PATN is a design for a machine problem solver which uses an augmented transition network (ATN) to represent planning knowledge. In order to explore PATN's potential as a theory of human problem solving, a linguistic approach to protocol analysis is presented. An interpretation of a protocol is taken to be a parse tree supplemented by semantic and pragmatic annotation attached to various nodes. This paradigm has implications for constructing a cognitive model of the individual and designing computerized tutors. Manual protocol analysis is tedious and informal; hence the design for PAZATN, an automatic protocol analyzer, is presented. PAZATN uses PATN as a generator for possible interpretations of the protocol, with bottom-up evidence biasing PATN toward plans which are likely to match the data. PAZATN is a domain independent framework for constructing specialized protocol analyzers. To apply PAZATN to a particular task domain, event specialists (ESP's) are needed which embody syntactically organized domain knowledge. ESP's for the LOGO graphics programming domain are defined and PAZATN's operation is hand-simulated on an elementary protocol for this domain. Forty-one references are listed. (Author/LLS)