Compellingness

intermediate2 min readcomputer science · education · psychology

Compellingness is the feeling that something is interesting, relevant, authoritative, or timely enough to deserve attention. The authors use early story-understanding research to explain how such feelings help select what to investigate; they do not establish whether a claim is true.

THE IDEA

GOING DEEPER

Man versus Machine

Some of the older research on the importance of feelings for processing information comes from an interesting source: computer scientists in the 1970s.

Back then, several linguists were trying to teach computers to understand stories. By the late 1970s, they had hit a problem. As described in Roger Schank’s foundational paper “Interestingness: Controlling Inferences,”2 their language models could read a story that involved a person walking down the street, getting a lemonade, and spotting a secret invading army in the woods, and the computer would spend as much time learning about the lemonade as the invading army. Schank proposed that something that humans do quite naturally—using the feeling of interest to direct attention—was not happening with the processing modules, and as a result, the computer sucked at interpretation.

Because we are interested in more than stories and events, instead of “interestingness,” we use the term “compellingness,” which covers not only things of interest but a mishmash of feelings regarding things like authority, relevance, and timeliness. But the underlying principle is the same. Properly conceived, our feelings can be a strength, not a weakness, and one that helps us outperform more methodical processes.

When to use it

Reading a news article and instinctively prioritizing the most relevant or surprising information over trivial details.
Deciding which topics to study deeper based on a sense of curiosity or relevance rather than a strict syllabus.
In a debate, notice which points feel compelling, then examine the evidence and assumptions behind them.
Reviewing a complex dataset and using gut feelings to identify patterns worth investigating before applying statistical methods.

FROM THEORY TO EVERYDAY LIFE

In practice

Computer vs Human Story Interpretation

In the 1970s, computer scientists found that language models reading a story about a person walking, getting a lemonade, and spotting a secret invading army would treat the lemonade detail as equally important as the invading army. Humans, guided by the feeling of interest, naturally emphasize the significant elements, demonstrating the power of compellingness.

✳

A MOMENT TO REFLECT

Put it to work.

When was the last time your feeling of interest led you to focus on a crucial detail that a systematic process might have missed?
How could you deliberately harness the feeling of compellingness to improve your decision-making or learning?
In what situations might relying on feelings of compellingness be a weakness rather than a strength?