Minimal viable sound systems for language evolution

Human vocal gamut covers 3000 unique speech sounds comprising the world’s languages, with each new speaker having to learn the sounds of its new language. Since large, expandable repertoires are facilitated by vocal learning, this capacity has long been considered a prerequisite for speech and language evolution. The postulation of a vocal learning and repertoire size ceiling has, however, never been tested, though such inquiry is needed: the prevalent vocal learning hypothesis rejects great apes as vocal learners, against evolutionary principles of shared ancestry and descent with modification, while failing to explain of how vocal learning would have otherwise emerged in the human lineage in the first place. Although collectively diverse, individual languages typically use two-digit repertoires of consonants and vowels. These repertoire sizes are tantamount to great ape repertoires, also composed by consonant-like and vowel-like calls. Furthermore, while new speakers must learn new vocalic and consonantal sounds, these do not depend on laryngeal (i.e., voice) control, as posited by the vocal learning hypothesis, but rather on supralaryngeal control. The tongue, lips and mandible must shape the oral cavity in novel ways to produce new formants and constrictions upon which vowel and consonant recognition depend, respectively. Novel acquisition of both call categories is now also well-documented in great apes. It appears the first language(s) never required vocal learning capacities nor repertoires larger than greats apes’ for full functionality. Instead, an increase in the ability to recombine existing sounds in novel ways seems to have been far more pivotal in the evolution of speech and language.