THE SCIENCE OF LEARNING TO LEARN

READING: SYSTEMS THINKING

One of the first studies of systems thinking occurred around the same time that Einstein published his theory of relativity. The study took place at the University of Chicago in the early 1900s, and as part of the project, psychologist Charles Judd had two groups of subjects fire some darts at a target submerged under water.

The first group of subjects simply practiced the procedure, repeatedly firing darts at the underwater target some four inches away. The second group executed the same procedure. But they also learned about the notion of refraction, or the way that light shifts when it’s under water.

Then Judd moved the underwater target to a spot twelve inches away, and while both groups did equally well at hitting the target at four inches, only the second group could hit the target with any accuracy at twelve inches.

As Judd argued, the students who understood the relationship between light and water were better able to hit the target in a different setting. They could use their learning to a new context. Because their knowledge was part of a system, their knowledge was more flexible--and they gained a lot more.           

Cognitive scientist Lindsey Richland has written a lot about this idea, and in a landmark paper, she argues that to build concepts, to solve problems, to engage in any sort of critical thought, people need to grapple with patterns within an area of expertise. Richland developed this idea after spending years hunting through a wide body of different academic fields—from math to history—showing that mastery is ultimately defined by systems thinking, by a sense of how structures work in an area of mastery.

“The underpinnings of the ability to do higher-order thinking really comes down to reasoning about relationships,” Richland told me when I visited her in her office at the University of Chicago.

If someone learns more about relationships and systems within math, they have deeper math reasoning skills.

Experts engage in this type of systems thinking a lot. In their fields, dedicated specialists understand how things come together, and so they can look past chaos and complexity and uncover the essence of an idea. Pablo Picasso is famous for once having sketched a bull using just seven lines. Great lawyers have similar skills and can easily find the key argument in a jumble of legal details. For another illustration, just think of the purified elegance of a pop song by the Beatles: By understanding relationships, the band made the complicated seem uncomplicated.

What's more, Richland showed that if people relate what they know, they develop sharper reasoning skills. So, for instance, if someone learns more about relationships and systems within math, they have deeper math reasoning skills. If someone finds out more about the way that historical details couple to each other, they have a richer historical understanding. "Effective learning comes down to thinking about relationships,” Richland argues.

As an example, take learning about the ocean. To develop reasoning, to create a systems understanding, Richland argues that people shouldn't overly dwell on stand-alone facts. Rather, they should examine questions like: What happens to the ocean if the level of salt goes up? What’s the difference between oceans and lakes? How do reefs impact ocean currents?

Like a type of reasoning pill, these are the sorts of questions that push people to develop their thinking about the field--and fully understand an idea or topic or skill. “You don’t just want to be memorizing a whole bunch of stuff,” Richland told me. "To learn effectively, people should be finding causes, finding analogues, finding differences.”

Richland developed her theory based on academic fields like physics and math, and after speaking with her, I was intrigued. So I thought I’d see if her argument extended to something a little less scholarly and signed up for a class on, yes, wine. There are, of course, a number of ways for people to hone their viniculture skills. Someone could globetrot though vineyards or attend workshop or even just sample a lot of wine.

But given Richland’s work, I slipped into a class on how to match wines with foods. I wanted to know: Would thinking about relationships give me richer insights, a better way to polish my knowledge?

Wine expert Amanda Weaver-Page taught the class on a rainy Friday night, and dressed in an all-white chef’s outfit, Weaver-Page started the class by detailing some wine basics. She spoke about issues of acidity, detailing the idea of tannins, which give red wines their sharp flavor. Texture was crucial, too. “Think of a light bodied wine as comparable to skim milk,” Weaver-Page said. “And a full body wine is more like whole milk.”

"Would thinking about relationships give me richer insights, a better way to polish my knowledge?"

Weaver-Page argued that the matching of a wine was about compliments. The food, in other words, should support the wine, while the wine supports the food, a type of nourishment ying yang. That's why lighter wines often go so well with lighter meals like fruit, while heavy red wines support something a grilled ribeye: “Take a light bodied wine with something texturally heavy like steak, and it’s going to overwhelm the wine.”

At first, I was generously skeptical of some of Weaver-Page's points. Like people talking about high-end art or fancy cars, there's a hefty does of exhibitionism that comes with wine talk. But then came the first pairing. A goat cheese salad matched with a Spanish Albarino wine, and the relationships between the two were clear, giving me an insight into the nature of wine that I had never had experienced before. The wine’s essence—soft and lime-like—seemed beyond question.

