Signals of Reality · Reality Is Not What You See

Reality: What Do We Actually See?

Seeing gives us access to the world through a biological process that selects, transforms, and interprets light. That makes perception fallible, not worthless.

A white shirt hangs partly in sunlight and partly in shade. The shaded cloth sends less light toward your eyes, yet it need not look like a different material sewn onto the bright half. You see a shirt with a shadow across it. That ordinary achievement contains a difficult problem: the light reaching you depends both on the surface and on what illuminates it.

Vision somehow has to work with the mixture.

This is a better starting point for asking what we see than the dramatic suggestion that everything is an illusion. Most of the time, perception lets us distinguish a step from a painted stripe, find a handle, and follow someone through a crowd. Its success deserves as much explanation as its errors. The interesting question is how a limited stream of physical signals supports such useful contact with a world larger than the stream.

Light arrives; a world appears

The front of the eye bends incoming light, and the lens helps focus it onto the retina. There, photoreceptors convert light into electrical signals. Signals pass through retinal circuitry and onward through the optic nerve; the brain participates in producing vision. This is a biological chain of transformations, not the delivery of a tiny finished photograph to an internal spectator. The National Eye Institute’s account of how eyes work describes the basic pathway.

The distinction matters because the object and the signal are different things. A cup on a table has a back, an interior, a weight, and a temperature. The light arriving from one viewpoint does not separately report all those properties. It carries information conditioned by lighting, angle, occlusion, and the eye’s sensitivity. Seeing the cup does not mean receiving an inventory of everything true about it.

Nor does it mean seeing only an unrelated private object. The light was affected by the cup. Moving your head changes the available view in ways constrained by the cup’s shape and position. Reaching for it can succeed or fail. A mediated relation can still be a relation to something outside you.

Consider how much the phrase “the cup is there” leaves open. You might locate it accurately while misjudging its size. You might recognize it while mistaking a reflection for a crack. Perception is not a single certificate of correctness attached to an entire scene. Different judgments drawn from the same view can have different levels of support.

Why interpretation enters

Imagine a simplified visual problem: a small dark patch in an image could be a dark surface under strong illumination or a lighter surface under weak illumination. One measured brightness cannot, by itself, settle which combination produced it. This is an ambiguity in the available evidence, before anyone begins arguing about philosophy.

Surrounding edges, changes across a surface, and clues about illumination can help. Researchers study how such contextual information contributes to perceived lightness. Michael Rudd’s 2014 cortical model offers one mechanistic account based on integrating information at edges. Its existence should not be confused with a final, universally accepted explanation of every lightness effect. The distinction between a phenomenon and a model proposed to explain it is essential here.

A second approach describes perception using probabilistic inference. In such models, ambiguous sensory evidence is combined with prior information about what is likely. David Knill and Alexandre Pouget’s review of Bayesian approaches explains why uncertainty is useful in computational accounts of perception and action. This is a modeling framework supported by particular findings, not proof that the brain consciously calculates probabilities or that every visual process follows one algorithm.

“Inference” can mislead if it suggests a person silently deliberating over every edge. The relevant processing generally does not wait for a verbal decision. You do not first see a list of brightness values, then choose to believe in a shirt. The organized scene is what ordinarily becomes available to experience.

That is also why learning the explanation for an illusion may leave the appearance intact. Knowing which interpretation is warranted and having an immediate visual impression are different achievements. Reason can correct the judgment without acquiring direct control over the appearance.

What an error can establish

Philosophers disagree about what perceptual experience fundamentally consists in and how it relates to the objects perceived. The Stanford Encyclopedia of Philosophy’s discussion of perception distinguishes several responses to the problems raised by illusion and hallucination. The neuroscience of signals does not, on its own, settle that dispute.

In particular, the fact that perception depends on a nervous system does not logically establish that the external world is unreal. An account of how access works is not automatically an argument that access fails. Conversely, everyday success does not establish that every apparent property belongs to an object exactly as it appears.

There is room between those claims for careful inquiry.

Suppose two people disagree about whether a mark on the cup is a crack. They can turn it, change the lighting, magnify the surface, or feel the mark. Each procedure introduces its own limitations, but it also makes different errors possible. Agreement across suitably different checks can strengthen the conclusion. Merely asking ten people to inspect the same ambiguous photograph may add less than one view from the other side.

This is a practical reason to value independent evidence. Confidence should grow because alternative explanations have become harder to sustain, not simply because the same appearance has been repeated.

The process does not supply a position outside all perception from which to certify the universe. It supplies something more modest and usable: ways to distinguish a claim that survives investigation from one that depends on a single misleading condition. Mediation does not flatten that difference.

The shirt remains a good example. A measurement of light from the shaded half and a judgment that the fabric is white can both be reasonable. They answer different questions: how much light reaches an instrument under these conditions, and what sort of surface is likely present. Confusing those questions can make an ordinary success of perception look like a scandal.

What we see is neither an exhaustive copy of the world nor an invention unconstrained by it. We encounter surfaces, distances, movements, and objects through capacities that have particular strengths and particular failure modes. The task is to learn when an appearance carries the information we need, and when to move, measure, or look again.

Further reading

The NEI introduction covers the biological pathway. Knill and Pouget discuss uncertainty in computational models; Rudd develops a specific lightness model. The SEP entry explains why the philosophical questions remain distinct from those mechanisms.

Share this article
Facebook WhatsApp X LinkedIn Medium
Quick navigationEsc