Remembrance of Things Past: Music, Autobiographical Memory, and Emotion

ABSTRACT - It sometimes happens that a piece of music becomes associated with an event from a person's life so that hearing the piece of music evokes memories of the original experience. The purpose of this paper is to examine this phenomenon empirically. More specifically, the study investigates what kinds of autobiographical episodes are triggered by music that respondents self-selected as examples of this phenomenon, what characterizes these instances, and what is the relationship between the emotions descriptive of the original experience and the emotions aroused by hearing the piece of music.

Citation:

Hans Baumgartner (1992) ,"Remembrance of Things Past: Music, Autobiographical Memory, and Emotion", in NA - Advances in Consumer Research Volume 19, eds. John F. Sherry, Jr. and Brian Sternthal, Provo, UT : Association for Consumer Research, Pages: 613-620.

Advances in Consumer Research Volume 19, 1992      Pages 613-620

REMEMBRANCE OF THINGS PAST: MUSIC, AUTOBIOGRAPHICAL MEMORY, AND EMOTION

Hans Baumgartner, Pennsylvania State University

ABSTRACT -

It sometimes happens that a piece of music becomes associated with an event from a person's life so that hearing the piece of music evokes memories of the original experience. The purpose of this paper is to examine this phenomenon empirically. More specifically, the study investigates what kinds of autobiographical episodes are triggered by music that respondents self-selected as examples of this phenomenon, what characterizes these instances, and what is the relationship between the emotions descriptive of the original experience and the emotions aroused by hearing the piece of music.

And before Swann had had time to understand what was happening, to think: "It is the little phrase from Vinteuil's sonata. I mustn't listen!", all his memories of the days when Odette had been in love with him, which he had succeeded, up till that evening, in keeping invisible in the depths of his being, deceived by this sudden reflection of a season of love, whose sun, they supposed, had dawned again, had awakened from their slumber, had taken wing and risen to sing maddeningly in his ears, without pity for his present desolation, the forgotten strains of happiness.

Marcel Proust (1913/1956), Swann's Way, p. 496

INTRODUCTION

Music is sometimes called the language of emotion, but it is not at all clear how music expresses, or is expressive of, the emotions. In order to elucidate the issues involved and to set the stage for the main focus of this paper, it might be helpful to briefly discuss two distinctions.

The first distinction is between what Pratt (1931) calls autonomous and heteronomous music. Adherents of the view that music is (or should be) autonomous argue that the "essence of music lies wholly within the borders of sound" (Pratt 1931, p. 205) and that the character of a piece of music is due entirely to intrinsic properties of the music itself. Those who regard music as heteronomous, on the other hand, argue that music possesses "content which is non-tonal" (Pratt 1931, p. 212) and that the character of a piece of music is often a function of extramusical associations. Drawing on the work of Charles Peirce, Dowling and Harwood (1986, Chapter 8), in their discussion of the relationship between music and emotion, make a similar distinction between iconic and indexical representation in music, where iconic representation (of emotions) refers to "patterns within the music itself ... of tensions and relaxations ... [which] mirror the form of emotional tensions and relaxations" (p. 205), whereas indexical representation refers to "direct association of a musical event with some extramusical object or event, so that emotions previously associated with the extramusical object come to be associated with the music" (p. 204).

The second distinction is between what Kivy (1989, Chapter XIII) calls musical emotivism and musical cognitivism. Proponents of musical emotivism argue that music arouses (or induces) emotions (e.g., sad music actually makes you sad), whereas adherents of musical cognitivism (such as Kivy) suggest that music merely embodies (or represents) emotions (e.g., sadness in music can be recognized because sad music has characteristic qualities). While the distinction between musical emotivism and musical cognitivism centers on whether music actually induces or merely represents emotions, the difference between autonomous and heteronomous music refers to whether music per se or extramusical objects or events associated with the music induce or represent emotions.

