PhD dissertation (University of Sheffield, 2019)
Where Am I? The Cognitive Architecture of Spatial Reorientation
It is hard to find something more familiar than space. When we perceive, or think about, the world, we do so spatially: we see that a plant is on the table, we hear a sound coming from the right, we wonder whether Jupiter is larger than Saturn. Familiar as it might be, the nature of spatial cognition has proven puzzling, and it has been the source of much debate among philosophers like Descartes, Leibniz, Berkeley, and Kant. In the last few decades, cognitive scientists have joined the conversation, bringing to the table fascinating discoveries about spatial cognition’s neural underpinnings, ontogeny, and evolution.
In my dissertation, I aim to show that we can make significant headway in explaining spatial cognition by combining philosophy and cognitive science. More specifically, my dissertation focuses on the type of spatial cognition involved in spatial navigation, the process of moving one’s body through physical space to get to a desired object or location. It discusses, as a central case study, the phenomenon of reorientation following disruption to spatial behavior. In typical reorientation studies, subjects rely on the shape of a room to find a desired goal after being blindfolded and rotated a few times by an experimenter.
The key contribution of my dissertation is the elaboration of a novel geometric-module theory of reorientation. That is, I defend and develop the hypothesis that reorientation behavior depends upon a geometric module, a specialized cognitive mechanism which operates only on geometric information about three-dimensional surfaces to guide spatial behavior. Accordingly, not only does my dissertation put forward a novel account of a key component of spatial cognition, it also contributes to two foundational debates in philosophy of mind and of cognitive science: the debate about the nature of mental representations, and the debate about the structure of the mind.
Chapter 1 discusses spatial cognition and the geometric-module hypothesis in relation to the two philosophical debates just mentioned. I make the case that the geometric-module hypothesis bolsters the following views: (i) that human and non-human animals’ mental representations encode high-level abstract properties about the world (namely, geometric properties), and not merely low-level sensory properties; (ii) that higher cognition is composed of modules, i.e., specialized cognitive mechanisms that operate with limited information from the rest of the mind.
Chapter 2 considers a powerful objection against the geometric-module hypothesis, which I call the explanatory inflexibility objection. The objection holds that models of spatial navigation committed to the geometric-module hypothesis do not have the required flexibility to deal with evidence that non-geometric cues can affect subjects’ search behavior in some experimental contexts. In response, I argue that there are strong theoretical and empirical reasons in favor of positing a new navigation mechanism whose normal operations can be disrupted by non-geometric cues in ways that account for the relevant findings.
The rest of the dissertation tackles the debate between the geometric-module framework and the most influential alternative explanatory framework of reorientation: the view-matching framework. The latter aims to explain subjects’ search behavior by appealing to snapshots, stored representations of the subject’s two-dimensional retinal stimulation at specific locations of an environment. The view-matching framework rejects the existence of geometric representations or modules in higher cognition. Each of the remaining chapters makes one specific contribution to the debate between the two frameworks.
Chapter 3 puts forward a detailed argument that, contrary to what many geometric-module and view-matching theorists alike suppose, existing evidence from psychology and neuroscience does not clearly favor either framework over the other. Rather, I argue that the debate has reached something of an impasse because each framework can provide systematic explanations for the type of psychological or neuroscientific results seen as favoring the other framework.
Chapter 4 develops a new problem — the representation selection problem — that offers the prospect of breaking the impasse by introducing a new type of explanatory consideration that theories of both types must address. The representation selection problem requires explaining how an agent can reliably select the relevant representation with which they initiate the reorientation process. It arises because the subjects of reorientation experiments have representations from multiple environments in memory while undergoing a trial, be they geometric representations or snapshots. I make the case that view-matching theories do not have the resources to properly address this problem by appeal to various behavioral, neurophysiological, and ethological findings.
