Reasoning as Memory.
Saturday, October 01st, 2011 [14:20 - 16:00]
SY_19. Reasoning as Memory
Thompson, V. 1 & Feeney, A. 2
1 University of Saskatchewan, Saskatchewan, Canada
2 Queen's University, Belfast, UK
As little as twenty years ago, reasoning researchers presented participants with unfamiliar tasks using artificial content, which were designed to minimize the role of knowledge, memory, and beliefs in reasoning and decision making. Today, the pendulum has swung the other way- reasoning research is being enriched by, and becoming evermore coherent with mainstream cognitive psychological theorising. The aim of this symposium is to showcase the ways in which theories of reasoning are informed by basic memory processes. For example, there is a growing body of research addressing the relationship between recognition memory and the processes engaged to perform a reasoning task. This symposium brings together three of the participants in a recent debate on whether inferences can be made about developmental changes in reasoning on the basis of recognition memory data (Sloutsky, Feeney, and Hayes). In this symposium, Sloutsky will present eye tracking data that is directly relevant to the issues raised in that debate, Hayes will consider whether recognition memory and inductive reasoning processes may be modelled in the same terms, and Feeney will attempt to arbitrate between accounts of adult inductive reasoning in part on the basis of recognition memory data. To illustrate the importance of associative processes in reasoning, Markovits and colleagues will consider processes in semantic memory and how they relate to predictions about deductive and statistical reasoning ability. Inspired by the literature on metacognition and memory, Thompson will show how frameworks such as dual process theories of thinking may be informed by consideration of metacognitive processes. We have chosen the participants so that a variety of types of thinking and types of memory process will be discussed. Although participants have not been chosen on the basis of their nationality, we consider the international nature of the symposium to be one of its strengths.
TALKS
SY_19.1 - The development of induction: Insights from recognition memory
Sloutsky, V.
Ohio State University, USA
Inductive generalization is a critical aspect of cognition because it allows using knowledge creatively in new situations. For example, upon learning that a cat is a carnivore, one would expect other cats to be carnivores too. Although many aspects of inductive generalization are present in non-human animals and exhibit early onset in humans, mechanisms of induction in humans and its developmental course are hotly debated. To examine mechanisms of induction across development, Sloutsky and Fisher developed the Induction-then-Recognition (ITR) task (Fisher & Sloutsky, 2005; Sloutsky & Fisher, 2004a, 2004b). Using ITR, Sloutsky and Fisher have argued that inductive inference undergoes a radical developmental transformation. According to this view young children base their induction on computation of perceptual similarity (if one cat is a carnivore, another cat should also be a carnivore because both cats look alike). In contrast adults, base their induction on determining if the two items belong to the same category (if one cat is a carnivore, another cat should also be a carnivore because both are members of the category ‘cat”). Some participants of the present symposium have challenged these conclusions. In particular, Hayes, McKinnon, & Sweller (2008) and Wilburn & Feeney (2008) presented arguments suggesting that developmental differences between early and mature induction could be rather modest, with both groups basing their induction on the common category. They further argued that ITR might have revealed developmental differences in visual attention rather than in induction. In this talk, I will present an overview of published research using ITR. I will then present new experiments in which 5- to 6-year-olds and adults performed ITR task, while their patterns of eye movements were recorded by an eye tracker. The patterns of eye movements support the idea of different mechanisms of induction in children and adults. At the same time, these findings are difficult to explain only by the development of visual attention.
SY_19.2 - Common and distinctive mechanisms in reasoning and memory
Hayes, B. 1 & Heit, E. 2
1 University of New South Wales, Australia
2 University of California, Merced, USA
This presentation reviews recent findings from a research program examining relations between inductive reasoning and recognition memory. The first experimental series [Heit, E., & Hayes, B. K. (2011). Predicting reasoning from memory. Journal of Experimental Psychology: General, 140, 76-101] examined how well responses on an inductive reasoning task predicted responses on a recognition memory task. A common stimulus set (e.g., pictures of dogs) was presented for study followed by a test set containing old and new members of the same category. In the recognition condition people were instructed to memorize study instances and respond “yes” at test to old items. In the induction condition people were instructed that study items contained a novel property (“beta-cells”) and to respond “yes” to test items that had this property. Across several experimental manipulations such as varying study time, presentation frequency, and the presence of stimuli from other categories, there was a high correlation between reasoning and memory responses (average r = .87). A second series extended the study of memory-reasoning relations to more complex forms of induction involving meaningful properties. In this case study items were members of a conjunctive category (aquatic mammals) and different inductive properties primed attention to either the aquatic dimension or the mammalian dimension. The relation between reasoning and memory responses was reduced when meaningful properties were used in induction; especially when additional time was given for test decisions. These results point to a common mechanism of exemplar-similarity underlying inductive reasoning and recognition memory, although induction with meaningful properties involves additional rule-based processes. A mathematical model, GEN-EX (GENeralization from EXamples), derived from exemplar models of categorization, is presented, that predicts both reasoning and memory responses from pairwise similarities among the stimuli, allowing for additional influences of deterministic responding.
