Context, interest relativity and the sorites Jason Stanley According to what I will call a contextualist solution to the sorites paradox, vague terms are context-sensitive, and one can give a convincing dissolu- tion of the sorites paradox in terms of this context-dependency. The reason, according to the contextualist, that precise boundaries for expressions like ‘heap’ or ‘tall for a basketball player’ are so difficult to detect is that when two entities are sufficiently similar (or saliently similar), we tend to shift the interpretation of the vague expression so that if one counts as falling in the extension of the property expressed by that expression, so does the other. As a consequence, when we look for the boundary of the extension of a vague expression in its penumbra, our very looking has the effect of changing the interpretation of the vague expression so that the boundary is not where we are looking. This accounts for the persuasive force of sorites arguments. Suppose we are presented with fifty piles each of which has one grain less than the pile to its left. On the far left is a pile we are strongly inclined to call a heap when presented alone. This pile is in the definite extension of ‘heap’. On the far right is a pile we are slightly inclined not to call a heap when presented alone. It is towards the end of the penumbra of ‘heap’. But starting with the left-most pile (pile 1), we may make progressive judgements of the form: (a) If that pile 1 is a heap, then pile 2 is a heap. (b) If pile 2 is a heap, then pile 3 is a heap. (c) Etc. (d) Etc. (n) If pile 49 is a heap, then pile 50 is a heap. Each of these conditionals would be one to which we would assent. By agreeing to this succession of conditionals, we eventually come to pile number 50, which we now would strongly feel inclined to count as a heap. The explanation for this, according to the contextualist, is that as we are presented with each conditional, the two piles are sufficiently similar so as to cause us to interpret ‘heap’ in such a way to count one as a heap if and only if the other is a heap. This has the effect of changing ANALYSIS 63.4 OCTOBER 2003 Analysis 63.4, October 2003, pp. 269–80. © Jason Stanley