Ordinary Language or Just Ordinary English?

17 Replies to “Ordinary Language or Just Ordinary English?”

1. Hi Nomy,
   Great topic 😉
   As it happens, there is a paper just published that compares English and Hebrew on “weak necessity”, or ‘ought’ vs. ‘must’; Aynat Rubinstein’s “On Necessity and Comparison”, in the December 2014 special issue of PPQ (edited by Mark Schroeder and myself).
   Hebrew isn’t the only language that lacks a dedicated term for ‘ought’. From what I’ve read, despite this it’s hard to find a language that doesn’t provide the resources to draw what looks to be the distinction between ‘ought’ and ‘must’. Kai von Fintel & Sabine Iatridou (“How to Say ‘Ought’ in Foreign”, 2008) find that in a wide range of languages, to say ‘ought’ you use constructions that compositionally express past necessity, and are commonly used to express counterfactual necessity (i.e. ‘would have to’). Alex Silk argues in some recent work that this directly reveals the meaning of ‘ought’. I’m skeptical of this: these constructions are ambiguous in those languages between a weak necessity and a counterfactual strong necessity reading, and counterfactual/past tense marking seems to be a generic tool for weakening the force of a term.
   I agree with you that if a philosophical view A is found in a population whose language doesn’t have feature B, then that’s strong evidence that feature B isn’t uniquely responsible for A. Is there a particular A and B you have in mind here?
   I don’t think that “findings that hang on the specifics of the language you speak” are problematic, if the questions you asked were influenced by the language you speak. (I like Nietzsche’s line: roughly, “Grammar is the metaphysics of the folk”).

2. Steve, thanks for answering! I will have to read your book now, and the studies about other languages.
   I don’t really get you when you say that findings that hang out on the specifics of the language you speak are not problematic “if the questions you asked were influenced by the language you speak”. What kind of “influence” are we talking about? Presumably, a Basque speaker and an English speaker can be interested in exactly the same ethical, metaethical, or moral psychological question, search for the truth about it, and differ meaningfully about whether, say, expressivism is true.The question “is expressivism true” is the same for both of them. We have all heard that the “100 words for snow” thing is false (and with it, one version of the “metaphysics of the folk” view).
   Here is a thought experiment. Or, first, here is a fact: in Basque the word used in ought statements is deeply and transparently connected to need (simple need, as in “cats need meat”). It’s pretty much mandatory to use a need-related verb in many ought statements that are not about need. Now for the thought experiment:
   Imagine that Basque becomes the language of a philosophically dominant nation. In that dominant nation, a bunch of Neo-Aristotelian virtue ethicists shows up one day, all excited by the fact that “We” in “Our Language” (or, “Ordinary Language”) connect normativity and need so inescapably: for doesn’t it suggest that the moral is what we need to have a good life? In such a scenario, it is next to impossible for us non-Basque non-Aristotelians not to intervene and ask rhetorically: just a moment, who is “We”? What do you mean “Our Language?” Is your language the only one that’s Ordinary? I think that would be a legit piece of criticism. The English speaker can legitimately tell the Basque here: you are prejudiced in favor of Neo-Aristotelianism, or at least make a bad argument for it, because you are taking a fact about the Basque language too seriously. To deny that is to risk being a sort of linguistic relativist.
   Now imagine that English speaking philosophers started talking like the Basque in the example…. Truly hypothetical of course!
   Now, again, I am ignorant of your book, and from your reply I get the impression that you don’t do that. All I am confessing to be confounded about here is your statement that findings that hang on the features of a specific language cannot be thereby problematic.

