Consider the cryptic voicemail message. “It’s me—call me back,” the person says, not bothering to leave a name or even a reason for calling. And although we can’t quite place who it is, we usually know at least one thing. “That was definitely a man,” we declare, proud to have winnowed the list of potential callers to approximately half the human race.
We have learned to associate men and women with different ways of speaking. But when did we learn this? Not as early as one might expect.
In a 2002 study, University of British Columbia researchers Michelle Patterson and Janet Werker exposed infants to two silent videos. The videos depicted faces—one male and one female—producing the same vowel over and over in synchrony. During the test phase of the experiment, the videos were accompanied by a soundtrack: a single vowel (produced by either a second male or female) playing in lockstep with the moving mouths. The researchers measured how long infants looked at each of the two videos. At eight months—but not at four-and-a-half or six—infants watched the face that matched the gender of the voice significantly longer than the face that did not match. In other words, the infants demonstrated some understanding that a male-sounding voice is likely to be produced by a male-looking face and a female-sounding voice by a female-looking one.
If eight months was as early as you expected, perhaps this will surprise you: at just four-and-a-half months, infants have already mastered the relationship between specific vowels and their corresponding lip shapes. They know when mouths producing “oo” are accompanied by a voice saying “ee,” and they don’t like it one bit. (Unless they’re bored, in which case they like it plenty.)
But even though the relationship between mouth shape and vowel is determined, insofar as one always comes with the other, the seemingly more obvious relationship between gender and voice is considerably looser.
Yes, some gender differences in speech have biological underpinnings. Post-pubescent men are, on average, larger than women, and they sound like they’re larger. Men tend to have bigger noses, throats, Adam’s apples, and faces; they also tend to have longer vocal tracts and larger, heavier vocal folds, both of which contribute to lower frequency (or deeper) voices. But there’s plenty of gender overlap to complicate the learning process. Some men have small faces, and some women sound like Vin Diesel.
Other gender differences have no biological basis whatsoever. Among speakers of American English, for instance, women’s vowels are generally more “dispersed” or distinguishable from one another than men’s. Women also tend to use more standard variants, preferring to go running instead of runnin’. Linguists have posited reasonable explanations for these differences, but they do not involve anything so straightforward as body size. And these cultural factors are probably even more variable than the biological ones. Some women may have deep voices, but many women take great pleasure in runnin’.
Interestingly, in the infant study, the researchers took care to eliminate the cultural cues that might signify whether a voice belongs to a male or a female. Rather than exposing the infants to words or even multiple vowels, researchers played a single vowel token over and over (which, depending on how you look at it, either made the infant’s job more difficult because he had less information to work with, or easier because there were fewer distractions). But it’s worth noting the extent of culture’s reach. Though biological factors constrain the range of frequencies that a speaker can access, precisely where in that range a speaker chooses to produce a vowel is affected by how she wishes to sound, which is to say culture, culture, culture. Indeed, according to a recent study published in the open-access journal PloS ONE, we are quite adept at spontaneously “masculinizing” or “femininzing” our voices, including our vowel production. Something to keep in mind when interpreting cryptic messages.
