The scent of literature

COGNITION AND EMOTION 2005, 19 (1), 101±119

The scent of literature Gerald C. Cupchik University of Toronto, Canada

Krista Phillips York University, Toronto, Canada This study examined the effects of hedonic odour context on cognitive and affective processes involved in reading literary excerpts. A total of 32 undergraduate subjects smelled pleasant and unpleasant odours, while reading eight passages that touched upon positive or negative subject matter and were either engaged or detached in style, in a fully counterbalanced within-subjects design. After each reading episode, subjects rated passage and odour on fourteen 7-point scales. In accordance with a hedonic congruence hypothesis, odours should affect the reading process when there was a fit between the hedonic tone of the textual properties (subject matter or style) and the odours (pleasant/unpleasant). Results supported this hypothesis showing that positive subject matter and engaged style combined with pleasant odour to bring the story to life and evoke the most images and feelings. Passages with negative subject matter were not readily affected by the hedonic odour context. These findings replicated earlier research involving paintings viewed in a hedonic odour context. Gender differences also revealed that males, compared with females, found episodes read in the context of unpleasant odours to be more involving and personally meaningful, and they could more readily identify with the characters.

The affective context within which an aesthetic work is encountered can have a strong effect on a person's experience. This hypothesis was proposed more than a century ago by Theodor Lipps (1900) who originated the notion of EinfuÈhlung or empathy. In particular, he claimed that the aesthetic evaluation of a painting Correspondence should be addressed to Dr. Gerald C. Cupchik, University of Toronto at Scarborough, Life Sciences Division, 1265 Military Trail, Scarborough, ON, Canada M1C 1A4. The authors would like to thank Dr Stephen Warrenburg of the International Flavors and Fragrances Corporation for providing all but one of the odours and offering continuing support for our project, both material and conceptual. The authors would also like to thank Dr Rachel Herz for obtaining synthetic sweat and Dr George Preti of the Monell Smell Institute who provided it, and Tonya Stokes-Hendriks who analysed the data. Dr Gary Leonard (Associate Professor of English Literature) and Ingrid Braun served as expert readers who formalised the engaged versus detached stylistic distinction. # 2005 Psychology Press Ltd http://www.tandf.co.uk/journals/pp/02699931.html DOI:10.1080/02699930441000139

102

CUPCHIK AND PHILLIPS

would be magnified when the viewer's mood matched the perceived mood in the work. While Hoege's (1984) attempt to test this hypothesis using the Velten verbal mood induction technique was not successful, Herz and Cupchik (1993) used odour to manipulate mood in order to modify the judgements of paintings because previous research (Rotton, 1983) had shown that an ambient unpleasant odour could diminish the evaluation of paintings. Results showed that a pleasant odorous context enhanced ratings of artistic quality for paintings with positive themes, whereas an unpleasant odour significantly diminished judgements of complexity and personal meaningfulness. Paintings with negative themes, which were strongly anchored and judged to be highly complex and personally meaningful, were resistant to the effects of odour context. By way of extension across aesthetic media, the experiment reported here examined the potential effects of odour on the reading experience and judgements of literary works. Of course odour has served as a prominent theme in literature per se. For example, Proust's (1922/1960) description that a madeleine biscuit soaked in linden tea made a ``shudder run through my whole body'' (p. 58) offered anecdotal evidence for odour-evoked autobiographical memories. Suskind's (1986/2001) novel, Perfume: The Story of a Murderer, made scent a dominating theme of the main character's obsessive life. But using odour as a hedonic context for the reading of literature has not been attempted and it can be hypothesised that the same principles would apply as in the case of viewing paintings. Literary works, like paintings, embody a unique combination of subject matter and style that make them different from prosaic discourse. In terms of modern reception theory, they are multilayered (Kreitler & Kreitler, 1972) and open-ended (Eco, 1962/1989). This stimulates an `èffort after meaning'' (Bartlett, 1932) which is ``the product of an interaction between the textual signals and the reader's acts of comprehension'' (Iser, 1978, p. 9). While some textual signals emphasise what protagonists think, know, or feel in the ``landscape of consciousness'', others focus on story grammar describing agents, goals, and situations in the ``landscape of action'' (Bruner, 1986). Either way, the `èffort after meaning'' yields a situational model (Graesser, Singer, & Trabasso, 1994) of the textual landscape that is coherent (Iser, 1978), placing experiential states and actions in meaningful contexts. This is demonstrated when readers, who are absorbed in personally meaningful passages, slow the pace of reading (Cupchik & Laszlo, 1994) and can sense the physical space in a story while visualizing the objects described (Braun & Cupchik, 2001). Thus, both cognitive and affective processes are affected, ranging from the experience of sensations, images, and meanings to general feelings of pleasure or suspense and deeper emotions elicited by situated characters and personally meaningful episodes (Cupchik, 1995, 1996). A theoretical framework is needed to account for the effects on literary experience of odour as a hedonic context. In relation to cross-modal interaction,

