual search. The results of the application of the tive Experience Scale (Bernstein & Putnam, 1986), which measures the extended diagnostics to new experiments are then described. cognitive dimension, two standard measures of eating disorders and ques (334) tions about dieting and weight. Express Saccades during the Encoding of Text? ALBRECHT (328) WERNER INHOFF & RICHARD TOPOLSKI, SUNY at Binghamton Perception of Symmetry: The Role of Features. SEONGHEE Reading, letter detection during reading, and copytyping require the en HONG, Stanford University, & M. PAVEL, New York University coding of text via a sequence of saccades and fixations. Effects of task (presented by M. Pavel)-Symmetry has been considered to be an im demands on fixation durations (saccade latency) were examined to de portant global feature facilitating object perception. Current models of termine whether visuospatial attention during the encoding of text af symmetry perception are based on the symmetry of either individual fects the prevalence of short-duration fixations. The results showed points or higher level features , such as lines or comers. To evaluate unimodal distributions when fixation durations during reading and letter these models, we compared judgment of 2-D stimuli defined by individual detection tasks were analyzed. A bimodal distribution characterized points or by patterns of points. Vertical symmetry perception, inves oculomotor activity during copytyping. tigated by the 2AFC method, was better when the points formed sim (335) ple features. We conclude that perception of symmetric features aids Attentional Reduction of (Processing) Noise in Perceptual Encoci in the perception of object symmetry. ing. PETER C. GORDON, Harvard University (sponsored by Kathryn 502 Posters 336-347 Saturday Noon Bock)-Performance of a distractor task influences perception of a con tirely viewable from one perspective produces an orientation-free current stimulus presented in a different modality by affecting the signal representation if the path is large, but an orientation-specific represen to-noise ratio in the encoding of perceptua1 features as described by an tation if the path is small (Presson, DeLange, & Hazelrigg, 1989). We additive cue model. The effect of reduced attention on different per show that the factor critical to orientation specificity is not path size ceptual features closely follows the effect of adding white noise to the but the perspective afforded by paths of different sizes. A horizontal stimulus. This supports the conclusion that one role of attention in per perspective produces an orientation-free representation and a vertical ception is to reduce processing noise in stimulus encoding. perspective, an orientation-specific representation. (336) (342) The Serial and Parallel Nature of Attention Sbifting: Biology and Category and Property Specificity in Category-Based Induction. Behavior. WALTER SCHNEIDER, JUDITH SHEDDEN, University SHARON LEE ARMSTRONG, Central University ofIowa (sponsored of Pittsburgh, & GREGORY McCARTHY, VAMC, West Haven-The by Lloyd Avant)-Osherson et aI. (1990) developed a model predict nature of attention shifting was examined in a positional scanning ex ing strength of categorical arguments ("Robins have an ulnar artery'" periment in which attention is attracted to one location and then scans "Birds have an ulnar artery") based on similarity relations among the to four other locations. Both reaction time and high-{!ensity evoked poten categories occurring in the premise(s) and conclusion. Their model was tial measurements were made. Attention shifting was examined in a serial tested using biological categories and properties. Research here tested character search and parallel color search. The biological basis of at the model across different kinds of categories (artifacts) and properties tention shifting is discussed. (behaviors) with mixed results. Prospects for a general theory of in (337) duction are discussed. The Distractor-Suppression Effect: Alternatives to Response In (343) hibition Theory. JIM CHEESMAN & MATTHEW POLLOCK, Uni Imagery and Creativity. MARGARET JEAN INTONS versity ofSaskatchewan (sponsored by James M. Clark)-In past studies, PETERSON, Indiana University-To study if and how imagery con we have observed that the negative priming effect diminishes and even tributes to creativity, we tested three conditions assumed to vary in their reverses to produce a facilitation effect as a function of practice. Two dependence upon imagery. In all three conditions, subjects were asked new experiments are reported that show that neither the speed-accuracy to combine four components (drawn from simple geometric forms and mechanism proposed by Neill and Westberry (1988) nor the argument capital letters) into novel, but potentially useful, objects. The data sup that subjects were predicting the occurrence of suppression trials can ported the view that imagery contributes to creativity and identified some be used to discount our results. We believe that the effects of practice constraints . are inconsistent with an attentional inhibition mechanism. An alternative (344) mechanism involving the coordination of response codes is proposed Effects of Context on Representational Momentum. TIMOTHY L. to account for the patterns of observed results. HUBBARD, Eastern Oregon State College (sponsored by Alliston K. (338) Reid)-The context surrounding a moving object influences judgment Right Hemisphere Lateralization of Arousal. ROGER WHITE of that object's final position. A square frame context surrounded a rotat HEAD & MARIO LIOTTI, University of Oregon (sponsored by Steven ing rectangle, and subjects judged final rectangle orientation. If the frame Keele)-Is there a specific cortical mechanism activated during sustained rotated in the direction opposite to the rectangle but vanished prior to attention? Behavioral studies of neurological patients and blood flow judgment, direction of representational momentum was reversed. If a data in normals suggest a right hemisphere lateralitation of arousal. The stationary frame was shown at judgment, direction of representational current study used reaction time and electrophysiological measures to momentum was strongly biased toward the frame orientation, regard test the prediction that a right hemisphere superiority would emerge when less of frame motion prior to judgment. normals are required to sustain attention over long delays. Both the anat- (345) 0my and the time course of arousal will be considered. Languages of Number: Language, Orthography, and Developmen (339) tal Changes in Number Similarity Judgments. KEVIN F. MILLER, Perceptual ProceBng and Recall for "Different" Items. RICHARD S. University ofminois at Urbana-Champaign-All mathematical ideas must CIMBALO, Daemen College, & ROBIN J. BRICKLER, SUNY at be expressed in some notation, and the most basic notation of all is the Brockport-Three experiments are described wherein a consonant is dis system of number names a language uses. The structure of number tinctive (red against a black ground) in a Sperling-type tachistoscopic narning systems and the orthographies used to describe them will be recall task. Overall recall advantages resulting from distinctiveness are discussed. Results of a recently completed study of effects of language difficult to explain using often-cited retrieval explanations. Our results and orthographic variation on number-similarity judgments by children extend earlier [mdings from short-term memory, where strategically in the U.S., Korea, and China will be presented. based explanations are more possible, to even briefer tachistoscopic (346) presentations, where these would seem to give way to more perceptual Representational Transfer in Math and Logic Problem Solving. explanations. LAURA R. NOVICK & CINDY E. HMELO, Vanderbilt University (340) (sponsored by James Pellegrino)-Previous research on transfer in Listening to the News: Knowledge Acquisition and Metamemory. problem solving has considered the transfer of a solution procedure (e. g. , SANDRA L. SCHNEIDER, University of South Florida, & SUZANNE K. a sequence of mathematical operations) from a source problem to an LAURION, University of Wisconsin-Madison (sponsored by James J. analogous target problem. We report the results of several studies that Jenkins)-This study provides evidence that (I) knowledge acquired from document and explore solvers' ability to transfer a problem representa a radio broadcast is affected by listener involvement (personal relevance) tion (e.g., a matrix, a hierarchy) from a source problem to a superficially but not by news item format (editorial vs. informational) and (2) confi dissimilar target problem that requires a very different procedure for dence ratings tend to be well calibrated with accuracy measures of solution. listeners' newly gained knowledge. The confidence-accuracy relation (347) ship is moderated by listener involvement, question difficulty, and in External and Internal instantiation of Abstract Principles Facili dividual differences. Overall, categorical shifts from certainty to par tates Analogical Transfer. ZHE CHEN, University of Kentucky, & tial confidence are easier for listeners to discriminate than other gradations MARVIN W. DAEHLER, University of Massachusetts (sponsored by in confidence. Marvin W. Daehler)-Abstract concepts or principles have often been (341) found inefficient in transfer. Using an insight problem-solving paradigm, The Role of Viewing Perspective in the Orientation Specif'lCity of the present research indicated that both extemaI exemplification (provid Cognitive Maps. TINA NOLAN & JEANNE SHOLL, Boston Col ing a concrete example of principle) and intemaI instantiation (encourag lege (presented by Jeanne Sholl)-A single look at a path that is en- ing learners to generate examples of a principle) facilitated analogical 503 Saturday Noon Posters 348-359 transfer. The fmdings shed more light on the mechanisms of analogical individual maze arms. Critically, the model does not include the guidance transfer: mental efforts devoted to generate concrete instances facili of choice behavior by a cognitive map. The performance of the model tate applicability of abstract principles in concrete contexts. will then be compared with the performance of male Sprague-Dawley (348) rats in a variety of new experimental situations. Subgoal Learning Not Aided by Studying Multiple Methods. (354) RICHARD CATRAMBONE, Georgia Institute of Technology (spon Material Matdling by a BottIenoIied DoIpbin. HERBERT L. ROIT sored by Timothy A. Salthouse)-Categorization research indicates that BLAT, DAVID A. HELWEG, University of Hawaii at Manoa, increased variability of exemplars leads to better transfer. This suggests PATRICK W. B. MOORE, & PAUL E. NACHTIGALL, Naval Ocean that problem solvers should best acquire subgoals for solving problems Systems Center, Hawaii Laboratory-A blindfolded dolphin was trained when they study examples demonstrating multiple methods for achiev to perform the matching-to-sample task using echolocation. The stimuli ing subgoals. An experiment using algebra word problems failed to con were identical O.3-liter aluminum bottles filled with different substances. firm this. Furthermore, learners were less likely to pursue old subgoals The dolphin selected the correct matching alternative with substantially requiring familiar methods in novel problems when new methods for above-chance accuracy. Characteristics of the dolphin's echolocation other subgoals were needed. Studying multiple methods did not reduce clicks and the corresponding echoes returning from each target will be this context sensitivity. described. (349) (355) Insights beyond Words: When Language Overshadows Thought. Parts Is Parts: Use 01 Fraction SyinIIok by a Cbim.. II7ft, SARAH T. JONATHAN W. SCHOOLER, STELLAN OHLSSON, & KEVIN BOYSEN, Ohio Stole University and Yerkes Primate Research Center-A BROOKS, University of Pinsburgh (sponsored by Lauren Resnick) chimpanzee was taught to label the relationship between symbols for Four experiments demonstrated that verbalization can disrupt insight fractions 1,4 and 'h, and portions of fruits. She was required to choose problem solving. In Experiments 1 and 2, retrospective verbalization the correct Arabic numeral (0, ',4 , 'h, 1, 2, or 3) when whole or parts of attempted insight problem-solving strategies reduced the probability of fruits were presented. Following acquisition, the chimp readily gener of successful solutions. In Experiment 3, concurrent verbalization simi alized use of fraction symbols to novel foods. A fmal test evaluated her larly impaired the solving of insight problems but not noninsight ability to generalize conceptual use of fractions with liquids (juice pre problems. In Experiment 4, concurrent verbalization remained disrup sented in glasses), including 1-3 full glasses, or glasses that were 'h tive even though subjects were encouraged to consider alternative ap or 1,4 full. proaches. These findings suggest that verbalization may overshadow (356) critical nonverbal insight processes. Peak Shift Obtained After Discrimination Training on tbe Dimen . (350) sion 01 Relative Numerosity. W. K. HONIG, Dalhousie University-In Action Planning: The Role 01 Prompts in Command Production. two experiments, pigeons were trained to respond to positive arrays of STEPHANIE M. DOANE, DANIELLE S. McNAMARA, WALTER 18 small red and 18 small blue elements, presented in randomized pat KlNTSCH, & PETER G. POLSON, University ofColoradn (sponsored terns on a response screen. They then discriminated these from arrays by Walter Kintsch)-Learners can acquire planning skills in domains such containing either more red than blue elements, or the reverse (but not as computer command production by receiving information in the form both). The full range of relative numerosities was presented in post of help prompts. Depending on the learner's background knowledge, discrimination testing. This test produced a marked peak shift; the birds comprehension of prompt information may range greatly, and comprehen responded more to patterns that contained more than 18 positive ele sion is required to translate prompt information into planning skills. Pre ments. The discrimination and the peak shift transferred well, on a sented are empirical and modeling data that support this claim and sug proportional basis, to arrays of 64 elements, presented in various gestions for help instructions designed to optimize learner comprehension. proportions . (351) (357) The Effect 01 Context on Judgments 01 Trutb. DEBORAH A. Hieran:bicaI and Nonbierarchical Orpnization and Rat Serial Pat PRENTICE, Princeton University, & ALBERT F . SMITH, SUNY at tern Learning. STEPHEN B. FOUNTAIN, JAMES D. ROWAN, & Binghamton (presented by Albert F. Smith) (sponsored by Ralph R. MICHAEL L. NERING, Kent Stale University-Rats learned either hi Miller)-Subjects read three true or three false biographical sketches erarchically or nonhierarchica11y organized serial patterns. The hierar of former presidents and then read a target sketch containing both true chically organized pattern was characterized by lower order rules relating and false statements. Subsequent true-false judgments about the target pattern elements within chunks and higher order rules relating succes depended on the truth of the context sketches. This pattern of results sively larger groups of chunks. In the nonhierarchical pattern, chunks was found with familiar and with unfamiliar facts about the target presi from the first and second halves of the hierarchical pattern were ex dent and when subjects evaluated the truth of statements about presi changed. Rats learned the hierarchical pattern faster than the nonhier dents before and after reading the sketches. Instructions emphasizing archical. Rats' error profIles will be compared to those of human sub accuracy eliminated the effect. jects perfortniog an analogous task. (352) (358) Response Amplilkation: An Automatic or Learned Reaction to Appetitive InstnunentaI Learning in the Ampbibian Bujo GmIaIIIIII. Failure? JASPER BRENER, SUZANNE H. MITCHELL, & AN RUBEN N. MUZIO, ENRIQUE T. SEGURA, Instituto de Bioiogia DREW B. SLIFKlN, SUNY at Stony Brook-Rats in a high-force group y Medicina Experimental, & MAURICIO R. PAPINI, Texas Christian pressed a beam with 10 g-force for food; a low-force group pressed University (presented by Mauricio R. Papini)-A conventional runway with 1 g-force. When the probability of reinforcement was reduced from procedure with access to water as reinforcer was used to study the per 1.0 to 0.25, immediate amplification of peak forces occurred in both formance of toads in several instrumental learning paradigms . Variables groups. Since the low-force contingency did not differentially reinforce manipulated include trial distribution, reward magnitude, and reward high peak forces, it is concluded that response amplification is unlearned. schedule. Some basic learning phenomena were readily observed (ac However, group differences in the form of amplification suggested that quisition, extinction, spontaneous recovery), but there was no evidence the process is modified by learning. of partial reinforcement and magnitude of reinforcement extinction ef (353) fects. The results will be presented in the context of comparative research A MapIess Model 01 Spatial Performance in the Radial-Arm Maze. on the paradoxical effects of reward. MICHAEL F . BROWN, JOANNA E. VONCULIN, & PATRICIA (359) RISH, Villanova University-A previously described model of choice Retention 01 Taste Aversions: Evidence lor Retrieval Competi in the radial-arm maze will be reviewed. The model assumes that choices tion. W. ROBERT BATSELL, JR., & MICHAEL R. BEST, Southern in the radial-arm maze consist of a sequence of yes/no decisions regarding Methodist University (sponsored by Michael R. Best)-Conditioned taste 504 Posters 360-371 Saturday Noon aversions are significantly weaker I day after conditioning than 3, 5, depressed respiration were obtained in 5-day-old chicks when the ther or 10 days after conditioning. The retention deficits at I day are not mal stimulus was decreased and morphine dose increased (5.0 mg/kg). due to nonassociative factors like toxicosis aftereffects or hydration dif Morphine hyperalgesia and respiratory depression were reversed by ferences. Instead, they seem to be mediated by retrieval interference naloxone, but naloxone alone produced hypoalgesia. provided by contextual associations. (366) (360) Using Fragmentary Patterns to Compare Models of Catqory Local Context during Training as a Modulator of Pavlovian Learning. JOSHUA B. HURWITZ, Harvard University (sponsored by Responding. ROBERT C. BARNET, NICHOLAS 1. GRAHAME, & William K. Estes)-This study compares the accuracy of a feature-node RALPH R. MILLER, SUNY at Binghamton (presented by Ralph R. network model and an exemplar model in predicting test-trial perfor Miller)-In a conditioned suppression task with rats, unsignaled USs mance after category training with either whole stimulus patterns or pat during Pavlovian training attenuated responding to a target CS. Extinc tern fragments. The network model's fit was superior to that of the ex tion of the training context partially restored responding to the CS, as emplar model, more so than in previous paradigms. However, to achieve did signaling the otherwise unsignaled USs with a "cover" stimulus. this superiority, the network model was modified to incorporate an aver Immediately surrounding each target CS-US trial with the excitatory aging function into its computation of category-node output. cover stimulus also attenuated responding to the target CS. In this case, (367) posttraining extinction of the cover stimulus, but not extinction of the Structured Imagination: The Role of Category Structure In Ex training context, restored responding to the target CS. These observa emplar Generation. THOMAS B. WARD, Texas A&M University tions implicate the temporally local context of CS training as the com Subjects created drawings and descriptions of imaginary animals. Most parator stimulus that modulates conditioned responding. of their creations were bilaterally symmetrical, had sensory receptors (361) and appendages, and included attribute correlations (e.g., feathered crea 1ime DiscrimiDation in CoIumbG /bIIa and Homo 1IIJlIIms. 1. GREGOR tures had beaks and wings). When creating variants on their initial animal, FETTERMAN, IUPUl, & PETER R. KILLEEN, Arizona State Uni subjects varied shape and appendages more across than within species. versity-We studied temporal discrimination in pigeons and humans, The fmdings illustrate the concept of structured imagination. In gener focusing on stimuli for less than I sec. A staircase technique was used ating novel exemplars of known categories, imagination is structured in Experiment I; the method of constant stimuli was used in Experi by principles that are characteristic of that category. ment 2. The data were consistent with the generalized form of Weber's (368) law. Experiment 3 ruled out brightness as the effective cue. Experiment 4 The Course of Forgetting in Connectionist Models. WILLIAM L. employed a modified staircase procedure; the results indicated a sub OLIVER, Florida State University (sponsored by George Weaver) stantial region over which the difference thresholds were constant. Forgetting caused by the learning of new information, probabilistic weight (362) decay, and contextual fluctuation was investigated in connectionist models Unconditioned Stimulus Factors in Sexual Approach Condition of recognition memory. The models were based on the backpropaga ing. MICHAEL OOMIAN & KEVIN HOLLOWAY, University of tion and mean field theory learning algorithms. Models were identified Texas at Austin-Male lapanese quail learned to approach a localized that displayed "exponential-power" forgetting functions for d' mea stimulus that signaled access to a female. Copulation with the female sures of recognition memory, which have been shown by Wickelgren facilitated approach conditioning, but learning also occurred with ex (1972) to describe human forgetting. Implications for theories oflong posure to the female behind a wire screen. Devaluations and revalua term retention and amnesia will be discussed. tions of the unconditioned stimulus (through changes in serum testoster (369) one levels) produced corresponding changes in conditioned approach Do the Beginnings of Auditory Words Have a Special Status? behavior, suggesting that the learning was mediated by 50S mechanisms. CYNTHIA CONNlNE, DEBRA TITONE, & DAWN BLASKO, SUNY (363) at Binghamton (sponsored by Albrecht Inhoft)-Three cross-modal prim "Pet Whipping Boy" Behavior in Low Dominant Captive African ing experiments are reported which investigated whether base words Elephants. ROBERT H. I. DALE & MELISSA R. SHYAN, Butler were activated by similarly sounding nonwords. Acoustic-phonetic University-Wild female African elephants live in herds with varying similarity was manipulated by altering linguistic features of the initial levels of consanguinity (matriarchies). Dominance hierarchies are sta (e.g., SEASON-ZEASON) or medial phoneme (LABEL-LAPEL) of a base ble, based on age, awareness of environmental resources, size, and ag word. Initial and medial nonwords showed significant priming effects gressiveness. The most dominant animal is not always the most aggres for associates to the base word. The results support a model in which sive. An artificially established herd (no known consanguinity) was auditory words are recognized on the basis of goodness of fit. studied at the Indianapolis Zoo to determine dominance hierarchy be (370) havior. Changes in dominance were studied when a new elephant was The Impact of Thinking-Aloud Protocol Analysis on the Writing introduced. These fmdings are compared to reports from other zoos. Process. SARAH E. RANSDELL, New College of the University of (364) South Florida (sponsored by Ira Fischler)-Protocol analysis, which in Rat and Human Temporal DIscrimination: A Signal Detection volves collecting spoken thoughts during problem-solving tasks, has been Analysis. 