What I'm Reading: Rhythm and Melody in Speech and Music

Music, Language, and the Brain, by Aniruddh Patel, is making me wish I had taken more psychology and linguistics courses. This week, I read about comparisons of rhythm and melody in speech and music. I was surprised to read that, despite our intuitions about rhythm, speech rhythm actually involves no periodicity (unlike music, which usually depends heavily on a beat).

You may be familiar with the idea that languages can be divided into two groups: stress- and syllable-timed languages. Speakers of stress-timed languages are thought to spend an equal amount of time between each consecutive stressed syllable. In syllable-timed languages, an equal amount of time is expected to be spent on each syllable. In a third group of languages, including Japanese, timing is thought to be based on regular temporal relations of morae, another type of sound unit.

Despite the widespread acceptance of these language groupings, no empirical evidence supports the idea that any languages are stress-, syllable-, or mora-timed. Thus, linguists must search elsewhere for speech characteristics that give rise to rhythm. Patel discusses a few of these characteristics, and shows how they can be used to compare music and speech rhythm within and among different languages.

Interestingly, in a study on French and English speech and music, Patel did find rhythmic correlations between the two domains, which differed between the languages. Patel attributes this to "statistical learning," whereby a composer has at his/her disposal a language-specific library of rhythms familiar from speech. (Of course, the composer could choose not to use these, but may do so consciously or subconsciously.)

The melody section also spent a lot of time discussing speech, and focused on the melodic contours of speech and music. A melodic contour can be thought of as a graph with time on the x-axis and pitch frequency on the y-axis; it is the "shape" of the melody through time, based on whatever frequency is being used at each moment.

While musical melodic contours involve sitting on the same pitch for some time, speech contours often consist of constant movement between pitches. Linguists have devised two main ways to simplify speech contours. In AM theory, important pitches in the contour are identified, and movement between them is dismissed as perceptually meaningless. In IPO theory, it's the movements between pitches that are viewed as important.


In a prosogram (shown above for a French phrase), a particular pitch is assigned to each syllable. Patel used prosograms to compare French and English speech and music. As with rhythm, he found that linguistic differences in speech melody were echoed by similar differences in musical melody.

Patel finishes up the melody section with a discussion of how asmusia and musical tone-deafness indicate the speech and musical melody processing in the brain overlap to a greater extent than previously thought. I'm going to devote a separate post to this later, as I find it particularly fascinating!

Musical notation source: Wikimedia Commons
Prossogram source: http://bach.arts.kuleuven.be/pmertens/prosogram/