Uncomfortable ideas and disfluent expression affect us similarly

Cognitive fluency

Integrating the research on cognitive fluency and cognitive dissonance can enrich our understanding of the cognitive strain (or excessive disfluency) produced by convoluted expression. I’ve extensively discussed  research on cognitive fluency-disfluency, whose basic lesson is that when a message is understood effortlessly it is more believable. Daniel Kahneman in his landmark work in cognitive psychology, Thinking, Fast and Slow (2011) at pp. 62 – 64, provides the following advice on minimizing cognitive strain in persuasive writing:

• Maximize legibility.
• Do not use complex language when simpler language will do.
• Make the message memorable.
• Choose sources with names that are easy to pronounce.

Cognitive dissonance

The term now part of the vernacular, cognitive dissonance, a social-psychology research program started by Leon Festinger in 1956, refers to our aversion to disharmonious ideas, but there’s unfortunately no quick way to understand what disharmonizes ideas. You have to grasp the concept from key experiments. I present two, displaying the breadth of the cognitive-dissonance concept:

In the “$1 and $20 experiment,” subjects performed a boring task, which they understood as the experiment’s real purpose, and they then sought to persuade another subject to participate on the ground that the experience was interesting. One group was offered $1 and the other $20 for their persuasive efforts (today’s values would be inflated by a factor of 7.5). Both groups subsequently evaluated the boring task’s enjoyability. The then-surprising result, as Festinger predicted, was that the subjects receiving $1 rated the boring task more interesting than did the subjects receiving $20. The counter-intuitiveness of the results is what made cognitive dissonance the most popular research program in social psychology in the 1960s: under the reigning reinforcement theory, the subjects in the $20 condition should have rated the task more interesting, since they were reinforced (rewarded) more for claiming it was interesting. Festinger had predicted the results by reasoning that the subjects in the $1 condition would experience more cognitive dissonance due to the disharmoniousness  between the two beliefs: 1) they had misrepresented a boring task and 2) they had done it for a mere dollar.

In another study, Festinger observed a group of fanatics who believed the end of the world was nigh and sought to prepare for it. When the world didn’t end, rather than relinquish their belief, they elaborated and deepened it by explaining away the disconfirmation and becoming yet more fanatical. To make their beliefs more harmonious, they construed the apparent disconfirmation as confirmation.

While my concern is to apply cognitive-dissonance research to cognitive strain, which is directly relevant to writing persuasively, cognitive fluency also clarifies cognitive dissonance, needing clarification because defining the harmoniousness that reduces dissonance is elusive. Social psychologist Eliot Aronson had proposed that cognitive dissonance comes from conflicts with self-concept, but recent research hasn’t supported this interpretation: choices affect beliefs even when the earlier beliefs are forgotten. (See Coppin et. al, I'm No Longer Torn after Choice: How Explicit Choices Implicitly Shape Preferences for Odors (2010) Psychological Science 21(4) 489 ‒ 493.) Cognitive-fluency theory suggests that disharmonious (dissonant) beliefs are beliefs whose understanding takes effort. They are disfluent beliefs, although the disfluency arises not from the manner of expression, as in cognitive-fluency research, but from the content of the beliefs.

Just as cognitive disfluency is useful in persuasion, so is cognitive dissonance, although the uses of dissonance, like those of disfluency, have been largely overlooked. To discourage the error of ignoring dissonance’s uses, I’ll offer a few obvious examples supporting the position that just as there’s an optimal level of fluency, so there’s an optimal level of dissonance needed to maintain a reader’s interest. It’s well known that skilled readers of fiction prefer complex to simple characters; paradox can be useful in exposition; and implausible beliefs and even logical contradiction have helped make religions popular—as with the Trinity doctrine.

