Description: The paper adresses an area of particular interest in the current european context but which has not been explored satisfactorily in the education literature: that of the evaluation of educational research. It identifies twelve recent trends in the evaluation of education research in Europe (from performance-based funding and institutionalisation of assessment to the development of peer review in eastern and central Europe) and it explores the benefits and perils of four types of assessment procedures (peer review, bibliometrics, econometrics and historiography) as they operate at a micro, mezo and macro level. Current evaluations of educational research (particularly those aimed at supporting funding decisions) often tend to operate from a rigid managerialistic perspective that employs narrow definitions of scientificity and productivity in research and largely ignores the epistemic specificity of the various fields, modes or genres of research, and the assumptions about knowledge with which they work (see 1, 2). The main thrust of the paper is to explore whether this neglected philosophical dimension could be recovered and embedded in future evaluations of research. Methodology: The paper stems from five years of research informed by ongoing debates in the philosophy of science, philosophy of educational research and policy studies. It joins philosophical reflection with qualitative and quantitative textual analysis of: official documents (at european national and organisational level) (2); the education submissions to the UK national Research Assessment Exercise 2001 (4); the contents of three refereed education journals (3); and interview transcripts with key players in the field. Conclusions: The paper argues that the recent developments in research evaluation have made clear a multiplicity of interests at stake (academia, policy, practice, public, media, market, etc.) and the tensions between the forms and criteria that each of these communities favours. The value and usefulness of each procedure/strategy for research assessment is therefore contextual. It has to do with the characteristics and goals of the process of evaluation and of the evaluator, as well as with the nature and claims of the research under evaluation and with the particularities of the publication/communication channel used. Moreover, the procedures and indicators of research evaluation are not purely technical matters, but they are also "cultural documents, imperfectly embedded in certain historical traditions" (5, p. 20), and thus they need to be read in their historicity and contextuality and in relation to the particular assumptions about knowledge and knowledge production that they embrace. Current discourses about research evaluation, fuelled by the mismatch policy-academe and by the financial, logistic and prestige-related rivalries between research communities and institutions (clearly illustrated with examples from the UK context - see 1, 3, 4), tend to play by the rules of administrative and political discourse and ignore the socio-cultural, historical and philosophical dimensions. The final part of the paper raises questions on whether these last dimensions of research assessment can be restored and whether such restoration can help a dialogue between research communities and the policy and practice arenas.