contrastive versus non- contrastive features in african american english child speech, ages 4-12....
TRANSCRIPT
Contrastive versus Non-contrastive features in African American English child speech, ages 4-12.
Janice Jackson, Atlanta GA & Barbara Zurer Pearson, Amherst MA
http://www.umass.edu/aae/
ASHA, Philadelphia November 20, 2004
Diagnosing Disorder: NON-Contrastive Features:
Past-tense copula, auxiliary (They were dirty, it was raining.)
Variables: double-wh (Who ate what?)
Long distance movement: (When did he say he hurt himself? vs. When did he say how he hurt himself?)
Theory of Mind syntax: (What did he say he caught? vs. What did he catch? He’s thinking about his cake.)
Non-word repetition (Modified to accommodate AAE phonotactics, eg. Koo-bow-faup.)
Identifying Dialect: with Contrastive Features:
Phonological features [f] for /th/: (“maf” not “math”)
Final consonant phonotactics: *CC# (“tes” not “test”)
Morphosyntactic features: Invariant 3rd person -s: (I have, you have, he have; I wash,
you wash, he wash, etc..)
Invariant past copula/aux: (he was, we was, they was)
Note: There are many other contrastive features, but they were not as reliable as indicators in our large-scale testing.
Data Source, DELV Tryout Research Participants: Recruited from two varieties, African-American English (AAE) and
mainstream American English (MAE). Criteria for membership in the AAE group were 1) African American background (by parent self-report); and 2) residence in a community of predominantly African Americans (by census report). Criteria for membership in the MAE group: 1) predominately European American background (by parent and examiner report); and 2) from communities which did not have a predominance of AA residents. (Note: 34 of the 475 reported AA ethnicity; 82 reported Hispanic ethnicity but spoke only one language in the home.) The parent education level of 87% of the sample was high school or lower (only 3% had finished college). The lower parent education levels were especially recruited because homes with lower education levels are thought to have higher levels of AAE usage (Craig & Washington, 1998). Similarly, the language impaired (LI) population was oversampled to allow comparisons between LI children and Typically-developing (TD) children with adequate statistical power.
By language variety and clinical status:
AAE4 5 6 7 8 9 10 11 12 Total
TD F 58 68 84 5 19 14 17 9 13 287
M 51 63 59 10 14 13 20 10 14 254
LI F 11 18 22 4 10 6 12 7 8 98
M 16 27 28 12 19 9 18 7 8 144
Total 136 176 193 31 62 42 67 33 43 758
MAE4 5 6 7 8 9 10 11 12 Total
TD F 32 35 47 6 14 12 10 10 16 182
M 28 26 26 10 6 8 11 8 15 138
LI F 8 6 10 3 8 7 5 4 4 55
M 13 26 24 6 11 5 9 2 7 100
Total 81 90 107 25 39 32 35 24 42 475
Data collected by The Psychological Corporation in conjunction with NIH Contract N01-DC8-2104 to Harry Seymour, P.I.
Sample Non-contrastive item: DELV Tryout Research
from the DELV-Screening Test (Seymour, Roeper & de Villiers, 2003, copyright TPC, 2000).
These children are planning a surprise party for their favorite teacher. Everyone was going to bring some food for the party, but the boy didn’t know what to bring. He asked the woman at the grocery store, “what shall I bring my teacher? “The woman told him his teacher loved bologna, so that’s what the boy decided to bring her.Who did the boy ask what to bring?
Diagnostic: Where AAE and MAE do not contrast:
Syntactic Bootstrapping / Fast Mapping
0
5
10
15
20
4 5 6 7 8 9
Age
Average Score /21
AAEMAE
Syntactic Bootstrapping / Fast Mapping
0
5
10
15
20
4 5 6 7 8 9
Age
Average Score /21
IMPAIREDTYPICAL
But where Typical Development and Impairment DO. From DELV Tryout Research, Seymour & Pearson (Eds.), 2004.
