Options
Linking musical metaphors, emotions and feelings of entrainment evoked by electronic music: From the physical properties of sounds to the experience of listening to music
Auteur(s)
Politi, Stefano
Editeur(s)
Didier Grandjean
Maison d'édition
Neuchâtel : Université de Neuchâtel
Date de parution
2024-01-01
Nombre de page
68
Résumé
This paper proposes a perspective on the experience of listening to music, where sounds are conceived as objects of auditory perception in interaction with the listener’s body. This interaction in considered to be a continuous process of attributing meaning by the listener, based on the possibility of action in accordance to the music; to possibility to move to – or to be moved by – the music, both metaphorically and literally. Musical meaning is, here, understood in terms of metaphors and emotions. In this paper, the relationships between metaphors and emotions are analyzed from the perspective of embodied (music) cognition, which considers the body as central for cognitive (musical) processing. The phenomena of the body synchronizing to the music, as a consequence of it’s interaction with it, is referred to as rhythmic entrainment (Trost et al., 2017). Within this theoretical framework, an enactivist standpoint affirms that action tendencies are intertwined with perception. According to such, physical (acoustic) properties of sounds are cognitively mapped through sensorimotor (bodily) representations: a process of attributing musical meaning to the interaction between the listener with the music. Listening to music necessarily implies processing sequences of sounds through time, often conceptualized in terms of metaphors of (bodily) motion – embodied musical meaning (Johnson, 1997). The motion of music through time is considered to be understood by the listener in terms of how this motion is reflected as motion in the body, whether though overt movements (as dancing) or covert sensations (as introspective emotions). Describing the experience of listening to music in terms of metaphorical and emotional language sheds light on how bodily motion is associated with musical motion, through the process of attribution of embodied musical meaning.
The core of this paper is based on an empirical study conducted with the aim to explore interactions between the physical (acoustic) properties of sounds and the experience of listening to music. By doing so, this study also aims to provide a scientific contribution to an embodied (music) cognition framework. Two categories of electronic music (ambient and dance) were considered as inclined to exert rather sedative or stimulative effects on the body of the listener. The study was composed of three phases, and combines tools from Computational Sciences and Music Psychology. The first phase consisted in the extraction and analyses of a set of acoustic features with the MIR toolbox (Lartillot & Toiviainen, 2007), from 16 electronic music excerpts (length of 30 seconds). In the second phase, 172 participants evaluated the music through an online survey. They rated items of the Geneva Musical Metaphors Scale (GEMMES) (Schaerlaeken et al., 2019), to assess musical metaphors, of the Geneva Emotional Music Scale (GEMS) (Zenter et al., 2008), and the two-dimensional emotional model of valance and arousal (Russel, 1980), to assess musical emotions, and of the Musical Entrainment Questionnaire (MEQ) (Labbé & Grandjean, 2014), to assess subjective feelings of entrainment. In both phases, data was analyzed with Principal Component Analysis (PCA), to obtain a few relevant groups of variables based on the correlations between them. The third phase consisted in computing linear mixed-effects models, in order to explore the interaction between the physical properties of music (acoustic features) and the experience of listening to music (survey ratings). Finally, results of all three phases were compared with those from a similar study on classical music (Schaerlaeken et al., 2022).
The first phase resulted in in two principal components of acoustic features; two groups of mathematical descriptors of sound-waves associated with perceptual characteristics of music. Dance music was associated with high timbre (in terms of frequency spectrum and energy), simple rhythms and regular structure, while the opposite was true for ambient music. The comparison with classical music indicated that acoustic features related to timbre and rhythm are core descriptors, independently of genre, whereas those related to structure/dynamics were particularly relevant only for electronic music.
The second phase also resulted two principal components, to describe the participant’s evaluations of the music. Dance music was most represented by elements associated with overt motion (i.e. body movements): metaphors of “Movement” and “Force”, emotions of “Power”, “Joyful Activation” and “Tension”, as well as arousal and feelings of entrainment. On the other hand, ambient music was most represented by elements associated with to covert motion (i.e. internal body sensations): metaphors of “Wandering”, “Flow” and “Interior”, emotions of “Wonder”, “Ten-
derness”, “Peacefulness”, “Nostalgia”, “Sadness” and “Transcendence”. This supports an understanding of embodied music cognition, where the differences between dance and ambient electronic music mirror the differences between stimulative and sedative influences of music on the body of the listener. Furthermore, the two groups of correlations were consistent with the results found with classical music, suggesting similarities in how meaning is attributed to the music through embodied experiences of motion.
