The role of action in perceiving and comparing functional relations - eScholarship
Ivan Vankov, Boicho Kokinov · Proceedings of the Annual Meeting of the Cognitive Science Society · 2010
The role of action in perceiving and comparing functional relations Ivan Vankov ([email protected]) Central and East European Center for Cognitive Science New Bulgarian University, 21 Montevideo Street, Sofia 1618, Bulgaria Boicho Kokinov ([email protected]) Central and East European Center for Cognitive Science New Bulgarian University, 21 Montevideo Street, Sofia 1618, Bulgaria Abstract There is growing evidence that even the most abstract capacities of human cognition are not entirely amodal and disembodied. The present study presents two empirical studies which aim to demonstrate that relational reasoning is grounded in our sensory-motor experience. Experiment 1 shows that the affordances of tool-like objects have an effect on comparing functional relations. Experiment 2 makes sure that this finding can not be explained by an automatic activation of motor systems. The results are interpreted as evidence that at least certain functional relations are perceived by simulating interactions with the environment. It is also asserted that the process of comparing such relations is constrained by the properties of the human body such as hand-dominance. Keywords: relations, situated cognition, embodiment, action, simulation, analogy Introduction Imagine you are asked to compare the relation between an axe and a wooden log with the relation between a meat chopper and a piece of meat. One way to solve this problem is to find out what relation holds in the first pair of objects, turn it into propositional form (e.g. ‘is used to cut(axe, wood)’), do the same for the second pair of objects and than compare the two symbolic structures. This is what many models of relational reasoning do (Gentner, 1983; Falkenhainer, Forbus, & Gentner, 1989). Some models can also establish a correspondence between distinct relational symbols by measuring their semantic similarity (Holyoak & Thagard, 1989; Kokinov, 1994; Hummel & Holyoak, 1997, 2003). However all these models do not address the problem of where the relational meaning comes from (how the propositions are encoded) and they all assume that the process of comparing relations is amodal and disembodied in nature. On the other hand, there is plenty of evidence that human cognition is inherently modal and constrained by the characteristics of the human body (Glenberg, 1997; Barsalou, 1999; 2008; Lakoff, 1999; Fisher & Zwaan, 2008). For example, it has been shown that the perception of a graspable object immediately activates a potential motor interaction with this object, even when it is task-irrelevant (Tucker & Ellis, 1998, 2004; Beauchamp, Lee, Haxby & Martin, 2002; Beauchamp & Martin, 2007; Buccino, Sato, Cattaneo, Roda & Riggio, 2009). Proponents of the embodiment theory claim that this phenomenon is not a mere side effect of spreading activation, but that motor programs are part of the representation of objects. These motor programs are used to simulate potential interactions with an object and determine its function. Similarly, the perception of a functional relation between two objects should require a mental simulation of the relevant interactions with the objects. For example, in order to comprehend the functional relation between an axe and a piece of wood, you would simulate grasping the axe and chopping the wood with it. And in order to compare two instances of functional relations you have to be able to compare the motor dynamics resulting from simulating the actions involved in each of the relations. Such an approach is justified by the study of Klatzky, Pellegrino, McClosky & Lederman (1993), which found that there is remarkable consistency in people's knowledge about the movements underlying functional interactions with objects. There is also evidence that sometimes people consciously try to detect the perceptual motor similarities of different situations in order to evaluate how analogous they are (Clement, 2009). Figure 1: An example of the stimuli used by V&K. Participants had to compare the relation between the objects in the left part with the relation in the right part of the screen. The affordances of the objects were manipulated by making them easier to be grasped with the left or with the right hand. In this example, both affordances are right. Recently, Vankov & Kokinov (2009) (henceforth V&K) proposed a model of grounding relational meaning in simulated interactions with the environment. According to the model, the motor dynamics resulting from these interactions is used not only to comprehend relations, but also to solve the role-filler binding problem (Hummel, 1999). The model makes two major predictions. First, it