[Wittrs] Re: some helpful guidelines for reading Wittgenstein's philo...
- From: Glen Sizemore <gmsizemore2@xxxxxxxxx>
- To: Wittrs@xxxxxxxxxxxxxxx
- Date: Sun, 9 Aug 2009 10:57:19 -0700 (PDT)
--- On Sun, 8/9/09, Glen Sizemore <gmsizemore2@xxxxxxxxx> wrote:
From: Glen Sizemore <gmsizemore2@xxxxxxxxx>
Subject: Re: [Wittrs] Re: some helpful guidelines for reading Wittgenstein's
philo...
To: Wittrs@xxxxxxxxxxxxxxx
Date: Sunday, August 9, 2009, 1:12 PM
The reply below was accidentally sent, so I will continue with what I was
saying...
--- On Sat, 8/8/09, Stuart W. Mirsky <SWMirsky@aol. com> wrote:
From: Stuart W. Mirsky <SWMirsky@aol. com>
Subject: [Wittrs] Re: some helpful guidelines for reading Wittgenstein' s
philo...
To: Wittrs@yahoogroups. com
Date: Saturday, August 8, 2009, 10:02 PM
--- In Wittrs@yahoogroups. com, "gerardoprim" <gerardoprim@ ...> wrote:
>
> (Stuart) When I was driving up through the Carolinas and suddenly realized
> the meaning of a sign that had initially seemed incomprehensible to me, I
> suddenly had certain images in my mind. There were no behavioral changes. I
> just kept driving. (The wiper blades and lights were already on.) But at one
> point I didn't understand and then I did. What occurred is not explainable as
> behavior unless you so broaden the term of "behavior" as to make it
> meaningless.
> (Gerardo) Why do you say that the broader concept of behavior would be
> "meaningless" ? It will be meaningful as long as it´s part of a language
> game, and new language-games come into existence all the time (PI 23). Also,
> cognitive science has changed the traditional grammar of many concepts (e.g.
> "representation" , "cognition", "computation" ), and possibly you wouldn´t
> say those concepts became "meaningless" . The narrower meaning of behavior is
> "publicly observable activity", the broader meaning includes also "privately
> observed activity". There´re many similarities that allow the extension of
> the concept, e.g. there´s a degree of public observability (e.g. talk aloud,
> talk whispering, talk privately to ourselves), conditioning principles can be
> applied to both overt and covert activities, covert activities are previously
> learned as overt behavior (covert reading comes from previous overt reading,
> mental math calculation comes from previous overt
math calculations, imagery comes from previous perception, self-talk comes from
previous overt talk), etc. You can say "I don´t like it" or "I´d prefer another
word", but you cannot say it´s "meaningless" .
>
> Regards,
> Gerardo.
>
SWM: Yes, you're right Gerardo. It has meaning the way you put it. But it still
fails a test that I think is important here. Terms do change all the time.
Language isn't static though it usually changes fairly slowly. However it is
possible to so broaden what we mean by a term as to extract it from its
original uses. If "behavior" means what is observable in a being capable of
acting (though we can and do talk of the behavior of rocks and gases, etc.
though that is irrelevant to behaviorist psychology) then to add in "and I also
mean all the mental events the being is having that prompt him/her/it to act or
that accompany the said acts or that may presage the entity's acting under
certain conditions (when a response is called for), then it no longer is about
the acts of the entity under study. It is now about the very things the usual
idea of behavior does not apply to.
GS: I find this argument somewhat bizarre. First, though, you slipped up above
when you said "If 'behavior' means what is observable in a being capable of
acting... [then private events are outside of this definition], " you must have
meant "If 'behavior' means what is PUBLICLY observable in a being capable of
acting..." Skinner's behaviorism (radical behaviorism) holds that the sorts of
events being called "private events" are, in fact, observable, but they are
observable only to the behaving individual. This, incidentally, does NOT
include much of what is called "mental" since much of what is called "mental"
is not observable even to the behaving individual! But more importantly, the
argument you use has the curious effect of cutting off a behavioristic account
of certain phenomena. You are, in effect, saying that if behaviorism attempts
to deal with certain issues it ceases to be behaviorism, and thus behaviorism
can have no account of certain
phenomena! But behaviorism is the view that behavior is the proper subject
matter of psychology, and that it can be a complete view, leaving nothing out.
