On Quine’s Translation Argument

Abstract

Quine's translation argumnent figures centrally in his views on logic. The goal of this paper is to get clear on that argument. It can be interpreted as an argument to the effect that one should never translate somebody’s speech as going against a law of the translator’s logic. Key to this reading of the translation argument is the premise that one should never translate somebody's speech such that their speech is unintelligible. Ultimately, it is my aim to reject this reading. I argue that only a weaker conclusion—one that says “not most of the time” instead of the stronger “never”—should be attributed to Quine. Accordingly, I propose and defend a weaker version of the first premise that better coheres with the weaker conclusion of the translation argument. Instead of the claim that one should never translate somebody’s speech such that their speech is unintelligible I argue that we should only ascribe to Quine the claim that one should not most of the time translate somebody’s speech in a way that makes it unintelligible. I go on to  sum up the results of my discussion and respond to a criticism of my reading.

1 A First Shot at Quine’s Translation Argument

In Word and Object, Quine writes:

To take the extreme case, let us suppose that certain natives are said to accept as true certain sentences translatable in the form ‘p and not p’. Now this claim is absurd under our semantic criteria. And, not to be dogmatic about them, what criteria might one prefer? Wanton translation can make natives sound as queer as one pleases. Better translation imposes our logic upon them, and would beg the question of prelogicality if there were a question to beg. (Quine 2013, 53)

The translation argument will begin with a moral drawn from Quine’s version of the principle of charity. But first, what are his views on the principle? In Word and Object, he writes, “one interlocutor’s silliness, beyond a certain point, is less likely than bad translation” (Quine 2013, 54).¹ He also claims, “For certainly, the more absurd or exotic the beliefs imputed to a people, the more suspicious we are entitled to be of the translations; the myth of the prelogical people marks only the extreme” (Quine 2013, 63). Similarly, in “Ontological Relativity,” Quine says, “[w]e will construe a neighbor’s word heterophonically now and again if thereby we see our way to making his message less absurd” (Quine 1969, 46). Based on these passages, Quine’s version of the principle of charity might be put bad translation is more likely than speaker silliness or absurdity.² When we understand “silliness or absurdity” as unintelligibility, which itself will be clarified in a later section of this paper, there are two morals that might be drawn from this principle:

Strong Moral One should never translate the speech of someone as unintelligible.

Weak Moral One should not the majority of the time translate the speech of someone as unintelligible.

As can be seen, the latter moral admits of exceptions while the former does not. This point will be of key importance in later sections of this paper. To return to the preliminary interpretation of Quine’s translation argument, the first premise will be the strong moral, the claim that one should never translate the speech of someone as unintelligible.

The second premise of the translation argument is the Quinian view that if a translation is to be intelligible then it will conform to the laws of the translator’s logic. This much is clearly attributable to Quine when he claims that translations of the form ‘p and not p’ are “absurd under our semantic criteria” (Quine 2013, 53, italics mine). To illustrate this, imagine that person A asked person B, “Are you happy?” Person B answers, “Yes and no.” Quine’s idea is that person A has to only understand B as claiming something like “Yes, I am happy in one sense but no, I am not, in another,” rather than as asserting straightforwardly something that conflicts with the law of non-contradiction. This is just to say that A must construe B in such a way that what he says conforms to her laws of logic.

From the strong moral drawn from the principle of charity along with the claim that intelligible translations conform to the laws of the translator’s logic, it follows that one should not, that is, one should never, translate somebody’s speech as going against a law of that translator’s logic.

2 Never vs. Not Most of the Time

Even if we do not ascribe to Quine the position that one should not translate others left and right as going against laws of logic, could we allow him the view that it can happen every now and again? That is, should the claim that one should never translate somebody’s speech as going against a law of the translator’s logic be taken to be the conclusion of the translation argument?

This sort of interpretation is primarily supported by a way of reading a handful claims in Quine. Recall that in Word and Object—but there are similar claims in other places too³—Quine writes that a translation that imposes “our logic upon [those who seem, at first pass, to be going against laws of logic] and would beg the question of prelogicality if there were a question to beg” (Quine 2013, 53, italics mine). Quine’s point here seems to be that really there is no such question to beg; it does not remain to be seen whether or not a prelogical culture can exist. This is perhaps because he thinks it simply impossible to have beliefs that go against laws of classical logic.⁴ Call this view the anti-prelogicality thesis.

