shape-to-color associations in non-synesthetes: evidence ... · -1.5 -1 -0.5 0 0.5 1...
TRANSCRIPT
Background
Shape-to-Color Associations in Non-synesthetes: Evidence for Emotional Mediation Michela Malfatti1, Karen B. Schloss2, Liliana Albertazzi1,3, & Stephen E. Palmer4
Objective Features vs Subjective Ratings
Conclusions
References and Acknowledgements
General Methods
Relations Between Color, Shape, and Emotion
1 Center for Mind/Brain Sciences, University of Trento 2 Department of Cognitive, Linguistic & Psychological Sciences, Brown University
3 Department of Humanities, University of Trento4 Department of Psychology, University of California, Berkeley
The Color of Line-Shapes
Color and Shape: Stimuli
0
10
20
30
40
50
60
70
80
(-)
(+)
(+)
(+)
(+)
(+)(+)
(-)
Saturatio
n
Lightness
Red/Green
Yellow/B
lue
Varia
nce
expl
aine
d
Stepwise Multiple Linear Regression
Objective Geometric Features Subjective Perceptual Ratings Subjective Emotional Ratings
Pred
icto
rs Curved(0)/Angular(1)
Open(0)/Closed(1)
Asymmetric(0)/Symmetric(1)
# of Line-Segments
Smooth(-)/Pointy(+)
Open(-)/Closed(+)
Asymmetric(-)/Symmetric(+)
Simple(-)/Complex(+)
Weak(-)/Strong(+)
Not-Angry(-)/Angry(+)
Passive(-)/Active(+)
Calm(-)/Agitated(+)
Sad(-)/Happy(+)
Safe(-)/Harmful(+)
# of Intersections
0
10
20
30
40
50
60
70
80
(+)
(-)
(-)(+)
Saturatio
n
Lightness
Red/Green
Yellow/B
lue
(+)
(+)
(+)
(+)(-)
Emotional Mediation Hypothesis: Colors are mapped onto shape through their shared emotional content.
Evaluate emotionalcontent of shape
People choose colors with matching
emotional content
People look ata shape
SATURATED
MUTED
LIGHT
DARK
4 Color Appearance Dimensions : Saturation, Lightness, Red/Green, Yellow/Blue
37 Colors (Berkeley Color Project) 44 Line-Shapes
4 Shape Features:Curviness: Curved, angularSymmetry: Asymmetric, symmetricClosure: Open, intersecting-once, intersecting>1# of Line-Segments: 2, 3, 8 line-segments
8 Hues: Red, Yellow, Green, Blue, Orange, Chartreuse, Cyan, Purple4 Saturation/Lightness levels (”cuts”): Saturated, Light, Muted, Dark+ 5 Achromatic Colors: White, Black, Light, Medium, & Dark Gray 1a 1c 2a 2c 3a 3c 4a 4c 5a 5c 6a
6c 7a 7c 8a 8c 9a 9c 10a 10c 11a 11c
12a 12c 13a 13c 14a 14c 15a 15c 16a 16c 17a
17c 18a 18c 19a 19c 20a 20c 21a 21c 22a 22c
Shape-Color Associations
Pick the 3 colors that are most consistent (and the 3 that are most inconsistent) with the shape in order (1st, 2nd, 3rd).
Shape Emotional Ratings
HappySad
Color Emotional Ratings
HappySad
Shape Appearance Ratings
PointySmooth
Group 1 Participants: 44 non-syns., determined by synesthete.org Group 2 Participants: 20 non-syns.
Rate each color/shape on 9 bipolar (emotional + other) scales: Rate each shape on 5 bipolar perceptual scales:
Sad/HappyCalm/AgitatedNot-Angry/Angry
Passive/ActiveWeak/StrongSafe/Harmful
Cool/WarmUnfamiliar/FamiliarUnpleasant/Pleasant
Smooth/PointyOpen/ClosedSimple/Complex
Asymmetric/SymmetricConvex/Concave
EMOTIONAL PERCEPTUAL
Geometric features of the shape (in particular its curviness and closure) in�uence the colors picked to go with it for di�erent Color Appearance Dimensions (e.g. more angular and intersecting line-shapes are consistent with more saturated colors).
Non-synesthetes have systematic associations between hues and shapes (Albertazzi et al., 2012). Here we extend the study of shape-to-color associations in non-synesthetes to other color attributes and a wider variety of shapes and shape features. Moreover, we test if such associations might be mediated by emotions, as music-to-color associations are (Palmer et al., 2013).
