The most interesting sort, however, of resistance to the idea that one is using an expression in an absolute or secondary sense is related not only to unclarity, but to indecision—irresoluteness—about the use of the expression. Such things happen in different places in philosophy: in particular, in discussions about self-legislation, self-deception, it generally happens in discussions about the soul, in discussions about moral necessity, about realism in literature, and sometimes in discussions about God.
To better see the sources of this irresoluteness, take the idea that we can see things in the imagination, which I take to involve a secondary use of ‘see.’ On the one hand, we insist on the idea of seeing here. “We need this term here.” What can be more literal? On the other hand, there are familiar problems with this idea, for it doesn’t carry all the logical-grammatical implications that talk of objects of sight normally carries. For instance, when I see a zebra with my mind’s eye, I am not thereby also committed to the idea that I could see clearly how many stripes it has. Asking me to count them would not even make any sense. And this indicates that seeing a zebra and seeing a zebra with one’s mind’s eye are two grammatically different things. That is, it indicates that talk of seeing in the imagination is not quite literal after all.
Now, instead of talking about imaginary objects of sight, we can simply talk of thinking about something—a zebra. We can talk—literally—of reflecting about zebras in general, or of contemplating a particular zebra; this would solve the grammatical problem I mentioned, and allow for all the literalness we want. It will, however, be very unsatisfying. For there is a significant difference for us between merely contemplating a zebra and actually imagining it—bringing the image to mind. We really need the idea of seeing here: Nothing else could capture the experience in this case of the image in our mind.
In general, then, in the kind of cases I have in mind, there is a perfect literal, non-figurative, way of using an expression; there might even be more than one. However, none of these uses is satisfying. None of them captures what we want. The alternative is to admit that one wants to use an expression in an absolute or secondary sense. Recognizing this, however, would mean that one has to accept that it is internal to what they say that perfectly capable speakers may take what they say to be nonsense. And if a philosopher is using an expression in such a way, it means that they would have to accept that their argument might not be capable of binding everyone—might not even be understood.
In the kind of cases I have in mind the philosopher is facing a dilemma: To make her argument binding, she would have to use an expression literally, but that would make the argument unsatisfying. To make her argument satisfying, the philosopher would have to use the expression in an absolute or secondary sense; but that would mean that the argument is not binding. The philosopher wavers.
It should not come as a surprise, then, that recognizing absolute and secondary uses of terms can be difficult. The temptation is great to insist that one has made a perfectly legitimate literal use of an expression, even when one has not; this would allow for one’s argument to be both satisfying and binding. The temptation is great to say things like: “for a creature who must constitute her own identity, it is equally true that acting is quite literally interacting with yourself”—Christine Korsgaard. (What makes saying things like this possible is that the figurativeness of secondary and absolute uses is different from the figurativeness of other metaphorical expressions. It thus makes it possible to truthfully deny that one is using an expression in any simple metaphorical way. But the difference between kinds of figurative expressions is a subject for another post.) Naturally, succumbing to such a temptation would also involve failure to recognize absolute or secondary uses of language.