SCENT OF LITERATURE

103

odours have been shown to have a more potent effect on words than vice versa. Odours become unified with words or images through a passive encoding process (Chu & Downes, 2000) that yields `à unitary, holistic perceptual event'' (Engen, 1991, p. 7). Fit or congruence appear to be critical qualities in the establishment of a unity between odours and perceptual or cognitive experiences. Wolpin and Weinstein (1983) have reported that images were clearer and more vivid when subjects smelled an odour that was congruent with the object being imagined. It has been argued that subjects spontaneously connect an odour with a visual source with which it fits only if it was below the threshold of awareness and they were unaware beforehand of its identity (Degel & KoÈster, 1999; Degel, Piper, & KoÈster, 2001). Integration of odours with naturalistic episodes is therefore established slowly and spontaneously through implicit, perceptual, and affective links. The fit of an odour with an event, either real or imagined, increases the coherence of an experience by integrating a sensory dimension (i.e., odour) with visual, auditory, verbal, and tactile cues. This process of implicit learning provides a foundation for odour-evoked emotional memories (Herz & Cupchik, 1992, 1995). A recent study has suggested that odours are associated with distinct cognitive and affective processes which could have separate effects on the reading experience (Cupchik, Phillips, & Truong, 2005 this issue). On the one hand, odours can be experienced as pleasantly soothing or energising, providing an affective context within which to experience a story. On the other hand, odours also evoked familiar sensations, images, memories, and generalised feelings which could serve as a cognitive scaffolding for the meaningful experience of a story. The primary goal of this study was to determine whether the relative fit or hedonic congruence between odours and narrative properties, both thematic and stylistic, would affect cognitive and/or affective responses. In line with the earlier research involving paintings and odours (Herz & Cupchik, 1993), it was expected that congruity effects would be stronger for the positive subject matter and pleasant odours. This would also be consistent with Ehrlichman and Halpern's (1988) finding that pleasant odours favoured the stimulation of memories which in this context, might enhance the reading experience. Further, an engaged narrative style emphasising the protagonist's perspective, compared with a detached narrative style, should make it easier for readers to become absorbed in the literary experience (Braun & Cupchik, 2001) and thereby have a positive effect congruent with that of the positive subject matter and pleasant odour. A secondary goal was to ascertain whether odours could have their effects on aesthetic experience even if they were above threshold and readers were explicitly aware of their presence. In this study, participants read short story passages while smelling odours and subsequently rated their reading experiences and the stimuli on verbal scales. Both subject matter and style of the excerpts from literary passages were systematically varied. The subject matter of each passage was either

104


positive or negative and the narrative style was either engaged or detached. The odours were not programmatically matched to the narrative episodes (e.g., a story reference to roses combined with the scent of roses) but were hedonically varied as pleasant or unpleasant and systematically counterbalanced against the stimulus materials across participants. Thus, the hedonic quality of the odour and not its referential value was critical in this study. It was hypothesised, based on the earlier study of paintings viewed in hedonic context (Herz & Cupchik, 1993), that hedonic congruence between the short story and odour properties would enhance aesthetic experience, whereas incongruity would diminish it.

METHOD Materials Odours. Six pleasant and six unpleasant odours were selected for use in the study based on pretesting for an earlier study (Cupchik et al., 2005 this issue). These odours were more or less familiar and represented a contrast comparable to freshness (clementine, rose, fresh cut grass, agarbatti incense, coconut, and relaxing pine) and decay (rotting leaves, mildew, synthetic sweat, smoke) or medicinal/preventative (mothballs, clean fresh pine or Vick vapour rub) odours. Eleven of the odours were provided by IFF (International Flavors and Fragrances Corporation) and one by the Monell Chemical Senses Center. The odours were diluted in diethyl phthalate polyethylene and absorbed into pellets, and were preserved in small opaque amber-coloured jars. They were transferred individually by tweezers into white squeeze bottles with flip nozzles and were replaced after losing their potency (generally after six subjects). Letter-number codes were glued to the bottoms of the bottles so that the experimenter could identify the odour stimulus. Literary passages. Excerpts from eight literary passages (mean number of words per excerpt = 107.25) by James Joyce, Margaret Atwood, and Alistair MacLeod were chosen based on ratings from earlier studies of literary text reception (Cupchik & Laszlo, 1994; Cupchik, Leonard, Axelrad, & Kalin, 1998). They represented the factorial combination of Subject Matter (positive and negative) and Style (engaged and detached) with one replication in each cell (see Table 1). Subject Matter was operationally defined in terms of theme and contrasted socially positive events with ones expressing negative emotions. The Style variable involved a contrast between two different ways of structuring the texts according to the assessments of two expert judges. The engaged style described a character's perspective on actions, setting, and events along with accompanying emotions and would have a positive effect making it easier for readers to empathise and find meaning in the unfolding episode. The detached style offered stark description without a strong sense of a controlling

