Accounts of mental content rooted in asymmetric dependence hold, crudely speaking, that the content of a mental representation is the cause of that representation on which all its other causes depend.1 To speak somewhat less crudely, such accounts, hereafter “AD accounts”, hold 1) that (a) any nomic relations which obtain between, on the one hand, a mental representation and, on the other hand, various causes and possible causes of that representation which are not its content, are dependent on (b) a nomic relation between that representation and its content, but 2) that the latter nomic relation is not dependent on any of the former relations. This paper argues that AD accounts are false. There is more than one AD account; there are a range of possible accounts which share that general form. Yet all of those alternatives are false. The problem is a dilemma defined by reference to three axes of difference among possible AD accounts. Any position along the first axis requires a problematic position along one of the other two. First, AD accounts may deploy nomic relations which differ in scope, which apply for instance to all individuals or merely to particular individuals and hence have in that sense relatively wide or narrow scope. Second, AD accounts may deploy different conceptions of the nature of the syntax of mental representations. Third, AD accounts may deploy different accounts of the grounds of the nomic relations which bind that syntax to target contents. I will argue that AD accounts which deploy wide scope laws cannot fit with a reasonably plausible account of the nature of the syntax of mental representations, and that AD accounts which deploy narrow scope laws cannot fit with a reasonably plausible account of the grounds of the nomic relations which bind that syntax to target contents.2 Section 1 is a brief exposition of AD accounts and of their motivation. It also sketches