Abstract
The innovation of iconic gestures is essential to establishing the symbolic vocabularies of signed languages, but might iconicity also play a role in the origin of various spoken words? Can people create novel vocalizations that are comprehensible to naïve listeners without the use of prior conventions? To test this capacity, we launched a contest in which participants submitted a set of non-linguistic vocalizations for 30 meanings spanning actions, humans, animals, inanimate objects, properties, quantifiers and demonstratives. The winner – who received a monetary prize – was judged by the ability of naïve listeners to successfully infer the meanings of the vocalizations. We report the results from the contest, along with a series of experiments and analyses designed to evaluate the vocalizations for: 1) their comprehensibility to naïve listeners; 2) the degree to which they resembled their meanings, i.e., were iconic; 3) agreement between producers and listeners in what constitutes an iconic vocalization; and 4) whether iconicity helps naïve listeners learn the vocalizations as category labels. The results show that contestants were able to create iconic vocalizations for a wide array of semantic domains, and that these vocalizations were largely comprehensible to naïve listeners, as well as easier to learn as category labels. These findings provide a compelling demonstration of the extent to which iconic vocalizations can enable interlocutors to establish understanding through vocalizations in the absence of conventions. This suggests the possibility that, prior to the advent of full-blown spoken languages, people could have used iconic vocalizations to ground a spoken vocabulary with considerable semantic breadth.