At its heart, the concept of a motivating reason is
deeply tied to concepts of rational interpretation like the
one found in Daniel Dennett’s description of the
intentional stance (Dennett, 1987). We take no position
here on the basis of, or constraints on, any specific
standards of rational interpretation.
Focus
As it is used above, the idea of an action’s focus is
intended to express a functionalist concept. When a
subject is performing an action it places itself into a
potential feedback loop with its environment. Its purpose
is to monitor the result of the action and to plan
adjustments to its course of action.
(7) The focus of an action is the ultimate entity being
monitored through the feedback channels taken to provide
indications of its status.
A subject may monitor the focus directly, or indirectly
by monitoring the status of some entity being used as an
indicator of facts about the focus. Because indicators are
made part of an extended guidance control system,
indications about the focus will cause in the subject
beliefs, decisions or equivalent states about further
appropriate actions or perhaps that action may cease.
When the focus is monitored through an indicator the
subject may have an indirect causal connection to the
focus or even no causal connection at all. An example of
an indirect causal connection to a focus would be an
engineer monitoring a gauge that is itself monitoring
engine pressure. Examples of focii to which there is no
causal connection are things like the time of day or a
mathematical operation on numbers. To monitor the first
we might monitor an indicator like a clock face and to
monitor the second we might monitor a progression of
numerals manipulated according to established rules. In
both of these cases the focus of the action is something
that is not present and to which the subject is not even
indirectly causally connected, but which can be monitored
nevertheless, despite the lack of causal causation, by
establishing a connection to something else that can be
manipulated to vary systematically with facts about the
focus.
Identifying the focus of an action in a given case
requires establishing the facts about what the subject is
monitoring in its circumstances, and understanding these
facts in terms of the subject’s motivating reasons.
Assumption of Information
An assumption of information is to be cashed out in terms
of facts about the actual operation of the representing
system (or subject). Beginning with an example will make
the concept easier to grasp. Imagine a computer
processing a user’s command to print a document. To do
this, the computer must determine to which printer it
should send its own commands. To guide this action, the
computer reads several character strings contained on its
hard disk, one identifying the printer and others with other
information about the printer. These strings guide it
regarding where it should send its print commands and
what protocol it should use to communicate with the
printer. From the perspective of the guidance theory, here
is the key fact: these character strings represent what they
do both because of the circumstances in which the
computer is reading them and also because of the
assumptions built into those circumstances. The computer
processes the strings as if they conveyed information
about the printer to which it sends its commands and
which communication protocol it should use. There is no
regress involved in claiming it makes this assumption,
because the assumption itself is not a matter of having
representational content. There is no representation inside
the computer with the content: I assume that this string
has information about the printer. Even more strongly, its
ability to make an assumption of information does not
require that the computer actually possesses information,
nor that it ever did.1 In the case described, the character
string the computer accesses could have been placed on
the disk via the output of a random number generator and
by coincidence be effective in directing it to the printer.
Even were that to be true, the string still would be
providing guidance and the computer would still be
making an assumption that the string contained
information about the correct printer. Therefore, the
ability to make an assumption of information does not
require an ability to have or obtain information.
Rather, the assumption of information about the printer
is a matter of know-how that is built into the architecture
of the computer: how it accesses representations, in what
circumstances it accesses them, how it reads and
interprets their structure, what actions it initiates and
monitors upon accessing them, how those actions cause it
to interact with the world, and so forth. We can provide a
candidate analysis of this know-how. To do this, we first
need to define, for any given token, the class of actions it
supports. The class of actions a token T supports is
relative to the kinds of circumstances C where the system
is prepared to use the token for guidance. It consists of all
the actions the system can initiate or modulate in C due to
its processing of T. Let us label this class of supported
actions Asupp.
(8) An action A is a member of the class of actions,
Asupp, supported by a token T used by a subject S in
circumstances C if, and only if, S in C would use T for
guidance regarding the initiation or manner of execution
of A.
We should think of the actions in Asupp as focus-neutral
descriptions of an action in need of association with a
focus in particular initiations. So, for example, if in some
circumstances a system is prepared to use a token for
guidance in running, the action running is the focus
neutral description. If the specific initiation of this action
occurs when the focus of the action is a bear, the focus-
neutral action “running” is initiated as the focus-specific
action “running away from a bear.” Actions obtain a focus
in the way discussed above.
Furthermore, since subjects do not initiate actions at
random, for each action in Asupp, there will be a (possibly
1 This, assuming that possessing information depends on causal
history and connection, which may not be the case.