EMERGENT COMPLEXITY
IN FIRST WORDS
Sophie KERN
Laboratory Dynamique du Langage,
Lyon, France
COMPLEXITY ?
What is complexity ? Lack of definition
Complex sounds vs Easy sounds
Infrequent vs. frequent in word languages
Late vs. early appearing in children
Early vs. late-mastered by children
>> Emergent complexity in first words = change from
easy to complex sounds from 12 to 24 months of
age
2
WHAT IS
EASY AT
12
MONTHS OF AGE
?
Continuity between babbling and first words (Oller et al., 1976 ;
Stark, 1980 ; Stoel-Gammon & Cooper, 1984 ; Vihman, Ferguson & Elbert,
1986 )
Same sound preferences
Within syllable preferences
labial consonants + central vowels [ba]
coronal consonants + front vowels [d]
dorsal consonants + back vowels [k]
Across syllable preferences
oral stops [p, b, d, g], nasal stops [m, n] and glides [w, j]
labials and coronal consonants [b, m, d, t]
mid and low front and central [ , æ,e , ϯ, a] vowels
more high-low variegation than front-back vowel variegation
more manner than place consonant variegation
“Frame then Content” (MacNeilage & Davis, 1993; 2000 )
3
LEXICAL
SPURT
?
First words production (around 12 months) (Fenson et al., 1993)
Until 50 words: slow acquisition of new words
After 50 word mark: Increase of lexical rate > lexical spurt
( Benedict, 1979 ; Bloom, 1973 ; Goldfield & Reznick, 1990 ; Nelson, 1973 ;
Poulain-Dubois & Graham, 1994; Dromi, 1987 et Mervis & Bertrand, 1995)
4
WHY ?
Conceptual development : non verbal categorization
(Gopnik & Melzhoff, 1992 ; Poulain-Dubois et al., 1995; Gershkoff-Stowe et
al., 1997 )
Memory retrieval (Dapretto & Bjork, 2000)
New learning principles > fast mapping (Behrend, 1990;
Markman, 1991; Mervis & Bertrand, 1995)
Increase of communicative needs (Clark, 1993)
Increase of articulatory control (Clark, 1993)
5
HYPOTHESES
Children go through a lexical spurt before 24 months of
age
Children produce more easy sounds and sound
structures than complex ones from 12 to 24 months of
age
Children’s productions are more complex after lexical
spurt than before
6
METHOD
Languages
Subjects
6 languages : French, Romanian, Dutch, Tunisian + Turkish,
Tachelhit
22 children – 4 FR, 3 RO, 4 DU, 4 TU, 4 TUR, 2 TA
Data collection
One hour of audio-video recording every two weeks from 8
months of age till 25 months of age in the children’s homes
12;15 to 24;15 divided in “before 50 words” and “after 50 words”
7
METHOD
Data processing
IPA transcriptions by native speakers of each languages
LIPP entering
Clan and Phon format (Yvan Rose, Brian MacWhinney, Carla
Peddle)
8
Nber of 1 hour sessions
Children
Early words
Later words
