Thursday, April 4, 2013

What do you mean "what's it mean?, what's it mean?"?

Philosophical Investigations 123 says that a philosophical problem has the form: "I don't know my way about." The German is Ich kenne mich nicht aus. According to my dictionary this might also be translated "I am stumped" or "I am at a loss." Auskennen means to know one's way about, to know what's what. Wittgenstein's words suggest feeling trapped, not knowing how to get out. Hence, presumably, the need for the liberating word. Language traps us, but the problem is really one of the will. So perhaps really it is the will that traps us. Or perhaps we cannot so easily distinguish between language and the will. It is our language, after all, and the problems of philosophy seem to arise, according to Wittgenstein, when we want to say things that in some sense we cannot say, or when we cannot find any words that mean what we want them to mean, or when our will is frustrated by language. Involuntary nonsense is a result of a certain kind of coming together (or not coming together, perhaps) of language and will. This means there is an ethical aspect to linguistic analysis: the will is involved, and willing can be evaluated as good or bad.

In this review of G. A. Cohen's Finding Oneself in the Other, Ralf M. Bader writes:
While Cohen does not provide an analysis of what it is for a statement to be unclarifiable, he puts forward a sufficient condition (which he attributes to Arthur Brown), namely "that adding or subtracting (if it has one) a negation sign from a text makes no difference to its level of plausibility: no force in a statement has been grasped if its putative grasper would react no differently to its negation from how he reacts to the original statement" (pp. 105-106). This proposal seems problematic insofar as the plausibility of the proposition that Fellows' lawn has an even number of blades of grass at a particular time seems to be equal to the plausibility of the negation of this proposition, yet neither of these propositions is bullshit. While the unfounded assertion of either statement might constitute an act of Frankfurt-bullshitting, in that the assertion would be groundless and as such not display a concern for truth, it would not be a case of Cohen-bullshitting since the statements are both perfectly clear and can be grasped (though not evaluated) straightforwardly.
I wonder about this. What could someone possibly mean by saying that the lawn has an even number of blades of grass on it? If they have some way of knowing then it makes sense. Perhaps the grass seed is planted in some way such that, would you believe it?, we know exactly how many blades there are (or perhaps we don't know the number but do know that it can only be an even number). Perhaps our new spy satellite is so good that it can even count the number of blades of grass on a lawn. But if we don't have anything like this kind of way of counting the blades, what could we possibly mean? (I don't mean that verificationism is true in general, but it seems to have some plausibility in this case.) I would not know what someone who said such a thing (i.e. something like "There is an even number of blades of grass on this lawn," assuming a normal, grass-covered lawn) meant. Because the assertion would seem so groundless and to show such a wild lack of concern for truth I would not feel that I could grasp the statement, nor that it was perfectly clear. There is a sense in which I know what "There is an even number of blades of grass on this lawn" means (it means what it says, I can paraphrase it in a Google-translate kind of way), but I would not know what someone who said it meant.

In his recent essay on "Rhees on the Unity of Language," Lars Hertzberg quotes Rush Rhees:
Understanding what is being said; not just understanding what the words mean. Objection: “those are synonymous.” I do not think so. You can understand what the words mean even when nothing is being said. 
This reminded me of the following: O. K. Bouwsma reports that Max Black once argued that in some circumstances a sentence such as “I am here” or “I exist” would serve no purpose but that, even then, if one were asked whether it was true or false one would say that it was true. To this, Bouwsma says, Wittgenstein responded, “No! No! Of course not, etc. Context determines use” (Bouwsma pp. 14-15). Rhees distinguishes between what is being said (by a person) and what words mean. Wittgenstein (reportedly and implicitly, as I understand him) denies that one can say whether a sentence is true or false regardless of its context. This does not contradict Rhees, although one might wonder whether they would have agreed on all this. Can words have a meaning but be neither true nor false? Of course they can if they don't form a statement (a question has meaning but no truth-value, for instance). But Hertzberg suggests that Rhees does have statements in mind:
I would suggest that what Rhees means by “understanding the words” is not simply recognising the single words but also recognising the sentence as an English sentence, getting a “sentence feeling,” as in a grammar exercise in which nothing is being said: “The cat is on the mat.”
It is neither true nor false that the cat is on the mat, because this kind of grammar exercise is a kind of bullshit in the technical sense: the concern is not with truth. "The cat is on the mat" has a sentence feeling, but does it have a meaning? I don't see how to avoid the answer Yes and No to this question. There is no question of the speaker's meaning because there is no speaker. Even if a teacher of English were to pronounce the sentence, she would not be saying (asserting) it. Qua grammar exercise it cannot be asserted, i.e. in grammatical exercises one demonstrates and practices grammar, one does not make, is not making, assertions. Such sentences are mentioned but not used, so to speak. So if by "the meaning of the sentence" we mean the use to which it is being put then it has no meaning. (The sentence is being used, but in a special kind of way.) But of course it would be odd to teach and learn English using meaningless sentences. These sentences do have a meaning in a thinner sense. They can be paraphrased, and we can imagine circumstances in which just those words in just that order could be used.

