• 1: Contiguity and Structure in Learning
• 2: Word and Text Comprehension
• 3: Visual Attention and Awareness

• 4: Reading Faces
• 5: Thinking and Reasoning
• 6: Tracing Processing through Fixations and Saccades

Wednesday 10-04, 17:00 - 18:30

• 7: Modeling Semantic Structure
• 8: Phonology in Speech and Reading
• 9: Object Discrimination

• 10: Codes in Memory
• 11: Shape, Space, and Language

Thursday 11-04, 15:30 - 16:30

• 12: Social Attitudes & Beliefs
• 13: Development of Concepts & Percepts

• 14: Strategy, Working Memory Capacity, and Inhibitory Control

Session 1: Contiguity and Structure in Learning (O1.1 - O1.4)

O1.1

Beyond spatio-temporal contiguity: Natural and learned audio-visual pairings use different neural integration sites

Gilles Pourtois^{1, 2} and Beatrice de Gelder^{1, 2}

¹ Tilburg University, The Netherlands
² Université de Louvain, Belgium

E-mail: g.r.c.pourtois@kub.nl

Studies of inter-sensory integration of auditory and visual information tended to assume that temporal and spatial contiguity is the single most important factor determining intersensory integration. In the present study we asked whether the specific type of audio-visual pairing also plays a role. If so, spatio-temporal contiguity is not sufficient for integration and neuronal sites of audio-visual integration will reflect content-sensitivity. We studied whether Transcranial Magnetic Stimulation (TMS) disrupted audio-visual integration as reflected in the redundant target advantage. Single pulse TMS applied over the left posterior parietal cortex at 50, 100, 150 and 200 ms disrupted integration at 150 ms and later of audio-visual shape/tone pairs (Learned condition) but not of voice/face pairs (Natural condition). Our result indicates that besides spatial and temporal contiguity, content specificity is an important determinant of audio-visual integration.

O1.2

Hierarchical coding of serially ordered spatial information: evidence from analyses of time to generate the next step in the sequence

Carlo De Lillo

University of Leicester, UK

E-mail: cdl2@le.ac.uk

The results of two experiments are reported. In the first experiment participants were presented with a configuration of nine identical icons, spatially grouped in three clusters of three icons each, on a touch sensitive computer monitor. The icons flashed according to sequences segregated by spatial clusters and, after a short delay, the participants were required to reproduce the sequences by touching the screen. An analysis of the latencies for touches corresponding to different ordinal positions in the sequence revealed longer latencies at cluster boundaries. These results corroborate previous findings and are compatible with the hypothesis that a two-level hierarchical representation underpins the reproduction of the sequences.

The second experiment was carried out to rule out the possibility that results of experiment one could have been accounted for by the relative length of the pointing movements. As in experiment one, the participants observed first a sequence of flashing icons on the screen. Then, segments of the sequence were presented. The participants were required to produce a single pointing movement to the icon corresponding to the next ordinal step in the sequence, with their hand starting from a resting position located at a fixed distance from the screen. Longer latencies emerged again for pointing responses directed to icons at ordinal positions corresponding to cluster boundaries. Taken together, the results of these two experiments provide strong evidence for the emergence of a form of spatial chunking in tasks requiring the encoding of serially ordered information within spatially structured patterns.

O1.3

High-order information reprocessing during post-training paradoxical sleep

Philippe Peigneux¹, Steven Laureys¹, Axel Cleeremans², and Pierre Maquet¹

¹ Cyclotron Research Center, Université de Liège, Belgium
² Cognitive Science Research Unit, Université Libre de Bruxelles, Belgium

E-mail: Philippe.Peigneux@ulg.ac.be

The brain areas needed for the practice of a serial reaction time (SRT) task are more active during subsequent REM sleep in subjects trained to the task than in non trained subjects [Maquet et al. (2000) Nature Neuroscience, 3(8), 831-836]. These experience-dependent cerebral reactivations during post-training REM sleep could reflect the reprocessing of recent memory traces. Here, 6 new subjects were trained to the same SRT task then scanned during the post-training night using PET. The only difference with the trained group in the Maquet et al. study was that the sequence of stimuli was at random. Therefore, both trained groups underwent similar visuo-motor basic learning, but only one trained group learned complex, high-order, rules underlying the sequence of stimuli during SRT practice. Hence, this group only could potentially reprocess high-order information during post-training sleep. We found that post-training REM sleep activity in the thalamus, occipital and premotor cortex increased more in subjects trained to a structured sequence (with rules) than in subjects trained to the random sequence. It suggests that experience-dependent cerebral reactivations during post-training REM sleep underlie the reprocessing of high-order information about the structure of the sequential material to be learned.

Supported by FNRS, FMRE, ULg Funds

O1.4

Perceptual or motor learning with artificial and fixed sequence structures

Natacha Deroost, Annemie Melis, and Eric Soetens

Vrije Universiteit Brussel, Belgium

E-mail: Natacha.Deroost@vub.ac.be

We investigate whether sequence learning in a serial reaction time (SRT) task is primarily perceptual or motor by varying the stimulus dimension on which the sequence structure is imposed. In a previous study, with fixed sequences, a two-choice task with symbolic S-R mapping was used. Learning only occurred in the condition where both dimensions were structured. There was no evidence for pure perceptual (location) or motor (colour) learning. In a subsequent study we replicated the experiment with an artificial grammar, using a four-choice task. To find out whether motor learning would occur, the structure was imposed on the relevant colour dimension, while location varied randomly. Unlike the first study, participants were able to learn the motor sequence. In the present study we investigated if the same results hold with a fixed sequence structure. Subjects again had to respond to colour, while ignoring the irrelevant location of the stimulus. The following three conditions were manipulated between-subjects: a repeating 32-element sequence was imposed either on the relevant colour dimension, on the irrelevant location dimension or on both stimulus dimensions. Results will be discussed in the course of the presentation.

Session 2: Word and Text Comprehension (O2.1 - O2.4)

O2.1

Comprehension complexity and corpus frequencies in noun phrase conjunction

Timothy Desmet¹ and Edward Gibson²

¹ Department of Experimental Psychology, Ghent University, Belgium
² Brain and Cognitive Sciences, Massachusetts Institute of Technology, USA

E-mail: timothy.desmet@rug.ac.be

A number of researchers have proposed that sentence comprehension is frequency driven, such that the ease of understanding a syntactic construction depends on its frequency of use (e.g., Mitchell, Cuetos, Corley, & Brysbaert, 1995). In apparent contradiction to such accounts, Gibson and Schütze (1999) showed that on-line disambiguation preferences do not always mirror corpus frequencies. When presented with the syntactic ambiguity involving the conjunction of a noun phrase to three possible attachment sites, participants were faster to read attachments to the first site than attachments to the second one, although the latter were shown to be more frequent in text corpora. In the present study, we investigated whether a particular feature in the items of Gibson and Schütze - disambiguation using the pronoun "one" - could account for the discrepancy they found. On the basis of an investigation of the Brown and Wall Street Journal text corpora and a self-paced word-by-word reading study, we conclude that there is no discrepancy between on-line preferences and corpus frequencies, and consequently there is no need to assume different processes underlying sentence comprehension and sentence production based on this syntactic ambiguity, as Gibson and Schütze had hypothesized.