1. S. COHEN & M. A. BRAMLEY, University of Windsor used extensively to reveal important on-line information about the writing Rat and human subjects had to differentiate long from short signals with process; however, its effects on such writing have received very little an immediate response (FI 0 sec) to the former and a delayed response attention. Transcripts of writing samples were compared when writers (FIIO sec) to the latter. Signal sensitivity, P(A), and bias, B, as a function produced a concurrent verbal protocol, no protocol, and when the writers of the difference between signal durations and of the delay between a produced a retrospective protocol based on watching a real-time replay signal and the opportunity to respond were investigated. Results sup of the original composition. port sca1ar expectancy theory (Gibbons, 1977) and prospective processing (371) of signal duration. Recovery from MBmaIyses or Gerden-Path Sentences. FERNANDA (365) FERREIRA & JOHN M. HENDERSON, University of Alberta (pre Morphine Hyperalgesia and Developmental Changes in Nocicep sented by 10hn M. Henderson) (sponsored by Peter Dixon)-Sentences tion in Domestic Fowl. RICHARD A. HUGHES, MARK BOWES, such as After the Martians invaded the town was evacuated are initially & KENNETH 1. SVFKA,Iowa State University-Morphine (2 .5 mgIkg) rnisanalyzed by the syntactic parser. We will present several experi produced hyperalgesia in 14-day-old chicks but not in 3-, 5-, or 7-day ments examining the process of reanalysis. Two conclusions emerge: olds. Respiration was lower in 3-day-old chicks than in older groups, First, reanalysis is more difficult when the head of the rnisanalyzed phrase but morphine depressed respiration at each age. Hyperalgesia and is far from the disambiguating word. Second, effects of verb bias are 505 Saturday Noon Posters 372-383 more pronounced when reana1ysis is difficult. These results suggest an frequency of look backs for computer-presented text varied as a func interaction between syntactic processing and thematic role assignment tion of whether critical sentences embedded within text segments were during reanalysis. consistent or inconsistent with preceding text, the reader's prior knowf (372) edge, or both. Look back as a recovery operation during comprehen Gestures Precede Speech: A Temporal Interval Predicted by Word sion may be reflected in what information is subsequently recalled and Familiarity • PALMER MORREL-SAMUELS, EDS Center for Machine recalled following a one-week delay. Intelligence, & ROBERT KRAUSS, Columbia University (sponsored (378) by Julian Hochberg)-Previous work suggests that conversational Studying Varieties of Lexical Structure with Word Association. gestures are perceived as meaningful because observers can utilize the JULIA C. JORGENSEN, Teachers College, Colwnbia University (spon semantic content of accompanying speech. The current research ana sored by George Miller)-Responses in continued word association are lyzes the temporal relation between gesture onset and onset of the ac viewed as reflections of lexical structure. Culturally important realms companying word. In our sample, gestures precede speech by longer of activity tend to be marked by sets of highly distinctive terms (such intervals when the accompanying word is unfamiliar. Results suggest as cooking vocabulary). We studied the conventionality and distinctive both gestures and words are accessed during lexical selection, and that ness of associative responses to words which denote scenes, scripts, ac gesture representations are generally less differentiated and more ac tions, and concrete objects in order to evaluate claims about the rela cessible than semantic representations. tive importance of such words in organizing human activity. (373) (379) Contextual InOuences 011 the Comprebension of Complex Concepts. When Is Love a Journey? What Metaphors and Idioms Mean. RICHARD J. GERRIG, Yale University, & GREGORY L. MURPHY, MATTHEW S. McGLONE, SAM GLUCKSBERG, & MARY BROWN, University ofnlinois-Our experiments investigated how readers under Princeton University (presented by Sam Glucksberg)-We asked peo stand novel noun-noun phrases like "vodka face." One view argues ple to provide interpretations of novel metaphors. Contrary to Lakoff's that readers simply access a discourse element described by the head proposals concerning conceptual root metaphors, people relied almost noun; another claims that readers form a new concept relating the two entirely on specific attributes of metaphor vehicles rather than on more nouns. The experiments supported the latter view. A description of the general conceptual analogies. In contrast, stable conceptual frameworks relevant relation was sufficient for comprehension of the phrases, as do seem to underlie idiom comprehension. These results imply that shown by reading times and a later test of readers' interpretations of metaphor use can be spontaneous and creative, while idioms (and perhaps the phrases. frozen metaphors) tend to rely on prestored conceptual analogies in (374) semantic memory. Forward Inferences about Specific Events during Reading. JOHN D. (380) MURRAY, CELlA M. KLIN, & JEROME L. MYERS, University of The Development of "Dating Scripts" in Young Children. M. DIANE Massachusetts (sponsored by Arnold Well)-Two experiments are CLARK & BRIAN NEWCOMER, Shippensburg University (sponsored reported which suggest that individuals make forward inferences about by PaulaJ. Schwanenflugel)-First-grade males and females were video specific events during reading. Prior studies have concluded that such taped while discussing a date for Barbie and Ken. Videotapes were filmed inferences are at best made minimally. Subjects showed a naming time in a play area that included Barbie, Ken, and three activity areas. Chil facilitation for target words representing a previously unmentioned cause dren manipulated the objects while responding to an interview protocol for an action stated in the last sentence of a narrative. A similar naming that probed for settings, activities, and conclusions of a date. An early time facilitation occurred in passages that were considerably shorter and generalized event representation of a date was clearly developed and for targets representing highly predictable consequences. was more similar between the sexes at this age than for adults. (375) (381) Effects of Speech Act Verb Pragmatic Strength and Speaker Per Doing It for Themselves: CoUege Experience and Appointment spective on Judged Communication Goodness. PAUL C. AMRHEIN, Keeping. SHARLENE W ALBAUM, Mount Holyoke College (spon University ofNew Mexico (sponsored by Henry Ellis)-The interaction sored by Will Millard)-Appointment keeping is one domain of prospec of speech act verb pragmatic strengths in requesting-committing dia tive remembering that may develop significantly during early adulthood. logs was investigated. Subjects judged dialog "communication good In this study, first- and fourth-year college students were surveyed regard ness" from requestor and committor perspectives. (I) Relative to weaker ing their use of appointment-keeping strategies. Results indicate that "non-face-threatening" requesting verbs, communication goodness was they differ in their reliance on strategies, with seniors reporting use of judged higher for stronger "face-threatening" requesting verbs when more accessible (and, therefore, more effective) cues. Results were ex paired with stronger committing verbs, but judged lower when paired plained in terms of changes in metacognitive awareness that come with with weaker committing verbs. (2) Stronger committing verbs increased the increased emphasis on independent living which is part of college life. judged communication goodness more when subjects judged from re (382) questor's than from committor's perspective. Age and Text Genre Effects on Capacity Expended While Read (376) ing. THOMAS V. PETROS, BARB K. BENTZ, TARA MILLER, & Modality mects 011 Coherence: Is It Better to Be Seen than Heard? DAVID TUPA, University of Nonh Dakota-The effects of age and ROGER J. KREUZ, RICHARD M. ROBERTS, ELIZABETH A. passage characteristics on processing capacity expended while reading BAINBRIDGE, & D. KRISTEN GILBERT, Memphis State Univer text was examined. Fourth grade, sixth grade, and college students' read sity (sponsored by Arthur C. Graesser)-Previous research has indicated narrative and expository texts from a computer. Subjects responded to that discourse coherence is affected by subjects' expectations: Highly secondary tones that periodically occurred while reading the texts. Older coherent plays and conversations became less coherent when their subjects required less capacity to process the texts than younger sub descriptive labels were reversed. Because these results were based only jects, and the effect of text genre and idea unit importance on capacity on written transcripts, the generalizability of this rmding is limited. The expended was greater for younger than older SUbjects. present research uses a broader range of presentation modalities: (383) (1) written transcripts, (2) audio presentation alone, and (3) audiovisual Visuospatial Precuing Effects in Young and Old Observers. LYNN presentation. Modality effects on the perception of coherence will be ZIMBA & DONALD J. TELLINGHUlSEN, University ofIowa (spon discussed. sored by James V. Hinrichs)-Age-related performance differences in (377) visuospatial precuing effects were investigated in terms of identifica Looking Back Over Things Once Parced: Text Versus Knowledge tion latencies to targets presented at attended versus unattended loca Inconsistencies. O. VERONIKA PRINZO, Fon Hays State Univer tions. Young and old observers made 2AFC decisions about line orien sity (sponsored by Joseph H. Danks)-Mean sentence reading times and tation or lexical nature of targets presented at low or high luminances. 506 Posters 384-395 Saturday Noon No differences in attention-effect magnitude were found between the edge of DMTS success. Sample, but not comparison, manipulation in· groups in either task under either luminance condition. Results suggest fluenced overall self-report accuracy. A signal detection analysis (sig some aspects of cognitive/perceptual function remain intact during the nal = successful DMTS response) found opposite effects of the two aging process. manipulations on report sensitivity. Only comparison manipuiation medi (384) ated a pervasive bias for reporting successful responses. The results sug Directed Forgetting Made DimcuIt. JONATHAN M. GOLDING, gest that metamemory can be differentially influenced by factors oper University of Kentucky, DEBRA L. LONG, University of California, ating at encoding and retrieval. Davis, & COLIN M. MACLEOD, University of Toronto, Scarborough (390) Campus (sponsored by Philipp Kraemer)-Two experiments investigated Constraints on Similarity Effects for Situational Frequency Judg semantic relatedness effects in directed forgetting. All combinations of ments of Words. CAREN M. JONES, Stanford University, & EVAN "forget" and "remember" cues (RtF, R/R, F/F, F/R) were used with HEIT, University of Michigan (presented by Evan Heit) (sponsored by both unidirectionally related (crab-leg) and unrelated (seat-rope) word John Jonides)-Multiple-trace memory theories (e.g., Hintzman, 1988) pairs. Among other findings, both recall and recognition showed facili predict that frequency judgments for a target word (e.g., tuna) will in tation for forget words only when they followed related remember words. crease with increasing presentations of a similar word (e.g., salmon). This demonstrates that directed forgetting cannot override preexisting Experimental results verified this prediction, showing a similarity ef semantic connections, placing a clear constraint on the domain of directed fect, but suggested two constraints. First, when tuna was not shown, forgetting explanations. presentations of salmon did not affect judgments for tuna. Second, the (385) similarity effect of each salmon presentation decreased with increasing Mental Representations of Spatial Relations: Combining Recog frequencies of salmon. nition Times and Distance Estimates. KARL F. WENDER & (391) MONIKA WAGENER, Universitat Trier-We combine two approaches Discourse Influences on Memory for Visual Forms. DEANNA to investigate spatial memory that have been used separately in the past. WILKES-GmBS & PETER H. KIM, Wesleyan University (sponsored Recognition times for places of a fictive city were measured under a by John G. Searnon)-Subjects examined abstract figures in conjunc priming condition and distances were estimated using a scaling proce tion with verbal labels, then conversed in pairs to arrange the forms dure. Nonmetric multidimensional scaling recovered solutions in two identically. Subsequent memory tests confirmed that verbal labels at en dimensions. A stochastic model is developed and tested that predicts coding can bias performance, but so can the basic pragmatic demands reaction times from the priming experiment on the basis of interpoint of referring in conversation. Recall was distorted most when partners distances from 2-0 space. shared the same perspectives on the referents, and least when their ini (386) tial labels conflicted. These results imply that "private" representations The Item-Specif'lClRelational Processing Distinction: Analysis of may be affected in the establishment of common ground. Cumulative Recall Performance. DANIEL J. BURNS, Lafayene Col (392) lege (sponsored by Todd R. Schachtrnan)-One limitation of the item Recognition Level and the Misinformation Errect: A Meta-Ana1ysis specific/relational processing distinction is the lack of good measures and Empiricial Investigation. MICHAEL P. TOGLIA, SUNY at um of item-specific and relational processing. Category access measures land, DAVID G. PAYNE, & JEFFERY S. ANASTASI, SUNY at are shown to be fraught with difficulties. An alternative technique for Binghamton-Examination of 32 experiments employing McCloskey and measuring item-specific and relational processing (within the context Zaragoza's (1985) modified recognition memory paradigm revealed that of hypennnesia experiments) is explored. It is shown that number of a significant misinformation effect (control-misled) tended to be ob items recalled during the early and late portions of a recall test differs served only with moderately high levels of control performance. This for each type of processing. observation may be a regression to the mean artifact. Computer simu (387) lations and an experiment (directly varying control performance levels) Is the "Flash" NeasoJIII"Y to Form a FWlbuIb Memory? CHARLES A. were conducted to investigate this possibility. Results of the simulations WEAVER m, Baylor University (sponsored by Lewis M. Barker) and the experiment are discussed in terms of their implications for mem As a laboratory demonstration, students were asked to remember an ory impairment hypotheses. ordinary event on January 16, 1991-they were essentially asked to form (393) a fake flashbulb memory. By happenstance, the U.S. bombing of Iraq Does Memory Reflect Statistical Regularity in the Environment? commenced the same day, triggering a real flashbulb memory. Com LAEL J. SCHOOLER & JOHN R. ANDERSON, Carnegie-Mellon parison of questionnaires completed immediately and three months later University (presented by John R. Anderson)-Anderson and Milson revealed minimal differences between true and fake flashbulb memories (1989) derived optimal performance functions for memory on the basis in accuracy, but substantial differences in the confidence with which of assumptions about memory's goals and a mathematical model of the these memories were held. statistical structure of the environment. Here, instead of using a mathe (388) matical model of the informational demands the environment places on Analogies, Similarities, and Abstraction. LESLIE J. CAPLAN & memory, we analyzed these demands directly. We studied three environ CARMI SCHOOLER, National Institute of Mental Health (sponsored ments: parental speech, newspaper headlines, and electronic mail. We by Carmi Schooler)-Subjects read about pairs of analogous domains, show that the performance functions of memory are optimal given the then rated and described their similarities. We manipulated encoding structure of these environments. complexity of the first passage, domain distance, and whether the anal (394) ogy was stated. Similarity ratings were highest in near domain/no anal Memory for Randomly F1IIed Matrices. THADDEUS M. COWAN, ogy and far domain/analogy cells. Near domains increased "surface" JEROME FRIEMAN, & CHARLES P. THOMPSON, Kansas State similarities. Providing analogy statements increased the number of University-The capacity of a mnemonist was tested by presenting num parallel concepts described. Complex encoding increased statements that bers singly in matrix form for matrices ranging from 3 x 3 to 10 x 10. the domains were isomorphic. Results support the hypothesis that these The numbers appeared by row or randomly. The subject rehearsed the variables affect abstraction. row presentation matrices after each row was completed and/or after (389) 15 numbers were given. The rehearsal times for the smaller randomly Metamemory as a Function of Memory Load at Encoding and presented matrices appeared when each row was completed. Rehearsal Retrieval. THOMAS S. CRITCHFIELD, Auburn University, & LEA T. strategies for the larger matrices were more complex. ADAMS, Rlinois State University (sponsored by Gordon M. Redding) (395) In a delayed matching-to-sample (DMTS) task, the number of stimulus Young Adults' Perceived Versus Actual Experiences of Forgetting. items was manipulated both at encoding (sample stimuli) and retrieval JUDITH A. SUGAR, Colorado State University-Young adults com (comparison stimuli), while trial-by-trial self-reports assessed knowl- pleted a memory questionnaire in which they rated their use of various 507 Saturday Noon Posters 396-407 memory strategies and estimated their average frequency of forgetting . BOURLIER, & SARAH WILLIAMS, Central Michigan University They subsequently kept track of their everyday experiences of forget (sponsored by John S. Monahan)-The role of response bias in cued ting. Estimates of the frequency offorgetting were very poor: subjects recall hypemmesia was examined by comparing the standard- and forced recorded approximately six times more incidents than they estimated cued recall procedures. The standard test asked subjects to recall the and there was no relationship between their estimates and the number studied words. The forced test required subjects to guess unrecaUed items. of incidents they reported. The net improvement was greater for the standard test. However, (396) hypermnesia was also observed in forced-cued recall. It was concluded Age Memory Changes a Necessary Consequence of Aging? that the cued recall hypermnesia is not an artifact of response bias. LEONARD GIAMBRA, HERBERT WEINGARTNER, DAVID (402) ARENBERG, PAUL COSTA, ALAN ZONDERMAN, JUDY FRIZ, Successive Attempts to Produce Interference in Associative Recog E. METIER, NaJionallnstituJe on Aging, & CLAUDIA KAWAS, Johns nition Memory. WILLIAM E. HOCKLEY & NADIA BARGHOUT, Hopldns University-When large decrements in recent memory do oc Wilfrid Laurier University-Previous research (Hockley, 1991 a, 1991 b; cur in the elderly, it may indicate pathology rather than normal aging. Murdock & Hockley, 1989) has demonstrated that short-term recogni Fifty-nine subjects who took the Benton Visual Retention Test (BVRT) tion memory for associations between random pairs of words is far more every six years during their fifth through eighth decades were examined resistant to interference effects than recognition memory for single words. for memory changes. Some individuals showed no change in BVRT The nature and degree of interference was varied in three experiments scores over 25 years, whereas others showed large changes. Other than in order to establish the conditions that produce interference in associa one DAT case, no differences were found in degree of chronic illness tive recognition. The results have important implications for current between those individuals who changed and those who did not change. global models of human memory. (397) (403) Effects of Arousal, Imagery, and Encoding in a Von Restorft'Task. Continuity between Recall and Recognition Revisisted. ZEHRA F. KINTA M. PARKER & LINDA WARREN DUKE, University of PEYNIRCIOGLU & LAURA J. MARGOLIS, American University Alabama at Binningham (presented by Linda Warren Ouke)-A previ High-frequency words are better recalled than recognized; low-frequency ous study found that recall of high-imagery words was more disrupted words are better recognized than recalled. We tested memory for eight by arousal than recall of low-imagery words. This study separated en letter high- and low-frequency words with 2, 4, and 6 letters given as coding strategy from the imageability of the stimulus in a von Restorff cues. At all levels of cuing, the patterns of performance resembled those paradigm. High- and low-imagery lists were presented to subjects who in free recall with fragment cuing instructions (explicit remembering), were assigned one of four encoding instructions (rote rehearsal, sepa whereas they resembled those in recognition with fragment completion ration imagery, relational imagery, or uninstructed). Earlier fmdings instructions (implicit remembering) . were replicated. Encoding strategy also interacted with arousal to af (404) fect recall. Frequency, Preference, and the Concurrent Process Model. MAT (398) THEW A. LEE, JANET L. SUNDBERG, & IRA H. BERNSTEIN, Age and Centrality Inftuence Recall and Susceptibility to Mislead University of Texas at Arlington (presented by Ira H. Bemstein) ing Postevent Information. KRISTEN E. McKEE & TERRY R. Preferential ratings are more positive for familiar as opposed to unfamiliar GREENE, Franklin and Marshall College (presented by Terry R. stimuli. There has been debate as to whether preference (affect) pre Greene) (sponsored by Richard S. Lehman)-First and fifth graders and cedes, succeeds, or operates concurrently with perceived frequency (cog college students watched a videotape during which a waiter tripped and nition) . Results obtained from judgments of both geometric forms and spilled spaghetti on a customer. Following the presentation, the ex faces support the latter, concurrent process approach. perimenter summarized the events. Central or incidental information (405) and true or misleading statements were embedded in the summary in Re-pIIred Prime and Target RepetitIon in I.exicaI DIDion. DAVID S. a 2x2 design. Two days later, subjects recalled the videotape and GORFEIN, Adelphi University, ANDREA BUBKA, St. Peter's College, responded to questions . Recall and susceptibility to misleading infor & STEPHANIE A. BERGER, Adelphi University-At the Psychonomic mation were influenced by age and centrality. Results are discussed in meeting of 1990, we reported that the amount of repetition priming is the context of eyewitness testimony research. decreased when a new prime is substituted on repetition. The aim of (399) the present paper is to examine whether the repetition of both prime Priming in a Very Large, HlghIy Organized Set of Numbers. and target contributes to the size of the repetition effect. An identity JEROME FRIEMAN, CHARLES P. THOMPSON, THADDEUS M. repetition condition was contrasted with a condition in which two related COWAN, RODNEY VOGL, & RAJAN MAHADEV AN, Kansas State prime-target pairs were re-paired on repetition to create two new related University-We tested a subject who had the fIrst 30,000 digits of pi prime-target pairs. Surprisingly, both of these conditions produced large committed to memory. The fIrst three experiments tested for priming and equivalent repetition effects. based on the location and similarity of strings of digits to target strings (406) in the set. The next three experiments showed that he has the digits or Suppression of Information In Word Recognition and Reading. ganized in strings of 10 digits and that the primary information for the GREG B. SIMPSON, HYEWON KANG, & MERILEE A. KRUEGER, location of these strings is contained in the fIrst three digits of each string. University of Nebraska at Omaha-In word priming, a meaning of a (400) homograph is suppressed if it is inconsistent with the meaning processed Lexicality and Generation Effects on Memory Span. RICHARD on an earlier trial. We present a priming experiment showing that the SCHWEICKERT, BETH ANSEL, & MARK A. McDANIEL, Purdue suppression is highly specifIc (other kinds of inappropriate information University-A basic equation for immediate memory is that span equals are not suppressed). A sentence-reading study is also presented, show speaking rate times trace duration, s = rt. We replicated the classic ing that the suppression does not depend on an explicit response to the word length effect that longer words decrease the rate, without chang homograph or any of its associates. ing trace duration. We also found two factors that increase trace dura (407) tion, when rate is constant: (a) lexicality and (b) generating words from Spontaneous Remembering: The Role of Indirect and Direct fragments. Word length and generation had multiplicative effects on Memory Tests. ROBERT F. BELLI, CreighJon University, JEFFERY J. speaking rate, in accordance with the multiplicative factors method of FRANKS, Vanderbilt University, & STACIE PANZA, Creighton Uni Seth Roberts. versity (sponsored by Jeffery J. Franks)-We tested indirect versus direct (401) memory upon "spontaneous remembering" (Franks et aI ., 1989). Ac Hypermnesia In Cued- and Fon:ed-Cued Recall. HAJIME OTANI, quisition consisted of word fragments and solutions; testing consisted HOWARD L. WHITEMAN, KELLY LOVELACE, MATTHEW of half "old" acquisition fragments, and half "new" fragments. In- 508 Posters 408-413 Saturday Noon direct memory involved fragment completion, whereas direct memory ceptual identification test. The results suggest some constraints on models involved recognition. Finally, the spontaneous discovery of an old versus of word identification. new classification rule was tested, which was not discovered following (411) indirect memory, despite reporting the ease of completing acquisition Mental Maniplllldon of Numbers by BIIiDpaIs. CHERYL FRENCK fragments. The rule was discovered following direct memory. MESTRE, CNRS, Universite de Provence, & JYOTSNA VAID, Texas (408) AdeM University (presented by Jyotsna Vaid) (sponsored by Steven M. Meaning Selective Access in Repetition Priming. J. VIVIEN BAIN Smith)-Automatic activation of number facts was compared in the bilin BRIDGE, University of Oklahoma (sponsored by Charles Genys) gual's two languages. Bilinguals in Experiment I verified whether a Several experiments examined the role of context in repetition prim probe number (e.g., eight vs. huit) was the sum of a preceding number ing, using the lexical decision task. Repetition priming disappeared fol pair (e.g., two jive vs. deux: cinq) and in Experiment 2 whether the probe lowing a simple change in the preceding context sentence, provided that had been presented in the pair. RT for negative answers was compared change also altered the perceived sense of the target. A similar context for the two languages and for two levels of absolute difference between change, failing to alter the perceived sense of the target, did not reduce true sum and probe (split). repetition priming. (412) Beyond the Edges of a PIcture. HELENE INTRAUB & DEBORAH (409) BERKOWITS, University of Delaware-Observers remember a close Repetition Blindness and Deafness: New Data and Theory. up as having shown more of a scene than was actually depicted (lntraub MICHELLE MILLER & DONALD G. MAcKAY , UCLA. (presented & Richardson, 1989). In two experiments, using recall and recognition by Donald G. MacKay)-Words repeated in close temporal proximity tests, this effect was obtained with close-up, close-up-inverted, proto are difficult to perceive if presented rapidly and visually (repetition blind typic, and wide-angle views of the same seven scenes within minutes. ness), but not if presented auditorily (repetition deafness) at equivalent Results suggest that a perceptual schema containing information likely rates using acoustic compression techniques. The present study presented to exist outside the picture's boundaries becomes active during scene computer-compressed words auditorily either in sentences or scrambled perception and becomes incorporated into the memory representation. into lists. Repetition deafness was observed for words repeated in scram (413) bled lists but not in sentences. These results suggested a new theory PIcture Naming Reveal'! the ~ Invariances Expected of a Shape to account for repetition deafness and repetition blindness. Recognition System. IRVING BIEDERMAN, University of Southern (410) California, ERIC E. COOPER, & PETER C. GERHARDSTEIN, The Effects of Similarity and Frequency on Repetition Priming. University of Minnesota-Priming of picture-naming RTs is virtually PATRICIA L. TENPENNY, Loyola Univmity, Oricago, & EDWARDJ. unaffected by a change in the primed picture's position, size, or orien SHOBEN, University oflllinois (sponsored by Eugene B. Zechmeister) tation in depth. In contrast, such image changes produce striking inter In a follow-up to an experiment presented last year, we investigated ference in making old-new shape judgments. A comparison of perfor the influence of studied neighbors on repetition priming for target words. mance on the two tasks may provide a behavioral, noninvasive technique We varied relative target-neighbor frequency, the study status of tar for horizontally splitting the extrastriate visual systems, with naming gets and neighbors, the font used for the targets at study, the font used reflecting the functioning of the ventral system and both dorsal and ventral for the neighbors at study , and the font used for the targets in the per- systems determining episodic judgments. 