If the anti-prelogicality thesis is attributed to Quine then it makes sense that the conclusion of the translation argument should be that one should all of the time avoid translating somebody’s speech as going against laws of the translator’s logic, given that the Quinian translator’s logic is classical. But this interpretation will conflict with a paradigmatically Quinian claim. Consider the following passage from “Two Dogmas of Empiricism:”

Any statement can be held true come what may, if we make drastic enough adjustments elsewhere in the system. Even a statement very close to the periphery can be held true in the face of recalcitrant experience by pleading hallucination or by amending certain statements of the kind of logical laws. Conversely, by the same token, no statement is immune to revision. Revision even of the logical law of excluded middle has been proposed as a means of simplifying quantum mechanics; and what difference is there in principle between such a shift and the shift whereby Kepler superseded Ptolemy, or Einstein Newton, or Darwin Aristotle? (Quine 1980, 43)

The idea is that there is a holistic web of belief that is itself subject to change. One can cling to a given claim even when facing recalcitrant experience that suggests its revision by revising a sufficient amount of the other parts of the web. Quine takes revision of the law of excluded middle in the face of quantum mechanics to be no different in principle from the scientific revolutions of Kepler, Einstein and Darwin. After all, “no statement is immune to revision” (Quine 1980, 43). Following Parent, call the view that any statement, including logical laws, may be revised the revisability doctrine (Parent 2008, 104).

There is a conflict between the revisability doctrine and the anti-prelogicality thesis. Given the revisability doctrine, laws of classical logic are revisable. If a law of classical logic were revised, sentences that go against that law would be able to make their way into the holistic web of belief. That is, one might come to believe sentences that go against that law of classical logic. Clearly this conflicts with the anti-prelogicality thesis, which just said that it is impossible for one to have a belief that goes against a law of classical logic.

Given this conflict, how should we read Quine? Should we see two Quines separated out over periods in his career?⁵ Should we limit the scope of the revisability doctrine? Another way of reading Quine⁶ allows for a weakening of the statement that one cannot have a belief that goes against a law of classical logic to the claim that it is all but impossible to do so, ascribing to Quine some limited version of the anti-prelogicality thesis. The view attributed to him would be only, say, that it is incredibly challenging to conceive of what evidence would outweigh evidence to the contrary and lead one to go against a law of classical logic.⁷

On one hand, this reading has the minor disadvantage of ascribing to Quine an overstatement of his case not only in Word and Object but also everywhere the text suggests it is flat out impossible to have a belief that goes against a law of classical logic. For example, on this reading, he did not really mean that there was no question whatsoever to beg of a prelogical culture, but instead, he meant that there is almost certainly no question to beg.

On the other hand, this reading will have the advantage of not ascribing to Quine a surreptitious rejection of the revisability doctrine, the doctrine itself being a Quinian view par excellance. For, given the conflict between the revisability doctrine and the anti-prelogicality thesis, if one wants to maintain an interpretation that holds Quine to the anti-prelogicality thesis then that interpretation will likely attribute to Quine an abandonment or limiting of his revisability doctrine. Moreover, Quine will have dropped or restricted the revisability doctrine in a not obviously marked way.

Which interpretation is preferable? It is a bigger interpretive leap to ascribe to someone the stealthy abandonment of a view with which they changed the landscape of philosophy than it is to ascribe to them some overstatements of conclusions they would draw. Who after all, especially in philosophy, has not once stated a conclusion ever so slightly stronger than they were entitled to? Because of this, it appears preferable to read a philosopher as somewhat incautiously overstating conclusions than it does to take them to be surreptitiously abandoning one of their major contributions to philosophy.⁸ Hence it seems clear that Quine should be read as committed only to the view that it is all but impossible to understand somebody as going against a law of classical logic in order to leave room for the revisability doctrine.

If Quine only holds only a limited version of the anti-prelogicality thesis then the conclusion of the translation argument can be weakened. Quine need not argue that one should never translate somebody’s speech as going against a law of the translator’s logic, since having such a belief is only all but impossible (assuming, again, that the logic of the Quinian translator is classical).

Moreover, the text lends itself more naturally to this sort of reading.⁹ Quine does after all say that it is “[b]etter translation” that “imposes our logic upon” apparently prelogical people who seem to be transparently going against laws of logic (Quine 2013, 53, italics mine). Better translation seems to leave room for worse translation. That is, for a worse translation that translates the apparently prelogical people as prelogical. Such translation might still be good or acceptable translation for some other reason. Sentences of science—like “Neutrinos lack mass”—when first heard, for example, may fail to be understood by the listener. Yet the reason to carry on trying to make sense of translations of these sentences despite their initial unintelligibility is that they might contribute to the advancement of our collective knowledge. But more on this in later sections.

3 How Much Charity Does the Translation Argument Need?

Recall that the first premise of the translation argument was the strong moral drawn from the principle of charity. This was the following:

Strong Moral One should never translate the speech of someone as unintelligible.

