ONT Re: interpreter & interpretant

[Jon Awbrey <jawbrey@oakland.edu>]

¤~~~~~~~~~¤~~~~~~~~~¤~~~~~~~~~¤~~~~~~~~~¤~~~~~~~~~¤

JA = Jon Awbrey
SR = Seth Russell

JA: I have even found myself accused of the dreaded sins
    of "agentism", "individualism", "psychologism" for
    doing no more than referring to the "interpreter"
    in the very same way that Peirce did, as a soppy
    embodiment of, and a way-station toward the more
    fundamental, but more difficult to explain,
    notion of an interpretant sign.

SR: What is an 'interpretant sign' ?

Shifting to Ontology List, as my ticket
to the Midway has been punched for today.

The term "interpretant" is short for "interpretant sign",
and so the two terms are commpletely synonymous.  It is
the name of a particular "slot" or a "relational role"
in a sign relation.  I typically write the three slots
in the order <o, s, i>, but the order is not all that
important so long as you always know what's what.

Here is what I personally consider to be the clearest and
the completest of Peirce's definitions of a sign relation:

| A sign is something, 'A',
| which brings something, 'B',
| its 'interpretant' sign
| determined or created by it,
| into the same sort of correspondence
| with something, 'C', its 'object',
| as that in which itself stands to 'C'.
|
| CSP, NEM 4, pages 20-21, & cf. page 54, also available at:
| http://members.door.net/arisbe/menu/library/bycsp/L75/L75.htm

More detail here:

http://suo.ieee.org/ontology/msg02121.html
http://www.chss.montclair.edu/inquiry/fall95/awbrey.html

SR: I would like to tease out a acceptable ontology of these concepts:

    1.  interpreter
    2.  interpretant
    3.  memories
          (examples might be
           a psychological state or
           a computer data base record)

SR: I think that we already have a acceptable concept of an Agent;
    in that I doubt that many ontologists would reasonably disagree.

!?

SR: Assuming this, then can't we say that an Agent
    is composed of an interpreter and memory states?

Calling something an Agent tells you next to nothing.
You have to specify "Agent of What" and then provide
a model of the process that is carried out by this
Agent of What.  I'm in systems engineering, so the
names "agent", "system", "representative point" are
all just equivalent and dummy terms for whatever it
is that moves through the states of a trajectory in
the relevant state space.  You really know little
until you define the state space and the laws of
time evolution or temporal transition through it.

For me, memory is an "aspect of state", in other words,
if the typical state x is in the state space X, then
the "memory subspace" M is specified by a projection
m : X -> M that tells you how much of the state is
used to store memory information.

SR: If the Agent is a running piece of software on a database of sumo:Formula,
    then it's interpreter would be a sumo:Procedure and it's memories might be
    sumo:Formula.

SR: Like in this picture:

SR: http://robustai.net/mentography/interpertant4.gif

¤~~~~~~~~~¤~~~~~~~~~¤~~~~~~~~~¤~~~~~~~~~¤~~~~~~~~~¤