Tom Roberts

If loneliness is a psychological state experienced by subjects, what does it mean to speak of things and places as exuding lonely atmospheres? In previous work (Osler et al. 2026), we argued that spaces seem lonely when they present “obsolete affordances”: action possibilities that remain unrealised due to material decay or an absence of human uptake. Here, we extend this framework to the digital realm. We argue that lonely atmospheres generated by dead game worlds, abandoned discussion forums, and forgotten websites similarly arise via their obsolete affordances. We show that digital obsolescence presents a unique phenomenological profile: unlike material ruins, which can be physically repurposed or reclaimed by nature, abandoned digital environments cannot. Through a series of case studies, we unpack the character of digital affordances, their experienced obsolescence, and show how this analysis enriches broader discussions of affordances, affect, and atmospheres.

Talkative artificial systems can sustain a realistic impression of intelligent agency, by generating quick, fluent, and context-sensitive strings of text in response to human input. Realism about AI treats this linguistic proficiency as evidence for an underlying suite of cognitive capacities, and so regards these technologies as subjects of psychological states. Firstly, I raise a challenge for the realist view from the perspective of embodied cognition: how might digital systems undergo interoceptive, affective, perceptual, and action- oriented mental states that seem to bear an important connection to the animate body? Secondly, I propose that the anti-realist view is consistent with taking a fictionalist attitude towards artificial speakers. When we converse with a talkative AI, we often imagine, pretend, or make-believe that our interlocutor is an autonomous agent with thoughts and experiences of its own, just as we do for more traditional fictional characters. These fictions are distinctive in being interactive, co-constructed, and self-involving; and in being capable of blending the world of the real with the world of the imaginary.

A realist view of artificial intelligence is one that holds that artificial systems can exhibit genuine cognitive powers of one sort or another. In this paper, I examine the further question of whether there exist artificial thinking things that are the bearers of these powers. That is, whether individual events or episodes of artificial cognition are ever bound together in virtue of belonging to the same (relatively robust and persisting) artificial thinker. I propose a dilemma for the AI-realist: either accept that artificial thinkers exist and take on the burden of explaining their conditions of individuation and persistence; or deny that artificial thinkers exist and accept that this places substantial limits on the array of cognitive phenomena that can be legitimately attributed to artificial systems.

While loneliness is typically conceived of as a psychological state experienced by agents, people often describe objects and places as lonely. Through an analysis of examples ranging from Ray Bradbury's automated house to abandoned buildings to digital decay, we emphasise how objects can be perceived as reaching out into the world for interaction and engagement and these invitations being left unfulfilled, hanging in mid-air. We suggest that we can understand this as a perception of what we call obsolete affordances—where we experience objects and places as lonely when we experience them as offering up action possibilities that remain unrealized or unrealizable due to the absence of human uptake or material deterioration. This approach emphasises the relational character of affordances while highlighting how loneliness can be embodied in material things, not just people. In doing so, we foreground how objects themselves can appear estranged or cut off from us, even when they offer an abundance of interactive possibilities that have been rendered obsolete.

Epistemic bubbles are informational structures that restrict a thinker’s access to a full range of relevant sources, giving rise to a narrow and one-sided doxastic perspective (Nguyen 2020). Here, I introduce the parallel concept of an affective bubble: a socio-informational structure that impinges in distinctive ways upon its occupants’ emotional states. Affective bubbles, which arise in both online and offline space, can undermine the fittingness, justification, and authenticity of an agent’s emotions by harnessing psychological mechanisms of attention, self-regulation, and imagination.

Kristiansen's "Feeling like a perpetual outsider: relationality in Social Anxiety Disorder" offers a valuable analysis of loneliness within social anxiety disorder (SAD). Although phenomenological psychopathology has given extensive attention to conditions like schizophrenia, depression, and disordered eating, a more nuanced phenomenological examination of SAD is needed (Bortolan, 2023; Tanaka, 2020; Trigg, 2016). Kristiansen's work addresses this deficit and contributes to broader philosophical and phenomenological discussions of loneliness, including recent work on loneliness within psychopathology (e.g., Krueger et al., 2023; Motta, 2021; Roberts & Krueger, 2021; Seemann, 2022). We propose that first-person narratives illuminate two additional, phenomenologically and clinically relevant aspects: bodily doubt and experiences of absence. Understanding these aspects enriches our descriptive understanding of SAD and potentially informs more effective treatment strategies.

