Slide 1 ASHA by lonyoo

VIEWS: 15 PAGES: 1

									ASHA 2007
“Honor the Past – Forge the Future"

November 15-17, 2007 Boston, MA
Fast Mapping

Dialectal Influences on Fast-Mapping Skills in African-American Preschoolers – A Pilot Study
Samantha Wyatt, Ruth Huntley Bahr, and Elaine R. Silliman
Visual Stimuli
•Real photographs of 19 familiar objects were chosen from a kindergarten vocabulary list to be included in the training task, and to be modified for the experimental task.
Ant Apple Ball Boy Bunny Cat
Table 1. Pre-primer words.

Communication Sciences and Disorders
Comparison of Task Performances by Dialect
AAE Speaking Children
100%

Procedures
•The testing procedures were conducted over two days (separated by no more than a week). All activities were randomized to account for fatigue. •Experimental items were recorded by an African American woman. •Each child only received two exemplars of each phonological feature during the fast mapping task sequence, for a total of 6 new words to learn. •Using established procedures as guidelines (Dollaghan, 1985; Costa, Wilkinson, McIlvane, & Das Gracas de Souza, 2001), the children were instructed that they were going to learn some new words. •Four experimental tasks were presented on the computer, using the ECoS/Win program:
Training Task Two known objects were displayed on the computer screen while hearing the prerecorded known word (e.g., cat, dog, etc.) The child was asked to choose between two known objects. The picture of an unknown object was displayed on the computer screen while hearing the pre-recorded non-word one time (e.g., stroot). The child was asked to click on the appropriate picture, given the choice between a known and unknown object. This is similar to the exposure task, except another novel object that was not given a label, a known object (cat), and the gray square were displayed on the computer. The previously trained word or its AAE variant (i.e., skroot for stroot) was presented. The child had to select the appropriate picture. The child was asked to name the picture and his/her responses were transcribed. The child saw the original object presented with two other unfamiliar objects and was asked to identify the original object.

•Fast mapping can be generally defined as a process in which some type of information about a word is stored (such as semantic, phonological, or syntactic), based on one or two exposures to the word (Carey, 1978; Dollaghan, 1985). •The process of fast mapping is as follows: •A gross phonological representation of the word is formed •A hypothesis about the word’s meaning is formed •A phonological-semantic link between the words is created
(Gray, 2006).

80%

% Correct

60%
Rec Com p Dialect

•Fast mapping allows children to hold onto many different preliminary representations of words. Each representation will be further completed as more information about the word is added
(Carey, 1978).

Cookie Crayons Cup Dog Fish Flower

Frog Girl Heart Monkey Pizza Sun Tree

40%

20%

0%

Cluster Reduction

Skr/Str

Devoicing

MAE Speaking Children
100%

% Correct

•Fast mapping tasks can provide insight into early lexical and semantic development and it believed to be a processingindependent task (Horton-Ikard & Ellis-Weismer, 2007).

Dialect and the Fast Mapping Process
•Differences in vocabulary growth in African American children have often been attributed to varying environmental factors, including parental vocabulary use and cultural knowledge, both of which are influenced by socioeconomic status (SES) (Bradford &
Harris, 2003; Qi, Kaiser, Milan, Yzquierdo, & Hancock, 2003; Qualls, O’Brien, Blood, & Hammer, 2003; Roberts, Burchinal, & Durham, 1999). •Studies by Storkel and her colleagues (2002, 2003, 2005) suggest

•Twelve of these photographs were altered using Microsoft Paint so that they no longer resembled the original item. •An additional picture of each of the target objects was found and altered. The additional pictures were used in the comprehension task. •These pictures were used to depict the non-words. •Each visual stimulus was displayed in the center of three vertical rectangles across a 15.0-inch display
Training task screen shot

80%

60%

Rec Com p Dialect

40%

20%

0%

Recognition Task

Cluster Reduction

SkrStr

Devoicing

•AAE speaking children showed more variability across tasks. Comprehension was a little more difficult. •All participants had difficulty with the Dialect task. They seemed to be using a whole word strategy for word identification.

Production Task

% of Children

that phonology plays a role in early lexical acquisition. •Other investigators suggest that phonology does not interfere with early word learning, but may influence word use as lexical density increases (Jarvis, Merriman, Barnett, Hanba, & Haitsma, 2004). •Dialect use involves phonological features that can impact phonotactic and density features of words, but the influence of dialect upon lexical acquisition is not well known.

