Go
back to Autism Articles & Chapters
Autistic
Behavior, Behavior Analysis, and the Gene - Part 2
Richard W. Malott1
Behavior Analysis Program
Department of Psychology
Western Michigan University
Abstract
This article reviews the negative behavior-analytic commentary on Drash and
Tudor’s behavior-analytic analysis of the etiology of autistic repertoires
and values. This article also asks that, in our effort to scrub it clean,
we not drown Drash and Tudor’s beautiful, but fragile, new-born, behavior-analytic
baby in hyper-methodological, hyper-scholarly bathwater
In a special issue of The Analysis of Verbal Behavior, several behavior-analytic
colleagues commented on the Drash and Tudor theoretical analysis
of autism as a contingency-shaped disorder (Carr & LeBlanc; Drash & Tudor
[a & b];
Hall; Hixson; Malott; Romanczyk & Gillis [2004]). Most
of the behavior analysts were quite critical of the Drash and Tudor theory
and not in a sympathetic way. We have been trained in grad school to review
articles as if we were critiquing a Psych. Review article resulting from
a long series of million dollar NIMH grants, rather than the speculations
of two colleagues who have been logging in quite a few years working with
children labeled as autistic. And I fear this heavy-duty, scholarly criticism
may have the unintended effect of killing this nascent behavior-analytic
theory in its infancy. Instead, I would encourage our behavior-analytic community
to help Drash and Tudor get their theory ready for prime time. Instead of
listing all the reasons why their theory is wrong and should be disregarded,
I think we would do better to list all the issues the theory still needs
to address and then help Drash and Tudor address those issues. And at the
same time, it might be the most important contribution to our the field since
the early Lovaas work (e.g., 1987), if these behavior-analytic autism experts
and the readers of this essay would actively participate in implementing
Drash’s (2004) brave, new research proposal, Preventing Autism Now:
A Possible Next Step for Behavior Analysis; this is a proposal to empirically
exam the role of the child’s early behavioral history in the acquisition
of autistic behavior and values. However, in the mean time, the present essay
will address some of the negative, behavior-analytic commentary on the Drash
and Tudor theory, because I can’t resist the immediate reinforcers
of criticizing the critics.
Terminology
I am skeptical of the very terms autism and disorder that
convention may almost be forcing Drash and Tudor to use, because, guard against
it though they might, talking about an autism disorder quickly
leads to a reification of autistic behavior, resulting in autism as
a thing that causes autistic behavior, and we end up on board the medical-model,
biological-deterministic band wagon with its circular trip to simplistic analyses
and easy answers to hard questions.
For example, even when outstanding behavior analysts acknowledge that
autism may represent a multiplicity of behavior problems with a multiplicity
of etiologies, they then go on to suggest that each of those behavior problems
may have its own neurobiological cause. However, a more behavior-analytic approach
might be to look at the contingencies responsible for each problem.
As a parallel, when considering a person who has a multiplicity of behavior
successes, such as speaking English, riding a bicycle, playing the piano, and
voting democrat, we wouldn’t say each component of this particular repertoire
of successes has its own neurobiological cause. However, if we were to label
this more-or-less random multiplicity of successes as the generation-w syndrome,
then researchers might be seduced into the search for the set of neurobiological
causes, just as the label autistic disorder may be seducing the best of behavior
analysts away from doing a thoughtful behavior-analytic interpretation of the
various ways early infant contingencies might wreak profound havoc.
Another parallel is the use of the term intelligence to describe a multiplicity
of behavior successes and problems. The reifying potential of this word could
be what leads even a world-class behavior analyst to suggest that intelligence
may be inherited in the form of speed of condition and the capacity to maintain
a large repertoire without confusion. But a more productive approach for behavior
analysts might be to do a careful analysis of the details of behaving intelligently
and the contingencies that facilitate the acquisition and maintenance of intelligent
behavior. Even concepts like speed of conditioning may address such a multiplicity
of conditions as to almost be a misleading reification, just as Skinner concluded
his response strength was.
Experimental Analysis vs. Informal Speculation
Yes, Drash and Tudor have not formally demonstrated a statistical correlation,
let alone a causal relation, between autistic behaviors and pathological
behavioral contingencies. What they have done is offer plausible suggestions,
based on their informal, though extensive, observations of autistic behavior
imbedded in pathological contingencies.
It is true, that Drash and Tudor are extrapolating well beyond their data set
consisting of a small number of single-subject research studies. But their
extrapolation is much less extravagant than that of Skinner when he extrapolated
from humble, little experiments with only three rats to an analysis of all
human nature.
And, it is true that Drash and Tudor’s analysis will have much more impact,
if it is supported with hard experimental and field-study data, but their analysis
is more empirical and less speculative than was that of Skinner (1957) when
he wrote Verbal Behavior. And it is true, that Skinner’s analysis did
lie dormant, until Jack Michael finally convinced a handful of graduate students
that Verbal Behavior was profoundly correct. Then those students started a
major new approach to working with the language deficits of children classified
as autistic (Sundberg & Michael, 2001). In other words, Drash and Tudor’s
contingency analysis may be a great start, even though it is speculative.
