For more than 20 years, belief research has been based on the premise that teachers' beliefs may serve as an explanatory principle for classroom practice. This is a highly individual perspective on belief-practice relationships, one that does not seem to have been influenced by the increasingly social emphases in other parts of mathematics education research. In this article, I use the notions of context and practice to develop a locally social approach to understanding the belief-practice relationships. It is a corollary of the approach taken that the high hopes for belief research with regard to its potential impact on mathematics instruction need to be modified.