3. Note I didn’t say “CANNOT be problematic”. I had in mind questions that we ask only because of idiosyncrasies of our particular language. I don’t actually think many, if any, of our metaethical questions are like this, because I think that our central normative concepts tend to be pretty universal.
   But here’s an example of the kind of thing I had in mind. I argue in my book that “hypothetical imperatives” of the form, ‘If you want to do A, then you ought to do B’ are a kind of (hybrid) relevance conditional (or “biscuit conditional”; e.g. ‘If you want biscuits, there are some on the table’). Suppose that’s right. Now consider the metaethical question, “How can it be that you can make it the case that you ought to do something just by desiring an end to which it’s the means?” If the relevance conditional analysis is correct, then learning this supports the following answer: actually, desiring an end doesn’t make it the case that you ought to pursue some end; it simply makes a certain kind of “imperative of skill” (in Kant’s terminology) relevant. Now suppose there’s a language L in which there are no relevance conditionals (and hence, hypothetical imperatives). The speakers of this language would never have asked that metaethical question in the first place. (They might, I suppose, but for them it would be unmotivated, because it wouldn’t have seemed to begin from a datum. Or so I’m assuming for the purpose of this illustration.)
   To address your Basque case:
   First, I take it you’re assuming that our concept of ‘ought’ isn’t necessarily connected to (our concept of) need. The question to ask, then, is whether the Basque’s concept of “ought” indeed is necessarily connected to need. If so, then they don’t have our concept of ‘ought’ at all, and it’s a mistake to translate the things they say as such. They just have different concepts from us, and it’s wrong to say that their linguistic analysis is leading them astray.
   But I suspect that the scenario you have in mind is rather one where the Basques are being led astray about their own normative concepts, through “taking a fact about the Basque language too seriously”. I definitely accept that semantics can be done badly, and that features of language can be misleading. For example, suppose that the Basque word for ‘ought’ is ambiguous between “ought” and “need”; perhaps “need” is the etymological origin of the term, the way that “own” is the etymological origin of ‘ought’. Perhaps your neo-Aristotelian Basques are overlooking the ambiguity IN THEIR OWN LANGUAGE, and taking the use of a common term to be more significant than it is. (An analogy with English: some metaethicists attach great importance to the fact that “reason” refers both to a consideration-that-counts-in-favor, and to an intellectual faculty, and argue that something is a reason just in case, and because, it is something that the faculty of reason is responsive to. I think this is a mistake prompted by “taking a fact about the English language too seriously” (although actually I think it just has the explanation backwards).
   I agree that language can lead us astray if we analyze it incorrectly. But the antidote to this is to become better linguists, not to give up on linguistics.

4. That’s a great example about the ambiguity of the term “reason”. But I suspect that a lot of why I find it convincing has to do with being aware of one language in which nothing resembling that ambiguity exists and another unrelated language in which “Reason” transparently means “the power to reason” (i.e the story goes in your direction).
   That is: I agree with you that there is good vs bad linguistics, and anything with too much etymology in it tends to be on the bad side, but I don’t think you can dismiss the separate issue of monolingual research vs cross-lingusitic research. Even the most brilliant and careful expert on the behavior of the English words ‘ought’ and ‘must’ isn’t going to convince me that he or she has cracked the secrets of metaethics once and for all if all he or she has is a fascinating and shrewd analysis of the behavior of ENGLISH words. I don’t think I can judge how seriously it is appropriate to take a particular fact about English without knowing if the fact is echoed in other languages spoken by people whose “common sense morality” is similar to that of English speakers. Some of these people, these days, speak non-Indoeurpean languages.
   On the other hand, I hereby express my doubt that you can find a language community to whom some version of the question “does my having a desire for chocolate imply, in itself, that I ought to take the means to get chocolate?” would be an unintelligible question or seem to have a trivial or obvious answer. A nation with no Humeans would be exciting to know!

