The term religion once referred to a scrupulous devotional practice, a form of worship. Over time, its sense has expanded to include many other aspects of life that are often considered part of our culture, including cosmological orders, moral codes, taboos, promises, curses, and social structures. The resulting taxon is very wide, and there is considerable disagreement over what is and is not religion. In the last few decades, scholars have shifted away from substantive or functional definitions of religion and toward polythetic approaches that treat the concept as a family-resemblance concept and look for crisscrossing and partially overlapping features that make a practice a member of the class we call “religion.”
This debate is complex. It cuts across the disciplines of anthropology, history, philosophy, psychology, sociology, and religious studies. It raises important questions about how we define concepts and what our responsibilities are in the use of those concepts. It also highlights a fundamental challenge: How can we sort practices into categories without sacrificing the integrity of those practices?
One answer to this problem is to treat a practice as if it had a particular essence. This approach is often called realist or lexical, and it is commonly applied to the concept of religion. But there is a problem with this approach. It makes the assumption that there is a defining property that distinguishes the practices we call religion from those that do not belong to it, and that this property is something that can be discovered by careful study.