Last updated March 29, 2018 at 10:28 am
When first impressions count, find out how to nail saying hello in French
French researchers have cracked the code of how different tones change the meanings of different words, and in the process created a computer program they’re calling the Ministry of Silly Speech.
Language is an incredibly complex thing. Not only do we listen to the words being said, but also how they’re being said. A minor change in pitch and tone can completely change the meaning of a sentence.
In a way it’s the ultimate first impression – when we meet someone for the first time and say hello it can come across as friendly or hostile, and in a split second we make that determination based on their intonation.
Just as we can immediately build a mental picture of what an apple looks like – round, green or red, with a stem, etc – we build a mental representation of others’ personalities according to the acoustic qualities of their voices.
However how those are built, and what they look like has been difficult to determine. Now, researchers have managed to visualise these mental representations to discover what makes a hello sincere, and make comparisons between different individuals.
The Ministry of Silly Speech
The researchers from the French National Center for Scientific Research, the Ecole Normale Supérieure, and Aix-Marseille University developed a computer program for voice manipulation called CLEESE. In honour of the legendary actor, it was them promptly nicknamed the Ministry of Silly Speech (although we think the Ministry of Silly Talks has a certain ring to it too).
This software can take the recording of a single word and randomly generate thousands of other variant pronunciations that all sound like something someone might realistically say, but each unique in their melody.
The team then played the recordings of the different pronunciations to participants in the study, and by analysing their responses worked out exactly what makes a hello seem sincere.
To sound determined, a French speaker must pronounce bonjour (French for “hello”) with a descending pitch, putting emphasis on the second syllable.
On the other hand, to inspire trust, the pitch must rise quickly at the end of the word.
By manipulating the tone of certain sections of words, the team was able to visualise the “code” people use to judge others by their voices. They also showed that no matter the sex of the listener or the speaker, the code stays the same.
By expanding the study to include English speakers and listeners, as well as non-native speakers, it will be interesting to see if the same patterns apply. You’d almost certainly assume that different patterns have different context in different languages, for example a fast rise in pitch at the end of words is usually considered to be questioning in English, however the differences and similarities could open up new ideas about the evolution of languages.
The Life of Brain
CLEESE could also be a crucial tool in understanding how words are interpreted by stroke survivors.
Following a stroke, people often perceive and interpret intonation differently, and the French team are now interested in using their program to understand this change better.
Whether for the purposes of medical monitoring or diagnosis, the researchers say they would like to use their method to detect anomalies in language perception and possibly make it a tool for patient rehabilitation.
In the meantime however, the researchers have made CLEESE freely available, and have asked research groups around the world to use it to tackle other questions of language. These could include questions around the structure of sentences instead of just single words, and the language being spoken.
However, CLEESE could also be an extremely valuable tool in understanding how emotions are understood and perceived by people on the autistic spectrum.
You can download CLEESE as a standalone Python module here .
The research has been published in Proceedings of the National Academy of Sciences