In the previous section I argued that the conclusion of the translation argument should only be that one should not the overwhelming majority of the time translate somebody’s speech as going against a law of the translator’s logic. But if we do not attribute to Quine the stronger conclusion, should the strong moral still be ascribed to him?

The strong moral is only required if one wants to read the translation argument with the stronger conclusion that one should never translate somebody’s speech as going against a law of the translator’s logic. It is thus needlessly heavy-handed as a premise in the translation argument given its conclusion is only that one should not the overwhelming majority of the time translate somebody’s speech as going against a law of the translator’s logic. For this weaker conclusion something like the following would suffice:

Weak Moral One should not the majority of the time translate the speech of someone as unintelligible.

For the remainder of this section I will argue that Quine would not have endorsed the strong moral and thus instead should be ascribed only something like the weak moral.

If Quine held the strong moral then he would not permit of exceptions. What I mean is that there should not be cases where Quine would say that one would have reason to translate somebody’s speech as something unintelligible. But I think there is an exception and I think it is one that Quine would readily recognize. Let me explain. Quine delineates a class of sentences which, when translated into a deficient language L, will be unintelligible. This is not a shortcoming of the translator but simply a deficiency in language L. Consider the following passage:

Thus who would undertake to translate ‘Neutrinos lack mass’ into the jungle language? If anyone does, we may expect him to coin words or distort the usage of old ones. We may expect him to plead in extenuation that the natives lack the requisite concepts; also that they know too little physics. And he is right, except for the hint of there being some free-floating linguistically neutral meaning which we capture, in ‘Neutrinos lack mass’, and the native cannot. (Quine 2013, 69)

And then a paragraph later:

Observation sentences peel nicely; their meanings, stimulus meanings, emerge absolute and free of residual verbal taint. Similarly for occasion sentences more generally, since the linguist can go native. Theoretical sentences such as ‘Neutrinos lack mass’, or the law of entropy, or the constancy of the speed of light, are at the other extreme. It is of such sentences above all that Wittgenstein’s dictum holds true: “Understanding a sentence means understanding a language” Such sentences, and countless ones that lie intermediate between the two extremes, lack linguistically neutral meaning. (Quine 2013, 69)

The sentence “Neutrinos lack mass” cannot be translated into the native’s language intelligibly because the native’s language is deficient in some way. Yet “Neutrinos lack mass” could also be of great importance to the native. There is an entire class of sentences like this, including the law of entropy and the constancy of the speed of light. These sentences, and others like them, are theoretical sentences. On one hand, theoretical sentences are unintelligible to someone with a deficient language; Quine’s native cannot understand “Neutrinos lack mass” due to the fact that “the natives lack the requisite concepts; also that they know too little physics” (Quine 2013, 69). On the other hand, they can be greatly valuable to our, and Quine's native's, collective knowledge.

If L is a deficient language then every translation of “Neutrinos lack mass” will be unintelligible. If one has reason to translate this sentence into L then translation will be done in a way that the sentence remains unintelligible. Is there still reason to translate (unintelligibly) “Neutrinos lack mass” into L? Surely one would want to express something like “We should learn the enriched language L + so that we can understand “Neutrinos lack mass,”” where L + is some language sufficiently powerful to render “Neutrinos lack mass” intelligible to its speakers. This seems so trivial a point that surely it would be acceptable to Quine.

If theoretical sentences are sentences that are in principle unintelligible in language L and one has reason to, nonetheless, translate them into L, then there are counterexamples to the strong moral drawn from the principle of charity. It thus makes sense not to ascribe the strong moral to Quine in the first place. As I pointed out, the translation argument can work with a weaker first premise (I offered the weak moral).¹⁰

Before moving on, it should be noted that in From Stimulus to Science, Quine writes, “‘Neutrinos lack mass’ is untranslatable into the English of 1930” (Quine 1998, 78). At first glance this appears to conflict with my claim that there is reason to translate theoretical sentences into a deficient language even though those translations will be unintelligible.