TABLE 1 Passages representing the factorial combination of positive and negative Subject Matter and engaged vs. detached Style Negative subject matter, detached style (a) Mr. Duffy abhorred anything which betokened physical or mental disorder. A medieval doctor would have called him saturnine. His face, which carried the entire tale of his years, was of the brown tint of Dublin streets. On his long and rather large head grew dry black hair and a tawny moustache did not quite cover an unamiable mouth. His cheekbones also gave his face a harsh character; but there was no harshness in the eyes which, looking at the world from under their tawny eyebrows, gave the impression of a man ever alert to greet a redeeming instinct in others but often disappointed. (b) MacRae comes down from the truck and leads Scott in a wide circle through the wet grass. He goes faster and faster, building up speed and soon both man and horse are almost running. Through the greyness of the blurring, slanting rain they look almost like a black-and-white movie that is badly out of focus. Negative subject matter, engaged style (a) Generous tears filled Gabriel's eyes. He had never felt like that himself towards any woman but he knew that such a feeling must be love. The tears gathered more thickly in his eyes and in the partial darkness he imagined he saw the form of a young man standing under a dripping tree. Other forms were near. His soul had approached that region where dwell the vast hosts of the dead. He was conscious of, but could not apprehend, their wayward and flickering existence. His own identity was fading out into a grey impalpable world: the solid world itself which these dead had one time reared and lived in was dissolving and dwindling. (b) One evening I went into the back drawing-room in which the priest had died. It was a dark rainy evening and there was no sound in the house. Through one of the broken panes I heard the rain impinge upon the earth, the fine incessant needles of water playing in the sodden beds. Some distant lamp or lighted window gleamed below me. I was thankful that I could see so little. All my senses seemed to desire to veil themselves and, feeling that I was about to slip from them. Positive subject matter, detached style (a) Her image accompanied me even in places the most hostile to romance. On Saturday evenings when my aunt went marketing I had to go to carry some of the parcels. We walked through the flaring streets, jostled by drunken men and bargaining women, amid the curses of labourers, the shrill litanies of shop-boys who stood on guard by the barrels of pigs' cheeks, the nasal chanting of streetsingers, who sang a come-all-you about O'Donovan Rossa, or a ballad about the troubles in our native land. (b) My grandmother gets up and goes for her violin which hangs on a peg inside her bedroom door. It is a very old violin and came from the Ireland of her ancestors, from the crumbled foundations that now dot and haunt Lochaber's shores. She plays two Gaelic airs. Her hands have suffered stiffness and the lonely laments waver and hesitate as do the trembling fingers upon the four taut strings. She is very moved by the ancient music and there are tears within her eyes. (Continued)

105

106

CUPCHIK AND PHILLIPS TABLE 1 (Continued)

Positive subject matter, engaged style (a) Farrington's eyes wandered at every moment in the direction of one of the young women. There was something striking in her appearance. An immense scarf of peacock-blue muslin was wound round her hat and knotted in a great bow under her chin; and she wore bright yellow gloves, reaching to the elbow. Farrington gazed admiringly at the plump arm which she moved very often and with much grace; and when, after a little time, she answered his gaze he admired still more her large dark brown eyes. The oblique staring expression in them fascinated him. She glanced at him once or twice and, when the party was leaving the room, she brushed against his chair and said O, pardon! in a London accent. (b) The bees buzz from the lilacs at the base of the house and bounce drunkenly against the window. The barn swallows with their delicately forked tails flash their orange breasts and dart and swoop after invisible insects. The dogs lie silently, moving only their eyes, conserving their strength as well. We are drowsy and waiting in the summer's heat.

consciousness, leaving the reader to work harder to infer the internal states of the characters from a distance, an overall negative effect.

Subjects and design A total of 32 undergraduate students at the University of Toronto, comprising 16 males and 16 females, participated in this experiment for course credit. The design included one between-subjects variable, Gender, and three withinsubjects variables, Odour (pleasant, unpleasant), Subject Matter (positive, negative), and Style (engaged, detached). A complex procedure was adopted to ensure full counterbalancing of odours and literary fragments. Familiarisation trials. The 12 odours (6 pleasant, 6 unpleasant) were organised into three groups (Groups 1, 2, and 3). Each group comprised 4 familiarisation odours (2 pleasant, 2 unpleasant) odours randomly chosen from the overall sample of 12 without replacement and 8 odours (4 pleasant, 4 unpleasant) that would be paired with the literary fragments. Consequently, all 12 odours were represented in the 4 familiarisation trials across the three stimulus groups. The familiarisation data were not considered in the analysis of the results. Odour 6 Literature combinations. The eight literary excerpts were divided into two matching 2 6 2 matrices (a and b) with each matrix representing the factorial combination of Subject Matter (SM; positive, negative) and Style (ST; engaged, detached). Random assignment was used to determine which of the two instances of each stimulus type (e.g., positive SM) would be represented in Matrix (a) or (b). Matrix (a) would be used for combinations of literary passages with pleasant Odours and Matrix (b) for pairings with unpleasant Odours.