509 Saturday Afternoon Papers 414-424 PERCEPTION I found but there was still no indication of storage of new information Embarcadero AB, Saturday Afternoon, 1:30-3:00 in semantic memory. 1:55-2:15 (419) Chaired by James T. Todd, Brandeis University Masked PrimIng Studies of EpI80dlc and Lexical Representation. DIANNE C. BRADLEY, Monash University (sponsored by W. D. 1:30-1:45 (414) Marslen-Wilson)-With masked (cf. unmasked) prime presentations, Memory for Position and Identity across Saccades. DAVID E. newly learned associations (e.g., fairy/SHARK) afford no reliable benefit IRWIN, University ofnLinois at Urbana-Champaign-A transsaccadic in lexical decision, but only in episodic recognition. However, priming partial report procedure was used to measure memory for position and between established associates (e.g., table/CHAIR) is equivalent across identity information across saccades. The results suggest that memory tasks. This latter effect of association is unlikely to have arisen within across saccades decays slowly, has a limited capacity, and is maskable the episodic system itself; and indeed, studies in form priming (e.g., for a brief time. In addition, it appears to retain identity information fistance/msTANcE) support an argument that access to episodic word better than position information. The implications of these results for representation can be lexically mediated. theories of information integration and perceptual stability across eye 2:20-2:35 (420) movements will be discussed. Sense from Nonsense in Epbodic Memory. J. W. WHITLOW, JR., 1:50-2:10 (415) Rutgers University-Efforts to understand priming and repetition effects Haptic Triangulation: Distal Layout Information from Spatia for codified events, such as words, have been complicated by the avail temporal Transformations. PATRICK A. CABE, Pembroke State ability of information in both lexical and episodic memory. This com University-Geometric analysis suggests that haptic information about plexity is readily apparent in the many dissociations reported for im the interval between two distal points is available via a string (with its plicit and explicit memory measures. However, studies using noncodified ends connected at those points) pulled into slight tension by an observer's events (pseudowords) yield a relatively uncomplicated picture of prim fmger actively exploring the elliptical path defined by the connection ing and repetition effects. For noncodified events, a simple episodically points (foci) and the constant length string. Naive observers make such based account appears sufficient for explicit and implicit tasks. interfocal distance judgments reasonably accurately. Thus spatiotemporal 2:40-2:55 (421) haptic/kinesthetic transformations afford usable distal layout informa Word Repetition as Eyes See It. MICHAEL E. J. MASSON, Univer tion. Some parallels to optical information are offered. sity of Victoria-A shared foundation for three word-identification par 2:15-2:35 (416) adigms-masked repetition priming, repetition blindness, and parafoveal Perceptual Asymmetries in a Haptic Texture Search Task. SUSAN previewing-is proposed on the basis of informational persistence that LEDERMAN, Queen's University, ROBERTA KLATZKY, University follows brief visual exposure to stimuli. Effects in these paradigms may of California, Santa Barbara, & CHERYL WILSON, Queen's Uni arise because informational persistence allows the visual system to as versity-A haptic equivalent of Treisman's visual search perceptual sign discontinuous views of a stimulus to a single episode. New evi asymmetry paradigm was performed with rough and smooth items, de dence is presented to support an orthographic basis for the contribution livered to the fmgers using a new custom-designed apparatus. A rough of informational persistence to masked repetition priming. target was consistently detected among varying numbers of smooth dis 3:00-3:20 (422) tractors more quickly than the reverse. The significance of different Implicit Memory lor Novel Visual Objects: Function and Struc search functions obtained with other manipuJations will also be discussed. ture. DANIEL L. SCHACTER, Harvard University, & LYNN A. This paradigm is being used to determine a reasonable set of tactile primi COOPER, Columbia University-Previous research has shown that prim tives for modeling the early stages of haptic processing. ing effects on a possible/impossible object decision task depend on en 2:40-2:55 (417) coding of global object structure. We found that study tasks that require Why Do We Touch Objects That We Can See? ROBERT A encoding of object functions produce quite different effects than do struc KLA TZKY, University ofCalifornia, Santa Barbara, SUSAN LEDER tural encoding tasks on object decision priming and explicit memory, MAN, Queen's University, & DANA MATULA, University of Califor respectively. Data are discussed with respect to structure/function dis nia, Santa Barbara-Subjects judged real objects on size, shape, weight, sociations in brain-damaged patients and in relation to the Gibsonian roughness, hardness, or temperature. We recorded not only the response notion of affordances. time, but whether touch was used, when it was initiated, and the dura tion of haptic exploration. Judgments ofstruClure (size, shape) appeared to use vision as a default, whereas difficult judgments of material used ATTENTION IV haptic exploration frequently and early. These results extend our model San Francisco A, Saturday Afternoon, 1:30-3:10 of haptic exploration to incorporate vision as an alternative "explora tory procedure." Chaired by Harold Pashler, University of California, San Diego REPETITIONIPRIMING EFFECTS m 1:30-1:45 (423) Embarcadero CD, Saturday Afternoon, 1:30-3:25 Processing ViwaI Stimuli Inslde and Outside the Focus of Attention. PAULA GOOLKASIAN, University ofNonh Carolino at Charlone Chaired by Jamie Campbell, University of Saskatchewan This research investigates attention to target and distractor stimuli that appear at locations from 0° to 15° to the right or left of a fixation point. 1:30-1:50 (418) The presentation location of the stimuli are varied so that in some con Semantic and Epbodic PrimIng In a Perceptual Identiftcation Task. ditions the target and distractor appeared together while in others they JEROEN G. W. RAAlJMAKERS, rNO Institute for Perception, RENE were presented at varied spatial locations. Results suggest that the in ZEELENBERG, University of Leidm, & CHRIS SCHRUNEMAKERS, fluence of a distractor on target processing is determined by its probable University ofNijmegen- We investigate the storage of new associations location and its compatibility with the target. in semantic memory. Three experiments are reported in which prime 1:50-2:05 (424) target pairs were repeated five times in a perceptual identification task. Saccadic Eye Movemeut and VIsual Selective Attmdon. BASKARAN Primes and targets could be semantically related, episodically related SUBRAMANlAM & JAMES E. HOFFMAN, University ofDelaware (unrelated but repeated), or unrelated. In two experiments, the episodic (read by James E. Hoffman)-The relationship between saccadic eye relation did not have any effect. In the third experiment, an effect was movements and covert orienting of visual spatial attention was inves- 510 Papers 425-436 Saturday Afternoon tigated in a dual-task study which required subjects to make a saccade 2:10-2:25 (430) to a prespecified position while detecting a briefly presented target let The Effect of the Uniqueness Point in Processlna Printed Words. ter, whose likely position was cued prior to each trial. The cue pro MONIQUE RADEAU,1OSE MORAIS, PHILIPPE MOUSlY, MARCO duced no benefit in target detection accuracy when it conflicted with SAERENS, & PAUL BERTELSON, Universiti Libre de Bruxelks the direction of the saccade, suggesting that saccadic eye movements The time course of lexical access in written word recognition is exam require spatial attention. ined by comparing words with early and with late uniqueness point (UP). 2:10-2:25 (425) Experiments using a normal (simultaneous) presentation of the letters Can an Eye Serve as an Effective Cue for Selective Attention? provided no evidence for sequential processing. Incremental presenta RUTH KIMCHI, University of Haifa, ORI TRAININ, & DANIEL tion of the word letters gave rise to a UP effect comparable in size to GOPHER, Technion, Haifa-In two experiments, subjects performed that obtained in an auditory study. Smaller UP effects were found with a target detection task under dichoptic viewing, using precuing proce syllable-by-syllable and phoneme incremental presentation. dure. The relative effectiveness of location precuing and eye precuing, 2:30-2:50 (431) and subjects' ability to utilize these two types of cues, were examined. Computational Modeling of Reading: Dual or Slop Route? MAX The results indicated that advance knowledge about the location of the COLTHEART, BRENT CURTIS, PAUL ATKINS, & ZOLTAN to-be-presented stimulus facilitated performance, and invalid informa SCHRETER, Macquarie University-Skilled readers can (a) read ex tion had an inhibitory effect. In contrast, advance knowledge about the ception words correctly; (b) read nonwords correctly; (c) perform visuaJ eye had no effect whatsoever. These fmdings suggest that an eye does lexical decision; (d) become surface dyslexic after brain damage; and not constitute an information channel for attention. (e) become phonological dyslexic after brain damage. Some theorists 2:30-2:45 (426) hold that these basic findings can only be explained by positing two Does Oculomotor Readiness Mediate Cognitive Control of Visual processing routes from print to speech (lexical and nonJexical). The 0De Attention? RevWted! AMANDA 1. PONTEFRACf & RA YMONO M. route connectionist Seidenberg-McClelland model challenges this claim. KLEIN, Dalhousie University (read by Raymond M. K1ein)-Are covert, We argue that this challenge fails, and introduce a dual-route computa endogenous shifts of visual attention accomplished by preparation to tional (DRC) model of reading. foveate the to-be-attended location? Klein (1980) described, tested, and (432) disconfmned this hypothesis. Given renewed interest in, and no fur (withdrawn) ther direct tests of, this hypothesis, we performed two new tests. We found that attending a location did not engender preparation to foveate HUMAN LEARN1NGIMEMORY m it and preparing to foveate a location did not engender an attentional Golden Gate A, Saturday Afternoon, 1:30-3:10 shift. Endogenous orienting is independent of oculomotor programming. 2:50-3:05 (427) Chaired by David G. Payne, SUNY at BinghamJon Va-patiaI Sustained Attention. PATRIZIA BISIACCHI & MADO PROVERBIO, University of Padua-Phasic orienting of spatial atten 1:30-1:45 (433) tion has received in recent years a great deal of interest. Our aim was Judging the Frequency of Action Commands. DARRYL BRUCE, to study spatial attention using a sustained attention task (i.e., main Saint Mary's University, & DAVID 1. MINGAY, NORC, University taining attention to visual locations throughout a series of events). We of Chicago-Individuals performed commands or imagined perform carried out two experiments in which subjects made speeded responses ing them. The commands were variations on a number of generic ac to peripheral stimuli in focused and diffuse attention conditions. Two tions. Frequency estimation and cued recall of the variations were tested. and four spatial locations were used in the first and second experiment, In comparison with individuals who imagined the commands, those who respectively. The results are discussed in the light of the models pro performed them typically showed poorer frequency estimation, better posed for describing properties of covert orienting mechanisms. recall, and higher correlations between judged frequency and recall, in dicating greater reliance on the retrieval of instances of generic actions LETTER/WORD PROCESSING n to estimate frequency. San Francisco B, Saturday Afternoon, 1:30-2:55 1:50-2:00 (434) Learning Names of Faces with Imagery Techniques. LOWELL D. Chaired by Keith Rayner, University of Massachusetts GRONINGER & JEFFERY STIENS, University ofMaryland Baltimore County-Videotapes of students given concrete (Starr) and abstract 1:30-1:45 (428) (Anderson) names were shown to subjects at a 20-sec rate under one String Length and Repetition Affect Lexical Decisions about Dis of three instructional conditions: (a) form an image of the name with oriented Letter Strings. KEVIN JORDAN & LAREE A. HUNTS a feature of the person, (b) form a bizarre image of the name with a MAN, San Jose State University-Comparison of studies reporting an feature of the person, or (c) learn the name of the person. The imagery influence of disorientation on words but not alphanumerics confounds techniques were better under both recall and recognition tests. stimulus set size (and hence repetition) with stimulus type. We report 2:05-2:15 (435) that when letter strings are not repeated, there are persistent effects of Tip-of-the-Tongue States and Blockers with Imaglnary AnimaI disorientation on lexical decision times. However, when a small, six Targets. STEVEN M. SMITH, Texas A&M University-After study item set of letter strings is repeated at each of six orientations, the ef ing a list of imaginary animals, subjects recalled the animaJ names, given fects of disorientation are absent or diminish quickly. The results may pictures as cues. More tiJHlf-the-tongue (TOT) states and recall failures indicate multiple representations with broad tuning for orientation. were reported when an incorrect imaginary animal name accompanied 1:50-2:05 (429) the cue than for real animal names, unrelated words, or letters. Name What Makes Targets Redundant? G. ROBERT GRICE, Univer reca1l was blocked only by other target names. The phonological similar sity ofNew Mexico-In golno-go reaction time experiments investigat ity of targets and blockers had no effect on TOT levels. ing redundancy phenomena with letter stimuli, the effects of target redun 2:20-2:30 (436) dancy were greater if the redundant targets were different members of Concreteness Effects in Free RecaU: On Relational and DIstiDc the same response class than if they were physically identical. Analysis tive ProcessIng. MARC MARSCHARK, University ofNorth Carolina also supported a process of coactivation rather than serial processing. at Greensboro, & LUCA SURIAN, MRC Cognitive Psychology Unit, Apparently, nonidentical stimuli may be quickly processed parallel to London-Recent research on imagery and memory suggests that con the level of response evocation. Similar results were not obtained with creteness effects may depend on joint distinctive (imaginal) and rela the choice RT procedure. tional (interitem) processing. The relative contributions of the two were 511 Saturday Afternoon Papers 437-447 distinguished using imagery and categorization tasks with related and closely and distantly related species. Research on song perception and unrelated lists of concrete and abstract words. Related abstract words memory for stored food in birds provides examples. were consistently reca1led as well as related concrete words. Unrelated 2:25-2:45 (442) words (the usual scenario) yielded concreteness effects only after both The Pbylogeny and Ontogeny of Cognition. H. S. TERRACE, relational and distinctive processing, a theoretically constraining result. Columbia University-The opportunity to study cognitive processes in 2:35-2:50 (437) animals provides a unique opportunity to investigate the phylogeny and Keywcrd I..ean*JK aod Readioa: Ea!iy Come, Ea!iy Go. MARGARET ontogeny of cognition. Much of our understanding of human cognition THOMAS, ALVIN Y. WANG, KATHLEEN GUEST, & PEGGY derives from the intensive study of adult human subjects. Serial learn NYLANDER, University of Central Florida-The keyword mnemonic ing provides an instructive example. Since Ebbinghaus, experiments on yields impressive gains in learning speed when compared with other serial learning have been performed almost exclusively on subjects who methods of studying second-language vocabulary. However, in several have had much experience learning lists by virtue of their formal and experiments, forgetting rates were actually greater for keyword learn information education. Recent research on serial learning in pigeons ing than for rote rehearsal. Possibly, preexperimental associations to and monkeys provides a new perspective on this fundamental skill, one the keywords regain their prominence over time and inhibit retrieval which does not require language and which appears to be phylogeneti of the keyword-translation image. Indeed, forgetting in keyword learners cally quite old. Of greater interest are qualitative differences in what was dramatically accelerated when they free associated to the keywords pigeons and monkeys know about a list as a consequence of having before vocabulary learning. learned to produce it. Discussion of these observations will place in per 2:55-3:05 (438) spective similarities and differences in animal and human cognition. Category Information and Verbal Recall. MA TIHEW J. SHARPS, 2:50-3:10 (443) California State University, Fresno, & MICHAEL TINDALL, North The Experimental Analysis of Cognition in Animals. ROBERT western University-Although items organized by category are gener COOK, Tufts University-The recent influence of cognitive science and ally better recalled than unorganized items, it was recently demonstrated ethology on the contemporary study of animal cognition is discussed. that this effect is largely confined to verbal stimuli. This finding may Whether or not the resulting changes in comparative cognition research reflect differences in the processing of item-specific and relational in represent an example of a scientific paradigm shift is evaluated. Some formation. If so, the mnemonic effect of organization for verbal stimuli evidence of the historical and current relations between the study of cog should diminish over long retention intervals, and this diminution should nition in animals and humans will be presented. Finally, several difficul be attenuated by information, presented at encoding, about stimulus or ties and problems with our current research strategies are considered. ganization. The present results verified these hypotheses. 3:15-3:35 (444) Human Models of Animal Behavior. RUSSELL M. CHURCH, SYMPOSIUM D: Brown University-Similarities between the behavior of humans and other COMPARATIVE COGNITION: TAKING STOCK animals in analogous tasks suggest that similar cognitive processes are Golden Gate B, Saturday Afternoon, 1:304:00 involved. Human studies provide an efficient way to collect large amounts of data with low variability, and animal studies provide a way to estab Chaired by Stewart H. Hulse, Johns Hopkins University lish generality, limit the range of plausible explanations, and develop explanations in terms of brain mechanisms. Examples will be taken from 1:30-1:35 (439) studies of timing that used similar procedures for humans and other Introductory Remarks. STEWART H. HULSE, Johns Hopkins animals. University-It has been 15 or 20 years (depending on when you start) 3:~3:55 (445) since we began to entertain seriously the idea that animals might be cog Discussion. DONALD A. RILEY, University ofCalifornia, Berkeley. nitive creatures, and a lot has happened in the interim. It seems appropri ate to take stock. Where has the field been? What are some possible JUl)GMENTIDECISION MAKING I developments in the future? What has been the impact of studies with Embarcadero AR, Saturday Afternoon, 3:10-4:55 nonhuman subjects on the search for facts and the development of the ories in the human domain? Are we making animals miniature, rather OIaired by Thomas S. Wallsten, University of North Carolina at OIapel HiU cognitively impoverished people? What is the evidence that nonhuman species are cognitively different from each other and from humans? These 3:10-3:25 (446) issues provide for lively discussion, and this symposium offers just that. Internal Attention and the Consideration of Decision Alternatives. 1:35-1:55 (440) MICHAEL E. DOHERTY & CLIFFORD R. MYNATT, Bowling Comparative Cognition: Toward a General Unclerstanding of Cog Green Stale University-An attention hypothesis is proposed to account nition in Behavior. EDWARD A. WASSERMAN, University oflowa for the failure to consider alternatives on "inferences" involving deci Human and nonhuman anirnals a1ike must adjust to complex and changing sions between unknown states of the world. It predicts that people will circumstances if they are to survive and reproduce. Advanced neural consider alternatives on "actions". involving decisions between courses mechanisms enable animals to remember the past, to act in the present, of action and on inferences in which attention is drawn to alternatives. and to plan for the future. Exploring the species generality of cognitive Subjects selected information about one alternative when given two processes in behavior is central to the field of comparative cognition. alternative inferences but about both alternatives when given actions. A comparative perspective may not only broaden, but deepen our un Manipulation of p(D/H) led to increased consideration of alternatives derstanding of cognition-in both human and nonhuman animals. with inferences but not with actions. 2:00-2:20 (441) 3:30-3:50 (447) Where is the Compariaon in Comparative Cognition? SARA The Accuracy-Informativeness Tradeoff in Judgment. !LAN SHETTLEWORTH, University of Toronto-The implicit purpose of YANIV, University of Chicago-In making judgments, individuals often most work on comparative cognition is to discover whether nonhuman communicate their belief or confidence via the graininess of their judg species process information in the same way as humans do. This research ments. For instance, in estimating arrival time, the communications program goes back to Darwin and Romanes and has the same theoreti "5:00 p.m." and "fiveish" have different grain sizes; the fIrSt is more cal weaknesses as earlier work on "animal intelligence." An alterna finely grained than the second. We propose that the selection of grain tive program compares ecologically significant aspects of cognition in size in judgment under uncertainty depends on a tradeoff between ac- 512 Papers 448-459 Saturday Afternoon curacy and informativeness and report studies in which subjects produced versity-Psycholinguistic investigations of morphological processing and judgments and evaluated others' judgments. representation have been restricted to typologically similar languages 3:55-4:10 (448) with relatively impoverished morphologies. This research explores the Reftection of Risk Preferences and Scale of Magnitude. KRISTINE time course of the morphological analysis of an agglutinative language KUHN, University ofIllinois, LOLA L. LOPES, & IRWIN P. LEVIN, (Turkish) by means of a series of on-line experiments (phoneme monitor University of Iowa (read by Lola L. Lopes)-We investigated the im ing and nonword decision). The results suggest early and immediate pact of numerical quantity on risky choice. Three groups of subjects activation of the phonological representation of the root and a sequen made choices among both gain and loss lotteries expressed in artificial tial analysis of the suffix string. currencies: small (EV = 1.(0), medium (EV = 1(0), and large (EV = 4:45-5:05 (454) 100,(00). For gains, the majority of subjects were risk averse in all Monitoring for Syllables In Morpbological.ly Complex Words. conditions. For losses, the majority of subjects were indeterminate for PIENIE ZWITSERLOOD, Max Pionck InstiMe, The Netherlonds (sp0n small and medium conditions and risk seeking for large conditions. sored by Arthur G. Samuel)-Experiments in Dutch and German showed 4:15-4:30 (449) that listeners are sensitive to the syllabic structure of spoken words of Group Versus Individual ProbabWty Judgment: Accuracy and their native language. Subjects monitored for the presence of visually Process. J. FRANK YATES & HUN-TONG TAN, University of or auditorily presented targets in disyllabic, polymorphemic spoken car Michigan-Along what dimensions do individual and group probabil rier words. A clear effect of syllabic match between target and spoken ity judgments tend to differ in their accuracy? What mechanisms ac word was obtained for derived and inflected words but no( for com count for such differences? To pursue these issues, subjects were asked pounds. A final study investigated whether acoustic or lexical informa to make judgments about general knowledge questions individually and tion is responsible for this difference. in triads. Groups were better at selecting correct alternatives and in the overall accuracy of their probability assessments. Notably, however, ATIENTIONV overconfidence was unaffected by group interaction. Analyses also sup San Francbco A, Saturday Afternoon, 3:20-5:00 ported a signed summation model of polarized group opinion. 