Spanish is spoken more quickly than Vietnamese. However, the amount of information conveyed per second in these two languages is almost the same. In an analysis of seventeen languages, researchers found that regardless of how quickly or slowly a language is spoken, the average “information rate” is around thirty-nine bits per second.
Each language’s information rate consists of its average speech rate multiplied by its information density. To determine speech rate, researchers from Hong Kong, France, New Zealand, and South Korea analyzed recordings of 170 native adult speakers reading text in seventeen different languages. To calculate information density, the researchers developed a measure for each language of how difficult it is, on average, to guess the syllable that starts a word, or the next syllable in a word already started. Information density of language is expressed in bits, the same unit of measurement we use for transmitting data on our computers and cellular phones.
The languages studied differed greatly in their speech rate and information density. For instance, speakers of Vietnamese, a “dense” language that conveys eight bits of information per syllable, use a slow speech rate (5.25 syllables per second), whereas speakers of Spanish, which conveys a comparatively less dense five bits per syllable, use a faster speech rate (7.7 syllables per second). Despite these different strategies, the two languages have nearly identical information rates: about forty-two bits per second. “Languages differ in their strategy of information encoding but their respective speakers can efficiently convey information by adapting their average speech rate with no effort,” explained François Pellegrino, one of the study’s authors.
It is not clear why languages seem to share a similar information rate, but Pellegrino and his team have a few hypotheses. It might say something about how our brains process speech—perhaps, this thirty-nine bits per second indicates an optimal rhythm or rate for processing. “Our study tells us that in a given language, there is a speech rate that is probably a sweet spot,” said Pellegrino—not too fast for listeners to understand in real time yet not so slow that listeners lose interest. (Science Advances)