Music and Emotion

Living apart together: a relationship between music psychology and music therapy

Annemiek Vink

This article was originally printed in Nordic Journal of Music Therapy, 10(2), pp. 144-158. The article is republished here with the kind permission from the author. 
 

Introduction

`Ladies and Gentlemen, my respected and beloved music friends! Musica animae levamen, music is medicine for our souls.' These are the opening lines of the book `Die musikalische Hausapotheke/The musical home medicine cabinet' written by the German musicologist Christopher Rueger (1991). In this work, Rueger describes a variety of musical recipes to cure an equally impressive variety of disorders. If we take a closer look into his laboratory, we can read that listening to Beethoven's Symphony No.2 helps to cope with feelings of depression and doubt. The popular Goldberg Variations of Bach will decrease sleeping disorders.

If it could only be so simple. For many lay people, these descriptions are their idea of what music therapy is all about. Many people have read about the `Mozart Effect' and a music therapist is often pictured in non-medical journals as being an equivalent of `a healthcare deejay'. There are, however, many disciplines involved in the study of the effects of music on the listener. It is the music psychologist who looks at effects of music in the general listener, e.g. what emotions can be induced by music in the listener? Cognitive theory studies how music is perceived and how cognitive schemata are activated when listening to music. How does our brain react to music? The music therapist uses music as a therapeutic medium with a variety of different client populations in yet another discipline.

For a layperson, these fields seem logically intertwined. Music induces emotions and this works therapeutically as various cognitions and feelings become activated, of course. In practice, there is still little collaboration between these fields of expertise. In fact, music researchers studying cognitive aspects and brain activity even state that it is far too early to be able to describe precisely how a healthy person reacts to music, both physically and psychologically. Before this knowledge is gathered, it is far too early to state that music is therapeutic, let alone to be considered as `medicine' as Rueger described. How will we ever be able to develop cogent theories from neurological research if we can't even agree on which halves of the brain processes various aspects of music experience (Kaufmann & Frisina, 1992)? Music therapists state on the other hand, that music cannot be considered as medicine as it is not merely the music from which the client benefits. Other factors contribute to the therapeutic effect as well, such as the group interaction, the interaction with the music and the therapeutic alliance. What are current research insights from the field of music psychology and can they be applied in music therapy, in an arts relationship?

To examine the mutual aspects of these disciplines, the focus in this review will be on the question how music is related to the experience of emotion. Everyone recognizes the ancient old idea that music affects man. But how? This question has turned out be an enormous difficult task to answer scientifically.

Theoretical orientations of musical emotions 

Music psychology research shows us that non-knowledgeable listeners respond emotionally to music as do music performers, both during playing and during listening to music. But how can these experiences be described? Is it accurate to state that `musical emotions' are experienced? Researchers discuss if it is legitimate to state that feelings experienced in relation to music are actual emotional states.

In general psychology, emotions are currently described as specific process-oriented response behaviours. To clarify the word process-oriented, an emotion is always directed at something, whether it be a particular circumstance or a person etc. Specific attributes within the emotion appraisal process are the personal involvement with the stimulus, the cognitive appraisal of the stimulus and a behavioural change towards the stimulus (Frijda, 1988). The presence of these characteristics and the degree to which they are represented in a person depend on the type of emotion experienced.

The concept of a `musical emotion' is difficult to define according to general emotion theories. A variety of different emotional responses and experiences can be found under the header of `emotion'. Although we can often read about the concept of `musical emotion' in the literature, regretfully how a `music induced emotion' can be defined is not clarified. Also the term `emotion' in the context of music listening is open for debate.

Often, emotions can be classified in order of their (biological) functions. A classical example is the experience of fear in this context. The cognitive appraisal that a certain situation is threatening leads to a behavioural change, e.g. fleeing or avoiding the situation. In a musical context there are no apparent biological functions. Although, everywhere on the world, music is played and listened to, we cannot state that we cannot live without the presence of music. As such, the intensity of an emotion experienced is difficult to explain in a musical context to directly result from music (Dowling and Harwood, 1986).

In everyday life, emotions are induced by concrete circumstances within a realistic setting, which we often cannot control, such as being involved in a car accident. This `real life' influence is often found lacking in a musical context. Although music can be frightening and threatening, there is always the sense of control, for instance by turning the music off. The use of the verb `emotion' is often therefore not acknowledgeable in a musical context. In recent music psychology literature we can see a clear shift in descriptions used to refer to musical experiences. In most of the research focussing on musical emotional responses, the focus has actually been on specific characteristics of an emotion: such as appraisal and arousal components. In some cases it is possible, according to Sloboda (1991), that music can induce intense emotions, but he states it is more accurate to describe `musical emotions' as mood states, experiences and feelings that are influenced by music listening. Music functions as a catalyst or stimulus for the experience of emotions, a characteristic often witnessed and used in music therapy practice.

The concept of `musical emotion' has many conceptual difficulties. Sloboda and Juslin (2001) describe how emotion theories such as Frijda's (1988) are beneficial to the study of musical emotions. They state that music psychologists generally are reluctant to turn to emotion psychology for theoretical guidance. On the other hand, they hold that it is interesting to note that psychologists themselves do acknowledge that art may evoke strong emotional response, but few scientists seem to have considered this problem worthy of study. Sloboda and Juslin (2001) state it is significant that one of the most influential analysis of music and emotion to date, namely that of Meyer (1956) is based on a theory of emotion.

