how to create sound symbolism
HOW TO CREATE SOUND SYMBOLISM: C R O S S - L I N G U I S T I C A N D E X P E R I M E N TA L E V I D E N C E F O R E S TA B L I S H I N G A C T I V E S O U N D S Y M B O L I C E F F E C T S
Inaugural Cultural Evolution Society Conference 13-15 September 2017 The Max Planck Institute for the Science of Human History
Niklas Johansson Lund University
SOUND SYMBOLISM • Previous studies on sound-meaning associations have usually included few meanings in plenty of languages or plenty of meanings in few languages (except Blasi et al. (2016))
• The present study dives deeper by more thoroughly examining: • The phonetic features of the sounds • The semantic features of the meanings • Both cross-linguistically (344 basic concepts in 245 language families) and experimentally (iterated learning)
• Lexemes transcribed and systematically grouped into sound groups • The percentage of the occurrence sounds in the concepts added together • Potential sound-meaning associations had to pass several statistical and areal tests
MACRO-CONCEPTS • 142 significant associations were found
• Semantic features shared between concepts which were also associated with similar sounds • 35 macro-concepts consisting of very different concepts with some common denominator [+BODY PART]
[+ANGLE]
NECK
ANGLE
[+BODY PART]
[+ACTION] BITE
THROAT
[+ANGLE]
[+ANGLE]
[+DESCRIPTOR] [+ANGLE]
[+BODY PART]
VEL-V CROOKED
VEL VEL+ -VV VEL+V
KNEE
VELAR
[+ANGLE]
Sound-to-sound mappings through a human filter
Onomatopoeia BIRD:
I-LIKE
Specific areas of the tongue exclusively responsible for different basic tastes Soft palate more sensitive to bitter taste (Sato et al. 2002) Onomatopoeia BIRD:
I-LIKE
Vocal gestures BITTER:
VELAR
LARGE: BIG, LONG, WIDE SMALL: SMALL, SHORT, LOUSE
Ohala’s (1994) frequency code Onomatopoeia BIRD:
I-LIKE
Vocal gestures BITTER:
VELAR
Frequency code BIG-SMALL:
LOW-HIGH FREQUENCY
MOTHER, BREAST, MILK
Nasals produced by infants while breastfeeding and [m, n] are acquired early (Sander 1972) Onomatopoeia BIRD:
I-LIKE
Vocal gestures BITTER:
VELAR
Frequency code BIG-SMALL:
LOW-HIGH FREQUENCY
Circumstantial MOTHER:
NASAL
A VESTIGE OR ACTIVE? • Are these associations just ultraconserved words (Pagel et al. 2013), inherited vestiges of a global proto-language (Imai & Kita 2014) or are they still an active part of language?
• Revisiting two of the most thoroughly investigated sound symbolic domains: SHAPE (ROUND-POINTY) & SIZE (BIG-SMALL) • What’s different? The typical experiment is flipped around • Participants sound symbolically shape words through cultural transmission
• Inspired by so-called Chinese whispers or telephone and conducted within the iterated learning framework (Kirby et al. 2015)
Cognitive Biases & Surrounding World
Hypothesis
Hypothesis
…e…one?
Telephone!
Output
Input
…a…pon…?
Mascarpone!
Output
Input
Ash apron?
