A Cross-cultural Examination of Raga Perception and Time-of-day

Rhythm Jain and Claire Arthur

Introduction

According to ancient Indian literature, the day is divided into eight time intervals called prahars– 4 for the day and 4 for the night, each one roughly 3 hours long. In Hindustani classical music (HCM), there is a historical association of ragas to a specific time of performance and doing so is believed to maximize that raga’s emotion conveyance. (Brabant & Toiviainen, 2014). Proposed by Pandit VN Bhatkhande, ragas map to one of these prahars (time intervals). For eg, Raag Bhairav is chosen to be performed in the morning (sunrise/daybreak) interval. Raag Bhimpalasi is recommended to be performed during the interval 12-3 pm. These associations are believed to be based the musicological traits of each raga and performers’ physiology during that time of day. In terms of relevant prior work in this area, a study by Agrawal et al. 2021., has examined associations between these time-of-day classifications and musical features of Hindustani raga performances (Agrawal et al. 2021).

However, although time theory highlights when to perform a raga, we were curious if some of these associations got carried onto the practice of listening. To the best of our knowledge such a study has not been conducted before. So, although prahar is related to emotion conveyance (i.e., performing the raga within the prahar is believed to maximize its emotional impact), we were specifically interested in whether the musical excerpts would be associated with a particular time of day by the listener. The main goal of this research was to understand if prahar is purely a learned cultural aspect of the performance tradition, or whether it can be perceived. In addition, we wished to separate the effects of enculturation as well as musical training. There have been several studies investigating the relationship between ragas and their evoked emotions. However, although prahar is related to emotion conveyance (i.e., performing the raga within the prahar is believed to maximize its emotional impact), we were specifically interested in whether the musical excerpts would be associated with a particular time-of-day.

Our research questions were the following:

  • Is prahar (time of day) a purely learned cultural aspect of the performance tradition, or can it be perceived?
  • What factors might influence the choice of a listener’s perception of time in music and what is the granularity of that perception?

Pilot

The pilot study was conducted to assist the larger more comprehensive study to gain insight on some of the factors which might influence listeners preference of time while listening to some music. In the pilot, we chose a total of 8 independent Hindustani classical stimuli and asked participants to mark their preferred time of day to listen to that stimulus. We kept the time granularity to 8 intervals in a day and allowed participants to select multiple intervals if they wished.

Figure 3. Example of the pilot study experiment interface

Based on the participant feedback and observations we decided to reduce the number of intervals from 8 to 4 as most participants associated music with a broader chunk of time. We decided to use just 4 maximally separated time intervals namely, morning afternoon, evening, and night. Experiment Design

For the final experiment, we decided to collect the following independent variables based on their likely impact on the listener’s perception of time –

  • Age
  • Sleep cycle
  • Lineage (whether the participant’s lineage was Indian or not)
  • Location in India
  • Musician/Non-musician
  • Familiarity and training in Hindustani, Carnatic and Western classical music

It is important to note while talking about the perception of raga is that a raga is different from a scale in western music because unlike a western scale, a raga is not just associated with the musical notes but also the note-to-note transition patterns. Due to this difference, we also aimed to investigate whether listeners (Indian and non-Indians) would associate the music of a Western modal system with a particular time of day. With this, we intended to incorporate a “true” cross-cultural aspect and added a scale-based control from a different musical system. Thus, we had a total of 18 stimuli including –

  • HCM Raga Improvisation (1 instrument)
  • HCM Raga performance in a concert setting (multiple instruments)
  • Instrumental Indian film music
  • Improvisation in Western Modes

Main Experiment and Analysis

Participants (N=71) were asked to select the time of the day they would most likely want to listen to each audio clip in a forced-alternative task. We had a majority of non-indian participants in our experiment. Participants were from India, North America and South America. Figure 4. A comparison of the lineage of the participants

We decided to fit a mixed effects logistic regression model because we are treating TOD as a categorical variable. In reality, time is continuous, but we presented time as intervals in the experiment. From the results of our significance testing using ANOVA tests we observed that none of the factors among age, familiarity, training and musicianship were of significance. However, there was a very slight increase in significance with the training levels in Hindustani classical music. This slight increase could have risen to significance if we had more data. Results

Most participants tended to select “Morning” or “Evening” TsOD more than afternoon or night for the stimuli in general. Indicating that most people tended to think in sort of the granularity of 2 intervals in a day.

For western stimuli, the Aeolian and Dorian modes received a majority of night time preference. For Ionion mode, the majority lay in morning and afternoon, whereas for Lydian it was evening and night.

For the Indian film stimuli, we observed that most of the listeners chose morning as the preferred time of listening for Bhupali (which could be attributed to the fact that the musical notes in raga Bhupali are analogous to major pentatonic scale.

Indeed, even listeners who self-identified as having 5-10 years of training were not significantly more likely to select the correct prahar.

Conclusion

Our biggest takeaways from our results were that surprisingly, even those with Hindustani classical training apparently were not very sensitive to the ‘appropriate’ prahar or time of day for the Indian stimuli. Also interesting was the seemingly large cross-cultural effect of minor mode (more-so than any of the Hindustani ragas). Minor is associated with darkness, and this has been shown in Arthur, 2018.

Overall, we felt that some more patterns (from information that we collected from the participants such as age or training in Hindustani and Carnatic music) could have surfaced if we had more data between this association of listening with time. In the future, we would also like to include stimuli from Carnatic music (although the Carnatic music style has very little significance of time of day), nevertheless seeing some results of listener’s perception in stimuli from that musical style could also be interesting.

For more information on this study, feel free to contact Rhythm Jain (rjain354@gatech.edu) or [Claire Arthur]/teammembers/Arthur_Claire.html (claire.arthur@gatech.edu)