SCENT OF LITERATURE

107

The 4 pleasant odours from odour Group 1 were then each randomly assigned to one of four cells formed by the factorial combination of Subject Matter and Style (e.g., positive SM, engaged ST). Similarly, the 4 negative odours from odour Group 1 were then each randomly assigned to one of four cells formed by the factorial combination of Subject Matter and Style (e.g., positive SM, detached ST). The same procedure was repeated for the 4 pleasant and 4 unpleasant odours from Groups 2 and 3. This procedure yielded two 2 6 2 matrices with each cell of each matrix comprising three odours representing the three odour groups (1, 2, and 3). The initial random assignment procedure yielded separate matching anchor matrices for pleasant (Matrix a) and unpleasant (Matrix b) odours. Each of the four cells in each matrix had a representative from each of the 3 odour groups. Three series of 8 odour/excerpt combinations were then derived. The first series comprised the 8 assigned literary excerpt/odour combinations from Group 1 in Matrices (a) and (b). An actual order of presentation involving Group 1 stimuli was randomly established with the constraint that two pleasant and two unpleasant odours appear in each of the first and second block of four trials. The same procedure was repeated for stimuli assigned to Groups 2 and 3. The odours were then rotated clockwise one cell in each matrix and the procedure was repeated three times. This was done through a single full rotation yielding 12 combinations of 8 odours and 8 fragments, 4 for each of the 3 odour Groups (1, 2, and 3). Twenty-four presentation orders were produced by reversing the assigned sequences. A male or female subject would receive the odour/fragment combination in the forward sequence and the second subject (female or male) would receive the odours in the reverse order. Twenty-four familiarisation trials involving each of the three Groups were also randomly chosen for presentation at the outset of the experiment (12 sequences 6 forward/backward).

Procedure Subjects were run individually in a 3 m 6 3.3 m windowless room with the door kept ajar. Air was circulated through vents in the ceiling and constant room temperature was maintained. Subjects were told that ``we are interested in people's reactions to odours'' and the odours were presented in the assigned sequence. In the Familiarisation phase of the experiment, subjects were asked for their `ìmpressions'' of four odours. They opened the cap of the squeeze bottle, smelled each odour, and then rated it on five 7-point scales indicating how soothing, pleasant, and energising it was (1 = not at all, 7 = extremely), as well as how many images and feelings it evoked (1 = none at all, 7 = a great many). In the Literary Reading phase of the experiment, subjects were asked ``to smell a series of odours while reading passages from short stories''. Other than saying

108


that ``we are interested in people's reactions to odours,'' no additional justification was provided and none was demanded by the subjects. Subjects were instructed to ``please open the cap and smell each odour while reading the given passage'' and were permitted to ``sniff the bottle up to three times in order to refresh your experience of the odour''. After reading the passage, they rated the odor on six 7-point scales and the passage on eight 7-point scales. The odour perception scales included the five listed above as well as ``Did the odour bring the story to life'' (1 = not at all, 7 = very much so). The story judgement scales measured how absorbed and involved the subject was in the passage, how easily the subject could imagine the scene and have a sense for the plot, how pleasing and personally meaningful the passage was (1 = not at all, 7 = extremely), as well as did the subject relate to or identify with any of the characters (1 = not at all, 7 = extremely easily), and did the subject experience a mood (1) or was the mood in the passage (7).

RESULTS Overview One goal of the study was to examine intercorrelations among the 14 verbal rating scales and determine whether the factor structure would distinguish affective from cognitively loaded responses to the literary passages. In the earlier study (Cupchik et al., 2005 this issue), affective and cognitive processes were clearly contrasted but the odors were decontextualised. The scales used in this study made reference to textual properties as well as perceived qualities of the odours. The central purpose of the study was to test the congruence hypothesis that hedonic similarity would enhance the aesthetic experience. If the congruence hypothesis did receive support, would this apply to both positive and negative conditions as well as encompassing cognitive and affective responses? Factor analysis. While there were only 32 subjects in the laboratory sample, each of their scale ratings reflected responses to eight reading episodes and this contributed to the stability of each verbal scale measure. The data were therefore aggregated for each participant across all eight of his/her reading episodes. A factor analysis was performed, with varimax rotation, to examine intercorrelations among the 14 verbal rating scales (see Table 2). Four factors were derived from the factor analysis. The first factor was labelled Meaningful Involvement based on the seven scales with loadings of .50 or greater; personally meaningful, identify with characters, involved, or absorbed in the passage, sense for plot, passage is pleasing, and odour brings story to life. Subjects who were absorbed in the passages and found them personally meaningful also identified with the characters and had a sense for the unfolding plot. The fact that odours could bring the story to life suggests their potential priming effect on cognitive activities involved in engaged reading.

