Background
Color-Grapheme Associations in Non-Synesthetes: Evidence of Emotional MediationChristopher Lau1, Karen B. Schloss2, David M. Eagleman3, and Stephen E. Palmer2
1College of Natural Resources, UC Berkeley; 2Department of Psychology, UC Berkeley; 3Department of Neuroscience, Baylor College of Medicine
Similarities Among Non-synesthetes and Synesthetes
Conclusions
References and Acknowledgements
General Methods
Colors of Graphemes
Do Emotions Mediate Color-Grapheme Associations?
As predicted by the Emotional Mediation Hypothesis:There is a clear mapping between the emotional content of graphemes and the emotional content of the colors chosen to go with those graphemes in non-synesthetes.
Although there are large individual di�erences in color grapheme associations among both synesthetes and non-synesthetes, there are shared systematic e�ects of saturation and lightness among both groups.
Future research: Does the emotional mediation hypothesis hold for synesthetes?
In non-synesthetes, colors are associated with graphemes that share similar emotional content (e.g., happy graphemes are associated with happy colors).
Eagleman, D. M. (in preparation). A large-scale analysis of color associations in colored sequence synesthesia. Rich, A. N., J. L. Bradshaw, and J. B. Mattingley. "A Systematic, Large-scale Study of Synaesthesia: Implications for the Role of Early Experience in Lexical-colour Associations." Cognition 98 (2005): 53-84. Schloss, K. B., Lawler, P. & Palmer, S. E. (VSS-2008). “The Color of Music.” Presented at the 8th Annual Meeting of the Vision Sciences Society, Naples, FL, May 2008.Simner, J., J. Ward, M. Lanz, A. Jansari, K. Noonan, L. Glover, and D. Oakley. "Non-random Associations of Graphemes and Colour in the Synaesthetic and Normal Populations." Cognitive Neuropsychology 22 (2005): 1069-085. Xu, Z., Schloss, K.B. & Palmer, S.E. (VSS-2010). “The Color of Faces.” Presented at the 10th Annual Meeting of the Vision Sciences Society, Naples, FL, May 2010.
AcknowledgementsWe thank Rosa Poggesi, Brian Alvarez, Thomas Langlois, Arielle Younger, Mathilde Heinemann, Madison Zeller, and Joseph Austerweil. This research was funded by the National Science Foundation grant (#BCS-0745820) to Stephen Palmer and a gift from Google.
Color-Grapheme Associations (CGA) were calculated for each grapheme in terms of the emotional dimensions (D) of the �ve colors that were most consistent with the grapheme (C) and the �ve that were most inconsistent with the grapheme (I).
Color-Grapheme Associations
Berkeley Color Project (BCP) 37 colors presented beside each grapheme (26 letters, 9 numbers).
Tasks:1) Click the 5 most consistent colors with the grapheme starting with the most consistent
2) Click the 5 most inconsistent colors with the grapheme starting with the most inconsistent
Graphemes presented one at a time with a bipolar emotional scale.
Tasks:Rate emotional content of each grapheme (blocked by emotional dimension): happy-sad active-passive angry-calm strong-weak good-evil
Grapheme-Emotion Associations Color-Emotion Associations
BCP-37 colors presented one at a time with a bipolar emotional scale
Tasks: Rate emotional content of each color (blocked by emotional dimension): happy-sad active-passive angry-calm strong-weak good-evil
HappySad
AHappySad
A
-50-100 1000 50
-50
0
50
100
A
B
C
D
E
F
G
HI
J
KLMNO P
QRST
U
V
W
X
Y
Z0
1
3
4
567
89 FH
J
246
Yello
wne
ss-B
luen
ess
Graphemes that are associated with similar colors among synesthetes and non-synesthetes tend to be the “desaturated” (grayish) graphemes
Eagleman (in preparation) showed that, although synesthetes have idiosyncratic color-grapheme pairings, the distance between pairs of graphemes is correlated across subjects. Do non-synesthetes show similar e�ects (i.e., correlated distances for grapheme pairs)?
