[From Bill Powers (920304.1000)]
Bruce Nevin (920304.1035) --
I hope all unusual circumstances are being coped with; nice to hear from
you again.
Fuzzy Logic:
I saw a demo of the fuzzy logic inverted pendulum controller. Pretty poor
control, as you mention. If there's a place for this approach, it has to be
at the higher levels -- I certainly wouldn't use it in place of analog
variables and quantitative continuous controllers (in other words, you're
right about its applicability to my description of "richness"). I have a
suspicion that fuzzy logic represents a step away from the purely digital
concept of brain function, caused by the discovery that logical systems
don't control continuous variables very well. Unfortunately, analog
concepts of simulation aren't taught along with digital concepts: if you're
adept at digital logic, you probably "don't do analog." If students became
adept at both approaches, I think they would find it much easier to know
when analog is appropriate, and when logic is needed for simulating higher-
level processes.
···
------------------------------------------------------------------
Language:
It appears that we keep the complexity of language control out of
awareness for good reason. There is just too much going on at once.
In order to describe it to ourselves and talk about it to one another >we
must use that which we would talk about. But using it and making it
wiggle real slow under controlled conditions so we can get a good look
are mutually exclusive.
This is what I was asking about. The problem here is that there isn't any
experimental or observational test to see whether the proposed invisible
processes are actually occurring unless you specifically, and slowly, make
them occur. You seem to be saying that the expansions are really taking
place, in words, but REAL FAST and not in awareness, during ordinary
speech. How do you tell whether this is a right, as opposed to merely
plausible, description of what is going on?
I think we can boil this down to a specific dispute. What I DO experience
is that a terse sentence results in incomplete perceptual meanings and that
I then complete them to eliminate dual or contradictory meanings. Having
done that, having perceived the ambiguity or incompleteness and having
remedied it in imagination, I then have many choices as to how to describe
the more complete meaning in detail.
You appear to be saying that the terse sentence is first recognized as
incomplete in terms of linguistic rules -- this word requires two arguments
and only one was supplied, and so on. Then the phrase is expanded to supply
the arguments, after which the phrase evokes the completed or unambiguous
perceptual meaning. An alternative is that linguistic analysis results in
recognition that a word is missing, after which a search for words with
acceptable meanings is done, selecting a missing word appropriate to the
meanings already evoked. In either case, the linguistic analysis occurs
before the completed meaning is experienced.
When I hear a sentence like "Rino Sanders hit safely," I don't think that I
first realize that "Rino Sanders hit" requires another word. I think that
the first result is an image of Ryne Sanders hitting -- swinging a bat --
and (when the object of hitting isn't mentioned, or even if it is
mentioned, before the name of the object is heard), the crack of the bat
against A BALL, the scamper for first base, the fielder getting the ball
back too late ("hit safely"), and so on. Bingo, the whole little scenario
pops into place and I know what the words mean. I don't detect any
intervening linguistic analysis of any flavor. In fact, if the sentence is
"Rino Sanders hit a car," I have to retract the whole immediate perceptual
interpretation and supply a different one, with a sensation of a big error
at the instant the word "car" is heard.
You seem to be saying that between "Rino Sanders hit .." and the appearance
of a completed perceptual meaning in my head, there is an ultra-rapid
invisible linguistic process that supplies a missing word in order to
satisfy the requirement of two arguments, and only then does the image of a
BALL enter my experience.
I would argue that the missing ball is supplied first, and that the
motivation for supplying it is that the meaning of hitting is incomplete
without the image of something doing the hitting and something being hit. I
think you are proposing that the missing word is noticed first, the
motivation for supplying it coming from the general requirement that the
operator in question is known to require two arguments, only one of which
has been supplied.
Is this what we're arguing about?
----------------------------------------------------
Sentence deconstruction:
(5) I'm trying to hear the processes going on in present time.
I recognize I as a word that can't be an operator--it can't be said
"about" some other word. This is a zero-order word, in OG parlance.
Does this recognition come before the meaning of "I" as the sense of a
person?
I recognize 'm trying as a reduction of am trying.
Does this recognition come between hearing "I'm trying" and getting sense
of a person trying to do something?
... I recognize try as a word that can be said "about" two other >words.
Two other words must be present (said or zeroed in a >reconstructable way).
Since the "other word" here is a whole phrase, "to hear the processes going
on in present time," or at least "to hear," are you saying that you get no
meaning out of "I'm trying" until something being tried is imagined or
heard?. Try it on this: "I'm wondering ............. what you mean." During
the dotted pause, doesn't "I'm wondering" convey any perceptual meaning?
I recognize that I is probably the first argument word that must be
present in order for try to be said. I recognize an operator-argument
dependency between try and I.