Lessons for persuasive writing

Now for what cognitive dissonance research implies about cognitive fluency. The research on cognitive dissonance conceives it as a drive to reduce an unpleasant arousal state: we’re motivated to reduce dissonance. (Kiesler and Pallak, Arousal properties of dissonance manipulations (1976) Psychological Bulletin, 83(6), 1014 ‒ 1025.) Cognitive-fluency researchers haven’t considered the motivation behind the preference for cognitive ease, but if cognitive-dissonance reduction is due to the motive that also enhances the believability of fluent messages, that has lessons for writers. The difference is that cognitive strain's unpleasantness motivates the reader to reject the disfluent expression, not only to find the fluent more credible. The analogy to cognitive dissonance suggests that when we disbelieve the disfluent, it’s because believing the disfluent is uncomfortable. Since we must believe to understand, unpleasant affect associated with the effort to understand prejudices the reader against the proposition itself even when it’s later expressed clearly.

The implication is that persuasive writers should avoid unwarranted disfluencies even when they're immediately clarified. If a concept is hard to understand without examples, prematurely presented conceptualizations undermine subsequent understanding. It's better to introduce the examples before the proposition they support.

Cognitive strain's unpleasantness supports using the method of successive approximations for introducing complex ideas. To use successive approximation, the writer presents a simplified concept that is subsequently elaborated in a series of changes, each simple enough to avoid cognitive strain.

The dialectic of clarity: Cognitive fluency vies with cohesion

Joseph M. Williams, my favorite writing authority, advises:

Some argue that the harder we have to work to understand what we read, the more deeply we think and the better we understand. Everyone should be happy to know that no evidence supports so foolish a claim, and substantial evidence contradicts it. (Williams, "Style: Lessons in Clarity and Grace" (9th ed. 2006) p. 221.) [HT: Vlastimil Vohánka, Comment to Cognitive Fluency: Simpler Isn’t Always Better.]

Williams’s summation is outdated after researchers reported that difficult texts promote better learning. (See Diemand-Yauman, C., et al., "Fortune favors the bold (and the italicized): Effects of disfluency on educational outcomes." (2010) Cognition.) Williams errs because he confuses Clarity, which measures the likelihood that the message succeeds, with fluency, which measures the effort the reader must expend to receive the message.

Social-psychological research points to two reasons for the divergence between Clarity and cognitive fluency (also called simplicity or cognitive ease). First, some matters are more intelligible and some beliefs more malleable when they engage an abstract (or “far”) construal level rather than a concrete (or “near”) construal level. Then, disfluent items will be more comprehensible. (“Construal-level theory: Matching linguistic register to the case's granularity,” in this blog, and “A taxonomy of political ideologies based on construal-level theory,” in Juridical Coherence, discuss construal-level theory.)

Second and more important practically, a trade-off between fluency and cohesion should discourage writers from always opting for the most fluent sentences, which must be bought at the expense of cohesion. The second reason explains disfluent clarity in legal writing better than the first because the writer should approach messages calling for abstract construal amelioratively, by imposing greater cohesion or greater Concision to complexify the text.

Joseph M. Williams himself best pinpoints the tension between fluency and cohesion.

The problem—and the challenge—of English prose is that, with every sentence we write, we have to strike the best compromise between the principles of local clarity and directness and the principles of cohesion that fuse separate sentences [ideas] into a whole discourse. But in that compromise, we must give priority to those features of style that make our discourse seem cohesive, those features that help the reader organize separate sentences [ideas] into a single, unified whole. (Williams, Style: Towards Clarity and Grace (1995) p. 48 [emphasis added].)

My change—“ideas” for “sentences”—emphasizes that consolidating the information contained in several sentences into a single sentence can contribute to cohesion.

The diagram below depicts the interactions leading from their causes to the two substantive Writing Virtues. (I save for another day the interaction between cohesion and omission, causing Concision).

Open in separate window

The diagram shows that writers must balance fluency and cohesion for Clarity; omission and cohesion for Concision. The centrality of cohesion, which contributes to both Clarity and Concision, helps explain why Williams is correct to stress cohesion over fluency (“local clarity”).

Confusion of fluency with Clarity—or at least, emphasis of fluency over cohesion—is the main technical deficiency promoted by the “plain writing” school.