It is possible to arrange the concepts of autonomous vs. heteronomous music and musical emotivism vs. musical cognitivism in a two by two table and to distinguish four cases of expressiveness in music. The topic of interest in the present paper is heteronomous music which arouses emotions, or the indexical induction of emotions (cf. Dowling and Harwood 1986, p. 204). The excerpt from Swann's Way by Marcel Proust quoted at the beginning is a well-known literary example of this phenomenon. Most readers, however, will probably have had a similar experience in which a piece of music conjured up images of some past autobiographical episode.

Writers on musical aesthetics, even those in the cognitivist camp, generally agree that a piece of music may acquire the capacity to arouse emotional reactions through association with significant personal experiences, but the expressiveness of music, it is argued, is then due to an accident of association and has (or may have) nothing to do with the "real expressive character of the music" (Kivy 1989, p. 157). For example, a happy piece of music could come to arouse sadness because of its association with a sad event. Furthermore, if expressiveness in music is a consequence of its indexical function, music is stripped of any special status as the language of emotion because any other object (a photo, a gift, even a smell) could also become associated with a significant personal experience and thus induce emotional reactions.

Although it is quite understandable why aestheticians would not get too excited about instances in which idiosyncratic associations endow music with emotive meaning, it nonetheless seems an interesting question to investigate this phenomenon empirically. The study reported below was designed as an initial examination of some of the issues involved. Specifically, the objectives of the study were (1) to assess how readily people can bring to mind personal experiences that have become associated with a piece of music, (2) to explore what characterizes these autobiographical episodes, the music that triggers them, and the instances in which the phenomenon occurs, and (3) to examine the relationship between the emotions descriptive of the original experience and the emotions aroused by hearing the piece of music.

METHOD

Subjects

Seventy-three undergraduate marketing majors (43 males, 30 females) participated in the study as part of a course requirement. The data were collected in one 45-minute session in which other experimental tasks unrelated to the present study were also administered. Participants were run in groups of at most 8 people in 12 experimental sessions.

Procedure and Measures

Subjects received a questionnaire and were told to complete it at their own pace. On top of the first page they were given the following instructions:

"Sometimes a piece of music becomes associated with an autobiographical episode from a person's life so that every time the person hears the music, (s)he is reminded of the original experience. Please take a minute and think about instances in which this has happened to you. Select the instance that is most salient in your mind and briefly describe the piece of music and the personal experience that the music reminds you of every time you hear it."

After subjects had answered this question, they were asked to rate their feelings toward both the piece of music and the personal experience on two seven-point scales (pleasant-unpleasant and favorable-unfavorable). The coefficient alphas of the two sets of measures were 0.93 and 0.99, respectively.

Subjects were then asked approximately how long ago the personal experience recalled had originally occurred, how vivid the recollections of the personal experience brought about by the piece of music were (seven-point scale from 'not at all vivid' to 'extremely vivid'), how emotional the recollections were when they occurred (seven-point scale from 'not at all emotional' to 'extremely emotional'), to what extent they relived the original experience when they heard the piece of music (seven-point scale from 'not at all' to 'a great extent') and to what extent the recollections were accompanied by imagery (visual images, sounds, smells, etc.) descriptive of the original experience (seven-point scale from 'not at all' to 'a great extent').

Finally, an attempt was made to capture in more detail the affective qualities characteristic of the original personal experience and the emotional tone descriptive of the piece of music associated with it. The question, however, was which instrument to use for the measurement of emotions (cf. Plutchik and Kellerman 1989).

In recent years, mainly through the work of Plutchik (1980, 1989) and Russell (1980, 1989), circumplex models of affect have become popular. In Russell's version of the model, 28 affect terms are arranged in a circular order, and their relative positions around the circle are described with reference to two underlying bipolar dimensions labeled pleasure-displeasure and degree of arousal (see also Mehrabian and Russell 1974).

It is interesting to note that as far back as the 1930s, Hevner (1935, 1936) proposed what is essentially a circumplex model of emotional epithets to describe music. In Hevner's scheme, 66 adjectives are classified into 8 groups arranged in a circular order. Even more interesting, the two dimensions underlying Hevner's circumplex structure correspond remarkably well to Russell's dimensions of pleasure-displeasure and degree of arousal.

