April Fool

[From Bill Powers (960403.0400 MST)]

Rick Marken (960402.1230) --

     My comments about the "uselessness" of conventional data were based
     on Bruce's April Fool's post where a conventional analysis of
     behavior (with no test to determine controlled variables) was used
     to reject a control model in favor of a reinforcement "model". I
     thought Bruce's post was a nifty (and sly) way of showing how easy
     it is to come to the wrong conclusions about behavior (like the
     conclusion that behavior is controlled by its consequences) when we
     look at behavior through the "wrong pair of glasses" (conventional
     rather than PCT).

One of the essences of a well-done April Fool's joke is that it has to
be presented believably. The Taco Liberty Bell joke was a classic: the
news report had Mary and me jerking upright and saying "What? They did
what? WHAT?!" The way this is done is to present "facts" that are
unbelievably shocking with such a straight face that the listener
accepts them as true. We both truly thought, for a few minutes, that
Taco Bell had bought the Liberty Bell and was going to move it to
California to use as their logo. We had been well and truly HAD.

Look at what Bruce said:

     The study was cleverly designed to pit reinforcement against
     control in the following way. Responses produce immediate
     reinforcement via delivery of a food pellet. However, those same
     responses act on the schedule to reduce the rate at which food is
     delivered over time. The animal can increase the overall rate of
     food delivery by reducing its rate of lever-pressing or,
     conversely, can decrease the rate of food delivery by increasing
     its rate of lever-pressing.

     If reinforcement is at work, the immediate delivery of food
     following a response should increase response rates; if control is
     at work, the animal should adjust its rate of responding downward
     to produce an optimal rate of food delivery, perhaps to the point
     of not responding at all on the lever. At this point in the study
     all the rats are responding at high rates, thus lowering the rate
     of pellet delivery far below the rate that would exist in the
     absence of responding, and far below the rate at which they consume
     food pellets when these are freely available.

At this point the control theorist is saying "What? They did what?
WHAT?" Note what Bruce is saying: the less the animals press the bar,
the more food they get. So if they don't press tbe bar at all, they get
the most food of all -- this is known as the free-feeding level. And if
they _do_ press the bar faster, they get less and less food until
finally they're getting none at all and pressing as fast as they can.

But look at the conditions:

     Responses produce immediate reinforcement via delivery of a food
pellet.

If this were true, then the faster the animal pressed the bar, the MORE
food it would get. But the immediately following statement is

     However, those same responses act on the schedule to reduce the
     rate at which food is delivered over time.

This clever "report on experimental conditions" actually contradicts the
first statement: how can the "rate of food delivery" decrease if the
response rate increases and every response produces an immediate
reinforcement? If you miss this contradiction, you have been set up to
accept as a fact the solemn announcement that

     At this point in the study all the rats are responding at high
     rates, thus lowering the rate of pellet delivery far below the rate
     that would exist in the absence of responding, and far below the
     rate at which they consume food pellets when these are freely
     available.

Diabolical! Masterful! Here, presented in the form of a carefully worded
scientific report, is a bald-faced LIE. There is no such study going on.
There are no rats responding at a high rate to get less food than they
would get without responding. In short, NONE OF THIS HAPPENED, just as
Taco Bell never bought or offered to buy the Liberty Bell.

Many years ago, Isacc Asimov, who was once a real biochemist, published
a fact-article in Astounding Science Fiction under the title, "The
endochronic properties of resublimated thiotimoline." He reported
finding that when thiotimoline was repeatedly purified by suhlimation
into a vacuum and recondensation on a cold plate, the time required for
it to go into solution when distilled water was added to it (the
"solution time") continually decreased. The interesting result that
appeared late in this study was that after a certain number of
resublimations, the solution time went to zero, and then became
negative. In short, the thiotimoline went into solution _before the
water was added_. This led to a series of interesting studies, ending
with one in which the experimenter went through all the motions of
adding water, but on a random schedule either did or did not add it. The
result was _neurotic thiotimoline_ which no longer knew whether to go
into solution, and finally refused to dissolve at all in any amount of
water.

An astonishing number of readers (well, perhaps not so astonishing
considering the venue) simply refused to believe that this article was a
hoax, even after John Campbell went to great lengths to assure them that
Isaac was joking. Many wrote in asking where they could buy
thiotimoline. After Campbell's explanation, people wrote furious letters
accusing Campbell and Asimov of hiding the truth, or of having given in
to Government restrictions and being forced to recant. Conspiracy
theorists went berserk. Those who fell for the joke the hardest
completely lost their senses of humor, and utterly refused to accept the
humiliation.

This article was one of my formative experiences, because I was hooked
until almost the end of it. Fantastic, I thought as I read. Marvellous!
Incredible! Time-travel is here! And when I finally realized that none
of this was real, I was embarrassed, humiliated, and crushed at my own
credulity. I was, as unbelievable as my friends, parents, and teachers
might have thought the notion, humbled. And I think I resolved somewhere
about that time never again to believe anything just because someone
told me in convincing terms it was true. That was when I started to grow
up.

The April Fool's joke, along with many other forms of artful and
entertaining lying, is probably a form of game that people have invented
to help themselves get over a natural tendency to believe anything they
are told. The game is to see how far one can lead other people into
believing more and more fantastic tales before they catch on. An
essential part of the game as children play it (although Taco Bell and
Bruce Abbott tastefully refrained from this) is the gloating that takes
place after the joke is revealed. The point is not just to humble the
credulous, but to rub it in and point derisively and indulge in coarse
laughter, all with the aim of producing the maximum embarrassment in the
victim, and the maximum schadenfreude in the perpetrator.

This tactic, I suspect, has survival value: it teaches the victims to be
properly skeptical and to consider starting to think for themselves. It
is also highly revealing of human frailty, for the results are not
always a shamefaced laugh and an admission of having been fooled. The
victims often nurture long-lasting resentment, and worse, they often
refuse to believe that they are the victims of a joke. If the story they
have fallen for is incredible enough, outrageous enough, sufficiently at
odds with common sense, they become incapable of admitting that the
story is only a story. They become True Believers and Defenders of the
Faith.

When this happens, we get groups of people telling each other the same
outrageous stories over and over, and _believing them._ Such people have
gone past the point of being able to admit that they have been tricked.
They keep increasing their investment in the April Fool's story until
the price of admitting even to themselves that the whole thing was a
made-up hoax has become greater than they are able to pay.

The April Fool's joke is strong medicine. The best ones play on our
hidden beliefs and intellectual dependencies; they mock institutions by
taking advantage of our belief in them, stretching that belief farther
and farther until -- in most cases, I hope -- something snaps. The most
effective April Fool's jokes are those that convince us of their truth
just because they are so very incredible. Reliable sources report that
the Pope has tearfully confessed to being a heroin addict. Mother Teresa
has been indicted for child abuse. The FBI is adding contraceptives to
the water supply in predominantly black cities. President Clinton keeps
a harem of six prostitutes on the third floor of the White House, and
Hilary hired them. Bob Dole's right arm is perfectly OK.

You can tell any story at all, if you have the knack of appearing
truthful, and people will believe you no matter how shocked and
horrified they are. They always knew, they say, that something like that
must be going on. Nobody, they say, could lie about something like THAT.
And now that you mention it, they heard something that absolutely agrees
with the story; the truth will out.

All this, of course, makes science a difficult process.

···

-----------------------------------------------------------------------
Best,

Bill P.