5. I agree that there are other reasons, besides our practices of uttering “hypothetical imperatives”, for being Humean. (I tried to hedge a bit for this reason). Though note that the question I suggested is one that even the Humean should find problematic: at best, what you ought to do depends on your overall preferences. Surely everyone can agree that merely having a (some!) desire for some end doesn’t automatically make it the case that one ought to take the necessary means. Yet that’s what “hypothetical imperatives” seem to imply, on their surface.
   I agree that cross-linguistic data can be very useful, and I’ve leaned on it myself. But I think we may disagree about how it’s useful. I think it’s useful mainly as a source of enlightenment about the meaning of our own language, and only thereby about the metaethical questions we ourselves ask. For example, the point about “reason”. Another example: the robustly cross-linguistic nature of the distinction between normative and nonnormative uses of (the words for) ‘must’ and ‘ought’ is strong evidence that it doesn’t reflect a deep ambiguity in our language, rather involving common concepts.
   However, I see this merely as one of many kinds of evidence about the meaning of our own language. And so I deny that anything is lacking if we have a totally convincing analysis of English words like ‘ought’.
   Here’s a dilemma. Either the other language is used to express the same thoughts that we express in English, or else it is used to express different thoughts. If the thoughts are the same, then a successful conceptual analysis of either of these languages will reveal to us the nature of the thoughts. If the thoughts are different, then they would seem to be about different topics. It would certainly be interesting (and to me, surprising) if people speaking other languages weren’t talking about what we ought to do, or what is good, but I don’t see how it could be relevant to the metaethical questions I’m asking. (Of course, this depends on exactly what the questions are.) In either case, I don’t see what could be lacking, with regard to my own metaethical questions, if I have a complete analysis of my own normative language.
   Consider the analogy of two different treasure maps/instructions. They might lead, by different paths, to the same destination. (E.g. 5 paces N, 2 paces W, 3 paces S, vs. 2 paces W, 2 paces N). That’s the situation I think we face when we compare (e.g.) English and Hebrew: the same thoughts, expressed in different ways. But you only NEED to follow one set of instructions to get to the treasure. Alternatively, the two maps might lead to different destinations. But if you’re hunting the treasure that corresponds to map A, then map B is simply irrelevant.

6. Throwing in my two cents: All should agree that understanding the linguistics literature on modals is crucial to doing semantics for modals in a natural language. One of the things I quite like about Steve’s book is the careful attention he gives to this literature. I think he’s also right to think that providing an empirically informed semantics for deontic modals in English should be of interest to metaethicists. After all, that’s the language many metaethicists use to express their intuitions and the nature of our intuitions is important to understand in order to understand what sort of evidential support they might plausibly provide ethical and metaethical theories.
   But I agree with Nomy that a basic constraint on such a semantics is that it should leave room for debates about topics metaethicists, ethicists, and ordinary folks are clearly able to have using this common language. And I don’t see how a semantics on which true ought claims (with a few extra whistles and bells) semantically entail true most reasons claim could be neutral on substantive normative debates about what we have reason to do. Nomy’s appeal to non-English languages makes this point, but we don’t need to talk about languages other than English to make it. We just need to reflect on the fact that two competent speakers of English could, with the type of contextual set up Finlay thinks is required, agree that Kim Jong-un ought to starve his people and disagree about whether he has any reason to, let alone that he has most reason to.
   There’s a related feature of Steve’s discussion that I don’t know quite what to make of. Often, as here, he talks about ‘our concept of ought, without quite explaining what a concept is, why we should think we have concepts in the sense he has in mind, and the connection between concepts and meanings. Certainly, when linguists investigate the meanings of modal expressions, they don’t talk in terms of concepts at all. So, I don’t quite understand the role of appeal to concepts in Steve’s methodology. (Apologies if these considerations were addressed in the earlier thread. I’m afraid I haven’t had time to read it just yet.)

