« previous entry | next entry »
Apr. 11th, 2011 | 11:40 am
music: John Powell - Battling the Green Death | Powered by Last.fm
Philosophy of language paper. Don't usually post school work here, but I'm a little proud of this one I think.
A. J. Ayer offers a simple, intuitively appealing account, of how sentences carry meaning: they are tautologies and express logical truths, or they are empirically verifiable and express a possible configuration of the world, or they are nonsense and express nothing at all. Having set out this criterion of meaningfulness, Ayer dismisses any philosophy purporting to “afford[ ] us knowledge of a world transcending the reality of common sense and science,”1 which turns out to be an extraordinarily large share of the discipline. This paper offers an interpretation of the somewhat hazy term “literal significance,” points out two problems with Ayer’s account and defends against these objections. It concludes that, though Ayer’s account is by no means without flaws, its snags and holes are not so bad as usually believed.
Ayer’s account of meaning concerns itself with the “literal significance of language,”2 and claims that metaphysical statements lack this sort of meaning. Although Ayer is very clear about what sorts of statements and questions lack literal significance – namely, those statements which pass the verifiability test: “Would any observations be relevant to the determination of its truth or falsehood?”3 – he is less direct about why these phrases lack literal significance (i.e., meaning) because of their unverifiability. The closest Ayer comes to giving us an answer is his introduction of his criticism of metaphysics:
One way of attacking a metaphysician who claimed to have knowledge of a reality which transcended the physical world would be to enquire from what premises his propositions were deduced. Must he not begin, as other men do, with the evidence of his senses? And if so, what valid process of reasoning can lead him to the idea of a transcendent reality? [...] But this objection would be met by a denial on the part of the metaphysician that his assertions were ultimately based on the evidence of his senses. [And even if they were], it would not follow that the assertions which he made concerning this non-empirical world could not be true.4
Ayer goes on to further develop the metaphysician’s counter-objections, including the claim that by saying there can be no metaphysical world, the anti-metaphysician is himself making metaphysical claims. This is not, Ayer is careful to say, applicable to his criticism. It is not, as Kant argued, that humans simply lack the capacity to navigate outside the phenomenal world – in which case the Metaphysician’s counter that one would have to have transcendent knowledge to deny the possibility of having transcendent knowledge would hold fine. Rather, it is logically impossible to have this sort of knowledge.5
But why should any of this suggest metaphysical statements are meaningless? Why aren’t they then “indeterminate” or “undecidable”? Ayer might implicitly be relying on an argument something like this. If a statement is metaphysical then it has to refer to either an object, action or relation which can not be observed in the empirical world, nor described purely in terms of logic. Now suppose someone, call him Oliver, sights the Platonic ideal of an ocelot living in his backyard.* Having checked the Audobon Guide to Local Wildlife, and finding no Platonic ocelots listed, Oliver calls up a nearby ecologist to notify them of this exciting discovery. “There’s the Platonic ideal of an ocelot in my backyard,” Oliver says. “Platonic what?” says the ecologist, “I’ve never heard of those.” What can Oliver say to clarify himself? That this ocelot is incorporeal and possesses precisely the necessary and sufficient features of ocelot-itude? The poor ecologist would only be more confused than ever! To what can Oliver point to explain the meaning of “Platonic ocelot?” What everyday language games exist in which, from observing the use of “Platonic ocelot” we can find its meaning? The problem with metaphysical statements and questions is that they contain language which, having no public use or empirically available example, can carry no meaning. With “red” we can point at apples, fire hydrants, and lobsters. For “greater than” we may play a language game of comparing the sizes of objects, or help someone with math homework. For “incorporeal, Ideal ocelot” we are out of luck. This is what gives us the criterion of verification: if a statement contains phrases whose meaning can not be clarified by example or practical use (empirical statements), or which is not given by definition of some arrangement of logical symbols (analytic ones), how can it be understood?