According to Meyer (1956) there are certain elements within the music, such as a change of melodic line or rhythm, that create expectations about the future development of the music. The expectation the listener has about the further course of musical events is a determinant for the experience of `musical emotions'. Music arouses expectations, which may or not be directly and immediately satisfied. He states: ...The greater the build-up of suspense of tension, the greater the emotional release upon resolution...(Meyer, 1956, p.28). The music arouses tendencies and this fulfils the conditions according to the arousal of affect. The role of emotion in this process is complex. The primary dimension on which emotion is experienced is in terms of a build-up of tension and release of tension, both within the music and in the listener. The more elaborate the build-up of tension to a certain climax, the more intense emotions will be experienced. When resolution follows, relaxation follows.

An important implication of Meyer's theory is that any conscious insight in this process reduces the emotional effect of music. This implicates that a `naive' listener without much music theoretical knowledge has a predominant affective response towards music whereas a trained listener reacts mostly cognitively while listening to music (Jansma& de Vries, 1995).

Within the music psychology research literature, Mandler's (1984) and Berlyne's (1971) cognitive theories of emotion are occasionally referred to. Mandler holds arousal as a prerequisite for the experience of emotion, but sees arousal in essence as of minor importance. Mandler attaches the more value to the cognitive appraisal of bodily experiences. Emotions, in line with Meyer (1956), are caused by interruptions of expectations about future events. The interruption, depending on the personal involvement, causes arousal for which an explanation is sought. Information processing is important for the experience of emotions. The experience of arousal is a result of a `mis-match' between existing knowledge structures and the integration of new information. This thought `matches' Meyer's idea, who states that as a result of unexpected musical changes, arousal is experienced. According to Mandler, negative emotions will be the result of an extreme mismatch between musical information and existent musical schemata. Accordingly, positive emotions are experienced when musical information matches existing cognitive schemata.

A comparable emotion theory is from Berlyne (1971). He hypothesized that when we listen to music, we take into account factors such as the complexity, familiarity and novelty of the music. The degree to which the music sounds familiar determines if the music is experienced as pleasurable or uncomfortable. The hedonic value is low when the music is entirely new to the listener, and progresses with increasing familiarity and will decrease again when the music is totally known. This process can be described with a reversed U-curve. According to Berlyne the experience of arousal is an important factor in the experience of emotion, equal to Mandler. But he goes further, he states that it is the listener's intention to experience arousal while listening to music. Musical preference is based on the preference to maintain the level of preferred arousal constant. The more complex the music, the more arousal will be experienced. The listener seeks activation, but to a certain extent. When the maximum is achieved he will avoid further activation. In a listener with more elaborate music knowledge structures, arousal will be less when listening to complex music. As such, a trained listener, will both according to Berlyne and Mandler, hold a greater preference for complex music than a naive listener as the threshold for experiencing arousal is higher. Fischer (1981) demonstrated that subjects who had taken amphetamine, a drug which causes arousal, preferred the more complex music. There are many experiments that backup the ideas of Berlyne. Subjects prefer melodies that are intermediate in the number of different pitches used and which have intermediate levels of complexity (Weinberger, 1998). Weinberger values Berlyne's theory of importance to therapy as it may explain why the same music can produce different emotions in the same person at different times, a known difficulty complicating research. He also states that the findings may be of help with further direct investigations of the brain substrates of music and emotion. A basic understanding of these relationships should also benefit future uses of music therapeutically.

The theories described above, are often referred to in describing the emotional impact of music. In general it is stated that Meyer's (1956) theory is comprehensive but problematic to test empirically and there are no research studies known which have directly tested Meyer' s assumptions in relation to musical emotions. Meyer himself acknowledged this problem by doubting the possibilities to test the existence of `musical emotions' in a scientifically controlled manner. Sloboda (1985) states in this context, that many researchers have the tendency to come up with their own theorizing about the effect of music, with the result that feedback for the development of a general theoretical framework explaining the effects is lacking. This can also be seen in music therapy research, many music therapists have studied or described if their own practice has any effect. The research studies are as such often very different in therapeutic orientation and difficult to compare related to client populations. 

Musical expression of emotion 

Are there certain elements in music, which may invoke musical experiences in the listener? Going back in the literature, we will find the impressive line of studies that Hevner conducted. Hevner (1935, 1936, 1937) is one of the first researchers who systematically studied which musical parameters are related to the experience of emotion. She adapted various short existing piano pieces and played them for the subjects, both the original version and an adapted version. In the adapted version, she constantly only manipulated one musical element, such as the mode (major or minor), the harmony (simple-complex), rhythm (steady-fluent), tempo (fast-slow) and the melody line (increasing- decreasing). Other musical parameters were held as constant as possible. After playing, she asked the subjects to indicate on the `Hevner's adjective circle', which emotion verb best described the emotional content of the piece. Tempo and mode had the strongest impact on the listener, when describing the experienced emotion in the music. Piano music played fast in major was described as cheerful and in contrast the slow piece in minor as dreamful and sensitive.

After Hevner, it has been repeatedly shown that various musical elements, with a particular mode and tempo, are expressive of emotion in music. Based on research studies such as these, it has been easily deduced that music induces emotion in the listener. This conclusion is however not warranted. The work of Hevner demonstrates which emotions are experienced in the music and not necessarily refer to the emotions experienced in the listener. Maybe so, but often it has not been the focus of research. On the basis of attribution based research, it can only be stated how musical elements are expressive of emotion and how the emotional content of the piece is valued.

A continuing problem that occurs in research such as Hevner's, is that she chose musical parameters that are representative for the music in general. Music often induces reactions and emotional responses in very specific moments in the music, such as by sudden changes in tempo or a climax of musical theme, as we will read later from Sloboda (1991). Statements about the tempo of a musical piece are merely descriptive of one characteristic of emotional expression of the music or the musical experience in the listener (Jansma & de Vries, 1995).