"Plain talk" writing: The new literary obfuscation

“Plain-talk” writing has replaced pretentious writing as the main stylistic mannerism impeding thought. More than a half century ago, George Orwell identified vague abstraction and stale imagery as contributors to political bedevilment: they are the means for making the vile acceptable by concealing its substance. The object of Orwell’s scorn hasn’t disappeared. Politicians and their sycophants still substitute high-flown cliché for penetrating depiction, but that form of literary dishonesty is, today, overshadowed by the abuse of cognitive fluency—by the cult of simplicity. This mode’s mainstay is the non-sequitur; its object of concealment, logical irrelevance; its mechanism, the short, plain sentence. When the new obfuscation becomes pedagogy, writing teachers present its virtue as that of writing as you talk; they call the style “conversational.” It demonstrates that concreteness and vagueness are entirely compatible.

Everyone knows you can’t write efficaciously the same as you talk. So, common sense revises the plain-talk project—using the simple and illogical expressional methods the advocates purvey. A writing blog, CopyBlogger, advises—to the applause of commenters—“Write like you talk, except better. Better words, better arrangement, better flow.” As if this advice were informative.

As a rule, no examples are given, and some of this style’s most ardent practitioners may deny their practice of “writing as you talk.” Writing teacher Wayne Schiess responded to Dr. George D. Gopen’s disparagement of this advice by calling his argument a straw man. Wayne had never heard this advice.

Blogger Luke Muehlhauser provides the rare express example of writing as you talk, and his example ably, if unwittingly, demonstrates how this approach to writing undermines lucid thought: (1), below, is Muehlhauser’s rendition of how a writer would ordinarily state a thought; (2) is Muehlhauser’s recommended rewriting, designed to combine the clarity of writing with the readability of talk:

(1) Perhaps the toughest intellectual work we must do regarding European reconstruction is to realize that it can be achieved through nonpolitical instrumentalities. Reconstruction will not be politics, but engineering.

(2) We have a tough job ahead of us. We need to figure out how to reconstruct Europe. It won’t happen with political forces. The European reconstruction will be a matter of engineering, not politics.

The plain-talk version, (2), is more cognitively fluent than is (1): it deftly hides the contradictions and vagueness baldly evident in (1). First, reference to “instrumentalities” in (1) impels readers to seek to identify them and calls readers’ attention to the merely negative characterization of the “instrumentalities” as “nonpolitical.” Second, the reader of (1) naturally demands to know how “we” are supposed to act through “nonpolitical instrumentalities,” when “politics,” after all, denotes our means for consciously coordinating the actions of numerous persons. Third, if realizing that Europe can’t be reconstituted through politics requires tough intellectual work (it actually was reconstructed through the very political Marshall Plan) the writer isn’t entitled to announce the conclusion in advance of the required work. These objections, occurring naturally to the reader of (1), make that version clear but hard to read. The reader tries to make sense of it, in the face of signals that (1) is false, and readers find known falsehood harder to understand than probable truth.

The “plain-talk” version, (2), expresses the same information contained in (1). The difference is that the clauses in (2) are poorly connected. Although (2) urges readers to figure out how “we” can reconstruct Europe, the inconceivability of collective action being nonpolitical is pushed from the foreground by replacing nonpolitical instrumentalities, through which we act, with nonpolitical forces, which happen. Furthermore, the unexpressed connection between, on the one hand, the conclusion about Europe’s nonpolitical reconstruction and, on the other, the intellectual work from which the conclusion follows, hides absurdity, that of announcing in advance a conclusion of work undone.

The integration fostered by (1)’s concision fosters skepticism of its flawed reasoning. The disjointed “conversational” style of (2) makes the flawed reasoning easier to overlook. Whether Muehlhauser prefers this outcome is unclear.

How new is cognitive fluency?

Except for the Baker law-review article discussed in the Disputed Issues entry on cognitive disfluency, and the Disputed Issues entry applying cognitive-fluency principles to citation formats, the legal-writing world has paid scant attention to the spate of cognitive-fluency research, which appraises simplicity’s benefits and drawbacks for document reception. Plain-language blogger Cheryl Stephens captures what may be the chary outlook of many legal writers:

Scientific research has expanded so much in the last 20 years that plain language practitioners could not keep up. Money for research is needed to ensure that plain language procedures take advantage of current scientific discoveries. The most significant of these seem to be in the new area of study: cognitive fluency.

