Marriages come in a very wide variety: if the reports of anthropologists and historians are to be believed, an extraordinarily wide variety. This includes some of the more unusual forms, including marriage to the dead; to the gods; and even to plants. This does suggest that few proposed marriage relationships would require 'redefining marriage': but on the other hand, it makes giving a general theory of marriage challenging. So one issue we should face is how accepting we should be of the reports: to what extent reported 'marriages' really are marriages. This paper defends the view that almost all of these reported marriages are in fact marriages, and suggests some theoretical approaches that may be generous enough to account for this.