4:35-4:50 (450) Chaired by John Palmer, University of Washington The Structure of Arguments In ProbebWty Ass mnt. SHAWN P. CURLEY, GLENN J. BROWNE, GERALD F. SMITH, & P. GEORGE 3:20-3:40 (455) BENSON, University of Minnesota-Reasoning is a central cognitive Linguistic and Conceptual Control of Visual Spatial Attention. activity in probability assessment, and one that has been understudied. GORDON D. LOGAN, University of Illinois-Conceptual relations We collected verbal protocol data as subjects performed several proba (above, below, left, right) were used to cue positions of targets in sim bility judgment tasks. In response, subjects constructed arguments toward ple and complex arrays. Time to report target identity varied systemat forming beliefs and their associated probabilities. Motivated and guided ically with the cuing relation. The results violated the hypothesis, im by a cognitive analysis of probability assessment, these data were ana plicit in current space-based theories of attention, that all display 1ocations lyzed for the structure of the subjects' arguments, including argument are equally available. Instead, they suggest that the availability of dis content, argument types, and argument strategies. play locations depends on their relation to the spatial reference frame that the subject imposes on the display. PSYCHOLINGUISTICS m 3:45-4:05 (456) Embarcadero CD, Saturday Afternoon, 3:35-5:10 FJreets of AtteIItioo on LaJcdl and DiI!ItIInce E...... YEHOSHUA TSAL & Lll.ACH MEVORACH, Tel Aviv Univemty--Subjects estimated Chaired by Michael K. Tanenhaus, University of Rochester the length of a vertical line and the distance between two vertically dis placed dots. For both types of judgments, estimates were consistently 3:35-3:50 (451) shorter when attention was directed to the stimulus than when attention Morphological PrimIng at Long and Short Lags in Visual Word was directed away from it. Results suggest that attentive and unatten Recognition. LAURIE B. FELDMAN, SUNY at Albany and Haskins tive judgments are mediated by two distinctive modes of processing. Laboratories, & DARINKA ANDJELKOVIC, University ofBelgrade 4:10-4:20 (457) Effects of orthographic similarity with and without morphological related Far Peripheral Cues Can Do Better Tban Near Cues. VERON ness were investigated in a lexical decision task with Serbo-Croatian ICA J. DARK & MICHAEL B. TEPIN, Iowa State University-Targets materials at lags of 10 and 0 items (SOA 3(0). Facilitation due to mor occurred at two locations either alone or with an X at the opposite loca phological relatedness was evident at long and at short lags. Inhibition tion. Cues occurring 100 msec prior to the target were central (20%), due to orthographic similarity (without morphological relatedness) was peripheral near (20%), or peripheral far (60%). Peripheral cues pro absent at lags of 10. At lags of 0, inhibition was weak and did not inter duced faster identification than central cues, with the fastest identifica act with relative frequency of prime and target. tion for far cues preceding a lone target. The distractor was more 3:55-4:15 (452) detrimental with far than near cues. Results suggest location priming Intonation Lut: The Utility of Dilferent Cues to Syntac:tk: Struc in addition to .. directing" attention. ture. VIRGINIA V ALlAN, Hunter College, CUNY, & ANDREA G. 4:25-4:40 (458) LEVITI, Wellesley College and Haskins Laboratories-College stu Suppression of the AbWty of Abrupt Onsets to Capture Visual dents learned an auditorily presented miniature artificial language in three Attention. HIDEYA KOSHINO, C. BRUCE WARNER, & JAMES F. experiments. We evaluated the roles of reference, marker frequency, mOLA, University ofKansas (read by James F. Juola)-In several ex and intonational cues to phrases (independent of pausing). Our results periments, the detection of a target among noise characters was affected suggest that learners utilize intonation only when referential support is by the validities of various spatial cues. Although abrupt onset of a sin lacking, and primarily when marker frequency-a cue to syntactic gle character was shown to be a powerful attractor of attention, in some structure-is low. We hypothesize that adult learners preferentially at conditions the prior presentation of a central or peripheral cue indicat tend first to referential cues to structure, then to high-frequency syn ing another position reduced the apparent attentional capture by the on tactic cues, and fmally to intonational cues. set character. The results are relevant for determining the commonality 4:20-4:40 (453) of automatic and voluntary attentional processes. Morphological Parsing in Turkish: An On-Line Experimental 4:45-4:55 (459) Study. UU FRAUENFELDER, Max Pionck Institute, The Netherlonds, Quantitative EEG during Dilferent Levels of Performance on a JORGE HANKAMER, University of California, Santa Cruz, JAKLIN Vigilance Task. DOMINIC A. VALENTINO, JAMES E. ARRUDA, KORNFILT, Syracuse University, & SUMRU OZSOY, Bogazi,i Uni- SHERRl M. GOLD, MICHAEL D. WEILER, MARY RIEDFORD, 513 Saturday Afternoon Papers 460-469 & LUZ TEIXEIRA, University of Rhode Island-Twenty-seven sub reactions are perfonned faster if there is spatial correspondence between jects perfonned an auditory continuous performance task (CPT) requiring stimulus and response, even if stimulus position is irrelevant for the task them to direct targets (consecutive, identical letters) in a series of let (Simon effect). Experiment I shows that the crucial variable in deter ters presented at a rate of 2/sec. EEG was sampled during early and mining the Simon effect is spatial correspondence between stimulus and late CPT performance. Detection accuracy decreased for all subjects action goal (i.e., intended action effect), the latter being manipulated between these periods. Comparisons of EEG power spectra showed dif by different instructions. Results of Experiment 2 indicate that non-goal ferences for frontal delta and theta frequencies and for frontotemporal related correspondences between stimulus position and position or ana and temporal beta frequencies (they each decrease). Alpha was unaffected. tomical mapping of the hand contribute to the Simon effect as well by decreasing or increasing its size. LETI'ERIWORD PROCESSING m 3:45-4:05 (465) San Francisco B, Saturday Afternoon, 3:30-5:05 Impulsivity and Neural Noise in Perceptual-Motor Performance. JENNIFER GLASS & DAVID E. MEYER, University of Michigan Chaired by Philip A. Allen, Cleveland State University (read by David E. Meyer)-Subjects who scored high, medium, or low on a self-report scale of impulsivity tried to hit spatial targets by mov 3:30-3:50 (460) ing either at prespecified velocities or as rapidly as desired. From pa Regularity Effects in Word Recognition ReconceptuaIized. rameters of obtained speed-accuracy tradeoff curves, it appears that high RICHARD K. WAGNER, WILLIAM L. OLIVER, & CAROL A. impulsives may prefer to move faster than low impulsives because there RASHOTIE, Florida State University-The ease with which a back is less neural noise in high impulsives' motor systems. This supports propagation model learns a pattern provides a measure of its regular some extant theories about how impulsivity and cortical arousal are ity. Regular patterns are those patterns that occur frequently and have related. few conflicting patterns in the training set. Using a backpropagation 4:104:30 (466) model, we derived phonological, orthographic, and orthographic-to Intersensory Facilitation with Multiple Stimuli: A Test of Com phonological regularity measures for 5,000 English words. We com peting Models. ADELE DIEDERICH & JEROME BUSEMEYER, pared these regularity measures to traditional measures of word charac Purdue University (read by Jerome Busemeyer)-An experiment was teristics and examined the extent to which they account for variability conducted investigating intersensory facilitation of reaction times where in naming latency. stimuli of three different modalities (light, tones in two different inten 3:55-4:15 (461) sities, and vibration in two different intensities) were presented, sepa The Role of Phonology in the Activation of Meaning. DEBRA rated by various interstimulus intervals. Three different models to ex JARED, McMaster University, & MARK S. SEIDENBERG, University plain the observed facilitation of reaction time are presented and tested: ofSouthern California (read by Mark S. Seidenberg)-Six experiments a separate action model, a superposition counter model, and a diffusion addressed the extent of phonologically based activation of meaning in model, the last two belonging to the class of coactivation models. visual word recognition, using variants of the semantic decision task 4:35-4:55 (467) developed by Van Orden. The results indicate that phonologically based Involuntary Capture of Spatial Attention Is Contingent on Con activation of meaning is limited to lower frequency words and nonwords. trol Settings. CHARLES L. FOLK, Villanova University, JAMES C. The studies also suggest that subjects cannot strategically control the JOHNSTON, & ROGER W. REMINGTON, NASA-Ames Research activation of phonological codes. Implications concerning the division Center (read by James C. Johnston)-Previous research suggests that of labor in dual-route models of word recognition are considered. involuntary capture of spatial attention is triggered by some cue properties 4:204:40 (462) (e.g., abrupt onsets) and not others (e.g., color). In these experiments, Phonetic Recoding 01 Phonologically Ambiguous Printed Words. however, abrupt onsets were used to find target stimuli. Newexperi RAM FROST, Hebrew University ofJerusalem-Speech detection and ments show that when color is used to fmd targets, abrupt onset cues matching simultaneously presented printed and spoken words were used do not trigger involuntary attention shifts and color cues do. We argue to examine phonologic and phonetic processing of Hebrew heterophonic that attention capture can be involuntary, but contingent on strategically homographs. The results suggest that the visual presentation of phono alterable control settings. logically ambiguous letter strings automatically activated both the dominant and the subordinate phonologic alternatives these letter strings REPETITION/PRIMING EFFECTS IV represent. This result provides further confrrmation for fast and man Golden Gate B, Saturday Afternoon, 4:10-5:10 datory phonetic recoding in reading. 4:45-5:00 (463) Chaired by Ira H. Bernstein, University of Texas at Arlington Phonology and Orthography in Visual Word Recognition. JONA THAN GRAINGER, CNRS, Paris, & LUDOVIC FERRAND, Rene 4:104:20 (468) Descanes University-Some masked priming experiments independently Cross-Modal Priming in Word Fragment Completion. WAYNE manipulating orthographic and phonological prime/target overlap are DONALDSON & ROBIN GENEAU, University ofNew Brunswick described. With nonword primes, phonology provided clear facilitation Why does auditory presentation prime visual word fragment comple over and above orthography. With word primes, facilitatory effects only tion? Possibilities include visualization of presented words and the use occurred in conditions of rnaximal prime/target overlap and were in of auditory infonnation on the visual test. Participants saw or heard words fluenced by the type of nonword distractors. Possible sublexical and and were prevented from cross-modal processing (i.e., those who saw postlexical mechanisms underlying facilitatory form priming effects in could not verbalize and those who heard could not visualize). Visual lexical decision are discussed. word fragment completion performance was reduced only for the visual group who could not verbalize, suggesting an auditory component on INFORMATION PROCESSING m the test. Golden Gate A, Saturday Afternoon, 3:20-5:00 4:25-4:45 (469) Specificity 01 Operations in Perceptual Priming. HENRY L. Chaired by Paula Goolkasian, University ofNorth Carolina at Charlotte ROEDIGER m, Rice University, SUPARNA RAJARAM, Temple University School ofMedicine, KAVITHA SRINIV AS, Boston College, 3:20-3:40 (464) & KATHLEEN McDERMOTI, Rice University-We report several Inverting the Simon Effect by Intention. BERNHARD HOMMEL, experiments in which priming on perceptual implicit memory tests is Max Planck Institute, Munchen (sponsored by Wolfang Prinz)-Choice highly dependent on the match of perceptual operations applied to study 514 Paper 470 Saturday Afternoon and test events. For example. picture fragments are primed more by GLAVIN. Wayne Slate University-Word, picture, and mixed picture pictures than by words and more by fragmented than by intact pictures. word pairs were presented at varying SOAs for superordinate cate Generating images of the relevant scene enhances priming. but gener gory decision. Pairs differed in the name/rhyme similarity of the ating the name of the concept from a semantic clue does not. Similar objects represented by the pictures and words. Name/rhyme sim results were obtained with verbal perceptual tests. ilarity influenced category decision for all pair types. Comparison 4:50-5:05 (470) with similar studies varying object shape similarity suggests some dif NamelRbyme SimIlarIty Effects on Category Dedsioas about PIc ferences in accessing and activating perceptual shape and lexical tures and Words. PATRICIA SIPLE & ANNEMARIE CWIKIEL- information. 515 Saturday Evening Posters 471-482 POSTER SESSION m lidity of the dimension. As a further test of this theory, the salience of Pacific Concourse East, Saturday Evening, 5:45-7:15 an irrelevant dimension was varied. As predicted by the theory, no ef fect was found on the utilization of a second relevant dimension. (471) (477) Some Subjects Are Not Influenced by How a Problem Is Framed. There's More Than 0Jt Reason for tbe Base-Rate Fallacy. DANIEL REISBERG & DAVID GOSSETI, Reed College-Subjects' CHRISTOPHER R. WOLFE, Miami University (sponsored by Micki preferences are dramatically influenced by how a problem is framed; Chi)-Subjects in two experiments (n = 192) received a series of Bayes in the extreme, subjects can be led to contradict themselves by shifts ian conditional probability problems and a menu of available informa in framing . However, not all studies of framing effects replicate these tion including the base rate, hit rate, and contropositive. Subjects gener results. We report data showing that some subjects seem systematically ally utilized the hit rate but ignored the base rate and contropositive. immune to framing effects; we document that this is not merely a con Measures of reasoning indicate that subjects ignore base rates due to sequence of some subjects being more careful in their decision making. misconceptions about the meaning of the base rate and hit rate, frequently Other possible factors potentiating framing effects are discussed. believing that p(AIB) = p(BIA). These "conversion errors" are as (472) sociated with the base-rate fallacy. An Alternative Interpretation of Framing Effect in Decision Mak (478) ing. JERWEN JOU, Georgia Southern University, JAMES SHAN A Rational Use of Positive and Negative Testing. ANGELA TEAU, & RICHARD JACKSON HARRIS, Kansas State University FRATIANNE & PATRICIA W. CHENG, UClA (preseOed by Patricia W. (sponsored by James Shanteau)-Tversky and K.ahneman (Science, 1981) Cheng)-Hypothesis-testing research has shown that people tend to test demonstrated that people showed a risk -aversion or a risk -seeking choice only positive cases of either the hypothesized condition or the target preference depending on whether a choice problem was presented in phenomenon. We examine positive testing in terms of a covariation a gaining or a losing frame. The authors suggested that such a reversal based causal reasoning and distinguish between a rational and irrational of preference across the two framings indicated inconsistency or irra use of positive tests. We report results showing that subjects obtain and tionality of human judgments. The present study suggested that atten evaluate all information necessary for a normative causal decision and tional dissociation rather than inconsistency may be an alternative in that they spontaneously relinquish ineffective positive tests and use nega terpretation of such choice behavior. tive tests when appropriate. (473) (479) Effects of Information Framing and Involvement in Investment On the Need to Take Action When Heading toward Defeat. ELDAR Decisions. IRWIN P. LEVIN, University of Iowa, ANTHONIUS SHAFIR & DALE MILLER, Princeton University-In a variety of BAGGERMAN, 1ilburg University, & GARY J. GAETH, University decision-making situations offering the options of action and inaction, of Iowa-Subjects distributed a sum of money across two stocks and subjects' preference for action is found to increase as a function of the a savings account. Realistic stock descriptions were framed in equivalent likelihood and gravity of an impending defeat. Contexts investigated positive or negative terms (e.g., chances of going up or chances of going include naturalistic exam-taking settings, as well as a number of hypothet down) for different subjects. Involvement was manipulated by using either ical decision scenarios. The propensity to act in panic situations is dis hypothetical investments or real monetary payoffs based on actual stock cussed in the larger context of rational decision making under pressure. performance. Consistent with an anchoring-and-adjustment model, framing (480) effects were found with low involvement but not with high involvement. Age and Sex Differences in Estimates of Long-Term Financial (474) Growth Functions. DOUGLAS A. HERSHEY, DARLA SWANN, & The Role of Emotion in Framing Effects. DEBORAH FRISCH & LESLIE BURKE, George Mason University (sponsored by David A. MICHAEL MURIAS, University of Oregon (sponsored by Douglas Walsh)-PeopJe have been found to make systematic errors in estimat Hintzman)-Framing effects demonstrate that seemingly minor changes ing the value of various types of long-term financial growth functions. in decision presentation influence subjects' choices. In two experiments, It is unclear, however, whether the ability to make accurate fmancial we examined the effect of framing on subjects' predictions of how they estimates varies as a function of age and sex. In the present study, 192 would feel if different outcomes occurred. Framing was found to in subjects who ranged in age from 20 to 79 made 24 estimates of long fluence choice to the extent that it altered the perceived desirability of term fmancial yield. Results indicate that neither age nor sex were con the different possible decision outcomes. These results suggest that, sistently associated with the accuracy of subjects' estimates. although framing effects violate utility theory, they may reflect accurate (481) and therefore normatively justifiable decisions. The Decision to Buy Housing: Mental Models and Expected Util (475) ities. DAVID A. WALSH, RUBY R. BROUGHAM, University of The Role of Computational Limitations in Producing "Neglect" Southern California, & HAZEL SPEARS, Princeton University of Prior Probabilities. MIRIAM W. SCHUSTACK , California State Expected utility theories propose that people's behavioral choices are University, San Marcos (sponsored by Patricia Worden)-The judgment based on their evaluation of the probabilities and the hedonic values they literature abounds with examples of departures from normatively associate with outcome consequences. The present research examined prescribed use of prior probabilities. Often, when information is provided the relationship between the expected utilities that 438 subjects assigned about the individual case being judged, effects of the prior probability to the options of renting and buying housing and their actual housing are inappropriately reduced or eliminated. My experimental findings behavior. Furthermore, we examined the influence of subjects' mental suggest that computational limitations play an important role in this models of financial knowledge on their expected utilities for consequences "neglect" of priors-people do not simply behave in accordance with associated with buying or renting housing. an erroneous belief or procedure whereby any case-specific informa (482) tion supersedes all prior-probability information. Contingent Valuation: Stability and Meaning of "Not Willing to (476) Pay" Responses. ROBERT M. HAMM, LYNNE BENNETI, KARL Salience of Irrelevant Information in a Probabilistic Environment. WUNDERLICH, & CHUCK HOWE, University of Colorado at JON M. BARNES & STEPHEN E. EDGELL, University ofLouisville Boulder-Citizens' willingness to pay (WTP) for gain is used to evalu (presented by Stephen E. Edgell)-Earlier work has shown that the phys ate public goods. Reliability of willing versus not willing responses was ical representation of a relevant dimension in nonmetric multiple-cue studied in a survey of WTP for increased water supply reliability. More probability learning affects the utilization level. This salience effect was than 25% of responses were reversed on telephone follow-up of mailed theoretically explained by errors in memory affecting the perceived va~ survey. Explanations for not being willing to pay were coded into 12 516 Posters 483-494 Saturday Evening categories that have different implications for value of the good. Sug (489) gestions for revised formats for WTP questions are made. Recognition Memory for Auditory Event Duration, MARILYN (483) SOLTZ, Haverford College-The purpose of this research was to ex A~ ProductIon aod Copitive Activity: A GIana! at the inter amine the effects of structural relationships on the recognition of event action. PHILIP H. MARSHALL, Texas Tech University-Using a durations. In Experiment I, subjects were asked to perform a series of laboratory analogue of a portrait photography situation, Marshall and perceptual ratings on computer-synthesized sounds that varied in their Thornhill (1990) determined the effects of certain subject matter charac internal structure (continuous; rhythmically predictable, unpredictable) teristics (novelty and affect) on the probability a photograph would be and ending (arbitrary, nonarbitrary) before performing a surprise mem taken and on the time course for making those decisions. This method ory test for the events' durations. Experiment 2 replicated this design of investigating aesthetic production was used in the present study to but relied on sounds from the natural environment. The results of both determine possible interactive effects between aesthetic production judg studies indicated that learning and subsequent remembering was superior ments and ongoing cognitive activity. for events containing a nonarbitrary ending and/or rhythmically pre (484) dictable structure. These findings are discussed in terms of a time esti Fonnat Changes on Ballot Propositions Affect How Voters Vote. mation model that emphasizes the role of event structure on cognitive DIANE F. HALPERN, DEBORAH CRISTAL WAITES, DAWN behavior. FISHER, & ERIC GUERRA, California SlJJJe University, San (490) Bernardino-Most voters are poorly informed about election ballot Attention Goes A-roving: Frequency Cuing Aids Frequency-Roved propositions that require them to make decisions about a wide range Auditory Intensity Resolution. LAWRENCE M. WARD, University of complex and diverse topics. We found that the information in these ofBrtitish Columbia, & SHUn MORI, Nagasaki University-Analogous propositions could be simplified with a modified outline format and that to spatial attention cuing in vision and frequency cuing in auditory de voters respond differently when costs are presented as cost per house tection, frequency cuing aids auditory intensity resolution in a roving hold or as cost for the entire state. frequency one-interval paradigm. Three experiments demonstrate an un (485) certain frequency effect for intensity resolution, and that both auditory Seeding the Knowledge-Base: A Method for AnaIyziDg aod improv cuing (a brief tone at the frequency at which the to-be-identified tone ing Real-World Estimation. NORMAN R. BROWN & ROBERT S. would appear) and informational cuing (a visual indication of the rele SIEGLER, Camegie-Me/Jon University (sponsored by Robert S. Siegler) vant frequency) improve performance over both uncued and uninfor Estimates of national population are typically inaccurate in absolute terms matively cued conditions. and moderately accurate in relative terms. In four experiments, we (491) "seeded the knowledge base" by presenting actual populations of a few Perceptual Factors In Aging and Auditory Selective Attention. countries. When arbitrary seed countries were presented, mean estimates JOAN M. McOOWD & DANA R. MURPHY, University ofSouthern for "transfer" countries improved, but rank order did not. However, California (sponsored by Joseph B. Hellige)-Twenty young and 20 older when countries were selected to disconfirm specific misconceptions, both adults received audiometric work-ups, then repeated words in three selec absolute and relative improvements were evident. Thus, range and rank tive attention conditions: targets in one ear without noise, with noise ordering factors influence quantitative estimation. presented to the other ear, and with noise presented to the same ear. (486) Presentation levels and signal-to-noise ratio were carefully controlled. Weighting of Features in JudgmeDts of Superiority aod Inferiority. An overall age-related deficit was obtained, but no age x condition inter KIMIHIKO Y AMAGISHI & JOHN M. MIYAMOTO, University of action. These results highlight the importance of controlling hearing dif Washington (presented by John M. Miyamoto)-We compared the ferences in studying auditory selective attention. weighting of attributes in superiority judgments ("How much better is (492) A than B?") to the weighting in inferiority judgments ("How much worse Induction Field Orpnization aod Sim~ Contrast. TlZIANO is B than A?"), where A and B are multicomponent stimuli like descrip AGOSTINI & WALTER GERBINO, University of Trieste (sponsored tions of apartments. An extension of Tversky' s contrast model predicts by Dennis Proffitt)-We argue that when a black-white induction field the impact of assessment type (superiority vs. inferiority) on weights is articulated into a complex figure-ground configuration, this can affect assigned to good and bad features of the superior and inferior options. subjects' simultaneous lightness contrast (SLC) judgments. In two ex The model was tested against the results of two experiments. periments, we manipulated a gray target belongingness with respect to (487) either the white or the black ground, and to the white or the black figure. Discrimination of Auditory Fractal Contours. MARK A. Results support the view that lightness judgments are independent from SCHMUCKLER, University of Toronto, Scarborough Campus, & the black-white factor, whereas they are influenced by figure-ground DAVID L. GILDEN, Vanderbilt University (sponsored by David L. belongingness of the target. Gilden)-Discrimination ability for a family of fractals known as scal (493) ing noises was examined. Given that many natural auditory and visual Laterality In Haptic Exploratory Strategies. JOEL FAGOT, sources can be characterized as scaling noises, they are of paramount AGNES LAGREUSE, & JACQUES VAUCLAIR, CNRS-LNF. Mar ecological significance. We found that listeners' discrimination of au seille (sponsored by Jacques Vauclair)-Forty-eight right-handed adults ditory pitch contours, derived from different scaling noises, varied as of both sexes had to monohaptically explore unseen meaningless stimuli a function of fractal dimension. Subsequent analyses compared these during 4 or 10 sec, and then to recognize their outline drawing. infor discrimination functions to theoretical characterizations of the information mation on hand