The third phase showed that acoustic features associated with the perception of timbre, rhythm and structure influence how ambient and dance music are evaluated by the listeners. As a general trend, higher frequency spectrum and more regular structure was correlated with higher ratings of overt motion elements. However, specific trends also apply to each category of electronic music. Ambient music was rated higher for covert motion and lower for overt motion in excerpts with lower energy, irregular structure and complex rhythm. For dance music, the peak of highest ratings for overt motion and lowest for covert motion in excerpts with moderate energy, regular structure and simple rhythm. Ratings for overt motion decrease and those for covert motion increase, if energy is too high, structure too regular and rhythm too simple. This pattern is not the reciprocal of the one found for ambient music. This may be a reflection of the importance of violations of musical expectations as a major component of musical meaning (Meyer, 1956). In this sense, dance music would have more potential to induce action, if containing a balanced degree of complexity and irregularity. Otherwise, the music becomes too predictable and the listener feels less inclined to move.
By exploring the link between acoustic features of sound and subjective evaluation of music, this study contributes to bridging the gap between MIR technology and music psychology (Acounturier & Bigand, 2013; Siedenburg et al., 2016). The choice of electronic music is also of scientific interest in itself, as it is a rapidly expanding contemporary genre of music, which is only marginally addressed in scientific research, in comparison to classical music for instance. Furthermore, the study contributes to a framework of embodied music cognition framework in two ways. First and foremost, the results are consistent with those of Schaerlaeken et al. (2022), linking feelings of entrainment with specific metaphors and emotions into two main clusters which broadly represent two modes of bodily interaction with the music (overt and covert motion). This study explores a different genre of music (electronic), and as such reinforces their ecological validity. Secondly, the conceptual framework mobilized to interpret the results provides tools to consider the role of the body in the construal of musical meaning. An enactivist perspective of cognition adequately accounts for differences between categories of electronic music (ambient and dance), with respects to diverging tendencies of action that the music can afford to the listener.
The degree to which the results of the study are generalizable depends mainly on two limitations. Firstly, all participants to the online survey belonged to WEIRD (Western-Educated-Industrialized-Rich-Democratic) population samples (Apicella, Norenzayan & Henrich, 2020). Considering the culturally-dependent nature of metaphors (Dancygier & Sweetser, 2014; Nuñez & Cooperrider, 2006; Zbikowski, 2008) and of emotions (Le Breton, 2021), the results cannot be generalized outide of WEIRD population. Secondly, the sample size of the music selection was relatively small (16 excerpts) and the range of different styles of music was limited and relatively polarized (ambient vs. dance). For the conclusions to be more robust, similar studies would have to be replicated with a wider variety and diversity of music.
The study, with it’s strengths and limitations, enlightens a few orientations for further research. The first points towards bridging the gap between MIR technology and music psychology, by promoting cross-disciplinary research and applications. It would be relevant for research in music psychology to explore different methodologies of MIR analysis, which are currently used to categorize large databases of music (Aucouturier & Pachet, 2003; Chen, 2014; Defferrand et al., 2016; Diakopoulos, 2009; Kirss, 2007; Lefaivre & Zhang, 2018; Nie, 2022). This approach would greatly benefit also from the integration of perceptual features (Aljanaki & Soleymani; 2018), as they provide a good mediator between the physical (acoustic) properties of sounds and the psychological phenomenology of listening to music. The second orientation refers to laboratory research. It would be relevant to investigate the neural correlates of rhythmic entrainment and action-perception coupling, while participants passively listen to music. This can be investigated by collecting EEG data of mu waves over the motor-related brain areas (Anderson, 2020; Fox et al., 2016; Hari, 2006; Pineda, 2005; Ross, 2022; Wu et al., 2016). This could further be integrated with physiological data (Bartlett, 1996; Gaston, 1951; Hodges, 2016; Matyja, 2016), in order to establish correlations between neural, physiological and subjective entrainment as well as emotional reactions. In addition, motion caption studies can be designed to further explore motor entrainment and specific movements associated with dancing to different styles of music (Toiviainen & Carlson, 2022). Finally, these previous orientations could be combined with ethnographical fieldwork in a fundamentally inter-disciplinary approach. This would imply contextually-situated qualitative research, based on the observation and analysis of individual movements and social interactions during live music performances. Unlike in a laboratory setting, this could shed light on phenomena social entrainment (Garcia, 2020; Witek, 2019). A mixed approach, able to integrate laboratory-based and fieldwork-based methodologies, would contribute to a holistic understanding of the experience of listening to music, from a perceptual, physiological, motor and social integrated perspective. Furthermore, intercultural studies could additionally shed light on universal vs. cultural aspects of music cognition. Bottom-up methodologies, similar to those involved in the creation of the GEMS and GEMMES, can be used to establish cross-cultural comparisons in musically-relevant metaphorical and emotional labels.