Sample CONTRASTIVE Items DELV Tryout Research
Morphosyntax
Point to the sleeping bags.) I see sleeping bags. (point to the bed). I see a bed. (Point to the boys.) The boys always sleep in the sleeping bags, (Point to the girl). But the girl always…
PhonologyI want you to say exactly what I say:1) Trial Item: I see a man 2) I see her brushing her teeth.3) I see a bird taking a bath.
From DELV (Seymour, Roeper, & de Villiers, 2003, copyright TPC, 2000)
Identification: Where AAE and MAE do contrast:
Average number of responses characteristic of AAE, not MAE.
From DELV Tryout Research, Seymour and Pearson, 2004.
Screener Identifier Items
0
2
4
6
8
10
12
4 5 6 7 8 9
Age
Average Score /16
IMPAIRED AAE
IMPAIRED MAE
TYPICAL AAE
TYPICAL MAE
Note POOR diagnostic value (of contrastive items):
AAE-Typically Developing and Language Impaired ambiguous through age 7.
MAE-Impaired and AAE-Typically Developing ambiguous throughout age range.
Around age 8 there is a sharp downturn in the number of AAE children who are classified by the DELV Screener as “strongdifference” from MAE.
The drop is especially sharp among the TD children (and also calculated by DELV passers versus failers) and a sharp rise in MAE speakers from 7 to 8.
Few LI children even with MAE input resemble MAE speakers. Most are “strong difference” from MAE.
Another illustration of the move toward MAE
AA children classified as MAE speakers (or as code-switchers), by age and clinical status:
AAE Dialect Density by Age (TD only)
0
0.2
0.4
0.6
0.8
1
4 5 6 7 8 9
Age
Percent of children
mae
somdif
strdif
AAE Dialect Density by Age (LI only)
0
0.2
0.4
0.6
0.8
1
4 5 6 7 8 9
Age
Percent of children
mae
somdif
strdif
Similar results with larger group of items and larger age range
(from Dialect Sensitive Language Test, DSLT) :
Predictable Non-MAE Responses
by Age and Dialect
0
4
8
12
16
20
24
4 5 6 7 8 9 10 11 12
Age
Average (of 32)
TD-AAE
LI-AAE
TD-MAE
LI-MAE
Further questions (using the more extensive data from the DSLT):
Do different contrastive features have different paths of development? Zero -ed, multiple negation, plural “s”; possessive noun (the horse’s bowl)?
Do the patterns of occurrence of the set of features and its subsets vary in AA children
whose parents have different levels of education?
from different regions of the country? by gender?
Q1:(Watch relationship of AAE-TD and MAE-LI)
Instances of zero past "-ed"
0
0.2
0.4
0.6
0.8
1
1.2
1.4
4-5 6-7 8-9 10-12
Age in years
Average (of 5)
AAE-LI
MAE-LI
AAE-TD
MAE-TD
Instances of Multiple Negation
0
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
4-5 6-7 8-9 10-12
Age in years
Average (of 3)
AAE-LI
MAE-LI
AAE-TD
MAE-TD
Possessive nouns (eg. horse’s)
Instances of Invariant Possessive Noun
0
0.5
1
1.5
2
2.5
4-5 6-7 8-9 10-12
Age in years
Average (of 3)
AAE-LI
MAE-LI
AAE-TD
MAE-TD
Instances of zero 3rd -s
0
0.5
1
1.5
2
2.5
3
3.5
4
4.5
5
4-5 6-7 8-9 10-12
Age in years
Average (of 5)
AAE-LI
MAE-LI
AAE-TD
MAE-TD
Both AAE-TD and AAE-LI higher than MAE
They was/they were
Instances of Invariant "Was"
0
0.5
1
1.5
2
2.5
4-5 6-7 8-9 10-12
Age in years
AAE-LI
MAE-LI
AAE-TD
MAE-TD
Do/don’t & have/has
Instances of Invariant "Do"
0
0.5
1
1.5
2
2.5
3
3.5
4
4-5 6-7 8-9 10-12
Age in years
Average (of 4)
AAE-LI
MAE-LI
AAE-TD
MAE-TD
Instances of Invariant "Have"
0
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
4-5 6-7 8-9 10-12
Age in years
Average (of 2)
AAE-LI
MAE-LI
AAE-TD
MAE-TD
Present tense copula or aux.