Then came the next wine—an Australian Shiraz. Weaver-Page paired it with grilled lamb with mint pesto, and again: The wine’s flavor was crystalline—rich, almost lewd, like something from a medieval carnival. When I posed Richland’s theory to Weaver-Page, she nodded in agreement. "Pairings give people a good introduction to the way that wine works," she told me.

In fact, Weaver-Page had a similar experience during her early years of culinary school. An instructor had given her a tannin-filled wine, which made her lips pucker like a grade school kiss. Then the instructor gave her a bite of cheddar. “The fat smoothed out that tannins. It tasted totally different,” she said.

When I left the class some two hours later, there were still gaps in my knowledge. Weaver-Page had also put a lot of thought into the pairings, and if I had picked up a bottle of plonk and matched it up with some McDonald’s fries, I would have had a very different experience. But I could also say with certainty that my thinking about wine had changed. I had a glimpse of what it was like to think like a wine expert, to see the world of wine in a more systemic way.

There’s another admittedly less risky way to learn a system within an area of expertise: Visually.

In many ways, this was the lasting contribution of John Venn. A professor at Cambridge during the late 19th century, Venn was fastidious, fond of detailed lists and rigorous designs. An amateur engineer, Venn created one of the first machines that could “pitch” a cricket ball and might be the only philosopher to strike out members of the Australian national cricket team.

Venn was fascinated by the intricacies of logic, and in a book published in 1881, Venn added an important twist to the study of logic, arguing for a visual approach. So instead of using text to describe a bit of logic, Venn proposed using circles. People needed to “visualize propositions,” Venn argued, and the classic Venn diagram looks like this:

When it comes to learning, Venn’s diagrams underscore an important point: People gain a lot by seeing a visual representation of a system of mastery. When we engage with some graphical form of relational expertise, we often develop important insights.

The cognitive sciences are pretty clear on this point, and in math, studies show that people often learn more if they rely on a some sort of visualization. In the same way, researchers in creativity recommend diagrams as a way to help people find new ideas. Same with science: When people see an illustration of something that they aim to learn, they often post higher outcomes.

Concept maps are a helpful example. A cousin of the Venn diagram, concept maps provide graphical way to grapple with a bit of knowledge. To get a sense of how concepts work--and how they promote a systems understanding--let’s go back to John Venn himself.

So, first, read this short biography of the British philosopher:

Born on August 4, 1834, John Venn is best remembered for inventing the Venn diagram. Early in his career, Venn helped popularize George Boole’s work on logic, which became the basis for computer programming. A University of Cambridge lecturer, Venn also developed a theory of probability—called the frequency theory. Today, just about every statistician relies on the approach. Venn died on April 4, 1923, and in 2014, tech firm Google promoted a version of a Venn diagram on the company’s home page to honor the British philosopher.

Then take a look at the same material represented in a concept map: 

Compare the two approaches to understanding Venn's biography, and it's clear that the concept maps helps people better understand relationships. The concept map suggests, for instance, that the fields of logic and computer programming have similar historical roots. The concept map also makes it easy to see that Venn was not a one hit, academic wonder. His writings also helped pioneer the field of computer science.

In the biographical text, however, these sorts of links are less clear. The linear nature of the text makes it hard to see these sorts of interwoven relationships. Certainly, I barely noticed them when I read the encyclopedia text for the first time.

One of the main benefits of the graphic organizers is that they make it easier for people to think in a more connected way...

Researcher Ken Kiewra has been studying different types of concept maps for years, and he argues that one of the main benefits of the graphic organizers is that they make it easier for people to think in a more connected way, showing deeper associations within an area of knowledge. “Graphics organizers help people put the pieces together,” Kierwa told me.

In his own life, Kierwa uses the learning tools all the time. At work, he uses the graphic organizers for any sort of writing or research project. At home, Kiewra also often relies on them to make important decisions, and he recently hauled out a version of a concept map to help his son sort out some decisions about college. “Things will just pop out,” he told me.  

When it comes to graphical representations like concept maps, technology can a lot. The technical devices that cause information overload can often help us map our way out of that overload.

The Atlantic’s James Fallows provides useful advice on this point. One of the nation’s most well respected journalists, Fallows often reviews information management software, and he has long sworn by a concept mapping software known as TinderBox. The tool helps organize files in a way that draws links across fields and topics, and Fallows describes it simply as a “software-for-thinking” program.