Another likely source of neglect is a prevalent belief that cognitive fluency is but a fashionable name for well-known effects. The cognitive-fluency results are new but not hard to understand, yet embodying the results in crisp recommendations is elusive, requiring an understanding of the tension between the writing Virtues Clarity and Concision, as their reciprocal modulation balances fluency and disfluency.

Cognitive fluency can seem like old hat because writers have long appreciated the value of minimizing mental effort for comprehension. Much of the recent findings’ novelty lies in in the advantages of disfluency; but even regarding fluency’s advantages, the research differs from traditional understanding, where avoidance of unnecessary complexity is based on the reader’s limited capacity to maintain multiple thoughts in a conscious state simultaneously, a rationale defining simplicity as well as justifying it. At least as long ago as 1852, when philosopher Herbert Spencer wrote The Philosophy of Style, this limited-capacity concept underpinned the rationale that the less capacity readers must allocate to decoding a communication, the more they can allocate to thinking about it. Readers were also expected to be less likely to misunderstand the simple, since it left spare capacity. The Disputed Issues entry “A rare shortcut to better writing” applied the hoary theory of limited-capacity attention to writing’s production, to explain how faster typing improves it. Science had seemingly vindicated the limited-capacity theory when psychologist George Miller published his finding that humans had a limited short-term memory capacity that varied between five and nine bits of information, as when a tester reads a digit series, one number per second, and few subjects will be able to remember more than nine or less than five. Miller’s finding this consistent limitation of conscious apprehension—Miller’s famous “magic number seven plus or minus two”—ensured that the digit-span test would remain part of standard intelligence testing, despite the low correlation with general intellect.

The past decade’s cognitive psychology retains the concept of working memory, but reconceptualizes it as the person’s skill in directing attention to recently conscious or related thoughts, which, hypothetically, are “activated” but unconscious. The subject’s preconscious thoughts—to use Freud’s term for ideation not conscious but amenable to being made so—are accessed in experiments where the subject is diverted from a memory task by subsequent attention-consuming operations. An easy test of this kind is given during standard psychiatric mental-status examinations, when the tester directs the patient to recall three words, which must be recited at the end of the examination, during which the tester elicits unrelated information. That the important component of working memory isn’t limited by fixed storage implies that we can’t deduce mnemonic efficiency from simplicity (which is to say, from cognitive fluency). Here’s an example—compare (1) and (2):

(1) Sentences can be short. They can also be long. This is a good thing. Lack of variety is wearying. It may drive you to distraction.

(2) It’s a good thing that sentences can be short or long, because lack of variety is wearying and may drive you to distraction. (H/T: Mark Nichol, Daily Writing Tips [for the examples].)

The four-sentence version (1) is simpler, its simple sentences bereft of complicating structural nuance. Speaking theoretically, the complex sentence (2) activates more unconscious ideas, inducing a more powerful working memory, not one limited to the simple sentences’ smaller ambit.

If the clearest prose isn’t the most fluent, if clarity is an optimum on the fluent – disfluent dimension, then the advantages of clarity aren’t those of simplicity. What, then, is the advantage of clarity? The answer might seem self-evident. Obviously, it might be thought, a writer wants to be clear so that he will be understood to mean what he does mean. Clarity means easily understood, the “obvious” thought continues, and the easier it is to understand, the more likely it will be understood. But this is fallacy. What requires less effort to understand is not, in logic or in fact, necessarily clearer, more likely to be understood—not if greater effort is forthcoming. This is the nontraditional conclusion on which cognitive-fluency and working-memory research converge.

Richard Posner versus Bryan Garner on citation formats: The verdict of cognitive-fluency research

Builds on “Cognitive disfluency: Simpler isn’t always better.”

