American Dialect Society Anaheim, CA Jan. 2006 The Impact of Dialect on the Rate and Order of Phonological Development Shelley L. Velleman*, Barbara Zurer.

American Dialect SocietyAnaheim, CA Jan. 2006

The Impact of Dialect on the Rate and Order

of Phonological Development

Shelley L. Velleman*, Barbara Zurer Pearson*, Timothy J. Bryant+ & Tiffany Charko@

*University of Massachusetts-Amherst+University of New Hampshire

@Agawam Public Schools

Research supported by

NIH contract N01-DC-8-2104* and

NSF Award BCS-0318135

*webpage:www.umass.edu/aae

Contact for information: [email protected]

With special thanks to

The Psychological CorporationThe Psychological Corporation,

who collected the data,

a host of dedicated graduate and

undergraduate students, and

our colleagues in the

UMass NIH Working Groups on AAE.

AAE: African American EnglishAAE: African American English

• Also called African American Vernacular English (AAVE), Black English, Ebonics, etc.– Spoken by many Blacks in the U.S. – Pronunciation in some respects similar to Southern

American English– Pronunciation and grammar in some respects similar to

West African languages– Shares many characteristics with other Creole English

dialects spoken by Blacks• Stigmatized in the U.S.• Children who speak AAE are often referred for special

education or speech-language pathology services

TerminologyWe are comparing MAE learners to “AAE

learners” BUTAAE learners are actually learning both

dialects; AAE is their 1st dialect, so we are making the

assumption that it will have the most impact on the order and rate of their phonological development

Terminology, cont.

“Match”: child’s form matches adult MAE form

“Non-match”: child’s form does not match adult MAE match

Terminology, cont.:

CONTRASTIVE CONTRASTIVE ELEMENTSELEMENTS

• Specific to AAE

• NOT characteristic of MAE

NONCONTRASTIVENONCONTRASTIVE ELEMENTSELEMENTS

• Common to AAE

and

MAE

Seymour & Seymour, 1977

Same phonemic repertoire (with possible exception of voiced “th”) but

• Interdental fricatives replaced by labiodentals or alveolars, depending on context

• Postvocalic liquids: Vowelized, absent; /r/ hyperarticulated (varies geographically)

• Final obstruents more weakened (devoiced, glottalized), especially alveolars

• str-, “shr-” skr- (lexical?)

Key Segmental Features of AAE Predicted to be Contrastive

Key Phonotactic Features of AAEPredicted to be Contrastive

Same structural repertoire but• Weak syllable deletion from iambics (or “stress

shift” to trochaic)• Final consonant clusters reduced at higher rate,

especially /___##C• Final obstruents and nasals omitted more

frequently, especially alveolars, especially /____##C

• Avoid sonority violations (lexical “metathesis”, very stigmatized)

Thus, phonotactic structures tend to be less complex

Impact of Ambient LanguageImpact of Ambient Language

Previous cross-linguistic research has shown that frequency of occurrence impacts rate and order of phonological acquisition:– Kehoe & Lleo, 2002– Demuth, 2002– Roark & Demuth, 2000– Pearson et al., 1995– Boysson-Bardies & Vihman, 1991

Research Question

What is the impact on rate and order of phonological development of learning two dialects that differ primarily with respect to frequency of occurrence, especially of complex phonotactic structures?

Hypothesis 1

• Frequency will impact rate and order of acquisition even in two dialects with the same phonemic and phonotactic inventoriesNon-contrastive elements same exposure in

both equivalent mastery in bothContrastive elements less exposure in one

dialect later mastery in that dialect

Hypothesis 2

• Phonotactic and segmental frequency will interactMost segments will be contrastive only

in marked environmentsFor AAE, only interdental fricatives will

be contrastive in all environments, marked and unmarked

Hypothesis 3• In the dialect with less exposure to more

complex phonotactic structures (AAE), phonetic development will outpace phonotactic development (in comparison to MAE).

– AAE will have more phonotactic non-matches to MAE than segmental; MAE vice versa

Study sample:

Children tested by The Psychological Corporation as part of the standardization process for the DELV.

