A Brief Introduction to the Guidance Theory of Representation



(11.2) It is a cognitive process and it (a) confirms a
representation that is in error3; or (b) disconfirms a
representation that is not in error; or (c) modifies a
value in a way that the subject later regrets; or (d)
recommends a course of action that fails.

(12) An action A fails in its intent because of R if, and
only if, (a)
A failed; and (b) A was taken with respect to
an entity
E; and (c) R provided guidance for A w.r.t. E;
and (d)
R has feature F; and (e) R with F represents that
E has property P; and (f) A failed because E was not P.
Note that the term “represents” in clause (e) is to be read
in light of the current theory of representation.

(13) A token representation R is in error for subject S
and action A in token circumstances C if, and only if, A
would fail because of R if taken by S in C.

The representation may be said to be in error for S
simpliciter if and only if the class of actions for which R
provides guidance in S’s circumstances C is dominated by
actions that would fail because of
R.

Comparison to Related Work

The guidance theory, broadly speaking, takes both a
naturalistic and a functional perspective on representation.
It is motivated by the same fundamental insight regarding
the epistemic importance of action and interaction as gave
rise to the theory of interactive representation (Bickhard,
1993; 1999). However, we offer a significantly different
development and formalization of this shared insight. For
instance, Bickhard’s analysis relies heavily on control
theory, cashes out representational content in terms of
‘environmental interactive properties’, and assumes some
version of process ontology. The guidance theory, while
compatible with these possibilities, does not require them.
Still, the relative advantages of these two analyses remain
largely to be determined. While there are many other
naturalistic theories of representation on offer, very few
adopt the functional perspective in as thoroughgoing a
way as we do. For instance, Dretske (1986; 1988) adopts
the functional perspective largely as a post-hoc fix to
what remains an information-content approach to
representation, so as to be better able to account for
misinformation. In contrast, Ruth Millikan does take the
functional perspective as the starting point for her theory
of representation, and the guidance theory thus bears the
most resemblance to hers (Millikan, 1984; 1993). Thus,
although the current article is meant only as a concise
introduction to the guidance theory, and is not the place
for any detailed comparisons with rival theories, it is
nevertheless worthwhile to say a few words about
Millikan’s theory in particular.

The resemblance between the guidance theory and
Millikan’s own biologically inspired theory is strongest
when she writes things like: “Cognitive systems are
designed by evolution to make abstract pictures of the
organism’s environment and to be guided by these
pictures in the production of appropriate actions.”
(Millikan, 1993:11) However, the impression of similarity

3 This clause in the definition is an embedded recursion, not a
circularity.

fades quickly as the details are examined. For while we
agree on this very general characterization of cognitive
systems, we differ as to the core point: that mental
representations must be pictures and, even when they are
pictures, we differ as to what makes such “abstract
pictures”
representations.

There are three main components to this very basic
disagreement. First, on our view, a given mental token is
a representation just in case it is standardly used by a
given organism to guide its behavior with respect to the
intended object; Millikan, in contrast, suggests that it is
only a representation if it is the result of (or consumed by)
a properly functioning system, performing the function it
was selected to perform: “It is not the facts about how the
system
does operate that make it a representing system
and determine what it represents. Rather, it is the facts
about what it would be doing if it were operating
according to biological norms.” (Millikan, 1993:10-11)

Second, and deeply related to the first, Millikan relies
heavily on the notion of such a “proper function” to
explain the possibility of representational error (a
representation is in error when the relevant
representation-producing or representation-consuming
system is not functioning according to biological norms).
In contrast, our theory allows for the possibility that a
system serving some function other than that for which it
was selected, or mal-functioning in some very lucky way,
could, in its use of mental tokens, be
representing just in
case (roughly speaking) the mental tokens in question
were being used to (successfully) guide the agent’s
actions with respect to the indicated objects. Rather than
analyze representational error in terms of mal- or non-
standardly-functioning systems, we cash it out in terms of
failure of action. Although we think representational
systems
did evolve, and attention to their evolutionary
history can help us understand how and why they function
as they do, we believe a system can sometimes
competently perform a function, including representing,
for which it was not selected, and in these cases its
unusual provenance should be no barrier to recognizing
this fact.

Third (and finally), whereas Millikan’s view of
behavior and action revolves around the function or
purpose of the organism or its parts (a movement by the
organism is only a behavior of that organism if it can (or
perhaps must) be understood in terms of the organism’s
proper function or biological purposes), our own
definition of action includes motor and cognitive
processes effected for a broader range of motivating
reasons. Although some element of teleology is
apparently necessary to ground the idea of a motivating
reason for acting, it is not clear to us that this must
necessarily be accounted for in terms of natural selection.
It could be the teleology of the subject itself, understood
as having a subjective purpose like maintaining its
homeostatic condition, pursuing hedonic value, or
maintaining adherence to a moral, political, or aesthetic
principle. A more detailed discussion relating the
guidance theory to some alternative theories, including
Millikan’s, can be found in (Anderson,
forthcoming-b).



More intriguing information

1. he Virtual Playground: an Educational Virtual Reality Environment for Evaluating Interactivity and Conceptual Learning
2. Public-private sector pay differentials in a devolved Scotland
3. The Complexity Era in Economics
4. The Employment Impact of Differences in Dmand and Production
5. Job quality and labour market performance
6. The name is absent
7. The Economics of Uncovered Interest Parity Condition for Emerging Markets: A Survey
8. The name is absent
9. The name is absent
10. Road pricing and (re)location decisions households
11. Momentum in Australian Stock Returns: An Update
12. The name is absent
13. Evolving robust and specialized car racing skills
14. The name is absent
15. CREDIT SCORING, LOAN PRICING, AND FARM BUSINESS PERFORMANCE
16. On the Relation between Robust and Bayesian Decision Making
17. The Importance of Global Shocks for National Policymakers: Rising Challenges for Central Banks
18. The name is absent
19. Life is an Adventure! An agent-based reconciliation of narrative and scientific worldviews
20. DIVERSITY OF RURAL PLACES - TEXAS