Since one of the goals of this research was to investigate how well the affective qualities descriptive of the personal experience corresponded to the emotional tone ascribed to the piece of music, it was decided to measure the two sets of emotions with the Russell (1980) and Hevner (1936) instruments. Russell's 28 affect terms, arranged in circular order from high arousal/pleasure to high arousal/ displeasure, are: aroused, astonished, excited, delighted, happy, pleased, glad, serene, content, at ease, satisfied, relaxed, calm, sleepy, tired, droopy, bored, depressed, gloomy, sad, miserable, frustrated, distressed, annoyed, afraid, angry, alarmed, and tense. Hevner's list of 66 adjectives was shortened to 24 by selecting three items from each of her 8 groups. The eight groups of three items, arranged in the same order as the Russell terms, are: majestic, vigorous, exalted; passionate, triumphant, exhilarated; merry, joyous, cheerful; light, graceful, playful; quiet, tranquil, soothing; sentimental, dreamy, longing; mournful, tragic, doleful; and serious, dignified, solemn.

To assess the emotional qualities of the personal experience recalled by a participant, subjects were asked, "Now we would like for you to tell us which of the following feelings are characteristic of the personal experience that you described on the previous page. Please tell us how much you felt each of the feelings listed below when the episode that the piece of music reminds you of first occurred?" Subjects responded on nine-point scales with endpoints of 'did not feel this feelings at all' and 'felt this feeling very strongly.' Similarly, to assess the emotional tone of a piece of music, subjects were asked, "The last thing we would like to know is how hearing the piece of music that you named on the first page makes you feel. Please tell us how much you feel each of the following feelings when you hear the piece of music?" Again, subjects responded on nine-point scales with endpoints of 'do not feel this feeling at all' and 'feel this feeling very strongly.'

The emotional descriptors selected from the Russell and Hevner instruments to measure the affective qualities of personal experiences and the emotional tone of pieces of music do not overlap and, for purposes of data collection, were listed in random order, but they supposedly tap the same two-dimensional structure of pleasure-displeasure and degree of arousal. The fact that the two sets of affect terms are distinct should reduce the problems associated with collecting all ratings at the same time, while the similarities in the underlying structure should make it possible to test whether the emotional tone ascribed to a piece of music is a function of the affective characteristics of the personal experience with which the music is associated.

TABLE 1

CLASSIFICATION OF PERSONAL EXPERIENCES

RESULTS

Subjects apparently had no difficulty bringing to mind an instance in which an autobiographical episode had become associated with a piece of music. Only three of the 73 participants (2 males, 1 female) did not report a piece of music and an associated personal experience, although they evidently had thought of one because they completed the rest of the questionnaire. However, these subjects were excluded from further analysis.

In what follows, the written responses to the open-ended question will be analyzed first, then subjects' descriptions of the personal experience and the piece of music which were collected with the standardized questions are discussed, and finally the self-report measures of emotion are examined.

Analysis of Verbal Responses

One question that can be addressed by looking at subjects' verbal protocols is what kinds of personal experiences become associated with pieces of music. Table 1 shows that the majority of episodes reported by subjects (64 percent) relate to past or current romantic involvements and times spent with friends. Past and current romantic involvements refer to recollections of times spent with a past or current boyfriend or girlfriend (e.g., memories of the first date, sexual experiences, etc.). Times spent with friends is similar to romantic involvements in that another person or other people are at the center of the recollection, but no romantic relationship is implied (e.g., parties, concerts, special events, and segments of one's life in which friends played a significant role, etc.). Vacations refers to recollections of past holidays spent with family or friends. Finally, the category of other experiences includes episodes which did not fit into any of the previous groupings and which were not mentioned frequently enough to warrant a separate classification (e.g., movies or concerts with no mentions of friends, a victory in a boxing match, the death of one's grandfather, childhood memories, etc.).

Table 1 also indicates that most personal experiences recalled by subjects (84 percent) were of positive valence. All recollections in the categories of current romantic involvements, times spent with friends, and vacations were positive, and only 6 out of 19 episodes in the other category were negative. Interestingly, memories of past romantic involvements were just as likely to be negative as positive.