SY_19.3 - Working memory, recognition memory and inductive reasoning
Feeney, A. & Leckey, M.
Queen's University, Belfast, UK
Category-based inductive reasoning involves the generalisation of properties from a sample consisting of members of one or more premise categories to members of a conclusion category. People are known to be sensitive to the diversity and size of the set of premise categories so that more diverse and larger samples result in stronger ratings of argument strength. Associative accounts of category-based induction explain both of these sensitivities in terms of a single parallel process that is fast and independent of working memory. Other accounts suggest that there may be multiple resource demanding processes in play. Feeney (2007) showed that sensitivity to both diversity and sample size is correlated with IQ, a finding that is problematic for associative accounts. Here we describe a study that directly investigated the role of working memory in sensitivity to diversity and sample size by placing some participants under cognitive load whilst they reasoned about pictorial premises. Because what people can remember about the contents of reasoning problems may have much to tell us about how they reason (see Sloutsky & Fisher, 2004) we employed an Induction-then-Recognition (ITR) paradigm where participants completed an unexpected recognition memory test after the reasoning task. Sensitivity to sample diversity but not to sample size was significantly affected by a secondary load. Sensitivity to sample size only was associated with faster reasoning times and with poorer recognition memory. These results suggest that sensitivity to sample size involves the application of a simple rule, whereas sensitivity to diversity requires more complicated similarity calculation. Thus, different inductive reasoning phenomena may require different explanations. We will discuss the implications of our results for dual process theories reasoning which predict, contrary to our findings, that correlations with IQ will often be accompanied by effects of a secondary load.
SY_19.4 - Reasoning, possibilities, and semantic memory
Markovits, H. , Forgues, L. & Brunet, M.
University of Quebec at Montreal, Montreal, Canada
Logical reasoning is implicitly taken to suppose use of algorithmic procedures that rely only on relatively abstract semantics. However, both Piaget (1987) and more recently, Byrne (2005), have pointed out that understanding logical reasoning cannot be done without taking into account people's ability to generate problem representations that involve cognitive possibilities. In the following, we first review evidence that in many cases, these possibilities require direct access to semantic memory traces. We then examine the idea that more flexible retrieval might characterize more logical reasoners using examples from both conditional reasoning and the Linda problem.
SY_19.5 - Intuition vs Reasoning: The Role of Metacognitive Experiences in Controlling Analytic Thinking
Thompson, V.
University of Saskatchewan, Saskatchewan, Canada
As little as twenty years ago, reasoning researchers presented participants with unfamiliar tasks using artificial content, which were designed to minimize the role of knowledge, memory, and beliefs in reasoning and decision making. Today, the pendulum has swung the other way- reasoning research is being enriched by, and becoming evermore coherent with mainstream cognitive psychological theorising. The aim of this symposium is to showcase the ways in which theories of reasoning are informed by basic memory processes. For example, there is a growing body of research addressing the relationship between recognition memory and the processes engaged to perform a reasoning task. This symposium brings together three of the participants in a recent debate on whether inferences can be made about developmental
changes in reasoning on the basis of recognition memory data (Sloutsky, Feeney, and Hayes). In this symposium, Sloutsky will present eye tracking data that is directly relevant to the issues raised in that debate, Hayes will consider whether recognition memory and inductive reasoning processes may be modelled in the same terms, and Feeney will attempt to arbitrate between accounts of adult inductive reasoning in part on the basis of recognition memory data. To illustrate the importance of associative processes in reasoning, Markovits and colleagues will discuss retrieval from semantic memory and how it relates to predictions about deductive and statistical reasoning ability. Inspired by the literature on metacognition and memory, Thompson will show how frameworks such as dual process theories of thinking may be informed by consideration of metacognitive processes. Collectively, these papers illustrate the “new paradigm” in reasoning that seeks to move away from reasoning research as an isolated enterprise to one that situates reasoning within the broader context of cognitive psychology.