[From: Bruce Nevin (Tue 930221 15:57:31 EST)]
( Bill Powers (950318.0100 MST) ) --
Mary and I have just seen the second of a PBS series on language
I saw part of this, by dint of mightily resisting my annoyance.
The argument for Universal Grammar (UG) as reflecting an innate,
biologically inherited mental organ specialized for language (part of the
postulated modularity of mind that I call neo-phrenology), goes roughly
as follows:
1. Look at how complicated language is.
2. See how many of the apparent complexities can be explained as outcomes
of abstract structures and rules manipulating them.
Notice that children overgeneralize valid rules by analogy ("look at the
gooses!" on the analogy of regular plurals like boxes), and then notice
that they do not make other analogical generalizations that might on the
face of it be plausible, but--aha!--no rules postulated for UG support
those other generalizations, so this confirms the psychological reality
of the abstract structures and rules of UG. An example they gave went
something like this:
I ate ham with eggs : I ate ham and eggs
What did you eat ham with? : *What did you eat ham and?
(The asterisk marks something that isn't acceptable as a sentence.)
3. There simply isn't enough data in the uses of language that children
hear for them to be able to arrive at these abstract structures and rules
by empirical induction. (This is the "paucity of stimulus" argument.)
4. And anyway infants and children don't have the intellectual capacity
to make the kinds of analyses and inferences that are involved.
It is true that the description of language in Generative Linguistics is
pretty hairy. However, most of the hair is inside the heads of the
practitioners. There is at least one alternative way of describing the
same observations without the abstract structures, where the rules and
operations affect words and word shapes, and to some extent perhaps
classes or categories of words (though that may just be an artifact of
how the rules are implemented in the domain of words). The same
regularities are reached by simpler means. The unacceptability of
utterances like "*What did you eat ham and?" has a simpler explanation.
Chomsky, however, has condemned approaches that are "data bound",
extolling the value of abstractness as a virtue in its own right. (That
this is rhetorically convenient passes unnoticed. More on this later.)
The demise in this way of (1) and (2) leaves (3) without foundation.
And anyway, the standard account is as though children learned language
in a social vacuum, ignoring work like e.g. _Child's Talk_ by B***
(sorry, lost his name: my synapses too, Bill, I guess!) on the ways in
which children learn through using language interactively with their
elders and peers to accomplish things. Staring the Generativists in the
face is one of their own causes celebres, the case of Genie, the little
girl who grew up in closeted isolation. She was exposed to language on a
television near her, but there was no interaction, there was no occasion
for her to use language to accomplish anything, so she didn't learn it.
C'mon! Why didn't her UG click in?
We actually don't hear much about (4) any more, as research into the
intellectual capacities of infants and children has pretty much scuttled
this claim. It had more prominence in the Generativist literature of the
1950s and 1960s.
What seems not to be
considered is that these underlying abilities that seem inborn may
reflect more basic properties of the brain that are indifferent as to
whether they are used to build language or to accomplish anything else.
Yes.
I heard on NPR of a report in Nature of research showing differences
between men and women in brain scan images while they were judging
whether or not pairs of unfamiliar (but familiarly spelled) nonsense
syllables rhymed. As I recall, a much larger area of the brain was
involved in women. The statement was made that women and men were doing
the same tasks, but using different neural structures to do them. There
was speculation whether this might have something to do with women's
better recovery rate from strokes.
Specialty by specialty, each person who studies human
behavior sees the ability to carry out that class of behavior as
something specifically existing inside the brain, like a specialized
organ.
Chomsky introduced this notion of innate ideas (or reintroduced it,
calling it forth, as it were, from its 17th-century grave) in the 1950s
and 1960s. I'm not aware of it's being taken up by sports psychologists,
and so on, but maybe so, what do I know. Chomsky's gotten away with it
because no one could prove him wrong. This is partly because he is such
a master of debate. It is partly because--and this is part of his
mastery of rhetoric--he keeps the terms of debate at a level of
abstraction that is very difficult to ground in testable data. And it is
partly because scientific approaches to language did not reach as far and
could not address the same issues, and philosophical approaches to
language had their own extraneous baggage in the way. (Chomsky, last I
heard, regards Generative Linguistics as a branch of philosophy with
affiliations to cognitive psychology, and says that it is pre-scientific.
I agree.)