Instances of zero "is" or "are" copula or auxiliary
0
0.5
1
1.5
2
2.5
4-5 6-7 8-9 10-12
Age in years
Average (of 3)
AAE-LI
MAE-LI
AAE-TD
MAE-TD
By regionAAE
4 5 6 7 8 9 10 11 12 TotalNorthCentral
42 57 57 4 26 4 13 6 9 218
Northeast 11 13 10 5 6 4 5 -- 1 55
South 73 100 120 22 30 33 45 27 32 482
West 9 6 6 -- -- 1 4 -- 1 27
Total 136 176 193 31 62 42 67 33 43 758
Non-MAE Responses by Region(AAE-TD only)
0
4
8
12
16
20
4 5 6 7 8 9 10 11 12
Age
Average
NC
NE
SO
Non-MAE Responses by Region(AAE-LI only)
0
4
8
12
16
20
24
4 5 6 7 8 9 10
Age
Average
NC
NE
SO
By parent education level
Fig. Non MAE Responses (of 34) by PED Level AAE-TD only
0
6
12
18
24
30
4 5 6 7 8 9 10 11 12
ages
1234
AAE TD4 5 6 7 8 9 10 11 12 Total
1 (< 8 yrs) 1 0 4 1 0 0 0 0 0 6
2 (<12 yrs) 27 21 32 4 4 4 5 2 2 101
3 (hs degreee) 63 83 79 6 26 19 22 13 19 330
4 (> 12 yrs) 13 25 25 4 3 4 10 4 6 94
5 (college deg) 5 2 3 0 0 0 0 0 0 10
109 131 143 15 33 27 37 19 27 541
F = 5.758, p < .001
Post hoc shows Level 2 significantly different from Levels 3 and 4.
By gender
Predictable Non-MAE Responses -by Gender (AAE-TD)
0
4
8
12
16
20
24
4 5 6 7 8 9 10 11 12
Age
Average (of 32)
F
M
Predictable Non-MAE Responses by Gender (MAE-TD)
0
4
8
12
16
20
24
4 5 6 7 8 9 10 11 12
Age
Average (of 32)
F
M
AAE4 5 6 7 8 9 10 11 12 Total
NorthCentral
42 57 57 4 26 4 13 6 9 218
Northeast 11 13 10 5 6 4 5 -- 1 55
South 73 100 120 22 30 33 45 27 32 482
West 9 6 6 -- -- 1 4 -- 1 27
Total 136 176 193 31 62 42 67 33 43 758
Both MAE and AAE (TD)
F < 1, n.s.
Summary AAE patterns diminish over time in test context, but
remain significant in over half the subjects Only minor effects of gender were found at this age
range. Only minor effects of region were found on this
selection of test items. The effect of parent education level in this population
was seen between finishing high school and not finishing high school.
Invariant verb conjugations in the present tense show the greatest difference from MAE among these features.
References Diagnostic Evaluation of Language Variation (DELV), Seymour,
Roeper, & de Villiers, Harcourt Assessments, Inc. 2003. Stockman, Newkirk, Guillory, & Seibert, Dialect Density and
Standardized Test Outcomes: Pilot Study, Asha 2003, Chicago, IL.
Seymour, H. N. & Pearson, B. Z. (Eds.) 2004, Seminars in Speech and Language, 25 (1).
Washington, J. A. & Craig, H. K. (2002). Morphosyntactic forms of AAE used by young children and their caregivers. Applied Psycholinguistics, 23, 209-231. (also Craig et al. 2003).
Acknowledgment: This work was supported by NIH Contract N01-DC-8-2104 to H. Seymour, PI with T. Roeper, J. de Villiers, P. de Villiers & B. Z. Pearson, Project Manager.