O2.2

Cognitive aging effects on integration processes during text comprehension

Laurence Demanet¹, Marie-Anne Schelstraete¹, Michel Hupet¹, and Guy Denhière²

¹ Unit Cognition and Development, U.C.L., Belgium
² C.N.R.S. Université d'Aix-Marseille, France

E-mail: laurence.demanet@psp.ucl.ac.be

In his model of text comprehension, Kintsch (1998) postulated integration processes allowing the incorporation of new information to already known information at different levels. On the first level, some parts of the already-built text representation are selected by the leading edge strategy (L.E.S.) and maintained in a limited capacity working memory buffer (Kintsch & van Dijk, 1975). This L.E.S. allows the readers to store in the buffer the most important and the most recent propositions of the current processing cycle. On the second level, the content of the buffer is supposed to be regularly updated and integrated in the long-term working memory (Ericsson & Kintsch, 1995) to progressively build a coherent text base. Finally, on a third level, some knowledge-based information stored in the long term memory- e.g. inferences- are recovered during the reading to ensure the comprehension. Many studies showed an age-related decline in text comprehension (Van der Linden & Hupet, 1994) that could be interpreted by integration problems (Hess, 1995). The purpose of the present research was to measure how aging could affect the three different levels of integration. Three groups of participants (young, old and very old) were compared in three different tasks: A Reading-Recognition-Comprehension Task to test the L.E.S. and the comprehension, a Verbal Reconstruction Task to assess integrative function of long term working memory and finally a Reading Task of sentence pairs varying in causal relatedness to study inferential processes. Results indicated that very old participants have difficulties to draw inferences and a reduced capacity to integrate information in long-term working memory.

O2.3

Deficits in regular past tense processing: Delayed activation of semantic representations

Catherine E. Longworth¹, William D. Marslen-Wilson² and Lorraine K. Tyler¹

¹ Centre for Speech and Language, Department of Experimental Psychology, University of Cambridge, UK
² MRC Cognition and Brain Sciences Unit, Cambridge, UK

E-mail: cat@csl.psychol.cam.ac.uk

Double dissociations between the regular and irregular English past tense (Marslen-Wilson and Tyler, 1998, Tyler, de Mornay-Davies, Anokhina, Longworth, Randall and Marslen-Wilson, 2002) imply separable systems processing morphophonologically complex and fully listed words. An alternative account (Joanisse and Seidenberg, 1999) argues that these double dissociations reflect impairments of phonology or semantics, with regular past tense deficits reflecting impaired phonological and spared semantic processing. We tested whether activation of semantic representations from regular past tense forms is normal in patients with regular past tense deficits. Healthy volunteers show equal semantic facilitation from verb primes irrespective of verb regularity and tense (Longworth, et al., 2001). If non-fluent aphasics have a normal time course of semantic activation from the regular past tense they should show similar priming. 4 non-fluent patients were tested using an auditory semantic priming/lexical decision task. Prime-target pairs (e.g. "jumped LEAP") were selected by manipulating semantic relatedness, verb tense and regularity. Verb stems showed semantic priming irrespective of regularity. There was no semantic priming for the regular past tense but normal priming for the irregular past tense. This suggests that patients with regular past tense deficits have an abnormal time course of semantic activation from the regular past tense.

O2.4

Effects of ambiguity: Further evidence from semantic categorization

Jennifer Rodd 

Centre for Speech and Language, Department of Experimental Psychology, University of Cambridge, UK

E-mail: jrodd@csl.psychol.cam.ac.uk

Semantic ambiguity can be used as a tool to investigate how the meanings of words are represented. Rodd, Gaskell and Marslen-Wilson (in press) report slower lexical decisions for words with multiple unrelated meanings (e.g. bark) compared with unambiguous words (e.g. cage). In contrast, they report a benefit for words with multiple related word senses (e.g. twist). I present a connectionist model that simulates these apparently opposite effects, in which retrieving word meanings is characterised as activating distributed semantic representations. The ambiguity disadvantage arises because of interference between different meanings, while the sense benefit arises because of differences in the structure of the attractor basins. Words with few senses develop deep, narrow attractor basins, while words with many senses develop shallow, broad basins. The model suggests that the sense benefit arises in lexical decision because only general semantic information is retrieved. It predicts that the sense benefit should be reversed when detailed semantic information must be retrieved. This prediction is confirmed in two semantic categorisation experiments. These findings are consistent with the view that the meanings of words correspond to stable states within a high-dimensional semantic space, and that variation in the meanings of words shapes the landscape of this space.

Session 3: Visual Attention and Awareness (O3.1 - O3.4)

O3.1

Irrelevant attention shifts produce a Simon effect

Wim Notebaert and Eric Soetens

Vrije Universiteit Brussel, Belgium

E-mail: wim.notebaert@vub.ac.be

In a Simon task subjects have to react to a non-spatial feature (e.g., colour) of the stimulus with a left or right response. Stimuli are presented to the left or right of the fixation cross. Reaction times (RTs) are faster when the irrelevant location of the stimulus corresponds with the response location. In the present study we conducted two experiments with auditory stimuli to further support the claim that the Simon effect is caused by an attention shift towards the stimulus location. In Experiment 1, subjects have to react to the pitch of a tone that is presented in the left or right ear, with a left or right response. Two response-stimulus intervals (RSI) are used in order to manipulate the attention shifts during the task. With a short RSI, there is no Simon effect for location repetitions because no attention shift towards the stimulus location is needed In Experiment 2, subjects had to react to the colour of a centrally presented visual signal. At stimulus onset, an irrelevant tone is presented in the left or right ear. The data demonstrate a Simon effect according to the irrelevant auditory signal. This demonstrates that even an irrelevant attention shift produces a Simon effect.

O3.2

Illusory perceptions of space and time preserve cross-saccadic perceptual continuity

Kielan Yarrow, Patrick Haggard, Ron Heal, Peter Brown and John C. Rothwell 

Sobell Department of Motor Neuroscience & Movement Disorders, Institute of Neurology, University College London, UK

E-mail: k.yarrow@ion.ucl.ac.uk

When glancing at a clock with a silently ticking second hand, observers sometimes perceive the first second to take longer than subsequent seconds do; time appears to be stretched following a saccade (Brown and Rothwell, 1997). A matching methodology was employed to further explore this temporal illusion (chronostasis). Subjects saccaded to a digital counter, with the resulting electro/infra-red oculogram used to trigger counter onset. The duration of the first count was varied, with subsequent counts taking one second. Subjects made a forced choice response (first count less than/greater than subsequent counts), with either a modified binary search procedure or logistic regression analysis employed to provide a single matched estimate. A control condition yielded a matched estimate in the absence of a saccade.

Experiment 1 established chronostasis' dependency upon saccade size (22º, 55º). Results suggest that a backward matching mechanism is being employed , with stimulus onset being predated to saccade onset (minus a constant). Follow up experiments indicated that chronostasis is genuinely saccade dependent. It does not disappear when preceded by active orienting of attention in the direction of the upcoming saccade, suggesting that attention is not the crucial factor (experiment 2). No chronostasis was observed when the counter moved to the point of fixation with no saccade (experiment 3). Chronostasis is critically dependent upon the brain's assumption that the saccade target (the counter) remains constant across the saccade. When one aspect of this constancy (position) is noticeably violated, chronostasis disappears (experiment 4).

Brown, P., and Rothwell, J.C. (1997). Illusions of time. Society for Neuroscience abstracts: 27th annual meeting. 23 (2): 441.4: 1119. Yarrow, K., Haggard, P., Heal, R., Brown, P. & Rothwell, J.C. (2001). Illusory perceptions of space and time preserve cross-saccadic perceptual continuity. Nature, 414, 302-305.

