Logo image
The song remains the same? The effect of oral transmission on folk melodies
Conference poster   Open access

The song remains the same? The effect of oral transmission on folk melodies

Joshua Albrecht and Daniel Shanahan
American Musicological Society Southwest Chapter Conference Proceedings, Vol.4
Spring 2015
pdf
Albrecht_AMS-SW_Proceedings64.13 kBDownloadView
Abstract and references Open Access
url
https://southwest.ams-net.org/wp-content/uploads/2021/12/AMS-SW_V4Spring2015Albrecht.jpgView
Poster Open Access

Abstract

Oral transmission is notorious for degrading a signal, as commonly seen in the game " Telephone. " Likewise, by comparing " families " of differences for Stephen Foster's " Oh! Susanna, " Spitzer (1994) traces the stemma of the song, generalizing four types of changes and hypothesizing that these tendencies are ubiquitous in oral transmission:

  • tendency to alter rhythms in order to clarify the beat
  • tendency for a salient harmony to draw the melody to the chord root
  • tendency to eliminate differences between parallel passages.
  • tendency to pentatonicize the melody at cadence points.

In this paper, we directly examine the tendency of orally transmitted folk song excerpts to pentatonicize cadence points through an experimental paradigm we're calling " Musical Telephone. " The final phrases of eight English and Irish folk melodies were selected, half of which ended with redo and half of which ended with ti-do. Each penultimate note was then altered using Melodyne to create an alternate version, resulting in sixteen melodies. The experimental hypothesis is that ti-do motion will be changed to redo significantly more frequently than the reverse.

In this experiment, fourteen teams of four music majors each played Musical Telephone with four of the resulting melodies, balanced between altered and unaltered and redo and ti-do cadences.

Of 240 total sung melodies, 29 (12%) altered the penultimate pitch. Ti-do motion was changed to redo motion in 20 (69%) of those instances, significantly higher than chance (p=.04). These results are consistent with the hypothesis, perhaps implicating a statistical learning effect in which redo cadences supplant ti-do cadences in memory because they are heard more frequently. Alternatively, it may be the case that melodies that descend to the final are more natural to sing because of physiological tendencies in which sub-glottal air pressure is reduced at the ends of phrases (T'Hart et al., 1990).

Details

Metrics

1 Record Views
Logo image