Introduction

What ties music and language together rhythmically ? 

• 300 pieces of work by English/French composers of classical music (Patel and Daniele, 2003)

  • nPVI of musical notes

  • rhythmic (stress) patterns reflect that of native speech

• Huron and Ollen (2003) found similar results for 7000 pieces of music from 12 different languages

Unanswered in previous studies

• Classical composers and professional musicians only constitute a small and specialized population.

  • extensive musical training, expertise, and elaborate cognitive framework

• Is such rhythmic connection between music and language more widespread?

Hypothesis for the present study

• Hypothesis: for amateur and novice musicians, native language would influence...

  • perception of speech rhythm

  • production of rhythmic variability

• Nonnative speech of another rhythmic class would be perceived as perceptually more variable
  • "language-specific listening" (Cutler & Otake, 1994)
    • using perceptual regularities of the native language
    • listening to nonnative speech -> perceiving variability/irregularity

Method

Design and procedure

• Musical rhythmic production: 

  • using two keys on a keyboard to produce "regular" and "English/French" tunes.
  • nPVI for notes
• Perception of rhythm in speech
  • rating 2 native and 2 nonnative speech recordings
    • nPVI: English = 52.2, French = 35.5
  • rhythmic variability on a 1-100 scale
    • 1 = regular = 500 IOI
    • 100 = random pattern of clicks with IOI from 100-900
• Recognition task (two weeks later, surprise!)
  • identifying participants own tunes (2AFC)

Results

Musical rhythmic production

• Significant interaction
  • English speaker's French tune is more variable than their regular tune

Perception of rhythm in speech

• Significant interaction
  • Both groups rated nonnative speech as more variable

Recognition task

• English speakers recognize their own tunes at rates above chance level for both types of tunes.

• French speakers recognize their own tunes at rates above chance level for both types of tunes.

• (The authors used Chi-square tests?!)

Discussion and Conclusion

the (un)familiarity effect

• Both English and French participants perceive nonnative speech as more variable
  • despite greater objective rhythmic variability of English
  • English speakers also produced more variable French tunes

the (un)familiarity effect (cont.)

• Explanation 1: unfamiliar speech is simply perceived as more variable
  • rhythmic expectations not met
  • overrides objective acoustic variability (high nPVI for English speech, but still less variable for English listeners)

• Explanation 2: paying attention to semantic messages in native speech so that rhythm is neglected
  • more attention to the nonnative leads to higher variability
  • Future work: low-pass filtered speech

The tunes

• Production of the novel tunes seem to be based on stable rhythmic representation
  • accurate in identifying own tunes

• Potential issue: playing familiar tunes?
  • but it would not explain the interaction (familiarity effect)
  • but it would be hard for novices to do so (with two notes)
  • but the produced tunes lack organized rhythmic structure
  • also, it should be more true for experts because they pay attention to musical structures

• Conclusion

• Evidence of a rhythmic relationship between music and language
• Even for musical amateurs and novices!
  • although not directly comparable to studies on experts because of task differences