SCENT OF LITERATURE

109

TABLE 2 Factor analysis of verbal rating scales Factor

Loading

1. Meaningful Involvement Personally meaningful Identify with character Involved in passage Absorbed in story Odour brings story to life Sense for the plot Pleasing passage

.91 .84 .83 .81 .71 .68 .56

2. Affective response Pleasant odour Energising odour Soothing odour

.89 .85 .83

3. Situational Model Mood in passage Imagine scene Sense for the plot

.83 .63 .58

4. Experiential Responses Odour-evoked images Odour-evoked feelings

.93 .77

Eigenvalue 6.46

1.97

1.61

1.09

The second factor involved direct effects of the odours on pleasing, energising, and soothing Affective Responses. This replicated the affectively oriented factor found earlier when odours were presented without any context (Cupchik et al., 2005 this issue). The third factor represented a detached Situational Model because of the three objectively oriented scales: mood was in the passage, ability to imagine the scene, and a sense for the plot. The fourth factor reflected Odour Experience combining images and feelings stimulated by the odours. Thus, the four factors encompassed both affective (Factor 2) and cognitive processes reflecting the facilitative influence of odors (Factors 1 and 4) and the dynamics of engrossed reading (Factor 3). Analysis of variance. The primary goal of the experiment was to determine whether odour would interact with literary properties to shape the aesthetic experience along cognitive and/or affective dimensions. Previous research (Herz & Cupchik, 1993) had underscored the importance of hedonic congruence between odour and theme in paintings, particularly for positive images. The hypothesis was that this process should extend to other cultural media, in this instance excerpts from literary passages. A series of ANOVAs were conducted on the raw ratings of the passage by participants on the 14 rating scales, treating

110


Gender as a between-subjects variable and Odour (Pleasant, Unpleasant), Subject Matter (Positive, Negative), and Style (Engaged, Detached) as withinsubjects variables.

Main effects Results showed that the pleasant odours had very powerful effects on subjects' reactions. Highly significant main effects were found for Odour (Pleasant, Unpleasant) on 7 scales (see Table 3); Pleasant Odour, F(1, 30) = 258.11, p < .0001, Soothing Odour, F(1, 30) = 168.25, p < .0001, Energising Odour, F(1, 30) = 73.31, p < .0001, and Pleasing Passages, F(1, 30) = 10.10, p < .003. The significant main effects for Odour-evoked Images, F(1, 30) = 25.37, p < .0001, Imagine the Scene, F(1, 30) = 6.54, p < .02, and Odour Brings the Story to Life, F(1, 30) = 8.64, p < .01, were incorporated in higher order interactions. The results reveal that subjects gave the pleasant odours higher ratings than unpleasant odours on the three basic dimensions of Affective Response (Factor 2); pleasant, soothing, and energising, which generalised to make the passages more pleasing. But pleasant odours also stimulated cognitive activities, evoking more images, making it easier to imagine the scenes that they helped bring to life. The pleasant odours were therefore generally more effective than the unpleasant odours at enhancing, not just affective states, but cognitive processes as well. Subject Matter had a direct effect (see Table 3) on cognitive and affective oriented scales from the Meaningful Involvement factor: Identify with Characters, F(1, 30) = 13.18, p < .001, Imagine the Scene, F(1, 30) = 15.28, p < .001, Sense for the Plot, F(1, 30) = 21.00, p < .0001, Personally Meaningful, F(1, 30) = 13.67, p < .001, Absorbed, F(1, 30) = 5.00, p < .03, Involved in Passage, F(1, 30) = 15.29, p < .001, and Pleasing Passages, F(1, 30) = 36.50, p < .0001, scales. The passages with Positive subject matter were judged more personally meaningful and it was easier for subjects to identify with the characters, imagine the scene, and sense the plot in those passages. Not surprisingly, subjects were also more absorbed and involved in the positive passages and found them more pleasing than the negative ones. Style affected cognitive and emotional scales representing three of the factors (see Table 3). These included: Identify with Characters, F(1, 30) = 10.49, p < .003, Personally Meaningful, F(1, 30) = 11.26, p < .002, Involved in Passage, F(1, 30) = 6.41, p < .02, Soothing Odour, F(1, 30) = 6.94, p < .01, and Pleasing Passages, F(1, 30) = 5.49, p < .03, scales. The main effect for Imagine the Scene, F(1, 30) = 4.95, p < .03, was also incorporated in a higher order interaction. Cognitively speaking, an Engaged style made it easier for subjects to identify with the characters, imagine the scene, and find the passages personally meaningful. From an affective point of view, the passages written in an Engaged

SCENT OF LITERATURE

111

TABLE 3 Main effects for Odour, Subject Matter, and Style Odour Rating scale Pleasant Odour Soothing Odour Energising Odour Odour-evoked Images Imagine the Scene Odour Brings Story to Life Pleasing Passage

Pleasant

Unpleasant

5.09 4.81 3.89 5.02 5.54 4.14 4.27

2.56{ 2.47{ 2.19{ 4.09{ 5.11* 3.47** 3.73**

Subject matter Rating scale Identify with Characters Imagine Scene Sense for the Plot Personally Meaningful Absorbed in Story Involved in Passage Pleasing Passage

Positive

Negative

3.81 5.71 5.57 3.71 4.55 4.38 4.63

2.86*** 4.94*** 4.56{ 2.88*** 4.08* 3.52*** 3.38{ Style

Rating scale Identify with Characters Imagine the Scene Personally Meaningful Involved in Passage Soothing Odour Pleasing Passage

Engaged

Detached

3.70 5.53 3.63 4.18 3.89 4.20

2.97** 5.12* 2.95** 3.71* 3.39** 3.80*

* p < .05; ** p < .01; *** p < .001; { p < .0001.

style were also generally more involving and pleasing, and subjects judged them to be more soothing.