Output
A VESTIGE OR ACTIVE? •
•
•
•
750 naïve participants were recruited online through Crowdflower •
5 conditions (BIG, SMALL, N/A, ROUND, POINTY)
•
10 participant chains per conditions
•
15 generations of participants per chain
Participants were presented with an arbitrary seed word (no overrepresentation of any sound group) and asked to repeat it
Condition stimuli: •
N/A (no information about the meaning of the word)
•
”The word you are about to hear means big/small” (text stimuli)
•
”The word you are about to hear means [pointy/round shape]” (picture stimuli)
The recording of what they produced was then played for the next participant in the same chain
[gi:mpra:lhu:s]
GENERAL RESULTS: MANNER/POSITION More hard palate sounds
Hard Palate-Lips Ratio (%) 40
30
20
10
0
-10
-20
G1
G2
G3
G4
G5
G6 N/A
More labial sounds
G7 BIG
G8 SMALL
G9 ROUND
G10 POINTY
G11
G12
G13
G14
G15
GENERAL RESULTS: MANNER/POSITION More hard palate sounds
Hard Palate-Lips Ratio (%) 40
30
20
10
0
-10
-20
G1
G2
G3
G4
G5
G6 N/A
More labial sounds
G7 BIG
G8 SMALL
G9 ROUND
G10 POINTY
G11
G12
G13
G14
G15
GENERAL RESULTS: FREQUENCY SCALE More low frequency sounds
High-Low Ratio (%) 40
30
20
10
0
-10
-20
G1
G2
G3
G4
G5
G6 N/A
More high frequency sounds
G7 BIG
G8 SMALL
G9 ROUND
G10 POINTY
G11
G12
G13
G14
G15
GENERAL RESULTS: FREQUENCY SCALE More high frequency sounds
High-Low Ratio (%) 40
30
20
10
0
-10
-20
G1
G2
G3
G4
G5
G6 N/A
More low frequency sounds
G7 BIG
G8 SMALL
G9 ROUND
G10 POINTY
G11
G12
G13
G14
G15
SEMANTIC TYPE IS MIRRORED IN PHONETIC TYPE Type dichotomous continuous
Domain SHAPE: ROUND-POINTY SIZE: BIG-SMALL
Stimuli visual text
Association manner/position frequency scale
• SHAPE grounded in the feel of the tongue and lips, and shapes created by vocal gestures • SIZE grounded in a mappable scale, pitch
CONCLUSION
• All identified macro-concepts (based on both cross-linguistic and experimental results) • Occur throughout human languages independent of family • Can still be actively extracted from language users • Grounded in some sort of mappable embodied experience
• Makes the role of sound symbolism as a bootstrapping mechanism present in the early stages of human language very likely
Thank you for your attention! And special thanks to:
Gerd Carling Lund University
Jon Carr University of Edinburgh
Arthur Holmer Lund University
Simon Kirby University of Edinburgh
Joost van de Weijer Lund University
Kenny Smith University of Edinburgh
Jennifer Culbertson University of Edinburgh
Inaugural Cultural Evolution Society Conference 13-15 September 2017 The Max Planck Institute for the Science of Human History
Niklas Johansson Lund University
SOUND SYMBOLISM • Previous studies on sound-meaning associations have usually included few meanings in plenty of languages or plenty of meanings in few languages (except Blasi et al. (2016))
• The present study dives deeper by more thoroughly examining: • The phonetic features of the sounds • The semantic features of the meanings • Both cross-linguistically (344 basic concepts in 245 language families) and experimentally (iterated learning)
• Lexemes transcribed and systematically grouped into sound groups • The percentage of the occurrence sounds in the concepts added together • Potential sound-meaning associations had to pass several statistical and areal tests
MACRO-CONCEPTS • 142 significant associations were found
• Semantic features shared between concepts which were also associated with similar sounds • 35 macro-concepts consisting of very different concepts with some common denominator [+BODY PART]
[+ANGLE]
NECK
ANGLE
[+BODY PART]
[+ACTION] BITE
THROAT
[+ANGLE]
[+ANGLE]
[+DESCRIPTOR] [+ANGLE]
[+BODY PART]
VEL-V CROOKED
VEL VEL+ -VV VEL+V
KNEE
VELAR
[+ANGLE]
Sound-to-sound mappings through a human filter
Onomatopoeia BIRD:
I-LIKE
Specific areas of the tongue exclusively responsible for different basic tastes Soft palate more sensitive to bitter taste (Sato et al. 2002) Onomatopoeia BIRD:
I-LIKE
Vocal gestures BITTER:
VELAR
LARGE: BIG, LONG, WIDE SMALL: SMALL, SHORT, LOUSE
Ohala’s (1994) frequency code Onomatopoeia BIRD:
I-LIKE
Vocal gestures BITTER:
VELAR
Frequency code BIG-SMALL:
LOW-HIGH FREQUENCY
MOTHER, BREAST, MILK
Nasals produced by infants while breastfeeding and [m, n] are acquired early (Sander 1972) Onomatopoeia BIRD:
I-LIKE
Vocal gestures BITTER:
VELAR
Frequency code BIG-SMALL:
LOW-HIGH FREQUENCY
Circumstantial MOTHER:
NASAL
A VESTIGE OR ACTIVE? • Are these associations just ultraconserved words (Pagel et al. 2013), inherited vestiges of a global proto-language (Imai & Kita 2014) or are they still an active part of language?