7. Thanks for your two cents, Jan, which are worth much more than their face value! Two responses:
   (1) I think your semantic neutrality point (“that a basic constraint…is that [a semantics] should leave room for debates [we] are clearly able to have using this common language”) is different from any of the points Nomy was making. I agree with this point, which (as you know but others may not) is being made forcefully by many people in the deontic modals literature at present. However, I think (and argue in my book, esp. pp. 163-4, 254-5) that my semantics respects this constraint, largely because it doesn’t constrain what ends we can adopt.
   It is true that I argue for analytic entailments between ‘ought’ and ‘most reason’. This isn’t metaethically neutral, but I think it is normatively neutral: it doesn’t entail anything normative, about what natural facts are reasons for which actions. I suspect you’re reading it as problematically committal because you’re assuming that ‘most reason’ always involves “full-blown” normativity, or what rationally matters–and, like me, you don’t think that every ‘ought’ matters (sorry if I’m misinterpreting you!) But I deny that: some oughts matter, some oughts don’t; some reasons matter, some reasons don’t. On my view, you can’t infer a reason that matters from an ought that doesn’t.
   (2) It’s true that linguists don’t generally talk about concepts. Mostly, they’re just interested in formally modeling the truth conditions of various expressions. However, I take it that linguists working on (e.g.) modals are also concerned to identify “the abilities of a person who has a complete grasp of the modal system” (to quote Kratzer). The rule that gets identified as the modal force of a particular term is presumably the rule which competent speakers are implicitly guided by in composing and interpreting modal utterances. Put another way: linguists are concerned to analyze semantic competence. And “semantic competence” is roughly equivalent to what I mean by “concept”. Why use different language? Well, one difference is that linguists are only interested in concepts that are conventional meanings of words (hence ‘semantic competence’ has a sufficiently broad extension), but philosophers may also be interested in concepts irrespective of language, for which ‘semantic competence’ would not be a felicitous label.

8. Thanks, Steve. That’s helpful. On the point about your analytic entailment claim: I would have thought that when philosophers are interested in reasons, they are interested in normative reasons. Some, indeed, might think that a fact’s being a genuine practical reason requires it’s being normative. Could you explain why philosophers interested in practical reasons should be interested in the kind you are talking about, the ones that are “analytically entailed” by ought-claims? There’s a risk here that you appear to be making an interesting claim about reasons in the sense that practical philosophers take themselves to be interested in, when you are in fact making a claim about reasons in some idiosyncratic sense. (If you don’t like my use of “normative”, above, and want to contrast it with what you’re calling “full-blown normatively” then I will ask a similar question about your use of “normative”, whether it is the same as the one philosophers interested in normatively have in mind.
   About concepts and semantic competence: Can you say a bit more? Do you think, for example, that a concept is whatever one grasps when one is competent with a term? (Or is it something with additional conditions or altogether different?) Do you assume, as your discussion suggests, that what is grasped by a person who is competent with some term is the same for everyone competent with that term? If so, why do you assume that there is such a thing? If not, then why do you speak of “our” concept of “ought”?
   Also, I would have said that semantic competence with some term T in some natural language L is whatever ability a speaker has to have in order to use T to communicate, coordinate, and collect information with other, similarly competent speakers. Would you agree with this? Or would you characterize it differently? Saying a bit more about each of these, concepts and semantic competence, would be a help.

