Geoffrey Nunberg - A Loss for Words (on vocabulary size)

A Loss for Words

Geoff Nunberg

"Fresh Air" commentary, Sept. 3, 2002

To linguists, it has a familiar sound -- someone claims that some primitive language has a vocabulary of only a few hundred words that leaves its speakers incapable of expressing abstract ideas and has to be eked out with gestures. People have said that about Cherokee, about Inuit, about Australian aboriginal languages -- a century ago, they were even saying it about Hungarian and Finnish. And yet when linguists look at those languages, they always turn out to have vocabularies that are more than rich enough to express the complexity of a people's experience.

Those claims have their echoes in the things that people have always said about the language of the poor and the underclass. Back in 1776, the Scottish philosopher George Campbell said: "As the ideas which occupy [the minds of the poor] are few, the portion of the language known to them must be very scanty." And until a few years ago, a lot of educators were saying that lower-class children spoke a "restricted code" that made abstract reasoning impossible. And yet both educated and uneducated people turn out to have vocabularies that are perfectly adequate to cope with the moral and mechanical complexities of daily life. Most of the edge that educated people have is in the words we need to talk about topics like politics, historical ideas, technical subjects, and the like.

Linguists have been battling these canards about vocabulary size for decades, so it's a little depressing to hear them trotted out again, all the more when they come from reading experts serving in the current administration. Not long ago someone pointed me to a recent speech by G. Reid Lyon, the chief of reading research at the National Institutes of Health. According to Lyon, "A 3-year-old child in an affluent family has a larger working vocabulary than the mother of a 3-year-old from a welfare family." And I found the same claim in different talks by Susan Neuman and Grover Whitehurst, two academics who are serving as Assistant Secretaries of Education.

The people who make that claim cite an influential study published in 1995 by Betty Hart and Todd Risley, two researchers at the University of Kansas who looked at parent-child interactions among different social groups. Hart and Risley found some striking differences. On average, professional parents talked to their toddlers more than three times as much as welfare parents did. And not surprisingly, that difference resulted in a big discrepancy in the children's vocabulary size. The average three-year-old from a welfare family demonstrated an active vocabulary of around 500 words, whereas a three-year-old from a professional family demonstrated a vocabulary of over 1000 words. Those differences get more pronounced as kids get older, and have an effect on their success in learning to read. In all, it's strong evidence for the need for early intervention -- by the time the low-income kids get to school and start to learn to read, they're already at an enormous disadvantage.

But how did that lead to the claim that welfare mothers had smaller vocabularies than the three-year-olds from professional families? Well, Hart and Risley did find that the average welfare mother used only about 1000 different words in talking to her kids over the several hours of parents' talk that the researchers recorded. But to put that in perspective, the average professional parents only used about 2000 different words in talking to their kids -- and that in considerably larger samples of speech. That scarcely means that the professional parents had 2000-word vocabularies, but only that parents of all classes tend to talk to kids in simple language. And when I talked to Betty Hart at the University of Kansas, she told me that the welfare mothers used far larger vocabularies when they were talking to their friends, their older children, or the researchers themselves -- conversations that weren't recorded for purposes of the study.

When you think about it, in fact, the claim that any normally functioning adult could get by on a three-year-old's vocabulary is absurd on the face of things. Welfare mothers may seem to live restricted lives by the standards of middle-class professionals. But they still have to function in a wide range of situations, whether it's a question of going to church to the supermarket to the welfare office, or simply sitting around chatting about friends or music or TV shows or sports teams, the way everybody else does.

I can't hear this sort of thing without being reminded that the word infant comes from a Latin word that means "not having the power of speech." In that sense, the claim that welfare mothers have thousand-word vocabularies is infantilizing in the literal sense of the term. It suggests that the mothers don't actually have enough language to be able to make sense of the world they live in. And from there it follows that these women simply aren't in a position to articulate their needs or make reasoned judgments about their lives -- that they literally can't speak for themselves.

Now I don't think that the Administration education honchos who have been repeating the claim about welfare mothers' vocabularies set out to deliberately distort Hart and Risley's research. My guess is that none of them actually read the study -- this has the sound of one of those third-hand factoids that are always making their way around the scientific grapevines. But even so, it's telling that they found this a credible sound-bite. Whether you're disparaging the vocabularies of welfare mothers or the Cherokee, the claim always carries an unfortunate tone of condescension. It's easier to ignore people's voices when you've decided they couldn't possibly have one.