These opportunities for future scientific research also reflect potential societal applications. Firstly, it contributes to the inclusion of perceptual and cognitive aspects in the classification of music. Music providers which deal with large databases of music could improve their services by adopting concepts as metaphors, emotions and entrainment in complement of a classification methodology purely based on acoustic features of sound. Music psychology has the potential to provide insight where MIR technology finds its limitations. Secondly, music composition and performance can surely benefit from a better understanding of how meaning is conveyed through music, and the role of the body in this process. Metaphorical and emotional concepts can aid musicians to better comprehend their own intentions of expression through music and how to do so through with, and through their bodies. Similarly, such aspects can also have a positive influence in the field of music education. Metaphorical and emotional language is useful to transmit knowledge and insight from the teacher to the student.
The core of this paper is based on an empirical study conducted with the aim to explore interactions between the physical (acoustic) properties of sounds and the experience of listening to music. By doing so, this study also aims to provide a scientific contribution to an embodied (music) cognition framework. Two categories of electronic music (ambient and dance) were considered as inclined to exert rather sedative or stimulative effects on the body of the listener. The study was composed of three phases, and combines tools from Computational Sciences and Music Psychology. The first phase consisted in the extraction and analyses of a set of acoustic features with the MIR toolbox (Lartillot & Toiviainen, 2007), from 16 electronic music excerpts (length of 30 seconds). In the second phase, 172 participants evaluated the music through an online survey. They rated items of the Geneva Musical Metaphors Scale (GEMMES) (Schaerlaeken et al., 2019), to assess musical metaphors, of the Geneva Emotional Music Scale (GEMS) (Zenter et al., 2008), and the two-dimensional emotional model of valance and arousal (Russel, 1980), to assess musical emotions, and of the Musical Entrainment Questionnaire (MEQ) (Labbé & Grandjean, 2014), to assess subjective feelings of entrainment. In both phases, data was analyzed with Principal Component Analysis (PCA), to obtain a few relevant groups of variables based on the correlations between them. The third phase consisted in computing linear mixed-effects models, in order to explore the interaction between the physical properties of music (acoustic features) and the experience of listening to music (survey ratings). Finally, results of all three phases were compared with those from a similar study on classical music (Schaerlaeken et al., 2022).
The first phase resulted in in two principal components of acoustic features; two groups of mathematical descriptors of sound-waves associated with perceptual characteristics of music. Dance music was associated with high timbre (in terms of frequency spectrum and energy), simple rhythms and regular structure, while the opposite was true for ambient music. The comparison with classical music indicated that acoustic features related to timbre and rhythm are core descriptors, independently of genre, whereas those related to structure/dynamics were particularly relevant only for electronic music.
The second phase also resulted two principal components, to describe the participant’s evaluations of the music. Dance music was most represented by elements associated with overt motion (i.e. body movements): metaphors of “Movement” and “Force”, emotions of “Power”, “Joyful Activation” and “Tension”, as well as arousal and feelings of entrainment. On the other hand, ambient music was most represented by elements associated with to covert motion (i.e. internal body sensations): metaphors of “Wandering”, “Flow” and “Interior”, emotions of “Wonder”, “Ten-
derness”, “Peacefulness”, “Nostalgia”, “Sadness” and “Transcendence”. This supports an understanding of embodied music cognition, where the differences between dance and ambient electronic music mirror the differences between stimulative and sedative influences of music on the body of the listener. Furthermore, the two groups of correlations were consistent with the results found with classical music, suggesting similarities in how meaning is attributed to the music through embodied experiences of motion.
The third phase showed that acoustic features associated with the perception of timbre, rhythm and structure influence how ambient and dance music are evaluated by the listeners. As a general trend, higher frequency spectrum and more regular structure was correlated with higher ratings of overt motion elements. However, specific trends also apply to each category of electronic music. Ambient music was rated higher for covert motion and lower for overt motion in excerpts with lower energy, irregular structure and complex rhythm. For dance music, the peak of highest ratings for overt motion and lowest for covert motion in excerpts with moderate energy, regular structure and simple rhythm. Ratings for overt motion decrease and those for covert motion increase, if energy is too high, structure too regular and rhythm too simple. This pattern is not the reciprocal of the one found for ambient music. This may be a reflection of the importance of violations of musical expectations as a major component of musical meaning (Meyer, 1956). In this sense, dance music would have more potential to induce action, if containing a balanced degree of complexity and irregularity. Otherwise, the music becomes too predictable and the listener feels less inclined to move.
