X-Authentication-Warning: Turing.Stanford.EDU: rickford owned process doing -bs Date: Wed, 19 Feb 1997 07:51:02 -0800 (PST) From: John Rickford
Thanks for sending us this. John drew my attention to it before. I was able to read the abstract a little more carefully this time, but would like to see the entire dissertation.
Quick reaction: It's terrific to have this kind of careful study. We really need more like this, and it's embarrassing that there were more studies like this a quarter century ago that there are today. Is this progress?!
The conclusion of this study that the children were all fluidly bidialectal is not matched by other contemporary studies (for instance, the Baratz 1969 paper which Ogbu referred to and had a handout on at the Berkeley session), so one would have to account for the different results. Was it in the instructions, or in the variables? Features like question embedding might produce different results than tense aspect features like be and BIN.
Also, although McWhorter pooh poohs it, there really is a question about whether today's Inner city AAVE speakers are not more isolated from SE [Std. English] speakers--Black and White--than their predecessors of a quarter century were. This is of course Labov's argument (divergence), and although I have argued against it in some places, there is certainly demographic evidence for increasing isolation, and as I told John, the kids I've worked with in EPA do seem to have little contact with SE speakers and to be more restricted to AAVE.
One striking thing is how well the highest educ/social group in this study did on all tasks. This kind of background is clearly a big plus on almost every measure--and probably has the best test taking skills overall. We know this generally of course, but the study shows it quite precisely.
On Wed, 12 Feb 1997, Susan Ervin-Tripp wrote:
>>Date: Mon, 10 Feb 1997 06:16:42 -0500 (EST) From: Lois Bloom
>>The question of "data" has come up regarding African-American children's relative facility with both standard English and the African- American English vernacular. Relevant data were reported by Elaine Lewnau in her 1973 dissertation, "Bidialectal Skills of Black Children," Teachers College, Columbia University. The dissertation committee consisted of myself, as sponsor, and Edmund Gordon at Teachers College, and Beryl Bailey, Chair of the Black and Puerto Rican Studies Dept. at Hunter College who served as consultant in preparing materials. While the research was done more than 25 years ago (and the children are now '30-something'), it is relevant to the debate and hopefully will stimulate further research of this kind.
>>Most of what follows here is quoted from the dissertation; only minor changes were made in the terms that were used then.
>>Elaine Lewnau and Lois Bloom
>>Bidialectal Skills of Black Children
>>Laura Elaine Bremer Lewnau, Ed.D.
>>Teachers College, Columbia University, 1973.
>>>From the Introduction:
>>It has long been recognized that at least two dialects are involved in the language experiences of most Black American children. The relative proficiency of the children in these dialects remains speculative although research has indicated that lower income Black children are most proficient in Black American English (BAE). The present study focused on the question of bidialectism in Black American children. The aim was to determine children's relative proficiency in BAE and General American English (GAE) as assessed by different measuring techniques.
>>>From the Procedures:
>>Three groups of children participated in the study. All were Black and lived in Black neighborhoods located in the Baton Rouge Metropolitan area. Group I included 25 children, ranging in age from 9, 3 to 12, 3 (mean = 11 yrs, 3 mos), from lower educational backgrounds. Their mothers had completed from 3 to 10 years of formal schooling, mdn = 7 yrs (of this group, 18 mothers had 8 or fewer years of schooling). Group II included 25 children (ranging in age from 9,0 to 11,10, mean = 10 yrs, 7 mos) from higher educational backgrounds, whose mothers had completed from 13 (1 year of college) to 16 yrs of schooling, mdn = 13.5 yrs (of this group, 6 had Bachelor's degrees). Children in Groups I and II lived in neighborhoods which were predominantly working class but also included middle income inhabitants, and they attended the same neighborhood public school. Group III included 19 children, ranging in age from 9,8 to 11,11 (mean = 10 yrs, 11 mos) from higher educational backgrounds. Their mothers' years of schooling ranged from 13 years to 16 years (with only 4 mothers having less than a Bachelor's degree and 8 mothers having at least a Master's degree). Children in this group lived in predominantly middle- and upper-middle income neighborhoods and attended a state university grade school. In sum, Group I were children from lower educational backgrounds (LEB); Group II children were from higher educational backgrounds and mixed economic neighborhoods (MxHEB); and Group III children were from higher educational backgrounds and higher economic neighborhoods (IsHEB).
>>Three tasks were used to judge relative proficiency in BAE and GAE. In a repetition task, each child heard 24 recorded sentences, each presented twice; they were asked to repeat exactly what they heard even if the sentences were different from what they might usually say. In an equivalence judgment task: 24 written pairs of sentence, each pair matched for content but differing in form (BAE and GAE), were given to the children to read as they heard an adult read the sentences aloud. They were asked to respond on an answer sheet by circling one of two responses to indicate whether the 2 sentences in a pair meant the same thing or meant something different. In a paraphrase task, the children each heard 2 paragraphs, one in BAE and one in GAE, each consisting of 8 conceptually simple sentences in story form, with the same number and types of linguistic features as in the other tasks. The order of presentation of BAE and GAE was random; subjects were asked to listen carefully and then to paraphrase the story; each paragraph was presented twice.
