The Peak-End Theory of Dessert: A New Philosophy of Wolfing

My daughter Kate eats desserts slowly -- has done so as long as I can remember. She is what I'll call an extreme savorer. In other words, she is a completely irrational moral monster, as I will now endeavor to show.

I'll admit that on a superficial analysis, Kate's approach to dessert appears wise. Someone bakes brownies. Everyone in the family receives a brownie of equal size. My son's is gone in a flash. My wife and I eat ours moderately quickly. Kate delicately saws off an edge and puts it on her tongue, waits a while, saws off another edge. Ten, fifteen minutes later, Kate is still enjoying her brownie while the rest of the family watches enviously.

At such moments I think, "Why don't I slow down and savor my dessert like Kate does? She obviously derives much more sweet pleasure from her slow ways!" Several days later, of course, it's "Yum, ice cream sandwiches!" munch munch munch and once again after my share is gone I find myself envying Kate's slower pace.

[Kate at work on a Trader Joe's dark chocolate peanut butter cup]

Now I'd rather not see myself as quite as terribly irrational as this pattern suggests. Fortunately, being well-trained in both philosophy and psychology, I have a wealth of theoretical resources from which to concoct a plausible justification of pretty much anything. Our task for today, then, is to demonstrate that I am right and Kate is wrong.

We need a philosophy of wolfing.

The Peak-End Rule

In a classic series of studies, Daniel Kahneman, Donald Redelmeier, and collaborators found that in retrospectively evaluating negative or painful experiences, people tend to disregard the duration of the experience. Instead, people evaluate their experiences mostly based on how the experiences felt at their peak and how they felt at the end.

Some of the results are startling: For example, in one experiment, ordinary colonoscopy patients were either given standard painful colonoscopy procedures or instead the same standard painful procedures plus the extra (but less severe) discomfort have having the colonoscope rest in their rectum unmoving for an additional three minutes at the end. Patients reported their pain levels in real time throughout the procedure. The peak level of pain was the same in the two groups, as was the overall pain during the main part of the procedure. Consequently, patients in the second group had more total pain: the pain of the main procedure plus an extra three minutes of discomfort. Nonetheless, patients in the second group retrospectively reported having experienced less pain, and they reported a less negative overall attitude toward the procedure.

What's more, the patients acted accordingly: Over the next five years, patients who otherwise were predicted to have a low propensity to return for another colonoscopy (patients with less past history of colonoscopies and no detection of abnormalities) were more likely to return for another colonoscopy if they were in the experimental group who had received the extended procedure than if they were in the control group who had received the standard procedure.

Kahneman and colleagues found similar results with participants asked to hold their hands in painfully cold water. Participants held one hand for 60 seconds in water that was 14.1 degrees Celsius (painfully cold but not damaging) (Procedure A) and also, either before or after, held their other hand for 60 seconds in water that was the same 14.1 degrees C and then kept it in the water for an additional 30 seconds while the temperature was slowly raised to 15.2 degrees (still uncomfortably cold) (Procedure B). When told that they could choose either Procedure A or Procedure B for the third trial, most chose Procedure B. They chose more pain over less.

According to the "peak-end rule", the retrospective evaluation of either a positive or a negative experience is an average of the quality of the experience at its positive or negative peak and the quality of the experience at its end, with little regard for duration. Despite the fame of this rule (including in guides to managing one's business, etc.), it isn't as thoroughly studied as one might expect, and findings remain mixed.

Still, let's assume that something close to the peak-end rule is true about the enjoyment of desserts: How fondly you remember your dessert is a function mostly of the average of your most pleasant bite and your last bite.

Wolves Will Remember Dessert More Fondly

It should now seem plausible that if you wolf down your dessert, you will remember it more fondly.

The peak will be better: Instead of modest bite after modest bite of moderate pleasure, you will experience the unrestrained joy of a giant bite of popping deliciousness all at once. Maybe the best part is the cherry atop the icing. The wolf will get that great mouthful of cherry, icing, and cake all at once, while the savorer will have no such moment of sudden decadent indulgence.

The end will also be better: We grow weary of even the best things over time. Although the twentieth bite of chocolate is still good, it's never as wonderful as the first few bites. By bite twenty your mouth is accommodated to the sweetness, and the pleasure is only a temperate, lingering continuation. The last bite will be more flavorful if you don't take too long in getting around to it.

Furthermore, there is a joy in not holding back. What could be more childish fun than just diving in, biting the whole head off the Easter bunny or shoving a great spoonful of cherry, whip cream, and cake right into your mouth? The savorer's experience will always be tainted with the slightly unpleasant feeling of self-restraint.

This figure compares the Wolf's and the Savorer's dessert experiences over time:

[click to enlarge and clarify]

Although the savorer enjoys dessert longer and the total sum pleasure (represented by the total area between the line and the x axis) might be larger, both the peak and the end are higher for the wolf. If peak-end theory is correct and duration is mostly ignored, then looking back on the dessert, the wolf will think, "wow, that was great!" while the savorer will think "okay, that was pretty good".

"But Peak-End Reasoning Is Irrational!"

Look, I know what you savorers are thinking. It's irrational to choose according to the peak-end rule. It doesn't make sense to tack some extra pain at the end of a colonoscopy just so that it doesn't conclude on quite so vividly painful a note. You should want to immediately withdraw your hand from the cold water rather than keeping it immersed longer while the water warms slightly. We should want less total pain, not more. It's a mistake to disregard duration as much as we do. And the wolfer, you think, is making exactly that mistake, while the savorer rationally sacrifices peak pleasure for enduring pleasure. If she does it rightly, the savorer derives more total joy from her dessert, wisely shaping her behavior to maximize not the peak or the end but instead the entire integral under the line.

This argument errs in two ways.

First, part of the pleasure of dessert is remembering it fondly and anticipating the next one. (Colonoscopies might differ in this respect.) If the wolfer sustains a more positive attitude toward dessert due to his memory of a great peak and a good end, the wolfer multiplies that pleasure in recollection and planning. "Wow, do you remember how great those brownies were? Let's make more next week!" Smelling the next batch in the oven, or putting the Swiss chocolate in his shopping cart, the wolf rekindles his greater bliss.

Second, a life with peaks and valleys is overall better and more choiceworthy than a life at steady medium good. Here I side with Nietzsche against the Stoics. This is, I think, especially true on the positive side, when the valleys are not too low. (I would not wish suicidal depression on anyone.) Given a choice between 2 2 2 2 2 and 1 0 10 -1 0 -- both summing to ten units of pleasure -- give me the second every time. Give me the wolf's sloppy peak bite off the top of the sundae, even if it's finished soon, over the savorer's monotonous, slow licking.

Envy Becomes Pity and Maybe Forgiveness

Now that I am thinking about these matters clearly, I see that our envy of Kate's slow ways is misplaced. As we sit there watching her still eating her brownie, we envy her because we imagine the pleasure of wolfing down the brownie that remains before her. But we should pity Kate instead: In her faux wisdom she will never know that wolfish pleasure. Not only does she deprive herself, but she negligently or recklessly or even intentionally (as part of her pleasure?) teases and torments us wolves, which makes her choice not only prudentially but morally wrong.

But Kate, I promise not to judge you too harshly -- I will forgive you, even! -- if you will share that last bit of the peach cobbler with me. Pretty please?

-------------------------------------

If you enjoy my blog, check out my recent book: A Theory of Jerks and Other Philosophical Misadventures.