Mind Matters editor Jonah Lehrer chats with Gary F. Marcus, New York University psychologist and head of the Infant Language Learning Center, about how computing, genetic biology and psychology together can help probe the wonders of human language development.
JONAH LEHRER: What first made you interested in studying the development of language in children?
GARY F. MARCUS: I came to language development through early exposure to computers, back in the late 1970s and early 1980s, when I was in grade school. I was one of those kids who took to programming like a fish takes to water, and very soon wanted to push the limits: What could I get a computer to do?
One obvious thing was to try get a computer program to understand language; I pretty quickly figured out just how hard that was—there was no way I was going to get my Commodore 64 to talk—but along the way I developed an abiding interest in human mind and how it managed to solve difficult problems. And to this day I still find it amazing just how good human children are at learning languages. No other creature comes close, and nor does any computer program, even now, 20 years later. How come our memories are so lousy relative to computers, yet our capacities for learning language are so good? That's the kind of question I like.
LEHRER: You pursue a multi-disciplinary approach in your lab, combining everything from linguistics to molecular biology to genetics. What are the advantages of such an approach? Any disadvantages?
MARCUS: I think it depends what you want to do. A lot of scientists succeed by studying extremely narrow questions, until they understand every last detail. But what I'm really interested is the big picture, in how things fit together. And the only way to do to that is to understand, at a serious level, how things like genes and brains relate to things like language and cognitive psychology. So for me, being a committed interdisciplinarian is simply the only option.
LEHRER: What's been the most surprising thing you've found in your years as a scientist? What data didn't you expect?
MARCUS: It's just amazing how engaged babies are. Your average six-month-old human is—by comparison to your average two-day old elephant—pretty ineffectual [physically]. Humans simply don't get up and walk the moment they crawl out of the womb, unlike newborns of so many other species. But study after study has shown that baby humans know a heck of lot about how the world works. They have, for example, a fairly firm grasp of basic physics, and they're constantly looking around them trying to understand the world.
In my own lab, for example, we found that your average seven-month-old baby can learn a basic language-like grammatical rule in the space of two minutes; if we give a baby a series of sentences like la ta ta, mi na na; she can pick up the underlying grammar (ABB) structure right away; by now, several other labs have replicated this basic finding. And it's amazing; we're not paying the babies who are in these experiments, and two minutes is not a very long time, but the babies can't help themselves. They want to know what's going on, and they work hard at trying to figure it out.
LEHRER: Your new book, Kluge: The Haphazard Construction of the Human Mind, critiques the notion that the mind is an "elegantly designed organ." Instead, you argue that we're a kluge, a cobbled together contraption that works, more or less, but has a lot of rough edges. What made you interested in writing this book?
MARCUS: In a lot of ways, Kluge grew out of something I discovered in the course of writing my last book, The Birth of the Mind, which is mostly about the relation between genes and psychology. When I went into it, I had two core beliefs—that our minds are richly structured from the moment that we enter the world, and that the mind was neatly carved up into a set of special-purpose modules that were well-adapted to specific tasks. By the time I'd finished writing the book, I realized that only one of those beliefs—the first—was likely be true. If you look at how the genome works, and how the brain develops, it's screamingly obvious that our minds really are richly structured from the outset. But it's also pretty obvious that most of that structure predates humanity, which means that most of our brain circuitry didn't really evolve specifically for, say, language or conscious reasoning.
Instead, as I immersed myself in biology, I realized that language couldn't have evolved out of whole cloth; it must have been a tinkering with stuff that came before. Once I realized that, I began to look for traces of our prelinguistic history in our current linguistic selves, and soon started encountered example after example of aspects of language that didn't really make sense, except in terms of historical accidents. Why, for example, do human languages have unintentional ambiguities? Computer languages don't have ambiguities; mathematical languages don't have ambiguities. Yet essentially all human languages do. Why are our languages so clumsy? Why are our memories so much worse than those of your average laptop? How could people continue to believe that Iraq had weapons of mass destruction 18 months after the invasion? These are the sorts of questions that led me to write Kluge.
LEHRER: Your book ends with a series of prescriptions for helping us make better use of our imperfect mind. Is there one prescription that you think is particularly important? Which piece of advice do you have the toughest time following?
MARCUS: The most important piece of advice might be the one that says, "Don't just set goals, make specific contingencies plans"—good advice that follows from the studies of my colleague [psychologist] Peter Gollwitzer [of New York University]. It's important because it's the best band-aid we've got for dealing with one of the more problematic kluges in our evolution: the split between a set of really ancient brain mechanisms that tend to be short-sighted and automatic and a more modern set of "deliberative" mechanisms that do their best to take the long view. As a species, we humans are the only creature that's smart enough to make long-term plans, but most of the time no matter how foresighted we might be, in the heat of the moment, our ancestral reflexive systems still tend to hold sway. By converting abstract goals (like a desire to lose weight) into specific if-then statements (e.g., "If I see the dessert menu, then I'll sit on my hands and discuss the election rather than choosing a dessert"), we can trick the older systems into following the sometimes-wiser goals of our more modern deliberative systems.
The hardest one? "Distance yourself," a bit of ancient wisdom that gets to the core of what we as humans trouble with. Our brains are built to react in the moment, and as a consequence we tend to overvalue whatever's happened in the last few seconds, relative to everything else that has happened or will happen. It’s sort of like studying yesterday's stock returns while ignoring everything else that happened in the five previous years. I'm not Buddhist, but there's much logic in detaching ourselves from the ups and downs of daily existence and moving toward a more balanced appreciation of life.