Relational Versus Attributional Mode of Problem Solving

Relational Versus Attributional Mode of Problem Solving? Svetoslav Bliznashki (valsotevs@gmail.com) New Bulgarian University, 2 Montevideo Street Sofia 1618, Bulgaria Boicho Kokinov (bkokinov@nbu.bg) New Bulgarian University, 2 Montevideo Street Sofia 1618, Bulgaria Abstract We argue that the concept of relational priming (e.g. Schunn 1996, Day 2007) can be extended from priming of specific relations to generating a cognitive state during which subjects are particularly likely to encode and use relations. We conducted an experiment in which three groups of subjects did different tasks before a target matching to sample task was introduced which contrasted a relationally versus an attributionally similar alternative. Subjects in one condition were asked to solve tasks involving relational reasoning while subjects in another condition were asked to tasks involving only attributes. As expected subjects in the first condition were more likely to pick up the relationally similar alternative while in the second condition the results reversed relative to a control group. In conclusion we argue that this study shows that encoding of relations can be a subject to unconscious context influence. Keywords: relational priming; context dependence; encoding of relations, cognitive state Introduction Since Gentner’s Structure Mapping theory (Gentner 1983) a great deal of research has been concentrated upon analogical reasoning in terms of mapping of higher-order relations. Although the mechanisms employed by the mapping process have been extensively studied little is currently known about the nature of the processes involved in relational encoding. This study attempts to scratch the surface of this complex matter by asking the question of whether the process of relational encoding is subject to certain external and internal context influences. While the answer to this question certainly would not reveal the nature of the encoding process it would hopefully tell us something about certain specific aspects of its functioning. Currently there is some agreement that the phenomenon of relational priming exhibits a somehow automatic (i.e. not subjected to voluntary control, external influence and conscious experience) nature (e.g. Kokinov 1996, Schunn 1996, Day 2007, Hristova 2009) 1 . The abovementioned studies employ different methodologies ranging from naturalistic-like settings (Schunn 1996) to Stroop-like interference Reaction Time paradigms (Hristova 2009). But see Spellman et al. (2001) whose results indicate that relational priming took place only when participants were explicitly instructed to pay attention to the relations existing between the stimuli (words) involved in the studies. There is however a common thread among these research projects – the use of specific relations. In other words all these (and other) studies concentrated on exploring the relational priming by using concretely represented, nameable relations. The same naturally holds true for relational priming in psycholinguistic research (e.g. Gagne 2005, Estes 2006). Instead of continuing this well established line of research we concentrated on the question of whether a global cognitive state can be induced in which people are more likely to encode relations in general. It can be said that we are still concerned with relational priming but we employ a rather broad, holistic and abstract definition of the phenomenon. We hypothesized that subjects confronted with tasks explicitly involving relational reasoning will subsequently be more likely to continue this style of reasoning when dealing with completely different tasks. We also hypothesized that subjects forced to encode and use attributes of objects will be considerably less likely to encode (and use) relations in subsequent tasks. In other words we argue that not only specific relations can be primed with similar other relations but also a “relational mode of thinking” can be induced by use of specific task requirements. Thus we claim that relational priming (and consequently “attributional priming”) is a much more complicated and abstract phenomenon than currently conceived by traditional research in the area. Another hypothesis related to the current study concerns the subjects’ ability to cope with the particular priming task. Since we argue that task requirements can possibly induce a particular cognitive state it follows that the degree to which this actually happens should depend upon a subject’s particular ability to successfully cope with the task at hand (the priming task). Experiment In the current experiment we tested three different groups of subjects in order to see whether prior tasks influence significantly relational and attributional reasoning during a target task. In the first condition subjects solved six different mental rotation tasks, in the second condition subjects solved six items from Raven Progressive Matrices test (e.g. Raven 2003) and in a third condition no task preceded the target task. The three groups are called attributional, relational and control conditions respectively.