Howard (1992) defines concepts as the information that a person has about a category, and argues for an eclectic theory of concepts on the basis of this definition. We argue that this definition is unacceptable and hence that eclecticism does not follow. First, the definition is circular as it stands. Secondly, when it is modified to avoid circularity, it implies conceptual holism, according to which concepts are not useful explanatory constructs in psychology. Thirdly, we argue that Howard's argument relies essentially on this unacceptable definition: alternative accounts of concepts, namely categorisational or representational views, do not support it. Having countered the argument for eclecticism, we then argue against it directly on methodological grounds.