The Male Voice as a Dishonest Signal:
Experimental Studies with Artificial Intelligence

This is the latest version of this item.

Abstract

The male voice appears to have been subject to substantial sexual selection, yet scholars disagree on whether this has mostly been driven by intersexual pressures (i.e. female choice) or intrasexual pressures (i.e. male-male competition). A further disagreement is over whether vocal masculinity is an honest signal – that is, whether masculine voices are cues to properties that would be beneficial in a mate (e.g. immunity, fertility) or detrimental in an opponent (e.g. physical size or strength). To shed light on these arguments, we ran eight studies over five chapters. In Chapter 2, we collected several vocal variables from the wider literature and attempted to measure their contribution to human vocal dimorphism. Ultimately, we found that variables which influence the depth of the voice (namely, fundamental frequency and formant frequencies) were highly useful in distinguishing male voices from female voices, but variables which affect vocal texture (e.g. roughness, breathiness) were not.

In the chapters that followed, we attempted to explain how this dimorphism may have been selected for. Chapter 3 presented a literature review of reasons to doubt that intersexual selection had played a leading role in the development of vocal masculinity. It was found that only some facets of vocal masculinity were attractive (i.e. low voice pitch) and only among certain populations (i.e. WEIRD populations). Furthermore, evidence was mixed as to whether vocal masculinity significantly improved a man’s likelihood of being selected for a second date. This review also examined the question of signal honesty, finding little reason to believe that vocal masculinity was an honest signal of some underlying quality that would be beneficial in a mate, such as immunity. In Chapter 4, we then presented two studies in which full-body videos were recorded of men counting to five and their vocal masculinity was manipulated digitally, to produce 12 variants of each video with different combinations of fundamental frequency and formant frequencies, each of which was then rated for attractiveness, fighting ability, physical strength, and social dominance. It was found that both fundamental and formant frequency manipulations significantly predicted all social perceptions; the men in the videos were generally rated as more attractive, stronger, better at fighting, and more socially dominant as vocal masculinity increased. A follow-up study found that the vocal manipulations significantly biased judgements of weight, but not height, suggesting that vocal masculinity may signal dominance dishonestly via an exaggeration of weight.

The final chapters delved more into the question of signal honesty. Chapter 5 demonstrated that men sounded more attractive and formidable, but less healthy, when suffering from a respiratory infection, suggesting that perceptions of the voice are somewhat incoherent and therefore vulnerable to dishonest signalling. We also found little reason to believe that preferences for vocal masculinity were motivated by preferences for health, as predicted by the Immunocompetence Handicap Hypothesis. Women with stronger preferences for masculine voices actually had stronger preferences for the voices of infected men, and women’s own vulnerability to disease only predicted stronger preferences for healthy, but not low-pitched, voices. Finally, Chapter 6 attempted to answer whether vocal masculinity was an honest signal of formidability, by comparing the voices of elite fighters against the voices of academics, all of whom had appeared as guests on the same popular podcast. This found no objective differences in vocal masculinity between fighters and non-fighters, although naïve listeners relied heavily on F0 and formant frequencies when judging the physical formidability of each man from his voice. Furthermore, naïve listeners did not rate fighters as sounding more formidable than non-fighters once the very tall fighters were excluded from the dataset. Overall, the findings of this thesis suggest that the male voice is a dishonest dominance signal. The radical changes that occur to the male voice during puberty do so in order to exaggerate physical weight and therefore formidability, but there is little evidence that these changes index actual formidability.

Available Versions of this Item

• The Male Voice as a Dishonest Signal:
  Experimental Studies with Artificial Intelligence. (deposited 17 Jun 2025 13:57)
  • The Male Voice as a Dishonest Signal:
    Experimental Studies with Artificial Intelligence. (deposited 24 Jun 2025 10:08) [Currently Displayed]