Mechanisms of Reference Frame Selection in Spatial Term Use: Computational and Empirical Studies

Abstract

Previous studies have shown that multiple reference frames are available and compete for selection during the use of spatial terms such as "above." However, the mechanisms that underlie the selection process are poorly understood. In the current paper we present two experiments and a comparison of three computational models of selection to shed further light on the nature of reference frame selection. The three models are drawn from different areas of human cognition, and we assess whether they may be applied to a reference frame selection by examining their ability to account for both existing and new empirical data comprising acceptance rates, response times, and response time distributions. These three models are the competitive shunting model (Schultheis, ), the leaky competing accumulator (LCA) model (Usher & McClelland, ), and a lexical selection model (Howard, Nickels, Coltheart, & Cole-Virtue, ). Model simulations show that only the LCA model satisfactorily accounts for the empirical observations. The key properties of this model that seem to drive its success are its bounded linear activation function, its number and type of processing stages, and its use of decay. Uncovering these critical properties has important implications for our understanding not only of spatial term use, in particular, but also of conflict and selection in human cognition more generally.

The effects of direction and orientation of located objects on spatial language comprehension

“…As a result, the spatial template may have been constructed not only on RO geometric information, but also on LO directional data. This hypothesis is congruent with recent modelling work (Schultheis & Carlson, 2017) and empirical studies (Li, Carlson, Mou, Williams, & Miller, 2011;Miller, Carlson, & Hill, 2011) showing that other aspects of the environment (e.g., contextual cues), such as the spatial arrangement of other objects surrounding the RO and the LO, may influence reference frame selection. To what extent the geometric information extracted from the LO affects reference frame selection was not tested in the current study and remains an intriguing aspect for future inquiries.…”

“…The results of recent computational modelling work suggested something different by showing that the information concerning the LO may be associated with the selection of a reference frame. The model presented in Schultheis and Carlson (2017) revealed that (a) it was not complete reference frames but parts of reference frames that were competing for selection, and that (b) it was likely that information determining which parts of a frame were selected was not restricted to the type of information commonly discussed in the literature (i.e., information pertaining to the absolute, relative, and intrinsic frame as introduced by Levinson, 2003) but also comprised contextual cues in the environment, such as, for example, axes of symmetry or the geometric properties of salient objects. It has also been shown that the geometric features of the LO were important for spatial language understanding (Burigo & Sacchi, 2013).…”

Social and configural effects on the cognitive dynamics of perspective-taking

“…For instance, when objects are presented in non-canonical vs. canonical orientations (e.g., an "upside-down" pumpkin), language users take longer to formulate descriptions of those scenes or to respond to instructions (e.g., "the pumpkin is above the strawberry") by placing objects at the correct location (Burigo & Sacchi, 2013). Recent computational modeling work (Schultheis & Carlson, 2017) further underscores the contribution of contextual information in the environment to reference frame selection, including configural cues about the axes of symmetry and the geometric properties of objects (e.g., about their direction and orientation).…”

Local Alignment of Frame of Reference Assignment in English and Swedish Dialogue

“…This preference towards intrinsic FoR has been demonstrated in object selection tasks, both when the object array was perceptually available or retrieved from memory [20]. A number of computational models have been developed to accommodate FoR underspecification in locative descriptions [15,25]. However, these studies and models have focused on the interpretation of a locative description in a one-off setting, as opposed to within the context of an ongoing dialogue.…”