It turns out that Henry Higgins was onto something: a recent study published by the National Academy of Sciences* finds that speech matters.

The study looked at the effects of job applicants’ speech on their success. The results present strong evidence that interviewers infer social and economic class from speech after hearing as few as seven words. An applicant presumed to be from a higher class is more likely to be offered a position, a better position, and/or a higher salary.

So, we can confirm social class on the list of differentiating traits that can disadvantage job seekers. Many will find this to be rather old news, but it is significant precisely because of how strenuously we try to level this playing field.

Given the collective and persistent choice to believe that America truly is the land of opportunity, we have tried for centuries to make it so. Even as prejudices survive, so have we tried to negate them. We have fought biases of all sorts all throughout our history, albeit more or less successfully. We have even outlawed some of those prejudices and codified protections for all sorts of people, if only on the basis of not impeding the common weal.

In our current and precarious age of severe and increasing income and wealth inequality, it behooves us to create the broadest possible access to participation in and contribution to our economy, if only so that we can create enough self-sufficiency and enough tax revenue to bring the government back to fiscal viability.

We try to create opportunity at the institutional level, by passing laws or filing lawsuits, or by creating institutions to help provide the prerequisites of access. For example, we create and fund—and require, to some extent—public education in order to create broadest access to a credential.

But as income and wealth gaps do not narrow and, in fact, widen despite our institutional efforts, we focus on personal efforts to create opportunity as well. For example, we encourage people to stay in school: we tout the robust statistical relationships between education and income that have held up for generations, and we offer decades of affordable debt to help fund it.

We are constantly exploring and promoting self-improvement: nutrition, health, exercise, sleep. A healthier person is, after all, more likely to make it into work as expected; a healthier family is less likely to cause a caregiver to call in sick. And the more self-improved a specimen you are, after all, the better your chances of success.

Technology has proven to be something of a leveler as well. It has created access for those disadvantaged by mobility or geography or even just a lack of social skills. Job applicants are much less identifiable online, and by submitting a resume, can focus the employer’s attention on experience, education, and skills—on the factors that should matter and not on the ones that should not.

Still, none of this institutional or individual encouragement of opportunity has made a more equitable society, as far as we can tell, and this report shows just how tenacious and persistent some biases can be. It seems that as soon as you open your mouth, the jig is up: a listener can describe your socio-economic status with alarming accuracy, and those of higher presumed status get the better opportunities.

So, speech joins the list of distinguishing characteristics that can put people at an economic disadvantage, if only because it betrays other factors that are the targets of bias or the nourishment of social stereotypes. This would be merely a matter of social aesthetics if it were not also a determinant of jobs, income, and wealth.

When “My Fair Lady” hit Broadway in 1956, audiences could enjoy the show with a smug belief that they lived in a country of more opportunity and less stratification than the old British Empire, as portrayed. America was enjoying a postwar prosperity that was reaching into the far corners of society. The GI Bill was profoundly expanding access to education and its promise. Trade dominance was creating employment and a comfortable middle class. The rising tide was floating all—or at least a significantly expanding number of—boats.

Eliza Doolittle’s “cold-blooded murder of the English tongue” was just comic relief. She rebelled against her caste, recognized her limitations, improved herself through hard work and education, and in the end, could pass for a princess. It seemed to be a fable from a happily far-off—and in many ways more backward—world. Now, it plays with disturbing new relevance.

*Michael W. Kraus, Brittany Torrez, Jun Won Park, and Fariba Ghayebi, “Evidence for the Reproduction of Social Class in Brief Speech,” Proceedings of the National Academy of Sciences, November 12, 2019.