O3.3

Visible persistence, informational persistence, and visual short-term memory in change detection tasks

Filip Germeys and Caroline Van Eccelpoel 

Laboratory of Experimental Psychology, University of Leuven, Belgium

E-mail: filip.germeys@psy.kuleuven.ac.be

Over the past ten years, research on 'change blindness', or the inability to detect changes to an object or scene, has grown steadily. Most of this work has focussed on the nature of our representations of the visual world. The striking inability to detect changes across blanks in the 'flicker' paradigm (an original and modified image are presented in alternation with a blank screen between them) and 'detection' paradigm (observers only receive one view of each image, separated by a blank) have mainly been interpreted as evidence for the idea that observers never form a complete, detailed representation of a scene. Furthermore, focussed attention is required to perceive change. In the absence of focussed attention, the contents of visual memory are simply overwritten by subsequent stimuli, and so cannot be used to make comparisons (Rensink, O'Regan, & Clark, 1997). Attention seems to insulate against the overwriting process, by transferring the item from a volatile to a more durable memory store (i.e. visual short-term memory, VSTM). The present study investigated the role of iconic memory (both visible persistence and informational persistence), attention, and VSTM in change detection. Five experiments using a change detection paradigm with arrays of letters, objects, or meaningless stimuli are presented.

Rensink, A., O'Regan, J. K., and Clark, J. J. (1997). To see or not to see: The need for attention to perceive changes in scenes. Psychological Science, 8, 368-373.

O3.4

Probing the prerequisites of experimental 'blindness'

Michael Niedeggen¹, Guido Hesselmann¹, Arash Sahraie², and Maarten Milders²

¹ Experimental Psychology II, Duesseldorf, Germany
² Department of Psychology, Aberdeen, UK

E-mail: michael.niedeggen@uni-duesseldorf.de

Using the technique of rapid serial visual presentation (RSVP), the conscious perception of coherently moving dots can be prevented. A brief transient 'motion blindness' is induced if subjects are asked to detect the presence of a local colour cue before switching attention to a global random-dot kinematogram. Essential for this effect is the presence of episodes of coherent motion (distractors) prior to the local colour cue. Using event-related brain potentials (ERPs) we examined which information processing stage is specifically affected by distractors. ERPs were separately analysed for three temporal distractor positions in the RSVP stream (early, medium, or late), and compared with two control conditions where the same motion stimuli were either relevant or irrelevant. A specific response was reflected in a frontal negativity at about 250 ms which increased with increasing number of motion distractors, and a centro-parietal positivity at about 350 ms which decrease with increased number of distractors. Early sensory processing, obtained in a posterior negativity at about 170 ms, remained unaffected. Our results suggest augmented activation in a frontal control system, triggered by distractors. This mechanism apparently prevents visual awareness, not by reducing the sensitivity in the visual cortex, but by gating to higher-order stimulus evaluation.

Session 4: Reading Faces (O4.1 - O4.4)

O4.1

The "Who Said What" Paradigm Revisited: Insights from Multidimensional Scaling

Christophe Labiouse

University of Liege, Belgium & Belgian NFSR Research Fellow

E-mail: clabiouse@ulg.ac.be

The original "Who Said What" paradigm was proposed by Taylor and colleagues (1978) and had provided an elegant tool to track issues related to social categorization. We propose some refinements to the original paradigm. A traditional element of the paradigm is that within-category differences are minimized and between-category differences are exaggerated. We show that this claim cannot be tested within the original framework. Therefore, instead of using real faces, we use morphed faces progressively modified by constant amounts between two face endpoints, and for which a priori baseline inter-similarities can be computed. In this way, we are able to test accentuation effects with the aid of multidimensional scaling. We use continua involving inter-ethnic faces. Evidence for social categorization and accentuation effects were found despite the absence of clear-cut categories, suggesting that people, even implicitly, exaggerate inter-categorical differences and reduce intra-categorical differences in order to cope with complex situations.

O4.2

Putting names to faces: No theory seems to work.

Mike Burton, Rob Jenkins and Allan McNeill

University of Glasgow, UK

E-mail: mike@psy.gla.ac.uk

It is well-established that retrieval of a person's name is generally harder (slower and more error prone) than retrieval of personal facts. Some recent theories have proposed that the difference between names and personal facts lies in the links between token makers and lemmas (e.g. Hollis & Valentine, 2001; Brédart et al, 1995). In essence, personal facts are stored in some semantic/conceptual system which in turn links to lemmas, whereas names have no semantic representation, and link to lemmas directly from token markers. This gives rise to the prediction that names should be easier to articulate than facts, when subjects are presented with a face. In two experiments we show that this prediction is not upheld. In common with other tasks comparing name and fact retrieval, subjects are slower to articulate names than facts, even when they are heavily over-trained on a small number of items (Experiment 1) or when the same words are used as facts or names (Experiment 2).

O4.3

Reading the mind from eye gaze

Andrew J. Calder¹, Andrew D. Lawrence¹, Jill Keane¹, Sophie Scott², Adrian M. Owen¹, Ingrid Christoffels³ and Andrew W. Young⁴

¹ MRC Cognition and Brain Sciences Unit, University of Cambridge, UK
² University College London, UK
³ University of Amsterdam, The Netherlands
⁴ University of York, UK

E-mail: andy.calder@mrc-cbu.cam.ac.uk

Baron-Cohen, S. (1997). Mindblindness: An essay on autism and theory of mind. Cambridge, Masachusetts: MIT press.

O4.4

The mentalistic significance of the mouth: A comparison of deaf and hearing children's use of mouth and eye information in schematic face reading.

L Jackson¹, R Campbell², and H Ellis³

¹ University of Reading, UK
² Human Communications Science, University College London, UK
³ Cardiff University, UK

E-mail: l.jackson@reading.ac.uk

In reading facial displays of emotion, deaf children may also attend more closely to mouth patterns than do hearing children (Jackson, 2000). We therefore hypothesised that a group of deaf children, predominantly from hearing families, may show idiosyncratic abilities in reading intention from schematic faces (The Charlie task - Baron-Cohen et al., 1995). After establishing that all these children were skilled at accurately indicating line of sight ('where is Charlie looking?'), some variants on the task were introduced. In Experiment 1, Charlie's line of sight was ambiguous, but the schematic mouth pattern indicated like ('smiley') or dislike ('sad') in relation to pictured objects of positive and negative valence. Both deaf and hearing children utilised mouth patterns to disambiguate eye gaze. Experiment 2 showed that deaf, but not hearing children, incorporated the schematic mouth pattern when inferring desire. This group of deaf children use the eyes to attribute mental states in a similar way to hearing children, despite indications that some deaf children have difficulty on ToM tasks dependent on reading 'the language of the eyes' (Jackson, 2000). They additionally take account of mouth patterns in reading faces to a greater extent than hearing children. Under these conditions, hearing, not deaf children, are blind to the 'mentalistic significance of the mouth'.

Baron-Cohen, S., Campbell, R., Karmiloff-Smith, A., Grant, J. & Walker, J. (1995). Are children with autism blind to the mentalistic significance of the eyes? British Journal of Developmental Psychology, 13, 379-398. Jackson, A. L. (2000). Do deaf children understand the language of the eyes? Paper presented at EPS conference, April 2000. Nottingham, UK.

Session 5: Thinking and Reasoning (O5.1 - O5.4)

O5.1

Does everyone "take-the-best"? Empirical tests of a ‘fast and frugal' decision heuristic.

Ben R. Newell, David R. Shanks and Nicola J. Weston

University College London, UK

E-mail: b.newell@ucl.ac.uk

The "Take-the-Best" (TTB) heuristic (e.g., G.Gigerenzer & D.G. Goldstein, 1996) states that when making a choice between two alternatives, people will base their choices solely on what they perceive as the most valid – or ‘best' – piece of information that discriminates between the alternatives. We report data from a series of experiments in which aspects of the experimental environment were manipulated to examine the parameters under which such a strategy operates. Clear evidence of TTB use was detected in all the experiments. However, the experiments also demonstrated that even under conditions contrived to promote the use of TTB – high cost of information, prior instruction as to the validity of cues, and a deterministic environment – there was a high proportion of behavior that was inconsistent with TTB. Together with the presence of large individual variability in strategy use in the experiments, these results question the validity of TTB as a psychologically plausible and pervasive model of behavior.