There was a great deal of variation with respect to the pieces of music that subjects had memories attached to. In 15 of the 70 cases the piece of music or the composer/interpreter were not named, although people generally mentioned what kind of music it was. With the exception of Gustav Holst's The Planets, Russian Christmas Music, and the National Anthem, all named musical selections were of the pop/rock/folk variety. Only one song was mentioned twice, and subjects mentioned 47 different singers or groups.

Phenomenological characteristics of indexical association

Table 1 indicates that 59 of the 70 personal experiences recalled were evaluated positively and 11 were evaluated negatively. Not a single person provided a neutral rating. More importantly, most recollections were strongly affectively charged. When the extremity of evaluations (i.e., the absolute value of the deviation from the midpoint) was analyzed, 49 of the subjects (70 percent) scored the maximum value of 6 (44 positive deviations and 5 negative deviations of 6 from the midpoint).

The picture is similar for the evaluation of music that triggers personal memories. Most ratings cluster around the endpoints of the scale, although only two musical selections were rated below the midpoint of the scale, and 41 subjects (59 percent) attain the maximum deviation from the midpoint.

The median time interval that had elapsed since the personal experience first occurred was 2 years, but for some people it was as short as a week and for others as long as 14 years. Subjects also indicated that the recollection of a personal experience brought about by the piece of music was rather vivid and emotional (means of 6.0 and 5.6, respectively, on seven-point scales). Furthermore, subjects tended to relive the original personal experience when they heard the piece of music (mean of 5.2) and reported that the recollection was often accompanied by imagery descriptive of the original experience (mean of 5.8).

TABLE 2

CORRELATIONS AMONG PHENOMENOLOGICAL CHARACTERISTICS

Correlations among the various measures discussed in the foregoing paragraphs are presented in Table 2. Subjects' evaluations of the piece of music and of the personal experience associated with it are strongly correlated at 0.68, which lends some preliminary support to the notion that feelings aroused by listening to music may sometimes derive from feelings associated with an autobiographical episode that the piece of music evokes. Table 2 also indicates that the more vivid and emotional the recollections triggered by a piece of music are and the greater the extent to which people relive the original experience, the more positive the evaluation of the piece of music will be.

Analysis of Experiential Emotions

Correlations were computed among Russell's 28 affect terms on which subjects rated the affective qualities of the personal experience they had recalled, and the resulting correlation matrix was subjected to a nonmetric multidimensional scaling analysis using the program KYST (Kruskal and Wish 1978). The stress values for one-, two- and three-dimensional solutions were 0.16, 0.11, and 0.06, respectively; the two-dimensional structure yielded the most interpretable solution, however, and it is depicted in Figure 1.

It can be seen that the 28 affect terms fall roughly into a circular order. With the exceptions of tired and bored, whose positions do not fit the hypothesized pattern, the structure proposed by Russell (1980) is replicated quite well. The horizontal axis reflects pleasure-displeasure and the vertical axis indicates degree of arousal.

Analysis of Musical Emotions

Correlations were also computed among the 24 emotional descriptors of music, and the resulting correlation matrix was subjected to a nonmetric multidimensional scaling analysis. The stress values in one, two, and three dimensions were 0.17, 0.11, and 0.06, respectively, and the two-dimensional structure yielded the most interpretable solution. The results are shown in Figure 2.

The hypothesized circular order is not obtained as clearly for the musical emotions as it was for the experiential emotions, but the arrangement of the 24 musical adjectives corresponds roughly to prior expectations. The horizontal axis again reflects pleasure-displeasure and the vertical axis indicates degree of arousal.

Correspondence between Experiential and Musical Emotions

In order to assess the degree of correspondence between the affective qualities descriptive of a personal experience and the emotional tone ascribed to a piece of music, it was necessary to reduce the number of experiential and musical emotions to a more manageable number. Principal components analysis was used to accomplish this.