-100 0 100-150
-100
-50
0
50
100
150
ABC
DE F
G HIJ
KL M
NOP
QR
S
TU
VWXY
Z012
3 45 678
9CG
A
-100 0 100-150
-100
-50
0
50
100
150
ABC DE
FG
H
IJ
K
LMNOPQRST
UV
WXY Z
0
1
2
34
56 7
89
Grapheme Emotion
CG
A
-100 0 100-150
-100
-50
0
50
100
150 A
BC
D
E
FG
HIJ K
LMN
OPQ R
STU VW
XYZ
0
1
2
3456
789
Grapheme Emotion
CG
A
-100 0 100-150
-100
-50
0
50
100
150
ABC
D
E
FG
H IJKLMN O
PQR
STUVW
X YZ 0
1
23
45
67 8
9
Grapheme Emotion C
GA
Grapheme Emotion
Avg. r = .38* Avg. r = .49** Avg. r = .63*** Avg. r = .58***ID r = .08* ID r = .06 ID r = .06* ID r = .16***
Redness-Greenness
Chromaticity Saturation and Lightness
25 50 75 100
-20
-100
-10
0
10
20
30
100
A
B
C
D
E
F
G
H
I
JKLM
N
OP
Q
RS
TU
V
W
X
Y
Z
0
1 2
3
4
5
6
7
8
9
Saturation
Ligh
tnes
s
For synesthetes and non-synesthetes, graphemes have similar saturations (r=.58***) and lightnesses (r=.49**).
10 20 30 40 50 60 70 80 35
45
55
65
75
85
A
B
C
D
E
F
G
H
I
J K
L
M
N
O
P Q
R
S
T
U
V W
X
Y
Z
0 1
2 3
4 5 6
7
8 9
Saturation
Ligh
tnes
s
As in verbal descriptions of colors (Rich et al., 2005; Simner et al., 2005), ‘G’ was greenish, ‘A’ was reddish ‘Y’ was yellowish, and ‘B’ was bluish, although there were large individual di�erences.
Prior research on verbal color-grapheme associations in non-synesthetes:Non-synesthetes tend to have similar associations between graphemes and colors, as tested using color name association tasks (e.g., Rich, Bradshaw, & Mattingley, 2005; Simner et al., 2005).
Linguistic explanation:The �rst letter of a color name is paired with its color (e.g., ‘G’ for green). The �rst letter of a strongly associated colored object is paired with the objects’ color (e.g., ‘A’ is red because of associations with apples).
Will such associations hold for a non-verbal color-association task?
Does the emotional mediation hypothesis hold for less emotionally expressive stimuli, such as graphemes?
‘X’ and ‘Z’ were desaturated and dark. ‘0’ (zero) was desaturated and light. Note that ‘O’ was more saturated (p<.01) and darker than ‘0’ (p=.07), even though they have similar shapes.
Saturation and Lightness of Graphemes Among Synesthetes (Eagleman, in prep)
G Y B X IA
Angry-Calm Strong-Weak Active-Passive Good-Evil
Schloss, Lawler, and Palmer (VSS-08): When choosing the colors that “go best” with orchestral music, participants choose the colors that best matched the emotional content of the music.
Choose the color whose emotional content best matches that of the stimulus
Evaluate emotional content of the stimulus
Experience the stimulus
The Emotional Mediation Hypothesis: When choosing colors to go with an emotional stimulus in a di�erent modality, people choose the colors that best match the emotional content of that stimulus.
Schloss, Lawler & Palmer (VSS-08)
