The Human Evolution Blog

It has been known for some time that, in mammals, certain vocal features correlate well with certain aspects of body size and shape. For example, the larger the animal, the deeper the voice. No big surprise there.

However, within a given species, and within each sex, correlations between vocal properties and body size are not so clear. For example, among male gorillas, the pitch of the voice does not correlate particularly well with body size. The same is true in humans. Body size only weakly predicts the depth of the voice and what little correlation there is seems to be a function of height alone, not weight or muscle mass.

But voices have many other attributes besides pitch. Researchers have found that features called formant frequencies correlate with some sex-specific physical traits in animals. In voice science, formants refer to the acoustic resonance of the vocal tract and are roughly proportional to the length of the vocal tract. In most mammals, including humans, taller individuals typically have longer vocal tracts than do shorter ones, and as a result, have lower formant frequencies and more resonant voices.

A recent study by Katarzyna Pisanski and colleagues posed the question with respect to humans: what information about our bodies is encoded in our voices?

To do this, they began with a multicultural sample of more than 700 men and women. They then measured 19 different features of their voices as well as eight different features of body shape and size, including obvious things like height, weight, and BMI but also the circumference of the chest, hips, and waist and the various ratios of those circumferences to each other (hip-to-chest ratio, waist-to-chest ratio, etc.) They then cross-analyzed the various vocal qualities and physical measurements to look for correlations.

They found some very interesting things. Firstly, the authors found that no acoustic measurements correlate well with height or weight except formant frequencies. This is what is seen in other mammals so it wasn’t a total surprise. The mechanism probably has to do with how our anatomy places constraints on the physical space that the vocal tract occupies in our necks.

Secondly, the authors found various physical features that weakly correlate with the character of men’s voice. For example, shimmer and jitter qualities were weakly correlated with the circumference of men’s hips and the hip-chest ratio. This, too, could be a function of anatomical constraints and  warrants further study.

However, the most fascinating thing that Pisanski and colleagues discovered was that several vocal features correlate with a very important aspect of body shape in women: waist-to-hip ratio. Those vocal features are jitter, shimmer, the standard deviation/variation in voice pitch, and, once again, formant frequencies.

Waist-to-hip ratio in women is not just any ole’ physical measurement like height or even chest size. This measurement is the most universally appreciated aspect of female physical attractiveness across all cultures. It is even independent of overall body size. Some find petite women attractive; some prefer full-bodied women. Some are attracted to their own race, while others are most attracted to women of a different race. Regardless of all that, most heterosexual men are attracted to a small waist relative to hips. (Or large hips relative to waist, as the case may be.) This is as close to a universally attractive feature as we have among human females.

Waist-hip ratio is also a surprisingly good indicator of overall health of women. A high ratio predicts risk of cardiovascular disease, type II diabetes, hypertension, and morbid obesity. This simple measurement has even been touted by public health officials as the best proxy measurement of overall health for women.

Waist-hip ratio also correlates very well with fertility and fecundity. With this in mind, it’s not at all surprising that heterosexual men are so drawn to women with a classic “pear-shaped” body. The evolutionary value of a mate who is in a state of good health and high fertility is obvious.

This is what makes Pisanski’s discovery of a correlation between certain vocal features and a low waist-hip ratio so intriguing. While the mechanism is not fully known, this correlation raises the possibility that women’s voice qualities could encode information about their overall health and fertility to potential mates. This would be valuable information for males when choosing women to pursue.

If this correlation is true, we would predict that voices with these same acoustic features would be rated as more attractive by men. Indeed, that has been found. Specifically, the same quality of formant frequencies found by Pisanski to correlate with waist-hip ratio is also frequently a component of voices that men find attractive.

The attractiveness of women’s voices has even been found to vary through the menstrual cycle such that they are most attractive during the, you guessed it, fertile period.

What does all this mean? The simplest explanation is that female-oriented attraction in humans has evolved to zero-in on certain vocal features and body shapes that correlate well with health and fecundity. This would have substantial value for males in their pursuit of evolutionary fitness, that is, many successful offspring.

However, that may explain the attraction, but not the correlation itself. What does waist-hip ratio have to do with health? What does it have to do with fertility? And most oddly, what do formant frequencies in the vocal tracts have to do with any of this?

The most likely explanation is that these are all shared affects of the reproductive hormones estrogen, progesterone, and prolactin. In the right balance, these hormones promote fertility, maintain cardiovascular health, and also direct body fat distribution around the hips, leading to the pear shape. The pear shape itself isn’t “healthy,” but rather it is the side-product, an indicator, of a healthy hormonal balance.

At the same time, absolute and ratio levels of these hormones could affect vocal features, as another side effect. In fact, this hypothesis has already been proposed and the monthly oscillations in vocal attractiveness add powerful support to this possibility. Pisanski’s work completes the circle by providing a key link in the causal chain: vocal qualities indeed correlate with waist-hip ratio.

So what we have here is a sophisticated (though likely accidental) system by which visual cues, in the form of body shape, and auditory cues, in the form of vocal features, transmit information about the health and fertility of adult human females. Not surprisingly, adult human men have evolved to read these cues.

-NHL