In a similar vein, writer Steven Johnson has long been a proponent of a concept mapping tool called DEVONthink. He argues that the software offers “connective power,” and it helps him spot relationships that he would not have uncovered otherwise. When Johnson uses DEVONthink, “larger idea takes shape in my head, built upon the trail of associations the machine has assembled for me.”

For my part, I’ve become a devotee of the writing software known as Scrivener. For me, it’s software-for-writing because it takes more of a concept mapping approach, offering a virtual cork board and a more networked management system. Not surprisingly, both Fallows and Johnson also rely on Scrivener, and at least like Fallows, I tend to use the software only for large projects like books. In other words, there has to be a lot of text to make the software worthwhile.

This last point is important. Because if we have lots of data, we need robust tools to sort our way through that data. If we have a lot of trees, we’re going to need some device to uncover the connecting forest. This is why we need to learn relationships, too. They’re ultimately what help us learn.

In this chapter, we’ve been pretty focused on relationships, studying ways to improve learning by looking for deeper systems. We’ve examined ways to improve learning by mixing up our practice—and glimpsed at the ways that activities like hacking can provide a type of understanding.

This is all important. But we’re also missing something. More specifically, we’re missing a way to understand how exactly skills and knowledge relate to each other, and it’s that idea that brings us to analogies, or the way that we learn through comparison. Put differently, relational thinking has a driver, and that driver is analogical thought.

Analogies are at the heart of understanding relationships, of grappling with systems of thought, and they can help us solve any sort of new or enduring issue.

Granted, analogies can often seem like an esoteric thing. They often spark memories of IQ tests (Nest is to bird, as doghouse is to ______) or bizarre turns of phrase like term “the pecking order.” But analogies are at the heart of understanding relationships, of grappling with systems of thought, and they can help us solve any sort of new or enduring issue.

As an example, let's consider Tom and Ray Magliozzi. For years, the two brothers had a radio show on WBUR in Boston in which they talked about car repair. Called Car Talk, the show typically featured the two brothers jawing like two teens in the back of math class—cracking bad jokes, teasing each other, throwing out double or even triple entendres.

“Don’t drive like my brother,” Tom would say.

“No, don’t drive like my brother,” Ray would say.

In between the slapstick humor and goofy jokes, the brothers would solve car problems. One day, for instance, a woman named Mary Gordon Spence phoned the Magliozzis. From her home in Texas, Spence explained that every time that she tapped on the breaks of her Mazda Tribute, there’d be a loud squeak. It’s “a high-pitched, one-note sound,” Spence told the brothers.

The brothers listened and then declared: There's problem with the power brakes vacuum booster.

This is impressive. To review the facts again: The brothers had never seen Spence’s car. They didn’t know if Spence's Mazda was leaking oil or had an old timing belt or if there was rust in the radiator. But still, the Magliozzis managed to solve the problem.

So what happened? What, mental trick did the brothers use to crack the problem?

Well, a lot of the answer goes back to analogy. Since the brothers couldn’t physically evaluate the Mazda, they made a comparison in their mind. They thought about other experiences in which they had issues with a Mazda with a power brake vacuum booster. In the simplest of terms, the brothers thought of an analogy.

To anyone who listened carefully to the show, this approach was pretty clear. When the Magliozzis helped a woman with rust on her old Subaru, they talk about rust on their old cars. When someone called in from Africa, the brothers discussed their own visits to Africa. And when a man’s electric winch died, the Magliozzis began detailing a similar problem, declaring “everything you say fits.”

At the heart of an analogy is a comparison.

To a degree, analogies can seem like just another type of relational thinking. But when it comes to learning, the approach goes deeper than that, and at the heart of an analogy is a comparison. More exactly, analogies make us find similarities and differences. They help us understand things that are new or different, and thus a very powerful learning tool, as we will see.

To get a better sense of how analogies help people learn, let’s consider this well-studied problem: Imagine, for a moment, that you’re a doctor, and a patient comes in one morning with a deadly tumor in her stomach. There’s no way to operate—the patient will suffer too much blood loss. Luckily, one of your colleagues recently created a tumor-killing ray—let’s call it the Vapor 3000—and with just one, long blast, the tumor will be gone.

There’s a crucial hitch, though. If you fire the tumor-killing rays at full blast everything around the stomach—intestines, liver, colon—will also become vaporized. In other words, you can’t shoot one huge blast to solve the problem. But then again, if you fire a weak blast from the Vapor 3000, nothing happens to the tumor. Just one low-power shot just isn’t enough.

So what do you do?