According to Desain and Honing (in Jansma &de Vries; 1995) false is also the expectation that musical parameters can be manipulated in a totally independent manner. They mention as an example, the manipulation of tempo. A piece cannot be simply played twice as fast. Also temporal relations between the notes accordingly need to be adjusted in order to let the music still sound natural. Still, with great consistency, it has been shown that musical characteristics, such as the mode, rhythm and tempo, are perceived by the listener to be expressive of emotion (for an overview of this topic see Gabrielson & Lindstrom; 2001).

In most of the music psychology research studies, only classical music is studied. In the music therapy literature we can read that listener's preferences range from heavy metal to classical music. Also in the context where musical emotions with children are studied, mostly classical music has been used as test material. The generalisability of the results, when only studying classical music, is therefore small. Still, there are reasons why researchers prefer classical music, as classical music is mostIy not accompanied by sung lyrics, which is the case for most of the popular music repertoire. For the earlier researchers, there were also interesting ethical reasons. Wing (1968) described that while developing his musical ability test:

"Jazz music was not included, as this would be unlikely to yield examples of really good harmony, would be likely to prejudice the authorities against the test, and would waste the children's time if they were listening to poor music" (Wing, 1968, p.37).

In more contemporary research studies, also popular music, jazz or religious music has been included. This is an important development for music psychology findings to be generalised to music therapy practice.

Regardless of the above-mentioned methodological problems, it is the common research outcome that listeners in general can describe with great accuracy which emotional expression is communicated. But of interest also for music therapy is whether these findings can be related to the experience in the listener. Which emotions or reactions are exactly induced in the listener through music? Various research studies have tried to answer this question with a variety of physiological and cognitive listening tests.

Physiological studies involving musical emotions. 

Both Berlyne (1971), Mandler (1984) as Meyer (1956) see arousal as an important factor in emotional experiences while listening to music. A typical example of physical experiences reported when listening to music is the experience of thrills. Goldstein (1980) illustrated that thrills were experienced by 96% of the subjects.

Already in the 18th century an effort was made to study the effect of music in relation to physiological changes. One of the most known researches of that time is from Gétry, who published in 1741 about the influence of music on the frequency of heart rate (in Dainow, 1977).

Often referred to in the music therapy literature, is the relation between music and physical changes in the clients. Smeijsters (1995) formulated the analogues process model, in which he described that musical characteristics can be perceived in the client's pathology and vice versa. In many research studies, it has been tried to establish if there is a relationship between for instance musical rhythm and physical rhythm, following Pythagoras idea that musical vibrations can bring about healing physical changes. The average beat rate in music equals almost exactly the average heartbeat, with 72 to 80 beats per minute. The precise relationship between musical rhythm and heartbeat is still unclear, although support can be found that the heartbeat follows musical rhythm. Stimulating music increases the heart rate and sedative music reduces the heart rate. Each type of music increases the heart rate. This increase is higher when listening to stimulating music than to sedative music.

A problem that occurs when measuring the heart rate is to establ itish that is also the measurement of the intended emotional change. An illustrative example was provided by Harrer (1977) who compared the heart rate of the famous conductor Herbert von Karajan while he was directing Leonora Overture No.3, to his heart rate while flying a sports aircraft. During flight-time, the heart rate peaked at 115 beats per minute while during conducting a maximum was achieved of 150 beats per minute. The maximum increase of pulse frequency, telemetrically recorded from Herbert von Karajan while conducting the Leonora Overture No.3, was obtained during those passages with the greatest emotional impact upon the conductor. These were the same passages that he singled out in subsequent conversations as being the ones he found most profoundly touching. At these moments the pulse rate increased for a short while, to twice the level of the initial value. When the tape was played back to him and a further pulse measurement was made, the same peaks were found. Changes however were much greater while he was conducting.

Next to changes in heart rate, changes in breathing can be perceived as a result of intense emotional experiences. The frequency of breathing and heart rate are normally to a certain extent co-dependent. An increase of muscle activity together with active breathing increases the hearth rate, as can be seen for instance during hyperventilation (Frijda, 1988). Ries (1969) found clear correlations between the respiration amplitude and the emotional response towards music. The relationship between the breathing amplitude and the subject's affective responses were highly correlated indicating that the more a subject reported liking the selection, the deeper his breathing became. Research studies involving respiration measurement are often not easy to interpret (Harrer, 1977). The main reason is that often only one characteristic of respiration is measured, such as the frequency or the amplitude. In general, it is stated that the frequency of breathing increases when listening to preferred music and that the breathing becomes deeper. The foremost problem however is that there are too many individual differences to be perceived in respect to irregularities of breathing and the depth of breathing, for instance as an result of extraneous factors such as smoking.

Many factors can influence physiological measurements. The amount of coffee a subject has drunken before the registration is for instance an influence in heart rate measurements. Often the amount of variables for which should be controlled for is too large to establish a reliable representative conclusion. Also is it an ancient old idea that each type of emotion is characterized by a specific pattern in physiological changes (Rimé et al., 1990). People/clients who are for instance fearful do not respond equally on physiological measures and we cannot estimate from a certain measurement that someone is experiencing `moderate happiness'. Also it is possible that people may well feel angry or sad without demonstrating any physiological changes. Stavenga (1979) compares the difference between mentally handicapped and normal adults in this context. Both can display the same physiological changes, while experiencing them entirely differently in a psychological manner. To return to Frijda (1988), the experience of arousal should not be seen as a key feature of an emotional experience. Still, it is interesting to see if this is also the case for musical experiences, as so many hold that arousal is indeed a key feature in the appraisal of music listening. Bever (1988) states that the perception of musical structures always arouses in relation to the appraisal of the aesthetic value of the music.