Gigerenzer, G., & Goldstein, D. G. (1996). Reasoning the fast and frugal way: Models of bounded rationality. Psychological Review, 103, 650-669.

O5.2

Insensitivity to prior causal knowledge in inference

F. J. López¹, A. Caño¹, P. L. Cobos¹, J. Almaraz¹ and D. R. Shanks²

¹ University of Málaga, Spain
² University College London, UK

E-mail: frjlopez@uma.es

In predictive causal inference people reason from causes to effects while in diagnostic inference they reason from effects to causes. Independently of the causal structure of the events, the temporal structure of the information provided to a reasoner may vary, e.g., multiple events followed by a single event versus a single event followed by multiple events. We report 2 experiments in which causal structure and temporal information were varied independently. The experiments use an overshadowing design that allows an unambigous evaluation of sensititvity to prior causal knowledge. The results reveal that inferences were influenced by the temporal but not by the causal structure of the task. These findings are relevant to the evaluation of two current accounts of causal induction, the Rescorla-Wagner (Rescorla & Wagner, 1972) and Causal Model theories (e.g., Waldmann & Holyoak, 1992).

Rescorla, R. A., & Wagner, A. R. (1972). A theory of Pavlovian conditioning: Variations in the effectiveness of reinforcement and nonreinforcement. In A. H. Black & W. F. Prokasy (Eds.), Classical conditioning II: Current theory and research, (pp. 64-99). New York: Appleton-Century-Crofts. Waldmann, M. R., & Holyoak, K. J. (1992). Predictive and diagnostic learning within causal models: Asymmetries in cue competition. Journal of Experimental Psychology: General, 121, 222-236.

O5.3

Utilising threat: Anticipated regret in deontic reasoning

Nick Perham and Mike Oaksford

Cardiff University, UK

E-mail: PerhamNR@cardiff.ac.uk

Three competing hypotheses were identified concerning the effects of varying threat material in deontic reasoning tasks, i.e., reasoning about what one should or should not do: attentional bias (Evans, 1996), decision theory (Manktelow & Over, 1991; Oaksford & Chater, 1994) and anticipated regret theory (Zeelenberg, van Dijk, Manstead, & van der Pligt, 2000). Experiment 1 discounted an attentional bias account because the effect of the threat material was observed for responses involving no explicit threat. Experiment 2 discounted standard decision theory, because it predicts similar effects when the threat material relates to a response that should be made as when it relates to a response that should not be made. However, equal and opposite effects were observed, which was only predicted by anticipated regret theory. The results showed that anticipated emotions affect deontic reasoning suggesting a productive theoretical synthesis of research in emotional decision making and deontic reasoning.

Evans, J. St.B. T. (1996). Deciding before you think: Relevance and reasoning in the selection task. British Journal of Psychology, 87, 223-240. Manktelow, K. I., & Over, D. E. (1991). Social roles and utilities in reasoning with deontic conditionals. Cognition, 39, 85-105. Oaksford, M., & Chater, N. (1994). A rational analysis of the selection task as optimal data selection. Psychological Review, 101, 608-631. Zeelenberg, M., van Dijk, W. W., Manstead, A. S. R., & van der Pligt, J. (2000). On bad decisions and disconfirmed expectancies: The psychology of regret and disappointment. Cognition and Emotion, 14(4), 521-541.

O5.4

Dissociating memory processes underlying misinformation effects in young children.

Robyn Holliday

University of Kent at Canterbury, UK

E-mail: r.holliday@ukc.ac.uk

This study investigated the contribution of automatic and intentional memory processes to 4- and 5-year-old and 8- and 9-year-old children's acceptance of misleading suggestions. Children were presented with an event and this was followed the next day by a post-event summary containing misleading suggestions that were either read to participants or were self-generated in response to semantic and perceptual cues. Individual children were then given either a Structured Interview or a modified Cognitive Interview. A short time later, children were asked a series of yes / no questions on their memories for the video story under two instruction conditions (cf. Holliday & Hayes, 2000). In the inclusion condition children reported whether they remembered details from either the video or the post-event summary phases. In the exclusion condition children were instructed to exclude post-event suggestions. Evidence of a misinformation effect was found in both age groups. Moreover, children were more likely to accept misled-generated items compared to misled-read items in the inclusion condition, but the opposite was the case under exclusion instructions. Process dissociation analyses (cf. Holliday & Hayes, 2000, 2001; Jacoby, 1991) revealed that both automatic and intentional processes influenced misinformation acceptance, but that suggestibility was predominantly due to automatic processes. These findings support a dual-process theory of misinformation acceptance in children (cf. Brainerd & Reyna, 1998). Competing models of the misinformation effect in children are evaluated in the light of these findings.

Holliday, R. E., & Hayes, B. K. (2000). Dissociating automatic and intentional processes in children's eyewitness memory. Journal of Experimental Child Psychology, 75, 1-42. Holliday, R. E., & Hayes, B. K. (2001). Automatic and intentional processes in children's eyewitness suggestibility. Cognitive Development, 16, 617-636. Jacoby, L. L. (1991). A process dissociation framework: Separating automatic from intentional uses of memory. Journal of Memory and Language, 30, 513-541.

Session 6: Tracing Processing through Fixations and Saccades (O6.1 - O6.4)

O6.1

Fixation positions on words in English sentences.

Sarah J. White and Simon P. Liversedge

University of Durham, UK

E-mail: s.j.white@durham.ac.uk

Two experiments investigated the effect of orthographic regularity on first fixation landing positions and refixations on words in English sentences. In Experiment One, the first or second letters of words were misspelled. The misspellings formed frequent or infrequent word initial trigrams. The critical words were embedded within the same sentence frame for all of the conditions. First fixation landing positions were significantly nearer the word beginning for all of the misspelled conditions compared to the correctly spelled condition. Also, the second of multiple fixations was more likely to be to the left of the initial landing position for two of the misspelled conditions compared to the correctly spelled condition. In Experiment Two all of the words were spelled correctly. The critical words began with either high or low type and token frequency bigrams and trigrams. The critical words were matched for word length and embedded within the same sentence frame up to the word after the critical word. For saccades launched from near launch sites, landing positions were significantly nearer the word beginning for words with infrequent initial letter sequences. The results will be discussed in relation to current models of eye movements in reading.

O6.2

Searching for information: Eye movements while attempting to verify statements about pictorial scenes

Geoffrey Underwood, Lorraine Bell and Kate Roberts

University of Nottingham, UK

E-mail: geoff.underwood@nottingham.ac.uk

Eye movements were recorded while 24 participants either searched for specific parts of a scene that would enable them to verify a previously read statement, or they first inspected the scene in preparation for the sentence verification task. The sentences made simple declarative statements about the events depicted in photographs of natural scenes. The order of presentation of picture and sentence was reflected in the fixation measures on both types of stimuli. When the sentence was read after the picture, average fixation durations and overall reading times were longer than when it was read before seeing the picture. The number of fixations, and overall inspection times on the picture were longest when they were seen before the sentence. These patterns reflect differences between general inspection and a specific search.