Female Male Total

AAE 286 251 537

MAE 182 135 317

Total 468 386 854

Other characteristics of the sample:• Selection criteria included demographics of

community of residence (predominantly African American vs. European American)

• Region: South (60%), North Central (25%), Northeast (6%), West (9%)

• Parent Education Level 77% ≤ HS

(overselected because AAE usage is higher in lower-income homes)

Overall Match Score by Region and Dialect

92

96

100

104

108

112

AAE MAE

# of matches (of 132)

WE

SO

NC

NE

Format• Sentence repetition• Target embedded in carrier phrase “I see: a mask; ..that fish breathe under water; ..a dentist”

66 words, each containing 2 segmental targets = 132 targets44 Contrastive: 88 Non-contrastiveCopyright 2000 The Psychological Corporation

Targets

Singletons Clusters*

Initial 21 19

Medial 1 17

Final 19 15

Total 41 51

*Non-morphological cluster targets

Singleton stimuli

Initial 21 types

31tokens

All (but /p/) Non Contrastive

Final 19 types

33 tokens

All Contrastive

Cluster Stimuli Types,

Tokens

Non-Contrastive Contrastive

CC initial 16, 21 br-, dr-, kr-, fr-, pr-, tr-, gr-,sm-, st-, sk-, sp-, kr-, kl-, gl

r-, “shr-”

CCC initial 3, 4 skr-, spl- str-

CC medial 14, 16 -nd-, -nt-, -st-, -l-,

-ld- -fr-, -sk-, -kt-,

-ft-, -br-, -t-, -rp-

CCC medial 4, 4 -r$d-, -$br-,

-n$tr-, -$str-

($=syllable break)

CC final 15, 19 -st, -sk, -r, -rd, -rt,

-rl, -rs, -lt, -nt, -ks,

-mp, -ft, -ld, -lt, -rf

Coding• Match to MAE target = 1• Nonmatch = 0

Phonetic (segmental) non-match:– Substitution– Distortion

Phonotactic non-match:– Omission (consonant or syllable)– Epenthesis (consonant or syllable)– Movement (consonant or syllable)

Results for elements predicted to be contrastive

Contrastive Elements by Age and Dialect

42

44

46

48

50

52

54

56

58

4 5 6 8 10 12

Age in Years

Number of Matches

AAEMAE

Results for elements predicted to be non-contrastive

Non-contrastive Elements by Age and Dialect

42

44

46

48

50

52

54

56

58

4 5 6 8 10 12

Age in Years

Number of Matches

AAE

MAE

H1: Comparison of non-matches per child by position

AAE MAE

Initial(non-contrastive)

2.81 2.86

Final(contrastive)

3.82 1.76

Non-contrastive Singleton Consonants by Age and Dialect

25

27

29

31

4 5 6 8 10 12

Age in Years

Matches (of 31 tokens)

AAE

MAE

Phonetic order of acquisition: Initial consonants

Initial Consonants

MAE Age of Mastery

AAE Age of Mastery

d 4 5

r, s 6 4

Voiced “th” 8 ≥12

Dialects differ at p=.014 but p=.952 without voiced “th”.

All other initial consonants, including voiceless “th”, acquired at the same time in both dialect groups.

Initial /r/ substitutions by age and dialect

p = .034 (chi-square)

Production of Final Consonants

Score for Final Singleton Consonants by Age and Dialect

0.5

0.6

0.7

0.8

0.9

1

4 5 6 7-89-1011-12

Percent of matches

AAEMAE

Phonetic order of acquisition: final C’s

Final Consonants

MAE age of mastery (90%)

AAE age of mastery (90%)

b, “j”, l 4 5

k, g, v 4 6

d, t 4 8

s, z 6 4

Voiced, voiceless “th”

10 ≥12

p <.0001 for age and dialect, even without voiced “th”.

Unexpected result: Non-morphological final /s, z/ mastered earlier by AAE learners

Production of Initial Clusters (N.S.)

Initial Clusters MAE Age of Mastery

AAE Age of Mastery

tr- 4 5

kl-, pl 5 4

kr- 6 4

gr-, pr-, sp-, st-

6 5

skr- 8 6

“thr-” (vless) 8 10

“shr-” 8 12

str- 8 ≥12

Initial Cluster Dialect Differences

Reminder: In AAE•str- skr e.g., [skrit] street•“shr-” skr- e.g., [skrImp] shrimp(Lexical?)

Note: Even in contrastive clusters such as these, /r/ itself is relatively preserved.