Physiological measurements are extremely popular though. Often music psychology and music therapy research outcomes are seen as too abstract, and defined clinical measurements are seen as a solution to communicate research to other healthcare professionals. Often we see a combination of approaches in modern-day research, the qualitative method to assess the individual experiences combined with physiological measurements. Through qualitative research these individual differences in experience can be more easily assessed.

With the arrival of new methods it is hoped that more insight will be gained in the future how musical structures are perceived and how emotional responses are related for instance to brain processes. According to Peretz (2001) musical emotions appear isolable in the human brain. She states that current evidence is pointing to the existence of a specific neural arrangement for certain musical emotions. One of the intriguing new findings is that subjects exhibit greater left frontal activity to music expressing joy and happiness (Schmidt & Trainor, 2001 in Peretz, 2001) and greater relative right frontal EEG activity to music expressing fear and sadness. Currently with new measurements such as functional Magnetic Resonance Imaging (fMRI) is studied in more detail which brain activity can be perceived when subjects are listening to music. But also physicians here state that there is a long way to go before accurately music perception can be measured.

To conclude, music does bring about physiological changes although it is not yet clear how these changes are related to musical characteristics. Clear outcomes in this context are of value for music therapy. If music produces physiological and psychological effects in healthy persons as listeners, then it may be assumed that people with known diseases respond in specific ways (Aldridge, 1996). If music is known to influence a physiological parameter such as the heart rate or blood pressure then Aldridge argues music can be used therapeutically for patients who have problems with heart disease and hypertension. These are interrelations from which both disciplines can benefit.

Cognitive emotional studies of music appraisal 

In line with the popular tradition in psychology, currently cognitive processes are studied, both in normal healthy adults and in diverse client populations. A variety of different cognitive tests have been used to gain more insight in the actual experience of the listener, whether he is a healthy adult, a routine listener, or a music therapy client. Still, this line of research is in its infancy. Most of the music psychology research studies make use of validated instruments from psychology studies. There are many logical arguments for this choice in terms of the representation of the study, the decrease in costs etc. But generally, these lists consist of emotion terms that are difficult to interpret in the context of music. Waterman (1992, in Jansma en de Vries, 1995) used a well-established list from Ortony, Clore and Collins (1988) to let subjects score which emotions they had experienced in response to music. This question is very aspecific, because there are many different ways how one can respond to music: as a product, admiration for the performance etc. Also it is unclear what is meant with `experienced in response to music'. There is a clear distinction between emotions experienced when playing music or listening to music, which is generally ignored in research. In table 1, the responses can be seen from some of the first year music therapy students at the Conservatory Enschede in the Netherlands, to the question how they experience the described emotions both during listening and playing music.
 
 

Table 1: The experience of emotions on the Ortony, Clore and Collins scale (1988) by music therapy students experienced when listening to music and performing music. |View table in full size|

As can be seen, most of the subjects experienced joy while listening to music. It is however difficult to understand whether this outcome was the direct result of the music listening, e.g. music induced emotion, or invoked through musical associations and attribution processes. Also, joy experienced during play has very different origins then joy experienced when listening to music. Jansma & de Vries (1995) give us another complicated interpretation with this scale. How should we understand the study outcome that many of the subjects have experienced hope or disappointment while listening to music? What do these results mean? Are they directly caused through the music, regardless of the style? Or are they the result of a lucky or disappointing expenditure in the cd-store. Emotions such as reproach are typically not experienced in a musical context.

Waterman (1992) demonstrates that most adult listeners have a variety of musical responses towards music and that a variety of extrinsic factors influence these responses. Meyer (1956) states it would be interesting to learn more how listeners, whether client or healthy adults, experience emotions directly from musical characteristics or through musical associations, the well known `Darling, they are playing our song' idea. To repeat Meyer's assumptions, he held it to be difficult to experimentally relate the effect of music to induced emotions in the listener. From the few experimentally controlled research studies it can be stated that some emotions are more easily induced than others and that effects are mostly found on the dimension of relaxation/tension and happiness/sadness (Clark & Teasdale, 1985). It remains unclear however, how musical parameters are related to these effects and open for debate is whether such outcomes could be of value for music therapy.

Typical for most of the emotion checklists is that fixed categories have been used. Problematic here is that people have a tendency to polarise their opinions on certain emotion traits, regardless of the emotion object. A method, which more freely gives room to the variety of emotional responses, is the free response technique. This method has been used on few occasions by music psychology
researchers in the context of music appraisal. Sloboda (1992) asked adults to recall any memories from the first ten years of their lives that involved music in any way. It was not asked specifically whether there was an emotional association with the music, just what musical experiences one recalled from childhood. From the free descriptions, it was determined that in 39% of the cases emotion was indeed a classifying characteristic. Also on adult age, emotion was an important factor in the free descriptions of music experiences. Sloboda (1992) let 67 regular music listeners describe in their own words, the nature of their most valued emotional experience to
music. Two dominant themes emerged in the spontaneous descriptions. First, people tended to use music as a `change agent' to alter their mood state (n-41). This was reported through statements such as `music relaxes me, when I am tense and anxious' and `music motivates and inspires me to be a better person'. Secondly, people reported that music was used as a method of catharsis to promote the intensification or release of already existent emotions (n-34). Example statements included `music releases emotions'
and `music helps me discover what I am actually feeling'. Sloboda states that the common factor to all examples was that music does not create emotion, but rather it allows a person access to the experience of emotions that are already `on the agenda'. A statement, all music therapists will recognize.