For the experiential emotions, five factors had eigenvalues greater than one and together accounted for 72 percent of the total variance. After performing an oblique rotation (PROMAX), the five factors were defined by the following items: EXPEMOT-I (aroused, excited, delighted, happy, pleased, glad, satisfied); EXPEMOT-II (serene, content, at ease, relaxed, calm); EXPEMOT-III (sleepy, droopy); EXPEMOT-IV (tired, bored); and EXPEMOT-V (astonished, gloomy, miserable, frustrated, distressed, annoyed, afraid, angry, tense, alarmed). The items defining a given factor were summed up and the reliabilities of the five composites are shown in the appropriate diagonal of Table 3. The items 'sad' and 'depressed' were not included in any of the composites because they had large negative loadings on the first factor and did not load very highly on any of the other factors.

For the musical emotions, four factors had eigenvalues greater than one and together accounted for 71 percent of the total variance. After performing an oblique rotation, the four factors were: MUSEMOT-I (vigorous, majestic, triumphant, exhilarating, graceful, dignified); MUSEMOT-II (exalted, merry, joyous, cheerful, light, playful, longing); MUSEMOT-III (passionate, tranquil, soothing, sentimental, dreamy); and MUSEMOT-IV (quiet, mournful, tragic, doleful, serious, solemn). The items defining each factor were again summed up and the reliabilities of each composite are shown as the diagonal entries in Table 3. It is apparent that the five experiential and the four musical emotion factors correspond closely to the way affect terms cluster together in the multidimensional scaling solutions.

FIGURE 1

MULTIDIMENSIONAL SCALING SOLUTION FOR 28 EXPERIENTIAL EMOTIONS

FIGURE 2

MULTIDIMENSIONAL SCALING SOLUTION FOR 24 MUSICAL EMOTIONS

TABLE 3

CORRELATIONS AMONG THE FIVE EXPERIENTIAL AND THE FOUR MUSICAL EMOTION FACTORS

Besides the correlations among the five experiential emotion factors and the four musical emotion factors, Table 3 also exhibits the cross-correlations among the two sets of factors. It is apparent that the emotional tone ascribed to a piece of music follows the affective characterization of the personal experience that the music evokes. If the autobiographical episode was described as exciting and happy (using the items with the highest loadings to define a factor, in this case EXPEMOT-I), subjects tended to characterize the music primarily as merry and longing (MUSEMOT-II), but also as majestic and triumphant (MUSEMOT-I) and tranquil and soothing (MUSEMOT-III). If the original experience was perceived as serene and calm (EXPEMOT-II), the piece of music was described predominantly as tranquil and soothing, but also as merry and longing and majestic and triumphant. Finally, if the affective qualities of the personal experience were called sleepy and droopy (EXPEMOT-III) or alarmed, astonished, and distressed (EXPEMOT-V), the music that triggered such recollections was characterized as mournful and doleful (MUSEMOT-IV).

Although the terms 'sad' and 'depressed' did not define a separate factor or load highly on EXPEMOT-IV, the two items were correlated 0.74 and had a correlation of 0.55 with EXPEMOT-IV. Furthermore, the composite of sad and depressed was correlated 0.68 with MUSEMOT-IV.

DISCUSSION

The results of this study suggest that the subjects who participated in this research had experienced a situation in which a piece of music had become associated with an event from their lives so that hearing the piece of music evoked memories of the original episode. Most of the instances of this phenomenon reported by undergraduate marketing majors involved relationships with past or present lovers or experiences with family and friends, and although some people described unpleasant episodes, overall there was a significant bias toward remembering happy events. Most personal experiences for which the phenomenon occurred were strongly affectively charged, and the recollections triggered by the music were described as vivid and emotional and as involving a reliving of, and being accompanied by imagery descriptive of, the original episode.

The results of the study also indicate that the emotional tone ascribed to a piece of music is a function of the affective characteristics of the personal experience with which the music is associated. First, there was a fairly strong correlation between a person's evaluation of the piece of music and his or her evaluation of the autobiographical episode. Second, there was more specific evidence that people's ratings of the affective characteristics of the personal experience corresponded to the feelings induced by hearing the piece of music.

How does the research reported in this paper relate to similar work in consumer behavior? There has recently been a flurry of research on the effects of music on consumer behavior (see the review by Bruner 1990 and the more recent papers by Alpert and Alpert 1991, Kellaris and Kent 1991, Scott 1990, Wallace 1991, and Yalch 1991). These papers deal with a variety of important issues relating to the use of music in marketing (music as an affective background in ads, music as a mnemonic device, music as a communicator of meaning, etc.), but they are not particularly pertinent to this study.

