The article argues for the possibility of translation manuals having an implicit internal structure. This structure is composed of specific methodological assumptions and techniques. Using the (N)-type and (G)-type distinction developed by Fuller for the study of scientific behavior, it is shown that these are incomplete characterizations of translation manuals. A more complete characterization must involve an analysis of how the presence or absence of methodological rules influences the interpretation of specific research questions. It is further argued that while Quine's (...) original indeterminacy thesis cannot be completely rejected, in some cases it can be modified to reflect the importance of the methodological constraints. Finally, it is suggested that the more critical analysis of translation manuals will benefit the on-going debates in the study of scientific behavior. (shrink)