In this dissertation, I examine the work involved in attempting to rebuild teacher education into an enterprise more directly focused on the practice of teaching. As a vehicle for doing so, I analyze a significant instance of teacher education reform: efforts associated with the Holmes Group at Michigan State University (MSU) in the 1980s and 1990s. The Holmes Group was a consortium of research universities that committed to a common agenda of improving schools, teaching, and teacher education, and Michigan State was a flagship site of its efforts. By investigating the experience of this large, national consortium on this particular campus, I attempt to uncover the problems of reform and to explore why the improvement of teacher education has been so elusive in the United States.

I argue that "constrained capability" is a core problem in teacher education reform and one for which leaders of improvement efforts should deliberately design. Although faculty members involved with the Holmes Group effort at Michigan State had great interest in improving teacher education and many resources for the work, collective uncertainty across the field about what teaching is and what is required to learn it and to teach it slowed and impeded their efforts. Lacking were learning opportunities that might have provided knowledge and skill essential for the work and channels through which new knowledge being developed at the time might inform the change effort. Although the faculty did alter the structure of teacher education at MSU and achieve several substantive changes, they did not fundamentally redesign the core curriculum for learning to teach. This analysis suggests that change efforts in teacher education must attend directly to enactors' extant knowledge and skill and create organizational learning mechanisms to support collective learning from and for the work of reform.