From newshub.ccs.yorku.ca!ists!helios.physics.utoronto.ca!news-server.ecf!utgpu!csd.unb.ca!morgan.ucs.mun.ca!nstn.ns.ca!news.cs.indiana.edu!att!princeton!shine.Princeton.EDU!harnad Mon May 25 14:06:45 EDT 1992
Article 5799 of comp.ai.philosophy:
Newsgroups: comp.ai.philosophy
Path: newshub.ccs.yorku.ca!ists!helios.physics.utoronto.ca!news-server.ecf!utgpu!csd.unb.ca!morgan.ucs.mun.ca!nstn.ns.ca!news.cs.indiana.edu!att!princeton!shine.Princeton.EDU!harnad
>From: harnad@shine.Princeton.EDU (Stevan Harnad)
Subject: Re: Grounding: Virtual vs. Real
Message-ID: <1992May21.045844.2833@Princeton.EDU>
Originator: news@ernie.Princeton.EDU
Sender: news@Princeton.EDU (USENET News System)
Nntp-Posting-Host: shine.princeton.edu
Organization: Princeton University
Date: Thu, 21 May 1992 04:58:44 GMT
Lines: 126

I'll try to answer some of the questions that have been raised
about grounding, why real transduction is essential, etc.:

(1) HUMAN TTT: The only grounding worth talking about is the grounding
of the symbols inside a robot that is capable of passing the Total
Turing Test (TTT), i.e., a robot capable of all the sensorimotor and
verbal interactions with the real world of objects, events and states
of affairs that we are capable of -- Turing-indistinguishable from us
in that respect, as a matter of fact.

(Animal TTT's would be nice, and logically prior to human TTT's, but
for methodological reasons, we cannot administer them with either the
ecological or the intuitive sensitivity that we have for administering
the human TTT.)

Note that the symbols of the robot are grounded in its TTT capacity;
remove the TTT capacity and you remove the grounding.

(2) DEDICATED SYSTEMS: So let's not talk any more about "grounded"
type-writers and "grounded" computers. Neither type-writers nor
computers can do what it takes to pass the TTT. They can't see things
and they can't manipulate them. Only robots are relevant.

(Nonetheless, as with the animal TTT, it would be useful to approach
the human TTT by a series of approximations, including "animal"
TTT-robots and even noncognitive "grounded" dedicated symbol systems
with limited robotic capacity. None of these, however, could have the
logical or methodological force of the human TTT, whose first "T"
stands for Total, not partial or approximate, which would defeat the
whole purpose of the TTT.)

(3) REAL TRANSDUCERS: Transduction is the transformation of energy
from one form into another. The specific kind of transduction that
is relevant to TTT robots is sensorimotor transduction: They must
be capable of receiving the "shadows" of objects cast on their sensory
surfaces -- and then taking it from there.

(The question is begged if the assumption is made that transduction is
trivial, just a thin surface that immediately goes into symbols:
Everything inside a robot or ourselves could in reality be analog all
the way through, or, as I believe, hybrid, with the symbolic part
grounded bottom-up in the analog).

But the minute you take away a robot's transduction capacity, you've
removed it's grounding, because a robot without sensorimotor capacity
cannot pass the TTT (hence it does not HAVE TTT capacity).

(4) "BRAIN IN A VAT": This is the point where people usually start
to think of brains in vats, but real brains in vats are not computers
with their trivial peripherals detached: They are mostly analog
re-projections of the sensory and motor surfaces. It must be left
COMPLETELY moot what would be left of the insides of a TTT-capable
robot if you removed its transducers. What DEFINITELY cannot be
presupposed (without begging the question) is that there would just be
a computer left in there -- and this is what most of the confusion
about virtual grounding and virtual reality arises from,

(5) TTT CAPACITY. Note that the criterion has always been TTT capacity
-- nothing about its causal history, nothing about the nature of the
actual objects in the world. A grounded (TTT) robot must simply be
capable of performing in the world in a way that is indistinguishable
from the way people do.

The human brain is grounded (because of whatever properties, so far
unknown, that give it TTT-power): It clearly has the capacity to pass
the TTT when attached to a musculoskeletal system. Its sensorimotor
"tranducers" are arguably a part of the peripheral nervous system, and
the brain consists of the peripheral and central nervous system. I have
no idea what a central nervous system without a peripheral nervous
system can do, and I don't care, because we know too little about what
EITHER of them actually does to put any substance in arguments based on
brains in vats.

So perhaps I should say the human BODY (brain and all) is grounded, to
short-circuit further sci-fi fantasies. It's grounded whether what is
stimulating the skin or the eyes is the shadow of a real object or just
computer generated video and vibrators, and whether the hands are
manipulating real objects or just joysticks connected to virtual
objects. The human body is grounded if it can pass the real TTT; and
even if from birth it's been exposed to nothing but artificial,
computer-generated stimulation, it's still grounded IF IT CAN PASS THE
TTT.

The exact same thing is true of a grounded TTT robot: If it does have
the CAPACITY to pass the TTT in the real world, it makes no difference
if it's sat all its life in a lab getting artificial input (to its
REAL transducers). The grounding comes from the CAPACITY, not
its exercise or the objects on which it is exercised (and of course
the TTT is just a TEST of that capacity -- let's not mix up the
capacity itself with our empirical criteria for verifying it's
there).

(6) HOW/WHY DOES TTT POWER GROUND SYMBOLS? Because in an ordinary,
ungrounded symbol system, whether it is static, like a book, or
dynamic, like a computer, symbols are manipulated only syntactically
(based on their shapes), yet those symbols are systematically 
interpretable as MEANING something, i.e., they have a semantics.
But that semantics is ungrounded, because it depends on our
interpretation. Independent of the interpretation we project on
it, there are no "meanings" in a book or a computer, because books and
computers are not the kinds of things that anything means anything to.

The TTT puts a second set of constraints on the symbols in a symbol
system, over and above the constraint that (i) they must be
interpretable, by us, as being ABOUT something (which is what makes
them a symbol system in the first place, rather than random gibberish).
That second constraint is that (ii) the system (now a robot) must be
able to pick out (discriminate, categorize, manipulate, identify and
describe) the objects that its symbols are interpretable as being about
in a way that is coherent with the interpretation and indistinguishable
from the way we do it: The meanings of its symbols are then grounded
directly in its robotic capacity rather than just parasitic on the
meanings we project unto them.

Is it now clearer why in order to keep this grounding intact, we cannot
eliminate the transducers and the TTT capacity and return to just a
symbol system that will merely bear the weight of a systematic
interpretation (e.g., as a virtual robot and its virtual world)? That's
all just ungrounded symbols again! Real TTT capacity is needed, and
sensorimotor transducers (and probably a lot of other analog structures
and processes) are essential for that.

-- 
Stevan Harnad  Department of Psychology  Princeton University
harnad@clarity.princeton.edu / harnad@pucc.bitnet / srh@flash.bellcore.com 
harnad@learning.siemens.com / harnad@elbereth.rutgers.edu / (609)-921-7771


