IN RECENT DISCUSSIONS OF THE PROBLEM OF LANGUAGE AND SOCIETY, BERNSTEIN (1961, 1964) EXPLORES THE HYPOTHESIS THAT SOCIAL RELATIONSHIPS ACT AS INTERVENING VARIABLES BETWEEN LINGUISTIC STRUCTURES AND THEIR REALIZATION IN SPEECH. HIS FORMULATION SUGGESTS THAT THE ANTHROPOLOGISTS' ANALYSIS OF SOCIAL CONSTRAINTS GOVERNING INTERPERSONAL RELATIONSHIPS MAY BE UTILIZED IN THE INTERPRETATION OF VERBAL PERFORMANCES. THIS PAPER ATTEMPTS TO CLARIFY THE SOCIAL AND LINGUISTIC FACTORS INVOLVED IN THE COMMUNICATION PROCESS AND TO TEST BERNSTEIN'S HYPOTHESIS BY SHOWING THAT SPEAKER'S SELECTION AMONG SEMANTICALLY, GRAMMATICALLY, AND PHONOLOGICALLY PERMISSIBLE ALTERNATES OCCURRING IN CONVERSATION SEQUENCES RECORDED IN NATURAL GROUPS IS BOTH PATTERNED AND PREDICTABLE ON THE BASIS OF CERTAIN FEATURES OF THE LOCAL SOCIAL SYSTEM. IN OTHER WORDS, GIVEN A PARTICULAR AGGREGATE OF PEOPLE ENGAGED IN REGULAR FACE TO FACE INTERACTION, AND GIVEN SOME KNOWLEDGE OF THE SPEAKERS' LINGUISTIC REPERTOIRE (GUMPERZ 1964), THE AUTHORS RELATE THE STRUCTURE OF THAT REPERTOIRE TO THE VERBAL BEHAVIOR OF MEMBERS OF THE COMMUNITY IN PARTICULAR SITUATIONS. (DATA ON VERBAL INTERACTION DERIVES FROM APPROXIMATELY TWO MONTHS' FIELDWORK IN HEMNESBERGET, A SMALL COMMERCIAL AND INDUSTRIAL TOWN OF ABOUT 1300 INHABITANTS IN THE CENTER OF THE RANA FJORD, CLOSE TO THE ARCTIC CIRCLE IN NORTHERN NORWAY.) THIS PAPER IS TO BE PUBLISHED IN "DIRECTIONS IN SOCIOLINGUISTICS," (JOHN J. GUMPERZ AND DELL HYMES, EDITORS) BY HOLT, RINEHART, AND WINSTON IN 1968. (AUTHOR/AMM)