This was first of all a tribute to B. F. Skinner on the hundredth anniversary of his birth. A good part of the presentation consisted of an overview of Skinner's life and work, and the significance of his achievements. In fact, it would be a good fit with the Tribute to B. F. Skinner presentations at this year's ABA Convention.
The talk was given at a new museum located in an old Minneapolis flour mill. Skinner's WWII work on Project Pigeon, the development of a guided bomb (Pigeons in a Pelican, 1960,1974) took place in an adjoining building in the same complex.
Dr. Peterson is a member of the faculty of the University of Minnesota Psychology Department whose office window happens to overlook the old mill site. Originally in the animal learning area, he is now involved in 'translational' work (he's a BCBA). He is active in the autism and dog training communities so the audience was both large and varied!
The title of
the talk refers to a discovery made by Skinner, Keller Breland and
Norman Guttman (at the time his graduate students) in the course of
this project.
Since this was a military project they spent a fair amount of idle
time waiting for task specifications and the like from their military
employers. During one of these lulls they decided to teach a pigeon
to bowl.
Now, those of us who are familiar with Skinner's work aren't too
surprised by this. He was noted for his skill at getting animals to
emit surprising complex behaviors.
Further, one would predict the manner in which he would go about
this: he would shape the behavior. After all, hadn't he already
taught the rat Pliny to emit a complex chain of responses through a
process of successive approximations (B of O, 1938, pp
339-340)?
And in Reply to Konorsky and Miller (1937) he also appeared to
be describing the process of shaping the lever press.
However, in The Shaping of a Behaviorist (1979 p268) Skinner
mentions that Pliny was in fact taught his chain by painstakingly
making gradual adjustments in the apparatus itself rather than by
what we'd now call "hand shaping": controlling the delivery of a
reinforcer directly by the experimenter rather than by some automatic
mechanism. It is likely the the description in Reply to Konorsky
and Miller is another instance of 'extension'; Skinner was
speaking hypothetically rather than from data.
Skinner's actual practice (not at all unique) for getting a subject to perform in an experimental apparatus was to put the pigeon or rat in the chamber at leave it until it chanced to operate the key or lever and be automatically reinforced. Most of us still train our research subjects this way (let the apparatus do the work).
However,
Skinner's pigeon showed no inclination to bowl; the response (swiping
at a wooden ball with its beak) was not emitted even once.
Finally, in desperation, they decided to reinforce anything that
looked a bit like a swipe. This was an easy step, since they had no
automated apparatus for this behavior and had to hand deliver
reinforcers anyway.
Skinner expresses their collective amazement at how rapidly the
pigeon acquired this unlikely behavior: "I remember that day as one
of great illumination (Shaping, p268)". The beginning of
shaping as we now know it!
Much of the preceding material can be found online in The Discovery of Shaping, Or, B.F. Skinner's Big Surprise, by Gail B. Peterson (http://www.behavior.org/animals/animals_discovery_shaping.cfm) and The World's First Look at Shaping: B.F. Skinner's Gutsy Gamble (http://www.behavior.org/animals/index.cfm?page=http%3A//www.behavior. org/animals/animals_worlds_first.cfm).
OK. Very interesting so far, but what (one might ask) does this have to do with verbal behavior?
From the rat
or pigeons' point of view, of course, there is no functional
difference in the source of the reinforcer; the contingency is the
same whether reinforcement is mediated by some aspect of the
inanimate environment or by a change in the behavior of another
living organism.
So far, no basic advance....
However,
note the phrase "mediated by .... a change in the behavior of
another living organism". This sounds very much like
Skinner's (1957) definition of verbal behavior!
Skinner had always been interested in language (not surprising for a
lapsed novelist), but his early interests were what you'd expect from
a former English major: antecedent focused studies of word use and
frequency, such as alliteration in Shakespeare's sonnets.
At about this time there was a change in the way that he approached
verbal behavior, culminating in the publication (after a long
gestation) of Verbal Behavior in 1957. This was the approach
that we know now -- very different from literature and linguistics
and focused on the consequences of behavior as well as its
antecedents.
Peterson suggests that it was the dramatic demonstration in the flour
mill of the power of one organism to directly change the behavior of
another that led to the change in the way that Skinner looked at
verbal behavior, and to a shift in his interest from the laboratory
and theoretical analysis to the applications of behaviorism to human
behavior on the cultural level.
I might add
(with Peterson's concurrence) that Skinner had another experience a
few years earlier that also influenced the way in which he approached
verbal behavior. This was the birth of his daughter Julie and her
acquisition of language. In The Shaping of A Behaviorist (p
239) Skinner marvels at the ease at which she acquires verbal
behavior: "'She's a swell little kid' I told Fred (Keller) 'and very
verbal. So far she has not successfully assailed my behaviorism.'
.... She was turning my attention to education." He goes on to talk
about writing a book to use "successive approximations" (his quotes)
to teach children complex verbal behavior.
At the time that Skinner was working on Project Pigeon his second
daughter Deborah would have been going through the same
process.
I'd suggest that it was the confluence of the demonstration of the power of mutual reinforcement systems and the earlier observation of his daughters' acquisition of verbal behavior that led to his future emphasis on the applications of behaviorism to complex (and verbal) human behavior in real world situations. In a sense, Schedules of Reinforcement (1957) was his swan song as a laboratory scientist; it was more Charles Ferster's product than his own. This is also described in the Journal of the Experimental Analysis of Behavior's special issue on the Harvard pigeon lab (2002, vol 77).
Peterson concluded by describing some of the many current applications of shaping, including teaching verbal behavior to autistic children (an interest of his). He points out that the goal is to teach autistic children "the behavior of a speaker".
All in all, a fine tribute to B. F. Skinner, and an insight into the genesis of Skinner's particular analysis of verbal behavior.
Paul K. Brandon
Minnesota State University, Mankato