Interactions A significant two-way interaction of Odour and Style was found for the Imagine the Scene scale which supported the congruence hypothesis, F(1, 30) = 5.71, p < .02. A combination of Engaged Style and Pleasant Odour made it easier to imagine the scene (see Figure 1).

112


Figure 1. Interactions of Odour and Style for the Imagine the Scene scale.

The most powerful test of the hedonic congruence hypothesis was found in significant three-way interactions of Odour, Subject Matter, and Style for Odour-evoked Images, F(1, 30) = 8.99, p < .01, Odour-evoked Feelings, F(1, 30) = 6.63, p < .02, and Odour Brings the Story to Life, F(1, 30) = 15.38, p < .001, scales. The results in Figure 2 all make the same point in support of the hypothesis and the interactions arise because virtually all of the activity appears to be concentrated in two key cells. The effects were strongest when Subject Matter was Positive, Style was Engaged and the Odour was Pleasant, and weakest when Subject Matter was Positive, Style was Engaged but the Odour was Unpleasant. Accordingly, subjects experienced the most Odour-evoked Images (Figure 2a) and Feelings (Figure 2b), and Odour Brings the Story to Life most (Figure 2c), when the hedonic odour context fit with the Positive Subject Matter and Engaging Style (i.e., odour was Pleasant), and was minimised when it did not fit (i.e., odour was Unpleasant). This interaction was not obtained when the Subject Matter was Negative, the ultimate anchoring effect.

Gender differences Significant interactions of Gender and Odour were found on five of the scales from Factor 1 which was Meaningful Involvement; Personally Meaningful, F(1, 30) = 7.33, p < .01, Identify with the Characters, F(1, 30) = 5.93, p < .02, Involved, F(1, 30) = 5.23, p < .03, Imagine the Scene, F(1, 30) = 6.55, p < .02, and Odour Brings the Story to Life, F(1, 30) = 5.09, p < .03. The results in Figure 3 show that differences between males and females were observed only

Figure 2. Interactions of Odour, Subject Matter, and Style for: (a) Odour-evoked Images; (b) Odour-evoked Feelings; and (c) Odour Brings the Story to Life.

113

114

Figure 3. Interactions of Odour and Gender for: (a) personally Meaningful; (b) Identify with the Characters; (c) Involved; (d) Imagine the Scene; and (e) Odour Brings the Story to Life.

SCENT OF LITERATURE

115

for Unpleasant Odours. Paradoxically, males found the Unpleasant Odours to be more personally meaningful (Figure 3a) and involving (Figure 3c) compared with the females. Males also identified with the characters (Figure 3b), imagined the scene (Figure 3d), and odour brought the story to life (Figure 3e) more readily in the Unpleasant hedonic context.

DISCUSSION This study was undertaken to examine the effects of hedonic odour context on the experience of reading literary passages. Previous research using paintings (Herz & Cupchik, 1993) had shown that congruence between the theme of a painting and the hedonic quality of an odour primed cognitively oriented judgements of quality, at least for works with positive subject matter. Pairing unpleasant odours with thematically positive paintings diminished judgements of complexity and personal meaningfulness. Paintings with negative themes were not affected, perhaps because they were very intense and this emotional anchoring made them less susceptible to contextual influences. Another study (Cupchik et al., 2005 this issue) established that odours can have distinct effects on cognitive and affective processes. Together, these studies imply that congruence between the hedonic tone of subject matter and odour would have an impact on cognitive processing of literary materials, especially for passages with positive themes. Factor analysis replicated the earlier finding in Cupchik et al. (2005 this issue) of a distinctive Affective Response factor comprising the pleasant, soothing, and energising scales which represent the three traditional dimensions of affect described by Wundt (1903; see Izard, 1971). Two of the three cognitively oriented factors covered different aspects of possible odour/narrative interaction; Meaningful Involvement and Experiential Responses. The Meaningful Involvement factor was interesting because it combined personally meaningful, identify with characters, involved, or absorbed in the passage, sense for plot, passage is pleasing, with odour brings story to life. The distinct combination of odour-evoked feelings and images separated out the Experiential influences of odours. The other factor, Situational Model, was uniquely relevant to literary text processing and described the detached understanding of characters engaged in purposive action (Graesser et al., 1994). Analysis of variance established the potent effects of the three experimental variables; odour, subject matter, and style. The pleasant odours enhanced reactions on the three scales of the Affective Responses factor, pleasant, soothing, and energising. They also facilitated cognitive processes, evoking images, helping subjects imagine the scenes, and bringing the story to life. Positive subject matter and engaged style also fostered absorption in the literary passages and related cognitive processes. Thus, subjects felt more involved, found the passages more personally meaningful and pleasing, identified with

