The Complexity-Extremity Effect
What chocolate chip cookies can tell us about how we should view ourselves—and others.
A couple weeks ago I was sitting at my favorite coffee shop in Oxford’s covered market. The covered market is a patchwork of shops and storefronts in the city center, selling mostly food and trinkets. I was drinking my coffee, reading, and watching other people order at the stall across from the café. It is called Ben’s Cookies.
The more I watched people buy their cookies, the more I came to see this transaction as a microcosm of people’s relationship with happiness and its acquisition. Some revel in the occasion. They show palpable signs of excitement, having planned all day to reward themselves for their more strenuous efforts. Or perhaps they’d just conceded “fuck it, I’m worth it” and made the decision spur of the moment. Others, usually alone, are more furtive. I watch them buy their cookie. When they see me watching, they glance away, avoiding eye contact in the way of people who have just publicly admitted to doing something private and unseemly, as one does when passing the next person waiting to use the toilet.
It occurred to me while sitting there that Ben isn’t really in the cookie game. People aren’t buying cookies. They are buying a guaranteed instant of pure bliss. Like a subtle heroin. The bliss lasts only momentarily but that blissful moment is a sure thing. There’s no pretense to longer lasting satisfaction.
Everyone conducts this transaction in a different way. Nearly all passersby gaze at the stall admiringly, contemplating the tantalizing possibility of momentary satisfaction. Some people are regulars. They need their fix. Some pass by and murmur the name, as if having stumbled upon some dimly recalled temple or church to which others have made a pilgrimage and related the glories of their journey back home.
During busier times, Ben’s Cookies often has a queue. The stall is manned by a single employee, who oversees a pile of cookies in an assortment of flavors. Business is good in the instant gratification economy.
When I sat down it didn’t occur to me to want a cookie. And once it did occur to me to want one, the prospect didn’t linger. It occurred to me a possibility, one I can live without. But then as I saw the consistent trickle of happiness purchasers, my defenses eroded. What was the point of forgoing my own moment guaranteed of joy? What do I gain by omitting that experience from the film role of my life? You know what, fuck it. I’m worth it.
I got a cookie. Double chocolate chunk.
Ben’s Cookies are not the best cookies on the planet. But they might be the most enjoyable. They are potentially the most concentrated in the active ingredients which bring on happiness. They are at once fluffy and dense. The chocolate is perfectly weighted. For each of the six bites it takes for the cookie to meet its demise, it’s like there’s nothing wrong with the world. But the thing is, they are unidimensional. And like a drug—based off my admittedly limited personal experience with heroin—it’s the kind of pure gratification which almost by definition cannot linger, cannot satisfy beyond the moment of consumption. There is no nuance to one of Ben’s Cookies.
In 1982, a psychologist at Carnegie Mellon named Patricia Linville published a paper on what she called the extremity-complexity effect. It is a paper I like very much, because in it she asks her participants to rate their evaluations of various chocolate chip cookies.
The cookies vary across six attributes. I just love how earnestly scientific her description of them is. They are:
number and quality of the chocolate chips,
degree of sweetness and richness,
degree of buttery taste,
fresh or stale,
soft or firm,
crispy or chewy
She also includes five distinct cookie types: “three different store-bought types, one delicatessen type, and one homemade type.”
In one condition of the experiment, she asked participants to rate the five different cookies across all six dimensions. “On each trial,” wrote Linville, “prior to tasting the cookie, subjects read the instruction sheet asking them to ‘think about this cookie in terms of the following characteristics’ for a 1.5 minute period” and listed the full spectrum of possible variation in cookie deliciousness. Participants rated each of the five cookies with an “overall impression” along a line labeled “favorable” to “unfavorable.”
Then, in a second condition, she gave participants the same prompt. But instead of listing all six attributes, she only gave them two to consider. After considering the cookies across these two dimensions for the minute and a half, participants gave their ratings.
What Linville found was that participants who considered fewer dimensions gave more extreme ratings. When they considered only two dimensions instead of six, participants were more likely to rate the cookies at the far ends of the favorability scale. This is the complexity-extremity effect. The more complex your concept of something is, the less extreme it tends to be. The less nuance the concept, the more extreme the judgment.
The brunt of Linville’s argument, alas, is not about cookies. She uses the cookies experiment as a low-stakes version of the effect, to show that it is consistent across many aspects of the human experience. But the larger point she makes in the paper is about people. The more complex our understanding of another person—particularly someone from a social out-group—the less extreme our feelings about them will be.
