The Chronicle Review

The Power of Mental Pictures

‘Framing’ may have fallen out of favor in Washington, but the field that birthed it—metaphor studies—is going strong

September 01, 2014

Around about 2005, you couldn’t shake a stick in Washington without hitting a political consultant who was focus-grouping your stick-shaking metaphor to see whether it provided better "framing" than his opponent’s. The inspiration for the framing craze, George Lakoff’s book Don’t Think of an Elephant!, argued that the Democrats lost in 2004 because they ignored the importance of frames: subconscious structures that determine why people vote the way they do, and that can be activated through abstract linguistic triggers like "family values" or "death tax." Lakoff, a Berkeley professor of cognitive linguistics, became a sought-after consultant in his own right, advising everyone from John Kerry to the U.S. Senate to local unions.

Over the years, framing lost its luster in Washington, as Obama-era liberals (at least publicly) declared their focus to be better policy, not fancier messaging, and the idea of spinmeisters tinkering with language to influence the nation’s subconscious began to feel a little creepy. Lakoff’s tendency to overreach when it came to suggestions for talking points didn’t help, either. In a 2006 review, the Harvard psychologist Steven Pinker ripped another of Lakoff’s books, Whose Freedom? The Battle Over America’s Most Important Idea, not just for "dubious" neuroscience but also for lousy political science. "One can imagine the howls of ridicule if a politician took Lakoff’s Orwellian advice to rebrand taxes as ‘membership fees,’" Pinker wrote in The New Republic. Lakoff responded angrily, calling the review "a vituperative and underhanded attack."

And yet although "framing" may not be the political buzzword it once was, the field that birthed it—metaphor studies, which draws on cognitive psychology, linguistics, and neuroscience to study the way figurative language structures our unconscious cognition—is still going strong. Metaphor studies dates back to the 1980 publication of Lakoff and Mark Johnson’s book Metaphors We Live By, which argued that metaphors aren’t just poetic figures of speech. Embedded in nearly every sentence we utter, metaphors thread together the abstract and concrete realms, a connection that dictates our thoughts, emotions, and behavior. When we talk about a relationship that's "hit the skids," we're using a culturally determined metaphor ("love is a journey") that may determine how we conceive of love itself. Lakoff and Johnson's work has been validated experimentally in many fields since 1980. And although when I spoke to Lakoff in April, he was dour about the state of metaphor studies, saying a lot of people were still stuck on an idea of metaphor from 2,500 years ago, his research is still applied across a variety of disciplines.

In fact, although Rahm Emanuel banned Lakoff from the Obama White House, research continues on the exact kinds of effects that Lakoff described in his political books—persuasion through metaphoric language—and it continues to show results. Lera Boroditsky, an associate professor of cognitive science at the University of California at San Diego, has written a series of papers on the effect of figurative language, particularly metaphors of space and time, on reasoning. One paper, written with Paul Thibodeau, an assistant professor of psychology at Oberlin College, showed that substituting just one word in a text about a crime wave ravaging an imaginary town—comparing crime to a "beast" instead of a "virus"—completely changed how readers responded to the problem. People who read that crime was a beast were far more likely to advocate putting more police on the streets or locking up criminals; people who read "virus" were far more likely to push for education and social reforms. And yet, when people cited the factors behind their decisions, no one mentioned the metaphor. "People love to think that they’re being rational, and all of us love to think that we’re basing our opinion entirely on facts," Boroditsky told me. "But in fact it was the metaphor that people overlooked."

Although lack of money led Lakoff to shut down his policy-oriented Rockridge Institute in 2008, other consultants continue to apply his theories. Anat Shenker-Osorio studied with Lakoff at Berkeley and worked with him at Rockridge, and in 2012 published a book, Don’t Buy It: The Trouble With Talking Nonsense About the Economy, that runs through the key economic metaphors of the last several years.

She says she understands some of the backlash against framing: "There was a lot of really great insight applied in terms of diagnosing what was wrong and what not to say, and then when it turned to the question of what to say instead, the tools weren’t as applicable."

