Given my interest in languages I found this post by Harry Baker at LiveScience on how cross-cultural the recognition of “iconic vocalizations” is fascinating:
In an online experiment, researchers exposed 843 participants, who spoke 25 different languages among them, to iconic vocalizations representing 30 meanings that would have been key for the survival of early humans. The participants then had to match the sound to one of six words, including the intended meaning.
The intended meanings for vocalizations were grouped into six main categories: animate entities (child, man, woman, tiger, snake, deer), inanimate entities (knife, fire, rock, water, meat, fruit), actions (gather, cook, hide, cut, pound, hunt, eat, sleep), properties (dull, sharp, big, small, good, bad), quantifiers (one, many) and demonstratives (this, that).
Researchers obtained these vocalizations through an online contest where, in exchange for prizes, people could submit basic sounds that they felt best represented different words. Everyone who submitted a vocalization spoke English.
In the experiment, people accurately identified the meaning of these vocalizations 64.6% of the time, on average. The most recognizable vocalization was that for “sleep,” which people identified with 98.6% accuracy. The least recognizable was the demonstrative “that,” with an accuracy of 34.5%, although it was still well over the 16.7% (one in six) expected by chance.
In general, people understood the vocalizations of actions and entities better than those for properties and demonstratives. “These recognizable sounds [actions and entities] are probably associated with these meanings across cultures,” Perlman said. “In others, there’s probably more variability over precisely what that sound is.”
Out of the 25 languages spoken by participants, speakers of 20 languages correctly guessed the meaning of each vocalization on average, speakers of four of the languages did so for all but one vocalization and speakers of the remaining language did so for all but two. The language speakers with the lowest accuracy were Thai speakers at an average of 52.1% and the best performing language speakers were English speakers with an average accuracy of 74.1%.
That certainly isn’t proof positive. It could be that 2/3s of the participants had some familiarity with the paralinguistic features of English, something that would hardly be surprising given the nearly universal exposure to American television. But it is interesting.
What I wish they’d do is repeat the experiment using submissions from native speakers of the Mongolian language or native speakers of Kituba. I also wish they had included recordings of at least some of these “iconic vocalizations”.
