Before describing our taxonomy in more detail, it is important to note two of its limitations. First, as with any taxonomy, ours can be accused of collaps- ing meaningful distinctions in the interest of parsi- mony. After all, taxonomies—like theories—are at- tempts to eliminate some of the complexity found in the real world (Bacharach, 1989). Our intention was not to capture every nuance of theory building and theory testing, but rather to create a tool that could be used to chart trends in theoretical contri- butions over time. Second, Figure 1 only captures what empirical articles are intended to do—it does not capture how well they actually do it. One could conceive of a third axis that captures how interest- ing a new construct is, how much a new relation- ship adds to a literature, how rigorously a theory is tested, or the degree to which the mere writing of a paper makes a contribution in and of itself, apart