We behavior analysts have long chanted the mantra of caring only about interventions,
about what works, not about etiology. And we have been so traumatized by our
mentalistic and psychodynamic heritage of the quagmire of speculative psychology
that we are afraid to even consider speculative behavior-analyses of etiology.
We even enforce compliance with this anti-theory mantra in our professional
journals. So appreciating the value of Drash and Tudor’s effort may require
a shift in our world view, our paradigm for appropriate scholarship.
Meaningful, but Subtle, Differences in Everyday Infancy
Drash and Tudor describe a set of pathological behavioral contingencies that
might result in autistic behavior. But others criticize their theory because
most of those pathological contingencies are present in the infancy of all
children, whether or not they acquire serious autistic repertoires; for example,
all parents pay attention to their screaming baby. But that just suggests
to me that we should be amazed that all children don’t acquire serious
behavioral problems. It’s like any behavior analyst who compares appropriate
classroom practices with actual classroom practices should be amazed that
any children learn to read, write, or count—a major grade-school miracle,
as far as I’m concerned. In other words, we need a more fine-grained
analysis of the differences in the contingencies and the interactions of
those contingencies that do and do not result in serious autistic repertoires.
To put it another way, none of us are impressed when an untrained teacher
says, “I
tried that behavior modification stuff with my students, and it didn’t
work.” We know enough about behavior modification to suspect that the
contingencies the untrained teacher tried were not close enough to the contingencies
we would use to achieve the desired results. Similarly, we know enough about
behavioral contingencies, in general, to suspect that the contingencies involved
in the early life of children with normal repertoires must not have been close
enough to those of children with serious autistic repertoires to have the undesired
results, even when comparing the behavioral histories of two children from
the same family.
Similarly, I don’t think we should be too impressed when an untrained
parent says, “I treated Susie and Johnny exactly alike, and Susie had
absolutely no problems, but Johnny has autism.” We
know enough about self-reports to suspect that even if we professional behavior
analysts were that parent, we should not be too confident in the validity of
our own report. It may be more-or-less impossible to casually monitor those
subtle behavioral interactions with any reliability. This is a case where independent
observers with their clip boards is crucial.
However, the difference between the classroom and the home is that behavior
analysts have filled volumes with experimental analyses of classroom contingencies
and barely a notebook with experimental analyses or even behavior-analytic
field studies of infant-rearing contingencies. So it takes a much greater leap
of faith to believe that successful child rearing, like successful classroom
teaching, lies in the subtle details of the contingencies. The angels as well
as the devil are in the details.
That some behavior-analytic autism experts have children
with autistic repertoires is interesting, but I don’t see it as very
convincing evidence for the biological basis of such repertoires. I’ve
know many professional behavior analysts who are skilled at doing behavior
analysis in the lab, in other peoples classrooms, and in other peoples homes.
But few have impressed me as being much more skilled than the average person
in sticking to their behavior-analytic guns when they are in their own classroom
or home. It is much easier to help others apply appropriate behavioral contingencies
in their lives than to apply them in our own, where we must now wear not only
the sole hat of behavior analyst but the multiple hats of teacher, parent,
spouse, friend, etc., enough hats to cause serious headaches.
Biological Determinism
Along the same lines, I don’t think we should be overly impressed with
the implications of a biological etiology when giving a drug to a child classified
as having ADHD disrupts his ADHD repertoire. We would not conclude that the
etiology of a pigeon’s fixed interval scallop is biological rather than
contingical, just because that scallop can be disrupted by drug administration.
So I think we should be reluctant to jump from the effects of drugs on ADHD
repertoires to the importance of biological determinism in the acquisition
of autistic repertoires and values2.
Bijou and Ghezzi’s suggest that a) some children may be hypersensitive
to the aversive qualities of auditory and tactile stimulation and b) the aversiveness
of such stimuli might form the basis of the contingencies that shape autistic
repertoires and values. But even if these interesting speculations are verified,
they still don’t point to a biological etiology of those autistic repertoires
and values in the usual sense of biological etiology. Such verification does
not mean there is a gene or a combination of genes that causes the child to
acquire autistic repertoires and values. That would be like saying there is
a genetic etiology of the acquisition of avoidance behavior in the Skinner
box, because electric shock is inherently aversive. It’s the contingencies,
not the gene, nor the neurotransmitters, nor any other biological entity, that
results in the acquisition of an avoidance repertoire; and it may be the contingencies
that result in the acquisition of an autistic repertoire3.
However, Bijou and Ghezzi’s speculation does suggest that we might do
well to screen infants for hypersensitivity to auditory and tactile stimulation
and carefully monitor the behavioral contingencies in the environment of those
infants found to have such hypersensitivities.
There are other possible candidates for pathological reinforcers and aversive-conditions
than hypersensitivity to auditory and tactile stimulation, for example, under
sensitivity to social reinforcers (such as a smile or verbal praise) and under
sensitivity to social aversive conditions (such as a frown or verbal disapproval).