As I mentioned, there was at least one exception to the generalization
about scientific approaches to language. This is the work of Zellig
Harris, from whom Chomsky learned about transformational grammar, then
went his own way with it. Harris never controverted Chomsky's views
because, consistent with his perception of what it is to be a scientist
and to do science, he avoided controversy and polemic; Chomsky has
thrived on controversy and polemic--and on Harris's negligence. Harris
died not quite three years ago.
I see language as an invention that caught on and spread like wildfire.
It is extraordinarily easy to learn and handle not because of any
particular language-related processes in the brain, but because the
capacities of the brain make language easy to learn and do
Yes. You would appreciate, I think, Harris's speculative suggestions
about the origin and evolution of language. They appear in his last
book, _A theory of language and information: a mathematical approach_,
(1992, Oxford), and in briefer and perhaps more accessible form in
_Language and information_ (1988, Columbia U. Press).
There is no reason that language could not have been independently
invented in more than one place and time. The evidence (existing
languages from which shared ancestor languages can be reconstructed, from
which further ancestors might be reconstructed) is disappearing
every year as minority languages cease to be used and their remaining
speakers die out.
ยทยทยท
-=+=-=+=-=+=-=+=-=+=-=+=-=+=-=+=-=+=-=+=-=+=-=+=-=+=-=+=-=+=-=+=-=+=-=+=-
If there are social
rules, for example, they exist only because people are capable of
perceiving rules and making behavior conform to them. This property is
totally independent of WHAT rules are put into effect.
I agree with this.
I observe a great concern for what the rules are in young children. I
rember this was mentioned in the PBS segment on language that I saw.
(Hagiography of Ferdinand de Saussure, with his notion of language as
being game-like. Images of children playing various games.)
There is something left out, however. It is implicit in the phrase
"capable of perceiving rules and making behavior conform to them."
Through history and for as far back in prehistory as we can project, back
to some postulated bootstrapping process at their origins, such as Harris
has described for language, it appears likely that every child learning
the ropes has perceived "rules" (speaking loosely) as something already
in existence, and has perceived others as "making their behavior conform
to them" (or sometimes making their behavior flout them).
(It is true that children may be told what a rule is, but most often they
infer it, and where there is a disparity between example and precept they
follow the example, making their own generalizations about what the rules
are, including precepts that are selectively followed.)
This is the capacity we have for "perceiving rules". But the rules are
pre-existent as an emergent property in the conduct of the people that
matter to the child. The child does not make them up out of whole cloth.
This means that the rules may be observed as perceptible human artifacts
in the world. It is true that we are projecting generalizations across
the behavior of individuals, behavior by means of which those individuals
autonomously control their private perceptions. But our justification is
that this is what people do. This is what the people do whom we are
observing.
-=+=-=+=-=+=-=+=-=+=-=+=-=+=-=+=-=+=-=+=-=+=-=+=-=+=-=+=-=+=-=+=-=+=-=+=-
A clue as to why people have cultures and languages and primates do not
came up in a National Geographic series on primate research that I saw a
week or two ago. They had shown some chimps using sticks to poke through
holes in a "beehive" and suck off honey. In the next segment, the
researcher had some candy on a table surrounded by some sort of cage with
vertical bars. He had a primitive kind of rake with three wide-spaced
tines resting on the table, the handle through the bars. The chimp had
the same, on an adjacent section of the table. If the tines were down,
they couldn't get the candy. The researcher demonstrated flipping the
rake over. With the back of the rake on the table top, it was easy to
pull the candy to one's hand. No matter how often he saw it, the chimp
didn't get it. The inference from this and other experiments was that
the primate perceives use of the tool to accomplish the end, but does not
attend to the manner of using the tool.
In the next segment, there is a similar setup with a small child. The
child fails with the tines down. Then the adult demonstrates flipping
his rake over. The child immediately, on the next turn, does the same,
even moving the rake through a curving path very much like that executed
by the adult's rake, to capture the object (a block) and pull it back to
the edge.
Doing things in the same manner as one's fellows do them. Restricting
the available degrees of freedom, reducing the variability of means for
accomplishing the given end. And then recognizing another as a fellow,
as a peer, as a member of the same social group, by the familiarity of
their mannerisms. In language, a great deal of variability is possible
in the pronunciation of words. Restricting the available degrees of
freedom in a particular way, consistently across the vocabulary, is
perceived as having a particular regional or social accent if it differs
from how you do it; it goes unnoticed if it is the familiar way that you
do it.
