Research

My work focuses mainly on the nature of geometric representations in human and non-human animals. Studying geometric representations can help us make headway on important philosophical issues, such as the following: the cognitive foundations of formal geometry, the nativism/empiricism debate, the nature and origin of spatial concepts, the status of the massive modularity hypothesis, the sources of cognitive flexibility, differences between human and non-human animal cognition, the role of language in human cognition, the relevance of evolutionary considerations in psychological theorizing, and the relationship between psychological and neuroscientific theorizing.

While working on my PhD dissertation, I have become convinced that a fruitful way to make progress on these issues is to get involved in assessing specific theoretical models of spatial navigation arising from work in psychology, neuroscience, and robotics.

PUBLICATIONS:

A. Duval. "The representation selection problem: Why we should favor the geometric-module framework of spatial reorientation over the view-matching framework", 2019, Cognition.

Many species rely on the three-dimensional surface layout of an environment to find a desired goal following disorientation. They generally do so to the exclusion of other important spatial cues. Two influential frameworks for explaining that phenomenon are provided by geometric-module theories and view-matching theories of reorientation respectively. The former posit a module that operates only on representations of the global geometry of three-dimensional surfaces to guide behavior. The latter place snapshots, stored representations of the subject’s two-dimensional retinal stimulation at specific locations, at the heart of their accounts. In this paper, I take a fresh look at the debate between them. I begin by making a case that the empirical evidence we currently have does not clearly favor one framework over the other, and that the debate has reached something of an impasse. Then, I present a new explanatory problem--the representation selection problem—that offers the prospect of breaking the impasse by introducing a new type of explanatory consideration that both frameworks must address. The representation selection problem requires explaining how subjects can reliably select the relevant representation with which they initiate the reorientation process. I argue that the view-matching framework does not have the resources to address this problem, while a certain type of theory within the geometric-module framework can provide a natural response to it. In showing this, I develop a new geometric-module theory.

A. Duval. & C. Côté-Bouchard. "The motivation problem of epistemic expressivists", 2023, Ergo. Preprint PDF

Many philosophers have adopted epistemic expressivism in recent years. The core commitment of epistemic expressivism is that epistemic claims express conative states. This paper assesses the plausibility of this commitment. First, we raise a new type of problem for epistemic expressivism, the epistemic motivation problem. The problem arises because epistemic expressivists must provide an account of the motivational force of epistemic judgment (the mental state expressed by an epistemic claim), yet various features of our mental economy seem to show that they can’t do so. Second, we develop what we take to be the most promising response to that problem for expressivists. We end by noting that this response faces an important challenge pertaining to the psychology of epistemic criticism and praise.

A. Duval. "Is there only one innate modular system for spatial navigation?", commentary on “Précis of What Babies Know” by Elizabeth Spelke, forthcoming in Behavioral and Brain Sciences.

Spelke convincingly argues that we should posit six innate modular systems beyond the periphery (i.e., beyond low-level perception and motor control). I focus on the case of spatial navigation (Chapter 3) to claim that there remain powerful considerations in favor of positing additional innate, non-peripheral modules. This opens the door to stronger forms of nativism and non-peripheral modularism than Spelke’s.

C. Klein & A. Duval. “Is pain asymbolia a deficit or a syndrome? Historical reflections on an ongoing debate”, forthcoming in Belgrade Philosophical Annual.

Nikola Grahek’s influential book Feeling Pain and Being in Pain introduced philosophers to the strange phenomenon of pain asymbolia. Subsequent philosophical debate around asymbolia has been partly taxonomic: the deep question is whether it is best understood as a specific neurological deficit or part of a broader syndrome. This paper looks to the history of asymbolia, positioning the origin of the term within broader historical trends. It shows that strange phenomena about pain and motivation have always presented interpretive challenges, and suggests that the current debate mirrors a historical split between German and French traditions. This does not resolve the debate, but does help place it within broader scientific and philosophical contexts.

WORK IN PROGRESS:

"In defense of hardwired flexibility, or: What rodents and humans can do without language" (under review)

Philosophers have argued that one of the biggest challenges for classical cognitive science is to explain how humans can entertain mental states that integrate contents across distinct domains. There are two influential models that have tried to address this challenge from within classical cognitive science: the hardwired model and the language model. The former states that the ability to entertain cross-domain mental states arises from fixed connections that put in direct contact a large number of domain-specific systems in higher cognition. The latter holds that the ability rather comes from the way in which domain-specific systems combine their outputs through the language faculty. In this paper, I reassess the most influential empirical argument in favor of the second model. It originates from behavioral studies of spatial reorientation in psychology. I maintain that not only does empirical research on reorientation not support the language model, it in fact provides a powerful case to uphold the hardwired model over the language model. To make this case, I draw on various neuroscientific findings about spatial reorientation in rodents and humans.