9. Good questions all, Jan. These are big issues, but I’ll try to be as concise as I can.
   On “normativity”: this is a recent, philosopher’s term of art, and in my view it jumbles together different things, and also means different things to different people. If we fix on the notion of guidance as central, we might distinguish between at least two different ways that we might judge a reason (or ought, or value property) to be “normative”. First, we might mean that it’s the kind of thing to function in a guiding role for some kind of agent/relative to some perspective or other. This gives us purely instrumental or “institutional” reasons (e.g. legal reasons, instrumental reasons of evil people). Second, we might mean that it functions in an important guiding role–where I would interpret this in a quasi-expressivist way as a guiding role that we ourselves endorse. For example, moral reasons. The latter is roughly what I meant by “full-blown” normativity.
   What kind of “normative reasons” are philosophers interested in? I think different philosophers mean different things by this. Humeans, instrumentalists, and relativists tend to mean the former (e.g. when Gil Harman denied that Hitler had any reasons to refrain from genocide, and Bernard Williams claimed that all reasons were internal). Absolutists, like Parfit, are only interested in the latter.
   I think that much of the metaethical debate over “normative reasons” is a consequence of equivocation, basically. I deny that I am making claims about reasons “in some idiosyncratic sense”; rather, I argue that our ordinary speech and thought about “reasons” is highly ambiguous (at the sentential, not the lexical level), and that this ambiguity is wreaking havoc in metaethics. My thought is that with an adequate grasp of the semantics (and pragmatics) of “reason” talk, we’ll be able to clear up this mess. (All this is to summarize what I argue for in “The Reasons that Matter”, AJP 2006).
   On concepts/competence: I may want to qualify some of the following if pressed, but in the interest of advancing the discussion…
   Yes, a concept is what one (tacitly) grasps when competent with a term. I’m happy to agree with your characterization of semantic competence.
   No, I don’t think that for any term, all competent speakers grasp the same concept. I don’t think that’s true of names, for example. Competence with a name merely requires that you know it’s a name of something, and employ some concept or other that individuates that something. Words that are like names in this way don’t have conceptual MEANINGS. (And so I wouldn’t talk about the “semantic analysis” of a name). But I think that many other kinds of words–including normative words like ‘ought’ and ‘reason’–have conceptual meanings, or are conventionally associated with concepts, a tacit grasp of which is necessary for being deemed semantically competent with those terms. (Other examples: indexicals like ‘I’, ‘tomorrow’, etc. One must grasp their “characters” in order to be semantically competent with them).
   I don’t assume that normative words have conceptual meanings, but I do argue for it (in my book, and in my methodology paper in progress). The primary evidence is that all speakers we’d deem competent with these words seem to be guided by the same rules in using them. For example, all competent speakers would judge p1 “good” for p2 iff (roughly) p1 raises the probability of p2. (But what about the extent of normative disagreement? I argue that this is fully explained by the combination of empirical disagreement and differences in preferred ends). Other evidence comes in the form of the extensive entailments that are alleged to be apriori; e.g. the practical relevance of reasons to action, the supervenience of the normative on the non-normative.

10. Hi Steve, I still disagree. Without crosslinguistic data we simply can’t know which facts about English should be regarded as significant for your purpose. Here is a dilemma.
    Suppose, hypothetically. It were to turn out that the behavior of ‘ought’ that you are talking about was unique to English. Two alternative explanations come to mind. One is that English speakers are profoundly different from the French,Basque, or Norwegian when it comes to normativity. A people that has no concept of ought is arguably even more different from us than the proverbial isolated tribe who leaves its old people to die in the snow, practices polyandry, allows its samurais to kill everyone they feel like, etc. it is just unlikely that such a distance exists between us and the Basques.. Average Basques tell their kids not to lie or steal or kill, just like we do, approve of putting murderers in prison, feel guilt and anger, have the same arguments about abortion as we do. When a Basque learns English she does not experience normative statements as mysterious – not even in the way that some American readers of daoist texts in Chinese experience “mysteriousness”. In short, no. I will not believe that Basques don’t have a concept of ought, much less that only English speakers have it. It is much “cheaper” to reject all your methodological views,
    What would the other horn of the dilemma be? Simply to conclude that if it were in fact unique to English, which apparently it’s not, the behavior of ‘ought’ that you discuss would be, like the reason/Reason ambiguity, considerably less important and significant.

