Fleming and Michel make a number of assumptions that I reject. Although commonly made, these assumptions reveal a kind of Cartesian materialism that leaves us stuck with hard problem and asking such unanswerable questions as “What is the function of consciousness?” and “Why did consciousness evolve?”

Dennett (Reference Dennett1991) described as Cartesian materialists those who claim not to be dualists but nevertheless believe in a finishing line beyond which previously unconscious contents “become conscious,” or a special place in the mind, or brain, where an observer experiences a stream of conscious perceptions and thoughts. The implication of this way of thinking is that consciousness is something additional to the brain processes on which it depends. Giveaway phrases of this view include: consciousness “emerges”; information (or processing, or contents, or perceptions – the terms vary) “reaches consciousness,” “becomes conscious,” or “requires consciousness”; there are “precursors of consciousness” and “consequences of consciousness.” Fleming and Michel use all these phrases without commenting on the implications.

Interestingly, they mention Dennett’s Orwellian interpretation of the cutaneous rabbit and other postdictive effects, but they miss the very point that Dennett was making. This was to contrast his Orwellian interpretation (in which the person was conscious of the first stimulus but the memory was then wiped out by the second), with his Stalinesque interpretation (in which the first stimulus was delayed on its way “up to consciousness” to allow time for the postdictive effect). His point was not to say that one or other has to be true but that there is no conceivable way of finding out. “This is a difference that makes no difference” (Dennett, Reference Dennett1991, p. 125).

Both interpretations, Dennett argues, fall into the trap of thinking that there is a moment at which “consciousness happens.” Clearly, Fleming and Michel believe there is, for example, when they refer to “the moment you consciously experience the apple.” These unquestioned assumptions lead Fleming and Michel to believe that consciousness must have a function of its own and to ask when and why it evolved. I disagree. If consciousness were distinct or separate from the functions of reality monitoring and model-based planning, which they discuss, this might be true, but I suggest that it is not, and this is why questions about the function and evolution of consciousness have become so troublesome (Blackmore & Troscianko, Reference Blackmore and Troscianko2024).

One further trap is to assume that consciousness can be measured or detected. This is important because if we could detect “consciousness itself” separately from other processes, then the problems referred to above would not apply. But we cannot. Fleming and Michel refer, in parentheses, to consciousness “as measured in humans” and cite Bayne et al. (Reference Bayne, Seth, Massimini, Shepherd, Cleeremans, Fleming, Malach, Mattingley, Menon, Owen, Peters, Razi and Mudrik2024), who search for tests for “phenomenal consciousness” or subjective experience, and ask when it arises in human development and which systems or organisms are conscious. A good C-test, they say, “should address the question of whether it ‘feels like’ something to be the target system” (p. 455), and should “probe consciousness itself” rather than other related phenomena (p 458). Yet in the end, all their C-tests are objective measures such as behaviour, comprehension, sensory abilities, and neural indicators. I suggest that no true C-test can ever be found because “consciousness itself” as something separate from all those functions does not exist. As with Fleming and Michel, their assumptions lead them to ask unanswerable questions.

Is there an alternative? I will suggest one, although my purpose is not to claim that this is the answer, but to show that other ways of thinking about consciousness that do not fall into these dualist traps are possible. I suggest that, rather than consciousness being something additional to brain function, it is simply “what it is like to be a predictive mental model.” This makes some sense for the following reason: if we ask “what is it like to be a bat?,” or any physical system such as a fish, frog, baby, or robot, we cannot answer. But if we ask, “what is it like to be the bat’s model of a bat?” we can answer. It is like whatever that model describes. This is what models of reality are; they are descriptions of a world.

Fleming and Michel refer briefly to predictive processing theories, pointing out that these do not distinguish between unconscious and conscious perception. Yet they continue to make this distinction, which, I suggest, is false. My suggestion is that there is something it is like to be all active predictive mental models, and the reason why we call some of them “conscious” and others “unconscious” is that the former are part of our current model of reality, or model of self in a world, while all the myriad other ongoing models are not.

This view rejects the idea that “conscious vision” has functions distinct from those of “unconscious vision,” and does away with many misleading assumptions. If “consciousness itself” has no function independently of reality monitoring, model-based planning, or any other functions, we need not ask when and why it evolved. Instead, we need to ask only when and why these other functions evolved. These questions we can answer, as Fleming and Michel show. This way forward can, I suggest, answer questions about consciousness without falling into Cartesian materialism.