A positive factor of such open-ended studies is that it can be studied if one refers to emotion when describing musical experiences and if these emotions are experienced directly in response to music. A disadvantage of such retrospective studies is the question of how reliable these descriptions of childhood experiences are. It can be assumed that regardless of the subject one describes of ones childhood, emotional descriptions will be given. If someone were to describe riding a bicycle in childhood years for instance, these terms were also likely to be found, simply because we evaluate all sorts of stimuli with affective responses. The question is again if they are actually emotional experiences directly invoked through music which are measured?

Researching emotional experiences directly related to music is complex, as has been demonstrated above. In the context of music listening, the effect seems to be related to earlier associations that are different for us all. In the studies described above it is clear that association processes are present, although it has not been the study focus. From many music therapy studies involving the effects of Guided Imagery and Music, these associational processes have been well demonstrated and documented.

Also in relation to specific musical characteristics, there are still many more layers than only structural characteristics with which the emotional response is associated, such as admiration for the performance. One single music piece can invoke a large variety of emotional responses: one of the powers of music, but a headache for the researcher.

Many researchers have tried to identify a single emotion in the listener when he is listening to music, such as sadness or joy. Emotional experience in the context of music is an interesting but difficult process to study. Music cannot simply make someone sad or happy. There are many other responses thinkable, which may occur at the same time. Exactly for these reasons, Meyer (1956) was sceptical about the possibility to ever identify clear relations between the music listened to and the responses invoked in the listener. He held it impossible that a listener can specify what elements in the music invoke for instance thrills. Sloboda (1991) studied this claim and argues that trained professionals can describe what elements and on what time they induce a response. The subjects all first described which physical changes they had experienced when listening to music. 90% of the respondents reported to have experienced thrills, 88% indicated to have laughed and 85% reported to have cried while listening to music. The listeners turned out to be very capable in stating which musical selections induced the responses and even more specific during which moments in the music. The specific fragments were related to the published scores and classified to the musical features they contained. Twenty passages related to the response crying which contained features such as melodic appoggiaturas and melodic or harmonic sequences. Shivers were related most frequently with new or unprepared harmony. These findings correspond with Meyer's (1956) and Mandler's (1984) expectations. The specific moments in the music were all associated with sudden, unexpected musical changes.

From this study we may conclude that there are indeed musical elements that can be described in association with the emotional experience. Still, the experiences have been described in retrospect. The physiological changes reported in this study, may bear no resemblance to the actual emotions experienced at the time when one was listening to the music, as Rime, Phillipot and Cisamolo (1990) found when comparing the emotions subjects recalled and experienced in everyday life.

Only on few occasions, musical responses were studied when subjects were actually playing or listening to music. Waterman (1996) compared differences in emotional responding to music between trained and untrained listeners. He derived 14 fragments from Sloboda's (1991) study, which were found to induce emotional responding. Fragments selected were jazz, pop and classical music. When listening to the music, the listener could activate a switch-button, when something in the music caused something to happen to him.

The outcome measure was the number of responses per bar of music. This was the task for the naive listener. The professional musicians performed the same pieces with also an inventive responsemode, which they could press to indicate reactions experienced. Waterman found that subjects could reliably indicate that the music moved them and did not find differences between the two groups in the degree that they were able to indicate events within musical extracts to be emotionally loaded. This was apparent when the music was known or unknown and to a lesser degree when the musical genre was found pleasing or displeasing.

Methodological problems are numerous in these sorts of approaches. Emotions are often of short duration and when they are studied in an experimental setting it can be questioned in what way they were influenced by task demands, such as in Waterman's study, pressing a button. The experimental setting holds an entirely different atmosphere to the couch at home. The described questionnaires can be validated to a certain extent by testing them in different populations and in different circumstances. Still, it is problematic that people have the tendency to select basic emotions to describe their experiences and they are less capable to describe the nuances (Frijda, 1988). According to Frijda, these questionnaires do not help us to gain any insight in how complex emotions can be understood, such as experienced when listening to music. What does one exactly experience when indicating that someone is happy or sad? Aldridge (1996) warns for the quantitative approach in music research. He holds that we should be careful not to organize nature according to concepts imposed on it and demonstrates the need for a more phenomenological, holistic understanding of musical experiences through qualitative research.

Standard questionnaires are difficult to use in a musical setting. Bartel's (1992) CART-M test (Cognitive-Affective Response Test) has been especially developed to measure musical responding. In this questionnaire there has also been made a distinction between different ways of responding: cognitively or affectively. Sloboda (1985) refers to humour, to clarify the influence of cognitive and affective stages in musical responding. We can only laugh about a joke when we have heard all of it. The cognitive stage is a necessary prerequisite for the affective phase to occur. Affective responses may or may not occur, someone may equally well have understood the joke but for some reasons could not laugh about it. With music something similar happens. When someone listens to music, he always makes a mental representation of the music and affect may or may not occur. Both within music and in emotion, these cognitive and affective phases can be distinguished. In the CART-M they are called a formal-intellectual dimension and an emotional-expressive dimension. With a semantic differential test it was studied if emotional responses were different among subjects when listening to different music styles. The responses to classical music were more cognitive than towards jazz and country music. This study implicates that more musical repertoires should be included than only classical music. Also it was demonstrated that the more familiar the music was, the more affective the response became. Also the amount of musical training is a determinant in the response. The subjects with music theoretical knowledge predominantly responded cognitively to the music and untrained listeners affectively. These findings again correspond to the theories of Meyer (1956), Mandler (1984 and Berlyne (1971).

