A fundamental assumption in some recently developed language teaching methodologies is that there should be a silent period at the beginning of second language instruction during which students do not speak the target language but develop and demonstrate comprehension in other ways. A number of studies and experiments have attempted to illustrate that language instruction is more effective when it begins with the silent period. However, data from natural language acquisition studies do not conclusively support the existence of a silent period in natural learning contexts, and the results of pedagogical research do not justify the general rejection of early target language output in the language classroom. An alternative that would satisfy the need of some students to communicate and the desire of others to remain silent would have an initial period of reduced output, limiting the pressure on students to communicate and limiting the types of communication used. This approach would accommodate individual differences but would not allow a prescribed progression through instructional materials. Instead, it would require teacher sensitivity and curricular flexibility. (MSE)