The Role of Agent-Patient Relations in Evaluating the Quality of an Analogy

MARTÍN A.; TAVERNINI, M; TRENCH, M.; MINERVINO, R.

Dijon

Congreso; Third International Conference on Analogy; 2013

University of Burgundy

Minervino, Oberholzer and Trench (submitted) demonstrated that evaluating the quality of an analogy sometimes consists in deciding whether two events are instances of a schema relational category (SRC). We carried out an experiment to find out: (a) if the assignment of two facts to a SRC depends on the relationship between their agents and patients, and (b) whether the less abstract is the SRC that the two events are members of, the more analogous such facts are considered. Participants were presented with sets of statements composed of a base analog (BA) and three target analogs (TAs). For each TA, they had to indicate to what extent they considered it to be analogous to the BA. Keeping constant the relations and objects, we varied the extent to which the agent-patient relationship in the TAs was similar to that of the BA. For instance, a BA such as The professor sent flowers to his student was followed by TAs such as The singer sent flowers to his fan (TA1), The child sent flowers to his grandmother (TA2) and The farmer sent flowers to the biologist (TA3). The agent-patient relationship in TA1 was chosen such that the SRC capable of encompassing BA and TA1 (seducing) was less abstract than the SRC capable of encompassing the BA and TA2 (pleasing). In turn, the agent-patient relationship of TA3 was chosen such that no shared SRC between BA and TA3 could be found (collaborating with a study maintains no resemble with seduction). Pairs of analogs that shared SRC were considered more analogous than those that did not. Furthermore, two events were considered to be more analogous when the SRC capable of encompassing them was relatively less abstract. An analysis of participants´ justifications provides concurrent support for our hypothesis. Implications for models of analogical reasoning are discussed.