| John Dilworth:
Research Overview 3. |
3. PRELIMINARY VS.
RESULTANT PERCEPTUAL REPRESENTATION
A. INTRODUCTION
Section 2 of this research overview provided three reasons why the interactive theory of perception (ITP) needs to be supplemented with a further theoretical structure involving two separable kinds of representational content--a Double Content theory of perceptual representation. Double content theory, with the aid of field theory, explains how and why aspectual content must be separated from the object-related perceptual content that is the predominant focus in cases of perception of objects that are not themselves representations.
But in addition, there is also a further dimension of considerations concerning the perceptual representation of properties of objects. Even after a cognitive double content analysis has separated out object-related from aspectual content, there still remains the issue of the worldly status of the relevant object-related content. There are two ways in which this object-related content F(x') could fail to represent an actual perceived object, plus some property it actually has. First, there may be no actual worldly object x that is in fact represented by content F(x'). And second, even if there is an actual object x that is thus represented, it may not actually have property F.
These two possible kinds of failure are related to the traditional distinction between perceptual hallucination (no object is seen at all) versus illusion (an object is seen, but with incorrect identification of one of its properties). However, that traditional contrast fails to adequately distinguish two significantly different epistemic stages in perceptual representation--what I shall here call preliminary versus resultant representation.
In the preliminary representation phase of perceptual processing, the cognitive system is applying further processing to the raw, object-related output F(x') of the initial double content processing. At the beginning of this phase, no decisions have been made as to the epistemic status of the putative object x' and the putative property F that object x' seems to have. For convenience, this initial content-based output F(x') will be described as a Putative Fact, in that, at the beginning of this preliminary stage, it seems to the perceiver as if there is some object x' that has property F--it seems to be a fact that object x' is F--whether or not there is an Actual Fact of some actual object x having property F that is currently being perceived.
In order to move from this preliminary, putative-fact-based representational phase to the final resultant representation phase of perception, a perceiver has to move from how things initially seem to her--a state whose content is the putative fact F(x')--to a state in which, in normal cases, she comes to believe that the putative fact is an actual fact. In this resultant phase the perceiver would, in typical cases, represent an actual worldly object x as having property F, as demonstrated by her F-related classification dispositions with respect to object x.
To summarize, from the broader perspective of the interactive theory of perception, double content analysis is a completely internal or preliminary kind of processing, in that its output is only a preliminary representation as discussed above, i.e., a putative fact of form F(x'). Nevertheless, in normal, non-hallucinatory cases, the resultant representation will be a state in which the perceiver Z becomes disposed--in view of the information provided by the putative fact--to classify some actually perceived object x as an F. Thus in normal cases, a resultant representation may be identified with the F-related classificatory dispositional state with respect to object x that is the relevant causal output as specified by the interactive theory of perception.
B. UTILITY OF THE CURRENT PRELIMINARY/RESULTANT (P/R) REPRESENTATION DISTINCTION
1) In distinguishing low-level animal representation from high-level, sophisticated propositional representation.
Fodor distinguishes low-level discrimination or sorting of objects from high-level conceptualized thinking about a kind of object. A parallel distinction can be provided using only the resources of the interactive indexing (II) theory as supplemented by two varieties of the current P/R distinction. In low-level discrimination the preliminary putative facts are relatively unconceptualized, and immediately cause resultant classification dispositions without any intervening epistemic processing. In contrast, in high-level propositional processing, relevant putative facts trigger sophisticated conceptual abilities that examine the epistemic credentials of how things seem to the perceiver, which strongly influence the relevant resultant classification dispositions. For further details see Secs. 6-8 of my draft paper "More on the Interactive Indexing Semantic Theory"
10/09 Draft: "More
on the Interactive Indexing Semantic Theory"
This article further explains and develops a recent, comprehensive semantic
naturalization theory, namely the interactive
indexing (II) theory as
described in my 2008 Minds and
Machines article "Semantic
Naturalization Via Interactive Perceptual Causality," (Vol.
18, pp. 527-546). Folk
views postulate a concrete intentional relation between cognitive states and the
worldly states they are about. The II theory eliminates any such concrete
intentionality, replacing it with purely causal relations based on the interactive
theory of perception (ITP). But
intentionality is preserved via purely abstract propositions about the world
that index, or correlate with,
appropriate cognitive states.
Further reasons as to why intentionality must be abstract are provided, along
with more details of an II-style account of representation, language use and
propositional attitudes. All cognitive representation is explained in terms of
classification or sorting dispositions indexed by appropriate propositions. The
theory is also related to Fodor's representational
theory of mind (RTM), with some
surprisingly close parallels being found in spite of the purely dispositional
basis of the II theory. In particular, Fodor's insistence that thinking
about an item cannot be reduced
to sorting dispositions is supported via a novel two-level account of
cognition--upper level propositional attitudes involve significant intermediate
processing of a broadly normative epistemic kind prior to the formation of
sorting dispositions. To conclude, the weak intentional realism of the II
theory--which makes intentional descriptions of the world dispensable--is
related to Dennett's 'intentional stance' view, and distinguished from strong
(indispensable) intentional realist views. II-style dispositions are also
defended.
2) In motivating a novel theory of truth, the Cognitive Correspondence Theory (CCT).
5/10 Draft: "Substantive
Truth via Sub-Propositional Correspondence"
Correspondence theories of truth are usually understood to hold that truth is a
matter of there being a correspondence between a complete truth-bearer, such as
a proposition, and some worldly item such as a fact. Nevertheless, there is an
ambiguity in the concept of a proposition, or propositional claim, that allows
for the possibility of a sub-propositional correspondence theory.
Intuitively, a proposition is true just in case the way it claims the world to
be corresponds with how it actually is. Understood sub-propositionally, the
relevant truth-condition is that a claimed or putative fact
corresponds with an actual fact. I provide some initial development of a
sub-propositional theory--the cognitive correspondence theory (CCT)--that
gives due recognition to the irreducibly cognitive nature of the relevant
concept of a putative fact as: how a worldly fact is conceived of, or
represented as, being.
Links:
Return to Home
Full Text Online Philosophy Journals Listing
Philosophy Dept. Faculty Page
Page Editor: John Dilworth
Last Updated:
June 04, 2010