-=+=-=+=-=+=-=+=-=+=-=+=-=+=-=+=-=+=-=+=-=+=-=+=-=+=-=+=-=+=-=+=-=+=-=+=-
But "social rules" are problematic to observe, precisely because of how
we perceive them and make our behavior conform to them. It is
exceedingly difficult to be "objective" about these observations.
Here's why, I think, or one kind of reason why.
* I need my rules for my own conduct to be self-consistent, so that I
don't experience conflict. I try to make a coherent system of them. I
project this system and its coherence as attributes of society.
* I want others to follow the same rules that I do, and I want to follow
the same rules that are followed by the people that matter to me,
because the rules provide me means for eliciting the cooperation of
others and avoiding conflict with others.
That is a two-edged sword if I ever saw one. On the one hand, when I try
to bring the rules to awareness and state them explicitly in language, I
tend to see the conduct of others as instances of the rules that I know.
(I am helped in this by my vocabulary, whose words "mean" for me just the
perceptions and perceptual values that I already know them to mean.)
When I perceive some bit of personal conduct as an exception to a rule, I
see the person as doing wrong, or as not knowing the correct rule. But
if I discover that it is my conduct that is the exception, I am very
likely to change my perception of what the rule is. Either way, I have a
lot of wishful thinking going for me.
This notwithstanding, useful work has been done. It should be possible
to look at the literature of anthropology and sociology and ask some
questions that are useful for PCT. For example, are there any
constraints on possible rules for social arrangements? Given certain
identified choices, are certain other choices unavailable?
Here's a movie plot in Japan. Young man and young woman fall in love.
His family arrange his marriage with a different person. After struggle
with feelings and values, young man tries to elope with young woman. We
see this as a romantic story with the young man as the hero overcoming
the rigid prejudices of a too strict family and seeking happiness and
fulfillment, as we all ought. In Japan, this is a tragedy, the young man
failing to conduct himself properly, and failing to fulfill his
obligations as we all ought. Those words--failing to conduct himself
properly, failing to fulfill his obligations--do not convey the complex
system of obligations that the Japanese perceive as part of the reality
into which they are born and come of age. (See Ruth Benedict, _The
Crysanthemum and the Sword_, the book she wrote instead of a book on
synergy, and then she ran out of time.)
-=+=-=+=-=+=-=+=-=+=-=+=-=+=-=+=-=+=-=+=-=+=-=+=-=+=-=+=-=+=-=+=-=+=-=+=-
The error here is to assume that for every regularity we see there must
be a corresponding cause -- the opposite of "systems thinking." If, for
example, we observe that there is a perception of social responsibility,
the assumption is that there must be some specific reason for this
perception, some social force or inherited goal of being socially
responsible. I'm not saying that you, Bruce, are necessarily making this
mistake -- the above suggests that you are trying to avoid making it.
But any concept that describes the properties of societies as anything
but emergent is surely making it.
Emergent in evolutionary origin. Emergent because there is and can be no
control system superordinate over the individuals who simultaneously
learn and recreate it. But the two cases differ because people leave
tracks and because people follow in another's footsteps. Even if their
construal of their perceptions is, to the mythical objective eye, as
foolish and misguided as Pooh Bear and Piglet tracking the Woozle round
and round the spinney (their own tracks), the need to make a construal,
and the process of doing so, is something essential about being human
for which PCT must account.
-=+=-=+=-=+=-=+=-=+=-=+=-=+=-=+=-=+=-=+=-=+=-=+=-=+=-=+=-=+=-=+=-=+=-=+=-
We talked a while back about changing habits and programs.
In early January, I started changing my way of tying my shoes.
To get a square not rather than granny not, so that the loops lie across
the instep instead of up and down, the second or loop half of the knot
has to be tied the reverse of the way that the initial overhand knot was
tied. If the left-hand strand crossed over the right to start the
initial overhand knot, then the right-hand loop has to cross over the
left to start the second overhand knot and complete the square knot.
It is easier for me to do the former than the latter--my left hand
passing the strand across to my right.
This became apparent when, a number of years ago, I learned from my wife
Sarah's grandmother a better way of tying a double knot (that doesn't
easily come untied). As a child, I had learned to tie yet another
overhand knot with the loops. This had the disadvantage of no longer
being a slip knot, harder to get untied when you want to. For a "Grandma
'Kee knot", you pass one loop around the other twice, through the same
opening, then pull it tight. This is a slip knot, but the double wrap
keeps it from coming loose accidentally.