Chapter 5 develops a new geometric-module theory which can provide a natural response to the representation selection problem. I motivate it by appeal to the findings discussed in Chapter 4, as well as by extending a well-known evolutionary argument in favor of the geometric-module hypothesis. The upshot of Chapters 3-5 is that we should favor the geometric-module framework over the view-matching framework. I conclude the dissertation by drawing new behavioral predictions from this geometric-module theory and by pointing out that it has stronger implications than standard geometric-module models for the two philosophical debates cited above.
In my dissertation, I aim to show that we can make significant headway in explaining spatial cognition by combining philosophy and cognitive science. More specifically, my dissertation focuses on the type of spatial cognition involved in spatial navigation, the process of moving one’s body through physical space to get to a desired object or location. It discusses, as a central case study, the phenomenon of reorientation following disruption to spatial behavior. In typical reorientation studies, subjects rely on the shape of a room to find a desired goal after being blindfolded and rotated a few times by an experimenter.
The key contribution of my dissertation is the elaboration of a novel geometric-module theory of reorientation. That is, I defend and develop the hypothesis that reorientation behavior depends upon a geometric module, a specialized cognitive mechanism which operates only on geometric information about three-dimensional surfaces to guide spatial behavior. Accordingly, not only does my dissertation put forward a novel account of a key component of spatial cognition, it also contributes to two foundational debates in philosophy of mind and of cognitive science: the debate about the nature of mental representations, and the debate about the structure of the mind.
Chapter 1 discusses spatial cognition and the geometric-module hypothesis in relation to the two philosophical debates just mentioned. I make the case that the geometric-module hypothesis bolsters the following views: (i) that human and non-human animals’ mental representations encode high-level abstract properties about the world (namely, geometric properties), and not merely low-level sensory properties; (ii) that higher cognition is composed of modules, i.e., specialized cognitive mechanisms that operate with limited information from the rest of the mind.
Chapter 2 considers a powerful objection against the geometric-module hypothesis, which I call the explanatory inflexibility objection. The objection holds that models of spatial navigation committed to the geometric-module hypothesis do not have the required flexibility to deal with evidence that non-geometric cues can affect subjects’ search behavior in some experimental contexts. In response, I argue that there are strong theoretical and empirical reasons in favor of positing a new navigation mechanism whose normal operations can be disrupted by non-geometric cues in ways that account for the relevant findings.
The rest of the dissertation tackles the debate between the geometric-module framework and the most influential alternative explanatory framework of reorientation: the view-matching framework. The latter aims to explain subjects’ search behavior by appealing to snapshots, stored representations of the subject’s two-dimensional retinal stimulation at specific locations of an environment. The view-matching framework rejects the existence of geometric representations or modules in higher cognition. Each of the remaining chapters makes one specific contribution to the debate between the two frameworks.
Chapter 3 puts forward a detailed argument that, contrary to what many geometric-module and view-matching theorists alike suppose, existing evidence from psychology and neuroscience does not clearly favor either framework over the other. Rather, I argue that the debate has reached something of an impasse because each framework can provide systematic explanations for the type of psychological or neuroscientific results seen as favoring the other framework.
Chapter 4 develops a new problem — the representation selection problem — that offers the prospect of breaking the impasse by introducing a new type of explanatory consideration that theories of both types must address. The representation selection problem requires explaining how an agent can reliably select the relevant representation with which they initiate the reorientation process. It arises because the subjects of reorientation experiments have representations from multiple environments in memory while undergoing a trial, be they geometric representations or snapshots. I make the case that view-matching theories do not have the resources to properly address this problem by appeal to various behavioral, neurophysiological, and ethological findings.
Chapter 5 develops a new geometric-module theory which can provide a natural response to the representation selection problem. I motivate it by appeal to the findings discussed in Chapter 4, as well as by extending a well-known evolutionary argument in favor of the geometric-module hypothesis. The upshot of Chapters 3-5 is that we should favor the geometric-module framework over the view-matching framework. I conclude the dissertation by drawing new behavioral predictions from this geometric-module theory and by pointing out that it has stronger implications than standard geometric-module models for the two philosophical debates cited above.