The 19th edition of The Bluebook contains more than 500 pages, and lawyers and legal-writing authorities both are divided on how religiously legal writers should follow its dictates. The first issue, dividing many trial attorneys from appellate attorneys, pertains to basics of citing cases and statutes: punctuation, sequencing, capitalization, and statutory-compilation abbreviations. Appellate attorneys usually conform strictly to these rules; trial attorneys deviate, believing that format errors don’t matter provided clarity and consistency are preserved. The second issue, dividing appellate attorneys, involves deviations from The Bluebook that the deviationists believe improve on it, the most common improvement being avoidance of most abbreviations in case names. This is to say, many trial attorneys deviate when they believe it doesn’t matter, whereas the dissenting appellate attorneys deviate because they believe their way is better. I conclude that attorneys best serve their clients when they follow most citation formats authoritative in their jurisdiction, since familiarity produces cognitive fluency, but for the same reason, they should curb excessive complexity, such as The Bluebook’s arcane abbreviations. If you’re keeping score, Garner wins the contest over the first issue—fidelity to the basics; Posner, the second issue—dismissal of arcana.

Posner (whom I declare wrong on the first issue) submits that most citation formatting is invisible to judges, hence irrelevant. (Richard A. Posner, The Bluebook Blues (2011) 120 Yale L. Journal 850, 853.) In fairness to Posner, he doesn’t directly address whether attorneys should follow The Bluebook; he advocates replacing it with less complex, less encompassing, more libertarian rules. The citation guide that approaches Posner’s preference is the Maroonbook, which governs the University of Chicago law review. Posner instructs his clerks with a short memorandum on formats and style. (Id., at p. 854.) I agree with Posner on the systemic question, but I take issue with his argument, which would justify the laxness of many trial attorneys. Posner maintains that it only matters that the format serves the two functions of citations: instantly revealing whether an authority is worth examination and clearly depicting its address. (Id., at p. 852.)

When applied to advice to practitioners, who’re stuck with their jurisdictions’ rules, Posner’s argument founders on scientific findings that “cognitive fluency”—ease of understanding—increases the believability of writings. (“Cognitive disfluency” has its place in persuasion, but the writer’s leeway to be disfluent shouldn’t be squandered on citation, rather devoted to concision.) Since familiarity makes for fluency, writers should mostly follow the official citation formats, familiar to judges. The “invisibility" of these formats is of little moment: many cognitive-fluency effects occur outside the reader’s awareness.

Posner’s view better accords with science when he objects to The Bluebook's needless complexity. Its “hypertrophy” (Posner’s word) includes post-adjudicative signals without legal significance, such as certiori denied in an old case. (Posner, supra at p. 851.) It also includes information supplied by context, such as the designation “dissenting opinion” within a discussion focused on the dissent. The greatest hypertrophy, accounting for much of The Bluebook’s expansion, consists of obscure abbreviations The Bluebook insists that legal writers apply to case names. The case name is often without significance, but scanning an unpronounceable letter series is “cognitively disfluent,” which will unconsciously prejudice the judge against the writer’s position.

Legal-writing authority Alan Dworsky, who maintains that correct formats signal competence, makes the argument for slavish submission to The Bluebook, including its case-name abbreviations:

Citation form is a litmus test of your credibility. Judging a writer's credibility is hard. Readers draw large inferences from small clues, and citation form is one place they look. Like spelling, citation form is either right or wrong. Especially in citations to commonly cited sources like cases and statutes, where a reader is likely to recognize an error, your citation form should be perfect. (Alan L. Dworsky (2d ed. 1992) The Little Book on Legal Writing at p. 75.)

Impliedly supporting strict Bluebooking, Bryan Garner, surprisingly, favors The Bluebook over an alternative—the legal-writing teachers’ ALWD Manual—regarding the Bluebook’s insistence on abbreviated case titles. Garner doesn’t explain the plus he awards, but the economic signaling argument is the strongest one I’ve seen for hyper-ardent Bluebooking. Judges are practical people—as long as the writer doesn’t flout a Formality—but conceivably, some law clerks fresh out of law review may accept rigorous Bluebooking as a competence signal. I expect these signals to be weightless when the writer systematically replaces most abbreviations with full names.

So, except when court rules expressly dictate formats—sometimes even then—brief writers shouldn't slavishly follow citation formats. The official formats should be followed presumptively, since they bear the advantages of expectation and familiarity, but not when disfluent, like The Bluebook’s arcane abbreviations.