Some lines of poetry rhyme to certain readers and not to others because of how their words are pronounced in different accents. And because the accent in which a person speaks is tied in significant ways to aspects of their social identity (such as their age, race, gender, and class), some poems rhyme in the voices of the privileged and not in the voices of the disadvantaged. This paper argues that appreciators can incur aesthetic harms when they are excluded from participating in aesthetic goods that are more fully available to others, and that these harms can constitute aesthetic injustices when they are part of a wider pattern of socio-economic hardship. Poems that embody objectionable gatekeeping practices can thus perpetuate aesthetic injustice.

Debates surrounding the nature of mental disorder have tended to divide into an objectivist camp that takes psychiatric classification to be a value-free scientific matter, and a normativist camp that takes it to be irreducibly values-based. Here we present an overlooked distinction between _status_ and _constitution_. Questions of the form “What is x?” are ambiguous between status questions (“What gives something the status of an x?”), and constitution questions (“Given that something has the status of an x, what is it made of?”). We elucidate this distinction in detail, and argue that normativism is uniquely well-placed to answer status questions while objectivism provides answers, where they are available, to constitution questions.

As technology improves, artificial systems are increasingly able to behave in human-like ways: holding a conversation; providing information, advice, and support; or taking on the role of therapist, teacher, or counsellor. This enhanced behavioural complexity, we argue, encourages deeper forms of affective engagement on the part of the human user, with the artificial agent helping to stabilise, subdue, prolong, or intensify a person’s emotional condition. Here, we defend a fictionalist account of human/AI interaction, according to which these encounters involve an elaborate practise of imaginative pretence: a make-believe in which the artificial agent is attributed a life of its own. We attend, specifically, to the temporal characteristics of these fictions, and to what we imagine artificial agents are doing when we are not looking at them.

In this paper, I draw a distinction between two types of deepfake, and unpack the deceptive strategies that are made possible by the second. The first category, which has been the focus of existing literature on the topic, consists of those deepfakes that act as a fabricated record of events, talk, and action, where any utterances included in the footage are not addressed to the audience of the deepfake. For instance, a fake video of two politicians conversing with one another. The second category consists of those deepfakes that direct an illocutionary speech act - such as a request, injunction, invitation, or promise - to an addressee who is located outside of the recording. For instance, fake footage of a company director instructing their employee to make a payment, or of a military official urging the populace to flee for safety. Whereas the former category may deceive an audience by giving rise to false beliefs, the latter can more directly manipulate an agent's actions: the speech act's addressee may be moved to accept an invitation or a summons, follow a command, or heed a warning, and in doing so further a deceiver's unethical ends.

How do questions concerning consciousness and phenomenal experience relate to, or interface with, questions concerning plans, knowledge and intentions? At least in the case of visual experience the relation, we shall argue, is tight. Visual perceptual experience, we shall argue, is fixed by an agent’s direct unmediated knowledge concerning her poise (or apparent poise) over a currently enabled action space. An action space, in this specific sense, is to be understood not as a fine-grained matrix of possibilities for bodily movement, but as a matrix of possibilities for pursuing and accomplishing one’s intentional actions, goals and projects. If this is correct, the links between planning, intention and perceptual experience are tight, while (contrary to some recent accounts invoking the notion of ‘sensorimotor expectations’) the links between embodied activity and perceptual experience, though real, are indirect. What matters is not bodily activity itself, but our practical knowledge (which need not be verbalized or in any way explicit) of our own possibilities for action. Such knowledge, selected, shaped and filtered by the grid of plans, goals, and intentions, plays, we argue, a constitutive role in explaining the content and character of visual perceptual experience.

