The sociolinguistic research to be described in this paper centers on contextual aspects of language variation. The domain of the research can be viewed as more encompassing than recent work in generative linguistics, which focuses on the sentence and its components as the units of analysis. On the other hand, it is more modest than the program for the study of communicative competence set forth by Hymes (1967). His framework for the study of the ethnography of speaking includes contextual and performance factors such as the following: setting (time and place); participants; intentions; topic; tone (e.g., mocking); the code and channel of transmission; norms of interaction (e.g., loudness, when to interrupt); and types of speech acts and events (e.g., conversations, curses, prayers). The domain of the present investigation takes the utterance of a sentence in a particular context and in a particular manner as the unit of analysis. Unlike Hymes', this approach restricts itself to verbal material as well as to single utterances. In that sense it also differs from, and, in fact, complements, the wider scope of the enthnomethodologists in conversational analysis (Sacks, Schegloff and Jefferson, 1974) or the research of others on social interaction (e.g., Duncan, 1973).This research utilizes structural aspects of language (i.e., the requesting strategies a speaker uses) as a window to social behavior in the same way Vygotsky conceived of language as a window to the mind. The notion of a requesting strategy is defined here by the semantic form of the utterance. Examples (1) through (5) provide a representative sample of the range of requesting strategies in English. The list is by no means exhaustive, since more than twenty such strategies have been identified for English (Fräser, 1977).(1) Would you lend me 35 cents?(2) Will you lend me 35 cents?(3) Can you lend me 35 cents? (4) I want 35 cents. (5) Give me 35 cents now.