She uses a similar experimental set up to show that when people are encouraged to judge other people on a larger set of attributes, they are less likely to resort to stereotyped, unrealistically rosy or derogatory evaluations of them.
In a follow-up study, published in 1985, Linville explored the extremity-complexity effect as it applies to one’s own conception of oneself. When we think about ourselves across a range of dimensions, does that encourage of us to develop a more stable, less extreme self-conception? She titled the paper “Self complexity and affective extremity: don’t put all your eggs in one cognitive basket.”
One implication of this idea, which she outlines in the introduction, is that we are more likely to experience swings in our self-conception—how favorably we view ourselves—when our identity is tied up in a single aspect of our lives. Right away, she singles out work. From her point of view, as a professor, if your self-conception is all tied up in the dimension of “scientist,” then how you feel about yourself on a given day is likely to fluctuate based on whether you just got a paper rejected from a journal, or whether you recently won an award, and so on. Where as if you have a multi-dimensional self-conception, you’re less likely to experience these big swings. If something goes wrong at work, then it’s a bummer—but isn't sufficient occasion to reevaluate your worth as a person.
Linville’s main study on this front looked at this effect in 59 undergrads. She began by taking a measure of their self-complexity. Each participant was given thirty-three index cards with a trait written on it (outgoing, rebellious, lazy, etc) and asked to sort them into groups reflecting how these traits apply to themselves. They didn’t have to use every card, and they could put any of the cards into multiple groups. Based on the way they organized these trait cards, Linville gave each one a “self-complexity” score. Then they used a computer to answer several more questionnaires, most of which had to do with how favorably they view themselves.
After completing the questionnaires, the computer presented them an error message—which was programmed by the researcher. The participant was told they would have to do the final assessment (an intelligence test) by hand. The researcher then left the room to check out the “error message” while the participant completes the test. When the researcher returned, she asked the participant whether they want to know how their scores compared to everyone else. She graded the test in front of them and gave them one of two reports: either that they are in the bottom 10% of scores, or in the top 10% of scores. Having delivered this news, the researcher discovered that their was a “glitch” in the computer program. She asked the participant to do the questionnaires one more time. This way, participants did the self-conception test twice: once before getting either good or bad feedback about one of their traits, and then again afterward.
Linville found that the participants who had a higher self-complexity score, were less likely to change their self-conception questionnaire answers in a negative way. In other words, the more dimensions they used to define their identity, the more robust they were to a hit to any single one of those dimensions.
I recently argued against thinking about oneself as a “good person.” Part of the this argument was against essentialist thinking—that the deepest truths about ourselves as simple and binary. No way. We’re complicated. And this series of experiments by Linville is another instance showing how getting into deeper connection with that complexity is beneficial.
But let’s go back to cookies. I think there’s a connection to be made between Linville’s arguments about self-complexity and the Ben’s-Cookies-style transactions for happiness. The connection, in short, is that simple evaluations lead to simple joys.
When it comes to chocolate chip cookies, the densest, gooiest, most chocolatey cookie might give us momentary satisfaction. I watched that occur over and over again and people went up to the cookie stall. But because that kind of gratification is so immediately attainable, it is also fleeting. Easy come, easy go.
Likewise, when it comes to ourselves. It’s easy to rely on simpler conceptions of our own identity as we move through daily life. Whether it’s thinking about oneself primarily as a “good person” or as a “scientist” or “entrepreneur” or whatever. And I think there’s a lot to be said for clarity of identity, for maintaining a center of gravity of what makes you who you are. But I think there’s also a lot to be said for embracing that complexity. A more nuanced understanding of oneself and others evens out our judgments. We’re less likely to swing wildly from favorable to unfavorable. It’s something that I’ve been thinking about recently in my own life. In any high-pressure career or environment, the pressure is always to double-down on what’s being measured as our key performance metrics. But the more we discard the other sides of ourself, the more those metrics govern our sense of well-being. It’s a difficult balance to find. But we must resist the urge to undercomplicate—to reduce our conception of ourselves and those around us to something that fails to capture the idiosyncratic nuance and makes us who we are.
P.S., This week’s episode of Cognitive Revolution features George Lakoff. George is one of the most cited cognitive scientists of all time, with his text Metaphors We Live By being one of the most influential works in the entire field. It is all about how metaphor provides the fundamental structure of cognition. (Say, for example… the complexity of cookies as a metaphor for the complexity of self.) You can find that episode here: