It is suggested that the static model of language which is prerequisite for contrastive analysis is inadequate in pinpointing potential difficulties in second language learning. The student learning graphemic-phonetic correspondences encounters two types of difficulties not exposed by contrastive analysis: dialectal difficulties and reading difficulties resulting from the particular strategy used by the student in learning to read his own language. Data are presented showing how certain dialect traits of Black English influence the learning of the relationship between standard German pronunciation and its written representation. Students taking part in this test were all in their first year of German study at Norfolk State College (NSC), Old Dominion University (ODU) or the University of Massachusetts (UMA). Their scores on reading/speaking task (performance) and on a hearing/writing task (perception) are the basic data of this study. Students of NSC consistently had the highest error quotients, and students at UMA consistently had the lowest. It is concluded that dialectal influence is a factor that must seriously be considered in teaching foreign languages; the dialect speaker has an additional difficulty that speakers of standard English do not experience. The number of factors causing difficulties in learning graphemic-phonetic correspondences in a foreign language can be complex, based as they are upon native-language dialect differences, reading strategies, and perhaps other variables. Therefore it is highly unlikely that contrastive analysis of the two standard languages can do more than approximate learning difficulties. An iconographic reading test devised during the study and tables showing error patterns on the tests are appended. (Author/CFM)