Music psychology meets Music therapy

Although the topic of music and emotion is central to both music psychology and music therapy, research findings from the respective fields are not commonly shared. Only some articles written on the topic of music therapy relate to how music therapy and music psychology could benefit from one and another.

Generally, in music psychology groups of healthy people are compared to each other, e.g. groups who have received formal music education and groups who have not. What would be interesting to know next is how do healthy people differ in this context to certain client populations, to understand more about musical emotions.

Music psychologists could for one benefit from music therapy studies in this area. By now it is more or less clear that people respond emotionally to music but interesting to know next what are the variations between different subjects. Smith (1989) studied for instance the listening preferences in subjects ranging from 18 to 89 in preferred loudness levels, a typical music psychology topic of relevance to music therapy practice.

A task for music therapists could be to study further how musical preferences are related to a variety of client populations. How do demented elderly for instance perceive music and emotions from a music psychological view. It would be an enrichment of the literature to compare these to music therapy findings. As music psychology generally focuses on healthy adults, it could be beneficial to music psychology research to study preferences and emotions in a therapeutic context with client populations such as the elderly and comatose people.

In this last section, some researchers will be highlighted which have incorporated music psychology findings in their music therapeutic research. Bunt (1995) is one of the few authors who mentions the common grounds between music psychology and music therapy in understanding musical development. For developing therapeutic goals and finding ways to attune to the child, it is important to know the capacities of children per se in understanding and perceiving music. When are children able to sing spontaneously? What musical qualities children adhere to when they have emotional experiences? Terwogt and Grinsven (1991) demonstrated that young children are able to recognize basic emotions in classical music. Here the music therapist can also benefit from general psychological findings. Children at the age of four are better in expressing these emotions than recognizing them (Field and Walden, 1982). At later ages children are more able to differentiate between the variety in positive and negative emotions (Manstead, 1993).

Most of the descriptions resemble the trend to music psychology: how is emotion perceived/experienced when one listens to music. In practice already in receptive music therapy, instruments such as derivates of the Hevner's adjective circle are used, to let the person become adapted to the therapy and to gather emotional connotations which the person has with the music for assessment purposes. Regretfully, these instruments are often not described in the clinical literature nor has their use and effect been documented. Meyer (2001) describes how often it is not the patterning of the music per se that evokes emotion, but the connotations that it has for different listeners.

This idea has been described in various music therapy studies. Smeijsters, Wijzenbeek and van Nieuwenhuijzen (1995) hypothesized that in the life of depressive patients, several values are very important (e.g., chaos, authority, death, religion). It was studied in an explorative study if specific musical excerpts can evoke these connotations. Depressive clients (n-44) listened during two sessions to 16 selected musical fragments (such as Bach (2 min), Symphony no. 2 E flat major, Part 1: Allegro di molto). Patients were asked to select, from a list of 14 values, those values (such as violence, family and relaxation) that came to mind as they listened to each excerpt. Results showed that the fragments indeed evoked the values but they found no one-to-one relationship. The authors underline the complexity of interpreting these results. In general, most studies are complicated by the fact that there are so many antecedents and responses that interact, which are difficult to measure. General emotions terms are thus likely to yield results, but often the nature of the response is unclear. Research is needed which is more directed at understanding the process of how emotions develop in music than the end-state `static' experience.

A method which is typically based on music psychological findings is Guided Imagery and Music developed by Helen Bonnny. The method is based on the assumption that the most appropriate music can be selected for healing purposes. The client listens to the music in a relaxed state and shares the experiences with the therapist. The music is selected for specific themes such as `positive affect'. Erdonmez Grocke (1999) has written extensively on the application of GIM, also in reference to a variety of client populations. Bonny states (1986) about the assumptions of GIM that research on the music-assisted imagery process (GIM) demonstrated that personal music preferences were not usually applicable for healing when used in therapeutic settings. This statement is one of debate for music therapy for which the music psychological findings are interesting. Others have demonstrated that music should be selected carefully in line with personal preferences. Gerdner (2000) demonstrated in her study the effects between listening to standard relaxation music compared to individualized music in its effect to reduce agitation in demented elderly. Both types of interventions were clearly more effective compared to baseline measurements. In all aspects, carefully selected music based on the person's preference and personal background was far more effective than standard relaxation music. Perhaps it could be hypothesized that when music meets collectively shared connotations such as presented in GIM, emotions can be invoked although with different associations. This would be interesting to research further.

In most of the research -both in music therapy and music psychology- is chosen for the option that people listen to music. Not many researchers have addressed the topic of the experience of emotions while performing or in music therapy while improvising. The processes present here are difficult to grasp with scientific hands. Bunt and Pavlicevic (2001) describe aspects of generating and receiving of emotions through music in the realms of clinical improvisation. They studied an outcome of music psychological research in a music therapeutic context. In the music psychology literature, specific instruments are seen as being more suitable than others to convey a particular emotion. Gabrielson and Juslin (1996) instructed nine professional musicians to sing or play -on violin, flute or electrical guitar- particular emotional expressions such as happiness, sadness and anger as well as without expression. In general, the listener easily understood which emotional expression was played or sung. The researchers concluded that the effect is influenced by the type of musical instrument or whether one sings an emotional expression.

