In a series of 4 studies published in the Canadian Journal of Experimental Psychology, my colleagues and I examined a phenomenon known as the Keats heuristic, whereby aesthetic statements are judged as more truthful than their less aesthetic paraphrases. We found that antimetabolic statements which follow an A-B-B-A pattern (such as “all for one, one for all”) were judged as more accurate compared to their non-antimetabolic equivalents. Fluency – or the speed of processing – explained the accuracy benefit they were afforded.
Two centuries ago, romantic poet John Keats equated beauty and truth, writing in his poem “Beauty is truth, truth beauty,—that is all ye know on earth, and all ye need to know.” Researchers have been catching up in recent decades. Empirical work has confirmed associations between aesthetics and credibility, perceptions of accuracy, and a general sense of moral goodness.
But why should judgments of morality and truth be biased by aesthetics?
Past work has demonstrated a fluency heuristic in which people’s judgments are biased by the degree to which a stimulus is processed fluently; fluently processed stimuli are judged as more accurate, familiar, and beautiful.
Relatedly, stylistic devices such as rhyme have been shown to both increase people’s understanding of a text as well as their perception of its accuracy. In one study, the researchers found that rhyming aphorisms (such as “Woes unite foes”) were perceived as more accurate compared to semantically equivalent non-rhyming paraphrases (such as “Woes unite enemies”). The authors ascribed this effect to a fluency heuristic, arguing that participants were misattributing the enhanced fluency of rhyming statements as evidence of their truthfulness.
Evidently, the style in which information is communicated influences perceptions of truth. But while it is becoming increasingly clear that subtle linguistic changes can bias numerous judgments, the influence of many stylistic devices remains unexamined. An interesting example is that of chiasmus, in which at least two linguistic constituents are repeated in reverse order, following an A-B-B-A pattern. John Keats’ famous expression – “beauty is truth, truth beauty” is in fact chiastic.
The most frequent subtype of chiasmus — the one manifest in Keats’ expression — is known as antimetabole, in which the reverse-repeating constituents are words. Antimetabole has been used extensively in both written and spoken forms, from literary classics (“All for one, one for all”), to Shakespearian plays (“Fair is foul, foul is fair”), and presidential addresses (“Ask not what your country can do for you, ask what you can do for your country”).
We thought that at least part of the explanation for the prominence of chiastic patterning is connected with the Keats heuristic and processing fluency.
To test this hypothesis, we recruited samples of 200 and 206 U.S. participants in Studies 1 and 2 respectively. For every antimetabolic statement (such as, “It’s nice to be important, but it’s more important to be nice”) a non-antimetabolic statement was generated by replacing content words with synonyms that would break the A-B-B-A pattern (e.g., “It’s nice to be important, but it’s more critical to be kind”).
A pilot study was conducted to establish that 1) antimetabolic statements were unfamiliar to participants and 2) that the statement pairs were judged as synonymous in meaning. For all studies, we only maintained statements that met this inclusion criteria.
Study 1 included a total of 18 statement pairs, while Study 2 included 25 pairs. Participants rated the accuracy of each statement, one at a time, on a scale from 1 (not at all) to 9 (very much). In Study 2, this was reduced to a 7-point scale.
Following all accuracy judgments, participants were asked whether statements that use parallel language describe human behavior more accurately than those that do not (yes/no). In Study 2 and onwards, this was reworded to “poetic” language. Participants concluded the study by responding to demographic questions (e.g., sex, ethnicity).
In Study 1, we found that antimetabolic statements were judged as more accurate compared to semantically equivalent non-antimetabolic statements. Nevertheless, this accuracy benefit was small, and was not replicated in Study 2.
In Study 3, we increased the incongruity between antimetabolic and non-antimetabolic statements. Instead of replacing content words with synonyms that break the antimetabolic structure, we modified either the entire first or second half of each antimetabolic statement. For example, the expression “Little things make perfection, but perfection is no little thing” was paired with “Small details make flawless results, but perfection is no little thing.”
Study 3 included a total of 205 U.S. participants and followed the same procedure as in Study 2, using 30 statement pairs. We once again found that antimetabolic statements were judged as more accurate compared to their non-antimetabolic paraphrases.
Study 4 sought to explain this phenomenon. Due to their reverse-repetitive structure, it could be that antimetabolic statements were being processed more fluently compared to their non-antimetabolic counterparts. Thus, in Study 4, we measured participants’ response times to all statements, as a proxy for processing fluency.
We conducted Study 4 in-lab as opposed to online, recruiting a sample of 186 undergraduate students from the University of Waterloo. We used statements from Studies 1 and 2, for a total of 37 statement pairs. Participants provided their ratings via keyboard.
As in Studies 1 and 2, we found that antimetabolic statements evoked stronger feelings of accuracy than non-antimetabolic paraphrases. The ease of processing associated with antimetabolic statements partially explained the accuracy benefit they were afforded. In other words, antimetabolic statements elicited faster response times than non-antimetabolic equivalents, and this predicted their accuracy ratings. Importantly, when controlling for the frequency of key content words, the observed effects remained unaffected.
It was also interesting that despite the majority of participants indicating that statements featuring poetic language are no more accurate than statements that do not, they demonstrated behavior which suggests otherwise.
A legendary example of antimetabole is in John F. Kennedy’s inaugural address, where he famously said, “Ask not what your country can do for you. Ask what you can do for your country.” Although he repeated this sentiment many times in numerous ways, this is the one that is best remembered and most widely reproduced.
Kennedy’s aphorism is not alone in its popularity. Chiastic expressions can be found in proverbs, music, literature, poetry, and religious texts—going back as far as language has been recorded. For example, lyrics like Snoop Dogg’s “My mind on my money and my money on my mind.”
Our findings support that the chiastic (A-B-B-A) pattern is an important contributor to their ubiquity, that the Keats heuristic is affiliated with the perceived accuracy of many such expressions, and that processing fluency is implicated in the Keats heuristic.
The research, “Beauty and Truth, Truth and Beauty: Chiastic Structure Increases the Subjective Accuracy of Statements”, was authored by myself (@ManeYakoubian), Alexander C. Walker, Konstantyn Sharpinskyi, Garni Assadourian, Jonathan A. Fugelsang, and Randy A. Harris.