strategies were recorded along with the measurement contained in these noises by modeling listeners' judgments. of accuracy. Results indicate: (l) strong sex and hand differences in (488) terms of exploratory strategies; (2) no hand difference in terms of ac Melodic Tone Duration and Perceptual Judgments. BERNICE curacy for the IO-sec span. It is concluded that the measurement of ac LADEN, University of Washington (sponsored by Earl B. Hunt)-Does curacy is not always appropriate to demonstrate laterality effects and tone duration influence perceptual judgments about music? Listeners sex differences. heard short chromatic melodies followed by a target chord and rated (494) how well the target chord fit the melody. Melodies were presented with The Role of Strategies in Choice Reaction 1bne Tasks, ROBERT S. several different durational patterns, including an isochronous one. McCANN, Sterling Federal Systems, CHARLES L. FOLK, Villanova Results provide evidence for an interaction between durational pattern University, & ROGER W. REMINGTON, NASA-Ames Research Colter and target chord. (sponsored by Cynthia NuU)-Subjects were cued to expect either a lex- 517 Saturday Evening Posters 495-506 ical decision task or a numbers inequalities task. Cuing effects were one stimulus in a multiletter display by a precue (0-200 msec before additive with manipulations of task difficulty, such as word frequency display onset). Stimuli varied in size, shape, and color. Subjects looked and numerical distance between the digits. These results suggest that for I, 2, or 3 specific cued stimulus features and indicated their presence when cues were invalid, processing on the actual task presented could or absence using choice RT. Results supported a gradual attention not begin until the incorrect task set was abandoned, and the correct process. Also, consistent with template theory, detection performance task set was loaded. Implications for strategic involvement in speeded increased as the features to be identified increased. tasks are considered. (501) (495) Symmetry, Perception, and Figure-Ground Parsing in Neglect. Attentional Mechanisms in Processing Hierarcbical Patterns. JON DRIVER, University of Cambridge, ROBERT RAFAL, Univer LYNN ROBERTSON, University of California. Davis. and VAMC. sity of California. San Diego. and VAMC, Martinez, & GORDON C. Martinez, & ROBERT EGLY, University ofCalifornia. Davis-Evidence BAYLIS, University of California. San Diego (presented by Gordon C. will be presented for two separate attentional mechanisms that can in Baylis)~Recent views of visual attention suggest that visual scenes fluence response time in analyzing hierarchical patterns (Navon, 1977). are first parsed into candidate objects to which attention is directed. One mechanism allocates attentional resources categorically to global The attentional deficit of a patient with RH damage supports this. and loca1levels, producing costJbenefit tradeoffs as target level probabil The patient neglected the left half of objects and was consequently ity changes over a block of trials. The other mechanism influences the unable to judge synunetry about the vertical. Nevertheless, symmetry diameter of an attentional window on a trial-by-trial basis, producing of shapes had normal effects on his figure-ground segregation. Both benefits for all targets within it. The relevance of these fmdings in inter sides of shapes were available for preattentive parsing, prior to the oper preting parieta1lobe damage and attentional dysfunction will be discussed. ation of neglect. (496) (502) Disengaging from the Disengage Mechanism: A Reinterpretation Superimposed Information Displays: Attentional Deficits and of Parietal Deficits. JONATHAN D. COHEN &MARTHAJ. FARAH, Potential Solutions. DAVID C. FOYLE, NASA-Ames Research Center, Carnegie-Mellon University (sponsored by James L. McClelland)-1n ROBERT S. McCANN, Sterling Federal Systems, & BEVERLY D. a spatially cued simple RT task, parietal-damaged patients respond ab SANFORD, San Jose State University (sponsored by Stephen R. Ellis) normally to targets presented in the affected field when preceded by Head-up displays (HUDs) superimpose display information on the out a cue in the intact field. The inability of such targets to disengage atten side world such that both information sources are in the pilot's field tion has been interpreted as evidence for a distinct "disengage" mecha of view. However, studies have shown that people may not process simul nism. We demonstrate, through the use of computer simulation models, taneously both the superimposed and out-the-window information. We that identical deficits can arise by lesioning a system characterized by report results suggesting that cognitive and perceptual cues may cause interactivity and competition, but which does not contain any "disen the two sources of information to segregate, yielding a divided atten gage" mechanism. tion deficit. Possible solutions involving the integration of display and (497) world information will be discussed. The Inftuence of Pbonotactic Constraints on Phonetic Coding. (503) CAROL A. WANNEMACHER & JAMES R. SAWUSCH, SUNY at Individual Differences in Speeded Classification: An Analysis of Buffalo (presented by James R. Sawusch)-Previous studies demonstrate Cognitive Style. LAURA FORD & BARBARA BURNS, University that the perception of ambiguous phonemes is altered by the surround ofLouisville (presented by Barbara Burns)-The goal of the present study ing phonetic context to produce percepts consistent with the phonotactic was to further characterize the individual differences in information constraints of the language. The influence of phonotactics was examined processing by highly analytic and holistic adults. Filtering. condensa using speeded classification of synthetic series in which the constrain tion, and grouping tasks were used with highly separable (circle size ing phonemes occurred either before or after the target. Results will vs. angle), separable (color vs. form), and integral (height vs. width be presented on the time course of phonotactic effects, the role of phoneme of rectangle) dimensional combinations. Cognitive style, as measured intelligibility, and their implications for models of speech perception. by Kagan' s MFFT, was found to interact with both stimulus structure (498) and task demands. Results are discussed in the context of current models Pbonological Priming Effects: A Function of Lexical Stress? of integral-to-separable processing. BEATRICE DE GELDER & JEAN VROOMEN, TIlburg University (504) (sponsored by Paul Bertelson)-Effects of phonological priming on an Interfenonce and Two Measures ofImplidt Memory. LAWRENCE M. auditory lexical decision task were examined in Dutch. Recognition of SCHOEN & BEN T. BROWN, Lake Forest College (sponsored by word targets is inhibited by word primes. The size of the effect is a Richard Jackson Harris)-Subjects were given a traditional interference function of the stress pattern of the prime. First-syllable-stressed primes task, learning three 20-item word lists. After each list they solved word lead to the strongest inhibition provided that the initial phoneme group puzzles (anagrams or fragments). Priming of puzzles occurred in either shared by prime and target is a syllable (as opposed to consisting of the first or third word list, with the primed puzzles appearing in either the same initial phonemes but not making a syllable). the first or third puzzle list. Interference from the explicit recall task (499) influenced solution rates of the implicit task, both for type of measure Combining IliOlable I'by!iI:al and Semantic Codes. PETER GROSSEN llIld among the three presentation combinations of prime/test. BACHER, PAUL COMPTON, MICHAEL I. POSNER, & DON (505) TUCKER, University of Oregon (presented by Michael I. Posner) Limits of Structural Complexity in Implicit Sequence Learning. How are different codes of a visual stimulus bound together into a com PEDER J. JOHNSON, JONATHAN REED, & PHILIP KRAGNES, mon representation? We have approached this problem for the visual University ofNew Mexic~ohen, Ivry, and Keele (1990) reported that and semantic codes of words with both cognitive and electrophysiolog when subjects were engaged in a secondary tone-counting task, implicit ical methods. We have found that the processing of semantic and visual learning of sequences was limited to first-order conditionals (e.g., each codes overlap heavily in time, while they maintain their separate anat response is determined by the previous response). Using this dual-task omy. We consider the role of attention in amplifying, speeding, and procedure. we found that subjects were able to learn both first- and conjoining these distinct but interacting computations. second-order conditionals (i.e., response determined by previous two (500) responses). Results are discussed in terms of those aspects of sequence All-or-None Versus a Graded Process Conception of Attention. complexity that are related to difficulty of implicit learning. LISA R. FOURNIER & CHARLES W. ERIKSEN, University ofRlinois (506) at Champaign (presented by Charles W. Eriksen)-To determine whether The Components of RecoUective Experience: Remembering and attention is a gradual or all-or-none process, attention was directed to Knowing. SUPARNA RAIARAM, Temple University School ofMedi- 518 Posters 507 - 518 Saturday Evening cine (sponsored by Michael J. Watkins)-The nature of recollective ex 88 college students. Subjects listed and prioritized their goals for the perience was examined in several experiments where subjects made following day and 2 days later, reported their accomplishments. Sub "remember" judgments (to items recollected vividly) or "lmow" judg jects also defined • 'planning" in a short essay. Self-reported good and ments (to items recognized on some other basis) in a recognition mem poor planners did not differ in the proportion of goals completed or in ory task. Results show that "remembering" is sensitive to conceptual sensitivity to goal priority. Planners' definitions of planning did differ. manipulations and "knowing" to perceptual manipulations. These results The importance of ecologically valid measurement in planning research indicate that dissociations similar to those obtained between perceptual is discussed. implicit memory and conceptual explicit memory tasks are observed be (513) tween remember and Imow judgments. Transfer from Computer-a-d Labonltory Environments to RaJ (507) World Tasks. COLLEEN M. ZEITZ, University of Pittsburgh (spon Prior Knowledge Affects Reports of the Dates of Events. VINCENT sored by Robert Glaser)-An important issue in the development of learn PROHASKA, Lehman College (sponsored by Karyl Swartz)-People's ing skills is the degree of transfer to novel contexts. Undergraduates prior Imowledge about the usual temporal intervals between events af who explored one, two, or three computer environments were compared fects their reports, from memory, of particular dates. Subjects read a to assess whether they developed skills that facilitated exploration in number of two-event vignettes whose dates created intervals that were a novel noncomputerized domain. Subjects who explored three environ less, greater, or equal to people's prior Imowledge of the usual inter ments discovered more laws, asserted fewer erroneous conclusions, and vals (established independently). After reading the vignettes, subjects collected more exhaustive data in a real-world laboratory. Protocol anal recalled the presented dates. Results indicated that the dates reported yses revealed that they abstracted general methods for scientific inquiry. were influenced by both the dates presented and prior Imowledge. (514) (508) Scientific Induction: Effects or Test Strategy and Multiple Wby We Recall More about a Tourist than a Thief. CATHERINE Hypotbe8es. ERIC G. FREEDMAN, Michigan Technological University HANSON, Temple University (sponsored by Nora Newcombe)-Changes (sponsored by Ryan Tweney)-Scientific induction was investigated using in schema activation are shown to lead to changes in the way prose Wason' s 2-4-6 task . Subjects tested either a single hypothesis or a pair material is organized into events and subsequently remembered. Sub of hypotheses. Subjects were also instructed to use either a positive jects who believed they were reading a story about a tourist not only or negative-test strategy. Induction improved when subjects employed parsed the story into more events than subjects who believed they were a combination of either a positive-test strategy and single hypothesis reading about a thief, but also recalled significantly more of the story. or a negative-test strategy and multiple hypotheses. A combination of Consistent with earlier work (Hanson & Hirst, 1989), recognition per multiple hypotheses and a negative-test strategy may have facilitated formance was not similarly affected by orientation. induction because subjects could elitninate incorrect hypotheses. (509) (515) Effects of Experiential Deprivation on Recognition Failure of Hypothesis Generation and Changes in Representation In Medi Recallable Famous Names. CHANDRA BHAL DWIVEDI & SUDHA cal DIagnosis. JOSE F. AROCHA & VlMLA L. PATEL, McGill SRIVAST A V A, lJanaras Hindu University (sponsored by R. Parasura University (presented by Vimla L. Patel)-This paper addresses two man)-Forty-<:ight experientially deprived subjects studied boldface Hindi related issues: changes in mental representation of clinical cases and name and non-name targets in the context of ordinary face cues. Fol the strategies that accompany those changes. Medical students were asked lowing the recognition failure (RF) paradigm, recognitions and cued to solve three clinical cases describing patients whose initial presenta recalls were obtained based on retrieval queries directed at episodic or tion indicated a single common disorder. Subsequent information, how semantic memory systems. The results (a) indicate significant main ef ever, indicated more strongly a different disorder. The results suggest fects of deprivation and famousness, (b) demonstrate the robustness of that as the subjects' experience with the clinical domain increased, their the RF phenomenon unaffected by degree of deprivation, and (c) fail initial disease space becomes fuzzier while their strategies become more to support the episodic-semantic distinction. focused. (510) (516) Perspective in Autobiographical Recall. JOHN A. ROBINSON & Strategies ror Hypothesis Induction. KEVIN DUNBAR, McGill KAREN SWANSON, University ofLouisville-Nigro and Neisser (1983) University, ANNE L. FAY, & DAVID KLAHR, Carnegie-Mellon contrasted field (F) and observer (0) perspectives in autobiographical University (sponsored by David KIahr)-The strategies used to induce recall and offered two explanations for the phenomenon. We tested their hypotheses from data were investigated. Subjects had to discover how hypotheses in two studies. Both received some support. Recent F and a new function key worked. There were two groups: one was told that o memories differed in several qualities. Old F and 0 memories did the key was a repeat key and the other that it was a mystery key. Sub not. Changing from F to 0 altered memory qualities. The reverse change jects in the mystery group used a strategy of inducing hypotheses by had no effect. Prior awareness of perspective and metacognitive accuracy contrasting a series of experimental outcomes, whereas the repeat sub were also assessed. People were aware of perspective but insensitive jects used single-<:xperimental results to induce hypotheses. to its features. (517) (511) Structure Versus Variety in Concrete Representations of Number. Source and Content Memory for Conversatiom. PAUL J. JURICA SUK-FONG TANG & JAMES W. STIGLER, University of Chicago & ARTHUR P. SHIMAMURA, University of California, Berkeley (presented by James W. Stigler)-AJthough children are thought to benefit (presented by Arthur P. Shimamura)-Subjects were asked to generate by using concrete representations of number, little is known about which answers from questions ("When was the last time you voted?") or to concrete representations benefit them most. In this experiment, third read statements ("Breakfast is my favorite meal of the day") presented grade children practiced quantifying computer-generated arrays of blocks by one of three sources (faces on a computer). Memory for content was that were either random and varying from trial to trial, or canonically better for questions than for statements. However, source memory was organized and nonvarying. Pre- and posttests used numerals to present better for statements than for questions. These findings support a limited a series of arithmetic tasks. Results showed significantly greater gains resource view of memory for conversations and limits the scope of gener for the group exposed to canonical nonvarying block arrangements. ation effects. (518) (512) Postsecondary Mathematical Education Has Little Impact on Everyday Planning: An Analysis of Daily Time Management. Mathematical Problem Solving. JOHN B. COONEY, University of DANIEL J. SIMONS, Harvard University, & KATHLEEN M. Northern Colorado-Contrary to its intention, postsecondary education GALOTII, Carleton College (presented by Kathleen M. Galotti)-We in mathematics does not appear to enhance students' memorial repre assessed planning, prioritization, and completion of daily activities in sentation of relevant problem information, perception of problem struc- 519 Saturday Evening Posters 519-530 tore, translation of problems into equations, or problem-solving accuracy. objects that depend upon whether the orienting task emphasizes elabora Advanced mathematics coursework, however, was related to algebraic tive or structural encoding of the drawings (e.g., Schaeter, Cooper, & proficiency. Consistent with predictions from the fuzzy-trace theory of Delaney, 1990). Two experiments indicate further dissociations when reasoning (Brainerd & Reyna, 1990), students' memory for problem structural encoding involves generation (construction) of targets com information was found to be independent of problem-solving accuracy. pared with conditions involving elaborative processing of targets. Ad The nature of students' functional representation is currently under ditionally, differential transfer of processing is observed by varying the investigation. format of targets from study to test. (519) (525) The Effect of Intuitive Representations on Math Strategy Verifi Multiple Sources of Spatial Information for Aimed Limb Move cation. JAMES A. DIXON & COLLEEN F. MOORE, University of ments. RICHARD A. ABRAMS, VICKI A. STEMMONS, & LINDA Wisconsin-Madison (presented by Colleen F. Moore)-Intuitive under VAN DILLEN, Washington University-Experiments are reported in standing of an acid mixture task was manipulated by instructing sub which subjects produced rapid arm movements toward a visible target. jects about one of five principles of acid mixture. Subsequently, sub The perceived location of the target was distorted in one experiment jects evaluated the answers to pairs of problems. Both problems in each by an induced motion illusion, and in another experiment by the eye pair had been solved the same way by an unspecified mathematical movements required of the subject. The limb's initial submovement strategy. Pairs of problem answers violated different principles, but were toward the target was more affected by the distortion than was the overall objectively the same distance from the correct answers. Subjects evaluated movement, suggesting that component submovements are programmed problem pairs that violated principles they understood more quickly than on the basis of distinct sources of spatial information. other problem pairs. (526) (520) The Effect of Variations in Analog Representation on Transfer Context Effects, Field Dependence, and Gender Differences in between Analogous Stories. C. A. CLEMENT, Eastern Kentucky Proportional Reasoning. CAROL A. LAWTON, Indiana-Purdue University, R. W. MAWBY, Kentucky State University, & D. E. GILES, University at Fort Wayne (sponsored by Bruce B. Abbott)-Few studies Eastern Kentucky University (sponsored by Rachel Joffe Falmagne) have systematically examined the effects of context on proportional Analogical transfer was examined in three experiments with college stu reasoning. In this study, college students were presented with the fol dents. Three manipulations were each found to improve transfer be lowing problem: 4 is to 6 as 6 is to _ ?, in either a mixture-of-liquids tween analogous stories. Transfer was significantly better when (a) analo or Piagetian-type cylinders context. Better performance was found on gous structural relations were relatively less embedded in domain-specific the mixture problem, although even subtle structural changes in this detail and language, (b) abstract rather than domain-specific verbs were problem affected performance. Verbal analytic ability, field dependence! used to convey key analogous relations, and (c) the domain-specific in independence, and gender also contributed significantly to performance. stantiation of an analogous structure was accompanied by a general term (521) that lexicalized this structure. Learning from Mistakes: Source Solution Failures and Analogi (527) cal Problem SolvIng. MARY L. GICK & SUSAN J. McGARRY, Carle Stage-Based Versus Continuous Models of Spatial Learning. ton University (sponsored by R. S. Lockhart)-Several transfer experi JOSEPH P. MAGLIANO, ROBERT COHEN, Memphis State Univer ments tested the hypothesis that the solution to a target insight problem sity, GARY L. ALLEN, University of South Carolina, & JAMES R. that typically results in persistent solution failures (due to inappropri RODRIGUE, University ofFlorida (presented by Gary L. Allen)-Stage ate problem representations) is facilitated if a source problem tends to based models of spatial acquisition posit an invariant representational induce initial (before feedback) analogous solution failures. Results sup format, while the continuous-based models posit flexible, experientially ported the hypothesis, and further indicated that surface similarity of based, knowledge acquisition. Adults viewed slides of a walk through words and diagrams between source and target problems had no effect. a small town and were instructed to attend to either landmark informa The relationships among failed solutions, problem representation, and tion, route information, or configural information. A control group was surface similarity are discussed. given no specific spatial instructions. Tests of landmark, route, and con (522) figural knowledge supported the stage-based model. The musory Transparency of Utterances: A Problem of Perspec (528) tive. BOAZ KEYSAR, University of Chicago (sponsored by William Mirror Tracing Is Learned Via a Series of Direction-Specific Asso Prinzmetal)-Subjects read scenarios specifying the meanings of other ciatiom. DANIEL B. WILLINGHAM, DAVID E. HUBER, JOHN L. wise ambiguous utterances. For each utterance, half of them were in SPEAR, Williams College, & JOHN D. E. GABRIEL!, Northwestern formed that speakers were being sarcastic and the other half were told University (sponsored by Richard o. Rouse, Jr.)