AAP 2026 CONFERENCE

Aphantasia, where individuals report lacking voluntary conscious visual imagery, has increasingly captured attention in empirical and philosophical literature. One fascinating aspect of this condition is: empirical findings suggest that aphantasics employ visual strategies to perform tasks—at least in some cases. The discrepancy between subjective reports and objective evidence motivates some researchers to defend the “unconscious” view, according to which aphantasics use unconscious visual imagery. In this talk, I will argue for the alternative, “conscious” view through the lens of metacognition.

When it comes to eliciting reliable introspective reports, consciousness research highlights a distinction between “visibility” measures and “confidence” measures. All studies on aphantasia use the Vividness of Visual Imagery Questionnaire (VVIQ) to recruit participants. As the VVIQ asks participants to only indicate the “vivacity” of their mental images, it closely resembles a visibility measure. I contend that in determining whether aphantasics indeed rely on unconscious visual imagery, we should further investigate their confidence in the accuracy of their imagery-task performance.

Few studies on aphantasia have included confidence ratings. However, intriguingly, aphantasics typically exhibit good metacognitive sensitivity. These findings provide supporting evidence that they engage in conscious, task-relevant visual imagery.

Humans are wishful thinkers: we are more likely to believe the things we want to be true (Windschitl et al., 2022). In contemporary psychology this phenomenon is known as desirability bias (Tappin, Van der Leer & McKay, 2017) and as a cognitive bias it is relatively understudied. From the perspective of scientific realism, desirability bias is irrational: reality is largely indifferent to our desires, but perception seems quite responsive to them. Desirability bias represents a key human cognitive vulnerability and therefore deserves investigation. There are a number of questions to ask of the phenomenon. How strong and robust is it? What moderates it? Can we reduce or enhance it? What might explain it? How does it relate to existing theories and literatures, such as self-deception, motivated reasoning and predictive processing? What is its relationship with (the much more famous) confirmation bias? How much certainty do we have about the above answers? I will review some of the literature that is relevant to these questions and discuss how we might go about working towards a satisfactory explanation for desirability bias.

Whether or not we have free will has long been a philosopher’s question, but in the last 50 years, neuroscientists have claimed to be able to weigh in on the problem. I begin by reviewing the philosophical landscape of free will, which is dominated by the question of whether or not determinism is true, and argue that neuroscience cannot provide evidence for or against the truth of determinism, so cannot bear on the problem in that respect. I then turn to another way in which neuroscientists have tried to provide evidence that we lack free will, by drawing on evidence that they interpret as showing that our conscious will is inefficacious. I argue that new ways of thinking about how to interpret that evidence undermines the conclusions that typically had been drawn. I close by considering how this new interpretation of the neural data may bear upon some standard philosophical positions in the free will debate.

People use their commonsense thinking about the past to inform their decisions. Intuitive historical thinking is therefore pervasive in the social and cognitive lives of humans. This type of cognition has not been systematically researched. Recent philosophical psychology is dominated by works that investigate cognitive tools used by intuitive historians – such as episodic memory, mental time travel, concepts of time, or causal reasoning – without directly studying intuitive historical thinking. To remediate this lacuna, we investigate intuitive historical thinking, referred to as ‘intuitive history’. We argue against the view that intuitive history can be reduced to any one of the cognitive tools used by intuitive history. The processes and phenomenology of intuitive history are linked to three types of interrelated activities routinely conducted by intuitive historians: managing historical information perceived as significant, which includes searching, gathering, storing, and updating information about the past; the interpretation of historical information, which may include the intuitive historian’s distinctive experiences, assumptions, emotions, and evaluations; and the use of historical information. Interpretative processes can be influenced by the assumption of pastness, singularity, reality, connectivity or causation, and significance. We review evidence suggesting that intuitive historians routinely use these assumptions to develop their inquiries into past entities.

Causation in classical computation is relatively simple – it involves a series of events, each causally dependent on its predecessor. Causation in neural networks can be more complex. One reason for this is the possibility of recurrent or re-entrant signals. This paper investigates this topic. First, drawing on the analysis of Anne Treisman, I look at the role that recurrence may play in binding – which in this context can be thought of as the combination of simple representations into more complex representations. Second, I analyse the kinds of causal patterns involved here through Luigi Pasinetti’s work, recently translated into English, on causal dependence and interdependence. The idea, roughly speaking, is that neural networks that bind simple representations into complex representations through some kind of recurrent activity will be comprised of states in relationships of interdependence, rather than the relations of casual dependence that characterise classical computation. In this way, it is suggested, Pasinetti’s concepts can be used to distinguish the causal patterns that characterise classical computation from those that characterise the kinds of neural networks described here.