O6.3

Bottom-up and top-down control in visual search

Wieske van Zoest and Mieke Donk

Vrije Universiteit Amsterdam, The Netherlands

E-mail: w.van.zoest@psy.vu.nl

Previous research suggests that visual selection is controlled either in a bottom-up or in a top-down manner. However, possibly both manners of control simultaneously determine visual selection. To date, no studies have been performed to estimate the relative importance of top-down and bottom-up control in visual search. In the present study observers were presented with search displays consisting of an array of line segments that were rotated at various orientations. The task of observers was to indicate the presence of a vertical line segment (the target) presented amongst a series of nontargets and one distractor. Manual reaction times and eye movements were recorded. By varying the absolute differences in orientation between the target, nontargets and distractors, relative target-distractor saliency and target-distractor similarity were independently manipulated to investigate the relative contribution of bottom-up and top-down control. The major result was that relative target-distractor saliency and target-distractor similarity affected search performance independently.

O6.4

Saccadic responses indicate parallel set activation during task-switching

Charlotte Golding, Tim Hodgson, and Chris Kennard

Imperial College, London, UK

E-mail: c.golding@ic.ac.uk

When participants switch between tasks they reliably incur a time penalty (switch cost) which may reflect a discrete processing event of task-reconfiguration that is completed by the end of the first trial after a switch (e.g. Rogers & Monsell, 1995 and Goschke, 2000). However, we (Golding et al., 2001) reported evidence of switch costs persisting beyond the switch trial. Here we investigate further subtle effects task-switching has on saccade trajectories. In a computer-generated card sorting test participants matched a central card to one of 3 outer cards according to a pre-cued perceptual dimension (shape-matching or colour-matching task). Switching disrupted efficient search, causing an increase in: the initial saccade's latency; the number of saccades and the bias of the initial saccade towards the relevant distractor. All of these switch costs persisted beyond the switch trial. Moreover, in both the switch and subsequent trial, the first saccade was frequently hypometric and followed by a secondary saccade to the target after a brief (<100ms) fixation. These hypometric saccades had abnormally high peak velocities, suggesting that they were truncated; their execution having been interrupted by the prior programming of the second saccade (McPeek et al., 2000). Such parallel processing of saccades to targets for both tasks is inconsistent with the serial stage models of task-reconfiguration.

Session 7: Modeling Semantic Structure (O7.1 - O7.3)

O7.1

Verbal fluency from common and ad hoc categories: evidence from dementia

Arlene J. Astell¹ and Romola S. Bucks²

¹ University of St. Andrews, UK
² University of Southampton, UK

E-mail: aja3@st-and.ac.uk

Category fluency, generation strategy, generation stability, fluency following prompting and category judgement are examined in three studies with people with dementia of the Alzheimer-type (DAT). The same eight categories are used throughout, four common and four ad hoc. As previously reported, category fluency is reduced in DAT relative to controls. However, it is more impaired for common categories than ad hoc categories. There is also differential usage of generation strategies, with the DAT participants making more use of experiential strategies than semantic ones. In addition, strategy prompting is found to assist in the generation of additional items in DAT. Category judgement is largely intact in the DAT participants for both common and ad hoc categories, even for low frequency or uncommon items. These findings taken together suggest that item knowledge is relatively well preserved in DAT. As such, observed difficulties in retrieving item information may be attributable to impairments in the processes that act on stored information in the verbal fluency task.

O7.2

Conceptual structure in the normal system: Domain differences in the time course of activation of correlated and non-correlated semantic information.

Billi Randall, Helen Moss and Lorraine K. Tyler 

Centre for Speech and Language, Department of Experimental Psychology, University of Cambridge, UK

E-mail: bjr22@cam.ac.uk

We have recently developed an account of conceptual representation in which the structure of concepts differs systematically across domains and categories of knowledge in terms of the relative distinctiveness of semantic properties and the patterns of intercorrelation among them. Support for the Conceptual Structure Account has been derived from patterns of impaired and intact knowledge in patients with category-specific deficits and from neuroimaging studies (Tyler & Moss, 2001).

The current study explored the model's implications for the time course of activation of conceptual information in the normal system. We predicted that the activation of distinctive properties of living things (e.g. elephant – has a trunk) would be delayed relative to that of shared properties (e.g. elephant –has legs) due to their paucity of correlations with other features. This prediction was supported in a speeded (respond-to-deadline) feature verification task. In a second experiment, where presentation was slower and there was no response deadline, no such differences between living and non-living things were observed, consistent with the view that in the normal system correlational structure affects the initial activation of information rather than its eventual state.

L.K. Tyler & H.E. Moss (2001). Towards a distributed account of conceptual knowledge. Trends in Cognitive Sciences, 5.6, 244-252.

O7.3

On the Use of Scaling and Clustering in the Study of Semantic Disruptions in Patients with Alzheimer's Disease

Gert Storms¹, Trinette Dirikx¹, Jos Saerens^{2, 3}, Sonja Verstraeten^{2, 3} and Peter P. De Deyn ^{2, 3}

¹Department of Psychology, University of Leuven, Belgium
²Memory Clinic, General Hospital Middelheim, Antwerp, Belgium
³Laboratory of Neurochemistry and Behavior, Born Bunge Foundation, University of Antwerp, Belgium

E-mail: Gert.Storms@psy.kuleuven.ac.be

In the past decade, several studies have used scaling and clustering techniques to document semantic deficits in patients with Alzheimer's disease. In this presentation we argue that many of the conclusions drawn from these studies are unjustified by the data. We review the methodology used in these studies and present data from simulation studies to further investigate the validity of their conclusions. We also present empirical data from patients and normal control subjects to demonstrate that analyses of the patients' proximity data do not provide unambiguous evidence for a generalized semantic deficit.

Session 8: Phonology in Speech and Reading (O8.1 - O8.3)

O8.1

In pursuit of the Syllabary: this Snark is a Boojum!

Stephen Monsell, Arie van der Lugt and Patricia Jessiman

University of Exeter, UK

E-mail: S.Monsell@exeter.ac.uk

Some speech production theorists, notably Levelt, Roelofs and Meyer (1999), have proposed that during the translation of abstract phonological sequences into articulatory motor activity, the speaker retrieves articulatory "gestural scores" for familiar syllables from a "syllabary". That the speech-motor system learns and retrieves frequent patterns rather than recomputes them is plausible and would solve part of the coarticulation problem. But evidence for the existence of a syllabary has been in short supply. We report three attempts to demonstrate the contribution of the syllabary to speech production. We failed to find more rapid picture naming for words whose constituent syllables had just been primed by a production in another word. We failed to find any difference in the rapid production of prepared utterances consisting of matched nonword sequences composed of frequent syllables versus syllables that do not occur in English. We found only a tiny difference in the latency for naming visually presented syllables of these two kinds, matched for orthographic difficulty. Either there is no syllabary, or the advantage it confers in production is very small.

Levelt W.M., Roelofs, A., Meyer, A.S. (1999) A theory of lexical access in speech production. Behavioral and Brain Sciences, 22, 1-75.

O8.2

The role of the rime. A cross-linguistic comparison of rime effects in reading English and Dutch

Dominiek Sandra¹, James Booth², Heike Martensen¹, Astrid Geudens¹ and Charles Perfetti³

¹ University of Antwerp, Centre for Psycholinguistics, Belgium
² Northwestern University, USA
³ University of Pittsburgh, USA

E-mail: sandra@uia.ac.be

Orthographies differ in the consistency of their grapheme-to-phoneme mappings. GP mappings are rather inconsistent in English (beat, bear) but quite consistent in Dutch. In English, consistency between orthography and phonology increases considerably at the rime level (fight, light). Such cross-linguistic differences might explain why rime effects were not found in Dutch (Geudens & Sandra, 1999; Martensen, Loncke & Sandra, in prep.). However, Booth & Perfetti's (2001) results suggest that even in English, rime effects do not reveal early recognition units in lexical processing (equal effects of rimes and non-rimes in brief identication) but depend on phonological assembly for reading aloud (effect in naming). We investigated this issue cross-linguistically, by comparing a language whose orthography-to-phonology mappings are quite reliable at the grapheme level (Dutch) to one in which such reliability emerges at the rime level (English). A masked priming paradigm was used to study the effect of VC primes on CVC targets, using brief identification and naming tasks (frequency-matching across languages). We obtained equal priming effects for Dutch and English in brief identification and no effects in naming. In Dutch, but not in English, a correlation between CV frequency and naming times was found. Follow-up experiments with CV primes are planned.