• Revisiting two of the most thoroughly investigated sound symbolic domains: SHAPE (ROUND-POINTY) & SIZE (BIG-SMALL) • What’s different? The typical experiment is flipped around • Participants sound symbolically shape words through cultural transmission
• Inspired by so-called Chinese whispers or telephone and conducted within the iterated learning framework (Kirby et al. 2015)
Cognitive Biases & Surrounding World
Hypothesis
Hypothesis
…e…one?
Telephone!
Output
Input
…a…pon…?
Mascarpone!
Output
Input
Ash apron?
Output
A VESTIGE OR ACTIVE? •
•
•
•
750 naïve participants were recruited online through Crowdflower •
5 conditions (BIG, SMALL, N/A, ROUND, POINTY)
•
10 participant chains per conditions
•
15 generations of participants per chain
Participants were presented with an arbitrary seed word (no overrepresentation of any sound group) and asked to repeat it
Condition stimuli: •
N/A (no information about the meaning of the word)
•
”The word you are about to hear means big/small” (text stimuli)
•
”The word you are about to hear means [pointy/round shape]” (picture stimuli)
The recording of what they produced was then played for the next participant in the same chain
[gi:mpra:lhu:s]
GENERAL RESULTS: MANNER/POSITION More hard palate sounds
Hard Palate-Lips Ratio (%) 40
30
20
10
0
-10
-20
G1
G2
G3
G4
G5
G6 N/A
More labial sounds
G7 BIG
G8 SMALL
G9 ROUND
G10 POINTY
G11
G12
G13
G14
G15
GENERAL RESULTS: MANNER/POSITION More hard palate sounds
Hard Palate-Lips Ratio (%) 40
30
20
10
0
-10
-20
G1
G2
G3
G4
G5
G6 N/A
More labial sounds
G7 BIG
G8 SMALL
G9 ROUND
G10 POINTY
G11
G12
G13
G14
G15
GENERAL RESULTS: FREQUENCY SCALE More low frequency sounds
High-Low Ratio (%) 40
30
20
10
0
-10
-20
G1
G2
G3
G4
G5
G6 N/A
More high frequency sounds
G7 BIG
G8 SMALL
G9 ROUND
G10 POINTY
G11
G12
G13
G14
G15
GENERAL RESULTS: FREQUENCY SCALE More high frequency sounds
High-Low Ratio (%) 40
30
20
10
0
-10
-20
G1
G2
G3
G4
G5
G6 N/A
More low frequency sounds
G7 BIG
G8 SMALL
G9 ROUND
G10 POINTY
G11
G12
G13
G14
G15
SEMANTIC TYPE IS MIRRORED IN PHONETIC TYPE Type dichotomous continuous
Domain SHAPE: ROUND-POINTY SIZE: BIG-SMALL
Stimuli visual text
Association manner/position frequency scale
• SHAPE grounded in the feel of the tongue and lips, and shapes created by vocal gestures • SIZE grounded in a mappable scale, pitch
CONCLUSION
• All identified macro-concepts (based on both cross-linguistic and experimental results) • Occur throughout human languages independent of family • Can still be actively extracted from language users • Grounded in some sort of mappable embodied experience
• Makes the role of sound symbolism as a bootstrapping mechanism present in the early stages of human language very likely
Thank you for your attention! And special thanks to:
Gerd Carling Lund University
Jon Carr University of Edinburgh
Arthur Holmer Lund University
Simon Kirby University of Edinburgh
Joost van de Weijer Lund University
Kenny Smith University of Edinburgh
Jennifer Culbertson University of Edinburgh
