The marshmallow at the end of the universe

Psychology Professor Michael Inzlicht has a confession to make. He’s been peddling shoddy wares – his words. And he’s feeling quite bad about the whole thing. The work wasn’t just intellectually weak. It did real harm. Though his own proposals to popularise his ideas were knocked back, the so-called “marshmallow effect” went viral.

As you’ll probably recall, in the 1960s, psychologist Walter Mischel ran experiments on children who were told they could have one treat (a marshmallow, an Oreo) now, or two if they waited. Years later, Mischel discovered striking correlations. Children who’d resisted the treat fared better: higher academic achievement, lower body mass.

A parable was born: delay the marshmallow, win at life. With self-control such a powerful predictor of future success Inzlicht and co sold that story to those who’d listen. They’d help them master their impulses and improve their lives.

Many years later they reviewed the evidence. And it showed “three surprising things”.

First, people with more consistent {long-term self-control habits} actually engage in less moment-to-moment struggle, not more. Second, exerting effortful control in the moment doesn’t reliably predict long-term success. And third, while people can sometimes {sustain self-control for a period}, these improvements tend to disappear faster than a New Year’s resolution in February.”

I respect Inzlicht’s candour. Yet even with his mea culpa I am not sure the real issue has been addressed. If the original picture was wrong, what exactly made it wrong? What kind of mistake was it — and what does it reveal about the way that knowledge was built?

The mistake

At its simplest, the mistake involved assuming causation from correlation.

But there are deeper problems.

Treating “self-control” as a unitary trait seems reasonable enough to modern sensibilities. But until modern times, such decontextualisation wasn’t really a thing. I think the decontextualisation is in large part an artefact of the scientistic method of modern academic psychology. Doing experiments and trawling existing data, it discovers regularities. These then stand in their reified glory. (For more on this see the next item below on abstractions).

In the case of self-control, there’s no shortage of richer contexts within which might understand the kinds of problems Inzlicht discusses. Aristotle framed self-control not as resistance to temptation but as the alignment of desire with reason. It is a habituated virtue, cultivated over time. It aims at eudaimonia—a flourishing life—not productivity or compliance with something imposed from without. At least in this context, William James’ ideas were similar.

So why weren’t ideas like these considered before Inzlicht and his colleagues raced off in the direction they did?

As he confesses, “We were rewarded for splashy findings, big effects, simple stories, and counterintuitive results. These were the kinds of papers that got published, cited, and funded.” He might have added that that’s what’s rewarded in the marketplace – for books and ‘content’ and short personal and professional development courses.

Science or practical wisdom?

There’s a deeper point. Aristotle distinguished between episteme and phronesis or in our language, between science and practical wisdom. The former is our knowledge of the world as it is. Practical wisdom is our capacity to operate in that world and to improve it for our purposes. Herbert Simon references this distinction when he distinguishes between science and design.

The natural sciences are concerned with how things are. Design, on the other hand, is concerned with how things ought to be, with devising artifacts to attain goals. … Engineering {and} medicine … are concerned not with the necessary but with the contingent – not with how things are but with how they might be – in short, with design.

Now I have a view – contrary to Simon – that all those disciplines by which we try to better understand human society (in which I’d include psychology, sociology, and economics) should be thought of as disciplines of practical wisdom, not as sciences. But I wont pursue that here.

What I will say is that, had Izlicht and his colleagues been proposing a new medicine or bridge building material, they wouldn’t have simply announced the new material together with some observed patterns. They wouldn’t have proposed using the new knowledge in practical settings until they’d identified the steps in the process from action to desired outcome and sought to test the efficacy and risks of the new approach all along that causal pathway.

My main point is not that​​ they should have taken fewer risks with those who took their advice – though Inzlich clearly thinks this. It’s that the way they’re operating isn’t really serious. It’s obviously unfit for purpose. Even if their initial research into whether their treatment worked or not was inadequate, if the process was at all serious, it would have quickly turned up clear evidence that it wasn’t working.

But that wasn’t how the discipline was built. And it continues to this day with practitioners at the commanding heights of the discipline of psychology throwing off ‘evidence-based’ odds and ends which people go out and market for all their worth.

Izlicht’s disowns the old model only to propose a new one – now split between “trait” and “state” self-control – as if this clears the confusion. But, as before, these abstract categories float freely, disconnected from any deeper suggestions of how the psyche works, or of the kind of insights into the phenomenology of the will that William James offered.