This study of rural primary teachers' life histories, their sources of motivation and their views of teaching, is an attempt to discover the conceptual frames out of which teachers conduct their lives and work in order to improve the theory and practice of teacher development The research methodology used in the study combined grounded theoty (Glaser and Strauss, 1967) with narrative and life history approaches (Goodson, 1992; 1995; Thomas, 1992; 1995). Comparative interview data was collected from sixty eight rural primary teachers, and ten teacher developers. advisers and educators. In a second phase of data collection, the preliminary findings were reopened to teacher educators, advisers, and inspectors in four workshops, as a way of testing emergent theory. Observation and participant observation were used in these workshops. Document analysis, journal writing and the methodological tools of coding data and writing analytical memos during data analysis fonned the basis on which theory was generated. As a study in life history method, the study sought to bring to the 'story of action' a 'theory of context', placing teacher narratives within their broader relationship to schooling, society, and history (Goodson, 1992a, p.241). The particular life histories of rural African teachers in South Africa marked them as both powerless and powerful; as victims and agent'; in post-Apartheid society. On the one hand, teachers represented themselves as victims of poverty and oppression, while on the other they regarded their aspirations to join the educated middle classes as having succeeded. They were 'somebodies'. Ambivalence played itself out at many levels in the narratives. This was most evident in the way in which teacher stories about career motivation and the practice of teaching shifted between the instrumental and the relational. Status, salary, 'knowledge' and results were held in tension with idealism, nation-building and pedagogic love in their narratives. The analysis of teacher 'frames' (Barnes, 1992) thus worked towards reflecting these tensions, ambivalences and contradictions. The study makes a contribution to knowledge in three areas. Firstly, it combines the methodologies of grounded theory, narrative and life history method in a way which allows for the expression and representation of teacher 'voice'. Secondly, narrative and life history methods are applied in a new context, with rural primary teachers in South Africa; and with a new purpose in that context, that is towards a theory of teacher development. The third contribution the research makes to knowledge is in its use of a combination of narrative and matrix analysis in theorising the complex and contradictory frames of teachers. This allows narrative methodology to move beyond the telling of teacher stories to a more complex level of analysis for which the matrix diagrams provide the conceptual tools.