Production of Final Clusters (p<.0001)Final Cluster MAE age of

masteryAAE age of mastery

-mp 4 4

-ks, -”ng”k, -rl 4 5

-rf, -nt 4 6, 8

-ld 4 10

-lt 4 >12

-rd 5 8

-rs 6 5

-rt 6 10

All others various ≤10 >12

H2 Results: Some elements contrastive in marked positions only

Singleton /n/ by dialect and position

0.88

0.9

0.92

0.94

0.96

0.98

1

Initial Final

% MAE matches

AAE

MAE

Elements contrastive in marked positions

Singleton /d/ by dialect and position

0.88

0.9

0.92

0.94

0.96

0.98

1

Initial Final

% MAE matches

AAE

MAE

/d/: less frequent in final position in AAE(glottalized, devoiced, omitted)

• 4 years difference between AAE & MAE in final position• 1 year difference between AAE & MAE in initial position

• More vulnerable in other marked contexts, e.g., more frequent non-match in unstressed syllables even in initial position (dusty vs. destroy)

InitialFinal

AAEMAE0.86

0.880.9

0.920.940.960.98

1

% match to MAE

Singleton /d/ by dialect and position

AAEMAE

H2 Results: Some elements contrastive in all positions

Singleton /eth/ by dialect and position

0.4

0.5

0.6

0.7

0.8

0.9

1

Initial Medial Final

% MAE matches

AAE

MAE

H3: Phonotactic vs. Segmental Non-Matches

Phonotactic Non-Match

Segmental Non-Match

AAE 2534 (49%) 2615 (51%)

MAE 514 (33%) 1046 (67%)

But that includes final consonants and final clusters, both of which tend to be omitted -- no surprise.

What if we focus our analysis only on initial clusters, which are:

• Not significantly different in % mismatches by dialect

• Not yet mastered by either group?

p<.0001

0

0.2

0.4

0.6

0.8

1

1.2

% of each type

AAE MAE AAE MAE AAE MAE

4 5 6

Mismatch Types by Age and Dialect(Initial Position)

subst

phon

Initial cluster mismatch types

Summary1. Certain segments (e.g., voiced “th”) and

positions (e.g., ___#) are contrastive between dialects

2. A deficit model is inappropriate: Frequencies of occurrence in the dialect influence order of acquisition

• MAE speakers acquire certain phonemes (t, d, interdentals) ahead of AAE speakers

• AAE speakers acquire certain phonemes (s, r) ahead of MAE speakers

Summary, cont.3. There are interactions between

phonotactic and segmental frequency effects (e.g., /d/)

4. Focus on learning complex phonotactics delays acquisition of more difficult segments (MAE); decreased attention to complex phonotactics lowers age of acquisition of later segments, even in more challenging contexts (AAE)

References• Boysson-Bardies, B., & Vihman, M. M. (1991). Adaptation to language: Evidence from

babbling and first words in four languages. Language, 67, 297-319.• Charko, T. & Velleman, S. (2003, July). The influence of dialect of children’s phonotactic

constraint rankings (ND children). Poster presented at the Child Phonology Conference, UBC.

• Craig, H. K. & Washington,J. A. (2004). Grade-related changes in the production of African American English. JSHR, 47(2), 450-463.

• Kehoe, M., & Lleo, C. (2002). The acquisition of syllable types in monolingual and bilingual German and Spanish children. Paper presented at the Boston University Conference on Language Development 27, Boston, MA.

• Pearson, B. Z., Navarro, A. M., & Gathercole, V. M. (1995). Assessment of phonetic differentiation in bilingual learning infants, 18 to 30 months. In D. MacLaughlin & S. McEwen (Eds.), Proceedings of the 20th Annual Boston University Conference on Language Development (pp. 427-438). Somerville, MA: Cascadilla Press.

• Roark, B., & Demuth, K. (2000). Prosodic constraints and the learner's environment: A corpus study. In S. C. Howell, S. A. Fish & T. Keith-Lucas (Eds.), Proceedings of the 24th Annual Boston University Conference on Language Development (pp. 597-608). Somerville, MA: Cascadilla Press.

• Seymour, H.N. & Pearson, B. Z. (Eds.), 2004. Evaluating language variation: Distinguishing dialect and development from disorder. Seminars in Speech and Language, 25 (1),

• Seymour, H. N., Roeper, T., & de Villiers, J. (2003, 2005) Diagnostic Evaluation of Language Variation DELV, Screening Test and DELV-Norm Referenced. The Psychological Corp., San Antonio, TX.

Questions?

[email protected]