A exists, for if it did not exist, no good would exist without evil, no greatness without littleness, no eternity without beginning; and the same thing would be true of perfection, which would not exist without imperfection, nor would justice exist without injustice, nor nobility without baseness, and so on for the others. But since goodness, greatness, etc. are concordant with being, and their opposites with privation, therefore one should not doubt that A exists, nor should one deny the existence in it of goodness, greatness, etc.; because if there were no goodness, greatness etc. in A, then it would be impossible for A to exist, since this existence is in accord with no being in which there is not immense goodness, greatness, etc., and in which, through bonification, there is no goodness in greatness, nor, through magnification, any greatness in goodness, and so on for the rest, which bonification is so great, etc., and which magnification is so good, etc., that it could only accord with a being that is A.
[Ramon Llull, Ars Demonstrativa, Distinction II, Part II, in Selected Works of Ramon Llull (1232-1316), Volume I, Bonner, ed. and tr., Princeton University Press (Princeton: 1985), p. 356.]
A is God. This is generally seen as an ontological argument, and allowing for the classificational mess that is the category 'ontological argument', this is probably right as intended; that is, it is an a priori argument for God's existence of some kind. However, I think there is a more fruitful way to think of it (and indeed, of most of Llull's arguments), which consists in seeing them as a sort of limit-converging transcendental argument argument concerned with the conditions of coherence for our thought about the world. That is to say, Llull's idea is that in remembering, understanding, and willing the world, we have these various unified intelligible domains -- for instance, insofar as remembering, understanding, and willing go, we find ourselves concerned with things that are able to be good (bonus), that are actually being good (bonificans, bonificating), and so forth. This only makes coherent sense if there is something that unifies these in some way; this is the dignitas (the axiom or principle), in this case goodness or bonitas. And so it goes, he holds, for a bunch of other things: greatness, eternity, power, wisdom, will (or love), virtue, truth, glory, perfection, justice, generosity, simplicity, nobility, mercy, and dominion. These are the dignitates that unify entire domains of thinking (remembering, understanding, willing) about and in the world. But how do these domains of thinking relate to each other? They can't be regarded as wholly separate, nor can they be regarded as inconsistent, as if entire domains of our thought were establishing that other domains of thought were wrong. The dignitates of these domains, at least, have to 'concord with being', to be the principles of domains of thinking about what is, and as such, have to have some kind of coherence with each other; since they concord with being, their opposites are privations of some kind. If you can never have being that is good without its opposite, then goodness (and thus the entire domain of thought of which it is the principle) has defective concordance with being. If goodness is inherently defective, however, that means that it is going to be defective in its coherence with others; for instance, if the greatest good is also defective in good, that means that its goodness has to be defective in greatness, or eternity (i.e., duration), or power, etc. The dignitates would then not have a complete coherence with each other, and the domains of thought that they unify would be inconsistent with each other.
Thus A is not God as such, but God specifically as principle of coherence among the dignitates, the principle that must exist if, for instance, the goodness-domain of our interaction with the world is to cohere completely with the power-domain, the truth-domain, etc. In the wheel of the dignitates, and of all our interaction with things, A is the central point that makes the wheel a wheel rather than a mess. If there is no A, Llull wants to say, then our thought about the world is ineliminably inconsistent and incoherent. It's not just a matter of goodness fitting imperfectly with wisdom, but with everything we think and do with regard to things having conflicting aspects. But in order to be the principle of coherence for the dignitates, A must be the limit case of each, in which all of the dignitates have perfect concordance: A is where goodness is great, eternal, powerful, wise, etc., and greatness is good, eternal, powerful, wise, etc., and power is good, great, eternal, wise, etc., and so on and so forth. All the domains of thought cohere, and are only able to cohere, because they all converge on A, the point where all their unifying dignitates are in complete concordance. Since incoherence of the domains of thought gives us endless contradictions and inconsistencies, A must exist.
In this sense, Llull's argument is somewhat like Kantian and Neokantian moral arguments: in those arguments we have a possible conflict, and practical reason authorizes postulating God to prevent the conflict, in those cases of some aspect of the moral domain and some aspect of the natural domain, from being insuperable in practice. Llull has a very different metaphysics and epistemology; there are many more domains than the moral and the natural; he takes the argument to be demonstrative for reality rather than a postulate for practice because the conflict and inconsistency he is trying to avoid is not merely practical; but the general structure is analogous. The dignitates posit A as a condition of their coherence, both in themselves and of their domains with each other; the fundamental conditions of remembering, understanding, and willing cannot contradict each other. In another way, it is like Aquinas's Fourth Way; indeed, while Aquinas wouldn't himself put it quite the same way, the Fourth Way has an explicit step that corresponds to Llull's 'concordance with being' point. In fact, I think the best way locate Llull's argument in 'argument space' is as an intermediate form between something like the Fourth Way and something like the Kantian moral argument.