Her approach, in her work at an Oakland-based political consultancy shop called ASO Communications, is that of a longtime campaign worker, not an academic: talking to people in the field and relying on focus groups and quantitative testing, which she says "wasn’t ... happening in the way it should be" in the headier early days of frame consulting. She’s teamed up with Thibodeau, who ran a study examining the effect of a metaphor that Shenker-Osorio says she’s "had a long diatribe against": income gap, which implies that poor and rich people aren’t interconnected, "when in fact the reason some people are as rich as they are is because other people are as poor as they are." Thibodeau found that describing the problem as an "imbalance" was much more effective at conveying the idea that the economy as a whole was in trouble.

Metaphors work in political discourse because they’re an excellent way of grounding abstractions. People don’t necessarily understand all the nuances of crime or the economy, but they do understand beasts and gaps. "The metaphors allow you an entry level into reasoning about something," Boroditsky says. "They organize information enough that you can relate it to something that you understand."

Even when the information is really abstract, metaphor can help us reason in this way. In their 2000 book Where Mathematics Comes From, Lakoff and Rafael Núñez of UC-San Diego argued that the metaphoric leap from concrete (a segment on a measuring stick) to abstract (the concept of addition or subtraction) is the basis of all mathematics, even at a very high level—suggesting that math is not quite so theoretical and objective as its practitioners would have you think. Núñez and his graduate student Tyler Marghetis published an article last year looking at the continuous function, a calculus equation that textbooks have sought, over the past 100 years, to define without any figurative language: "proceeding," "moving forward," or having a "direction." Instead, the books described continuous functions as a purely abstract set of numbers. But Núñez and Marghetis found that when grad students in math described those same functions, they used words and gestures indicating that they were grounding these abstractions with concrete metaphors of dynamic motion, tracing lines in the air as they indicated functions were "increasing" or moving "to the left."

"That’s this notion that we get from high-school English class that metaphors are imprecise or evocative. And what we’re finding is that metaphors are highly precise," Marghetis told me. "They allow you to take precise reasoning and the understanding that you have of concrete domains and then export them to be used in more abstract domains. So in a way, metaphor actually allows you to think precisely about things that otherwise ... you might have trouble getting a grip on."

People use metaphor in many other situations to understand and convey difficult ideas—in teaching, religious parables, the law (a language rich with metaphor, from "building a case" to all the combative imagery of the courtroom), economics (bull and bear markets, the economy as an engine or a pump or an ecosystem), and also in medicine and clinical psychology. The field of narrative medicine, interested in the story-telling relationship between doctors and patients, has brought new attention to metaphor. Susan Sontag’s 1978 book, Illness as Metaphor, explored how 19th-century depictions of tuberculosis and 20th-century language of cancer stereotyped patients as passive victims. "The most truthful way of regarding illness ..." she wrote, "is the one most purified of, most resistant to, metaphoric thinking."

The sweeping (and metaphorical) notion that you could clean up thought, removing all the messy emotion-laden metaphors, is nothing new. As the writer James Geary details in his book I Is an Other: The Secret Life of Metaphor and How it Shapes the Way We See the World, appeals to a purer, nonmetaphoric language go back to Thomas Hobbes and John Locke, who wrote in An Essay Concerning Understanding, "all the artificial and figurative applications of words eloquence hath invented, are for nothing else but to insinuate wrong ideas, move the passions, and thereby mislead the judgment."

But Sontag’s urgent attack on the imagery of disease, both in Illness and Metaphor and her later book AIDS and Its Metaphors, was important for patients and scholars alike. And medical metaphors, particularly for cancer, continue to carry some troublesome connotations. After Richard Nixon declared a "war on cancer" in 1971, the powerful metaphor instantly spread, with all its associated images of white-coated generals battling hordes of invading cells, while the bodies of soldiers—cancer patients, victims once again—fell by the wayside. The metaphor hasn't just stigmatized the "losers" who succumbed, it has also affected treatment, says Rishi Goyal, who teaches in emergency medicine at Columbia University who also teaches comparative literature: "Because of metaphors like that, there’s been a push for more aggressive treatment in cancer—even treatment that has dubious to no benefit—because in the military metaphor, more is better."

