I'm thinking of using this as my MA thesis, so I've taken it down.
The problem of how best to explain the truth-making relation is a vexed one for truth-maker theory. As Raimi points out in his introductory survey, theories of this relation face four main difficulties:
An adequate definition of the truth-maker relation must satisfy at least four conditions. It should not fall victim to any of the following problems: (i) the problem of counterintuitive truth-makers; (ii) the problem of excluded truth-makers; (iii) the problem of missing truth-makers; and (iv) the problem of unnecessary truth-makers. A definition falls victim to the first problem if it classifies as truth-makers for a certain proposition entities that are intuitively not truth-makers for this proposition. It falls victim to the second problem if it fails to classify as a truth-maker for a certain proposition an entity that intuitively is a truth-maker for this proposition. It falls victim to the third problem if it fails to account for any truth-maker for a certain proposition that intuitively has a truth-maker. Finally, it falls victim to the fourth problem if it classifies a truth-maker for a proposition that intuitively has no truth-maker. (Truth and Truth-Making, pp 13-4)
In this paper I propose an account which I hope will not succumb to any of these problems. Section 2 sketches a couple of the major accounts that have been given of the truth-making relation, and explains what their problems are. In Section 3 I explain the basic ideas behind my own proposal, where I introduce the idea of reference-making, and use it to account for the idea of truth-making for subject-predicate sentences, taking a truth-maker to be a reference-maker for a sentence. In Section 4, I give a quasi-formal account of how it can be applied to truth-functional compounds, quantified sentences, and modal sentences. Section 5 gives a reply to Greg Restall’s arguments that logical considerations lead quickly to the trivialization of the truth-making relation: that everything is a truth-maker for every true truth-bearer. I show that this does not hold for my approach, and in the process show how it avoids problem (i). Next, in Section 6 I discuss some of its philosophical implications. Then, in Section 7, I show how my account, contrary to first appearance, can be tweaked to avoid truth value gaps. Section 8 answers objections to my views. Finally, I conclude the paper in Section 9.