Proposals that gesture played a pivotal role in the evolution of language have been highly influential. However, there are many differences between gestural origin theories, including different definitions of ‘gesture’ itself. We use a cognitive semiotic approach in order to categorize and review these theories. A semiotic system is a combination of signs or signals of particular type, defined by characteristic properties, and the interrelations between these signs/signals. Signal systems like spontaneous facial expressions and non-linguistic vocalizations are under less voluntary control than sign systems. The basic distinction relates to the question of whether gesture played an exclusive role in early stages of language evolution (monosemiotic theories), or whether other semiotic systems were involved as well: polysemiotic theories. The latter may be equipollent, where language and gesture are considered equally prominent from the onset, or pantomimic, where gesture played the main but not exclusive role in breaking from predominantly signal-based to sign-based communication. We conclude that pantomimic theories are the most promising kind.