Fair enough for words like “noumenal” or “homoi-ousios.” But how does Ayer fit “the world of sense-experience [is] altogether unreal,” which expresses a metaphysical proposition in fairly common language, into the nonsense category?6 Each of those words looks like it has a fairly well established, practical meaning. An answer might be an odd use of the word “real.” If, out of some mean-spirited impulse to upset children, you tell a three year old “unicorns aren’t real,” they might set out to find one, and if they did in fact return with a unicorn you would have to admit that unicorns were real. In the upsetting-small-children-though-disilusi
It is tempting – but mistaken – to say that Ayer’s account is self-negating. At first glance, the claim that if a statement is neither empirical nor analytic it is senseless appears to condemn itself. That is, “Only those sentences which satisfy the criterion of verification have meaning” does not appear to satisfy the criterion of verification. It certainly isn’t analytic unless we’re defining “meaningful” to be “satisfies criterion of verification,” but then it becomes hard to deny competing ideas of meaning. And it does not appear to be empirically verifiable because, to date, we haven’t invented a literal meaning detector which makes a sharp buzzing noise at the utterance of nonsensical statements or some other device. If Ayer’s central claim was unverifiable, the entire project would collapse. But this contradiction relies on the assumption that meaning is something removed from the empirical world. If we accept that meaning is, instead, to be found in daily language games the claim that the kind of statements Ayer condemns are meaningless becomes verifiable after all. What facts would be relevant for the determination that they were? I think either of the following observations would do: i) A large sample of observers not acquainted with certain classes of empirical statements and metaphysical statements regularly come to a clear and agreed-upon interpretation of the empirical statements but not the metaphysical ones. If one hundred people of diverse backgrounds and talents come to a consensus in how to use words from Halliday and Resnick’s Fundamentals of Physics but not Pseudo-Dionysius the Areopagite’s On the Divine Names, then the claim that Pseudo-Dionysius’ language does not have literal meaning appears more probable, or ii) standing questions in one area of empirical enquiry (say, chemistry) or analytic reasoning (group theory) are resolved more quickly than standing problems in a metaphysical discipline. This version of Ayer’s account however, even if not self-contradictory, is a little bit circular. It does not require meaning be defined as “satisfaction of the criterion of verification,” but it does rely on meaning not being itself a metaphysical or transcendental object. If a metaphysician were to claim that the meanings of words were metaphysical objects, I can think of no Verificationist reply that would not beg the question.
Up until now I have discussed the criterion of verification as if it were universal. This is not obviously true. Because Ayer uses a weak principle of verification – that there are some experiences which could corroborate and cast doubt upon a proposed factual proposition – but does not specify bounds on what experiences carry evidentiary weight, he leaves it open that each individual might adopt their own. This is an unappealing result. Extreme liberality on what can count as evidentiary experience might allow the sensation of an epiphany to stand as evidence of all manner of religious or metaphysical claim, undoing Ayer’s main thesis. Even within set and established bounds of what can count as evidentiary experience, if two people do not have identical definitions on all points this would lead to the possibility of conversations where the same statement would be meaningful when one person said it but senseless when their conversation partner heard it. To be charitable and fair, I think this possibility is probably very remote. If all that is necessary is there be some set of possible observations relevant to determining if a statement is or is not true, then I think it would be hard to find two distinct plausible rules of evidence that do not simply agree about what is necessary for something to be proven, but where one admits a whole class of experience as evidence that the other does not so that there could be a proposition on which the rules disagree about its meaningfulness. Having sat here for a while no example comes to mind. Still the possibility is unnerving and risks making the existence of meaning in language individualized and subjective.
