Determinacy of Translation

Determinacy of Translation

The following seems logically possible: a speaker’s use of the word “rabbit” is accompanied by rapid changes in its meaning and denotation—at one moment meaning rabbit and at another meaning undetached rabbit part. This will not be apparent to an observer, since assent behavior will remain constant in the presence of rabbits (and their undetached parts). Perhaps the speaker’s brain is being manipulated from moment to moment by a mad scientist. Should we take this hypothesis seriously as a possible interpretation of the speaker’s words, declaring it on a par with the usual assumption of semantic constancy? Or should we take seriously the hypothesis that a speaker’s use of “+” is accompanied by rapid alternations over time between addition and quaddition as its meaning? Clearly, that would be bizarre, though we have to admit that such a thing is logically possible. Why not? Because the world does not change like that over time. Granted, it could, if circumstances were conducive. But we have no evidence of that and can reasonably reject the idea (it’s just another piece of hyperbolic skepticism). We can reasonably assume a certain uniformity over time—stability, constancy. But if that is right, can’t we make the same assumption across speakers? It is logically possible that another speaker means by “rabbit” undetached rabbit part, but that would be a breach of uniformity that we have no reason to countenance; we have no evidence that the other speaker means something so different from us. He is a human being like us, with the needs and senses we have, living the same kind of life, so why should he differ so oddly from us in his linguistic nature? We can dismiss the alternative hypothesis as just so much skeptical exaggeration, not warranting dramatic claims such as that meaning is inherently indeterminate. Just as we can assume constancy of meaning across time, so we can assume constancy of meaning across individuals—assuming we have no evidence of any peculiarity in the present instance. That is just sound methodology, common sense. Even if the behavior of the speaker underdetermines the commonsense attribution, it is still acceptable to stick with it. We would need some startling new information to make us question our normal procedure (as that a mad scientist is manipulating the brain of the other). And questioning our background assumption of uniformity across time and across individuals of the same species (etc.) looks like an excessive response to the problem of induction. Nature is uniform and we are entitled to rely on this assumption in governing our descriptions of the things of nature.
But this short way with indeterminacy claims will not satisfy proponents of that position, so it would be desirable to undermine on general grounds their confidence that the problem is real. First, consider inverted spectrum cases: it seems logically possible that a person’s color perception might vary over time, switching from red to green (and conversely) from moment to moment. Yes, but it is hardly likely in the absence of any positive reason to believe it—short of some radical skeptical doubt. Why? Because human beings are similar in their perceptual capacities and responses, so we can reasonably assume no departure from the norm. Granted, we can’t observe other people’s minds, but reasonable principles of psychological attribution allow us to make uniform attributions. Or consider the possibility that the human language of thought varies over time and from person to person—having one sort of grammar at one time and another at a different time, or varying from person to person. Sure, that is a logical possibility—we can’t be certain it is not so—but it is an outlandish hypothesis with nothing to be said in its favor (unless we discover that it is actually so for some reason). In all these cases there is a certain kind of empirical underdetermination, which allows skepticism to get a foothold, but that is no reason seriously to entertain bizarre alternative hypotheses about what is going on. Phooey, we might say. Still, the defender of indeterminacy might insist that the mind is special; it alone allows for these kinds of underdetermination problems. Is that true? What if electrons were to switch their charge with protons over time, alternating from negative to positive, but with no change in the observable behavior of the elements (I’m not saying this is really physically possible, only that it is an intelligible hypothesis)? Should we take such a possibility seriously? No, because the world doesn’t work like that: it stays stable, uniform–it doesn’t just switch for no reason. Similarly, it would be bizarre to suggest that some atoms obey the usual charge distribution while others invert it (inverted charge cases)—even if the evidence couldn’t conclusively establish otherwise. It isn’t as if the two hypotheses are equally likely! Or again, it is logically possible that some parts of space are Newtonian and some parts Einsteinian, but we can confidently rule this out given that space must have a uniform nature across its entire extent. It is even logically possible that space switches from Newtonian to Einsteinian across time, being in the Einsteinian state whenever we apply empirical tests; our observations don’t logically exclude this possibility. But none of this warrants the idea that space might really vary in these ways, or that it is indeterminate between the two conceptions of space. Empirical underdetermination has no such drastic consequences; and uniformity assumptions rule out these outlandish possibilities. So, psychological indeterminacy is no more credible than physical indeterminacy (actually less so, given the availability of the first-person perspective). This is a position frequently urged (e.g., by Chomsky), but it is worth seeing how hyperbolic the Quinean line really is. It inflates mere logical possibility into serious science. Natives (i.e., fellow humans) are no more likely to mean undetached rabbit part by “rabbit” than newly observed electrons are likely to be positively charged. Maybe in some possible worlds these things are so, but in the actual world we can safely dismiss them. There is a fact of the matter about what people mean and we can generally tell what it is.[1]

[1] There is nothing especially inscrutable about meaning and reference, any more than there is about syntax and pragmatics. We can tell that an utterance has a certain grammatical form and a certain illocutionary force, and we can tell what it refers to equally. This is why we can actually translate foreign languages—which ought not to be possible under Quinean strictures. And isn’t it strange that other parts of language allow for translation but not the few cases discussed by Quine? Some meanings are held to be knowable and determinate, but some are not—why? Really the doctrine should be called “indeterminacy of some translation”. You can’t even derive the indeterminacy of color words from Quine’s considerations—or mathematical language, moral language, mass terms, and more.  Some meaning, then, is perfectly determinate.