-Neurophysiological that speakers were truthful. This information was unknown to the ad studies indicate that simple visuomotor associations and tracking skills dressees. These overhearers thought that addressees perceived sarcasm both rely on posterior parietal and premotor cortices. We sought evi only when they knew it was intended. They could not disregard their dence that tracking tasks are in fact learned via a series of direction privileged knowledge even though it was irrelevant for addressees, be specific visuomotor associations. College students traced a straight line cause it made the utterance appear "transparent." in a single orientation via a mirror. Some subjects mirror traced more (523) lines in other orientations, but they did not benefit from the extra train Parts and Wholes in Face Recognition. JAMES W. TANAKA, ing, indicating that the skill is direction specific. Oberlin College, & MARTHA J. FARAH, Carnegie-Mellon Univer (529) sity (sponsored by Martha J. Farah)-Are faces recognized using more The Representation of Color and Form in Memory. AURA holistic representations than other types of stimuli? We define a "holistic" HANNA & ROGER W. REMINGTON, NASA-Ames Research Center representation as one without internal part structure, in contrast to ob (sponsored by Geoffrey Loftus)-We investigated the effect of color jects with explicit part representations. Explicitly represented parts should on recognition memory for geometric shapes. Subjects studied sequences be more easily identified than nonrepresented parts. Consistent with the of stimuli composed of colored or monochrome geometric shapes and holistic hypothesis, features in normal faces were disproportionately more were tested for their recognition of shape only. Performance was best difficult to recognize relative to features in scrambled faces, inverted when the same colors were present at study and test regardless of whether faces, and nonface objects when recognition of wholes was equated. they were assigned to the same geometric shapes. These results are con (524) sistent with the hypothesis that color and shape are represented separately Measures of Memory for Possible and Impossible Figures: Study in memory. to-Test Transfer. BRIAN J. LYMAN, Trinity University (sponsored (530) by Paula T. Hertel)-Recent research has found implicit/explicit mem McCoIlougb FJrects in a TranswitdJing Paradigm. R. M. YAREMKO, ory dissociations for line drawings of novel possible and impossible 3-D San Diego State University-McCollough color aftereffects (MEs) were 520 Posters 531-542 Saturday Evening induced in a transwitching (similar to occasion setting) classical condi then different processes or coping strategies may come into play. Under tioning procedure to test the hypothesis that MEs result from Pavlovian these circumstances, subjects must resort to using internally generated processes or from specific feature detectors. Subjects were given red representations of response strings, and performance suffers. vertical/green-horizontal bar training with one background context, al (537) ternating with green-vertical/red-horizontal patterns on a distinctive The TIme Course of Perceptual and Conceptual Processes in Prim background in four 4O-min sessions. Posttests indicated MEs were de ing. MARY SUSAN WELDON, University of California, Santa Cruz pendent on presence of a context but not the context associated with (sponsored by Henry L. Roediger 01)-Two experiments provided evi induction. dence that perceptual information becomes available before conceptual (531) information during priming on an implicit data-driven test. Primes were Circadian Rhythm of Semantic Memory Activation in Dream presented visually (words), auditorily, or pictorially. Later, priming on Production. CORRAOO CAVALLERO, Universita'di Bologna (spon the word fragment completion test was measured. In Experiment I, word sored by David Foulkes)-Retrieval efficiency from semantic memory fragments were exposed for 12 sec, 5 sec, I sec, or 500 msec. All primes appears to vary according to a circadian cycle characterized by a trough produced priming at the longest exposure time, but only visual words at about 4 a.m. and a peak late in the afternoon. The present experi produced priming at the shortest time. Further evidence that perceptual ment investigates whether semantic memory involvement in dream information becomes available fastest was obtained in Experiment 2, production follows this circadian cycle. Results indicate that semantic in which cumulative response functions were examined. memory is maxirnally involved in the production of dreams in second (538) REM (i.e., approximately when retrieval efficiency from semantic mem Stem Completion Versus Cued Recall: The Role or Response Bills. ory of sleep-deprived subjects is at its worst). EYAL M. REINGOLD, University of Toronto. Erindale Campus, & (532) PHILIP M. MERlKLE, University of Waterloo (sponsored by Philip M. Hiding Valuables: Age Differences in a MnemonicaUy Risky Be Merikle)-Graf and Mandler (1984) found that under shallow encod havior. ALAN S. BROWN & TAMARA A. RAHHAL, Southern ing conditions. stem completion is a more sensitive indicator of mem Methodist University-A questionnaire assessed subjects' experiences ory relative to cued recall, whereas under deep encoding the reverse with hiding valuables and having difficulty locating them later. Age dif pattern was observed. In the present experiments. we demonstrate that ferences were found between younger and older adults for the type of differences in response criteria across tasks may account for these results: item hidden, reason for hiding the item, how long ago the item was hid the original pattern was obtained when subjects were discouraged from den, number of searches for the item, who eventually found the item, guessing during cued recall (conservative criteria) but not with forced and time taken to fmd the item. There were no age differences in per cued recall instructions (lenient criteria). cent of items recovered and where an item was found. (539) (533) Cue-Compounding Versus Cue-Decompounding of Complex Cross Talk and Computatiooal Momentum in Cognitive Processes. Causes. MICHAEL R. WALDMANN, Univusity of Frankfurt (spon CAROL A. VAREY, University of Waterloo (sponsored by Daniel sored by Norman H. Anderson)-Current theories of causal induction Kahneman)-The Stroop effect is a paradigmatic example of subjects see learning as a bottom-up process that is biased toward additive in performing more cognitive operations than required. A series of ex tegration of independent causes. Experiments will be presented which periments in which tasks were varied within a block of trials demon show that identical causes are integrated differently depending on the strated a new source of unnecessary processing. Cross-talk effects in effect of the causes. Furthermore, the results demonstrate that subjects reaction times suggested that subjects were performing the previous task have a strong preference for using additive integration rules when decom on the new stimulus in addition to the new task. This "computational posing compounds of multiple causes even when they would use an aver momentum" effect was observed in feature-verification tasks and aging rule when compounding them. same-different judgments. (540) (534) Speciticity of Priming in Nonverbal Tests. KA VITHA SRINIV AS, Question Asking in Tutoring Sessions. ARTHUR C. GRAESSER, Boston College (sponsored by James R. Pomerantz)-Two experiments JOHN D. HUBER, & NATALIE K. PERSON, Memphis State were conducted to investigate whether priming is specific to the view University-We analyze the questions that students and tutors ask dur ing angle of an object. Subjects studied objects photographed from either ing tutoring sessions on topics in research methods (e.g., statistics, oper usual or unusual viewing angles. At test, subjects were required to name ational defmitions, and interactions). A very small percentage of the brief presentations ofthe same or different viewing angles. In general, questions were well specified. The types of questions and question mech priming was greatest when objects were presented in the same viewing anisms were quite different for students and tutors. We will discuss these angle at study and test, supporting the transfer-appropriate processing results in the context of natural conversation, instructional pedagogy, account of priming. and intelligent tutoring systems. (541) (535) Strategic and Automatic Mechanisms of Syntactic Priming in Effects of Circadian Rhythm Optimality on Reactions to Adver Speech Perception. AVITAL DEUTCH & SHLOMO BENTIN, tising. RICHARD JACKSON HARRIS & JULIA CANDACE Hebrew University of Jerusalem (sponsored by Shlomo Bentin)-The POUNDS, Kansas State University-Subjects read a series of 12 brief interaction between facilitation and inhibition components of syntactic advertisements and afterward rated (I) the truth of a critical claim which priming and attention was examined in speech identification. Relative had been either stated directly or pragmatically implied in the ad, and to a neutral condition, the identification of masked targets was equally (2) their likelihood of purchasing the product. Results showed that both inhibited and facilitated by syntactic context. Presenting the congruency the truth of the claims and the likelihood of purchase were rated higher conditions in separate blocks reduced the inhibition by 20%. increas by subjects in their circadian rhythm optimal time of day (e.g., morn ing the lSI between the context and the target to 350 msec reduced in ing people tested in the morning). hibition by 9%. Facilitation was unaffected by either manipulation. The (536) results suggest the involvement of two cognitive mechanisms in syn Human Behavioral Variability and Symbolic Support. LOUIS G. tactic priming. LIPPMAN, Western Washington University, & LINDA R. TENNISON, (542) Washington State University-Humans were reinforced for producing List Subset Effects and the Tulving-WioJeman Function. CARL A. binary response patterns that differed from a set number (lag) of prior BARTLING, McNeese State Univ~rsity-Three experiments provided strings. Visual displays conveyed different amounts of information about results inconsistent with the Tulving-Wiseman function. With type of the current strings. Results suggest that when memory limits are word pair a within-subject variable, obtained probabilities of recogni challenged (high lag value) or if external symbolic support is limited, tion failure of recallable words exceeded predicted probabilities from 521 Saturday Evening Posters 543-554 the Tulving-Wiseman function for noun-adjective, cue-target pairs. The with free-recall and cued-recall tests, and the magnitude of the effect reverse was true for adjective-noun pairs. Additional research on the was similar across test types. In some cases, the rates of item recovery relation between recognition and recall is suggested before supporters and intertest forgetting were differentially affected by the type of test. elevate the Tulving-Wiseman function to the status of a law. The results have implications for retrieval-based accounts of hypermnesia. (543) (549) Time Course of Retrieval: Implications for Models of Memory. Separating Strategic and Data-Driven Components of SkIlled Per SCOTT D. GRONLUND, MARK B. EDWARDS, & EUGENIE L. formance. ARTHUR F. KRAMER, University ofRlinois, & DAVID L. BERTUS, University of Ok/oJroma (sponsored by Francis T. Durso)-To STRAYER, University of Utah-The contributions of strategic and data link memory models (predict asymptotic accuracy, but not how that result driven factors to skilled performance were separated by manipulating is achieved) and retrieval models (predict the latency of processing) re the predictability of two classes of stimuli in a series of memory search quires specifying how information accumulates over time using the tasks. When consistently mapped (CM) and variably mapped (VM) response signal paradigm. A series of experiments establish an equiva stimuli were blocked, large performance differences were observed. lence between the retrieval of item and position information and be However, when CM and VM trials were mixed, performance differ tween associative and order information. Also, the retrieval of item in ences were substantially reduced. The dynamics of strategic shifts be formation is unaffected by dividing attention while the retrieval of tween automatic and nonautomatic processes will be discussed in the associative information is delayed. presentation. (544) (550) Retrieving Order and Item Information from STM: Serial and Autobiograpbical Memories for Emotional Events and Pbysical Parallel Mechanisms. BRIAN McELREE, University of California, Feelings. PETER E. MORRIS, Lancaster University (sponsored by Irvine, & BARBARA ANNE DOSHER, Columbia University (spon Eugene Winograd)-Autobiographical memories of emotional events sored by Barbara Anne Dosher)-Serial and parallel models for retrieving (e.g., being angry or thrilled) are used in a series of studies to evaluate order information from short-term memory were tested with reaction theories of the nature of emotions and to elucidate the properties of the time (RT) and speed-accuracy tradeoff (SAT) variants of a relative judg emotional memories themselves. Scherer's theory of stimulus evalua ment of recency task. RT means and distributions suggest that order tion checks predicts well the properties of the memories. Memories of information was retrieved by a self-terminating mechanism. Large dy emotional events are also compared with those for physical feelings (e.g., namic differences observed in SAT retrieval functions indicate a serial being cold or wet). Differences reflect the personal implications of emo self-terminating mechanism. In contrast, an SAT forced-choice item tional events. recognition study demonstrated that item information was retrieved by (551) a parallel mechanism. Speed of Mental Processing in IllV-Infected Intravenous Drug (545) AblL'lers. DAVID DORFMAN, Mount Sinai School ofMedicine, PAUL Associative Recognition and Recall in Mixed and Pure Frequency RINALDI, VAMC, Bronx, LEONARD HANDELSMAN, & Lists. STEVEN E. CLARK & RICHARD E. BURCHETT, Univer CHARLENE BANG, Mount Sinai School of Medicine (sponsored by sity of California, Riverside (sponsored by George J. Andersen) Thomas Toppino)-Like other subcortical-type dementias, AIDS De Previous research has shown that the word-frequency effect reverses mentia Complex (ADC) is characterized by slowed mental processes. for associative recognition: Common, high-frequency (HF) words are This slowing is difficult to discriminate from the motor slowing fre recognized better than rare, low-frequency (LF) words. By comparing quently accompanying HIV infection. HIV-infected intravenous drug performance in mixed and pure frequency lists, the current experiments abusers performed the Sternberg reaction time task. Using an additive test the hypothesis that the HF advantage is due to recall-like retrieval factors analysis, we were able to discriminate motor from mental slow processes. Results showed that frequency mixing eliminates the HF ad ing. The results suggest that, as in other risk groups, ADC is unlikely vantage in cued recall, but not for associative recognition, contrary to to occur early in HIV infection. the recall hypothesis. (552) (546) Age Differences in Text Memory: Continuity or Discontinuity in Interference Context and Short-Term Recall of Sentences. Wll.. Failure? JOELLEN T. HARTLEY & KIKU ANNON, California State LIAM S. MAKI & DAVID GEORGESON, North Dakota State University. Long Beach-If text presentation speed is titrated so that University-Potter and Lombardi (JML, 1990) suggested that immedi older and younger adults recall a single sentence equivalently, can equiva ate recall results from conceptually based reconstruction using recently lence be maintained as number of sentences increases? We varied the activated words. Sentences were followed by unstructured word lists; number of sentences in a text and looked for a critical point where age intrusions occurred if a list contained a synonym (lure) of a word in differences emerged. There was only weak evidence that a catastrophic, the sentence. The same effect was found in the present study. More age-related failure occurs in the integration of increasingly larger amounts over, intrusions were frequent even when the interfering material was of information. Further analyses bearing on this conclusion will be changed from a list to a coherent sentence in which the meaning of the discussed. lure was preserved. (553) (547) Memory Constraints in Early LeJdcal Acquisition. MIRELLA Noninformative Peripberal Cues and Motor Facilitation. DAPRETTO, ELIZABETH UGON BJORK, & ROCHEL GELMAN, CAMILLE-AIME POSSAMAI, CNRS-LNF, Marseille (sponsored by UCLA (presented by Elizabeth Ligon Bjork)-The relationship between Jean Requin)-Many studies have shown that a noninformative leading increasing retrieval abilities and linguistic advances in the second year cue decreases response time to a subsequent target presented in the same of life was investigated by comparing the performance of 18-26-month region as the cue. Some authors have assumed a perceptual origin of olds in the receptive and productive stages of lexical development on this facilitation, which lasts about 150 msec. During the last few years, a memory task involving recall of both verbal and nonverbal informa I have collected a series of data which suggest that the origin of this tion about pairs of related and unrelated objects. The results support facilitation is, at least in part, in the postperceptual, decisional, and/or the hypothesis that enhanced retrieval abilities underlie the advances in motor processes. word production witnessed in children toward the end of the second year. (548) (554) Hypermnesia Occurs in Both Free Recall and Cued Recall Tests. A Practical Application of "Old" Psychophysics in Dermal Warmtb DAVID G. PAYNE & JEFFREY S. ANASTASI, SUNY at Bingham Sensation. THOMAS HECKMANN & DAVID M. ROESSER, General ton-Payne and Roediger (1987) suggested that hypermnesia would typi Motors Research Laboratories-Magnitude estimation is preferred for cally not be obtained with cued-recall tests because the tests provide evaluating dermal warmth sensation but is inappropriate for rapidly de strong retrieval cues. In several experiments, we obtained hypermnesia termining how well different materials protect people from solar ex- 522 Posters 555-565 Saturday Evening posure. We used simultaneous foreann exposure and "wanner than" LIE B. COHEN, University of TeJUJS at Austin (presented by Leslie B. responses to evaluate wannth sensations under automotive window Cohen)-An object-exarnining task was used to investigate infant at glasses. Estimates of Weber's law for dermal wannth corresponded to tention to form/function correlations. Following familiarization with two the most recent published account (50 years ago), validating our proce toys, 10-, 14-, and 18-month-old infants received two trials, one with dure for purposes of product engineering, financial decision making, a familiar toy, in which the form/function correlation presented during and evaluation of solar simulators. familiarization was maintained, and one with a new toy, in which the (55S) correlation was violated. The 18-month-olds, but not the younger in An Analysis of Ability/Performance Relationships as a FlIIlCtion fants, increased responding to the latter toy , indicating recognition of of Practice and Age. WENDY A. ROGERS, Memphis State Univer the correlation. sity, ARTHUR D. FISK, & CHRISTOPHER HERTZOG, Georgia In (561) stitute of Technology (presented by Arthur D. Fisk)-The relationships Sequential Effects in Bimanual and Unimanual Serial Two-Choice between cognitive/speed abilities and performance on consistently Reaction Time. ERIC SOETENS & JOHAN HUETING, University mapped (CM) and variably mapped (VM) visual search tasks were as of Brussels (sponsored by Johan Hueting)-Previous research has shown sessed for 70 young (age 20) and 70 old (age 70) adults. LISREL tech that sequential effects in two-choice reaction time are caused by two niques were used to assess the influence of general, fluid, and crystal independent mechanisms, automatic facilitation and subjective expec lized intelligence, working memory, perceptual speed, semantic memory tancy. In a follow-up experiment, using unimanual responses, we found access, and psychomotor speed abilities on search performance. Age the same pattems of higher order effects as for bimanual responses. How differences and CMIVM differences are discussed in terms of abil ever, the first-order alternation effect, which is typical for subjective ity/performance relationships. expectancy, did not appear in the unimanual condition. These data im (5S6) ply that subjects' performance for alternations over repetitions is purely Infant Recognition as a Function of HabituationlTest Format. a response property. BEVERLY J. RODER, Fitchburg State College, THOMAS H. SCHUL (562) lNG, Harvard University, & EMILY W. BUSHNELL, Tufts Univer Incentive Processes Determine Instrumental Performance after a sity (sponsored by Emily W. Bushnell)-Four-month-olds saw repeated Shift in Primary Motivation. BERNARD BALLEINE, University of presentations of a visual target and a comparison stimulus until habitu Cambridge (sponsored by Mark E. Bouton}-Several experimenls examined ated. Then, in a recognition test, they saw the target with a novel stimulus. the role of incentive learning in instrumental performance following a Stimulus presentations were either sequential or simultaneous through shift in primary motivation. Rats trained to perform an instrumental action out each habituation/test sequence. Results indicated that the opportu when nondeprived increased their performance when subsequently food nity to compare stimuli facilitated recognition. Infants recognized difficult deprived but only if they had previously consumed the reinforcer when (dull, complex) and easy (bright, simple) targets when stimuli were pre deprived. Similarly, when trained deprived and tested nondeprived, per sented simultaneously, but only easy targets wben stimuli were presented formance was only reduced if animals were preexposed to the reinforcer sequentially. when nondeprived. These experiments indicate that instrumental perfor (557) mance following reinforcer-revaluation depends upon incentive learning. Children's Metalinguistic Representation of Linguistic Units_ (563) RICHARD P. MEIER & ANN REPP, University of TeJCQS at Austin AnalogIcal Codes, Common Codes, and Temporal Samples in (sponsored by Lawrence M. Parsons)-We examined children's meta Many-to-One-Matching. ANGELO SANTI, STEVE BRlDSON, & linguistic abilities by asking them to isolate syntactic constituents. Chil MARY JO DUCHARME, Wilfrid Laurier University-Pigeons were dren, aged 3'h or 5, were trained to echo subject NPs out of modeled trained to match temporal and color samples to line comparisons in many sentences. Children were assigned to one of two training conditions: to-one matching. There was no choose-short effect for temporal sam they were trained either on one-word or multiword subject NPs. Sub ples during delay testing. Following the acquisition of new comparison sequently, all children were tested on stimuli of both types. Byexamin associations to one sample dimension, asymmetrical transfer effects oc ing children's responses on test stimuli, we gain insight into how they curred. Interference tests produced evidence of temporal summation when represented the constituents modeled during training. either temporal or color presamples preceded temporal target samples. (558) These data are discussed with reference to both analogical coding and Event Representation and Preschool Children's Conversational common coding hypotheses. Skills_ M. GRAZIA CARELLI, BEATRICE BENELLI, University of (564) Padua, & TIMO MANTYLA, University of Stockholm (sponsored by Self-Control and Postponed Reinforcer Consumption. DA VlD A. Timo Miintylii)-The effects of scripted knowledge and contextual sup CASE & EDMUND FANTINO, University of Califomia. San Diego port on children's conversational skills were examined in two experi (sponsored by Ben Williarns)-Studies have reported lower self-control ments. Pairs of preschool children interacted in realistic settings that in pigeons and young children with immediate access to reinforcers than varied in terms of event familiarity and contextual richness. The results in college students who accumulated reinforcers until session end. The showed that although both event farniliarity and contextual richness were effect on self-control choice of postponing utilization of reinforcers was important for children's conversation, high levels of event familiarity examined in pigeons. Total choice response rate was reliably higher with did not compensate for the absence of contextual support. conventional within-session consumption of food. However, self-control (559) preference was not significantly affected, tending to decline with post Age Changes in the Processing of Implicitly Activated Informa poned consumption counter to the prediction based on subject -correlated tion. CATHY L. McEVOY, University of South Florida-Younger and differences. older subjects were compared on cued recall of words varying in num (56S) ber of associates theoretically activated during study. Type of study con Changeover Contingencies and "Sensitivity" to Reinforcement text cue (meaning-related, unrelated, or no context cue) determined on Multiple Concurrent Schedules_ LEON R. DREYFUS, DEE whether older subjects showed either diminished or comparable effects DEPORTO-CALLAN, & SARAH ANNA PESlLLO, Vassar College of number of associates, relative to younger subjects. The results sug (sponsored by J. Gregor Fetterman)-On a multiple concurrent sched gest age changes in the processing of implicit representations, but also ule with different changeover contingencies in the components, response suggest one condition in which comparable processing may occur. ratios underrnatched reinforcer ratios under a changeover delay and over (560) matched under a fixed-ratio requirement. An interpretation of these dif Infant Attention to an Object's Correlated Form and FlIIlCtion. ferences in overall response allocation, in terms of local contingencies LISA M. OAKES, University ofIowa, KELLY L. MADOLE, & LES- arranged by changeover requirements and a failure to fmd differences 523 Saturday Evening Posters 566-575 in slopes of the generalized matching equation computed early or late jects from another category. The experimenter asked the children to select in sessions, raises questions about the status of the "sensitivity" pa all four objects under one category. The majority of l-year-olds showed rameter of the generalized matching equation. categorical behavior despite poor comprehension. Three-year-olds showed (566) near perfect comprehension on all categories. The nature of initial Orthogonal Cues and Dimensional Contrast. JOHN M. HINSON categories will be discussed. & LINDA R. TENNISON, Washington State University-Two experi (571) ments examined the effects of orthogonal cues on dimensional contrast Unconscious Recognition in Prosopagnosia: An Alternative Ex obtained during visual wavelength discrimination by pigeons. Signal planation. MARTHA J. FARAH, RANDALL C. O'REILLY, & ing characteristics of visual stimuli were manipulated by adding an au SHAUN P. VECERA, Carnegie-Mellon University-Prosopagnosics, ditory stimulus as either a redundant or conditional cue. Range of who appear unable to recognize faces, may nevertheless respond ap wavelength stimuli was also varied. Redundant cues enhanced visual propriately to faces when tested using certain indirect methods (e.g., discrimination, while conditional cues reduced discrimination. These priming). This has been interpreted as evidence that brain damage did results may be due to changes in attention near the dimensional border not affect face recognition but rather conscious access to face recogni between positive and negative stimuli. tion' implying separate neural substrates for consciousness. We present (567) a computational model of face recognition which, when face recogni Enbancement of Conditioned Reinforcement during Repeated Ex tion units are lesioned to abolish overt recognition, shows preserved posure to Cocaine. MARC N. BRANCH, University of Florida performance on the indirect tasks. Pigeons keypecked in two-part sessions. In the first, a second-order (572) scbedule operated; in the second, extinction was in effect. In some ses Rule-Like Learning Bebavior from an Exemplar-Based Network. sions, a brief stimulus associated with food could be obtained by key JOHN K. KRUSCHKE, Indiana University (sponsored by James T. pecking in the second part of the session. When an intermediate dosage Townsend)-Proponents of rule-based theories of category learning ad of cocaine was given before each session, responding during extinction duce two classic phenomena (among others): (1) High-frequency ex that produced brief stimuli gradually increased to levels as high as those ceptions to rule-describable categories can show three-stage (U-shaped) maintained by food reinforcement. learning (e.g., Rumelhart & McClelland, 1986). (2) Intradimensional (568) feedback reversals are easier to learn than extradimensional (nonrever Selective Flavor Neophobia in Mongolian Gerbils. RODERICK sal) shifts (Kendler & Kendler, 1962). In this poster, I present new hu WONG & CHRISTOPHER McBRIDE, University of British man data and show that ALCOVE (Kruschke, 1990a, 199Ob, in press), Columbia-In contrast to studies indicating absence of neophobia to fla an exemplar-based connectionist network, exhibits both effects. vored liquids in gerbils, we present data indicating the prevalence of (573) this phenomenon to specific flavored foods. Gerbils in the experimen Negative Occasion Setting in Flavor-Calorie Conditioning. RON tal group were tested with a nut flavored with either NaCl, sucrose, MEHIEL, CATHERINE McCARTHY, & DEBRA A. ZELLNER, citric acid, or quinine crystals and were compared to control animals Shippensburg University (sponsored by Debra A. ZeUner)-Rats were tested with an unflavored nut. The results indicated that sweet-, salty-, conditioned to prefer a flavor (S 1) paired with calories over a flavor and sour-tasting nuts elicited longer oral latencies than plain nuts. How (S2) paired with water. Then a novel flavor (S3) was paired with either ever, there was no evidence of neophobia to bitter-tasting nuts. SI or S2. When tested for preference between the water-paired and (569) calorie-paired flavors (SI-S2), the presence of S3 decreased the prefer Learning Modes, Feature Correlations, and Postencoding Anal ence for the calorie-paired flavor. Thus, S3 served as a negative occa yses. WILLIAM D. WA TTENMAKER, University ofPittsburgh (spon sion setter. sored by James F. Voss)-Wattenmaker (1991) found that incidental (574) category learning tasks produced greater sensitivity to correlated fea Producing Confusions in Source Monitoring: A Reality Monitor tures than intentional category learned tasks. To explain this result, it ing Approach. TRACEY L. KAHAN, Santa Clara University (spon was hypothesized that the incidental tasks preserved more information sored by Marcia K. JoImson)-Three experiments reliably produced con about the individual exemplars, and that the correlations were computed fusions between memories for imagined and perceived events. A from the stored exemplars. The present research provides direct sup computerized mental rotation task was used to present 12 seen and 12 port for this interpretation, and provides additional evidence that stor imagined letters and numbers. Each subject (N = 120) then received ing examples can be valuable for addressing unanticipated needs and one of two incidental memory tests. Old/new recognition and source goals. identification measures were thus obtained for characters rotated clock (570) wise (Experiment 1) and counterclockwise (Experiment 2) and for Comprehension of Superordinate Category Terms Among Young characters not rotated (Experiment 3). The major findings are discussed Japanese-Speaking Children. M. MICHAEL AKIYAMA, University in light of recent models of reality monitoring and source monitoring. of Michigan-Dearbom-Sixteen l-year-olds, 16 2-year-olds, and 16 3- (575) year-olds in Japan were tested on the comprehension of four superordinate Research Support from the National Science Foundation. JOSEPH categories (i.e., vehicle, clothing, food, and animal). They were shown YOUNG, NSF Program Director for Human Cognition and Perception, four basic-level objects from one superordinate category with four ob- & FRED STOLLNITZ, NSF Program Director for Animal Behavior. 