Do we have to explain everything?

Behavior analysts are used to explaining. Whether as researchers or practitioners, we see our job as understanding behavior. This mandate implies that we should be able to explain what is going on with behavior, both as a general phenomenon and in particular instances. After all, analyzing behavior involves studying it carefully, breaking it down into meaningful units and identifying the variables that influence them. With this approach, it’s understandable if we feel an obligation to come up with credible explanations when challenged, even if the challenge is entirely our own.

So explaining behavior is part of our job description, but are there no limits to this burden? How far must we go? How detailed must we get? What about when the explanation can only be speculative because the best we can do is refer to general principles? Do we have to explain every implication embedded in colloquial dialect? Do we have to explain everything?

What deserves an explanation?

The seemingly inexhaustible supply of vernacular terms apparently referring to behavior provides boundless opportunities to translate them into a more science-based terminology. Skinner modeled this practice in much of his writing, Science and human behavior (1953) and About behaviorism (1974) being good examples. These discussions can be enlightening by encouraging readers to appreciate real aspects of behavior that are masked by the mentalistic implications of everyday dialect. Such translation exercises explain why an ordinary term referring to behavior is misleading and instead describe what is really going on with behavior, sometimes offering alternative terminology.

The apparent ease with which we can do this masks what Box 2.5 calls a “translation trap.” Just because we are faced with a term that implies something about behavior doesn’t mean it deserves a translation, much less a full explanation. There are literally thousands of such terms, typically overlapping in vernacular connotations. Most involve mentalistic insinuations and lack any clear evidentiary basis for their apparent behavioral referents. Terms labeling traits, attitudes, and moods, for example, suggest something about behavior, but trying to translate each of them into scientific terminology is a fool’s errand.

As a discussion in the preface points out, our science-based dialect does not allow a one-for-one substitution of technical terms for associated vernacular terms. The kind of translating we can do is much more complicated than pulling out a dictionary that offers a scientific term for each plain language behavioral reference. Everyday terms are usually superficial, conversationally efficient, and don’t require special training. They also usually depend on an embedded mentalistic framework that can be avoided only by extended discussion.

Nevertheless, it’s often tempting to take on the challenge, if only because everyday behavioral terms tend to serve as nouns in colloquial syntax and therefore seem to refer to actual “things.” True, behavior analysts can certainly wax poetic about what we “mean” technically when people use everyday dialect to talk about behavior, but this exercise should not suggest that everyday terms have some veridical association on a one-to-one basis with distinctive behavioral phenomena. Just because we say that someone is expecting something does not mean that there is such a phenomenon as expecting that needs to be explained, for instance.

We never know enough

It’s one thing to explain how behavior works in general, but trying to explain specific instances is often impossible. It’s not that our science and its philosophical framework isn’t up to this task, but that we are inevitably short of information when it comes to particular moments of behavior. This is an explanatory limitation for all sciences, however, so we shouldn’t feel inadequate. An expert in fluid dynamics cannot predict exactly where a specific leaf falling from a tree will land not because of shortcomings in the science of fluid dynamics but because the details of the variables operating in that instance cannot reasonably be known.

Even the most expert behavior analyst is similarly limited in explaining exactly what is going on with a particular instance of behavior at a particular moment. Although there may be sufficient evidence to make a reasonable speculation, it can be no more than an educated guess. In any specific case, we just cannot know enough about the history of the behavior at issue, nor the status of all of the variables operating at that moment, to allow a detailed explanation with any certainty that it is correct.

It is important to acknowledge this limitation, lest we be over confident in our abilities. It is particularly important that others understand this constraint when we are challenged to explain a particular episode of behavior. This shortfall has nothing to do with the adequacy of the underlying science however. After all, we trust the science of fluid dynamics well enough to hop onto an airplane and fly away, confidant it will not fall out of the sky. We have similar confidence about our ability to change behavior in targeted ways, given adequate control over the relevant variables.

It’s OK to say you don’t know

Finally, one of the defining features of being a scientist – or a practitioner delivering a science-based technology – is being perfectly content to admit that we don’t know everything. This precaution should be seen more as a welcome relief – even with a note of pride – rather than as an admission of failure. Science operates at outer limits of our understanding of nature, where we ask about what we do not yet know. Admitting that we still have things to learn shouldn’t be seen as depreciating what we already know, as much as respecting the limits of scientific knowledge.

When you are challenged to explain what is going on with a particular behavior, it’s fine to give it a shot, but you should feel no compunction to portray too much confidence. In fact, your explanation should be replete with caveats about what you don’t know. Describing what the underlying science has to offer, in tandem with qualifications about the variables that might be operating in the case at hand, is likely to come across as more convincing, if not more humble, than pretending that you know it all.

References

Skinner, B. F. (1953). Science and human behavior. New York: McMillan.
Skinner, B. F. (1974). About behaviorism. New York: Alfred A. Knopf.