FR1
19
14
FR2
14
10
FR3
23
4
FR4
19
14
75
42
RO1
16
-
RO2
21
6
RO3
23
-
60
6
DU1
18
5
DU2
11
17
DU3
12
16
DU4
15
19
56
57
TU1
15
6
TU2
19
7
TU3
22
8
TU4
22
6
Tunisian total
78
27
ALL
269
132
French total
Romanian total
Dutch total
9
Nber of segments
Children
Early words
Later words
FR1
2,300
14,948
FR2
2,423
11,387
FR3
4,939
3,637
FR4
2,079
19,571
11,741
49,543
RO1
5,078
-
RO2
557
7,805
RO3
2,368
-
8,003
7,805
DU1
4,305
3,833
DU2
1,333
9,460
DU3
2,101
19,655
DU4
1,189
31,504
8,928
64,452
TU1
2,282
2,230
TU2
2,515
8,538
TU3
4,492
10,626
TU4
4,717
3,532
Tunisian total
14,006
29,722
ALL
42,678
151,522
French total
Romanian total
Dutch total
10
CUMULATIVE
VOCABULARY
French
Tunisian
600
550
500
450
400
350
300
250
200
150
100
50
0
300
250
200
150
100
50
0
9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
FR1
FR2
FR3
9
10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
TU1
FR4
TU2
TU3
TU4
Romanian
Dutch
200
800
750
700
650
600
550
500
450
400
350
300
250
200
150
100
50
0
150
100
50
11
0
9
10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
DU1
DU2
DU3
9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
DU4
RO1
RO2
RO3
RATE
OF ACQUISITION
80
70
60
50
40
30
20
10
0
Romanian
74
50
43,5
36
36
<50
>50
7
8
5
6,5
Nber of words
Nber of words
French
43,5
40
30
<50
20
10
>50
4
3,25
0
FR1
FR2
FR3
FR4
RO1
100
81
80
54
60
20
<50
34
8,5
6,510,5
14
8
0
DU1
DU2
RO2
RO3
Tunisian
DU3
DU4
>50
Nber of words
Nber of words
Dutch
40
2,5
50
39,8
40
30
<50
20,5
20
10
10
5,5
6
TU1
TU2
12,5
5
4,25
0
TU3
TU4
>50
12
CONSONANTS MANNER
Stops + glides (vs. other manners)
100
95
FR1
FR2
Percentages
90
FR3
85
FR4
RO1
80
RO2
75
RO3
DU1
70
DU2
DU3
65
DU4
60
TU1
TU2
55
TU3
50
TU4
Early words
Later words
13
CONSONANTS MANNER
Stops + glides (vs. other manners)
100
95
FR1
Percentages
90
FR2
85
FR3
FR4
80
RO3
75
DU1
DU2
70
DU3
65
DU4
TU1
60
TU2
TU3
55
TU4
50
14
Early words
Later words
CONSONANT - PLACE
Labials + coronals
100
Percentages
FR1
95
FR2
90
FR3
FR4
85
RO1
RO2
80
RO3
75
DU1
DU2
70
DU3
65
DU4
TU1
60
TU2
Early words
Later words
TU3
TU4
15
CONSONANT - PLACE
Labials + coronals
105
100
FR1
95
Percentages
FR2
90
FR3
FR4
85
RO3
80
DU1
DU2
75
DU3
DU4
70
TU1
65
TU2
TU3
60
Early words
Later words
TU4
16
VOWELS
Vowels from the left inferior part (vs. other V.)
100
FR1
90
FR2
FR3
80
Percentages
FR4
70
RO1
60
RO2
RO3
50
DU1
DU2
40
DU3
30
DU4
20
TU1
TU2
10
TU3
0
TU4
Early words
Later words
17
VOWELS – LATER
WORDS
Vowels from the left inferior part (vs. other V.)
80
70
60
50
55,7
67,1
61,5
44,8
40
30
20
10
0
French
Romanian
Dutch
Tunisian
18
VOWELS
Vowels from the left inferior part (vs. other V.)