The sense in which Quine says “Neutrinos lack mass” is untranslatable is not the same, however, as the sense that I was using. Recall the related passage from Word and Object:

Thus who would undertake to translate ‘Neutrinos lack mass’ into the jungle language? If anyone does, we may expect him to coin words or distort the usage of old ones. We may expect him to plead in extenuation that the natives lack the requisite concepts; also that they know too little physics. And he is right, except for the hint of there being some free-floating linguistically neutral meaning which we capture, in ‘Neutrinos lack mass’, and the native cannot. (Quine 2013, 69)

The point here is that the native cannot translate intelligibly the theoretical sentence. This was because the native’s language and conceptual framework were deficient. In this first sense, the sentence “Neutrinos lack mass” cannot be translated. But this is not the same sense of “translation” as Quine's native’s translation in the claim, “I don’t understand the sentence “Neutrinos lack mass,” please provide me with the tools to learn physics so I can come to understand it.” In this sense, the theoretical sentence is translatable. Hence, when Quine claims, “‘Neutrinos lack mass’ is untranslatable into the English of 1930,” we should read “‘Neutrinos lack mass’ is untranslatable intelligibly into the English of 1930” (Quine 1998, 78). Just as the native can say sentences along the lines of, “I want to understand “Neutrinos lack mass,”” the English speaker of 1930 might do the same.¹¹

But should an interpretation of Quine allow for unintelligible translations? It seems to me that it should. This is because a translation that is bad because it is unintelligible still counts as a translation. Recall that rendering the native as asserting something of the form “p and not p” in fact counted as a translation, but just as a bad one (Quine 2013, 53). That is, “wanton” translation is still translation.

4 Quine’s Translation Argument (Really)

With the exegesis of the previous two sections in tow I now present what I take to be Quine’s translation argument. I formulate it as follows:

• 1. One should not the majority of the time translate the speech of someone as unintelligible.

• 2. Translating someone as going against a law of the translator’s logic renders their speech unintelligible.

• 3. One should not, the majority of the time, translate somebody’s speech as going against a law of the translator’s logic.

Premise (2) is preserved from section one. It is exactly the same as the second premise in our first run through of Quine’s argument.

Step (3) is the conclusion that I argued should be attributed to Quine in the second section of this paper. The reasons were its prima facie plausibility as a reading of the conclusion of Quine’s translation argument and the fact that the alternative stronger conclusion was supported, at least primarily, by readings of Quine that would move towards having him drop the revisability doctrine. Though people do change their views, it is interpretively questionable at best to ascribe to an author a clandestine desertion of a thesis with which they shook the philosophical world.

Claim (1) is a weakened version of the moral drawn from the principle of charity. This version fits better with the weakened conclusion. A stronger first premise, one that said to never translate somebody’s speech in a way that renders it unintelligible, was less than ideal for two reasons. First, there was a class of sentences delineated by Quine that provided reason not to ascribe to him the strong moral, namely theoretical sentences which were unintelligible to those with deficient languages on one hand yet valuable on the other. Because of this, I argued, people have reason to translate them in a way that preserves their unintelligibility. Second, given the weakened conclusion the stronger moral is excessive as a premise of the translation argument.

5 Unintelligibility

An important but not yet explicit premise in my argument has been that the sense in which a sentence that goes against a law of a translator’s logic is unintelligible is the same as the sense in which a theoretical sentence might be unintelligible to, say, Quine’s native with a deficient language. But are these two types of sentences unintelligible in the same sense?

An interlocutor might argue that there are in fact two distinct senses here. On one hand, the deficient language of Quine’s native can be enriched so that sentences like “Neutrinos lack mass” come to be intelligible. On the other hand, perhaps there is no possible enrichment of a translator’s language so that a sentence that goes against a law of that translator’s logic will come to be intelligible. If this is so, it will be clear that we have really been working with two senses of “unintelligibility” all along.¹²

The claim that a sentence that goes against a law of logic cannot come to be intelligible through language enrichment might be conceded to my interlocutor. This concession will only be problematic if it entails that there are two senses of “unintelligibility” at play. It does not, however. Returning to the text we can see how this concession can be accepted while a single sense of “unintelligibility” is preserved. First, let us clarify the relevant sense of “unintelligibility.” Recall that Quine allows that by some sort of change in the holistic web of belief sentences that go against laws of logic can be held true. In “Two Dogmas of Empiricism,” Quine famously says this much (Quine 1980, 43). Hence I take the correct sense of “intelligible” and “unintelligible” to be something like intelligible/unintelligible with respect to a holistic web of belief. Since there is no difference in principle between a change of scientific theories and a revision of logical ones, it makes sense that this sense of “intelligible” and “unintelligible” would apply uniformly both to sentences that go against a translator’s laws of logic and to theoretical sentences.

Next, we can accept that theoretical sentences come to be intelligible by language enrichments while sentences that go against a law of a translator’s logic fail to do so. Though this surface level difference is present, there is a deeper analogy. Sentences that go against a law of a translator’s logic will come to be intelligible, if they do so at all, because of a change in the holistic web of belief. Similarly, theoretical sentences will come to be intelligible by a change in the holistic web of belief, namely, by an enrichment.