perceptions as unconscious influences

Philosophers have generally considered visual perceptions to be closely related to, or even to be samples of, surfaces of surrounding objects. Thus, vision was thought to be rather like smell, and as direct and immediate as touch. However, with the invention of the camera obscura, and the related discovery that the object world is imaged optically in the eyes, it became clear that patterns of light in the eyes (retinal images) are transmitted to the brain by coded electrical signals (action potentials), which are then, somehow, read as objects having very different and far richer properties than the optical pictures in the eyes. It became clear, at least to physiologists, that there must be a great deal going on in the brain in order to produce perceptions from sensory signals: that our perceptions are created in us, and that they may be very different from the corresponding objects of the external world, as described by physicists. Furthermore, there are illusions, of many kinds, which are hard to reconcile with a 'direct' or 'immediate knowledge' account of perception.

An alternative account — very largely due to Hermann von Helmholtz (1821–94), is that perceptions are conclusions of unconscious inductive inferences. Introducing this idea, he compared visual perception from images with the arbitrary signs of language, and how we learn language. Thus, Helmholtz (1866) points out that, as with meanings of words, 'the concept of the normal meaning of frequently repeated perceptions can come about with immutable certainty, lightning speed and without the slightest meditation'. Helmholtz continues:



Helmholtz stresses that the association of word with meaning, and of sensation with meaning in perception, comes from regular experiences of the connection with no (or few) exceptions. In this way, meanings are built up inductively (from many instances to a conclusion that is not logically necessary), both for language and for perceptions. These associations are, presumably, built up largely by interaction with objects — so that, for example, the patterns of the grain of wood become associated with a hard substance which can be dropped without breaking, and which can be cut with a saw; and the transparency of glass becomes associated with potentially dangerous brittleness; in consequence, we behave very differently with wooden as opposed to glass objects. Helmholtz suggested that we come to our ideas of the physical form of objects inductively, by combining visual experiences from many viewpoints, following the rules of perspective. Comparing these 'inductive conclusions' with the scientific method, Helmholtz says: 'Inductive conclusions are never so reliable as well-tested conclusions of conscious thought. ... False inductions in the interpretation of our perceptions we tend to label as illusions.' Of these he says: 'Obviously, in these cases there is nothing wrong with the activity of the organ of sense and its corresponding nervous mechanism which produces the illusion... . It is, rather, simply an illusion of the judgement of the material presented to the senses, resulting in a false idea of it.' He goes on to say: 'These unconscious conclusions derived from sensation are equivalent ... to [his italics] conclusions from analogy.' He attaches a lot of weight to active structuring of perception, which is especially evident in conditions of dim illumination, or when complex crystals or other structures are viewed stereoscopically:





In this connection, phenomena of ambiguity (which occur especially with vision and hearing) are clearly very important: 'Without any change of the retinal images, the same observer may see in front of him various perceptual images in succession, in which case the variation is easy to recognize.' For these and other reasons, Helmholtz thinks of perception as given by learning, and as empirical. It is not passive acceptance of stimulus patterns, but rather projection (though not merely geometrical projection) from internally organized knowledge of objects and processes. In current terminology, this may be termed use of stored knowledge 'top down', for interpreting or reading sensory signals, as originating from particular objects. Similarly, we project our meanings of words on what we describe. (The emphasis on processes searching for the best reading, or interpretation, on the available evidence, we call 'active', in comparison with 'passive' accounts, such as that of J. J. Gibson.)

It seems a natural development to extend Helmholtz's account by calling perceptions (which, not altogether happily, he called 'perceptual images'), perceptual hypotheses (Gregory 1970, 1981). This suggests useful analogy with hypotheses in science and the ways they are developed, used, and tested. That is not to say that perceptual hypotheses and scientific hypotheses are formed or used or tested identically, but they do seem to share striking similarities. Since the methods of science are open to inspection, they may provide a basis for considering processes of perception, which (as Helmholtz realized) are exceedingly hard to discover by direct methods, because they are hidden in brain processes that are only beginning to be understood, and also because they are unconscious.

(Published 1987)

— Richard L. Gregory

Bibliography
• Helmholtz, H. von (1866). Treatise on Physiological Optics, vol. iii (3rd edn.). Trans. and ed. J. P. C. Southall, 1925, section 26. (Repr. 1962.) (Also given in Warren, R. M., and Warren, R. P. (1968), Helmholtz on Perception: Its Physiology and Development.)
• — —  (1896). 'The origin of the correct interpretation of our sensory impressions'. In Treatise on Physiological Optics (2nd edn.).

---