But whatever the faults in the individualized account of Ayer’s theory of meaning, they have nothing on the problems with assuming the standard for empirical verification is universal. Ayer provides no universal standard for what experiences are or aren’t admissible as evidence, but he can hardly be blamed for that. The failure to resolve each and every problem of epistemology is a very common shortcoming of philosophical works. But let’s say that Ayer enjoyed academic success beyond his wildest dreams, that every single person on Earth found Language, Truth, and Logic so compelling they converted to verificationism and dedicated all of their energy to helping sort out what the new universal standard for empirical verification ought to be. Being practical-minded, the people of Earth start with a few example propositions which seem like they’ll represent the more troublesome cases. Once these particular classes of statement are sorted out, it will be smooth sailing. The person with the most legible handwriting on Earth writes the first proposition, P on an enormous blackboard in English, Mandarin Chinese, Spanish and few other commonly-spoken languages, and then begins taking down possible experiences which would suggest the truth or falsehood of the proposition. After ten or so are on the board someone in the back raises their hand “Could you clarify what you mean by P? I’m not sure if the experience I’m about to suggest would actually be relevant.” Silence. How can anyone clarify the meaning of a statement (which might not have a meaning at all) if they don’t know what experiences will corroborate it? How can they draw up a list of evidentiary experiences relevant to a statement if they don’t know how to interpret its meaning? A universal theory of verification relies on a robust epistemological theory, which for practical application to propositions relies on a well-developed and meaningful language. Ayer’s theory might work if it concerns only some high-level definition of meaning distinct from the meaning of everyday language, so that it’s possible to build up propositions which are verifiable (and therefore have higher-level meaning in Ayer’s theory), but then we would be discussing a meaning different from the one which we set out to understand.
The derivation of the verification principle from meaning as a non-metaphysical concept partially rescues Ayer from this problem as well. The verifiability of statements isn’t a consequence of their meaningfulness in the sense of causation, but a convenient byproduct of the fact that a meaningful (non-analytic) statement needs to be empirical (so that it can be used productively in day-to-day language games or have its meaning given by visible example), and if a statement is empirical it can be supported by experience. This doesn’t quite save debates over which rule of interpretation to use,† and so agreeing upon a fixed standard for evidentiary experience is out, but it gives us the possibility of individual definitions of evidentiary experience. This also still leaves an incomplete idea of what counts as “relevant to the determination of [...] truth or falsehood,” and the objection that it is hard for a verificationist to say why epiphany isn’t evidence for religious phenomena stands, but this is hardly unique to verificationism since asking why epiphany can’t count as evidentiary experience is the same thing as asking why we should believe certain rules of inference rather than others, which is the sort of question that can cause mischief no matter what system is under discussion.‡
So the objections to Ayer’s account that still stand are these: i) that it is unable to disprove, or even argue against, competing theories of meaning (those which say meaning is in some way metaphysical) ii) that debate over a universal evidentiary standard is doomed under verificationist accounts to be unproductive iii) that in accounts where individuals set their own evidentiary standards of relevance, it is possible for a person to make a meaningful statement which, when heard by another person, becomes meaningless and iv) there’s no clear way to exclude standards using the criterion of verification which allow metaphysical claims through the back door, despite their lack of meaning (i.e. the criterion of verification gives false positives). How to fix these remaining problems, if they can be fixed at all, I do not know. I imagine the criterion of verification could be reformulated to avoid iv), and as I have already said I think iii) so rare that it would not be of practical concern. i) and ii) seem inescapable, but next to criticisms that Ayer’s project is self-negating, they also seem much less serious.
* Of course, if it is a Platonic ideal of an ocelot it wouldn’t actually live anywhere in this world, but humor me for the sake of the example.
† This probably requires some elaboration since I say below that the same objection is not fatal to the case where individuals hold their own evidentiary standards. The account in which there is an agreed universal standard for the relevance of an experience to determining the truth of a statement assumes this standard comes about from public discussion, and not some innate human preference for a particular flavor of empiricism or very lucky coincidence that everyone picks the same private standard. Discussion over what standard to adopt however is fruitless because advocating the superiority of one standard over another relies on metaphysical claims (assuming advocating the superiority of a standard is saying it better matches types of experience to things in themselves). By contrast if each person individually adopted a standard there would be no need for this debate at the cost of a presumed loss of uniformity.
‡ I do not know whether asking “why don’t we use these rules of inference?” counts as an analytic question or not. It is not meaningless the way metaphysical statements are, but suggests a different collection of true propositions (generated by taking your list of experiences and determine a new list of conclusions on the basis of this new set of rules of inference). It is certainly not empirical, and if it is not analytic a further category of meaningful statement might need to be added.
1. A. J. Ayer, Language Truth and Logic (New York, Dover Publications)(1952 ed.), 33.
2. Id., 35.
3. Id., 38.
4. Id., 33-34.
5. Id., 34-35.
6. Id., 38.