Scientific rumors

[From Bill Powers (950113.0930 MST)]

Bruce Abbott (950112.0830 EST) --

RE: movement control

What you have in this literature is a story that is mostly unsupported
by evidence and strongly guided by the way these people think behavior
OUGHT to work. There is, of course, evidence, but much of it is like a
piece from a jigsaw puzzle: a great deal of imagination is required to
make it seem to go with other pieces. Also, the language of neurology
has a nice stock of terms that sound important but mean nothing. For
example, you cite this:

     These cells receive visual, orbital, and neck afferent information.
     The integration of the information from these different sources
     generates a neural code representing the location of an object with
     respect to the body and the head (Andersen et al. 1985b).

Both sentences are probably true. It is probably true that parietal
cells receive signals from those other areas. It is probably true that
Anderson et. al. said what they said. But what is "integration?" What is
a "neural code?" What is "representation?" And when a cell "receives
information," what kind of information is received? The time at which a
signal occurred? The rate of change of some variable? The logical
relationship between two variables? A weighted sum? When information is
received, what is done with it? Is it stored in a card file until
someone can read it?

There is a whole vocabulary for talking about processes in the brain
that allows a person to construct bridges between isolated observations
without seeming to leave any gaps. This is a skill much like the one
that an experienced public speaker can draw upon when suddenly asked to
say a few words when he isn't sure where he is, who the audience is, who
the other speakers are, or what they have been talking about.

"My friends -- and I am sure that in this distinguished gathering there
are many whom I would be pleased to count as friends and allies -- the
occasion that brings us together is an important one, and one that well
deserves marking with ceremony. There are those who say that assemblages
of this sort are a waste of time and money, but all you have to do is
note the distinguished personages both behind the podium and in the
audience to see that responsible people disagree. There is more than
reputation and empty show here; we have a purpose in meeting, a goal, an
objective, a reason. It is the same sort of reason and purpose that has
brought souls together throughout history. I am reminded of what Abraham
Lincoln said about Alexander Hamilton, on the occasion of his wife's
birthday. Honest Abe was leaving the Illinois State House when he was
accosted by ....."

And there you have two minutes of a stirring speech being put together
on the spot while the speaker scans his audience for informative
reactions. The speaker assumes that everyone else knows the subject
matter and what has been said; he counts on them to supply their own
confirming details that will put meaning into his generalizations. With
any luck he will leave the audience with the impression of having heard
something relevant and important, even though he has not actually said
anything at all. "So let us go on, with hope for the future and
appreciation of the past, but most of all remembering what Abe Lincoln's
wife said: it's the present that counts (friendly laughter). Thank you."

Then we have the trick of converting an opinion into a fact:

     Psychophysical observations by Morasso (1981) have suggested that
     this planning stage is carried out in extrinsic coordinates that
     represent the motion of the hand in space.

Note that the observations suggested this, with Morasso merely serving
as the deliverer of the message. It would be much less impressive to say
"A guy named Morasso thinks that ...". This is part of a very old
tradition in science, where the scientist is merely a trained observer
who relays the self-evident facts to those who can use them. When an
author cites someone in this manner, Morasso (1981), the rule is that
you are supposed to assume that the author has read the citation,
understood it, and verified that the conclusion is inescapable. Thus the
mere mention of the name and the fact that the opinion has been
published is supposed to lend more weight to the author's use of the

In fact, it often happens that when you look up the citation you find
the same statements followed by (Jones et. al. 1980), which leads to
(Petersen 1975) and so on, with nobody actually taking responsibility
for the "fact." This is especially true of statements about feedback and
control processes. The authors, not having any first-hand acquaintance
with control theory, cite conclusions published by well-known names in
the field who do not have any first-hand acquaintance with it, either,
thus turning the process of citation into a sort of formalized rumor

What matters in peppering a paper with terse citations is to show that
you are a member of a vast group in which these matters have been
considered and conclusions have been reached on which we can build
further. What these people who are cited actually said, and whether
their conclusions were forced on them by observation or merely invented
to try to make sense of confusing data, is of no importance. What
matters is to give an impression of solid backing of one's own guesses.

I suspect that this is an important reason for the resentment that
springs up so readily when PCTers say that basic ideas in traditional
behavioral sciences are mistaken. If that's true, it removes the
comforting support of all these citations and leaves one standing alone
before the world, faced with the necessity of defending one's guesses
without help from authorities.

Just think how many arguments in how many papers would collapse if it
were generally known that Taub's de-afferentation experiments show that
deafferentation completely ruins motor control! Taub was actually trying
to disprove a claim that nobody had ever made: that without feedback,
there can be no motor action. The fact is that deafferented organisms
can still tense their muscles and cause limb movements. But they have
lost all kinesthetic control. Nevertheless, we have Bizzi citing
deafferentation experiments to show that feedback isn't necessary!
That's how the scientific rumor mill works.



Bill P.