Xu, Schloss & Palmer (VSS-10)
Choose the color whose emotional content best matches that of the grapheme
Evaluate emotional content of the graphemeA
40 50 60 70 80 90 100 110 120 130 140 15010
20
30
40
50
60
70
80
90
100AB
ACAD
AE
AF AG
AH
AIAJ
AK
AL
AMAN
AP AQ
AR
ASAT AUAVAW
AY
A2 A3
AA5A6
A7A8
A9
BC
BD
BE
BFBG
BHBI
BJBK
BL
BMBN
BPBQBR
BS
BT
BUBVBW
BY
B2
B3B4
B5B6
B7
B8
B9
CD
CE
CF
CGCH CI
CJCKCL
CM
CN
CP
CQCR
CSCT
CU
CVCW
CY
C2C3
C4C5
C6C7
C8
C9
DE
DF
DGDHDI
DJ
DKDL
DM
DN DPDQ
DR
DS
DTDU
DVDW
DY
D2
D3
D4 D5
D6D7
D8
D9EF
EGEH
EI
EJ
EKEL
EMEN
EP
EQERES
ETEU
EV
EWEY
E2
E3
E4
E5 E6E7E8
E9
FGFH FI
FJ
FK FL
FM
FN
FP
FQFR
FS
FTFU FVFW
FY
F2
F3F4
F5F6F7
F8
F9
GHGI
GJGK
GLGM
GN
GP
GQ
GRGS
GTGU
GVGW
GY
G2G3
G4G5
G6G7 G8
G9
HI
HJHK
HL HM
HN
HPHQHR
HS
HTHUHVHW
HY
H2H3H4 H5
H6H7H8H9
IJ
IK
IL
IM
IN
IP
IQ
IR IS
IT
IU
IVIW IY
I2I3I4I5
I6I7I8
I9
JKJLJM
JNJPJQ
JR
JS
JT
JUJV
JWJY
J2J3
J4J5
J6J7J8 J9KL
KMKN
KP
KQKR
KS
KT KUKV
KW
KY
K2K3
K4K5K6
K7
K8
K9
LM
LN
LP
LQ
LRLS
LT LULV
LWLY
L2L3L4L5
L6
L7
L8
L9
MN
MPMQ
MR
MSMT
MUMV
MW
MYM2 M3
M4M5
M6
M7
M8
M9
NP NQNR
NS
NTNU
NVNWNY
N2
N3N4N5N6N7
N8
N9PQ
PR
PS
PT
PU PVPW
PYP2
P3
P4 P5
P6
P7
P8
P9QR
QS
QT
QU
QV
QW
QYQ2
Q3
Q4 Q5
Q6
Q7
Q8Q9
RSRTRURV
RW
RY
R2 R3R4
R5R6
R7
R8
R9
STSUSV
SWSY S2
S3
S4
S5 S6S7
S8S9
TUTV
TWTY
T2T3T4
T5T6
T7
T8
T9
UVUW UY
U2
U3U4U5
U6 U7U8
U9 VWVY
V2
V3
V4V5
V6
V7V8
V9WY
W2W3W4W5
W6W7
W8W9
Y2Y3
Y4Y5Y6
Y7
Y8
Y9
23
24 25
262728 29
34
35
3637 38
39
45
46 47
4849
5657
58
59
67
68
69
78
79
89AOAX
AZA0A1BOBXBZ
B0
B1
COCXCZ
C0C1
DODXDZ
D0
D1
EOEX
EZ
E0E1
FOFX
FZ
F0F1
GOGX GZ
G0
G1
HOHX HZ
H0H1
IOIXIZ
I0I1
JOJXJZJ0
J1KO
KX
KZ
K0K1
LOLX
LZ
L0 L1
MOMX
MZM0
M1
NONXNZN0
N1OP
OQ
OROS
OT
OU
OVOW
OX
OY
OZ
O0
O1
O2O3
O4 O5O6 O7
O8O9
PXPZ
P0P1
QX
QZ
Q0
Q1RX
RZR0
R1SXSZS0
S1
TXTZ
T0T1
UXUZ
U0
U1VXVZ V0
V1
WX
WZ
W0
W1XY
XZ
X0
X1
X2X3X4 X5
X6X7
X8X9
YZY0
Y1
Z0
Z1
Z2
Z3Z4
Z5Z6Z7Z8
Z9
01
0203
04 05
06 07
0809
12131415
161718
19
Dis
tanc
es b
etw
een
Gra
phem
e Pa
irs fo
r Syn
esth
etes
Distances between Grapheme Pairs for Non-Synesthetes
All pairsr = .44***
Only desaturated graphemesr = .47***
No desaturated graphemesr = .20***
Average Distance (CIELAB) between Each Pair of Graphemes for Synesthetes vs. Non-synesthetes
Calm Angry
Calm
Ang
ry
StrongWeak
Stro
ngW
eak
ActivePassive
Activ
ePa
ssiv
e
GoodEvil
Goo
dEv
il
CD = (5c1,D + 4c2,D + 3c3,D + 2c4,D + c5,D)/5
ID = (5i1,D+4i2,D+3i3,D+2i4,D+ i5,D)/5
CGAD = CD - ID
most consistent color
most inconsistent color
-100 0 100-150
-100
-50
0
50
100
150 A
BC
D
E
FG
HIJK
LM
NOP
QR
STUV
W
X
Y
Z0
1
23
45
6
789
CGA HS
Grapheme Emotion
Avg. r = .81*** ID r = .16***
Happy-Sad
HappySad
Hap
pySa
d
Experience the grapheme
Computing CGAs
Participants: 42 non-synesthetes, as determined by the www.synesthete.org initial questionnaire.