Over the past forty years, psychologist Keith Holyoak has presented this problem to hundreds of different people. The riddle has come to define his career, in fact, and the answer rests on a concept known as convergence. Specifically, the solution is to fire short blasts of rays from the Vapor 3000 at the tumor from various angles.

There are a number of ways to help people arrive at this solution, and people with a background in engineering have an easier time. Not surprisingly, advice helps a lot, too, and if someone like Holyoak gives someone a tip, they're much likely to find the answer.

But what Holyoak has shown over the decades is that analogies provide one of the best ways to help people learn. They dramatically improve people's ability to crack the riddle, helping us recognize patterns. Holyoak first demonstrated this fact some four decades ago using the tumor problem—and evidence for his argument has grown far more robust over time.

Most recently, Holyoak showed some subjects an animation that depicts an analogous solution to the tumor problem. Think of multiple cannons in a circle firing on a castle, and after seeing the video, people were far more likely to provide the right solution. “The continuous representation forced people to think more in terms of the analog,” Holyoak told me when I reached him in his office at UCLA

As a learning tool, analogies require some attention, to be sure. Holyoak recommends, for instance, that people rely on a source analogy that they know well. The idiom—“it cuts like a knife,” for instance, works as an idiom because people are pretty familiar with knives.

If you own chickens, for instance, the expression--"the pecking order"--makes sense since chickens have pretty strict social hierarchies. (Indeed the expression, "the pecking order" appears to have entered English at a time when many people typically owned backyard chickens.)

When using analogies to learn, people should outline the exact similarities between the two things or ideas. In the tumor problem, for example, people solve the problem more easily if the analogs are presented next to each other, if not side by side, according to Holyoak. To give a similar but different example, consider that canons are to castles like a Vapor 3000 is to a tumor.

But perhaps the most important thing about analogies is that they help us understand new concepts and ideas. They give people a way to understand something that they're not particularly familiar with, and people should use analogies to get their heads around something new in the same way that we can use Latin to understand Italian or Spanish to grapple with Portuguese.

The phrase Uber but for…  is a great example, and people will often reference the car sharing company to describe start ups. The company Blue Apron has presented itself as the Uber for high-end cooking. The dry-cleaning company DRYV has been described as Uber but for dry cleaning. There’s also now an Uber for haircuts—and an Uber to shuttle kids around.

Smart marketing companies know this, and they are famous for using analogies to introduce new products. The insurance firm State Farm has relied on the jingle: Like a Good Neighbor, State Farm is There. Politicians do this all the time, too, and policymakers sold the notion of a “three strikes” crime law based on a baseball analogy, as writer John Pollock argued in his wonderful book Shortcut.

Analogies can also serve a bridge between two ideas or concepts.

You can think of analogy, then, as the rightful mother of invention. It's a way for us to develop new ideas, to create unexpected links, and it turns out that the history of creativity is littered with analogical twists. Johannes Gutenberg invented the printing press after seeing a wine press. The Wright brothers studied birds in order to build the world’s first airplane. Twitter is half SMS, half social media.

Analogies can also serve a bridge between two ideas or concepts. Most people are familiar with Romeo and Juliet, for instance, and so an analogy makes it easier to explain the musical West Side Story: Just think of a 1950s version of Romeo and Juliet set in New York City.

Another example is the C.S. Lewis novel The Lion, The Witch and the Wardrobe. One easy way to explain the plot is to reference the Bible, and the book is a fantasy-novel version of the New Testament. The film Thelma & Louise? Actress Susan Sarandon starred in the 1980s blockbuster, and she describes it well: It’s a “cowboy movie with women instead of guys.”

The bottom line is that analogies help us understand. As a learning tool, they give us a more complex sense of a body of knowledge, and if you want to get better at anything, look for an analogy. For an example, take social media. Let’s imagine that we want to improve how we use Twitter and Facebook.

So an analog to social media is face-to-face interaction, and it’s pretty clear what works in offline engagement. First, no one likes someone who dominates the conversation. Second, we generally like people who are good listeners. Third, if someone is giving us money-saving advice, we tend to pay attention.

So apply these same three principles to social media. First, don’t post too much on social media: People don’t enjoy hearing about where you parked your car. Second, make sure to comment on other people’s posts, and generally speaking, people will reciprocate. Third, share things that will help out others.

Even weak analogies have their own sort of power...

Granted, analogies don’t always work. Sometimes there’s not much of a similarity. It’s hard to make a robust link between, say, the President of the United States and a set of car keys, or a goldfish and Mount Kilimanjaro.