116


characters, and could imagine the scene more readily when either subject matter or style were positive. The bias in favour of positive themes has been observed elsewhere in terms of reactions to paintings (Herz & Cupchik, 1993) and funny versus violent magazine photographs (Cupchik & Izadpanah, 1998), as well as success at recalling films (Boltz, Schulkind, & Kantra, 1991), social narratives (Isen, 1987), rock music videos (Cupchik & Saltzman, 1999), and qualities of characters in short story excerpts (Phillips & Cupchik, 2004). The centrepiece of this study was an examination of the interaction between the top-down interpretive processing of literary texts and the bottom-up priming effects of odours. According to the hedonic congruence hypothesis, a match between the hedonic tone (i.e., pleasantness) of the text properties and odours would facilitate cognitive responses to the texts, whereas an absence of fit would have deleterious effects. Support for the congruence hypothesis was found in different aspects of response involving odour-evoked images, generalised feelings, and odour brings the story to life, but only for positive stimuli. Congruence between positive subject matter, an engaged style, and pleasant odour yielded the greatest number of odour-evoked images and feelings, and was most effective at bringing the stories to life. But these effects were destroyed if the odour was unpleasant and incongruous with the positive textual properties. These results therefore replicate those found with paintings which were also restricted to positive motifs and applied primarily to cognitive processes. The results also fit with the idea of `ìmplicit learning'' (Degel & KoÈster, 1999; Degel et al., 2001) which in this case means the integration of an odour with a literary passage so that they become a coherent unit. However, the results show that odours do not have to be ambient or subthreshold in order for a relationship with a semantic structure to be established. The primary point is that an attempt needs to be made to achieve coherence or integration across the sensory/verbal boundary and this is a function of how the participant approaches the event or episode. In order for implicit learning to take place, the focus cannot be on the odour as an isolated stimulus to which one must attend and identify, and then append to the text as in a paired-associate task (Chu & Downes, 2000). From the perspective of gender differences, males were more responsive than females to the unpleasant odours, finding them more personally meaningful and involving, and they could more readily identify with the characters in the short story excerpts. Why were males generally more responsive to the unpleasant odours? First, it has been suggested that hemispheric laterality is tied to affective valence, with the left hemisphere favouring positive emotions and the right hemisphere tied to negative emotions (Davidson, 1992; Silberman & Weingartner, 1986). Second, there are hemispheric differences in odour perception and recognition. Herz, McCall, and Cahill (1999) found that familiar and mildly pleasant odours were named more accurately when presented to the left nostril (and therefore to the left hemisphere) and were rated as more pleasant when sniffed through the right nostril and therefore processed in the right hemisphere.

SCENT OF LITERATURE

117

Zucco and Tressoldi (1988) also observed better perceptual processing of odours in the right hemisphere. Third, to the extent that females have privileged access to verbal centres in the left hemisphere (Safer, 1981), they should be more sensitive to pleasant odours and affects that are better represented in the left hemisphere. Males, who are generally more lateralised than females, and who are responsive to perceptual processes centred in the right hemisphere (McGlone, 1980), should therefore be better able to develop coherent images and associations involving unpleasant odours which are more readily perceived in the right hemisphere. While the meaning attached to positive and negative odours is socially and personally grounded, neurological structures may provide an underpinning for sex difference in the associative process. In sum, this experiment examined the effects of hedonic context on the reading experience. While odours had a direct facilitative effect on experiences of pleasant, soothing, and energising affects, interactions with subject matter or style were observed for cognitive or interpretive processes relating to images, generalised feelings, and bringing the story to life. Within this framework, hedonic congruence governed the facilitative effects of interactions between top-down (i.e., interpreting subject matter or style) and bottom-up (i.e., odour) processes, particularly for positive hedonic qualities. One might expect that this principle of hedonic congruence would extend to memory for the details of literary passages as well, and this is indeed the case (Phillips and Cupchik, 2004). The combination of positive subject matter and pleasant odour was reflected in more accurate recall of character details, while pairing negative subject matter and unpleasant odour resulted in more accurate recall of setting details. Thus, odour as a hedonic context has an enhancing effect on both the experience and recall of literary texts. Manuscript received 21 June 2002 Revised manuscript received 10 February 2004

REFERENCES Bartlett, F. C. (1932). Remembering: A study in experimental and social psychology. Cambridge, UK: Cambridge University Press. Boltz, M., Schulkind, M., & Kantra, S. (1991). Effects of background music on the remembering of filmed events. Memory and Cognition, 19, 593±606. Braun, I., & Cupchik, G. C. (2001). Phenomenological and quantitative analyses of absorption in literary passages. Empirical Studies of the Arts, 19, 85±109. Bruner, J. (1986). Actual minds, possible worlds. Cambridge, MA: Harvard University Press. Chu, S., & Downes, J. J. (2000). Odor-evoked autobiographical memories: Psychological investigations of Proustian phenomena. Chemical Senses, 25, 111±116. Cupchik, G. C. (1995). Emotion in aesthetics: Reactive and reflective models. Poetics, 23, 177±188. Cupchik, G. C. (1996). Suspense and disorientation: Two poles of emotionally charged literary uncertainty. In P. Vorderer, H .J. Wulff, & M. Friedrichsen (Eds.), Suspense: Conceptualizations, theoretical analyses, and empirical explorations (pp. 189±197). Hillsdale, NJ: Erlbaum.