Cognitive disfluency: Simpler isn't always better

“Simplicity” is the new mantra in endeavors influenced by psychology’s half-decade-old wave of cognitive-fluency research. Understanding why the mantra is an over-simplification should also help counter some simplistic plain-writing advocates.

Cognitive fluency refers to mental ease, including ease of understanding, and it has been found to have striking benefits for at least some forms of communication, such as marketing, where fluent writing and branding have sometimes more than doubled sales. Experiments manipulate the cognitive fluency of written material by using small words, short sentences, familiar concepts, and clear fonts, among other means. The resulting fluency often dramatically increases the message’s acceptability. Not only is it better understood but is also better liked.

As in Julie A. Baker’s application of fluency research, presented in her article “And the Winner Is: How Principles of Cognitive Science Resolve the Plain Language Debate”—which claims cognitive-fluency research best answers legalese’s proponents—appliers of these scientific results have largely overlooked the recent research focus on cognitive disfluency, which also carries striking benefits,. (HT: Lawrence B. Solum, The Legal Theory Blog [Baker essay].) While Baker nods to research favoring disfluency, her argument that cognitive-fluency research argues for plain writing discounts it. The pro-fluency research surprises by the strength of fluency’s effects, including the positive effect of a corporation’s simple name on its stock prices. But the disfluency research has its share of startling results, such as the recommendation—flying in the face of plain-writing dogma—that school texts should be harder to read. Disfluent writing, experimenters report, is read with greater retention. Similarly, cognitive psychologists point out that disfluent company names build greater customer loyalty, despite or because of the customers’ initially reluctant embrace. Disfluent writing engenders reasoning deeper and more abstract and thinking less stereotypical (an explanation for asiatic style’s effectiveness with an unfriendly court).

Fluency’s bipolarity is important for legal writing because it shows that cognitive fluency isn’t identical to or even correlated with the writing Virtue Clarity: fluency isn’t necessarily Virtuous. This interpretation differs from most marketers adopting fluency as their mantra and from Baker, according to whom fluency is virtuous except in narrow circumstances—most notably, summarizing opponent theories. The research favoring some disfluency demonstrates how fluency diverges from Clarity: often, somewhat disfluent writing is clearer than the most fluent—when, for example, Clarity benefits from reader’s construing the topic abstractly. (Construal-level theory is another topic of considerable research in cognitive psychology.) Fluency theory offers to integrate the accounts of truth and beauty, and the analogy between the two implies that, as with beauty, the optimal fluency is a compromise between fluency and disfluency, although where that compromise is found may depend on the subject, such as whether it is best understood when construed abstractly or concretely. A recent explanation of fluency effects maintains that absolute fluency isn’t the relevant factor when fluency enhances understanding or persuasiveness; but rather the relevant factor is fluency relative to readers’ automatic expectations.

The desirability of fluency in that pure form advocated by the marketers and some plain-writing proponents—those who would be flummoxed to discover that students might learn more from harder-to-read textbooks—is rebutted by counter-examples, such as these: If sheer fluency is desirable, why do legal writers struggle to avoid brute repetition? If we are attracted to the fluent, how is it an author like William Faulkner—with his complex, lengthy sentences—is the most effective American fiction writer? The automatic expectations of fluency and disfluency are yet uninventoried, but I conclude long sentences and other complexities create an expectation of disfluency, contrasting with the relative fluency of a well-constructed long sentence, whereas blatant repetition creates an expectation of fluency, offsetting fluency's perception.

When to modulate fluency versus disfluency remains subject to the writer’s intuition, although the psychological research can help refine a writer’s choices. But direct advice on how to modulate fluency versus disfluency follows from the Disputed Issues writing-Virtues framework. To write disfluently for Clarity’s sake, tilt the language toward greater Concision, rather than decrease fluency by arbitrary means, such as unclear typography or convoluted sentences. The advantages of selective disfluency don’t justify legalese, which not only is overly disfluent but also verbose.

See also: Richard Posner versus Bryan Garner on citation formats: The verdict of cognitive-fluency research and How new is cognitive fluency?