Human and animal cognition is grounded in embodiment, which involves more than physical instantiation. Embodiment means an active relationship of dependency upon a real environment, such that the creature’s cognition is what it needs to be in order to exist in that environment. Realness is how the mind experiences that environment (and behaves toward it) in that embodied relation of dependency. But the adequacy of the organism’s cognition for survival does not guarantee its truth or objectivity, which are human ideals established in response to inevitable error.

Science is a cognitive strategy to correct for the limitations and errors of natural cognition. Through technology, it enhances the human organism’s agency in regard to its environment—which we humans call the real world. Yet there remains a blatant gap between our direct and natural cognition, given by the senses, and the scientific understanding of the world. The latter is indirect, inferential and abstract, and “opaque” in the sense that cognitive processing in science is interpersonal and open to conscious scrutiny and direction. In contrast, while also inferential, natural cognitive processing is private and “transparent” in the sense that we have little conscious access or control over it. Helmholtz referred to it as “unconscious inference,” on the analogy of conscious thought processes. But the parallel between them works both ways: scientific cognition can be viewed as a higher-level version of natural perception.

To some extent, we choose our behavior, but scarcely our sensory perception, which seems involuntary and self-evident. From the point of view of an ideal cognitive agent, this aspect of natural perception can be viewed as a disability for which science can be viewed as a compensation. What if this corrected scientific version of reality could be directly incorporated into some ideal cognitive agent’s perception, so that it had the natural experiential quality of immediacy, transparency, and self-evidence? Instead of conceptually understanding the world in terms of scientific abstractions, what if we humans could perceptually experience it through the eyes of our best science? What would it look like? Is that even a coherent idea? To put the question another way, what if a cognitive agent could upgrade its sensory perception—based on unconscious inference—to correspond to its best conscious modeling of the world?

We perceive water, for example, as a continuous substance, while our science knows it is made of discrete molecules. Our visual perception is not acute enough to see molecules in the way that we see macroscopic objects, so we imagine them as small objects of some sort, but which lack the individuating detail that allows us to identify specific macroscopic objects. As it happens, the physical limits of observation and measurement do not allow the detection of such detail, even should it happen to exist on the micro scale. Rather, atoms are “hairless” in the way that black holes are thought to be: characterized only by a small number of parameters. In other words, like black holes, atoms are idealizations. To experience them “directly” would be to see them as stark Platonic forms, not as ordinary objects magnified, which could be potentially rich with more detail. Unlike macroscopic objects, which can differ perceptibly, all molecules of a kind are identical by definition. What is the relation between our sensory perception of water as a continuous fluid and this ideality of water molecules? They both involve acts of cognition—sensory or scientific. Yet the first seems to reflect a limit of human cognition while the second seems to be reflect a property of reality itself. I suggest that they are not so very different.

The fundamental act of mind is a sort of assertion, whereby some property is “filled in” experientially. A paradigm example is the visual blind spot, which is experienced as a continuous visual field despite a gap in enervation on the retina. Another familiar example is the apparent motion effect whereby the individual frames of a motion picture are experienced as continuous. Both cases can be viewed as inability to detect something that is really there—in spite of which it is experienced as there. Alternatively, both can be viewed as the assertion of something that is not there. It would not serve the human organism to be aware of a gap in the visual field, nor of the gap between frames of film below a threshold of speed. Similarly, it would serve no purpose for a creature on our scale to perceive the individual molecules of water, and appropriately it is not equipped with the sensory means to do so. Instead, the mind asserts the continuity of that medium, inventing the experience of a clear liquid, despite the discreteness and “solidity” of its molecules.

However, it does serve a purpose to know that water consists of discrete molecules. Such knowledge enables chemistry. But would it serve chemistry to treat individual molecules as non-identical in the way that macroscopic objects—such as coins—appear? (As physical individuals, coins can have identifying marks that allow them to be distinguished. But as units of value they are identical by definition.) To treat individual molecules as identifiable objects would make mathematical treatment awkward if not impossible. The question is rendered moot by the fact that the physics (of light, for example) does not permit us to closely inspect individual molecules like macroscopic objects. It is convenient, therefore, to idealize them as identical and featureless. Idealization is the conceptual version of perceptual “filling in.” Uniformity is asserted where it may or may not exist.

We live at a certain mesoscopic scale, to which our consciousness is adapted. We can imagine occupying a different scale, such as the very small or the very large—moving among molecules or among galaxies. In this “fantastic voyage” to a different realm, our consciousness carries with it the mentality adapted to our actual scale, according to which molecules or galaxies could plausibly seem to us like natural objects of sensory perception. In truth they are not. To imagine perceiving them that way simply projects onto them knowledge actually gained by means of telescopes, microscopes, and experimental apparatus—not with the natural senses. While constantly being improved, the resolution of such devices is limited by real physical factors.

The world can only be viewed through the cognition of some particular embodied agent, whether that is by means of natural senses or extended with technology. Yet, we are tempted to imagine an objective view as what the world would look like if no one in particular was looking. (By default, that tends to be simply the world as we humans perceive it, in contrast to the perception of other creatures.) Alternatively, we might be tempted to imagine it as the view of a disembodied agent, such as the mind of God. While an embodied organism is cognitively limited by its biology, an agent freed of biological constraints has no bodily need to conceive “reality” at all. A disembodied agent would have nothing at stake, no need for certainty, and no basis for decisions. While the organism is cognitively a prisoner of its biology, an agent completely freed of those constraints has no need to confront “reality” at all. It would seem to us to exist in the world, but neither the world nor its own self would exist for it.

What, then, should objectivity mean? Well, it could mean what is invariant among embodied cognitive systems. Mathematics, for example, may represent what is common to all human cognizers:, consisting of the most general features of natural reality as it appears to us. This would include features such as “objectness” (number), grouping characteristics (sets), and patterns that can be formulated as rules (axioms and operations). Such  features are characteristics of our particular environment, on a solid planet with liquid water and transparent atmosphere. They may also reflect features of our embodiment, such as hands that can manipulate and even manufacture “objects,” and minds that can perform manipulations on simulated objects. While mathematics is deemed to represent the most general properties of the world, science may depend more on the particulars of our embodiment. For example, the concept of force in physics—and the sensation of acceleration—are grounded in muscle proprioception.

If there is no absolute (read: disembodied) objectivity, can we at least define a relative one, in which one embodied perspective can be more objective than another? That seems to be one function of reflexive self-consciousness, which enables humans to transcend a given framework by stepping beyond it to a meta-level. “Thinking outside the box” is even a measure of intelligence as we conceive it: the ability to try a different approach in solving a problem, not to be stuck in an approach that doesn’t work. This ability to step back and take a broader view is a critical power of the cognitive self. Perhaps it is a measure even of how deeply that self exists. Descartes may have been mistaken to assume that the mere appearance of experience confirms the existence of a self. The fact that the world seems real is biologically built in, and is no guarantee of either objectivity or of the subjective self. On the other hand, the ability to question perception, belief, and apparent reality implies the capacity to self-transcend. That at least implies an additional level of agency.