An ELT Glossary : The Input Hypothesis / The Output Hypothesis

Definition : The Input Hypothesis is one of the five hypotheses in the Monitor Model proposed by Krashen in his work in the 70s and 80s.
For Krashen, acquisition was dependent only on input. If you received the right sort of input, your brain  would automatically work on it to acquire the language. The "right sort of input" was seen by Krashen as being at the "i+1 " level. ie comprehensible at the acquirer's current competence ("i") but containing "new " language which would be the next structure to be acquired ("+1"). The Natural Order hypothesis stated that this wouldn't be just any  language that happened to crop up but would be a specific morpheme/structure which was "next on the list" for acquisition. As we have only a hazy view of what the list might be, Krashen argued that we couldn't plan to introduce the next item, but that by providing "roughly tuned input" - ie input that was comprehensible for the learner but contained a certain amount of new items, made comprehensible by the already understood language, gesture and visuals etc, the "next" item would naturally crop up. 
For Krashen  then, productive use had no part to play in acquistion - and in fact he suggested that beginners needed a "silent period" in which they weren't asked to speak but just to comprehend.

Criticisms : As with the Acquisition/Learning Hypothesis, many people have disagreed with this, including Swain who posited that negotiating meaning – ie “noticing” failures in communication and attempting to formulate language in a way that is comprehensible -  was also a factor in acquisition. She called this the Output Hypothesis

Further reading : Baker, C. and Prys Jones, S. Encyclopedia of Bilingualism and Second Language Acquisition