There appear to be at least two general avenues of access to non-propositional approaches toward understanding thinking or thought processes. Both are rooted in the theory of semiotic in the sense of the theory of signs advanced by C. S. Peirce:
(1) PEIRCE'S THEORY OF SIGNS: According to Charles S. Peirce, a sign is something that stands for something (else) in some respect or other for somebody. There are three basic kinds of signs, namely: ICONIC SIGNS, which stand for something (else) by virtue of resembling that something in some respect or other; INDEXICAL SIGNS, which stand for something (else) by virtue of being causes or effects of that something; and SYMBOLIC SIGNS, which stand for something (else) by virtue of some habitual or conventional association. Icons thus include photos, statues, drawings and paintings when they are realistic or naturalistic, etc. Indices include ashes in relation to fire, fire in relation to ashes, red spots and an elevated fever in relation to the measles, etc. Symbols include red lights at intersections, words in natural languages, sentences and paragraphs, and such.
(2) MINDS AS SEMIOTIC SYSTEMS: By inverting and generalising Peirce's theory of signs, I have introduced a theory of mind, according to which minds are the kinds of things that can use signs; hence, minds are semiotic (or ``sign-using'') systems. I take it that there can be at least three basic kinds of minds, namely: ICONIC MINDS, which are capable of using icons; INDEX- ICAL minds, which are capable of using icons and indices; and SYMBOLIC minds, which are capable of using icons, indices and symbols.
There also appear to be at least two other ``higher-order'' kinds of minds, namely: transformational mentality exhibited by the capacity for (formal) deductive and inductive reasoning and metamentality exhibited by the capacity for using signs of one kind to stand for signs of another, especially in criticism in suggesting how signs might be used more efficiently, effectively, or reliably to achieve their purpose.
The key point relative to the current discussion is that these approaches emphasise the role of icons (things that look like, taste like, smell like, feel like or sound like other things) as fundamental to thought processes of all kinds, where higher modes of mentality (including both indexical and symbolic) presuppose iconic mentality. Moreover, iconic mentality already presupposes a point of view, in the sense of how things are to be taken in regard to ways in which they resemble one another. A system is CONSCIOUS (with respect to signs of specific kinds) when it has the ability to use signs of those kinds and is not inhibited from the exercise of that ability. COGNITION occurs when a system that is conscious (with respect to signs of those kinds) encounters signs of specific kinds within a suitable causal proximity, where COGNITION is the effect of a causal interaction between those signs and the system's other inner states, which may include other mental states, motives, beliefs, etc.
These considerations suggest that the notion of a PROPOSITION may harbor ambiguity. Propositions, presumably, are assertive and correspond if not to declarative sentences themselves then to the meanings that they assert. Images, smells and the like are not inherently assertive, however, even though the effects of COGNITION that their presence may bring about (within a CONTEXT consisting of a cognitive system's other relevant states) will tend to have an assertive character. Consider, for example, seeing someone coming toward you who looks like a friend, yet upon closer examination, you discover it is someone else instead. Initially, the image interacted with your CONTEXT to produce the impression that your friend was in the vicinity as an initial cognition, but subsequently a more detailed image interacted with your CONTEXT producing the impression that it was someone else instead as an alternative inference. Notice that we are thinking in terms of images, smells, and such even though they do not - as icons per se - have an assertive content.
These considerations may also shed some light on recent work in this general area. A catalogue from MIT (just received) offers the following prose in promotion of STARMAKING:
The title alludes to Goodman's famous defence of the claim that because all true representations of stars and other objects are human creations, it follows that in an important sense the stars themselves are made by us. More generally, the argument moves from the fact that our right representations are constructed by us to the claim that the world itself is similarly constructed.Icons and indices, however, are natural signs in the sense that the resemblance and cause-or-effect relations upon which their representational capacity depends are not created by us and exist apart from any sign user. By contrast, their use as signs only depends upon sign users having noticed those relations (of resemblance or of cause-or-effect) in order to use them as signs. Symbols thus appear to be the only kinds of signs that must be created rather than discovered, but even here it makes little or no sense - and is generally false - that what symbols stand for are created by us as mere constructions of the human mind.
Hopefully, more adequate distinctions (such as those that the semiotic conception provides) will enable us to deal with nonsense like these promotional claims - which appear to have an appeal to the public and even to some professionals which is far greater than their actual significance would warrant - and to better understand the nature of thought processes. The images starlight creates in the retina, for example, occur as an effect of a causal process that no more qualifies as ``our creation'' than do resemblance relations between specific images that are created by light from different stars. In case any of you may be inclined to explore this approach in greater detail, therefore, I list the following among other references.
James H. Fetzer