Abstract: What is the nature of semantic memory? Philosophers and cognitive scientists have long held that semantic memory stores invariant knowledge structures to be retrieved as such. In this paper, I argue that this conception of semantic memory is likely false. In particular, I argue that if episodic and semantic memory share causal mechanisms, and episodic memory is (re)constructive, then semantic memory is likely constructive too. I review evidence that suggests that episodic and semantic memory are subserved by a domain-general system that supports representing and navigating relations among various kinds of stimuli, including space, time, events, and semantic relations. I then review the supposed hallmark properties of constructivism in episodic memory and show that they appear in semantic memory as well. To increase the inductive support for my proposal, I show how the view predicts some of the evidence others have marshaled in favor of a constructivist semantic memory system. Finally, I close by providing a proof of concept for the view on offer, the semantic pointer architecture.

Abstract: Some embodied theories of concepts state that concepts are represented in a sensorimotor manner, typically via simulation in sensorimotor cortices. Fred Adams (2010) has advanced an empirical argument against embodied concepts reasoning as follows. If concepts are embodied, then patients with certain sensorimotor impairments should perform worse on categorization tasks involving those concepts. Adams cites a study with Moebius Syndrome patients that shows typical categorization performance in face-based emotion recognition. Adams concludes that their typical performance shows that embodiment is false. Moebius patients must draw on amodal (non-embodied) emotion concepts. In this paper, I review face-based emotion recognition studies with Moebius patients yielding conflicting results and diagnose these conflicts as a difference in experimental design. When emotion labels are provided, patients have typical performance, but when labels are not provided patients are severely deficient. I then show how an embodied, psychological constructionist view of emotions predicts and explains these performance differences. The upshot is that embodied theories of concepts are vindicated. 

Abstract: A central task of developmental psychology and philosophy of science is to show how humans learn radically new concepts. Famously, Fodor has argued that such learning is impossible if concepts have definitional structure and all learning is hypothesis testing. We present several learning processes that can generate novel concepts. They yield transformations of the fundamental feature space, generating new similarity structures which can underlie conceptual change. This framework provides a tractable, empiricist-friendly account that unifies and shores up various strands of the neo-Quinean approach to conceptual development.

Abstract:In this article, we explore various arguments against the traditional distinction between episodic and semantic memory based on the metaphysical phenomenon of transitional gradation. Transitional gradation occurs when two candidate kinds A and B grade into one another along a continuum according to their characteristic properties. We review two kinds of arguments—from the gradual semanticization of episodic memories as they are consolidated, and from the composition of episodic memories during storage and recall from semantic memories—that predict the proliferation of such transitional forms. We further explain why the distinction cannot be saved from the challenges of transitional gradation by appealing to distinct underlying memory structures and applying our perspective to the impasse over research into ‘episodic-like’ memory in non-human animals. On the whole, we recommend replacing the distinction with a dynamic life cycle of memory in which a variety of transitional forms will proliferate, and illustrate the utility of this perspective by tying together recent trends in animal episodic memory research and recommending productive future directions.

Abstract: Patient RB has a peculiar memory impairment wherein he experiences his memories in rich contextual detail, but claims to not own them. His memories do not feel as if they happened to him. In this paper, I provide a causal model of RB's phenomenology, the self-attentional model. I draw upon recent work in neuroscience on self-attentional processing and global workspace models of conscious recollection to show that RB has a self-attentional deficit that inhibits self-bias processes in broadcasting the contents of episodic memories to the global workspace. Typically, self-related contents enjoy a higher level salience than other-related contents. Elimination of bias toward self-related contents diminishes the salience of those contents to the level of other-related contents. Because the typical high salience of self-related content is necessary for the feeling of ownership, RB lacks the feeling of ownership. I also discuss potential applications of the self-attentional model to other psychopathological cases. 

Abstract: Philosophers and cognitive scientists alike have recently been interested in whether cognition extends beyond the boundaries of skin and skull and into the environment. However, the extended cognition hypothesis has suffered many objections over the past few decades. In this paper, I explore the option of control extending beyond the human boundary. My aim is to convince the reader of three things: (i) that control can be implemented in artifacts, (ii) that humans and artifacts can form extended control systems, and (iii) that perhaps extended control ought to be preferred over extended cognition. Using the objections to extended cognition as constraints on my own extended theorizing and the example of autofocus systems in cameras, I decompose and localize the components of an autofocus system that realize the central properties of control from a plausible theory of control in the literature. I then provide criteria according to which control can be extended in a system. Finally, I consider how this theory of extended control ought to be preferred to theories of extended cognition. 

