Why Are They Talking So Fast?
The impression of increased speech rate in a second language
Posted Feb 24, 2013
Post written by François Grosjean.
Languages are spoken at about the same rate even though there is a lot of variability due to the speakers themselves, the situation, the topic being talked about, and so on. Many years ago, my University of Paris colleague, Alain Deschamps, and I measured the speaking rate of a group of English and French speakers in an interview situation. The average rates we found were quasi-identical: 176 words per minute for English and 174 words per minute for French. (The rates in syllables per minute were not significantly different either).
Even when languages are totally different from one another, a spoken language on the one hand, and a sign language on the other, similarities are found if one examines the right level. Salk Institute Professor, Ursula Bellugi, and her colleague, Susan Fischer, found that when someone is signing, as compared to speaking, less signs are produced per minute than words. This makes sense as the hands-arms-body articulators are much larger in sign than the tongue-jaw-lips articulators in speech. However, and this remains a major finding of their research, when you count the number of propositions (basic ideas) per minute, the rate is identical in sign and speech.
Why is it then that when we listen to a language we do not master well, we feel that the rate is faster than in our native language? This leads to a general impression that speakers of language X (fill in your preferred second language) always talk fast. University of Geneva researcher, Sandra Schwab, and I examined this precise question in a study that was published in the journal, Phonetica.
We asked a group of native speakers of French, and a group of native speakers of German who had studied French for several years and used it occasionally, to listen to short stories in French read at a fast, a medium and a slow rate. They were asked to estimate the rate of each story (using magnitude estimation) and to answer five comprehension questions.
The first thing we found was that, overall, second language speakers did indeed give higher estimates than native speakers. But, more interestingly, the difference between the two groups was not the same at all rates. At a slow rate, the groups gave practically the same estimates. They started diverging at a normal rate (the estimates of the second language speakers were higher), and they were quite different at a fast rate (here the second language speakers gave much higher estimates). In sum, the faster the rate, the greater the difference between native and non-native speakers.
To start understanding why it is that we feel that speakers of a language that we do not master well speak faster, we examined the results of the comprehension assessment of the non-native speakers. We wanted to see if there was a relationship between the level of comprehension in a second language and the estimates of speech rate. A correlational analysis showed that there was. We found a significant negative correlation between comprehension scores and rate estimates: the lower the score, the higher the rate estimate. This was true when the stories were presented at a slow rate and at a medium rate. We didn't find the relationship at a fast rate: all non-native speakers found that rate extremely rapid however good their comprehension was.
So it would seem that our increased estimate of speech rate in a second language, at least when it is either normal or slow, could be due, in part, to the fact that we are trying to understand what is being said. The less we understand, the more we feel that the rate is high. Future studies will want to show how this estimate of rate evolves over time as oral comprehension improves. If there is progress, then estimates of speaking rate may start resembling those of native speakers, or at least of fluent bilinguals.
Photo of the young business team from Shutterstock.
Sandra Schwab & François Grosjean (2004). La perception du débit en langue seconde (Speech rate perception in a second language). Phonetica, 61, 84-94.
"Life as a bilingual" posts by content area.
François Grosjean's website.