O8.3

Lexical Bias in Spoonerisms: a 'related beply' to Baars, Motley, and MacKay (1975)

Robert J. Hartsuiker¹, Martin Corley¹, and Heike Martensen²

¹ University of Edinburgh, UK
² University of Antwerp, Belgium

Phonological speech errors occur more often when they form real words than nonwords. This 'lexical bias' effect has been taken as evidence for phoneme-to-word form feedback in word production, but also as evidence for pre-articulatory editing. The latter account seemed to be supported by the finding that the effect is modified by context (Baars et al., 1975). We further explored such context effects in order to test the editing account. We used a speeded naming task of nonword pairs ("pood gath") to elicit spoonerisms (responding "good path"). Critical items were all nonwords and either resulted in words when spoonerized (pood gath) or not (pooth gad). The critical items were either embedded in mixed lists of words and nonwords or in a pure nonword context. In the mixed condition, nonword pairs with lexical outcome (pood gath) were spoonerized more often than nonword pairs with nonlexical outcome (pooth gad). This lexical bias effect disappeared in the pure nonword condition. Inconsistent with an editing account, we observed that lexical context increased the number of word outcomes, rather than decreased the number of nonword outcomes. We tentatively propose the alternative explanation that context influences the probability to misread nonword targets as real words.

Session 9: Object Discrimination (O9.1 - O2.3)

O9.1

Context-dependent asymmetries in stimulus comparisons by monkeys

Hans Op de Beeck^{1, 2}, Johan Wagemans², and Rufin Vogels¹

¹ Laboratory of Neuro- en Psychofysiology, University of Leuven, Belgium
² Laboratory of Experimental Psychology, University of Leuven, Belgium

E-mail: hans.opdebeeck@med.kuleuven.ac.be

Similarity is a core concept in theories of object recognition, categorization, and reasoning. It is often conceptualized as a geometric distance in a multidimensional stimulus space. However, research in humans has revealed that similarity judgments involve more than a simple distance calculation (e.g., Tversky, 1977, Psychological Review). Contrary to the central asumptions in geometric models, similarity judgments are asymmetric when stimuli differ in their prototypicality or salience. For example, most people judge Mexico to be more similar to USA than USA to Mexico. Here we provide the first evidence for marked asymmetries in nonhuman species. We trained two rhesus monkeys in a temporal same/different task with a parametric control of stimulus prototypicality. The monkeys made most errors ('same' responses) when the first stimulus in a trial was less prototypical than the second one. This effect of prototypicality became remarkably fine-tuned as the monkeys acquired more experience with the stimulus set: In the beginning of the experiment the asymnmetries were sensitive to prototypicality in the context of the total set of stimuli, whereas the prototypicality for more local regions in stimulus space mattered in later sessions. We conclude that asymmetric similarity judgments also occur in non-verbal animals, and that they do so with a high sensitivity for stimulus statistics.

O9.2

Advantage of low visual acuity of the retina in 2-month-olds infants: a statistical categorization model

Martial Mermillod¹, Robert M. French¹, Paul C. Quinn², and Denis Mareschal³

¹ University of Liège, Belgium
² Washington & Jefferson College, USA
³ Birkbeck College, UK

E-mail: mmermillod@ulg.ac.be

From the earliest moments of postnatal development, perceptual maturation is believed to place a functional constraint on the ability to recognize or categorize stimuli from the environment. Using a computer simulation of retino-cortical development, with a bank of Gabor wavelets providing the output of the V1 complex cells, we showed that reducing the range of spatial frequencies from the retinal map allows better categorization performances for a statistical model. The simulations provide a pattern of statistical data showing that a statistical categorization model may be more efficient with low visual acuity retina. Moreover, we have found that this effect interacts with the type of category in such a way that super-ordinate level categories are enhanced by the simulated low visual acuity of 2-month-old infants whereas basic level categories are less or equally categorized with the low visual acuity. These data are convergent with empirical data showing that 2-month-olds infants acquired earlier and more easily super-ordinate level categories than basic level categories (Quinn & Johnson, 2000).

O9.3

Interactions between view changes and shape changes in picture-picture matching

Rebecca Lawson¹, Heinrich H. Bülthoff² and Sarah Dumbell¹

¹ University of Liverpool, UK
² Max Planck Institut, Tübingen, Germany

E-mail: R.Lawson@liverpool.ac.uk

Four studies presented pictures of different morphs of novel, complex, three-dimensional objects, similar to objects which we must identify in the real world. We investigated how viewpoint changes influence our ability to discriminate between morphs. View changes had a powerful effect on performance in picture-picture matching tasks when similarly shaped morphs had to be discriminated. Shape changes were detected faster and more accurately when morphs were depicted from the same rather than different views. In contrast, view change had no effect when dissimilarly shaped morphs had to be discriminated. This interaction between the effects of view change and shape change was found for both simultaneous stimulus presentation and for sequential presentation with interstimulus intervals of up to 3600ms. The interaction was found following repeated presentations of the stimuli prior to the matching task and following practice at the matching task as well as after no such pre-exposure to the stimuli or to the task. The results demonstrate the importance of view changes relative to other task manipulations in modulating the shape discrimination abilities of the human visual recognition system.

Session 10: Codes in Memory (O10.1 - O10.2)

O10.1

Semantic coding in the recall of picture and word lists shown at RSVP and STM rates

Veronika Coltheart, Stephen Mondy and Robyn Langdon

Macquarie University, Sydney, Australia

E-mail: veronika@maccs.mq.edu.au

Short-term memory serial recall provides evidence for the use of phonological codes with little, if any, involvement of semantic coding. However, when sequences of words or pictures are visually presented at high rates (8-12 items/second, termed RSVP - rapid serial visual presentation) there is evidence of early extraction of conceptual and semantic information. Such results led Potter (1993, 1999) to propose a model of Conceptual Short-term Memory. We report experiments which investigated the influence of thematic relationships among items in picture and word list recall when items were shown at RSVP rates (8/second) and low STM rates (1/second). The list items were selected from a large pool, and the nature of the semantic relationship among list items varied from list to list. Semantic relatedness benefited item recall, even when items were phonologically similar, at high and low rates of presentation. This is evidence for early involvement of semantic codes as hypothesized by the theory of Conceptual STM and it also indicates a role for semantic coding in STM.

O10.2

Interactions between long- and short-term phonological memory: Evidence from an implicit phonological learning paradigm

Steve Majerus¹, Martial Van der Linden², Thierry Meulemans³, Ludivine Mulder³

¹ Research Fellow-FNRS, Neuropsychology Unit, University of Liège, Belgium
² Cognitive Psychopathology Unit, University of Geneva, Switzerland
³ Neuropsychology Unit, University of Liège, Belgium

E-mail: smajerus@ulg.ac.be

Verbal short-term memory (STM) is influenced by phonological and lexico-semantic knowledge stored in long-term memory (LTM), as evidenced by better STM for words than nonwords. This knowledge is highly over-learned and acquired early during development. In this study we show that STM can also be supported by very recently and implicitly acquired phonological information. In experiment 1, 20 adult and 20 eight-year-old participants passively listened to a thirty-minute sequence of CV syllables while they were performing a picture drawing task. The succession of the consonants and vowels and the CV syllables was governed by an artificial grammar, which the participants were unaware of. After this implicit phonological learning phase, a nonword repetition task was presented comprising nonwords of increasing length, which were either legal or illegal relative to the artificial grammar. STM performance was significantly better for legal than for illegal nonwords, in adults and children. In experiment 2, STM performance was similar for legal and illegal nonword lists when not preceded by a phonological learning phase. Our results suggest that STM can be supported by recently and implicitly acquired phonological information and directly reflects subtle changes in phonological LTM. Implications for current models of STM and LTM are discussed.