11. Hi Nomy. I think this argument is a bit unfair. You ask me to suppose, hypothetically, that the “behavior of ‘ought'” is unique to English. Then you protest “No, I will not believe that Basques don’t have a concept of ought”. But I’m not asking you to believe that; I don’t believe it myself.
    A lot depends here on what you mean by “the behavior of ‘ought’ that [I am] talking about”. Distinguish between two different kinds of behavior: what it represents, vs. the means by which it represents it. (In terms of my earlier analogy: the destination vs. the set of instructions you follow to get there). If you mean us to suppose that Basque has no equivalent to English in the former way (what it represents), then I think it follows that Basques don’t genuinely have thoughts (at least, that they express in language) about ought. However, like you, I think that supposition is highly implausible, and for the kinds of reasons you give. By contrast, I do take it that many languages, including Basque and Hebrew, differ from English in the second way, in the means by which it represents it. But this doesn’t imply that Basques don’t have a concept of ought (even if they don’t have a dedicated word for ‘ought’). In either case, I don’t see why it would be IMPOSSIBLE to know what the concept of ought is without data about other languages (as your opening paragraph seems to claim). Surely if we know what ‘ought’ means, then we know what the concept of ought is, and surely we can know what ‘ought’ means without first learning about other languages.
    A concrete example (from pp. 91-2 of my book). Consider the English expression ‘most reason’. This construction is apparently idiosyncratic to English; at least, I’m told (by Gunnar Bjornsson) that the directly corresponding construction does not occur in Swedish. But Swedish employs an alternative construction that seems to mean the same thing, translating literally as “strongest reasons”. This cross-linguistic data is certainly useful in pointing us toward what ‘most reason’ means (Swedish seems a lot more transparent, in this case, whereas the English construction appears idiomatic.) But I think it’s possible to work out what ‘most reason’ means without access to facts like this about Swedish.
    In the case of ‘ought’, we know that English has a dedicated term to express the concept, while some other languages use the literal equivalent of ‘had must’/ ‘would have to’ (von Fintel & Iatridou 2008). But the evidence is they express the same concept.
    A couple of other points. (1) It doesn’t necessarily follow from the fact that (e.g.) the Basques make substantively the same “practical judgments” (broadly construed) that the English do, that they have the same concept of ‘ought’. Conceivably, different concepts could play the same practical role in different communities. Take for example Mackie’s 1977 book (or Harman 1996, for that matter): first he claims that ordinary people mean something by (moral) ‘ought’ that isn’t instantiated in the world, then he goes on to make roughly parallel judgments using the same vocabulary, but allegedly with a new meaning. There are real metaethical questions about whether people who apply ‘ought’ to the same things mean the same thing by it. (I happen to think they do).
    (2) Conversely, it doesn’t necessarily follow from the fact that some group or individual makes radically different “practical judgments” from us, that their word ‘ought’ means something different from ours. One alternative explanation, for example, is expressivism. (I think Matt Bedke has recently argued that this is a better explanation of the extent of normative disagreement). The explanation I prefer is rather semantic incompleteness/relativism (following Harman).

12. Jan and Steve, I’m interested in something that came up in your posts. (Actually, I’m really interested in all of this, but leave that to the side.)
    Jan aid “But I agree with Nomy that a basic constraint on such a semantics is that it should leave room for debates about topics metaethicists, ethicists, and ordinary folks are clearly able to have using this common language.”
    I wonder if this, in full generality, actually is a constraint. Consider an analogous case from another area. It is hardly a constraint on a semantics for mathematical language—roughly fol plus some jazzy stuff—that intuitionists and classical mathematicians be able to carry on debates they sometimes seem to have. And the respective semantics have quite significant first-order mathematical consequences. Likewise, it’s not obvious to me that certain fundamental normative terms, like “rational”, “ought”, or “reason” mean the same in my mouth as they do, for example, in Parfit’s—and not just because he’s British and fancy. And it’s not obvious to me there is a neutral way to frame the difference between our ways of talking. It would be nice if this were true of a semantics, but I’m not sure it approaches the level of a constraint…anyways. Interesting. I hadn’t thought to put my worry this way before. This seems yet another reason for me to get cracking on your book, Steve!