Comprehension Task/Dialect Task

Performance on the Production Task
100%

80%

60%

MAE AAE

Research Questions
• Does the use of dialect influence the fast mapping of novel stimuli in preschool children? • Does the degree of dialectal impact vary as a function of specific phonetic features, i.e., final consonant deletion, final consonant devoicing, and [skr] for [str] or task?

40%

Recognition Task

20%

Nonwords
Twelve nonwords were created for the fast mapping tasks. •The nonwords contained the following features of AAE (Bailey &
Thomas, 1998; Craig et al., 2003; Pollack, 2000; Stockman, 1996; Wolfram & Schilling-Estes, 1998; Washington & Craig, 2006):

0%

•A summarization of the stated requests and the criteria for a correct response for each task using the non-word dold:
Task Training Task Computer’s Request “Click on the apple, if you don’t see the apple, click on the black square.” “This is a dold. Click the screen.” Correct Response Participant clicks on the apple, if it is displayed. If it is not displayed, participant clicks on the black square. Participant clicks the screen.

Final Cluster Reduction

Skr/Str

Final Devoicing

Participants

Group MAE* AAE

N 15 5

Male 7 2

Female Mean Age SD (yrs.; mos.) (in mos.) 8 3 4;8 4;9 15.86 10.31

*Mainstream American English (MAE) speakers

•Final consonant cluster reduction (last becomes las) •Final consonant devoicing (bed becomes bet) •Skr/str (street becomes skreet) •These features were selected because of their frequency of occurrence in AAE and the potential for words incorporating these features to have ambiguous phonological representations. •The nonwords were developed by manipulating 1 or 2 phonemes from a real word (Storkel, 2003; Storkel & Morrisette, 2002). •Four different nonwords were generated for each AAE feature, for a total of 12 words.

•About 50% of the children were accurate in producing the target feature. •Children who spoke AAE experienced more difficulty producing final consonant clusters.

Conclusions
•Phonetic differences in the non-word did not deter the child from selecting the picture of the trained object during the dialect task. This was true, regardless of PPVT score (Kruskal-Wallis ANOVA = ns). •Preschool children appear to be focusing more on the whole word rather than phonetic details in their fast mapping responses, which is consistent with the segmentation (Fowler, 1991) and lexical restructuring (Metsala & Walley, 1998) hypotheses of lexical development. •While statistically non-significant, the differences in task performance from the recognition task to the comprehension task were greater for the AAE speaking children and more obvious in the skr/str and final consonant devoicing rules. Dialect use may be interfering here. •The production findings support previous fast mapping results, where children were only required to approximate the target word. Our scores were lower because we targeted the correct production of a specific phonetic feature and children often produced the dialectal variant. •The idea that fast mapping tasks are process-dependent measures and may be useful in testing children from lower SES (Horton-Ikard & EllisWeismer, 2007) was not supported. Our results indicate that fast mapping is susceptible to dialectal influences, which are language-dependent. This hypothesis merits further investigation, perhaps with older children who are better able to segment words.

Exposure Task Recognition Task

•The Diagnostic Evaluation of Language Variation (DELV, Seymour, Roeper, & de Villiers, 2003– 2005) was administered to determine if the children were dialect speakers. The results indicate the degree of language variation as either a strong variation for Mainstreamed American English (MAE), some variation from MAE, or strong variation from MAE, which would classify them as speakers of African American English (AAE). •The Peabody Picture Vocabulary Test-4 (PPVT-4, Dunn & Dunn, 2007) was administered as a measure of vocabulary recognition. Group MAE* AAE Mean Standard Score SD 105 95 13.26 6.27

Feature Final Cluster Reduction

Nonword [dold]  [dol] [casp]  [cas] [seft]  [sef] [mund]  [mun] [straet]  [skraet] [stral]  [skral] [struf] [skruf] [strub]  [skrub] [gid]  [git] [bab]  [bap] [piv]  [pif] [nug]  [nuk]

“Click on the dold. Participant chooses the If you don’t see dold (defined nonword). the dold, click on the black square.”

Backing in /str/ clusters

Comprehension “Click on the dold. Participant chooses the Task If you don’t see black square. (Rejecting the dold, click on both displayed pictures). the black square.” Dialect Task “Click on the dol. Participant chooses the If you don’t see black square (undefined the dol, click on nonword). the black square.” “What’s this?” Participant says “dold.”

Final Consonant Devoicing

Production Task


								
To top