I often tell my students (particularly my introductory, non-philosophy-major students) things about philosophy that I do not strictly believe. For example:
1. I tell them that nothing counts as philosophy unless there is a definite argument, with clearly identified premises supporting a conclusion. Just stating your views without defending them is autobiography, not philosophy; even if you state them in a compelling and attractive way, without an argument it’s just rhetoric and still not philosophy.
Yet I don’t really think that’s exceptionlessly true. Some ways of putting forward a view in a compelling and attractive way amount to drawing the reader’s attention to the right sorts of consideration and so getting them to see what they otherwise might not. In my book that counts as giving reasons even if there’s no formally identifiable argument. Thus Nietzsche, Kierkegaard, and Wittgenstein, for example, all count by my lights as doing philosophy even when arguments in the literal sense are thin on the ground; and Rand’s novels and Sartre’s plays would still count as philosophy even if you cut out the explicit arguments in the speeches.
2. I also tell my students that in criticising an argument it is not appropriate simply to attack the conclusion; they have to identify a false premise or fallacious inference kn the argument. But I don’t believe that either. The Moorean strategy of showing that an argument must be unsound if it has an unacceptable conclusion is perfectly legitimate; you can know that an argument is unsound even if you haven’ yet identified at what step it goes wrong.
3. Finally, when my students become antsy about the use of fantastic or science-fictional examples in philosophy (such as Plato’s Ring of Gyges or Thomson’s violinist) I tell them that the unrealistic nature of an example cannot be an objection to an argument that uses it unless the argument falsely states that the example is realistic. Thus when the unrealistic example is confined to the antecedent of a counterfactual conditional (as in “If people seeds could float through your window and start growing in your carpet, it would be morally permissible not to water them,” to pick another example from Thomson), the fact that the antecedent is unlikely or even impossible is no objection to the proposition as a whole, whose truth-value we can still assess.
Cicero makes the same point:
If we have made even the least proficiency in philosophy, we ought to be thoroughly persuaded that, even though we could escape the view of all gods and men, still nothing ought to be done by us avariciously, nothing unjustly, nothing lustfully, nothing extravagantly … For this reason Plato introduces the well-known story of Gyges, who, when the ground had caved away on account of heavy rains, passed down into the opening, and saw, as the story goes, a brazen horse with doors in his sides. Opening these doors, he saw a man of unusual size, with a gold ring on his finger, which drawing off, he put it on his own finger (he was a shepherd in the king’s service), and then repaired to the company of the shepherds. There, as often as he turned the part of the ring where the stone was set to the palm of his hand, he became invisible, yet himself saw everything; and was again visible when he restored the ring to its proper place. Then, availing himself of the advantage which the ring gave him, he committed adultery with the queen, and by her assistance killed the king his master, and removed by death those whom he thought in his way. Nor could any one see him in connection with these crimes. By means of the ring he in a short time became king of Lydia. Now if a wise man had this ring, he would not think himself any more at liberty to do wrong than if he had it not; for it is right things, not hidden things, that are sought by good men. Here, however, certain philosophers [Cicero means the Epicureans], by no means ill-disposed, yet somewhat deficient in acuteness, say that this is only a fictitious and imaginary story that Plato has told – as though, forsooth, he asserted that such a thing took place or could have taken place. The meaning of this ring and of this example is as follows: If no one would ever know, if no one would ever suspect, when you performed some act for the sake of wealth, power, ascendency, lust – if it would remain forever unknown to gods and men, would you do it? They say that it is impossible. Yet it is not utterly impossible. But I ask, if that were possible which they say is impossible, what would they do? They persist, awkwardly indeed; they maintain that such a thing could not be, and they stand firm in this assertion; they do not take in the meaning of the phrase, “If it were possible.” For when we ask what they would do if they could conceal what they did, we do not ask whether they can hide it; but we put them, as it were, on the rack, that if they answer that they would do what seemed expedient if assured of impunity, they may confess themselves atrociously guilty; and if they make the contrary answer, that they may grant that whatever is wrong in itself ought to be shunned.
But are counterfactual conditionals with impossible antecedents always legitimate? Probably not. In cases where the antecedent is merely empirically rather than conceptually impossible, I stand by their legitimacy. (Thus I agree with Cicero about the ring of Gyges, and with Thomson about the violinist and the people seeds – all legitimate examples in my view.) But when an antecedent is actually conceptually impossible (where that includes nominally empirical claims that are among the presuppositions of a concept’s applicability) it’s no longer so clear that the resulting counterfactual (or perhaps countersensical) conditional even has a truth-value.
So for example, if someone were to say, “Okay, so you don’t think God could make it morally obligatory to torture an innocent person to death; still, suppose, counterfactually, that God could and did make it obligatory – then would you do it?” I think we are entitled to dismiss the question as nonsensical, just like the question “If spheres were at the same time cubical, would they roll?”
Now the Epicureans that Cicero is complaining about think, in effect, that the nonexistence of rings of Gyges is one of the empirical presuppositions of the applicability of moral concepts; I think they’re wrong about that, but it’s not a crazy thing to think (well, no crazier than consequentialism generally), and given that they think that, their dismissal of Plato’s example isn’t as groundless as Cicero makes it seem. In any case, contrary to what I tell my students, there can be grounds for objecting to the use of impossible situations even when the situation turns up only in the antecedent of a counterfactual conditional.
So why do I present these rules to my students as exceptionless when I think they actually have exceptions? Because a) introductory philosophy students are overwhelmingly more likely to err on the side of neglecting arguments and dismissing weird examples than they are to err on the other side; and b) the reasons for the exceptions are even more complicated to explain than the reasons for the rules, and sufficit diei malitia sua.