In philosophy arguments can either serve to establish a theory or tear it down. Here I am going to focus on the arguments that are meant to establish philosophical positions as true and leave discussing arguments against theories for another time. This might seem redundant, after all isn’t an argument still an argument no matter what it is about? But in philosophy at least the two are quite distinct. Arguments against theories proceed by tearing the theory down, by showing that its claims don’t measure up to some standard (whether it be accuracy or logical consistency). In contrast, arguments for theories focus on working upwards from simpler ideas so as to establish them as true. And so such arguments will proceed by different methods and be subject to different criticisms.
But before we bother getting into the details about how arguments for a theory might proceed let me come out right away and make a contentious claim, that arguments for particular theories can be unnecessary and detrimental, and that it is best to evaluate a theory by what it claims rather than how we might be led to hold that theory. Now if I was to follow my own advice perhaps I would simply skip to talking about the implications of this claim and its merits. I wish this was possible, but unfortunately my brief claim about is not actually a complete theory that can be evaluated on its own, but rather it is a consequence of a few claims that are a bit smaller in scope, and which are themselves consequences of a larger theory (which, when we examine all the details, will be one of the more acceptable ways of arguing for a philosophical claim, so this claim isn’t as self-defeating as it may first appear).
My sub-claims involve a division of arguments for philosophical theories based on the ultimate premises that they rest on, and then drawing conclusions about the value of theories that might result from such arguments and as to how valuable that argument actually is.
First we have arguments that proceed only from logical tautologies and constructed definitions. A constructed definition, in this context, is a definition that we accept as an axiom, and not one that is intended to reflect in any way to reflect the nature of the world. For example, the definition “a bachelor is an unmarried man” is a constructed definition because it “constructs” the meaning of bachelor from other terms. Generally if it doesn’t make sense to question whether the definition is a good definition then we are probably dealing with a constructed definition. It is simple enough to see that an argument that proceeds from such premises can never make claims about the real world. Which is fine if we are doing mathematics, which is perfectly happy to have all its claims be about abstract systems, but not if we are doing philosophy, which is supposed to be about the world (lest it collapse into mathematics).
Secondly we have arguments that proceed from other theories about the world that are justified because they are intuitive, or so it is claimed. Obviously the strength of the theory resulting from the argument can only be as strong as the premises, so in this case it is at best an intuitive theory itself. But just because something is intuitive doesn’t mean that it is true, so really such arguments haven’t added much to the theory in question. It still needs to be defended on the basis of being a good explanation, and surely being a good explanation is better than being intuitive. And so such arguments are redundant.
Finally we have arguments that proceed from one or more other theories about the world that are themselves thought to be good explanations. Let us first consider the sub-case where the theory in question is the logical consequence of a single such other theory about the world. If it is truly consequence of such a theory then it immediately is as justified as the other theory is, so in that sense the argument does serve a purpose. However, despite having this justification, it still needs to be evaluated on its own merits. It may be that this derived theory is in fact a poor explanation, that it runs contrary to what we know in some way. In that case the derivation serves as a reductio ad absurdum of the first theory, showing that the first theory is actually flawed in some way (if it has such problematic implications) and needs to be modified. And being a consequence of multiple theories, no matter how well those theories are established, doesn’t put the theory derived in this way on much better footing. There is always the troublesome possibility that a theory derived correctly from several theories, each of which is well known to be a good explanation, can be a bad explanation. From science the example that springs to mind is that when quantum mechanics and general relativity are put together in a logical way nonsense results. Of course in this case it is not clear which of the theories should be modified in response, it is even possible that they are good explanations in their own right, but due to their construction are simply incompatible with each other. So no matter how perfect the theories that the new theory is derived from are there is still the possibility that it may be a poor theory. Again, this means that the new theory must be evaluated in light of its merits, and cannot be endorsed based on the argument for it alone.
So, putting these cases together, it seems that it is almost always the case that a theory needs to be evaluated on its own merits, no matter how it was argued for. But we may want to make the occasional exception when we are dealing with theories that follow from theories that are well known to be good explanation, when the derived theories are very simple extensions (simply because rigorous evaluation is likely to be a waste of time). Now let me turn this back on the very claim that I made at the beginning of this section. Obviously that claim is really a simple conjunction of the claims that in each particular way a theory may be argued for it still needs to be evaluated on its own terms to some degree. This is simple enough that it seems reasonable to allow that it can be considered a good claim, so long as the claims made in each case are good. But what about those claims? Well they are really applications of a larger theory as to the nature of philosophy, that it attempts to provide good explanations, and that theories can be evaluated as to how well they explain. Given that, argument for a theory is largely irrelevant to how well the theory explains what it sets out to. So in a sense the claims here aren’t the right things to evaluate at all; what is to be evaluated is this larger theory about philosophy. And obviously that isn’t an enterprise for the current moment, although eventually it must be done for the sake of consistency, and to justify all the things that I have said follow from it. It is enough then to note here that this claim can be seen to serve as evidence that this theory about philosophy is a good one; think of the innumerable failed philosophical theories. Each had arguments supporting it that seemed sound to the authors who proposed them. Yet almost all of them have been discarded as false, not because the arguments for them were refuted (although sometimes that was done as well), but because they were shown to be unacceptable in some way, either by conflicting with experience or by implying something absurd, or were simply surpassed by better theories. In any case why they were rejected ultimately had nothing to do with the arguments for them. So the arguments for them acted at best as a rhetorical device, to get people to consider them seriously. But they are a misleading rhetorical device, because ultimately the quality of the argument has little to do with the quality of the theory. Thus recommending that we focus more on the explanatory qualities of the theory than on the argument for it seems reasonable by historical lights, since it is those qualities that ultimately matter. And so the historical progress of philosophy seems to support the more general theory about philosophy that the specific claims made here follow from, since it seems to provide evidence that at least its recommendation in this case as to how to do philosophy is reasonable.
As a final remark I would like to note that, while the argument for a theory is largely philosophically irrelevant (unless we are interested in conducting a reductio against the claims it supposedly follows from), I wouldn’t encourage giving up on them. Although the argument doesn’t serve a strictly philosophical purpose it does have practical merits. Trying to argue for a philosophical theory involves showing that the theory follows from simpler principles, each of which is reasonable. Developing philosophical theories in this way is likely to lead to better theories, because the best theories generally built around a simple core. Additionally arguing for a theory tends to feedback into the construction and explanation of the theory itself, hopefully by eliminating some of the vagueness and hand-waving.
On Philosophy 