significant and original, but noted that "unfortunately, the part that is significant is not original, and the part that is original is not significant."
I do not intend to describe the study in detail, but several points are directly relevant to the topic of this paper. There was some skepticism at first about trying to define the dissertation in a discipline-free way. Some people doubted that they could have productive discussions about the concept of the dissertation across fields as different, for example, as chemistry and classics. As it turned out, there was agreement across all fields that the dissertations served two purposes: to allow students to demonstrate that they could do whatever people in that field did when they did research, and to produce research that constituted a (significant? original?) contribution to knowledge. These two very different skills are generally considered necessary conditions for the award of the Ph.D. degree in most countries and educational systems.
There was, however, some divergence of opinion (mainly by discipline, although occasionally by institution) about what constituted appropriate Ph.D. research. At some institutions, laboratory research on topics that rank high in national funding priorities was considered most appropriate, since they were thought to represent a kind of "peer" consensus of significance. In some areas, research in theoretical areas was valued more highly than research on practical problems. As stated by one physicist, "Ideally, a dissertation project would be self-contained, would allow individual initiative to flourish, and would address philosophically interesting and non-trivial issues." In practice, many styles of research in physics are not compatible with these "ideas." A graduate dean put it more generically: ''Whether the student works alone or on a team, the research project should be an original, theory-driven investigation characterized by rigorous methodology and capable of making a significant contribution to knowledge about the subject under study."
Historically, research has been the characterizing element of doctoral education. Whatever else was done, research was always the centerpiece. Doctoral students had to complete, as in Ireland, "[a] substantial thesis making a significant, original contribution to knowledge." 2 This was the only requirement for the degree, and much the same held true for almost all European and Asian universities. (It is interesting and perhaps revealing to note that in these countries, demonstration of the ability to do original and independent research was referred to as "preparation for an academic career.") Although some or even most students might take courses, the main tasks (as stated, for example, in Japan) "are to submit a doctoral dissertation and pass an oral examination within three years."3 In some countries, notably China and Japan, it is possible to obtain the degree without being at the university, by submitting evidence of research accomplishments acceptable to the faculty. There are variations of this in other countries, including the United States.
The idea of required courses, or of a graduate program involving both research and coursework, has been associated primarily with the United States and Canada and is gradually becoming the norm in most places. The principal reasons for this are related to the desire to provide students with a broader background so that they will be better prepared for a wider variety of career options, and to ensure some general understanding of their fields, particularly as the number of students increases. As Stuart Blume points out, "The attractiveness of the North American model has derived from the fact that it has seemed able to ensure effective and efficient training of researchers on a much greater scale than has been usual