13. Hi Jack, good question. I won’t presume to speak for Jan, but I agree that we certainly shouldn’t be ruling out apriori the possibility that some people (folk or philosophers) are talking past each other with some of this language. (In fact, I suggested above that this is the case with some metaethical debates about reasons).
    I agree with Nomy and Jan that it should at least be a default presumption that the correct semantics won’t rule out as incoherent first-order normative judgments that at first glance look like substantive normative positions. But your post makes me wonder whether Jan meant to say that it is a substantive constraint on a semantics of normative language that it allow for familiar METAethical debates. I have trouble seeing how that “constraint” could be violated. Normative language is here the object language, not the metalanguage, and it doesn’t seem that a semantics for the OL could constrain what we may say in the ML. It might render some ML/metaethical claims false, but it can’t be that a semantics of normative language needs to be metaethically neutral. (My guess is that Jan didn’t mean to say any such thing).

14. My main worry, which I think you’re agreeing with, is that it is a (defeasible) default presumption, not a (n indefeasible) constraint, that our semantics be reasonably first-order-normatively neutral.
    On the second point, I agree that if the object language and metalanguage are cleanly distinguished, then giving a semantics for the OL in the ML can’t constrain the ML. Especially if we think of the job of a semantics as merely assigning truth-conditions. I was rather worried that a semantics for first-order normative language, interpreted as something like providing meanings when suitably interpreted, might not accommodate all potential metaethical views , at least in a unperverted form. The comparison with intuitionistic mathematics again is useful—we can do intuitionistic mathematics, sort of, within classical mathematics, but it’s (by intuitionistic lights) a perversion of their view. (There are a number of disanalogies here as well. But they don’t matter for the illustrative point.) So it might be that giving a formal semantics for normative language excludes certain otherwise seemingly coherent metaethical positions by virtue of their not fitting into that semantics in any natural way. This strikes me as an open and interesting question (i.e. I’m frantically trying to think of a devastating example.) But I am guessing here that I’m not saying anything you disagree with.

15. Jack, I think we’re in agreement, though I confess that the mathematical analogy isn’t helpful to me; I don’t know enough about the relevant debates. I do think that some metaethical claims (and philosophical claims more generally) are literally incoherent, though this needn’t be transparent. (I devoted a chapter of my PhD dissertation to defending this claim.)

16. Hi Steve, I agree that you don’t need to know about other languages in order to understand the expression ‘to have most reason’. I just think that without some awareness of what other languages do it is quite easy to read too much into our use of this (or any) particular expression. I still have to be convinced that any amount of care taken while treating data from one language can prevent us from such over-reading. I also don’t think you need anything resembling a word for ‘ought’ in order to understand what ought is, but I am not sure how I would have handled that fact if I treated linguistic evidence as very, very important in philosophy, the way you do. I don’t think knowing something about non-Indoeropean languages is an urgent need for just any philosopher, but I do suspect that it might be a real need for anyone who takes data about linguistic behavior to be important enough to topple entire schools of metaethics.
    OK, I’ll go buy your book!

17. Hi Nomy, perhaps we don’t disagree as much as I thought. I completely agree that “without some awareness of what other languages do it is quite easy to read too much into our use of this (or any) particular expression”. For this reason I’m always on the lookout for available comparative data. However, for what it’s worth, so far I’ve found this data only supporting the hypotheses I’ve reached on the basis of my armchair study of English, which perhaps explains my greater personal confidence that armchair methods can be viable by themselves. (Not that I expect anyone else to accept this on my say-so. But I can give examples on request.)
    I would probably have been just as skeptical as you that the armchair study of language could yield substantive metaethical results, before I tried such a study, and found (as it seems to me) that our ways of expressing ourselves in English contain rich clues that all point in a consistent direction. I didn’t start out such a linguistically-oriented philosopher: I followed a few interesting connections, and got sucked down the rabbithole. “The proof is in the pudding”. But you’ll have to be your own judge of that, if you read my book.