My dissertation, defended in May 2024, took the form of three individual papers. Below I summarize each of the papers; they both challenge and defend significant findings in cognitive science, offering original contributions to ongoing controversies in the science of the mind. 

CHAPTER 1: Constructing Embodied Emotion with Language: Moebius Syndrome and Face-Based Emotion Recognition Revisited (see entry on embodied concepts below for preprint)

One strand of the embodied cognitive science research program claims that mental concepts are embodied. One popular way to understand this claim is that concepts are represented in a sensorimotor manner, typically via simulation in sensorimotor cortices. For example, my thoughts about dogs, according to this view, would involve simulating petting, smelling, or seeing dogs through reactivation of areas of the brain that implement those actions. Fred Adams (2010) has advanced an empirical argument against such embodied concepts reasoning as follows. If concepts are embodied, then patients with certain sensorimotor impairments should perform worse on categorization tasks involving those concepts. Adams cites a study with Moebius Syndrome patients, who suffer from facial paralysis, that shows typical categorization performance in face-based emotion recognition. Adams concludes that their typical performance shows that embodiment is false. Moebius patients must draw on non-embodied emotion concepts. In this paper, I review face-based emotion recognition studies with Moebius patients that show conflicting results, and diagnose these conflicts as emerging from differences in experimental design. When emotion labels are provided, patients have typical performance, but when labels are not provided patients are severely deficient. I then show how a recently developed embodied view of emotion predicts and explains these performance differences. My argument vindicates theories of embodied concepts.

CHAPTER 2: What’s in a Color?: Labels, Gesture, and Synesthesia (Draft available on request)

The label superiority effect, wherein subjects provided with labels in discrimination and categorization tasks outperform those subjects without labels, is a well known finding in cognitive psychology. A commonly accepted explanation for this effect is that language, and in particular, category labels, make perceptual representations more categorical. Perception is more categorical when between-category differences are more salient than within-category differences, as when the boundary between red and yellow is more salient than reddish-orange and reddish-yellow. Categorical perception involves the abstraction, or removal, of category-irrelevant features. Is language special in its ability to augment perceptual representations through abstraction? In this paper, I argue that it is not. I examine cases of olfactory-color synesthetes, who associate odors with colors. These individuals show a capacity for odor categorization and identification that far exceeds that of normal individuals. I argue that these results challenge the claim that language is special in its ability to augment representations through abstraction. I then present evidence that, I argue, reveals a common mechanism subserving representational abstraction across language and synesthesia.

CHAPTER 3: The Constructive Episodic-Semantic Memory System (Draft available on request)

What is the nature of semantic memory? Philosophers and cognitive scientists have long held that semantic memory stores invariant knowledge structures, such as the year of Lincoln’s assassination, or the currency used in France. The received view distinguishes semantic memory from episodic memory, which is thought to encode incidents particular to one’s life, such as the kind of cake you ate at your twelfth birthday or the location where you honey-mooned. In this paper, I argue that this distinction is illusory. In particular, I argue that the neural mechanisms that realize episodic and semantic representations grade into one another, exhibiting the metaphysical phenomenon of transitional gradation. I then motivate the consensus view regarding the constructive nature of episodic memory. That view says that episodic memory stores “traces”, as opposed to the entirety of an event. At retrieval, the trace recruits various sensorimotor, affective, and spatial representations to reconstruct the target event as faithfully as possible. If it is true that episodic and semantic memory systems grade into one another, and it is true that episodic memory is constructive, then we have good reason to think that semantic memory is constructive too. I close the paper by providing independent evidence that semantic memory is constructive, and developing a proof of concept for the view on offer.

1. “Are Concepts Stored in a Symbolic Format?” (3000 words)

I argue that the inference (popular among Language of Thought theorists) from the observation that thought is compositional to the conclusion that concepts are stored in a symbolic format is not valid by appealing to a distinction between the storage and deployment of concepts. (Draft available on request).

2.  “What’s That Smell I Remember So Well?” (3000 words)

I argue that olfactory representations are high-dimensional, getting compressed through interfacing with language while remaining potent cues for episodic memory.

3.  A paper on aphasia and machine learning.