What would you think of this:
I immediately suggests a person; try immediately suggests the person in
process of doing something unspecified. This knowledge, that something is
not yet specified, is immediate and perceptual. In a formal linguistic
analysis, we say that the person and the trying are to be classified as an
operator and an argument. The sense that something is still missing,
perceptually, suggests to us that another argument is called for by this
particular operator-word. This in turn suggests a rule: some operators
require two arguments. The logic then follows: this is such an operator; it
has less than two arguments; therefore there is a missing argument to be
supplied.
Here's an analogy without the linguistic analysis: You see a picture of a
man leaning far forward holding onto a rope that goes backward over his
shoulder and then straight horizontally out of the frame. Without talking
about it, you know that there is something out of the frame resisting the
pull of the rope. Maybe the other end is tied to a post, or a donkey, or
another man pulling the other way. So you can see the man, and you can see
that he is pulling, but you can't see what he is pulling on. Nevertheless,
you know that he is pulling on SOMETHING, via the rope. If the man is
progressing forward, you wait in anticipation of seeing what the rope will
pull into the picture.
Now you could create a description of what you see: the man is pulling with
the rope. The omission of WHAT he is pulling is clear, not because of the
words, but because of the picture and its wordless sense of something on
the other end of the rope resisting the pull, which you imagine. You could
say "The man is pulling ...." and wait for whatever is on the other end of
the rope to appear before you name it: "... an elephant!" Or you could use
a word that refers to a missing perception, an error signal: something.
What led you to expect something to appear? Was it the knowledge that the
name of this action is an operator requiring the name of an agent and the
name of an object? Or was it the wordless comprehension that there was
something outside the picture resisting the pull? I suggest it was the
latter: you can't know what kind of operator "pulling" is until you see its
perceptual meaning. There's nothing in the word itself to indicate how many
arguments it takes, or even whether it's an operator or an argument.
Before you can say whether an operator-word requires one argument or more,
you have to know what specific perception is meant by it. "Shout", if it
refers just to a vocal spasm, takes no argument. "The man shouted hello" is
clearly an operator that takes two arguments. It isn't the word, but the
perceptual meaning that "takes arguments." And the arguments "taken"
aren't words, but perceptions of configurations, events, and so on that are
related by the action described. Words are classed as operators and
arguments because of the way their meanings are related; their meanings are
not related as they are because the words are classified as they are, but
because the perceptual world works the way it does.
------------------------------------------------------------------
This leads me to an observation that's been a-building for some time while
I lacked a way to talk about it. It's the difference between a descriptive
rule and a prescriptive rule.
Here's a neutral example of a descriptive rule: Bodes's law. Take the
numbers 0,3,6,12,24 ... (doubling each time after the three). Add 4 to get
4,7,10,16,28 ... . Divide by 10; the numbers are very nearly the distances
of the planets from the Sun in astronomical units (Earth's distance =
1.00). When Uranus was discovered, it was found at 19.19 A.U., where the
rule of Bode's Law said it should be at 19.6 A. U. This led to a hunt for
planets at the missing position, 2.8, between Mars and Jupiter, and the
discover of Ceres, the first known asteroid. Pluto was discovered at 39.5
A.U., fitting Bode's Law which predicted it at 38.8. Unfortunately, Neptune
comes between Uranus and Pluto (most of the time), between two numbers in
the sequence.
This rule is very impressive, except that there's no known reason WHY the
planets should approximate this series -- except perhaps that their
distances are bound to approximate SOME series, which someone was bound to
discover.
An example of a prescriptive rule: a knight moves two squares ahead or back
and one sideways, or two squares sideways and one ahead or back. Why does
this rule hold true? Because behavior in chess is deliberately adjusted to
keep the moves of knights in conformity to this rule. If someone moves a
knight in any other way, the opponent or any onlooker will cry "Illegal
move!" and force the move to be corrected. We therefore observe that
knights always move according to this rule.
Neptune does not "obey" Bode's Law, but this "illegal position" does not
give rise to any action that makes it legal again. So Bode's law only
describes. The movements of chess pieces do, literally, "obey" rules, for
if a move violates a rule it will be set right. The rules of chess are
prescriptive.
Now the question I have for Bruce Nevin and Avery Andrews is this: are the
rules of grammar that you are helping to develop descriptive, or
prescriptive? In short, is there an underlying system that forces language
to exhibit these rules, or are these rules like Bode's Law -- interesting
but fortuitous approximations with no necessary basis in the natural world?
It seems to me that if any set of rules is prescriptive, we will find that
deviations from the rules are corrected. Would this not be a way of
choosing between rival grammars?
----------------------------------------------------------------
Best,
Bill P.