524 Papers 576-586 Sunday Morning MOTOR CONTROL cates that the fine timing adjustments made to match rates between I Embarcadero AB, Sunday Morning, 8:00-10:15 and 5 Hz persist into the minute range. The biases attributable to this aftereffect are distinct from the biases that Collyer, Broadbent, and Chaired by Gordon M. Redding, Illinois State University Church have interpreted as evidence for discreteness in motor timing. 8:00-8:15 (576) Speed and Accuracy of Aimed Hand Movements toward Shifted LETfERlWORD PROCESSING IV Targets. HOWARD ZELAZNIK, Purdue University-The "smooth· Embarcadero CD, Sunday Morning, 8:00-10:50 ness" of corrective movements towards a target that can be shifted I cm toward or away from the subject's starting position was examined. Eight Chaired by Mark S. Seidenberg, University of Southern California subjects performed line drawing movements toward a .8-cm target 20 cm from a home position. Target shifts could be either asymmetrical or sym 8:00-8:20 (582) metrical. The shift occurred coincident with movement onset. Subjects IndmduaI DlffI!l'l!DlB in Lexical F1ueucy: JIIaer Is Betta-. DAVID B. were able to move towards the shifted target regardless of type of pos PISONI, MARY JO LEWELLEN, STEPHEN D. GOLDINGER, & sible target movement. Movement time effects were very small. BETH G. GREENE, Indiana University-Investigations of individual 8:20-8:35 (577) differences frequently show that "high-verbals" surpass "low-verbals" Separable Components of 3-D Tracking witb Misalignment be on various tasks. Proposed differences between groups typically focus tween Display and Control Axis. STEPHEN R. ELLIS, MITCHEL only on the automatized subprocesses of reading. Using word familiar TYLER, NASA-Ames Research Center and University of California, ity ratings, we identified two groups of subjects and compared them Berkeley, WON SOO KIM, & LAWRENCE STARK, JPL-Subjects on several tasks requiring extensive lexical processing, such as lexical performed 3-D tracking with misaligned display lcontrol frames of refer decision, naming, and recall. The groups did not differ consistently on ence on perspective displays. Their RMS tracking error in orthogonal every task. Results suggest that individual differences in lexical fluency components exhibits a visual component linearly separated from a visual extend beyond early subprocesses of reading. motor component. The visual component appears proportional to the 8:25-8:40 (583) average screen projection of the orthogonal components. The visual Determinants of Word Translation. A. M. B. DE GROOT, Univer motor component may not exhibit characteristics expected were it due sity of Amsterdam-This study addressed the question of which word to an added time lag caused by mental rotation. Practice allows sub characteristics determine performance in translating words from the na jects to adapt to multiple misalignment angles. tive to a second language. The following variables were included: fre 8:40-9:00 (578) quency of the stimulus word; frequency of the to-be-produced transla Symmetry Breaking Dynamics of Coordination. JOHN J. JEKA tion; familiarity, imageability, context availability, and ease~f- 525 Sunday Morning Papers 587-597 readers do not extract any semantic information from below the line 8:45-9:05 (592) of text they are reading; however. the visual information below the line Relative LIIIIIinaJM:e Is Not Derived from Absolute Lwnlnance. of text being read modulates the speed of reading. JAMES SCHUBERT & ALAN L. GILCHRIST. Rutgers University 9:45-10:00 (587) (read by Alan L. Gilchrist)-Thresholds were measured for detection Impact of Word Shape on Word Recognition. PHILIP A. ALLEN of the direction of change of overa1lluminance level in a homogeneous & BENJAMIN WALLACE. Cleveland Stale University-We conducted (Ganzfeld) and a patterned visual field. and the direction of change of two lexical decision experiments examining case alternation. Experi the edge ratio in a two-part visual field. The results indicate that detec ment 1 compared consistent. lowercase presentation to mixed-case tion of absolute luminance level (I) occurs. but grossly. in the absence presentation. Experiment 2 compared consistent. uppercase presenta of spatial and temporal gradients. (2) is not aided by the presence of tion to mixed-case presentation. For both experiments. letter shapes were spatial gradients. and (3) is not the basis for relative luminance detection. always familiar. but word shapes were unfamiliar for mixed-case presen 9:10-9:30 (593) tation and familiar for consistent-case presentation. There was a mixed On FiIUng in the BHod Spot and Other Medical Marvels. V. S. case disadvantage for both experiments. indicating that subjects used RAMACHANDRAN. University ofCalifornia, Son Diego-We designed word-shape information rather than letter-shape information to recog several novel stimuli to explore the "filling in" of the blind spot. nize words presented in consistent case. (I) Several small rings were viewed with one alone positioned on the 10:05-10:20 (588) blind spot. This ring looked like a homogeneous disc and popped out Parafoveal Versus Foveal Processing: Effects of Phonemic Prim preattentively. (2) A vertical illusory contour composed of horizontal ing and Motivation. EVA DREIKURS FERGUSON. Southern Rlinois lines was placed on the blind spot. The illusory line was completed rather University-Edwardsville-Parafoveal and foveal word processing are than the horizontal lines defining it. (3) A cross made of intersecting different. Previous studies found significant word characteristic effects black and white lines was presented. The lines competed for comple and facilitation from motivation (hunger vs. satiation) for foveal. not tion unless one was made longer. parafoveal. word recognition. Now. three parafoveal rhyming-list studies 9:35-9:50 (594) with cortical magnification. versus three foveal studies. showed: for first The Effect of Pitched Environment on Apparent Height: Task De time. word frequency. typically significant with foveal processing. sig pendence. ARNOLD E. STOPER & ABIGAIL B. BAUTISTA. Califor nificantly affected parafoveal processing also; although motivation facili nia State University, Hayward-Both apparent eye level and apparent tated foveal processing. unexpectedly phonemic priming occurred sig height of a standing target object have been previously shown to be in nificantly with parafoveal but not foveal processing. Implications are fluenced by the pitch of a structured visual environment. The effect of discussed. pitch on apparent height was strong when apparent height was indicated 10:25-10:45 (589) either by pointing to a position on a vertical standard or by verbal esti Orthographic Rime Frequency Effects Are Independent of Word mation. Now we report that this effect aIrnost disappeared when a match Frequency. JUDITH A. BOWEY & ANDREW J. SHATrE. University ing task was used to indicate apparent height. of Queensland (sponsored by Michael S. Humphreys)-Orthographic 9:55-10:15 (595) rime frequency was defined as the number of monosyllabic words within Visually Perceived Eye Level, Visually Perceived Vertical, and the the Ku~ra and Francis word count sharing the same orthographic rime. Great Circle Model. LEONARD MATIN & WENXUN LI. Colum excluding nonwords and proper noons. Experiment 1 investigated the poten bia University-Pitching a frontal plane containing two vertical lines tial interaction of orthographic rime frequency and word frequency within in darkness systematically influences the elevation visually perceived a word naming task. Rime frequency and word frequency were both as eye level (VPEL). but the orientation perceived as vertical (VPV) significant. but did not interact. In Experiment 2. nonwords with high is unaffected. However. when two lines are pitched by equal amounts frequency rimes were read faster than those with low-frequency rimes. in opposite directions, VPV changes linearly with pitch but VPEL is unaffected. Identical results with both manipulations were obtained with PERCEYI10N n equivalent stimuli in an erect plane. These results are predicted by the San Francisco A, Sunday Morning, 8:00-10:20 Great Circle Model. Chaired by Michael Kubovy, University of Virginia JUl)GMENTIDECISION MAKING n San Francisco B, Sunday Morning, 8:00-10:15 8:00-8:20 (590) Stimulus-Onset Asynchrony Is Not Necessary for Motion or Chaired by Valerie F. Reyna. University of Arizona Metacontrut. VINCENT DI LOLLO. University ofAlberra-Coherent directional motion is seen if a translated image is displayed in two se 8:00-8:10 (596) quential frames (Fl and F2). In a related paradigm-metacontrast Modeling Strategies in Dynamic DecIsIon MakIng. ALEXANDER J. masking-Fl contains the target and F2 the mask. Although temporal WEARING, CHRIS PIVEC, & MARY M. OMODEI, University of succession is considered essential in both paradigms. I obtain both co Melbourne-Determining the effectiveness of strategies in dynamic de herent motion and metacontrast masking with simultaneous onsets of cision making is constrained by the difficulty of identifying strategies F 1 and F2. provided that F2 outlasts F 1. Current theories can account subjects use. and ensuring consistent behavior. We have simulated a for the motion outcomes but not for the masking outcomes. firefighting task. Artificial subjects were developed that model various 8:25-8:40 (591) strategies. These were tested in a variety of task environments. We report Apparent Motion and Attention. ARlEN MACK. BENYU TANG. the effects of various task factors (properties of fires and appliances. New Schaal for Social Research. IRVIN ROCK. University of Califor information delay. complexity) on the performance of artificial subjects. nia, Berkeley. & WlLLANN STONE. New Schaal for Social Some comparisons with human subjects are presented. Research-Features of visual stimuli shown to be processed preatten 8:15-8:30 (597) tively are not perceived under conditions of inattention (Mack et aI .• Representativeness and the Disjunction Fallacy. MAYA BAR 1990; Rock et aI ., 1990). Using the technique described at these meet HILLEL & EFRAT NETER. Hebrew University ofJerusalem-Under ings last year. we examined whether this is the case for apparent mo conditions of certainty. people prefer to categorize items into basic-level tion. Results suggest that short-range apparent motion is more likely categories. We show that under uncertainty. the level at which one catego to be perceived without attention than long range. This is consistent with rizes items interacts with representativeness. The narrower of two the view that the latter entails more cognitive processing than the former. representative categories is preferred. whereas between unrepresenta- 526 Papers 598-608 Sunday Morning tive categories, the broader is preferred. When categories are ranked a peak-procedure experiment, two measures of performance change by probability, or by willingness to bet, this pattern results in a dis simultaneously, and their time courses show that two different kinds junction fallacy. The pattern is hardly higher, however, in tasks in which of associative learning are happening at the same time. A variety of evi it is not fallacious, as when ranking by suitability, or by tendency to dence suggests that the two changes correspond to the two kinds of learn predict. ing assumed by selective-attention theorists: learning what dimension 8:35-8:50 (598) to attend to, and learning what response to make. A Foolhib Consistency Is the Hobgoblin of LiUIe Minds. BARBARA 8:20-8:40 (604) MELLERS & LISA ORDONEZ, University of California, Berkeley Fixed-Duration Treatment Shocks, Feedback, CS Preexposure, and One of the most perplexing violations of expected utility theory is the Fear: Where's the Cognition? D. CHRIS ANDERSON, CHARLES R. preference reversal phenomenon: Subjects' preferences for risky op CROWELL, University ofNotre Dame, NEIL R. BOYD, Pennsylvania tions can change rank order depending on the response mode (e.g., choice State University, & JOSEPH TORREZ, University ofNotre Dame-In vs. selling price). Six new experiments investigate the robustness of the Experiment I (four groups), preexposure to a feedback stimulus offset phenomenon. Preference reversals continue to occur even when sub its shock-treatment-attenuating effects on posttreatment activity, skin jects are directly confronted with their own inconsistent behavior. resistance, and escape/avoidance performance. In Experiment 2 (II 8:55-9:10 (599) groups), preexposure to the feedback stimulus factorially was crossed Portable Measurement of Price and Income Effects with Audi with preexposure to the shock chamber. The findings of Experiment I tory Stimuli. JUSTIN ENGUSH & JAMES ALUSON, Indiana Univer were duplicated only when preexposure combined both the chamber cues sity (read by James Allison)-lndividual consumers interacting with a and the feedback stimulus. Preexposure to either without the other did portable computer allocated budgets to two alternative goods available not offset the impact of feedback. Furthermore, in the four instances at various behavioral prices. The goods were two different tones sounded where feedback effects were manifest, none entailed complete attenua when the consumer pressed a key a specified number of times. The tion of the symptoms of shock treatment. Three more studies are reported response to Slutsky-compensated budgets revealed the tones as substi showing that ST promotes fear "loading" to both the chamber cues and tutes, and revealed each tone as a normal good. The results validate to the shocks per se of the treatment regimen. this experimental method as a means of studying basic economic 8:45-9:00 (60S) processes in individual humans. Unusual, Strange, and Special Aspects of Local Contrast. JOHN C. 9:15-9:35 (600) MALONE & DEANNA L. TIMMERMANN, University of Tennessee Expert Versus Novice Decision Makers: It's Not How Much In at Knoxville-Strangely, some researchers have been unable to produce formation but What Kind of Information That Matters. JAMES local contrast. We sought to determine what conditions maximize and SHANTEAU, Kansas State University, & RUTH H. PHELPS, U.S. minimize the effects. Pigeon subjects showed few or no effects when Army Research Institute, Boise-Experts are often assumed to make use faced with alternating VIlEXT components signaled by key colors. Some of more information than novices-the "information-use hypothesis." effects appeared when the duration of S + was varied, but very strong However, recent studies report experts rely on the same or less infor and durable effects resulted when stimuli were difficult to discriminate, mation than novices (e.g., expert livestock judges have 5.3 significant the identity and duration of S - varied, and the presentation sequence cues vs. 5.2 for novices). But the information used by the experts was changed daily. more diagnostic. Thus, experts are better able than novices to discriminate 9:05-9:20 (606) between what is relevant in a given context and what is not. Contingent Adaptatioo to Color. SHEPARD SIEGEL, LORRAINE G. 9:40-9:55 (601) ALLAN, & LISA UNDERS, McMaster University-Evidence that con Acquired Modularity for Assigning Relations in Problems. ditioning contributes to adaptation to various stimuli (e. g., drug adminis STEPHEN K. REED & ROBERT ZELMER, San Diego State trations or exposures to a cold environment) is provided by demonstra University-We examine modularity for assigning relations in word tions of "contingent adaptation": the display of adaptation is especially problems by determining whether a decision regarding the arithmetic pronounced in the presence of cues previously paired with such stimuli. relation between two traveled distances is influenced by the two speeds We demonstrate contingent adaptation to color. We hypothesize that of travel. The decisions of psychology students were significantly biased the contingent color aftereffect (McCollough effect) is the CR that medi by whether the goal of the problem was to [md the slower or the faster ates contingent chromatic adaptation, as other compensatory CRs mediate of two speeds. The decisions of mathematics students were significantly contingent adaptation to other stimuli. influeoced only by the correct arithmetic relation between the two distances. 9:25-9:45 (607) 10:00-10:10 (602) Contingency Manipulations and the McCollough Errect. LOR Mental Arithmetic: Japan Versus United States. MARK H. ASH RAINE G. ALLAN, SHEPARD SIEGEL, & TOM EISSENBERG, CRAFf, Cleveland State University, & OSAMU ISHIHARA, Nihan McMaster University-According to a conditioning analysis of the University-We report the first studies of a cross-cultural project ex McCollough effect (ME), orientation stimuli become associated with amining simple arithmetic performance among Japanese and U.S. col chromatic stimuli. Contrary to this interpretation are reports that the lege students. Subjects rated the difficulty of basic addition and sub ME is not degraded by decreasing the contingency between the puta traction facts, then produced answers to these in both pure and mixed tive CS (grid orientation) and the putative UCS (color). In fact, as blocks. Although both groups show standard mental arithmetic effects, predicted by the Rescorla-Wagner model, when a "superordinate con the group differences in ratings and reaction times suggest a large educa text" is present, contingency manipulations affect the ME as they af tional and cultural impact on performance. fect other types of conditional responses. 9:50-10:00 (608) Signaled Avoidance of Proboscis Extension in Restrained Honey ASSOCIATIVE LEARNING: ANIMAL m Bees. CHARLES I. ABRAMSON, SUNY Health Science Center at Golden Gate A, Sunday Morning, 8:00-10:05 Brooklyn, & BRIAN H. SMITH, Ohio State University-Differential conditioning was used to assess the ability of harnessed foragers to dis Chaired by Stanley J. Weiss, American University criminate between two odors-()ne associated with sucrose and the other with sucrose plus a 10-V ac shock. The results indicated that most sub 8:00-8:15 (603) jects learn to extend their proboscis to the odor associated with sucrose Peak-Procedure Evidence for a Separately Changeable Dimension and to withhold proboscis extension to the odor associated with sucrose Selection Process. SETH ROBERTS & AFSHIN GHARIB, Univer plus shock. Analysis of the individual learning curves reveal several sity of California, Berkeley-When the probability of food is varied in types of response patterns. 527 Sunday Morning Papers 609-620 ~L~G~MORYW periments, subjects were trained in either a read or a generate condi Golden Gate B, Sunday Morning, 8:00-10:30 tion. In Experiment 1, subjects performed simple and difficult multipli cation problems. A generation advantage occurred only for the difficult Chaired by Mark A. McDaniel, Purdue University problems. In Experiment 2, subjects learned to associate nonword vocabulary terms with common English nouns. A generation advantage 8:00-8:15 (609) occurred, and in both conditions subjects using mnemonic strategies Gross Improvements in Text's Memory Representations witb showed superior performance. The results are explained in terms of a Minimal Misconception-Driven Revisions. BRUCE K. BRITfON & procedural account of the generation advantage. PAMELA TIDWELL, University of Georgia-Our goal was to stra 10:1&.10:25 (615) tegically target each punctate text change toward a central misconcep Effects of Caffeine on Automaticity in Human Information tion in the readers' mental representations, to get the maximum Processing Tasks. WING HONG LOKE, National University of learnability-improvement bang for each text change. We used cogni Singapore-The present study examined whether caffeine facilitates low tive maps of Air Force recruits (N = SO) mental representations after difficulty nonautomatic performance relative to placebo but decreases reading to discover their misconceptions, including previously un high-difficulty nonautomatic performance relative to placebo, and that suspected misconceptions. Misconception-driven text revisions in 6 of it does not affect automatic performance across levels of task difficulty 43 sentences in the I,OOO-word text improved to near ceiling (r = .8-.9) relative to placebo. The results supported an instance of the predicted the match of new recruits' (N = 1(0) representations to the text's author three-way interaction and also demonstrated the viability of using phar and seven subject matter experts. macological agents to test concepts of automaticity and other mecha 8:2~:35 (610) nisms of information prncessing. Tuning and Capacity Limitations in Feedback (Instrumental) LearnIng. LARRY E. ROBERTS, DENISE PRESTON, & BOB VTTL, PSYCHOPHYSICS McMaster University-We studied feedback arrangements that permit Embarcadero AB, Sunday Morning, 10:25-12:45 ted instrumental adaptation without spatially directed action. Response differentiation and self-report of feedback behavior were observed, but Chaired by Gregory R. Lockhead, Duke University subjects spent little time looking at what they were doing. When feedback was given in an auditory modality, subjects closed their eyes as training 10:25-10:35 (616) progressed. Feedback and perceptual tasks interfered with one another Sweet Taste of Water Induced by Uttle Red Pills (Pilula rul¥IlulG). when presented in the same (visual) modality. Tuning by a centripetal DEBRA A. ZELLNER, Shippensburg University, SCOTI PARKER, convergence of kinesthetic and exteroceptive pathways is suggested. American University, & CARINA TORNOW, Shippensburg Univer 8:40-9:00 (611) sity (read by Scott Parker)-Sucking on lin-tan, a lapanese-made breath Isolating Primary Memory Processes: Expectancy and Item Value freshener pill, makes water taste sweet. It also elevates ratings of the in Recall. THOMAS F. CUNNINGHAM, St. Lawrence University, sweetnesses of aqueous sucrose solutions. Postpill sweetness ratings of AllCE F . HEALY, University ofColorado, ROBERT E. TILL, Univer pure water exceed ratings of a .075 M sucrose solution before the pill. sity of North Dakota, DAVID W. FENDRICH, Widener University, The effeCt lasts for at least 2 min. lin-tan does not affect saltiness rat & CHRISTINA Z. DIMITRY, St. Lawrence University-In two experi ings of NaCI solutions. Neither lin-tan's taste nor its aftertaste is sweet. ments, subjects recalled one of two letter segments following a digit Gymnema sylvestre administered prior to the pill abolishes the effect. filled retention interval. In Experiment I, precues correctly informed 10:40-10:55 (617) or misinformed subjects conceming which letter segment would be tested. A Novel Model for QuaIIa. STANLEY A . KLEIN, University of In Experiment 2, one segment was cued as important but the uncued California, Berkeley-Progress is slowly but steadily being made on men segment was recalled first. Recall performance decreased very rapidly tal models of the "objective," testable aspects of subjective experience. when subjects were misinformed or when the segment was unimpor The one area that has continued to escape being pinned down by psy tant. The retention functions were compared to predictions of the Estes chologists and philosophers is quaIia, the "feel" of an experience. This perturbation model. talk shows how von Neumann's insight into the movable split of du 9:05-9:20 (612) alistic quantum mechanics provides the key to developing a surprisingly DIrect Coding for Frequency of Occurrence. JOHN JONIOES, consistent language for handling the qualia problem. University ofMichigan, & CAREN M. JONES, Stanford University 11:00-11:15 (618) Evidence for the contribution of direct frequency codes to frequency Visible Persistence as Error. GEORGE WOLFORD, Dartmouth judgments is demonstrated. Subjects were shown words whose frequen College-There are several counterintuitive findings associated with visi cies were later judged. Encoding conditions were manipulated by hav ble persistence. For example, visible persistence is inversely related to ing subjects generate either numerical or word associates to the stimuli. stimulus intensity and to stimulus duration. I will present additional Frequency estimates were poorer when the generated associates were counterintuitive findings and a new way of looking at visible persis numerical, though free-recall data indicated that the encoding condi tence that renders the various findings less counterintuitive. tions led to equal learning of the words. 