Recently, studies have looked at the effect of metaphor in doctor-patient communication, finding that patients report understanding their doctor better when more metaphors are used. Scientists have also analyzed metaphors in hopes of helping doctors improve diagnoses when a patient says something vague like, "My chest feels tight." Metaphor, Goyal says, can be ignored because it "seems like an ornamentation or an exaggeration or hyperbole." But it’s often a crucial way for the doctor to enter into the patient’s subjective experience.

This is particularly true in clinical psychology, where the only way for a patient to indicate symptoms is through language. Metaphors have always been important to psychologists, all the way back to Freud’s cigars. Over the past 20 years, more fields of psychotherapy have become interested in how the analysis and restructuring of patients’ metaphors can help them heal. Everett McGuinty, a clinical therapist in Ontario, has run studies showing the powerful effect of metaphor therapy on young people with autism. He begins by drawing out his patients’ metaphors, discovering their own underlying thought structures. Then he proceeds to neutralize the metaphors. If a patient describes his anxious thoughts as a monkey on his back, McGuinty will ask, "What does it look like, is it your friend, does it have a name? Could we train it? ... Could it become a pet in some ways? Are there any good things about it?" This kind of approach, says Lynne Angus, a psychologist at York University who has studied the use of metaphor in clinical psychology, can be really effective: "Instead of having something attacking you, you then start to befriend it. It starts giving you agency over the experience."

The only time language is truly emptied of metaphor, completely literal, is when a computer’s speaking. Figurative language snarls up computers—think of cyborg Arnold staring at the kids in Terminator when they call him "a couple cans short of a six-pack." When Anatole Gershman, a professor of computer science at Carnegie Mellon University, was in graduate school, he worked on a computer program to extract events from text, and the computer read the headline, "The Death of Pope John Paul Shook America," after Pope John Paul I died. "Our program interpreted that there was an earthquake in America," Gershman told me.

Recently, Iarpa, the intelligence community’s research wing, launched its own "metaphor program" with the goal of teaching computers how to identify and read figurative words. Heather McCallum-Bayliss, the program manager, is a theoretical linguist steeped in Lakoff who told me she wants to use metaphor to understand "what underlying beliefs can be attributed to cultures and how those beliefs vary across cultures." While it’s unsettling to hear a member of the intelligence community remind you that "metaphors can shape the way you think," the teams funded by Iarpa have included some of the most well-respected researchers in computer science and metaphor studies: Boroditsky, Gershman, and Ekaterina Shutova, a computational linguist at Berkeley.

Shutova described her approach as machine learning: addressing the underlying theories of metaphor in a text and building models based on those, rather than trying to hand code manual descriptions of each potential metaphor that could appear, or use a tagging approach, so that the program drops a flag each time it sees figurative language. The systems are still in a testing phase, but McCallum-Bayliss says so far they’ve been very successful, teaching programs to identify metaphors in the four languages they’re looking at (American English, Mexican Spanish, Farsi, and Russian) 70 percent of the time.

None of the researchers I spoke to wanted to speculate about how Iarpa would apply its work, except in vague terms like "data mining." McCallum-Bayliss mostly spoke generally about how it would broaden cultural understanding and eliminate bias by transferring analysis from human to machine. But clearly, systems able not just to extract and analyze metaphors in other languages but also to read figurative text (i.e., all text) much more quickly and more effectively would be hugely powerful tools.

Lakoff finds the Iarpa project exciting, although he says that the government has "a lot of catching up to do, and they’re not putting a whole lot of money or effort into funding the people who’d need to do it." But metaphor studies seems to be moving along with or without Washington’s help.

Britt Peterson is the Word columnist for The Boston Globe’s Ideas section and a writer on culture and ideas in Washington.