When we listen to speech, our brain processes different information concerning vowels and consonants. In fact, it is mostly consonant sounds that guide us to identify words within the flow of speech, according to a recent study by researchers of the Scuola Internazionale Superiore di Studi Avanzati (SISSA/ISAS) in Trieste and other institutes.

Speech is not just a series of sounds; speech also elicits a series of representations, such as syllables, vowels or consonants, which our brains identify as such from the very early onset of language acquisition. Listening involves a kind of statistical computation based on the fact that it is more probable that certain sounds follow particular syllables.

In the article, published in the June issue of Psychological Science, a journal of the American Psychological Society, researchers explain how consonants and vowels play a different role in these computations.

The researchers discovered that the representations guide when and how the statistical computations are carried out by listeners. Specifically, the study shows that consonants serve mainly to distinguish among words, whereas vowels tend to carry grammatical information. According to researchers, listeners are sensitive to this difference.

"Other scientists conjectured that it could be possible to learn a language simply using the incredible statistical capacities of the brain. Instead, in our work we studied which computational limits language imposes to this system for statistical calculus," explains Luca Bonatti, one of SISSA researchers of the team. The study has been carried out with dozens of French speakers. The choice of this language is not accidental: these kinds of experiments are possible only with idioms containing many vowels, like French.

"It has been known that humans can execute statistical computations in many different domains. However, they are not able to do it with vocalic structures. It seems that there could be a sort of specialization of tasks imposed by language areas to the whole brain," Bonatti states.

The theory works only within linguistic systems. Interestingly, living organisms like non-human primates who do not have language, seem to behave in the opposite way: "Other researchers showed that primates can compute statistics on vowels but not on consonants. We think that this behavior, which is the opposite to what our human participants did, can be explained because animals consider vowels as simple sounds without any grammatical value, and hence they are free to compute statistical relations among them", Bonatti says. "Instead, they ignore consonants completely because they cannot consider them as having linguistic import, treat them as simple noise, and hence disregard them entirely."

Source: American Psychological Society