Yet even weak analogies have their own sort of power, and when people toy with a concept or category, they can spark creativity. Some comedians like Steven Wright built their careers out of toying with analogies. “It's a small world, but I wouldn't want to paint it,” Wright once stated.

Jerry Seinfeld wasn’t all that different: “I was the best man at the wedding. If I’m the best man, why is she marrying him?”

Same for the Magliozzi brothers, and no long after the first call from Mary Gordon Spence about the Mazda, the two brothers reached out again to her again. The brothers wanted to make they sure that they provided Spence with the correct answer to her car problem.

“So was it the vacuum brake booster?” one of the brothers asked once they got her on the phone.

“You know, I would not have called y’all,” Spence said. “Unless you were going to tell the truth, and you were right on. So right on.”

Spence did have one complaint, though. Without the noise from the brakes, she couldn’t tap out songs anymore like Jingles Bell. “I get so bored now driving down the street.”

The brothers laughed and thought of an analogy, a new twist on the problem.

“I'm going to suggest that you take up the harmonica.”

As a learning tool, analogies work because they make us ask a specific set of questions: How are these things similar? What makes them different? How are they comparable?

In other words, analogies help us understand categories. They make us think about groups and what constitutes a group. When people say that apples and oranges are both fruits, for instance, they’re relying on a type of analogical thinking. They’re matching up the attributes of apples and oranges—both have seeds, come from trees, have a type of flesh—to declare them to be fruits.

Another example is dogs. While a furry malamute and a five-pound pug look almost nothing alike, we have no problem calling them both dogs because we understand the analog that connects them. We understand that both animals have certain things in common—social mammals with noses, tails, legs, and sharp teeth.

Analogies help to sharpen the distinction between different ideas or things.

We’ve come across the value of similarity before when we discussed Robert Goldstone and the problem with the king and his four daughters—and one of the reasons that people should mix up their learning is that it pushes people to think in commonalities. Specifically, we get a better sense of the category of fruits if we come across various fruits. Likewise, we better understand the category of dog, if we come various dogs.

Another way to understand this idea is that analogies help to sharpen the distinction between different ideas or things. They provide a compare and contrast approach to learning.  

Take the radio show Car Talk, which we discussed in the previous section. At first glance, the show might seem pretty revolutionary, at least for NPR, given the goofy jokes and practical subject matter. But make a comparison, and it turns out that Car Talk was a pretty normal outgrowth of NPR's programming history. Before Car Talk, for instance, there was Garrison Keiler’s Prairie Home Companion, which was also part vaudeville, part stand-up comedy.

Another example is Einstein, who we discussed at the start of this chapter. We can learn a lot by comparing him against other great physicists. By seeing similarities and differences, people get a keener sense of a thing. Relative to other top physicists, for instance, Einstein was also much less of a dedicated mathematician. Einstein’s contemporary Paul Dirac had equations named after him. But Einstein, not so much.

Still doubtful? Take a study that occurred some years ago at a business training sessions when a group of managers and aspiring managers all piled into a room. Like so many business training seminars, there was a training packet with some sample cases provided to the group, and they were supposed to read the cases, which revolved around the notion of contingent contracts.

As an approach, contingent contracts are generally pretty helpful. When the contract is conditional on certain actions or outcomes, both parties typically have more flexibility. But for all sorts of reasons, people tend not to use contingent contracts in actual negotiations. People are not aware of them--or they just don’t understand them. This training aimed to address that issue, and all of the individuals had to read the training packet before they began to role-play their negotiation.

The use of the more analogical prompt had a dramatic effect.

A few psychologists oversaw the training, and they slipped one small tweak into the session. One half of the consultants would just “describe” the case studies. The other half had to “think about the similarities” of the cases.

It wasn’t much of a difference, only a few words, really. But the use of the more analogical prompt had a dramatic effect. It pushed a compare-and-contrast approach, and the second group was almost twice as likely to use contingent contracts. They also understood the underlying idea a lot more.

Dedre Gentner was one of the psychologists who worked on the negotiation training study, and I met up with her recently. It was in the hallway of a drab conference hotel. We were both getting coffee.

When I indicated an interest in analogies, Gentner pointed at me excitedly. "If we see the same thing over and over, that's a good way to get started. But if you don't see more dissimilar things, basically, you'd better stay in same village your whole life."

"But analogies are hard,” I countered.

Gentner nodded. "But analogies are what allow you to take knowledge on the road."