>>Four linguistic features were used in constructing the BAE and GAE materials: negation; habitual "be," remote past "been," and possessive marker. The materials in the 3 tasks were tape recorded by a female black speaker and tasks were administered by Elaine Lewnau, who grew up in a predominantly Black community in Baton Rouge.
>>>From the Results and Conclusions:
>>All data were subjected to Analyses of Variance and/or Scheffe tests, with p < .05 the accepted level of confidence. The results are presented in 20 tables and so cannot be adequately presented here. Following is a brief summary (only results that reached statistical significance are reported).
>>1. In the imitation task, children from lower educational (LEB) and mixed higher educational backgrounds (MxHEB) were better able to imitate sentences with BAE linguistic features than sentences given in GAE; however, they imitated a majority of the features from both BAE and GAE.
>>2. Children from higher economic backgrounds (IsHEB) imitated linguistic features from BAE as well as GAE; they also imitated a majority of the linguistic features from both GAE and BAE. They imitated the BAE linguistic features as well as children from the LEB and MxHEB groups, and they were superior to both groups in imitating features of GAE.
>>3. Children from IsHEB performed as well in both dialects in the imitation task, although they were better at imitating GAE sentences than were the other two groups, and children from both LEB and MxHEB were better at imitating sentences in BAE than sentences in GAE.
>>4. In the equivalence judgment task, children in the IsHEB group were better able to judge the sentence pairs than the LEB group, but the MxHEB group did not perform differently from either of the other two groups. When sentences were separated into equivalent and non- equivalent pairs for further analysis, the difference between the IsHEB and LEB groups was accounted for by the non-equivalent sentence pairs; the IsHEB group was better than the LEB group in judging the non-equivalent sentences. Children in all three groups performed significantly better when judging equivalent pairs of sentences than non-equivalent.
>>5. When paraphrasing the stories in the third task, children from LEB and MxHEB were as likely to maintain just as many features in stories presented in BAE as in GAE. They also remembered as many details from stories presented in one dialect as in the other. These two groups performed similarly to each other in maintaining features and recalling details when paraphrasing.
>>6. Children from IsHEB maintained more of the relevant linguistic features when paraphrasing stories presented in GAE than in BAE. However, they remembered as many details from stories presented in one dialect as in the other. They also maintained a similar number of BAE linguistic features in paraphrasing as the other two groups, but they maintained more GAE features when paraphrasing than did the other two groups.
>>7. The IsHEB group remembered more details while paraphrasing than did children in either of the other two groups, regardless of the dialect in which they heard the story.
>>8. All differences between the IsHEB group and the other two groups in imitation and paraphrasing were accounted for only by the GAE presentations.
>>9. The 3 groups imitated the features "be" and "been" with similar ease on the Imitation task, but these same features differentiated between the LEB and IsHEB groups on the Judgment task. Negation and Possession distinguished the 3 groups in GAE Imitation, but performance with these two features was similar in the Judgment task. In the Paraphrase task, the IsHEB children differed from both of the other groups.
>>The evidence suggests that these Black children from different educational background and economic neighborhoods were capable of bidialectism and, at least potentially, capable of code-switching.
>>Children from all 3 groups showed facility with both dialects in the paraphrase task, with no differences between them in the number of BAE features maintained when paraphrasing the BAE story. Children from the LEB and MxHEB maintained as many GAE features as BAE features in paraphrasing 2 different stories, while the children from IsHEB maintained more distinguishing features in the GAE story than in the BAE story and more than were maintained by the other 2 groups.
>>The children used 'their own words' to respond in the paraphrase task; they did not have to maintain the dialect of the story, providing at least an indirect test of code-switching. An analysis based on the number of features switched from one dialect to the other when paraphrasing revealed that the IsHEB children showed a clear preference for GAE, making more dialect switches than children in both the LEB and MxHEB groups and more switches than the other groups when paraphrasing the BAE than the GAE stories; but the groups did not differ when paraphrasing the GAE stories. Only the IsHEB group made more switches while paraphrasing the BAE story; the LEB and MxHEB children were as likely to switch from BAE to GAE as from GAE to BAE.
>>The 3 tasks each assessed different abilities and different levels of linguistic complexity (and task difficulty). Thus, results from different tasks are required when studying bidialectism and code-switching. Performance on imitation tasks, which had been the principle procedure used in assessing BAE at that time, could not be conclusive by itself for determining presence or absence of bidialectism in natural speech.
>>Finally, this study suggests that both educational background and neighborhood, as well as factors other than dialect, like difference in cultural values and experiences, influence the language children bring with them to the schools and their performance on tasks like these and other academic tasks.
>>For further details, please contact:
>>Prof. Elaine Lewnau
>>College of Education >>Baton Rouge, LA 70813 >>email: firstname.lastname@example.org