The so-called “species problem” has plagued evolutionary biology since before Darwin's publication of the aptly titled Origin of Species. Many biologists think the problem is just a matter of semantics; others complain that it will not be solved until we have more empirical data. Yet, we don't seem to be able to escape discussing it and teaching seminars about it. In this paper, I briefly examine the main themes of the biological and philosophical literatures on the species problem, focusing on identifying common threads as well as relevant differences. I then argue two fundamental points. First, the species problem is not primarily an empirical one, but it is rather fraught with philosophical questions that require—but cannot be settled by—empirical evidence. Second, the (dis-)solution lies in explicitly adopting Wittgenstein's idea of “family resemblance” or cluster concepts, and to consider species as an example of such concepts. This solution has several attractive features, including bringing together apparently diverging themes of discussion among biologists and philosophers. The current proposal is conceptually independent (though not incompatible) with the pluralist approach to the species problem advocated by Mishler, Donoghue, Kitcher and Dupré, which implies that distinct aspects of the species question need to be emphasized depending on the goals of the researcher. From the biological literature, the concept of species that most closely matches the philosophical discussion presented here is Templeton's cohesion idea. BioEssays 25:596–602, 2003. © 2003 Wiley Periodicals, Inc.