Session 11: Shape, Space, and Language (O11.1 - O11.2)

O11.1

The cultural relativity of shape categories: Goodness had nothing to do with it.

Debi Roberson¹, Jules Davidoff² and Laura Shapiro²

¹ University of Essex, UK
² Goldsmiths College, University of London, UK

E-mail: robedd@essex.ac.uk

The Gestalt theorists of the early twentieth century proposed a psychological primacy for circles, squares and triangles over other shapes. They described them as ‘good' shapes and the Gestalt premise has been widely accepted. Rosch (1973), for example, found that speakers of a language lacking terms for any geometric shape nevertheless learnt paired-associates to these ‘good' shapes more easily than to asymmetric variants. A cross-cultural investigation sought to replicate Rosch's findings with the Himba of Northern Namibia who also have no terms in their language for the supposedly basic shapes of circle, square and triangle. We found no advantage for these ‘good' shapes in the organization of categories. It was concluded that there is no necessary salience for circles, squares and triangles. Indeed, we argue for the opposite because the general absence of straight lines and symmetry in the perceptual environment should rather make circles, squares and triangles unusual and, therefore, less likely to be used as prototypes in categorization tasks. We place shape as one of the types of perceptual input, like colour, that is readily susceptible to effects of language variation in categorisation tasks.

O11.2

Spatial discourse: Processing, memorizing, and using navigational instructions

Michel Denis

Groupe Cognition Humaine, LIMSI-CNRS, Université de Paris-Sud, Orsay, France

E-mail: Michel.Denis@limsi.fr

Spatial discourse takes place in a variety of environmental contexts. The value of analyzing the cognitive processes involved in the processing of this type of discourse is to reflect the mind's capacity to translate verbal information into visuo-spatial knowledge, and reciprocally its capacity to express non-linguistic spatial representations in a linguistic format. The studies that we have conducted in the past years have focused on the processing, memory, and use of navigational discourse, in both laboratory and field settings. These studies provide information on the cognitive resources called upon during the processing of the descriptive and the procedural parts of spatial discourse. Because of the visuo-spatial nature of navigational information, the hypothesis that individual imagery capacities affect the processing of this type of discourse has been examined and has received substantial empirical support.

Session 12: Social Attitudes & Beliefs (O12.1 - O12.2)

O12.1

A Recurrent Connectionist Model of Attitude Formation

Frank Van Overwalle

Vrije Universiteit Brussel, Belgium

E-mail: Frank.VanOverwalle@vub.ac.be

This presentation discusses a recurrent connectionist network of current dual-process approaches of attitudes, focusing on the common processing principles that may underlie both central and peripheral processes. Major findings in attitude formation and change involving both forms of processes are reviewed and modeled from a connectionist perspective. The majority of these processes are illustrated with well-known experiments, and simulated with an auto-associative network architecture with linear activation update and delta learning algorithm for adjusting the connection weights. All of the phenomena considered were successfully reproduced in the simulations. Moreover, the proposed model is shown to be consistent with algebraic models of attitude formation (Fisbein & Ajzen, 1975). The discussion centers on how the particular simulation specifications may be used to develop novel hypotheses for testing the connectionist modeling approach and, more generally, for improving and unifying theorizing in the field of attitudes as well as social cognition in general.

O12.2

Openness to Experience and Boundaries in the mind

Alain Van Hiel and Ivan Mervielde

Ghent University, Belgium

E-mail: alain.vanhiel@rug.ac.be

The present research investigates whether Openness to Experience and Boundaries in the mind are related to conservatism. In the first study, a positive relationship between several scales of the Boundaries in the mind questionnaire and indicators of conservative beliefs were obtained in an adult sample (N = 78) as well as in a sample of political party activists (N = 44). In the second study, these relationships were replicated in an adult sample (N = 225). Moreover, two dimensions representing Boundaries in the mind were identified, one positively related to Openness to Experience and negatively related to conservatism, and the second showing high positive correlations with Neuroticism. The exceptional strong correlations between some Boundaries in the mind facet scales and conservatism are discussed, as well as the relationships between Openness to Experience and cultural and economic conservatism.

Session 13: Development of Concepts & Percepts (O13.1 - O13.2)

O13.1

Cognitive factors and concept learning: A developmental approach.

Sabine Gelaes and Jean-Pierre Thibaut

Université de Liège, Belgium

E-mail: sgelaes@ulg.ac.be

Many experiments have shown that in concept learning, children reach classification criterion faster when the criterion correspond to their preferred dimension. In this experiment, we tried to replicate this phenomenon with 4, 6 and 8 year-old children and to link their categorization performance with different cognitive factors (attention, memory, cognitive flexibility, cognitive inhibition and planning). First, we assessed the hierarchy of dimension preference for two sets of stimuli composed of 3 dimensions (colour, number and shape for the first set and texture, number and shape for the second set). Then, children participated in two concept learning tasks, one in which their preferred dimension was the relevant feature and one in which their least preferred dimension was relevant. Contrary to previous experiments, there was no difference between the two concept learning tasks. For the 4 year-old children, memory factor was associated with performance. Cognitive inhibition factor was the factor correlated with performance for the 6 year-old children. For the 8 year-old children, there was only an association between performance and a cognitive factor, in this case planning factor, in the condition of learning on the least preferred dimension.

O13.2

Perception of Partial Collision Events in Infancy

Peter Dejonckheere¹, Ad Smitsman², and Leni Verhofstadt-Denève¹

¹ Ghent University, Belgium
² University of Nijmegen, The Netherlands

E-mail: peter.dejonckheere@rug.ac.be

Studies about object perception have revealed fundamental insights in the mechanical interactions between objects in infants. Several studies showed a developing understanding in infants of the way in which relations between object properties constrain the course and outcome of interactions between those objects (o.a. Aguiar & Baillargeon, 1999 ; Aguiar & Baillargeon, 1998 ; Sitskoorn & Smitsman, 1995). In this study we first questioned whether infants of 6, 9 and 12 months old perceive whether the rims of a box would provide passing through. Infants were habituated to block movements approaching a box lied in different positions, all providing enough room to pass through. In a subsequent test phase, identical block approaches were presented with the exception that the rims of the box were widened. Consequently, two collision actions and one control action (non collision) emerge. The results showed that infants of 9 and 12 months old perceived whether the width relation between block and box specified passing through. In addition we found that both the age of the infant and the amount of overlap between the block and the opening of the box played a significant role in the perception of these passing through interactions.

Session 14: Strategy, Working Memory Capacity, and Inhibitory Control (O14.1 - O14.6)

O14.1

What makes an insight problem? An investigation into the role of strategy-goal congruency.