Two other streams of research are more relevant to the present concerns. One is the work by Baumgartner, Sujan, and Bettman (in press) on the role of consumers' autobiographical memories in information processing (see also Thorson and Friestad 1989). In a series of studies, Baumgartner et al. (in press) show that consumers' autobiographical memories involving products and product usage experiences are affectively charged and that the retrieval of these memories influences ad and brand evaluations. The present study indicates that some pieces of music may be rather effective in cuing autobiographical memories and that the recollections triggered by music are strongly, and generally positively, affectively charged. This suggests that music in advertising may impact ad and brand evaluations through the retrieval of autobiographical memories. Although there was a lot of diversity with respect to the pieces of music for which subjects reported recollections of autobiographical events, the present study does not allow drawing conclusions about how frequently associations with popular pieces of music are formed. A fruitful avenue for future research would thus be to investigate whether the inclusion of well-known tunes in ads triggers memories of past personal experiences which arouse emotions beyond what would be predicted from the mere popularity of a piece of music.

A second area of research that deals with issues similar to the present concerns is the work of Holbrook and Schindler (1991) and Havlena and Holak (1991) on nostalgia. Nostalgia is often defined as a "yearning for the past" and refers to the "bittersweet" memories people have of past personal experiences (frequently from the period of adolescence and early adulthood). As noted by Havlena and Holak (1991), there has recently been a boom in the use of nostalgia in marketing, and advertisers often try to evoke nostalgic feelings through music, jingles, and other stimuli. In a study dealing with the development of musical tastes, Holbrook and Schindler (1989) found that subjects expressed the strongest liking for music that was popular when they were in their early twenties. The authors argue that one reason for this peak in preference in early adulthood might be that people associate pieces of music from this time with emotionally powerful events such as fraternity parties, school dances and other social gatherings. Many of the associations reported by subjects in the present study may be classified as nostalgic recollections of past events, and the results show that most memories are "sweet" rather than "bitter." Furthermore, the participants in the study were all juniors or seniors and thus in their early twenties, and most of them (78 percent) described pieces of music that had been popular, and experiences that had occurred, during the previous four years. Given that the music tended to take on the emotional tone of the personal experience with which it was associated and given that most episodes recalled were happy events, the present findings suggest that preference for popular music, especially when tracked over time, may not be a function of the music per se, but a consequence of the personal associations people have with the music.

The research reported in this paper only scratches the surface of many of the issues raised. Furthermore, several limitations of the study should be kept in mind. First, the findings that emerged are based on data from a relatively small sample of undergraduate marketing majors, they are correlational in nature, and they may not generalize to other segments of the population. Second, instances of the phenomenon of indexical induction of emotions through music were self-selected, and the ratings of both the affective qualities of the personal experience triggered by music and the emotional tone of the piece of music associated with the personal experience were collected at the same point in time, which could have increased the correspondence between the experiential and musical emotions. One way of overcoming this deficiency would be to obtain ratings of the emotional tone of the piece of music at a later point in time and to mask the connection to the earlier study by gathering data on several pieces of music. Third, the structure underlying Hevner's (1936) shortened instrument for the measurement of musical emotions was not very clear-cut and more research seems necessary to refine the list of adjectives. Furthermore, since the Russell (1980) and Hevner (1936) instruments probably do not measure the emotions associated with a personal experience and a piece of music equally well, the correlations reported in Table 3 should be interpreted with care. If the research were conducted in two waves, the same scale could be used for the assessment of the affective quality of a personal experience and the emotional tone of the piece of music and this problem could be circumvented. The present findings are preliminary, but the results thus far look encouraging and the phenomenon of indexical induction of emotions through music is interesting so that further investigation of the topic seems warranted.

REFERENCES

Alpert, Judy I. and Mark I. Alpert (1991), "Contributions from a Musical Perspective on Advertising and Consumer Behavior," in Advances in Consumer Research, Vol. 18, eds. Rebecca H. Holman and Michael R. Solomon, Provo, UT: Association for Consumer Research, 232-238.