It holds that what is "spected" in introspection IS behavior (or, perhaps,
stimuli generated by behavior). Hence the term "radical behaviorism, " where
"radical" means "thoroughgoing. " Unlike Watsonian behaviorism (I make a
distinction between Watsonian and methodological behaviorim), which denies the
existence of "mental imagery*," and methodological behaviorism, which
acknowledges the role of subjectivity but rules it out of scientific
consideration, radical behaviorism argues that what is observed in
introspection IS behavior. Thus, rather than being a sort of concession to
mentalism, what one gets is a kind of "super behaviorism" (i.e. RADICAL
behaviorism) . Radical behaviorism is MORE behavioristic than Watsonian or
methodological behaviorism - especially methodoilogical behaviorism which is, in
fact almost indistinguishable from mentalism. As Gerardo has pointed out (and
he has done an excellent job) is that in a behavioristic account, private
behavior STILL follows the laws of overt behavior. Also, as he has pointed out
(as have I), there are a few things that suggest the cogency of this position:
first, there is the simple fact of private talking (or reading silently, for
example). Do you really want to argue that when we read silently we are not
behaving? That the eye movements are the only thing that is behavior and what
we "hear" silently is NOT behavior? Second, related to this, is that the energy
level of behavior is a differentiable property of behavior. That is, by
arranging contingencies of reinforcement (the main explanatory principle of
radical behaviorism) one can produce response classes where the energy level is
quite low. We can't, of course, arrange contingencies based on complete
non-observability (hence the similarity to to
Wittgenstein' s treatment of the impossibility of a "private language"). But
the point is that behavior can occur with a range of energy levels - why can't
it occur at a level that renders it unobservable to anyone but the behaving
individual? Indeed, the cultural evolution of "silent reading" seems to
constitute just such an occurrence. The historical record indicates that, in
fact, there was a period of time when "silent reading" did not occur. It has
been suggested that it (silent reading) arose after the advent of libraries.
When there are a room full of people reading different texts out loud, reading
accurately is difficult. This lead to a situation that fostered ever less
energetic vocal behavior until the public aspects were removed by differential
contingencies of punishment. In some people, remnants of "reading out loud"
remain; it is sometimes difficult to get children to read silently without
"moving their lips." Finally, the most important
aspect of treating private events as behavior stems from the fact (and this is
not unrelated to my previous points) that, even when unobservable, the
hypothesized behavior comes about because of contingencies. That is, the total
behavioral episode (and I will give an example immediately below) is best,
given a respect for parsimony, viewed as behavior. One of my favorite examples
of this is the "directed forgetting" literature with pigeons. I will describe
this phenomenon, but I am not going to try to describe specific experiments,
though my description will closely parallel experiments that have been done.
This will be somewhat long-winded. Consider delayed match-to-sample experiments
(DMTS). In such experiments, the center key (in operant experiments with
pigeons, they often peck a plastic strip that is mounted behind a circular hole
in the operant chamber - the palstic is translucent, and when a colored light
is turned on behind them, the key
appears as a colored circle) is transilluminated with a color, and there are at
least two different colors used. At he beginning of a trial the center key (of
three horizontally- arranged keys) is, say illuminated "red." The pigeon pecks
this key and the light is turned off. After some delay (on the order of
seconds), the two "side keys" are illuminated. One key is the same color as the
now extinguished "sample" and the other is some other color. If the pigeon
pecks the red key, food is delivered, if it pecks the "wrong" key, no food is
delivered and there is often a "blackout" for some period of time (i.e.,
negative punishment). If the delay is not too long between the sample and the
choice stimuli (10 s is often too long for a pigeon), the pigeon will peck the
"correct" key above chance levels. If the delay is on the order of a few
seconds, the accuracy can be quite high. This, of course, is often held to be
an animal model of "working memory." One of
the first things to notice (though this is not my main point) is that in order
to get pigeons to respond correctly at longer delays, the delay must be
gradually lengthened. One can not just go from a zero s delay to a four s
delay. That is, the pigeon must be trained, and the training involves gradually
changing the contingencies of reinforcement, just as is done when behavior is
trained in other circumstances. But, once the pigeon is trained, one can do
something interesting; one can add a stimulus, say a tone. On trials where the