¹ Lancaster Unversity, UK
² School of Public Administration, University of Victoria, Canada

E-mail: E.Chronicle@lancaster.ac.uk

We have previously demonstrated that part of the difficulty of some insight problems (such as the nine-dot problem) stems from the adoption of rational local strategies that lead to impasse. Here, we explore the problem factors that influence strategy formation. A single experiment is presented, using as its basis the six-coin problem, for which we have independent evidence of the use of certain operators. Four variants of the problem were presented to subjects, with four different goal conditions: (i) abstract, in which the goal was specified only verbally, (ii) concrete-congruent, in which the goal was visually presented, and application of the operator would not result in impasse, (iii) concrete-incongruent, in which application of the operator would likely result in impasse, and (iv) both concrete goals together. Participants in conditions (iii) and (iv) solved the problem significantly more often than those in condition (i). Furthermore, there was a marked preference in all participants for concrete-congruent goals. These data support the notion that goal-directed strategy formation may - somewhat paradoxically - underlie impasse in certain classes of problem.

O14.2

Strategy use in relation to mathematical ability: Developmental delay or deficit?

Joke Torbeyns¹, Lieven Verschaffel², and Pol Ghesquière²

¹ K.U. Leuven (Research Assistant FWO - Vlaanderen), Belgium
² K.U. Leuven, Belgium

E-mail: joke.torbeyns@ped.kuleuven.ac.be

In this study we investigated the variability, frequency, efficiency, and adaptiveness of young children's strategies in the domain of early arithmetic, using the choice/no-choice method (Lemaire & Siegler, 1995; Siegler & Lemaire, 1997). Seventy-seven beginning second- and third-graders, divided in three groups according to age and mathematical ability (MA), solved a series of 36 simple additions and subtractions up to 20. The first group consisted of 26 second-graders with strong MA, the second of 25 second-graders with weak MA, and the third of 26 third-graders with weak MA. All children solved the series of problems in four conditions. In the first condition, children could choose between retrieval, decomposition up to 10 (e.g., 8 + 3 = 8 + 2 + 1 = 10 + 1 = 11), and counting (e.g., 8 + 3 = (8), 9, 10, 11) to solve each problem. In the second, third, and fourth condition, children had to solve all problems with one particular strategy, respectively retrieval, decomposition up to 10, and counting. The results demonstrate that second-graders with weak MA use the same (retrieval, decomposition, and counting) strategies as their peers with strong MA, but differ in the frequency, efficiency, and adaptiveness with which they apply these strategies. Furthermore, our results indicate that the development of children with weak MA is, compared to normally progressing children, marked by a developmental delay: Third-graders with weak MA use the same strategies as second-graders with strong MA, and execute these strategies equally frequently, equally efficiently, and also equally adaptively. Finally, this study illustrates the value of Siegler's theoretical and methodological framework to describe young children's strategy use in the domain of early arithmetic.

Lemaire, P., & Siegler, R.S. (1995). Four aspects of strategic change: Contributions to children's learning of multiplication. Journal of Experimental Psychology: General, 124, 83-97. Siegler, R.S., & Lemaire, P. (1997). Older and younger adults' strategy choices in multiplication: Testing predictions of ASCM using the choice/no-choice method. Journal of Experimental Psychology: General, 126, 71-92.

O14.3

The role of working memory in the carry operation of mental arithmetic

Stijn De Rammelaere and Ineke Imbo

Department of Experimental Psychology, Ghent University, Belgium

E-mail: Stijn.DeRammelaere@rug.ac.be

We investigated the role of working memory in the carry operation of mental arithmetic. The value of the carry and the number of carries were manipulated, so that there were problems with 1 or 2 carries (= number of carries) of which the value could be 1 or 2 (e.g., in 183 + 122 + 601 = 906 there is one carry with a value of 1; in 249 + 388 + 269 = 906 there are two carries with a value of 2). These problems were solved in a control condition and in experimental conditions in which the phonological loop and the central executive were loaded. We found that the number of carries determines the difficulty of a problem and that executive processes contribute to this effect. An important new finding is that also the value of the carry is an important variable and that the phonological loop is responsible for maintaining the value of the carry. This leads us to the straightforward conclusion that the central executive is responsible for the number of carries, whereas the phonological loop handles the value of the carry.

O14.4

Working memory capacity in causal reasoning: blocking background knowledge

Wim De Neys, Walter Schaeken, and Géry d'Ydewalle

KULeuven, Belgium

E-mail: Wim.Deneys@psy.kuleuven.ac.be

We present two of the first experiments that examine the role of working memory in reasoning with realistic, causal conditionals. A Dutch adaptation of the OSPAN (La Pointe & Engle, 1990) test was used to measure working memory span. A first experiment (n=66) established that high span participants (top quartile of first-year psychology students) were better at retrieving disabling conditions from semantic memory than low span participants (bottom quartile). Retrieving a disabling condition is known to suppress the valid Modus Ponens (MP) and Modus Tollens (MT) inferences (Cummins, 1995). In Experiment 2 (n=60) we presented participants a conditional reasoning task where the number of possible disabling conditions of the conditionals was manipulated. Participants did not receive explicit deductive reasoning instructions. Results showed that, despite the better disabler retrieval, high span participants rejected MP and MT less than low spans. However, inferences in both span groups were affected by the conditionals' number of possible disablers. These data show that both high and low span reasoners use their background knowledge when drawing conditional inferences. However, when this background knowledge conflicts with logic, high span reasoners manage to block this impact more efficiently. Implications for current reasoning theories and the debate on human rationality are discussed.

La Pointe, L. B., & Engle, R. W. (1990). Simple and complex word spans as measures of working memory capacity. Journal of Experimental Psychology: Learning, Memory, and Cognition, 16, 1118-1133. Cummins, D. D. (1995). Naive theories and causal deduction. Memory and Cognition, 19, 274-282.

O14.5

How to know what to inhibit: Action-effects are used in response-suppression

Bernie Caessens¹, Karen Mortier², and André Vandierendonck¹

¹ Dept. of Experimental Psychology, Ghent University, Belgium
² Cognitive Psychology Section, Vrije Universiteit Amsterdam, The Netherlands

E-mail: bernie.caessens@rug.ac.be

Although many experiments have investigated the time-course of response inhibition mechanisms, few have been directed towards identifying the representations involved in the inhibition process. In the present study we started from the idea that effect codes are the cornerstones of both preparatory and inhibitory activity. This was investigated by means of a series of dual-task experiments in which subjects selectively inhibited a signalled manual response (Task 1) and quickly prepared and executed a congruent or incongruent response in a different modality (Task 2). The results showed that after successful inhibition, congruent responses were executed more slowly then incongruent responses. The consequences for behavioural inhibition are derived and discussed.

O14.6

Aging effects in inhibitory control over no-longer relevant information during a garden-path sentence task

Valentine Charlot and Pierre Feyereisen

Unit Cognition and Development, UCL, Belgium

E-mail: Charlot@neco.ucl.ac.be

The goal of the present work was to analyse the assumption according to which cognitive ageing is associated with inefficient inhibitory mechanisms (Hasher and Zacks, 1988). In contrast, Burke (1997) assumed that a decline of memory for context could better explain the age differences. We used a modified version of the Hartman and Hasher (1991) task. Two experiments were conducted, differing by the kind of indirect memory test used. In the first stage, older and younger adults read a series of sentences whose final word was a member of semantic different categories (e.g.: flowers). Then, subjects received the same list of sentences, each missing its final word, and they had to complete them by re-using the previous words (inclusion), or by a new word (exclusion). Finally, two indirect memory tests - perceptive identification (exp. 1 & 2) and category exemplars generation (exp. 2) - and a direct memory - recognition (exp. 1) or free recall (exp. 2) - were used to examine the view of impaired inhibitory mechanisms in ageing: do older adults show equal priming of inhibited (exclusion) and facilitated (inclusion) words? The findings of the first experiment are consistent with an inhibitory framework. The second experiment is being analysed.