"What is geometry for navigating mammals?"

Behavioral and neuroscientific experiments strongly suggest that mammals can encode in memory some aspect of the geometry of the three-dimensional surfaces of the environments that they visit. Assuming that we take such findings at face value, an important question arises: What is the nature of that geometric information? This paper compares the two main answers that have been proposed in the literature in cognitive science about spatial navigation. The global-shape view holds that mammals encode mathematical parameters that track the overall shape of three-dimensional surfaces with respect to a metric coordinate system. This was the dominant view in the literature until the development of a new view, about a decade ago, which I call the distance-and-direction view. It holds that mammals only encode distance and direction information related to specific three-dimensional surfaces. In this paper, I argue that, contrary to what many researchers assume, the global-shape view is superior to the distance-and-direction view. First, I provide a new interpretation of the main behavioral and neuroscientific findings widely cited in favor of the distance-and-direction view. Second, I put forward two new arguments for the global-shape view by discussing neurobiological findings that have been overlooked in the context of the debate. I end by drawing implications of the new analyses proposed in the paper for three important issues in the cognitive science of spatial navigation and mathematics: (i) the question of how to understand the distinction between boundaries and landmarks, (ii) the debate about the existence of cognitive maps, and (iii) the issue of characterizing the ontogenetic and phylogenetic origins of humans’ ability to do formal geometry.

“A neurobiological argument for the geometric module”

One widely discussed idea about the source of navigation abilities in human and non-human animals is the geometric-module hypothesis. It states that many species, including humans, possess a module that operates exclusively on representations of the global geometry of three-dimensional surfaces to guide behavior. The hypothesis arose from a specific type of behavioral experiments known as reorientation experiments. Despite its initial popularity, the hypothesis has come under significant pressure in the last fifteen years from a number of objections and from the formulation of alternative models of reorientation experiments. In this paper, I offer a wide-ranging analysis in its favor. I start by considering an objection against the hypothesis that hasn’t been discussed at any length in the scientific literature even though it seems to cut right to its heart. It is that the geometric-module hypothesis appears at odds with well-known neurobiological findings that navigation-related cells in the mammalian hippocampal formation are sensitive to non-geometric information. To answer this objection, I provide a new cognitive architecture for spatial navigation in mammals. I then make the case that any plausible model of spatial reorientation — and of spatial navigation more generally — will need to be committed to this architecture to account for the relevant neurobiological findings. In doing so, I reject all extant theories of reorientation that deny the geometric-module hypothesis.

“The case for nativism about geometry”

This paper provides three detailed arguments for nativism about geometry based on research about spatial navigation. The first two have been hinted at in the philosophical and scientific literature, but have never been developed in detail. The third one is new. The first one is that evidence of sensitivity to geometric properties of spatially-tuned neurons in mammals soon after birth strongly supports a nativist account of the origin of these cells. The second one is that, insofar as the influential geometric-module hypothesis about spatial reorientation is true, it is difficult to explain how the module could have been acquired through general-purpose learning systems. The best explanation is that it is innate. The third one is that, insofar as the geometric-module hypothesis is true, we need to assume that attendant navigation mechanisms dealing with additional geometric properties are innate as well. On the resulting picture, human and non-human mammals have multiple navigation systems that track distinct geometric properties of various types.

“Reviving strong nativism about formal geometry”

Nativism about formal geometry is the view that we possess a significant amount of innate cognitive structure underpinning our ability to do formal geometry. It has had numerous proponents in the history of philosophy — most notably Plato, Descartes and Kant — as well as many detractors. Yet, nativism about formal geometry has gone virtually undiscussed in contemporary philosophy, despite great advancements in cognitive science over the last 40 years that reveal the sophistication of the cognitive and neural underpinnings of our geometric abilities. Thankfully, Elizabeth Spelke and colleagues have recently developed a detailed, empirically well-motivated, and influential framework for thinking about the origins of our ability to do formal geometry. It supports a moderate kind of nativism according to which the ability arises from two innate cognitive systems that deal with limited geometric content. I use their account as a foil to develop a more strongly nativist framework about formal geometry, which is more in line with the views of Plato, Descartes and Kant. I maintain that many of the concepts underpinning formal Euclidean geometry — concepts like POINT, LINE, PLANE, STRAIGHT and CONGRUENT — are innate and come from the internal resources of navigation mechanisms.