As usual, we assume we have a theory that is highly suspect or otherwise obviously inadequate, and we want to avoid conclusive refutation. Suppose our theory is that moral claims are true if and only if Tim Tebow asserts them (the Gator Command theory).
One reasonable way to evaluate a claim about which you are in doubt is to define the claim and then try to argue against it (no kidding!). There are innumerable ways to argue for or against a claim, but one method stands out for how cheap and easy it is, and thus is to be preferred over all others. And that is to demand a theory of the objection to our position and then counterexample (or otherwise refute) that definition.
The correctness of a theory of X is only weakly relevant to the truth of X. If you can develop a good theory of X you can help dispel doubts that X might be an illusion or that the reasons for belief in X are an error of some kind. But a failure to have a theory of X does not mean that X is not real. Indeed, theorizing about the nature of X must come after recognition of the existence or reality of X. Moreover, it is virtually impossible to give a coherent, complete theory of any philosophically interesting item X (or even many uninteresting ones).
Wittgenstein, as we know, made this point, and it doesn't apply just to the Socratic questions about piety, virtue, etc. It applies in even mundane contexts. There is no set of necessary and sufficient conditions for something to be a game, but it does not follow that there is no such thing as a game. Play the game game yourself. Try to think of anything that all and only games have in common. There are lots of necessary conditions (e.g. being an activity), and lots of sufficient conditions (e.g. being scrabble), but no set of conditions that are jointly sufficient but individually necessary. If you play this game, and find it frustrating, you may realize that even being fun is not necessary for something to be a game.
Or, consider sports. Is there anything that all and only sports have in common? Here's a definition of "sport" from freedictionary.com "Physical activity that is governed by a set of rules or customs and often engaged in competitively." (http://www.thefreedictionary.com/sport)
This is too broad. Construction work is a physical activity, and it is governed by rules. Note the open-ended part of the definition "often engaged in competitively". Just about any physical activity can be engaged in competitively. There are competitive eating contests (hence, physical activities), but eating isn't a sport.
In retrospect, it is obvious that demands for a theory will often fail to be met, not because the item in question does not exist, but because either we cannot figure out what set of conditions make it what it is, or because there is no set of such conditions. This error is obvious, yet it is often overlooked.
My first example is from Russ Shafer-Landau in his introductory article on Ethical Subjectivism in Reason and Responsibility. Shafer-Landau makes exactly this demand-for-a-theory argument against moral realism, saying that one reason against it is that there is no adequate theory of objective moral facts. But the fact that we cannot completely define and explain the existence of objective moral facts does not mean there aren't any; in fact, it's not even much evidence that there aren't any. This inability to give a theory of moral facts could simply reflect our well-known inability to give a complete and correct theory of much of anything.
Here's a case in which a philosopher defends a problematic theory. Brian McLaughlin, writing in the Encyclopedia of Philosophy, on the topic of Epiphenomenalism, the view that the mind or mental properties (or types) are causally irrelevant, notes that many critics of epiphenomenalism claim that it creates insuperable problems of reference to and knowledge of the mental. But, McLaughlin claims, such a criticism would require that a causal theory of reference or knowledge be correct, and there are problems with causal theories of reference and knowledge.
The problem with McLaughlin's claim is obvious. To say that it is necessary that some particular (concrete, abstract, whatever) be causally relevant to one's belief/utterance in order for that belief to count as knowledge/for that utterance to refer is not to claim that a causal theory of knowledge or reference is correct. I may think that theories of any such thing are impossible to attain, but I may also think that causation is necessary but not sufficient for knowledge or reference (to concrete particulars). So the lack of causation would undermine the possibility of knowledge/reference even though there would not be a causal theory of either.
I don't have a causal theory of sitting, but I'm fairly sure that a necessary condition for one to be sitting on something is that the thing you are sitting on be causally related to you. I don't have a "height" theory of basketball prowess, but I'll wager that being over 3 feet tall is a necessary condition for making it in the NBA (damn you, NBA!).
My final example comes from an article, "Teleology and the Nature of Mental States," in the American Philosophical Quarterly by Scott Sehon.
Sehon argues for a teleological view of action-explanation and, based on that theory, argues against functionalism. The discussion relevant to my point is Sehon's response to the obvious Davidsonian point that nothing can constitute an explanation of behavior unless it is causally responsible for that behavior. Only causation can turn the and of "Jamie was jealous and broke up with his girlfriend" to the because of "Jamie broke up with his girlfriend because he was jealous."
Sehon claims that in order to support the claim that causation is necessary for a correct teleological explanation for behavior, one must believe that a causal analysis of teleological explanations is at least possible. But, he claims, one cannot give such an analysis, primarily because causation is not sufficient for teleology. In other words, one’s mental state can cause a behavior in a deviant way so that we would not say that the behavior was done intentionally but the mental state nonetheless caused it.
There are two glaring problems with this argument. First, no one need claim that there is a causal analysis of intentional action in order to think that causation is necessary for that action. The second problem, related to the first, is that to say that X is necessary for Y (while not implying that there is a theory of Y in terms of X) does not mean that X is also sufficient for Y. So, showing that X is not in fact sufficient for Y has literally no bearing on the claim about X’s necessity for Y.
So, why would anyone make these arguments? Why demand a theory of whatever-it-is that is thought to render your theory inadequate? Because it’s easier than addressing the real issue. Sehon could very easily show that causation is not necessary for an explanation of action if he could give an example in which the supposed intention is not actually a cause but does nonetheless provide a correct intentional explanation. Obviously, there is no such example.
Now we can protect our theory from refutation by always demanding a theory of our objector for whatever theory we propose. So, now, when someone objects that our Gator Command theory of morality implies that Tebow would have no reason for his commands, we can say: To conclude that the Gator's commands are not based on reason is to assume that there could be an objective theory of reasons in morality against which the Gator's commands are to be judged. But given that there is no such theory, this objection provides no basis to undermine our view. Now, declare victory and go home.