Bunt and Pavlicevic (2001) studied how therapists judged short improvisations on emotional intentions such as happiness, sadness, tenderness, anger and fear. Two groups of music therapists were included. The first group consisted of music therapy trainees and the second group of music therapy researchers and teachers. Members of each group were asked to improvise freely on one of the described emotions, after which the improvisation was rated in terms of the presence for each emotional state. The listening music therapists were able to describe more or less accurately the intended emotions. Problematic is the case, as it is in music psychological research that it is likely that no singular emotion is solely present but when someone is playing anger, sadness is also experienced in the music, which contributes to the variety in scores. This is one of the major topics, which should further future research: what are the interrelations between the variety of musical parameters related to the variety present in musical responding.

Coda

Far-reaching conclusions cannot be made about the relationship and the direction of the relationship between music and emotion. There are still too many methodological and theoretical problems to be able to clearly describe this relation, both in music psychology and music therapy. Ruud (2001) describes that one should think that the music therapist would be the first to know why music is so effective in music therapy. How difficult it is to describe this and to demonstrate this scientifically has been shown in this paper. As there are so many aspects still unknown, possibly a strong partner can assist in finding new answers: both for music psychology and for music therapy.

In this article some findings of music psychology and music therapy research studies were discussed which share the same topic: music and emotion. The paper is certainly full of blanks, as this is a topic that deserves a book (series:), but possibly it stimulates discussion to fill in the blanks.

Still, the fields of music psychology and music therapy are relatively young and it is only since recent dates that the relation between emotion and music is researched more thoroughly. Sloboda and Juslin (2001) state that after a period of neglect, because of the listed methodological problems, the topic of music and emotion is again at the forefront of music psychology. They illustrate this with the sudden increase in music and emotion papers at various conferences.

According to Sloboda (1985) it is important to integrate existing knowledge from a variety of disciplines to be able to further theoretical development. He also mentions aptly the tendency that researchers want to come up with a theory of their own. Mostly this complicates integrating findings. Luckily, more and more people build their ideas upon already existing knowledge. Ansdell (2001) sums some of the recent trends in musicology which he sees as of importance for music therapy such as the study of music as a process as well as a structure. The latest European Music Therapy conference in Naples in 2001 included musicology presentations that demonstrated the interrelations between disciplines. Possibly, on the next music therapy conferences also music psychologists will be present. It is important to focus on these interrelations, as current research is so often divided in separate specializations. Often a musicology approach is to focus on musical elements and to ignore psychological characteristics and vice versa the therapist often attaches less value to musicological aspects.

Currently, the impression on conferences and from the literature is that music therapy and music psychology are two entirely different fields, which they of course are. But in each relationship, it is their differences that attract and which should be elaborated more, such as the question how music and emotions are related. I hold that the two partners in this young relationship can benefit a great deal from their `living apart together' relationship.

References

Aldridge, David (1996). Music therapy research and practice: from out of the silence. London: Jessica Kingsley Publishers.

Ansdell, Gary (2001). Musicology: misunderstood guest at the music therapy feast? In Aldridge, G. Di Franco, E. Ruud, T. Wigram (Eds.). Music Therapy in Europe. Rome, Ismez/Onlus.

Bartel, Lee R. (1992). The development of the cognitive affective response test. Psychomusicology, 11, 15-26.

Benenzon, Rolando O. (1981). Music therapy manual. Springfield (Il): Charles C.Thomas.

Berlyne, Daniel Ellis (1971). Aesthetics and psychobiology. New York: Appleton Century-Crofts.

Bever, Thomas G. (1988). A cognitive theory of emotion and aesthetics in music. Psychomusicology, 7 (2),165-175.

Bonny, Helen (1986). Music and Healing. Music Therapy, 6A, I, 3-12.

Bunt, Leslie (1995). Muziektherapie. In: Evers, Jansma, Mak, De Vries (Eds). Muziekpsychologie. Assen: Van Gorcum.

Bunt, Leslie & Pavlicevic, Mercedes (2001). Music and emotion: perspectives from music therapy. In: Sloboda, J.A & Juslin, P.N. (Eds.) Music and Emotion: theory and research (pp. 181-205). New York: Oxford University Press.

Clark, David M. & Teasdale, John D. (1985). Constraints on the effects of mood on memory. Journal of Personality and Social Psychology, 48, 1595-1608.

Dainow, Elliot (1977). Physical effects and motor responses to music. Journal of Research in Music Education, 211-219.

Dowling, W. Jay & Harwood, Daniel L. (1986). Music and Cognition. Londen: Academic Press.

Droh, Roland & Spintge, Ralph (1987). Musik in der Medizin: Neurophysiologische Grundlagen, Klinische Applikationen, Geisteswissenschaftliche Einordnung. Berlin/Heidelberg: Springer-Verlag.

Erdonmez-Grocke, Denise (1999). A Phenomenological Study of Pivotal Moments in Guided Imagery and Music (GIM) Therapy. [online] Music Therapy World Available from: http://www.musictherapyworld.net/archive/SQL/index.php3 [Accessed November 2001].

Evers, Frank, Jansma, Mireille, Mak, Peter & Vries, Bart de (1995). Muziekpsychologie. Assen: Van Gorcum.

Field, T.M. & Walden, T. (1982a). Perception and production of facial expressions in infancy and early childhood. In H.W. Reese & L.P. Lipsitt (Eds.). Advances in child development and behavior (Vol. 16). New York: Academic Press.

Fischer (1981). Biological time (pp. 357-382). In J.T. Fraser (Ed.). The voices of time. Amherst: The University of Massachusetts Press.

Frijda, Nico (1988). De Emoties (The Emotions). Amsterdam: Uitgeverij Bert Bakker.

Gerdner, Linda A. (2000). Effects of individualized versus classical relaxation music on the frequency of agitation in elderly persons with AD disease and related disorders. Intern. Psychogeriatrics, 12, 1, 49-65.