SCA
Sco
re
Lightness Yellow/BlueSaturation Red/Green
most consistent color
most inconsistent color
Shape-Color Association (SCA Score): SCAd,s = Cd,s - Id,s
The weighted average of the colors picked as most consistent with the shape (Cd,s) minus the weighted average of the colors picked as most inconsistent with the shape (Id,s), along a given dimension (d). Id,s = (3i1,d,s+2i2,d,s+ i3,d,s )/6
-200
0
200
Lig
ht
Dar
k
Curved
Angular
-200
0
200
Red
Gre
en
Curved
Angular
-200
0
200
Sat.
Un
sat.
Curved
Angular
Support for the Emotional Mediation Hypothesis: There are strong correlations between emotional ratings of each line-shape and emotional ratings of the colors consistent with that shape, except for Sad/Happy.
Principal Component Analysis (PCA) of the Dimensions showed that 83% of variance could be explained by 1 Component that roughly corresponded to Not-Angry/Angry (PC loadings: Not-Angry/Angry = .99; Safe/Harmful = .99; Calm/Agitated = .99; Passive/Active = .93; Weak/Strong = .91; Sad/Happy = -.62).
Multidimensional Scaling of Shapes based on Emotional Ratings
p < .001
0.2
0.0
0.4
0.6
0.8
1.0
Corr
elat
ion
(r)
Sad/Hap
py
Calm/Agit
ated
Not-Ang
ry/Ang
ry
Passiv
e/Activ
e
Weak/S
trong
Safe/Harm
ful
Dimension 1
Dim
ensio
n 2
-1.5 -1 -0.5 0 0.5 1 1.5-0.8
-0.6
-0.4
-0.2
0
0.2
0.4
0.6
0.8
1a
1c
2a
2c
3a
3c
4a
4c
5a
5c
6a
6c
7a
7c
8a
8c
9a
9c
10a
10c
11a
11c
12a
12c
13a
13c
14a14c
15a
15c
16a
16c
17a
17c
18a
18c
19a
19c
20a
20c
21a
21c
22a
22c
-2
Angry
Happy
Not-Angry
Sad
The color of a line-shape could be better predicted by its Subjective Ratings rather than its Objective Features. Both Subjective Perceptual Ratings and Subjective Emotional Ratings explained a good amount of variance.
0
10
20
30
40
50
60
70
80
(+)
(+)
(-)
(+)
(+)
(-)
(+)
(+)
(-)(-)
Saturatio
n
Lightness
Red/Green
Yellow/B
lue
- Systematic associations between shapes and colors in non-synesthetes were con�rmed.
- We investigated what Shape Features are mapped to what Color Appearance Dimensions speci�cally , and we found main e�ects of the curviness and closure of line-shapes in particular on the saturation and lightness of the associated colors.
- We further found support for the Emotional Mediation Hypothesis: line-shapes and colors picked to go together were associated with many overlapping emotions, which mainly re�ect the Not-Angry/Angry emotional dimension. When making shape-to-color associations, subjects may evaluate the emotional content of the shape and then choose the colors that have a similar emotional content.
Open Int.=1 Int.>1
Closure
Open Int.=1 Int.>1
Closure
Open Int.=1 Int.>1
Closure
-200
0
200
Yello
wB
lue Curved
Angular
Open Int.=1 Int.>1
Closure
Smooth/Pointy
Open/Closed
CORRELATIONS Sad/Happy Calm/Agitated Not-Angry/Angry Passive/Active Weak/Strong Safe/Harmful
-0.53 0.55 0.61 0.35 0.67 0.60
-0.07 0.49 0.44 0.62 0.45 0.49
Albertazzi L., Da Pos O., Canal L., Micciolo R., Malfatti M., & Vescovi, M (2012). The Hue of Shapes. Journal of Experimental Psychology: Human Perception and Performance, 39, 37-47.Albertazzi L., Malfatti M., Canal L., Micciolo R. (in press). The Hue of Angles: Was Kandinsky right? Art and
Perception.Palmer, S. E., Schloss, K. B., Xu, Z. R. M., & Prado-Leon, L. R. (2013). Music-color associations are mediated by
emotion. Proceedings of the National Academy of Sciences, 110, 22, 8836-8841. Whiteford, K. L., Schloss, K. B., & Palmer, S. E. (2013). Music-Color Associations from Bach to the Blues:
Emotional Mediation in Synesthetes and Non-synesthetes. Presented at the 13th Annual Meeting ofthe Vision Science Society, Naples, FL, May 2013.
Acknowledgements: We thank Will Griscom (at UC Berkeley) and Kelly Whiteford (at University of Minnesota) for support throughout the development of the project.
Cd,s = (3c1,d,s+2c2,d,s+ c3,d,s )/6