摘要：In the literature on conceptual change and students' intuitive preconceptions about the natural world, a long-running paradigm debate pits the Coherence perspective against the Fragmentation perspective. The Coherence perspective attributes stable, robust, context-independent concepts or intuitive theories to students, while the Fragmentation perspective describes students' reasoning in terms of the context-dependent activation of finer-grained knowledge elements that are more loosely organized than "theories." Over the past few years, this paradigm debate has entered the literature on students' personal epistemologies, their views about the nature of knowledge of knowing. Epistemology researchers have generally assumed student epistemologies to consist of robust, globally applicable cognitive structures - epistemological theories, belief systems, or stages - that fit into the Coherence perspective. However, cognitive frameworks that fit largely or wholly into the Fragmentation camp recently began to challenge this consensus. And two years ago, Smith & Wenk published the first large-N study specifically designed to experimentally distinguish the Coherence and Fragmentation perspectives. Smith & Wenk conclude that their data favor Coherence over Fragmentation. With the emergence of this paradigm debate in the epistemology literature, a crucial methodological issue becomes, What counts as evidence for Coherence at the expense of Fragmentation, or vice versa? This article attempts to further the scholarly discussion of this issue. To do so, I use a critique of Smith & Wenk's study as a launching point from which to make more general methodological arguments.