We introduce two concepts—social certainty and social doubt—that help to articulate a variety of experiences of the social world, such as shyness, self-consciousness, culture shock, and anxiety. Following Carel's (2013) analysis of bodily doubt, which explores how a person's tacit confidence in the workings of their body can be disrupted and undermined in illness, we consider how an individual's faith in themselves as a social agent, too, can be compromised or lost, thus altering their experience of what is afforded by the social environment. We highlight how a loss of bodily or social certainty can be shaped and sustained by the environments in which one finds oneself. As such, we show how certain individuals might be more vulnerable to experiences of bodily and social doubt than others.

Loneliness is a near-universal experience. It is particularly common for individuals with (so-called) psychopathological conditions or disorders. In this paper, we explore the experiential character of loneliness, with a specific emphasis on how social goods are experienced as absent in ways that involve a diminished sense of agency and recognition. We explore the role and experience of loneliness in three case studies: depression, anorexia nervosa, and autism. We demonstrate that even though experiences of loneliness might be common to many psychopathologies, these experiences nevertheless have distinctive profiles. Specifically, we suggest that: (i) loneliness is often a core characteristic of depressive experience; (ii) loneliness can drive, and even cement, disordered eating practices and anorectic identity in anorexia nervosa; iii) loneliness is neither a core characteristic of autism nor a driver but is rather commonly experienced as stemming from social worlds, environments, and norms that fail to accommodate autistic bodies and their distinctive forms of life. We aim to do justice to the pervasiveness of loneliness in many — if not all — psychopathologies, while also highlighting the need to attend to psychopathology-specific experiences of loneliness, agency, and (non-)recognition.

Traditionally, the sense of touch—alongside the senses of taste and smell—has been excluded from the aesthetic domain. These proximal modalities are thought to deliver only sensory pleasures, not the complex, world-directed perceptual states that characterize aesthetic experience. In this paper, I argue that this tradition fails to recognize the perceptual possibilities of haptic touch, which allows us to experience properties of the objects with which we make bodily contact, including their weight, shape, solidity, elasticity, and smoothness. These features, moreover, may be indicative of how well-suited an object is for its function, and in feeling them we can thus feel the positive aesthetic quality of functional beauty.

The thesis of the extended mind (ExM) holds that the material underpinnings of an individual’s mental states and processes need not be restricted to those contained within biological boundaries: when conditions are right, material artefacts can be incorporated by the thinking subject in such a way as to become a component of her extended mind. Up to this point, the focus of this approach has been on phenomena of a distinctively cognitive nature, such as states of dispositional belief, and processes of planning and calculation. In this paper, we aim to expand the scope of ExM by considering the case for extended affectivity. We begin by clarifying the central commitments of ExM, before investigating its applicability to a range of affective phenomena, both dispositional and occurrent. We argue that proponents of ExM should also accept that the vehicles of emotions, moods, sentiments, temperaments, and character traits can extend beyond skull and skin.

In addiction, apparently causally significant phenomena occur at a huge number of levels; addiction is affected by biomedical, neurological, pharmacological, clinical, social, and politico-legal factors, among many others. In such a complex, multifaceted field of inquiry, it seems very unlikely that all the many layers of explanation will prove amenable to any simple or straightforward, reductive analysis; if we are to unify the many different sciences of addiction while respecting their causal autonomy, then, what we are likely to need is an integrative framework. In this paper, we propose the theory of “Externalist” or “4E” – for extended, embodied, embedded, and enactive – cognition, which focuses on the empirical and conceptual centrality of the wider extra-neural environment to cognitive and mental processes, as a candidate for such a framework. We begin in Section 2 by outlining how such a perspective might apply to psychiatry more generally, before turning to some of the ways it can illuminate addiction in particular: Section 3 points to a way of dissolving the classic dichotomy between the “choice model” and “disease model” in the addiction literature; Section 4 shows how 4E concepts can clarify the interplay between the addict’s brain and her environment; and Section 5 considers how these insights help to explain the success of some recovery strategies, and may help to inform the development of new ones.