tone is present, the trial no longer ends with the usual stimuli. Instead, for
example, both side keys can be illuminated with a white light, and pecking
either key results in reinforcement. Or, the trial may simply end with a
blackout. An interesting thing happens when one does this. If one turns on the
tone, but presents the choice stimuli anyway (clever, huh), the pigeon's
accuracy falls to 50% (i.e., chance levels). It
does not remember what color the sample was, even at delays that it usually
"handles" quite well. What is the interpretation of this effect? I think the
parsimonious answer is this: in DMTS the animal learns to behave in some
particular SUBTLE way when the sample is presented. This SUBTLE behavior
continues through the "retention interval" and it is this behavior that
discriminatively controls the choice response. When the tone is presented, the
behavior that occurs during the retention interval is not reinforced (either
because no food can be obtained, or because the reinforced response is
irrelevant to the sample), and it undergoes extinction, but it is only
extinguished in the presence of the tone. Now, what is compelling here, is that
what the pigeon does during the retention interval (if, indeed, it is DOING
something) is not easily observed, and efforts to find "mediating behavior" has
often failed. Interestingly, in the few cases where the
mediating behavior IS easily observed (like, the pigeon turns around to the
left when the sample was red, and turns to the right when it was green), there
is no "forgetting function!" But the main point is that, even though the
mediating behavior cannot be observed publicly, [this is where I left off], the
"alleged behavior" adheres to the same laws as publicly-observable behavior. A
stimulus is presented that "signals" that the behavior will go unreinforced,
and the behavior undergoes extinction. When the choice stimuli are presented
anyway, accuracy falls to chance levels. If, whatever is going on is regarded
as "mental," the fact is that the this "mental" DOING obeys the same laws as
observable behavior. In order to see this clearly, let's say that we conducted
a different experiment. Say we had two keys mounted below the three keys in the
experiment I just described. In this experiment, the pigeon is trained
explicitly to peck the lowest left-hand
key when the sample is red, and the lower right-hand key when the sample is
green. At the end of the "retention interval," the original side keys are
illuminated. It is worth noting that there would be no appreciable forgetting
function (and I'll discuss the significance of this later - not necessarily in
this post since it is already long). Here, it is obvious that the mediating
behavior is, in fact, behavior. If we now introduce the tone "telling the
pigeon that pecking the lower keys will go unreinforced," pecking those keys
will cease, and the probe trials will result in chance levels of responding.
The main point here is that the same operations that result in the extinction
of explicitly-trained mediating behavior also result in the extinction of the
alleged private behavior that is not explicitly trained and not easily
observed. One may insist that the "mental" activity that is not publicly
observable is of a different sort from the
publicly-observable mediating behavior, but one does so at the expense of the
more parsimonious, non-dulistic interpretation.
*Watson argued that "thinking" was "subvocal speech," so he did not deny all
subjectivity.
SWM; If behavioral pyschology includes our mental events (ideas, thoughts,
feelings, beliefs, images, logical connections, etc.) then how can we continue
to call it behavioral?
GS: I hope I have given an answer to this question. Whatever the animal is
doing in the retention interval is not observable, but it is affected in the
same way as behavior in the second sort of experiment. No? One might ask "How
can we not call it behavior?"
SWM: Now I take it you and, perhaps, Glen are arguing that that is precisely
what Skinner really meant (or what his followers now mean) and as I have said I
am no expert on psychology or behaviorism or Skinner so I don't want to be seen
here as passing judgment on that claim. Certainly, I must reiterate, I do NOT
believe that any real behaviorists ever really thought they don't have the
kinds of mental events I've described here despite how many of their claims
might sound. So it is possible that your interpretation is correct and they do
mean more by "behavior" than merely overt (observable) actions.
But if THAT is the case it looks to me that they are confused because they are
so redefining "behavior" as to make it no longer about behavior but about all
the things most of us usually recognize as being other than behavior. Perhaps
this debate is really just about terms and meanings though?
GS: Yes, the debate is partly about meanings. But why? I assert that, in part,
the reason is that one must abandon the comfortable assertion that behaviorism
cannot account for phenomena that are interesting. And, ironically, when
behaviorists do account for such phenomena, it is claimed that it is "not
really behaviorism." What utter nonsense!
Cordially,
Glen
Other related posts:
