Cognitive processes underlying spatial belief revision

Cognitive processes underlying spatial belief revision Leandra Bucher (leandra.bucher@psychol.uni-giessen.de) Antje Krumnack (antje.krumnack@psychol.uni-giessen.de) Jelica #ejasmic (jelica.nejasmic@psychol.uni-giessen.de) Markus Knauff (markus.knauff@psychol.uni-giessen.de) Justus-Liebig University, Experimental Psychology and Cognitive Science, Otto-Behaghel-Str. 10F, 35394 Giessen, Germany Abstract New information sometimes contradicts what is believed about certain states of the world. To integrate contradicting information, reasoners have to revise existing beliefs. In the course of belief revision they need to decide which beliefs to retain and which ones to retract in order to regain consistency within current belief states. What guides belief revision has been studied in the non-spatial domain. Based upon previous work on spatial reasoning, we develop hypotheses about the cognitive processes of belief revision in spatial reasoning. Spatial beliefs are considered to be based on spatial mental models that are subject to variation during revision. We provide empirical evidence that models are varied according to information provided by inconsistent statements rather than processes vital for construction of initial models. Furthermore we show that revising spatial models follows dissociable principles from constructing initial spatial models. Keywords: Spatial reasoning; mental models; belief revision; spatial cognition; relational reasoning; spatial relations Introduction Imagine the following situation. You look for an apartment to buy when two friends of yours tell you about a nice place offered by the local real estate agency. A says: “There is parquet floor either in the lounge or in the bed room, but not in both rooms.” B says: “There is parquet floor in the bed room.” As a matter of fact you get the information from the real estate broker: “There is parquet floor in the lounge.” Given the broker knows what he sells, either A´s or B´s statement must be wrong. Which statement do you believe is true – A´s or B´s? Whom do you believe – A or B? Processes that guide this decision are subject to belief revision research. Belief revision is an everyday task that describes the process reasoners perform in order to regain consistency when confronted with new information that is inconsistent with existing beliefs. Generally, it requires knowledge about the entities reasoned about, e.g. such that two entities cannot reside within the same spatio-temporal coordinates. Furthermore, belief revision might be affected by numerous factors such as familiarity with the entities or trustworthiness of speakers uttering the information (Wolf & Knauff, 2008). Here, we study belief revision in reasoning about objects that are neutral regarding these factors. Specifically, we look at factors that play a role in the initial construction of spatial mental models and carefully examine their potential role in processes during revision of these initially constructed models. The processes that guide belief revision in the spatial context have not been investigated so far. Based on what is known about reasoning with spatial mental models (De Vooght & Vandierendonck, 1998; Klauer, 1998; Byrne, 1998; Schaeken, Girotto, & Johnson-Laird, 1998; Goodwin & Johnson-Laird, 2005), we develop hypotheses about cognitive processes that might be vital for belief revision in spatial reasoning. Empirical evidence culminates to the assumption that spatial reasoning relies on the construction, inspection, and variation of spatial mental models (Knauff, Rauh, & Schlieder, 1995; Rauh, Schlieder & Knauff, 1997, Knauff Rauh, Schlieder, & Strube, 1998). We briefly summarise relevant work on spatial reasoning and come up with hypotheses about belief revision in the spatial domain. Subsequently, we report empirical evidence from two experiments that tested these hypotheses and discuss implications of the results. Construction and inspection of spatial mental models Consider the following spatial description that determines the linear arrangement of three objects: The apple is left of the pear. The mango is left of the apple. Sentences of this kind are called premises, from which a mental model can be incrementally constructed by successively integrating information. Starting with the first premise (1), the two objects are arranged according to the relations and result in the following model (M1), (M1) apple – pear Successively, the information from the second premise (2) is integrated, resulting in the model (M2):