Embed Size (px)
Citation preview
Cross-Modal Relations between Preference for Harmony and Emotional ContentWilliam S. Griscom and Stephen E. PalmerDepartment of Psychology, University of California, Berkeley
Background
Research Questions
Stimuli & Methods
Emotional Associates of HarmonyWithin Domain Preference for Harmony (PFH)
Conclusions
References & Acknowledgements
Previous research has shown that individual aesthetic preferences are correlated across domains (e.g., Eysenck, 1940). Recently, research has shown that preference for color combinations is largely driven by color harmony, but individuals vary in degree of preference for harmony (PFH) (Schloss & Palmer, 2010).
How well does individual preference for harmony in one domain predict preference for harmony in other domains?
Are personality factors or expertise relevant?
How does cross-domain preference for harmony relate to emotional content?
Cross-Domain Preference for Harmony (PFH)
Color and musical harmony are best predicted by high scores on the Calm-Angry dimension (r2=.41 and .60), whereas shape simplicity and goodness-of-fit are best predicted by high scores on the Strong-Weak dimension (r2=.67 and .39).
-0.1 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8012345678910
Preference-Harmony Correlation (r)
Subj
ects
Experiment 130 Psychology students30 Art Practice students30 Music students
All stimuli were rated first for preference and then for harmony, blocked by domain, using a bipolar line-mark rating scale.
Experiment 220 Psychology students
All stimuli rated on four basic emotional dimensions:
1- Happy/Sad2 - Angry/Calm3 - Active/Passive4 - Strong/Weak
Eysenck, H.J. (1940) The General Factor in aesthetic judgements. British Journal of Psychology, 31:1 94-102.Garner, W.R. (1970) Good patterns have few alternatives. American Scientist, 58, 34–42.Palmer, S. E. (1991) Goodness, Gestalt, Groups, and Garner: Local symmetry sub groups as a theory of figural goodness. In G. Lockhead & J. Pomerantz (Eds.), The perception of structure: Essays in honor of Wendell R. Garner. Washington, D.C.: APA. Schloss, K. B. & Palmer, S. E. (2011) Aesthetic response to color combinations: Preference, harmony, and similarity. Attention, Perception, and Psychophysics, 73, 551-571.
Thanks to Karen Schloss, Rosa Poggesi, Kyle Jennings, and many more. Funding for this research was provided by Amy’s Natural Foods, the National Science Foundation (BCS-0745820) and Google.
Preference for harmony seems to be a domain general explanation for individual differences in aesthetic preference.
Formal training has an effect of reducing preference for harmony in relevant domains. In addition to these domain specific effects, participants with more training showed lower general preference for harmony.
Across domains, harmony is associated with positive emotional valence. Color and music harmony are associated with calmness, while spatial harmony is associated with strength.
Cross-Domain Correlations in PFH Group Differences in PFH
Harmonious Disharmonious
Preference for Color Harmony (CH)
Preference for Musical Harmony (MH)
Preference for Shape Simplicity (SS)
Preference for Goodness of Fit (GF)
Happy/Sad Calm/Angry Active/Passive Strong/Weak
Happy/Sad Calm/Angry Active/Passive Strong/Weak
Happy/Sad Calm/Angry Active/Passive Strong/Weak
Happy/Sad Calm/Angry Active/Passive Strong/Weak
Shape Simplicity
Musical HarmonyGoodness of Fit
Color Harmony
Cor
rela
tion
(r)
Cor
rela
tion
(r)
Cor
rela
tion
(r)
Cor
rela
tion
(r)
-0.5
0
0.5
1
****
**
**
*
*
*
**
**
r = .97**r = .95**
r = .47*r = .80**
0.43**
0.37**
0.60**
0.39**
0.46** 0.32**
-0.5
0
0.5
1
-0.5
0
0.5
1
-0.5
0
0.5
1
In all domains, average preference predicts average ratings of harmony, but there are large individual differences.
Individual differences are correlated across domains. Groups show domain-relevant differences in PFH.
N=908 parametersGoodness of Fit=.983
N=9012 parameters
Goodness of Fit=.965
GeneralFactorModel
EmpiricalTrimming
Model
Color(CH)
Frame(GF)
Shape(SS)
Music(MH)
-
-
-
-
Color(CH)
Frame(GF)
Shape(SS)
Music(MH)
Structural Equation Models
Confirmatory Factor Analysis shows good fit between data and models with a single general factor for PFH.
# S
ubje
cts
# S
ubje
cts
Average r = .44** Average r = .16**
Average r = .57**Average r = .36**
Ave
rage
With
in-G
roup
PFH
(r)
r (Preference, Harmony) r (Preference, Harmony)
r (Preference, Harmony) r (Preference, Harmony)
. . .
56 color pairs(Schloss & Palmer, 2010)
35 spatial positions(Palmer, 1991)
Bach
14 classical piano pieces
22 dot configurations(Garner, 1970)
Schoenberg