By exploring the link between acoustic features of sound and subjective evaluation of music, this study contributes to bridging the gap between MIR technology and music psychology (Acounturier & Bigand, 2013; Siedenburg et al., 2016). The choice of electronic music is also of scientific interest in itself, as it is a rapidly expanding contemporary genre of music, which is only marginally addressed in scientific research, in comparison to classical music for instance. Furthermore, the study contributes to a framework of embodied music cognition framework in two ways. First and foremost, the results are consistent with those of Schaerlaeken et al. (2022), linking feelings of entrainment with specific metaphors and emotions into two main clusters which broadly represent two modes of bodily interaction with the music (overt and covert motion). This study explores a different genre of music (electronic), and as such reinforces their ecological validity. Secondly, the conceptual framework mobilized to interpret the results provides tools to consider the role of the body in the construal of musical meaning. An enactivist perspective of cognition adequately accounts for differences between categories of electronic music (ambient and dance), with respects to diverging tendencies of action that the music can afford to the listener.
The degree to which the results of the study are generalizable depends mainly on two limitations. Firstly, all participants to the online survey belonged to WEIRD (Western-Educated-Industrialized-Rich-Democratic) population samples (Apicella, Norenzayan & Henrich, 2020). Considering the culturally-dependent nature of metaphors (Dancygier & Sweetser, 2014; Nuñez & Cooperrider, 2006; Zbikowski, 2008) and of emotions (Le Breton, 2021), the results cannot be generalized outide of WEIRD population. Secondly, the sample size of the music selection was relatively small (16 excerpts) and the range of different styles of music was limited and relatively polarized (ambient vs. dance). For the conclusions to be more robust, similar studies would have to be replicated with a wider variety and diversity of music.
The study, with it’s strengths and limitations, enlightens a few orientations for further research. The first points towards bridging the gap between MIR technology and music psychology, by promoting cross-disciplinary research and applications. It would be relevant for research in music psychology to explore different methodologies of MIR analysis, which are currently used to categorize large databases of music (Aucouturier & Pachet, 2003; Chen, 2014; Defferrand et al., 2016; Diakopoulos, 2009; Kirss, 2007; Lefaivre & Zhang, 2018; Nie, 2022). This approach would greatly benefit also from the integration of perceptual features (Aljanaki & Soleymani; 2018), as they provide a good mediator between the physical (acoustic) properties of sounds and the psychological phenomenology of listening to music. The second orientation refers to laboratory research. It would be relevant to investigate the neural correlates of rhythmic entrainment and action-perception coupling, while participants passively listen to music. This can be investigated by collecting EEG data of mu waves over the motor-related brain areas (Anderson, 2020; Fox et al., 2016; Hari, 2006; Pineda, 2005; Ross, 2022; Wu et al., 2016). This could further be integrated with physiological data (Bartlett, 1996; Gaston, 1951; Hodges, 2016; Matyja, 2016), in order to establish correlations between neural, physiological and subjective entrainment as well as emotional reactions. In addition, motion caption studies can be designed to further explore motor entrainment and specific movements associated with dancing to different styles of music (Toiviainen & Carlson, 2022). Finally, these previous orientations could be combined with ethnographical fieldwork in a fundamentally inter-disciplinary approach. This would imply contextually-situated qualitative research, based on the observation and analysis of individual movements and social interactions during live music performances. Unlike in a laboratory setting, this could shed light on phenomena social entrainment (Garcia, 2020; Witek, 2019). A mixed approach, able to integrate laboratory-based and fieldwork-based methodologies, would contribute to a holistic understanding of the experience of listening to music, from a perceptual, physiological, motor and social integrated perspective. Furthermore, intercultural studies could additionally shed light on universal vs. cultural aspects of music cognition. Bottom-up methodologies, similar to those involved in the creation of the GEMS and GEMMES, can be used to establish cross-cultural comparisons in musically-relevant metaphorical and emotional labels.
These opportunities for future scientific research also reflect potential societal applications. Firstly, it contributes to the inclusion of perceptual and cognitive aspects in the classification of music. Music providers which deal with large databases of music could improve their services by adopting concepts as metaphors, emotions and entrainment in complement of a classification methodology purely based on acoustic features of sound. Music psychology has the potential to provide insight where MIR technology finds its limitations. Secondly, music composition and performance can surely benefit from a better understanding of how meaning is conveyed through music, and the role of the body in this process. Metaphorical and emotional concepts can aid musicians to better comprehend their own intentions of expression through music and how to do so through with, and through their bodies. Similarly, such aspects can also have a positive influence in the field of music education. Metaphorical and emotional language is useful to transmit knowledge and insight from the teacher to the student.
Identifiants
Type de publication
master thesis
Dossier(s) à télécharger