Goldstein, Avram (1980). Thrills in response to music and other stimuli. Physiological Psychology, 8, 1, 126-129.

Gabrielson, A. & Juslin, P.N. (1996). Emotional expression in music performance: Between the performers intention and the listeners experience. Psychology of music, 24, 68-91.

Gabrielson & Lindstrom (2001). The influence of musical structure on emotional expression (pp. 223-249). In: Sloboda, J.A & Juslin, P.N. (Eds.). Music and Emotion: theory and research. New York: Oxford University Press.

Harrer, G. & Harrer H. (1977). Music, emotion and autonomic function. In M. Critchley and R.A. Henson (Eds.). Music and the Brain: Studies in the neurology of Music (pp. 202 -216). Londen: William Heinemann Medical Books Limited.

Henkin, Robert (1957). The prediction of behavior response patterns to music. Journal of Psychology, 44, 111-127.

Hevner, Kate (1935). The affective character of the major and minor modes in music. American Journal of Psychology, 47, 103-118.

Hevner, Kate (1936). Experimental studies of the elements of expression in music. American Journal of psychology, 48, 246-268.

Hevner, Kate (1937). The affective value of pitch and tempo in music. American Journal of Psychology, 49, 621-630.

Jansma, Mireille & Vries, Bart de (1995). Muziek en emotie (pp. 204-222). In: Evers, Jansma, Mak, De Vries (Eds). Muziekpsychologie. Assen: Van Gorcum

Kaufmann, William & Frisina, Robert (1992). Preface: the fusion of neuroscience and music. Psychomusicology, 11, 76-78.

Mandler, George (1984). Mind and body: Psychology of emotions and stress. New York: Norton.

Manstead, Anthony S.R (1993). Children's representation of emotions. In C. Pratt & A.F. Garton (Eds.). Systems of representation in children: Development and use (pp. 185-210). Chichester: John Wiley & Sons Ltd.

Meyer, Leonard B. (1956). Emotion and meaning in music. Chicago: Chicago Press.

Meyer, Leonard B. (2001). Music and Emotion: distinctions and uncertainties (pp. 341-361). In: Sloboda, J.A & Juslin, P.N. (Eds.). Music and Emotion: theory and research. New York: Oxford University Press. 

Ortony, Andrew, Clore, Gerald L. & Collins, Allen (1988). The cognitive structure of emotions. New York: Cambridge University Press.

Peretz, Isabelle (2001). Listen to the brain: a biological perspective on music and emotion (pp. 105-135). In: : Sloboda, J.A & Juslin, P.N. (Eds.) Music and Emotion: theory and research. New York: Oxford University Press.

Phillippot, Pierre (1992). Peripheral differentiation of emotion in bodily sensations,physiological changes, and social schemata. Unpublished doctoral dissertation,Catholic University of Louvain.

Ries, Harold A. (1969). GSR and breathing amplitude related to emotional reactions to music. Psychonomic Science, 14 (2), 62-64.

Rimé, B, Philippot P, Cisamolo D (1990). Social schemata of peripheral changes in emotion. Journal of Personality and Social Psychology, 59:1 38-49

Rueger, Christoph (1991). Die Musikalische Hausapotheke. Geneve/Kreuzlingen: Ariston Verlag.

Ruud, Even (2001). Musicology and musicotherapy. In: D. Aldridge, G. Di Franco, E. Ruud, T. Wigram (Eds.). Music Therapy in Europe. Rome, Ismez/Onlus.

Sloboda, John A. (1985). The Musical Mind: The cognitive psychology of music. Oxford: Oxford University Press.

Sloboda, John A. (1991). Music structure and emotional response: some empirical findings. Psychology of Music, 19, 110-120.

Sloboda, John A. (1992). Empirical studies of musical affect. In M. Riess-Jones & S. Hollerman (Eds.). Cognitive bases of musical communication (pp. 33-45). Washington: American Psychological Association.

Sloboda, John A. & Juslin, Patrik N. (2001). Psychological perspectives on music and emotion (pp. 71-105). In: Sloboda, J.A & Juslin, P.N. (Eds.) Music and Emotion: theory and research. New York: Oxford University Press.

Smeijsters, Henk (1995). Handboek Muziektherapie. Melos: Heerlen.

Smeijsters, Henk, Wijzenbeek, Gaby and Nieuwenhuijzen, Niek van (1995). The Effect of Musical Excerpts on the Evocation of Values for Depressed Patients. Journal of Music Therapy, XXXII (3), 167-188.

Smith, David S. (1989). Preferences for Differentiated Frequency Loudness Levels in Older Adult Music Listening. Journal of Music Therapy, XXVI (1), 18-29.

Stavenga, Anke (1979). Muziektherapie: Een Inventarisatie van theorieën en werkwijzen. Zeist: Dijkstra.

Terwogt, M. M., & Van Grinsven, F. (1991). Musical expression of moodstates. Psychology of Music, 19, 99-109.

Waterman, Mitch G. (1992). Emotion in Music: Towards a new methodology for the investigation of appreciation. International Journal of Psychology, 27, (3/4),189.

Waterman, Mitch G. (1996). Emotional response to music: Implicit and explicit effects in listeners and performers, Psychology of Music, 24, 53-67.

Weinberger, N.M. (1998) Understanding Music's Emotional Power. [online] MuSICA Research Notes, Volume V, Issue 2, Spring 1998. Available: http://www.musica.uci.edu/mrn/V5I2S98.html#understanding [Accessed November 2001].

Wing, Herbert D. (1968). Tests of musical ability and appreciation. London: Cambridge University Press.