If we say that something depends on something else, then what we say is true just in case there is a dependence of the something upon the something else. In particular, it is often the case that sentences expressing such a dependence consist of depends on flanked by indirect questions (IQs). The purpose of the present paper is to bring game-theoretical semantics 1 (GTS) to bear on the problem of providing a semantical analysis of depends on flanked by IQs. I shall assume that the reader is familiar with the basic ideas of GTS; the appendix of the present paper contains some of these basic ideas, stated very briefly. I say that I wish to examine the problem of "providing a semantical analysis of depends on flanked by IQs". I cannot say that I wish to examine the semantics of IQs, or the semantics of depends on. The reason for this is that, as Jaakko Hintikka (the father of GTS) is fond of remarking, GTS is an "outside-in" semantical apparatus, as opposed to the more common methods, which are "inside-out". 2 This means that whereas more familiar approaches (e.g., Montague semantics) determine the interpretation of a given sentence inductively, on the basis of the already established interpretations of its ingredients (i.e., by "putting the sentence together"), GTS proceeds by a stepwise reduction of the sentence to simpler ones. This procedure is not exactly one of "taking the sentence apart", since interpretation of the sentence is not generally referred to its ingredient sentences, but nevertheless it is "outside-in" in the sense that the first step of interpretation concerns the outermost operator of the sentence, whereas on the more common approaches it is the final step which concerns the outermost operator. The contrast is more happily captured, but still only roughly, by saying that GTS is a "diminishing complexity" method, but the more familiar methods are "increasing complexity" methods. This is not quite accurate, since either sort of method assumes, in interpreting non-atomic sentences, an assignment of values to atomic ones. Due ~o this peculiarity of GTS, we cannot expect, as a matter of principle, that a given IQ, for instance, can be assigned a meaning of its