ABSTRACT
There is only one condition under which it is sensible to argue that a given definition of group solidarity is preferable to others: when it is a component of a theory whose explanations and predictions about empirical phenomena are superior to those of alternative theories. Such a theory may treat solidarity as a determinant that explains variation in other phenomena; or other factors may be invoked to explain variation in solidarity; or both. Whichever the case, no scientific value inheres in defining solidarity this way or that, for definitions alone do not explain anything. Worse, it is far too common in our discipline to find candidate definitions for given terms being evaluated using arbitrary criteria that as often as not filter out potentially useful but counter-intuitive candidates, and admit those with intuitive appeal only because they feel warmly familiar or were espoused by an “authority” or a skilled rhetorician.