My habitual way of tying my laces was to do the initial overhand knot the
easy way (left hand passing to right, which then manipulates the turn).
That meant that the double wrap of the loops was up to my left hand, and
that was definitely not the easiest way.
For about a month, I had to consciously think about doing the reverse of
my usual method. I noticed what came up naturally as the first move, and
deliberately did the opposite.
Around the middle of this month, February, I went through a period of not
quite a week of being at a loss which way to begin. I had to think
through the rationale, looking ahead to which way I wanted the second
half of the knot to come out, and starting the initial overhand knot the
opposite way.
This weekend, I noticed that I now start the desired way automatically.
-=+=-=+=-=+=-=+=-=+=-=+=-=+=-=+=-=+=-=+=-=+=-=+=-=+=-=+=-=+=-=+=-=+=-=+=-
One morning this weekend, I had come in from feeding and watering the
animals. I picked up my robe and slippers where I had left them after
showering in the downstairs shower and getting dressed, earlier. The
slippers were in my left hand, and my robe slung over my left arm. But I
hadn't yet taken off my jacket. When I went to the closet to hang it up,
I couldn't take it off in my usual way by grasping one sleeve with the
other hand. I imagined shrugging the jacket off my shoulder over my free
hand, but anticipated that wouldn't work well because it's a down jacket,
not heavy enough to fall readily off a shrugged shoulder. In imagination
I experienced myself vigorously shrugging my shoulder and flailing my
arm. Not very comfortable. So I paused. Then I noticed that my left
hand, without my conscious choice, had transferred half the burden (the
slippers) to my right hand. So I pulled my right sleeve down off my
shoulder with my now freed left hand, transferred the slippers back, and
slipped my right arm out; transferred both objects to my right and
completed the divestment.
As I reflected on this, it seemed that surely there were more efficient
ways to do it. But something happened without my awareness, which
changed the conditions of the problem, and I just went from there. I
wondered if perhaps the transfer of my slippers to my right hand was an
instance of reorganization.
-=+=-=+=-=+=-=+=-=+=-=+=-=+=-=+=-=+=-=+=-=+=-=+=-=+=-=+=-=+=-=+=-=+=-=+=-
I have a clear recollection of there being a prior post from you, Bill,
responding to my sketch of linguistic procedures, but it has disappeared
from sight.
This sketch was a restatement of some of the initial procedures
identified by Harris in the mid 1940s, published finally in his 1952 book
_Methods in Structural Linguistics_. It sets the conceptual and
methodological groundwork for the work that I would like to do in PCT
phonology. An actual proposal would have to start from there--about all
I said was that we need means to disturb parameters in a way that
participants can resist, controlling a perception of speech.
I tried to lay these preliminaries out clearly because last year we got
stuck on whether people have perceptions of contrast, whether they're
just controlling perceptions of words and the appearance of their
controlling contrasts is just a side effect, and so on.
I remember that this first post had contact info for the sound blaster,
and you said you would send something else about it. Thanks--much
appreciated!
I also recall that you made some comment about the standard set of
phonological features used in Generative phonology. In my effort to be
concise I led you to think that I embrace these wholeheartedly. Sorry.
I don't. I share your qualms about premature closure as to what are the
controlled perceptions.
A problem varying parameters of a speech synthesizer is that they may not
be just what you can vary with the organs of speech. The intermingling
of sensors, effectors, and environment within the vocal tract seems well
beyond our present means to model directly (just the number and
relationships of muscles in the tongue makes a "little mouth" model seem
a distant prospect), so we have to make do, but there will be a
persistent problem interpreting results: just what is it that a subject
has controlled, and what relation does it bear to what is controlled
during speaking?
-=+=-=+=-=+=-=+=-=+=-=+=-=+=-=+=-=+=-=+=-=+=-=+=-=+=-=+=-=+=-=+=-=+=-=+=-
( Bill Leach 950217.20:10 EST(EDT) ) --
Thanks for your concern, Bill. I think we're OK.
-=+=-=+=-=+=-=+=-=+=-=+=-=+=-=+=-=+=-=+=-=+=-=+=-=+=-=+=-=+=-=+=-=+=-=+=-
And now I must be quiet for a while.
Bruce