According to the evaluativist theory of bodily pain, the overall phenomenology of a painful experience is explained by attributing to it two types of representational content—an indicative content that represents bodily damage or disturbance, and an evaluative content that represents that condition as bad for the subject. This paper considers whether evaluativism can offer a suitable explanation of aversive auditory phenomenology—the experience of awful noises—and argues that it can only do so by conceding that auditory evaluative content would be guilty of widespread error. Defending such an error-theory, moreover, comes with several explanatory costs.

In this paper, we develop an analysis of the structure and content of loneliness. We argue that this is an emotion of absence-an affective state in which certain social goods are regarded as out of reach for the subject of experience. By surveying the range of social goods that appear to be missing from the lonely person's perspective, we see what it is that can make this emotional condition so subjectively awful for those who undergo it, including the profound sense of being unable to realise oneself, in collaboration with others.

Recent work on extended mind theory has considered whether the material realizers of phenomenally conscious states might be distributed across both body and world. A popular framework for understanding perceptual consciousness in world-involving terms is sensorimotor enactivism, which holds that subjects make direct sensory contact with objects by means of their active, exploratory skills. In this paper, I consider the case of emotional experience, and argue that although the enactivist view does not transfer neatly to this domain, there are elements of emotional consciousness whose physical underpinnings include parts of the extra-bodily environment.

Externalist theories hold that a comprehensive understanding of mental disorder cannot be achieved unless we attend to factors that lie outside of the head: neural explanations alone will not fully capture the complex dependencies that exist between an individual’s psychiatric condition and her social, cultural, and material environment. Here, we firstly offer a taxonomy of ways in which the externalist viewpoint can be understood, and unpack its commitments concerning the nature and physical realization of mental disorder. Secondly, we apply a strongly externalist approach to the case of Autistic Spectrum Disorder, and argue that this condition can be illuminated by appeal to the hypothesis of extended cognition. We conclude by briefly considering the significance this strongly externalist approach may have for psychiatric practice and pedagogy.

I argue that it is possible for a subject to undergo experiences of emotional absence, during which she becomes aware of her own failure to be moved by the world around her. Just as a part of one's body feels numb when it manifestly fails to incur the ordinary sensory consequences of transactions at the surface of the skin, so an individual feels emotional absence when her affective condition manifestly fails to vary in predictable ways as she navigates her surroundings. Experiences of emotional absence, such as feeling numb with shock or grief, feeling unamused, or feeling fearless, are not simply flat or neutral states of awareness, but can bear psychological and epistemic significance for the agent.

The project examines the relationship between perception and action, and is divided into two parts. The first establishes a detailed philosophical critique of recent sensorimotor or enactive approaches to perception, targeting in particular the work of Alva Noë. In the second part I defend what may be called an 'action-space' account, according to which conscious experience is constituted by an agent's representing his surroundings in such a way as to enable a certain suite of actions. The enactive approach, I argue, misconstrues the relationship between perception and action and fails in its aim to provide an explanation of consciousness. It faces difficulties, too, when it comes to illusion, hallucination and non-visual perception. The action-space model, by contrast, drawing upon work by Andy Clark, Daniel Dennett and Philip Pettit, has the resources to provide a reductive, functionalist account of phenomenal consciousness; an account that locates consciousness where we want it - in the service of fluid world-engagement by embodied, active perceivers. Thus the perception/action interface is taken to be less direct than on the sensorimotor interpretation, but is nonetheless deep and important. The approach I endorse, furthermore, is consistent with and informed by empirical results from the cognitive sciences, including work on embodied, situated cognition and dual-streams analyses of visual processing.

Two varieties of aesthetic virtue are distinguished. Trait virtues are features of the agent’s character, and reflect an overarching concern for aesthetic goods such as beauty and novelty, while faculty virtues are excellences of artistic execution that permit the agent to succeed in her chosen domain. The distinction makes possible a fuller account of why art matters to us—it matters not only insofar as it is aesthetically good, but also in its capacity as an achievement that is creditable to an individual, and as a reflection or embodiment of virtuous motives.