11:2&.11:35 (619) 9:25-9:40 (613) Aerial Visual Acuity of a Risso's Dolphin at Two Distances. PAUL E. The Rhyme-Word MnemonIc Technique with Elementary School NACHTIGALL, Naval Ocean Systems Center, Hawaii Laboratory, & Children. RICHARD KRINSKY & SUZANNE G. KRINSKY, Univer lEFFREY L. PAWLOSKI, Science Applications International sity of Southern Colorado-In two experiments, the rhyme-word Corporation-Minimum angles of resolution for the vision of a Risso's mnemonic device was employed with elementary school-age children. dolphin were determined by presenting varying line width Ronchi rul In Experiment 1, fourth and fifth graders recalled reliably more target ings at distances of 1 and 2.5 m from the animal's eyes. Behavioral 75% items that were associated with rhyme words 10 months after original thresholds based on both two-alternative forced-choice and successively learning. In Experiment 2, fifth graders at risk for academic failure, presented go/no-go procedures demonstrated that, like the bottlenosed when compared with classmates who were not at risk, were similar in dolphin, Risso's dolphins, in air, resolve detail better at 2.5 m than they their immediate memory and disproportionally superior in their delayed do at 1 m. memory for target items associated with rhyme words. 11:40-11:55 (620) 9:45-10:05 (614) Measuring the Subjective MagnItude of Brain Stimulation Reward. A Generation Advantage for Multiplication Skill and Nonword . C . R. GALLISTEL, MATIHEW LEON, & TERRY A. MARK, Vocabulary Acquisition. DANIELLE S. McNAMARA & AilCE F. UCLA-We used rats' tendency to match the allocation of their time HEALY, University of Colorado (read by Alice F . Healy)-In two ex- between two levers to the relative net rates of return to measure the 528 Papers 621-631 Sunday Morning effect on the subjective magnitude of reward of varying current, pulse & JENNIFER WILEY, University of Pinsburgh-Statements of U.S. frequency, and train duration. Subjective reward magnitude grows as senators made during the Gulf Crisis debate were obtained from the Con a steep sigmoidal function of stimu1ation strength (current x frequency). gressional Record and analyzed with respect to metaphor and analogy It grows as an approximately linear function of train duration to an abrupt usage. The senators compared the political situation to movies, politi saturation at a duration of about 1 sec. cal events and personalities, historical analogies, and gambling. Usage 12:00-12:20 (621) was also related to a person's political position. Personalization of is Wave Similarity Theory or Same-DIfferent (Equality) Judgments. sues was also observed. The ftndings are considered in relation to one's STEPHEN LINK, McMaster University-Wave similarity theory describes political motives and to cultural standards. the way in which stimuli are compared in order to make judgments of 12:20-12:30 (627) same-different or equal-unequal. Such judgments are common to many Effect or Reading Strategies ror Proc:edural Text on Recall, Com psychological tasks that measure response times and response propor prehension, and Task Performance. CAROL BERGFELD MILLS, tions. The new theory accounts for the empirical results and permits Goucher College, VIRGINIA A. DIEHL, Western Illinais University, estimation of the psychological function for the feeling of eqUality. LIEN-CHONG MOU, University ofMaryland, & DEBORAH P. BIRK 12:25-12:40 (622) MIRE, U.S. Anny Human Engineering Laboratory-Fifty subjects read Is Measured Channel Capacity Limited by the Absolute Identifi procedural texts using one of three reading strategies: (I) they performed cation Paradigm? KENNETH A. DEFFENBACHER, University of the task described by the text while reading, (2) they looked at the 0b Nebraska at Omaha, FRANCISJ. CLARK, University ofNebraska Med ject involved in the task while reading, or (3) they only read the text. ical Center, M. ELIZABETH DAVIS, & KONNEY J. LARWooD, After reading the text, they recalled the text, answered true-false ques University ofNebraska at Omaha-Subjects either pointed to visual tar tions, and performed the task. The different strategies resulted in dif gets with the not-visible outstretched arm or absolutely identified (AI) ferent patterns of results for the three performance measures. the same targets with an integer. The AI paradigm showed a limit of 12:35-12:45 (628) about seven identifiable targets, irrespective of the number of targets, Cognitive Control in ReadIng Expository Text: Goal Manipuladons whereas the pointing paradigm showed no such limit. We argue that and Working Memory. PAUL WHITNEY & ROBERT BRAMUCCI, the AI paradigm estimate of channel capacity was imposed by the para Washington State University--One means for investigating whether digm itself, a limitation not imposed by any limits on precision of tar "schemata" affect encoding or retrieval is to test for recall of previ get acquisition. ously unrecalled information after a perspective shift. With narratives, we found that the effects of a perspective shift vary with working mem LANGUAGE PROCESSING ory span. We extended this research to expository text by manipulating Embarcadero CD, Sunday Morning, 11:00-12:50 high- and low-span readers' goals at encoding and retrieval. The results are relevant to understanding how readers adapt to their own working Chaired by Monon Ann Gemsbacher, University of Oregon memory constraints. 11:00-11:15 (623) PERCEPTION m The Relative Ease or Writing Narrative Texts. RONALD T. KEL San Francisco A, Sunday Morning, 10:30-1:00 LOGG, MERlLEE KRUEGER, & ROSE BLAIR, University of Mis souri-Rolla-The narrative writing task, we hypothesize, is most com Chaired by Andrea R. Halpern, Bucknell University patible with the structure of conscious thought. Hence, composing a narrative text should demand less cognitive effort, occur more fluently, 10:30-10:50 (629) and yield a more coherent dncument than composing persuasive and Are Texture Segregation by Form and Color Independent or Each descriptive texts. Analyses of secondary task reaction times and sub Other? MICHAEL KUBOVY & DALE COHEN, University of jective ratings confinned that narrative writing is least effortful. An anal Virginia-In gestalt detection (= RSVP + texture segregation), we ysis of coherence ties and document length indicated that persuasive writ embed in a sequence of distractor textures a target texture, which is ing exhibited the least cohesion and fluency. the only one to have a boundary. On some trials the target's boundary 11:20-11:35 (624) is defmed by color, by form, or by both. When more than one feature Locating Objects From Memory or From Sight. DAVID J. deftnes boundaries, they sometimes coincide. The subject reports the BRYANT, Nonheastern University, & BARBARA TVERSKY, Stan location of one boundary. Detection of a dual-feature coincident boundary ford University (read by Barbara Tversky)-Subjects learned two kinds is better than probability summation of the individual single-feature of spatial arrays from observation and were later probed for objects by boundaries. locations from memory or from perception. Memory reaction times 10:55-11:15 (630) yielded the spatial framework pattern found for arrays learned from Shape Recognition Contributions to the Organization or 3-D Dls descriptions (Bryant, Tversky, & Franklin, in press; Franklin & Tversky, plays. MARY A. PETERSON & BRADLEY S. GmSON, University 1990) but perception reaction times were different, indicating that these of Arizona-We examined whether shape recognition processes influence spatial mental models are not like internalized perceptions. the initial and sustained organization of 3-D displays, using stereograms 11:40-11:55 (625) in which binocular disparity specifted that one of two regions (equal Word Recognition in Continuous Speech. PATRIZIA TABOSSI, in area but not in meaning) was forward. Results showed that the Universita'di Bologna, DONIA SCOTT, Brighton Polytechnic, & meaningful region was seen in front, even when disparity specifted that CHRISTINA BURANI, CNR, Rome-Two cross-modal experiments the meaningless region was forward. These effects were obtained for investigate word identification in continuous speech. A word (e.g., SAND) contrast, but not random-dot, stereograms, constraining the nature of followed by another one (e.g., AWNE) whose first syllable forms with critical shape recognition mechanisms. the preceding word a longer item (e.g., SANDAL) facilitates the recog 11:20-11:35 (631) nition of a target associated to that item (e.g., SHOE) (Experiment 1). The Perceptual Identiftcation or Image Contours. JAMES T. TODD The effect, however, disappears when the ambiguity is removed (e.g., & PETER SCHNITTMAN, Brandeis University-Optical contours can SAND CASTLE) (Experiment 2). The results suggest that recognition cannot arise from a variety of physical causes including smooth occlusions, always be a sequential, word-by-word process. specular highlights and abrupt discontinuities of surface reflectance, 12:00-12:15 (626) orientation, or illumination. Our research has investigated the abilities Metaphor and Analogy Use by U.S. Senators in the Gulf Crisis of human observers to identify these contour types in natural images Debate. JAMES F. VOSS, JOEL KENNET, TONYA SCHOOLER, with minimal amounts of contextual information. In contrast to the re- 529 Sunday Morning Papers 632-642 cent arguments of Cavanagh and Leclerc (1989), our results indicate ORIEL STRICKLAND, University ofNorth Carolina at Chapel Hill that image contours can be reliably identified even within scenes that Two experiments demonstrate subjects' propensity to judge statements contain no recognizable objects. "true. " A model applied to judgments of complementary pairs of state 11:40-11:55 (632) ments in multiple paradigms separates covert judgment from response Auditory Pemived DisIaDI:e oIFamiIIIIr Speedt Sounds. DONALD H. selection. It is supported in a third experiment that shows the bias for MERSHON & JOHN W. PHILBECK, North Cm-olina State University- "true" to be one of response assignment relative to underlying judg Digital recordings of a man and a woman using different vocal styles ment. Another experiment asks within the context of the model whether were presented from a stationary loudspeaker to 192 blindfolded male covert judgments are biased in the true-false relative to the pair and female listeners in an acoustically dead space. Included were shouts, comparison paradigm. whispers, and a normal conversational style. Playback levels were ad 11:15-11:35 (638) justed to avoid extraneous sound level cues. The shouting voice was An Adaptive Strengtb-Comparison Model of Predicting Basket reported as appearing farthest, the whispering voice closest. The con ball Winners. EVAN HEIT, PAUL PRICE, University of Michigan, versational voice was intermediate. Auditory perceived distances may & GORDON BOWER, Stanford University (read by Gordon Bower) be affected by past experience. For each game of a simulated basketball season, subjects predicted the 12:00-12:15 (633) point-spread, then received feedback about the actual outcome. Sub "Adaptation" to Displacement Prisms Is Sensorimotor Skill Ac jects' predictions showed sensitivity to the win-loss record of each team, quisition. GEOFFREY P. BINGHAM, MICHAEL M. MUCHISKY, the strengths of each team's past opponents, and home-court advantage. & JENNIFER ROMAK, Indiana University-We show that adjustment Trial-by-trial performance was accounted for by a mathematical model to glasses is difficult only on the first day because one acquires a sen incorporating an adaptive learning rule that estimated each team's strength sorimotor skill. We measured trajectories in a rapid targeting task, al and a response rule that predicted outcomes by comparing team strengths. ternating sets of trials with monocular viewing and with monocular view 11:40-12:00 (639) ing through a 10° displacement prism. Trials were performed within TheoredaII Impliratiom 01 CbildreD's Decision Making. VALERIE F. a set to criterion accuracy. The number of trials per set decreased over REYNA & DANA NARTER, University ofArizona-Data demonstrat sets. Trajectories revealed continuous sensorimotor control along trajec ing phenomena of inconsistent choice in children (e. g., framing effects tories of duration of I sec or less. and the Allais paradox) are used to evaluate theoretical interpretations 12:20-12:40 (634) of adult data. The existence of such phenomena in children challenges The Acquisition of "Dual Adaptations" and "Adaptation Sets." theories whose process assumptions are implausible for immature sub ROBERT B. WELCH, NASA-Ames Research Center, BRUCE BRIDGE jects. We examine the relationship between specific cognitive abilities, MAN, SULEKHA ANAND, & KAITLIN BROWMAN, University of especially compensatory thought, and inconsistent choices, and we also California, Santa Cruz-In two experiments, we confirmed that repeat consider the developmental path that leads to adult performance: the edly alternating between two different prismatic displacements eventuates computation-intuition shift. in the ability to adapt and re-adapt more quickly-a "dual adaptation." 12:05-12:25 (640) We then asked the question: What is the underlying basis of this adap meets of Multiple Plays on Evaluations of Gambles. DOUGLAS H. tive flexibility? The answer, as demonstrated in our second experiment, WEDELL, University of South Carolina, & ULF BOcKENHOLT, Uni is an "adaptation set" (analogous to a "learning set"): Subjects have versity ofnlinois-Previous research has demonstrated reductions in de acquired the capacity to adapt more readily to prismatic displacements cision anomalies (preference reversals and certainty/possibility effects) never previously encountered. when gambles are represented as being played multiple times. The present 12:45-12:55 (635) research investigated effects of representing simple or more complex Local-Global Processing as a Function of Visual Field Locus. bets as being played once or multiple times on preferences elicited by STEPHEN CHRISTMAN, University of Toledo-Processing of Navon different response modes. Additionally, two techniques were used to type hierarchical letter stimuli in the upper versus lower and left versus infer subjective utility and probability functions. Theoretical interpre right visual fields was examined. In Experiment I, a lower visual field tations of the effects of the multiple play representation are discussed. advantage (VFA) was found for global processing, while an upper VFA 12:30-12:45 (641) was obtained for local processing. No significant left-right differences Judging the Effectiveness of Interventions. NIGEL HARVEY & were found. In Experiment 2, significant left and lower VFAs were ob FERGUS BOLGER, University College London-Sequences of a diag tained for global processing, while an upper VFA was again obtained nostic index were simulated with autoregressive algorithms. At some for local processing. point in half these series, a "drug" changed the overall level of the index. Subjects were poorer and slower at assessing drug effectiveness JUDGMENTIDECISION MAKING m in trended and cyclical series. They may subject the raw series to time San Francisco B, Sunday Morning, 10:25-12:50 consuming error-prone global analyses to filter out these overall pat tern components before assessing intervention effectiveness. We there Chaired by Lola L. Lopes, University of Iowa fore studied effects of presenting the series in small segments. 10:25-10:45 (636) ANIMAL COGNITION n Generation of Random Binary Series in Strictly Competitive Golden Gate A, Sunday Morning, 10:15-12:55 Games. DAVID V. BUDESCU, University of Haifa, & AMNON RAPOPORT, University ofArizona-Numerous studies have shown that Chaired by Sarah T. Boysen, Ohio State University human subjects are poor judges and generators of random sequences. In particular, they associate randomness with series including too many 10:15-10:30 (642) runs. These studies are criticized on several grounds and a new para A Spencian Model of Categorization. EDWARD A. WASSER digm is proposed based on performance in two-person zero-sum games. MAN, University of Iowa, & S. L. ASTLEY, Cornell College-We Results of an experiment, in which the regular "biases" associated with present a model of categorization based on Spence's 1937 theory of dis random series generation are almost completely eliminated, are reported. crimination learning. In addition to Spence's postulates of conditioned 10:50-11:10 (637) excitation, conditioned inhibition, and stimulus generalization, we pro Separating Response Bias from Judgment in Statement Veriflca pose that naturalistic stimuli are locatable along categorical dimensions. tion. THOMAS S. WALLSTEN, CLAUDIA G. GONZALEZ, & This framework explains most known facts of categorization in animals, 530 Papers 643-654 Sunday Morning including category learning and transfer to new examples, errors of THOMPSON, University of SoU/hem California-The present experi categorization, learning of true and pseudocategories, and the effects ment was designed to establish a highly localized memory trace within of repetition and category size on learning and generalization. the cerebellum, using classical conditioning of discrete motor responses. 10:35-10:45 (643) Stimulating electrodes were chronically implanted in rabbit cerebellum, Categorization and Jdentiflc:adoo, K. GEOFFREY WHlTE & AONA to provide an electrical CS activating cortical parallel fibers and thence McPHERSON, University of Otago-We examined category and identity Purkinje cells, and an electrical US activating white matter underlying matching by pigeons. Responses to a probe stimulus were rewarded ac lobule HVI and eliciting unconditioned responses. Preliminary results cording to whether the probe was identical to or in the same category indicate that associative conditioning occurs using this paradigm. (color or shape) as a preceding target stimulus. Increasing the retention interval between target and probe stimuli produced a greater discrimi HUMAN LEARNINGIMEMORY V nation decrement for identity matching than for category matching. We Golden Gate B, Sunday Morning, 10:40-12:55 conclude that categorization is based on matching of features and is faster and more accurate than identification. Chaired by Judith A. Sugar, Colorado State University 10:5&-11:05 (644) Development of Backward Associations during Establishment of 10:40-10:50 (650) Forward A.ssocIatiom by PIgeons. THOMAS R. ZENTALL, LOU M. Remembering the Earthquake: "Flubbulb" Memory for Ex SHERBURNE, UniversityofKenlUC/cy, & JANICE N. STEIRN, Georgia perienced Versus Reported Events. STEPHEN PALMER, CHARLES Southem University-Following conditional discrimination (DMTS) train SCHREffiER, & CRAIG FOX, University of California. Berkeley ing involving differential outcomes (food vs. no food), one of the two Within 3 days after the 1989 California earthquake, University of Califor samples was replaced by "food" and the other by "no food" presentation. nia, Berkeley, students recalled their experiences both for a direct ex Transfer data from three experiments indicated that pigeons developed perience (the earthquake itself) and a reported event (the Bay Bridge backward associations between correct comparisons and their associated collapse). Memory was retested 1.5 years later. High-i:onfidence, vir outcomes during original training. These results cannot be attributed tually error-free recall was obtained for several questions about the ex either to sample-outcome associations or to response mediation (i.e., perienced earthquake, but performance was significantly worse for cor the similarity between sample responding during training and testing). responding questions about the reported bridge collapse, despite greater 11:10-11:30 (645) concern. The results will be compared with memory by students at Emory Relative Size Discrimination by an African Grey Parrot. IRENE M. University and University of California, Santa Cruz, for the same PEPPERBERG, University of Arizona-Some laboratory studies sug earthquake. gest that birds respond relationally only after learning about absolute 10:55-11:10 (651) values of stimuli. A parrot, Alex, learned to answer questions concern Remembering the Earthquake: "What I Experienced" Versus "How ing the relative size of certain exemplars and transferred to novel situa I Heard the News." ULRIC NElSSER, EUGENE WINOGRAD, Emory tions that did not provide information about the absolute values of stimuli. University, & MARY SUSAN WELDON, University of California, He answered (accuracy of78.7%) vocal questions about which was the Sonta Ouz-Shortly after the 1989 California earthquake, students at larger or smaller exemplar by vocally labeling its color or material, or Emory University recorded how they had first heard about it; three weeks saying "none" for equally sized exemplars. after the quake, students at University of California, Santa Cruz, recorded 11:35-11:55 (646) how they had actually experienced it . Tested 1.5 years later, some of Transitive Inference in Pigeons. JANICE N. STEIRN, Georgia the Emory subjects made gross recall errors despite high confidence; Southem University, & THOMAS R. ZENTALL, University ofKentucky no Santa Cruz subjects made such errors. Affect was uncorrelated with Pigeons' use of transitive inference was examined in a task involving recall in either sample. These data will be compared with direct and successively presented simple hue discriminations of the form A + B - , indirect recall of the same earthquake by Berkeley students. B+C-, C + D-, D+ E-. When nonadjacent, nonendpoint stimuli (BD) 11:15-11:35 (652) were then tested, strong evidence for transitive inference (i.e. , prefer The Role of Autobiograpbical Memories in TraIt Judgments about ence for B) was found. Explanations of pigeons' performance on a tran Self and Others. STANLEY B. KLEIN & JUDITH LOFTUS, Univer sitive inference task were examined. sity of California, Santa Barbara-We report studies exploring the role 12:00-12:15 (647) of behavioral exemplars (autobiographical memories) in trait judgments Delayed Auditory Feedback Disrupts Bird Song Production. about self and others. The studies use a priming paradigm to examine JEFFREY CYNX & UTA VON RAD, Rockefeller University-Adult whether (a) retrieval of trait-relevent autobiographical memories facili bird song production, unlike juvenile song development, has rarely been tates a judgment of the trait's descriptiveness and (b) a trait descrip approached as a cognitive phenomenon. Both ethological and neuro tiveness judgment facilitates retrieval of trait-relevant autobiographical ethological studies have suggested that it is an autonomous motor pro memories. We fmd considerable facilitation for judgments about other, gram. However, the use of delayed auditory feedback indicates that adult but none for judgments about self. We discuss the implications of these song production, like human speech, is amenable to cognitive processes findings for exemplar and abstraction models of social judgment. that govern the relation between perception and production of learned 11:40-11:50 (653) vocalizations. Aging and Time- Versus Event-Based Prospective Memory. 12:20-12:35 (648) GILLES O. EINSTEIN, Furman University, & MARK A. McDANIEL, Periodc RfspInIe PaUems in Tempni Seardl BeDavior. HILARY A. Purdue University-We tested for age differences in prospective mem BROADBENT & RUSSELL M. CHURCH, Brown University (read ory. In the first experiment, subjects were asked to perform an action by Russell M. Church)-Rats on an A schedule received food prob every 10 min (a time-based task presumed to be high in self-initiated abilistically. Following food, responses occurred near the end of the retrieval); in the second experiment, subjects were asked to perform next interval, suggesting that food delivery resets an interval timer. Fol an action whenever a particular word was presented (an event-based lowing omission, bursts of responses occurred throughout the interval, task presumed to be low in self-initiated retrieval). Age differences were suggesting the timer is not reset after omission: animals must search found only with the time-based task, thereby suggesting (a) important through time for the next reinforcer. Spectral analysis (Walsh transform) differences between time- and event-based prospective memory tasks indicated a strongly significant periodicity in the bursts. This periodic and (b) that self-initiated retrieval is an important determinant of age ity may indicate an endogenous oscillator that controls searching. differences in prospective memory. 12:40-12:50 (649) 11:55-12:10 (654) Classical Conditioning witb Intracerebellar Electrical Stimulation Evidell(e for Using PartIal Information to Reconstruct Dates of as CS and US. PAUL G. SHINKMAN, University ofNorth Carolina, Events. CHARLES P. THOMPSON, Kansas State University, JOHN J. RODNEY A. SWAIN, ALAN F. NORDHOLM, & RICHARD F. SKOWRONSKI, Ohio State University at Newark, & ANDREW L. 531 Sunday Morning Papers 655-656 SETZ, Ohio Stale University-We present a theoretical distribution of selves for a test containing free-recall, cued-recall, and specific errors based on the use of partial information in reconstruction of dates knowledge reconstruction. Results are discussed in terms of these time of personal events. The theoretical distribution is compared to data from bound variations as well as in tenns of different levels of information a diary memory study. The model captures the major characteristics processing. of the data. 12:30-12:50 (656) 12:15-12:25 (655) An Extreme Statistics Model for ForgettiDg and Retrieval. TAROW When and How to Prepue Best for Test about tbe Professor's Lec INDOW, University of California. Irvine-Forgetting and retrieval of tures? DANIEL SLOM, KARIN HEGAR, DIETER HAENGGI, & relevant items from long-term memory are interpreted as Weibull GERHARD F. STEINER, University of Basel (read by Gerhard F. processes. A parameter (the exponent) gives useful insight about the Steiner)-In a series of three experiments, college students took notes structure of underlying memory supports and cues to lead to retrieval. from ordinary lectures, reprocessed their notes according to different Methodology of parameter estimation in independent and nonindepen time schedules (spacing and retention intervals were varied), as well dent data points is discussed. Memory of social events over a span of as reprocessing procedures (massed vs. spaced), thus preparing them- 30 years and very short retrieval processes are used as data. 532