Series de tres términos y modelos de representación

Reasoning yields a conclusion from percepts, ideas, or assertions. Deduction yields valid conclusions if the premises are true. Some deductions depend on reasoning about relations. For example, «Anne is better than Sue: Julia is worse than Sue:: Who is worst?» Early laboratory studies suggested that reasoners construct mental diagrams (spatial model) although linguistic analyses are also involved (linguistic model). Two additional processing models were proposed some years later: the mixed and the levels of representation models. These models are based on a combination of mental images and linguistic processing. The processing time and the difficulty level of 32 three-term series of single solution (16 positive comparative and 16 negative equatives) are predicted from the specific cognitive processes postulated by the models. The present study assesses 980 participants to compute the processing time and the difficulty level of 32 three-term series. A regression analysis is computed to predict processing time and difficulty level. The spatial model is the better predictor of both processing time and difficulty level. This finding is the just opposite of the observed in previous research. Some implications are discussed.