Action may indeed rely on unconscious sensory processing, but that may not be true of perceptual recognition in aquatic species, which in mammals relies on brain circuits that are paradigmatically associated with conscious processing. Perhaps limited sensory horizons even make model-based conscious processing especially important for finding prey and mates, avoiding predators, and navigating underwater environments.

Evaluation of biological significance in the visual cortex occurs along multiple parallel pathways to provide fast but coarse processing. This was critical when visual horizons were expanded by the sea-to-land transition, which meant broader visual exposure to a range of survival threats. Response latencies in the macaque cortex and postdictive effects of evolutionary-relevant stimuli in humans attest to fast visual processing.

We discuss the functions and evolution of conscious vision. Conscious vision, we argue, operates too slowly to be suited for immediate actions, but instead evolved for offline cognition. We trace the emergence of conscious vision to the water-to-land transition, where larger terrestrial sensory horizons allowed animals to benefit from model-based planning. This shift drove the evolution of “reality monitoring” – the capacity to determine whether internal signals reflect external reality or endogenous activity uncoupled from sensory input. Following higher-order theories of consciousness, we associate consciousness with this reality monitoring function and discuss novel empirical predictions.

Determining the functions of consciousness is difficult, in part because the term “consciousness” subsumes many different sensory modalities and aspects of cognitive processing. Fleming and Michel make an excellent start by forwarding compelling arguments regarding functions for conscious vision. I consider how their picture might generalize beyond vision, by looking at two areas: conscious proprioception and conscious mental imagery.

The commentaries collectively provide a thoughtful and constructive analysis of resource-rational contractualism (RRC), helping both to push the theory forward and clarifying its limitations. We use these as a springboard to revise and improve upon the ideas presented in the target article. Our replies are organized in three main themes: possible extensions, challenges, and wider applications.

The authors make a timely argument that the temporal profile of consciousness is an under-exploited constraint on theories of consciousness. However, the exact timing matters for the evolutionary hypothesis. A rival hypothesis is that in aquatic environments, it already had the function of representing distal objects, with a time-delay to consciousness in the 100–200ms range.

Fleming and Michel propose that as our evolutionary ancestors made the water-to-land transition, model-based, conscious planning and thought developed alongside the model-free, unconscious, reflexive behaviors that were sufficient in aquatic environments. But a century of research has failed to build a compelling body of replicable evidence that behavior can be controlled unconsciously in humans, severely undermining this theory.

Why might people make negative moral judgments about others’ consensual exchanges? Within triple theory, these should count as mutually beneficial agreements and therefore enjoy uncontroversial moral approval. We propose two explanations: observers of an exchange may (1) do their own cost-benefit analysis or (2) prioritize moral norms. Additional theorizing is required for either of these to align with resource rationality.

Speed of consciousness results may contribute to the popular notion that our consciousness is inefficacious. I argue that Fleming and Michel’s claim that consciousness is slow, coupled with their idea that consciousness marks a predictive world model, instead supports a view of agency in which we are consciously engaged with the world as it unfolds, both in perception and response.

We support the proposal and advocate for greater continuity between in-the-moment moral reasoning and culturally-evolved moral norms. The target article construes culturally-evolved norms as rigid byproducts of blind selection and therefore out of scope. A more integrative account can be achieved by viewing norms as culturally-evolved joint programs that reflect a history of resource-rational reasoning by the people who learned them and passed them on.

A central plank of Fleming and Michel’s thought-provoking paper is that postdiction reveals a lower bound for the speed of consciousness, showing that perceptual awareness is slow and motivating their perceptual reality monitoring theory of consciousness. The plank cannot bear the weight: Postdiction neither demonstrates a lower bound for the speed of consciousness nor shows that awareness is slow.

Levine et al. argue that morality solves the problem of finding mutual benefit in interdependent choice. We explore how legal and institutional design can reduce cognitive burdens and support agreement-finding—explicit, implicit, and simulated. From contract law to government and organizational policy, we highlight how aligning systems with resource-rational contractualist principles can support individuals in finding mutual benefit.

Fleming and Michel suggest that conscious vision reflects a reality monitoring (RM) mechanism that evolved to separate planning from perception. Here, I critically examine the RM decision policy implied by the authors – a slow, winner-take-all strategy tagging only one first-order representation as “real.” Specifically, I explore how expectations about environmental or reality-tagging stability may mask alternative RM decision policies, with important implications for learning and the evolutionary emergence and function of consciousness.

The contractualist approach to morality holds a great deal of promise. However, it is not clear how it applies to situations in which the idea of negotiation is itself morally abhorrent. In particular, it is not clear how it applies to the thick relationships that exist between friends and family, which form a large part of the moral universe.

Explaining altruistic behavior may require expanding the bargaining framework. In particular, accommodating altruism may require greater emphasis on bargaining under normative rather than actual conditions. Fleshing out the account of resource-rational bargaining under normative conditions raises important questions for future research.

This commentary critiques Fleming and Michel’s “sensory horizons” framework, arguing that it underestimates the fragility and fallibility of conscious vision. Drawing on clinical and neuroscientific evidence, we highlight frequent failures in reality monitoring and affordance perception in psychopathology. We propose that a revised model must integrate predictive processing and neural dysregulation to fully account for both normal and abnormal conscious experience.

The authors neglect the most distinctive and puzzling features of human moral experience and thus fail to offer a theory of moral cognition at all. Humans experience moral norms and obligations as externally imposed on us in ways we are not free to modify, invent, or ignore, in sharp contrast to the motivations and outcomes of bargaining or prior agreements.

In consciousness science, theoretical predictions are often untestable, such as claims about phenomenal consciousness in other beings. This evidential underdetermination, in combination with the perceived moral significance of consciousness, puts consciousness science at risk of becoming a marketplace of rationalizations: a field that produces theories that reaffirm social practices and conventions.

While Levine et al. restrict themselves to explaining interpersonal moral judgments using a resource rational contractualist lens, we believe that their ideas could be extended to produce a theoretical account of personal virtues by taking into consideration the fact that people are able to think about their future selves. We suggest how such intra-individual contractualist explanations could be formulated.

Elite athletes perform visually guided actions under severe temporal constraints. Reconsidering conscious versus unconscious vision, we present evidence from elite athletes that challenges the conceptualization of unconscious action guidance and propose instead distinguishing seeing-for-movement and seeing-for-thought. Elite athletes appear to use both in a coordinated manner when performing visually guided actions.