118


Cupchik, G. C., Phillips, K., & Truong, H. (2005). Sensitivity to the cognitive and affective qualities of odours. Cognition and Emotion, 19, 121±131. Cupchik, G. C., & Izadpanah, S. (1998). Reactions to funny, violent, and sexual images in magazines. Spiel, 17, 15±32. Cupchik, G. C., & Laszlo, J. (1994). The landscape of time in literary reception: Character experience and narrative action. Cognition and Emotion, 8, 297±312. Cupchik, G. C., Leonard, G., Axelrad, E., & Kalin, J. (1998). The landscape of emotion in literary reception: Generating and receiving interpretations of James Joyce. Cognition and Emotion, 12, 825±847. Cupchik, G. C., & Saltzman, M. (1999). Recognition memory for serious and light rock music videos. Empirical Studies of the Arts, 17, 59±72. Davidson, R. J. (1992). Emotion and affective style: Hemispheric substrates. Psychological Science, 3, 39±42. Degel, J., & KoÈster, E. P. (1999). Odors: Implicit memory and performance effects. Chemical Senses, 24, 317±325. Degel, J., Piper, D., & KoÈster, E. P. (2001). Implicit learning and implicit memory for odors: The influence of odor identification and retention time. Chemical Senses, 26, 267±280. Eco, U. (1989). The open work (A. Cancogni, Trans.). Cambridge, MA: Harvard University Press. (Original work published 1962) Ehrlichman, H., & Halpern, J. N. (1988). Affect and memory: Effects of pleasant and unpleasant odors on retrieval of happy and unhappy memories. Journal of Personality and Social Psychology, 55, 769±779. Engen, T. (1991). Odor sensation and memory. New York: Praeger. Graesser, A. C., Singer, M., & Trabasso, T. (1994). Constructing inferences during narrative text comprehension. Psychological Review, 101, 371±395. Herz, R. S., & Cupchik, G. C. (1992). An experimental characterization of odor-evoked memories in humans. Chemical Senses, 17, 519±528. Herz, R. S., & Cupchik, G. C. (1993). The effect of hedonic context on evaluations and experience of paintings. Empirical Studies of the Arts, 11, 147±166. Herz, R. S., & Cupchik, G. C. (1995). The emotional distinctiveness of odor-evoked memories. Chemical Senses, 2, 517±528. Herz, R. S., McCall, C., & Cahill, L. (1999). Hemispheric lateralization in the processing of odour pleasantness verus odor names. Chemical Senses, 24, 691±695. Hoege, H. (1984). The emotional impact on aesthetic judgments: An experimental investigation of a time-honored hypothesis. Visual Arts Research, 10, 37±48. Isen, A. (1987). Positive affect, cognitive processes, and social behavior. In L. Berkowitz (Ed.), Advances in experimental social psychology: Vol. 20 (pp. 203±253). New York: Academic Press. Iser, W. (1978). The act of reading. Baltimore: Johns Hopkins University. Izard, C. E. (1971). The face of emotion. New York: Appleton-Century-Crofts. Kreitler, H., & Kreitler, S. (1972). The psychology of the arts. Durham, NC: Duke University Press. Lipps, T. (1900). Aesthetische EinfuÈhlung. Zeitschrift FuÈr Psychologie und Physiologie der Sinnesorgane, 22, 415±450. McGlone, J. (1980). Sex differences in human brain asymmetry: A critical survey. Behavioral and Brain Sciences, 3, 215±263. Phillips, K., & Cupchik, G. C. (2004). Scented memories of literature. Memory, 12, 366±375. Proust, M. (1960). Swann's way (S. Moncrieff, Trans.). London: Chatto & Windus. (Original work published 1922) Rotton, J. (1983). Affective and cognitive consequences of malodorous pollution. Basic and Applied Social Psychology, 4, 171±191. Safer, M. A. (1981). Sex and hemisphere differences in access to codes for processing emotional expressions and faces. Journal of Experimental Psychology: General, 110, 86±100.

SCENT OF LITERATURE

119

Silberman, E. K., & Weingartner, H. (1986). Hemispheric lateralization of functions relation to emotion. Brain and Cognition, 5, 322±353. Suskind, P. (2001). Perfume: The story of a murderer. New York: Vintage International. (Original work published 1986) Wolpin, M., & Weinstein, C. (1983). Visual imagery and olfactory stimulation. Journal of Mental Imagery, 7, 63±74. Wundt, W. (1903). Grundriss der Psychologie. Stuttgart: Engelmann. Zucco, G. M., & Tressoldi, P. E. (1988). Hemispheric differences in odour recognition. Cortex, 25, 607±615.