100
90
FR1
FR2
80
Percentages
FR3
70
FR4
60
RO3
DU1
50
DU2
DU3
40
DU4
30
TU1
20
TU2
TU3
10
TU4
0
Early words
Later words
19
RESULTS- INTER-SYLLABIC COOCC
Coronal + front
FR
RO
Labial + central
Dorsal + back
<50
>50
<50
>50
<50
>50
1.23
1.08
1.43
1.27
1.8
1.56
(1LF)
(1LF)
1.02
1.34
(1CB)
1.25
1.92
0.49
1.41
(3CB)
DU
TU
1.17
0.98
(1DF)
(1DC+1DF)
DF: 1.37
1.19
1.17
(1DF)
1.17
1.50
0.72
0.96
0.98
(3CC+1DC)
(1LB+1CB)
(4LB)
1.86
1.84
1.32
(2CB)
20
RESULTS- INTER-SYLLABIC COOCC
Coronal + front
FR
RO
Labial + central
Dorsal + back
<50
>50
<50
>50
<50
>50
1.23
1.08
1.43
1.27
1.8
1.56
(1LF)
(1LF)
1.02
1.34
(1CB)
1.25
1.92
0.49
1.41
(3CB)
DU
TU
1.17
0.98
(1DF)
(1DC+1DF)
1.19
1.17
(1DF)
1.17
1.50
0.72
0.96
0.98
(3CC+1DC)
CC:1.17
(1LB+1CB)
LB:1.18
(4LB)
LB:1.36
1.86
1.84
1.32
(2CB)
21
RESULTS- INTER-SYLLABIC COOCC
Coronal + front
FR
RO
Labial + central
Dorsal + back
<50
>50
<50
>50
<50
>50
1.23
1.08
1.43
1.27
1.8
1.56
(1LF)
(1LF)
1.02
1.34
(1CB)
1.25
1.92
0.49
1.41
(3CB)
DU
TU
1.17
0.98
(1DF)
(1DC+1DF)
1.19
1.17
(1DF)
1.17
1.50
0.72
0.96
0.98
(3CC+1DC)
CC:1.17
(1LB+1CB)
LB:1.18
(4LB)
LB:1.36
1.86
1.84
1.32
(2CB)
22
INTRA-SYLLABIC VOWEL CHANGE
Height (vs. backness)
80
FR1
FR2
70
FR3
FR4
Percentages
60
RO1
50
RO2
RO3
40
DU1
DU2
30
DU3
20
DU4
TU1
10
TU2
TU3
0
Early words
Later words
TU4
23
INTRA-SYLLABIC VOWEL CHANGE - LATER
WORDS
Height (vs. backness)
80
70
60
50
55,7
67,1
61,5
44,8
40
30
20
10
0
French
Romanian
Dutch
Tunisian
24
INTRA-SYLLABIC VOWEL CHANGE
Height (vs. backness)
80
70
FR1
FR2
Percentages
60
FR3
FR4
50
RO3
40
DU1
DU2
30
DU3
TU1
20
TU2
10
TU3
TU4
0
Early words
Later words
25
INTRA-SYLLABIC CONSONANT CHANGE
Manner (vs. place)
90
FR1
80
FR2
FR3
Percentages
70
FR4
60
RO1
RO2
50
RO3
40
DU1
DU2
30
DU3
20
DU4
TU1
10
TU2
0
TU3
early words
later words
TU4
26
INTRA-SYLLABIC CONSONANT CHANGE
Manner (vs. place)
80
Percentages
70
60
FR1
50
FR3
FR2
FR4
40
RO3
30
DU3
20
TU2
TU1
TU3
10
TU4
27
0
early words
later words
CONCLUSION
Lexical spurt ?
11 children (84.5%)
2 RO <50 words, 1 DU and 1 TU had a linear development
More easy sounds and sound structures ?
Stops +glides > others
Labials+coronals > others
Vowels from the left inferior part of vocalic space > others (80% first words, 69% later
words)
Expected CV co present except for Dutch
Slight preference for height changes in later words
No preference for manner changes over place changes
More complex sounds and sound structures after lexical spurt
Decrease of stops and glides in 77% of chidren
Decrease of labials+coronals in 69% of children
Decrease of vowels from the left inférior part of vocalic space in 77% of children
No clear changes in the CV cooccurrences between early words and later words
No decrease of height changes
No decrease of manner changes
28
WHAT NEXT ?
Influence
of articulatory constraints on early word
segmental and structural forms ?
Compare the complexity of targeted words by children with the
complexity of their actual productions of those words.
Influence
of input (word frequency and
neighborhood density) on first words ?
Correlate the words produced by children to their frequency in
CDS and Adult speech samples. We will also calculate the
correlation between word frequency and order of acquisition.
Correlate the words produced by children to their neighborhood
density in CDS and Adult speech samples.
29
MANY THANKS
TO
Barbara Davis, University of Texas at Austin, Austin, USA
Inge Zink, University of Leuven, Leuven, Belgium
The children and their families
30
MANY THANKS
TO
You for your attention
31