Baumgartner, Hans, Mita Sujan, and James R. Bettman (in press), "Autobiographical Memories, Affect, and Consumer Information Processing," Journal of Consumer Psychology, 1.

Bruner, Gordon C. (1990), "Music, Mood, and Marketing," Journal of Marketing, 54 (October), 94-104.

Dowling, W. Jay and Dane L. Harwood (1986), Music Cognition, Orlando, FL: Academic Press.

Havlena, William J. and Susan L. Holak (1991), "'The Good Old Days': Observations on Nostalgia and its Role in Consumer Behavior," in Advances in Consumer Research, Vol. 18, eds. Rebecca H. Holman and Michael R. Solomon, Provo, UT: Association for Consumer Research, 323-329.

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, 246-269.

Holbrook, Morris B. and Robert M. Schindler (1989), "Some Exploratory Findings on the Development of Musical Tastes," Journal of Consumer Research, 16 (June), 119-124.

Holbrook, Morris B. and Robert M. Schindler (1991), "Echoes of the Dear Departed Past: Some Work in Progress on Nostalgia," in Advances in Consumer Research, Vol. 18, eds. Rebecca H. Holman and Michael R. Solomon, Provo, UT: Association for Consumer Research, 330-333.

Kellaris, James J. and Robert J. Kent (1991), "Exploring Tempo and Modality Effects on Consumer Responses to Music," in Advances in Consumer Research, Vol. 18, eds. Rebecca H. Holman and Michael R. Solomon, Provo, UT: Association for Consumer Research, 243-248.

Kivy, Peter (1989), Sound Sentiment: An Essay on the Musical Emotions, Philadelphia, PA: Temple University Press.

Kruskal, Joseph B. and Myron Wish (1978), Multidimensional Scaling, Sage University Paper series on Quantitative Applications in the Social Sciences, 07-011, Beverly Hills and London: Sage Publications.

Mehrabian, Albert and James A. Russell (1974), An Approach to Environmental Psychology, Cambridge, MA: MIT Press.

Plutchik, Robert (1980), Emotion: A Psychoevolutionary Synthesis, New York, NY: Harper.

Plutchik, Robert (1989), "Measuring Emotions and their Derivatives," in Emotion: Theory, Research, and Experience, Vol. 4, The Measurement of Emotions, eds. Robert Plutchik and Henry Kellerman, San Diego, CA: Academic Press, 1-35.

Plutchik, Robert and Henry Kellerman (1989), Emotion: Theory, Research, and Experience, Vol. 4, The Measurement of Emotions, San Diego, CA: Academic Press.

Pratt, Carroll C. (1931), The Meaning of Music, New York, NY: McGraw-Hill.

Proust, Marcel (1913/1956), Swann's Way, New York, NY: Modern Library.

Russell, James A. (1980), "A Circumplex Model of Affect," Journal of Personality and Social Psychology, 36, 1152-1168.

Russell, James A. (1989), "Measures of Emotion," in Emotion: Theory, Research, and Experience, Vol. 4, The Measurement of Emotions, eds. Robert Plutchik and Henry Kellerman, San Diego, CA: Academic Press, 83-111.

Scott, Linda M. (1990), "Understanding Jingles and Needledrop: A Rhetorical Approach to Music in Advertising," Journal of Consumer Research, 17 (September), 223-236.

Thorson, Esther and Marian Friestad (1989), "The Effects of Emotion on Episodic Memory for Television Commercials," in Cognitive and Affective Responses to Advertising, eds. Patricia Cafferata and Alice M. Tybout, Lexington, MA: D.C. Heath, 305-325.

Wallace, Wanda T., "Jingles in Advertisements: Can they Improve Recall," in Advances in Consumer Research, Vol. 18, eds. Rebecca H. Holman and Michael R. Solomon, Provo, UT: Association for Consumer Research, 239-242.

Yalch, Richard F. (1991), "Memory in a Jingle Jungle: Music as a Mnemonic Device in Communicating Advertising Slogans," Journal of Applied Psychology, 76 (April), 268-275.

----------------------------------------