[ExI] Symbol Grounding: Sets, Searle, and the Seat of the mind

[Will Steinberg]

Thinking about all this symbol manipulation and Searle stuff.

Imagine there are two general and distinct sets of information in the human
brain--nominal and sensory objects.  Nominal objects are what we think of as
true objects like an ice cream cone or a block of wood.  Sensory objects are
descriptive objects we use to describe nominal objects--milky, white, sweet.

When experiencing something for the first time, we only have sensory
objects.  Our sense organs provide information to the brain, and if this
information does not mathematically translate into a sobject we already
have, it uses something (be it an algorithm or an RNG) to create a string
representing it.  This also necessitates the creation of a new nobject, if
this sobject is found to not be associated with any nobjects (which it is
not, because the sense is a new one).  Therefore we have established rules
already: to physically identify and confirm the newness of nobject, we first
must be sure the sobject is new.  If it is not, some prior nobject has
produced it.


In this way the relatedness of nobjects can be approximated by sobjects in
common. Milk is associated with sobjects {white, liquid, milkflavor,
milksmell, usually cold} as well as many society-produced sobjects that are
closer to abstract notions than senses.  Ice cream, having sobjects {white,
milkflavor, milksmell, sweet, solid that melts, always cold} can be mapped
to milk using a few strong ties, notably the powerfulness and uniqueness of
the milk-something sobjects (showing that uniqueness of the sobject is in
direct correlation to ease of identification) and the use of
symbol-manipulating guidelines (i.e. a cold solid often will produce a
liquid because it is a frozen form.  We learn this early in life.  In fact,
we even must learn the guideline that sharing of sobjects can be a measure
of relatedness.  These guidelines, since they are learned, constitute a new
sort of processive object--adhering to our nomenclature, a probject.

Coming to things like Wernicke's and Broca's areas: a failure Wernicke's
area produces semantic failure, meaning the algorithm to correctly select
symbols has been compromised.  It would be interesting to note whether
substituted symbols are produced by a semi-predictable pattern or whether
they are a more general shuffling of symbols; this could offer a clue as to
how the strings for each symbol are arranged.  A failure in Broca's area
means the algorithm to correctly arrange symbols has been compromised,
perhaps leading to the conclusion that each object has a syntactical tail of
information responsible for placement.    It is almost like looking at
portions of DNA responsible for sorting.

This is the reason why Searle's CRA fails.  To approximate a human, the man
needs more than to internalize the rules; he needs to have a base object
manipulation method that can combine a changing bank of these three objects.

So a conversation with a human unfolds as such:

A: Hello. [Conversation starter a la methionine)
B: Hello. [Recognition of conversation]
A: How is your wife? [Analyzed--"How is" is a probject mapping nobject wife
to sobjects happy, sick, pregnant]
B: She is well. ["She is" is the language produced in response to probject
"How is" + sobject "feminine" associated with nobject "wife"; "well" is
sobject currently associated with "wife"]
A: That is good. ["That is {sobject}" is part of a societally mediated
probject used in order to communicate empathy.  A series of probjects, given
inputs "wife" and "well" figure that, will make person A produce a socially
acceptable response.]
B: *Checks Watch.* Oh!  I've got to be going! [simple time check algorithm,
checks against sobject "meeting at ten" associated with nobject "meeting",
uses probjects to realize that sobject "my watch says 9:50" means person A
must leave now in order to get to his meeting]
A: Alright.  Goodbye. [societal probject maps sobject "heard: 'I've got be
be going' ", parses it into understandable sobjects, and produces a societal
response.]
B: Goodbye [Recognition again]
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.extropy.org/pipermail/extropy-chat/attachments/20091220/7d8b1e32/attachment.html>