Why thoughts aren’t causes

Marshmallow MushroomsI was asked this week by a couple of my students why I prefer the behaviour analytic ways of thinking, compared, say, with the standard information processing approaches of cognitive psychology. I’m not a diehard behaviourist, but I do lean quite a bit toward functional contextual ways of thinking, and I’ll be honest and say B.F. Skinner is my all-time intellectual hero. (I’ve also been involved in qualitative research and I teach psychometric scale development, so please don’t write me off as ‘one of them’.)

There is a lot of rubbish spouted about behaviourism, often by people who should know better. Claims that behaviourists deny the existence of internal psychological events like thoughts and emotions might not be ridiculous if you’re thinking about the behaviourism of of John B. Watson, but virtually no behaviour analysts today are thinking about him. Watson’s behaviourism is often called methodological behaviourism and it stands in stark contrast to Skinner’s more recent radical behaviourism. Skinner explicitly was interested in thinking and feeling. Indeed, he wrote an entire book about it (which arguably lead to the falling out of favour of this school of psychology).

Claims that behaviour analysts routinely punish their clients into compliance are simply bullshit. (I’m using the word here in the sense of Harry G. Frankfurt’s classic text where he defines bullshit as making knowledge claims when you have insufficient familiarity with the knowledge domain. So there.)

So what are the main features of modern behaviourism, and why do I think it’s a useful way to conduct science? (I’m going to say ‘modern behaviourism’ to lump together radical behaviourism and closely related philosophical frameworks like functional contextualism.) There are lots of important features. Functional contextualists make the assumption, for instance, that it’s important to very clearly define the scope of the behaviour you’re analysing; since the world doesn’t come already pre-quantised it’s important that we state plainly how we, as scientists are choosing to chop it up.

For me, though, the two most important features of modern behaviourism are these:

  1. Mental events are not considered causes of behaviour. They are simply types of behaviour themselves and are themselves to be explained.
  2. Behaviourists look to be able to predict and influence behaviour. Prediction alone is not enough.

This is the very nub of the difference between much of applied cognitive psychology on the one hand and modern behaviourism on the other. Though I wouldn’t want to over-play it, this difference often leads to quite different interpretations of the same phenomena, and in my view, the behaviourist account is usually more hopeful.

Take the classic marshmallow experiments. You’ve probably heard of them. In a series of experiments in the late 1960s and early 1970s at Standford, Mischel, Ebbesen and colleagues sat children in front of a marshmallow and told them that if they didn’t eat it right away, they’d get more marshmallows later. This is a classic delayed gratification paradigm. If the child can resist, she’s rewarded with even more sweeties. Some kids wait. Some gobble the sweets. Over the years, Walter Mischel and colleagues built an impressive body of work examining individual differences in this ability to delay gratification. For example, in 1988 they published a longitudinal paper showing that “children who were able to wait longer at age 4 or 5 became adolescents whose parents rated them as more academically and socially competent, verbally fluent, rational, attentive, planful, and able to deal well with frustration and stress.” Impressive stuff.

What are we to conclude from such research? Well, the phrase ‘individual differences’ appears many times in these papers. Though the researchers occasionally look at contextual factors, such as whether the children were encouraged to think about the flavour or the shape of the treat, mostly they are clearly describing this phenomenon as a function of some internal ability the child has. And that’s precisely how the media interpret such findings too. Time Magazine, for instance, use phrases like, “show an underlying inability to exert self-control in adulthood.” What does it suggest to us, as applied psychologists and educators? At best, that we should encourage people to pull themselves up by their bootstraps and develop ‘self control’. At worst, that some people just don’t have what it takes. Elsewhere in the popular media the findings are interpreted as suggesting that we can develop better willpower techniques if we try really hard to work on our own minds.

Fast forward half a century. Celeste Kidd and colleagues at the University of Rochester repeated the experiment with a contextual manipulation. For some of the kids, the marshmallow experiment was preceded by the experimenter offering some crappy crayons for an art project, saying she’d go and get better ones, then returning empty handed. For some kids, the experimenter followed through. Then, just to drive the message home, in the first ‘adults are unreliable’ condition, the experimenter offered a sticker and said she’d come back with more, better stickers soon. Then didn’t. In the ‘adults follow through’ condition she came back with some awesome stickers.

At this point, your common sense is telling you what happened when the kids were then asked to wait, staring at a marshmallow, whilst the experimenter went to get more. You might imagine yourself in the situation. You might imagine yourself thinking, “she lied before, so she’s lying now”. Regardless of what you imagine happening in the child’s head, the focus of this experiment on the context of the behaviour leads us naturally to different ideas about its implications. We’re no longer thinking that the marshmallow-munching kids show an “underlying inability”. Instead, we realise that kids raised in an unreliable family environment would learn a generalised set of behaviours to take what’s available now, and discount promises of future reward. With this later set of results, we’re imagining all sorts of family-based interventions to help children become more “academically and socially competent”.

Dr Kidd describes herself as a cognitive scientist, so why am I using the excellent work she and her colleagues did as an example of how behaviourism is the best thing since the web? It’s simple. Cognitive psychology took half a century to come up with some robust findings that environmental context plays a powerful role in guiding these sorts of ‘willpower’ behaviours. Dozens of papers and thousands of person-hours have gone into exploring whether this or that personality characteristic is associated with waiting for the second marshmallow. I humbly suggest that modern behaviourism would have got us there faster.

Remember what I said earlier about causality? Behaviourists do not accept mental events as causes of other behaviours. Explanations invoking children’s willpower or other ‘individual differences’ would be complete non-starters for most behaviourists. Any behaviour analyst coming across this phenomenon would immediately have started looking for contextual events, outside of the child’s own skin, that seem to influence the behaviour. These events might have been patterns of reinforcement within the family context. Early on, had Mischel taken a behavioural stance, he would have asked what experimental manipulations to the procedure and to the environment in which the children found themselves would encourage them to wait for the second marshmallow.

B.F. Skinner wrote an almost utopian novel, Walden Two, published in 1948, to show how the application of radical behaviourism could lead to an improved, more harmonious society. The book has been roundly criticised, but I’d like to suggest that Skinner was right in at least one regard. If we look to internal mental phenomena like ‘willpower’ to explain our behaviour we risk developing a very pessimistic outlook on the world where people have simply to put up with their lot in life. If we focus on how external circumstance influences behaviour we immediately start to build ideas of how to support people to develop more useful behavioural repertoires.

Despite the bad rap, modern behaviourism is inherently the most hopeful school of psychology. Perhaps I’m a hopeful psychologist, but it’s this, more than anything else, that draws me to modern behaviourism.