However, we know enough about the acquisition of learned reinforcers and learned
aversive conditions, that we might readily suspect the lack of those particular
crucial reinforcing and aversive values is a result of a lack of effective
pairing of neutral social stimuli (e.g., smiles) with effective reinforcers
and a lack of effective pairing of other neutral social stimuli (e.g., frowns)
with effective aversive conditions. So we needn’t rush to the search
for neurobiological causes of these deficits. In fact, when working with children
labeled autistic, the first thing we behavior analysts do is try to provide
effective pairing of neutral social stimuli with reinforcers so as to establish
those learned reinforcers. (However, we seem to actively avoid effective pairing
of other neutral social stimuli with aversive conditions, so as to establish
those learned aversive conditions. And this might be a serious omission in
our behavioral curricula, if we want the children to ultimately be sensitive
to the socially inappropriateness of various subtle behaviors.)
Other Issues
It is true that we behavior analysts should guard against too insular a perspective
that results in our ignorance of the large amount of biological research
dedicated to finding the neurobiological basis of autism. However, we should
also guard against being uncritically swept away by the rising tide of biological
research. We should view proposed biological etiologies, as the explanations
of last resort until we have thoroughly exhausted our search for pathological
behavioral contingencies. If it works, the contingency analysis is the most
parsimonious, elegant explanation of behavior problems of all sorts, because
it makes use of no hypothetical processes or reification and relies only
on known, well-documented concepts, principles, and processes.
If pathological behavioral contingencies cause autistic repertoires and values,
then it is likely that replacing the pathological contingencies with appropriate
contingencies would replace the autistic repertoires with appropriate repertoires.
(But the success of this contingency replacement is likely, only if the appropriate
contingencies plus remedial contingencies are implemented with the skilled
help of behavior-analytic autism experts.) However, it is a logical error called
affirming the consequent, if we argue the reverse, that successful implementation
of appropriate and remedial contingencies proves that pathological contingencies
created the pathological repertoires in the first place. But, it is not a logical
error to argue that the success of such behavioral-contingency interventions
strongly suggests that pathological behavioral contingencies are good candidates
for the cause of the problems.
Most behavior analysts have had a tremendous amount of high-quality training
in research methodology and a relatively small amount of training in conceptual
analysis; probably as a result of a misunderstanding of Skinner’s objection
to hypothetico-deductive theory construction. As a result we must take great
care not to miss the conceptual forest for the methodological trees. For example,
in writing reviews, we must take care that we not fall into the trap of comfortably
writing critiques of research methodology, with little regard for conceptual
integrity or conceptual significance of the literature we are reviewing. I
suggest that the Drash and Tudor analysis has great conceptual integrity and
significance.
Conclusions
I ask my colleagues to apply their highly developed research muscles to helping
Drash and Tudor carry the heavy load of empirically determining the role
of behavioral contingencies in the etiology of autistic repertoires and values.
I ask my colleagues to actively participate in implementing Drash’s
(2004) brave, new research proposal, Preventing Autism Now: A Possible Next
Step for Behavior Analysis.
References
Drash, P. W. (2004). Preventing autism now: A possible next step for behavior
analysis. Retrieved, January 1, 2004 from http://www.DickMalott.com.
Carr, J. E., & LeBlanc, L. A. (2004). A comment on Drash and Tudor’s
(2004) operant theory of autism. The Analysis of Verbal Behavior, 20, 25-29.
Drash, P. W. & Tudor, R. M. (2004-a). An analysis of autism as a contingency-shaped
disorder of verbal behavior. The Analysis of Verbal Behavior, 20, 5-23.
Drash, P. W. & Tudor, R. M. (2004-b). Is autism a preventable disorder
of verbal behavior? A response to five commentaries. The Analysis of Verbal
Behavior, 20, 55-62.
Hall, G. A. (2004). Towards a balanced account of autism etiology. The Analysis
of Verbal Behavior, 20, 37-43.
Hixson, M. D. (2004). Autism as a contingency-shaped disorder of verbal behavior:
Evidence obtained and evidence needed. The Analysis of Verbal Behavior, 20,
49-53.
Lovaas, O. I. (1987). Behavioral treatment and normal educational and intellectual
functioning in young autistic children. Journal of Consulting and Clinical
Psychology, vol. 55, pp. 3-9.
Malott, R. W. (2004). Autistic behavior, behavior analysis, and the gene. The
Analysis of Verbal Behavior, 20, 31-36.
Romanczyk, R. G., Gillis, J. M., (2004). Commentary on Drash and Tudor: An
analysis of autism as a contingency-shaped disorder of verbal behavior. The
Analysis of Verbal Behavior, 20, 45-45
Skinner, B. F. (1957). Verbal behavior. New York: Prentice-Hall
Sundberg, M. L., & Michael, J (2001) The benefits of Skinner’s analysis
of verbal behavior for children with autism. Behavior Modification, 25, 690-724.
Author’s Notes
I wrote this article as an assignment in Linda Leblanc’s Psychology
697. Autism: Etiology, Assessment, and Behavioral Intervention, fall
2004; and I would like to thank Linda for her thoughtful, helpful suggestions.