Perhaps we could say they have a potential meaning, but everything has a potential meaning. So maybe we should say they have a standard meaning (or a literal meaning, which need not be the most common meaning since some words, e.g. 'literally', are most often not used literally). Then does "I am here" or "I exist" have this kind of meaning? Yes, I think so. But "I am here" doesn't have a clear standard meaning. It's literal meaning is tautological, but it's most common use would, I think, be much the same as that of "Here I am!" or "Cooey!" And cooey is neither true nor false. "I exist" has no common use outside of philosophy. But there it is taken to be true, at least usually. Is it therefore true? It seems to be, but Wittgenstein was apparently reluctant to say so. More than reluctant. He was apparently insistent that we not say that such sentences are true. Perhaps because he had a bit of Rorty's idea in him that to call a sentence true is to pay it a certain kind of compliment. I doubt he had more than a little bit of this idea in him, but it might have been enough.  

Do I have a point? What ties all these ideas together? Philosophy seems to involve a relationship between the self (the puzzled philosopher), other people (who created and sustain the language, and help determine what is correct and what incorrect), language itself (which has a limited kind of independence of its creators), and the world (which language, among other things, describes). In the grass case the philosopher is puzzled by how a speaker, his sentence, and the world (the nature of lawns, for instance) could go together. Is the speaker lying? Bullshitting? Kidding? Roughly the problem is: how can we be charitable to the speaker in this case? (There might also be questions of charity towards the world or towards language. Skepticism is a lack of faith, which is a kind of lack of charity. Or is that more bull?) And questions of meaning are not wholly separate, or separable, from questions about people, because words only have meaning in a full sense when used by people. (And on the question of the relation between language and the world, see Cora Diamond's recent essay "The Skies of Dante and Our Skies.") This is all a bit half-baked, but perhaps I will be able to finish baking at least of it later.


  1. It seems to me that what ties these together is the issue as to whether normal listening to (or reading of) language involves the constant interpretation of what's being said. This is a vexed question that lies at the heart of the rule-following discussion, and my interpretation (irony intended) of Wittgenstein is as follows:

    In the normal flow of language we don't interpret what's being said. We simply understand it. BUT! if it subsequently turns out that we understood wrong then the rules of grammar allow us to say "I interpreted you as meaning x, but now I see I was wrong". What we mustn't do (what will get us into trouble) is to assume (as an inference, or kind of theory) that some sort of interpretation must have been going on all the time - or else how could I have mis-interpreted what was said? Wittgenstein's point is (I think) that perverse as it sounds, there's no interpretation involved in hearing (eg) "The cat was on the mat" unless it subsequently turns out not to mean what we thought it did (eg "cat" turns out to refer to the 1970's Chelsea goalkeeper Peter Bonetti). We're tempted to reply "But it can't be like THAT! Surely the case where we're mistaken PROVES that interpretation is always going on?"

    And here we're imagining some kind of mental mechanism which regulates meaning in the same way that (we think) a mathematical function regulates the product of an argument. But we're imagining it! It's a theory. And I think Wittgenstein would reply: no matter how compelling that theory is, our grammar doesn't bear it out. That's not how the word "interpretation" is used. To believe that language must function like that (to believe that we as human beings must function like that) is to be under the spell of a certain picture of mechanism.

    Anyway, that may or may or not be right, but the thing about your post that grabbed my attention was the headline. I assume you were quoting Wittgenstein's tendency (in his younger years) to respond to "What do you mean by...?" by bellowing "WHAT DO YOU MEAN 'What do you mean'?"

    I immediately thought "which song is that quoting?" The Fall were the prime suspects (though Talking Heads were also in the frame for a bit). The video at the end of course confirmed it to be The Fall - though I wouldn't have guessed it to be the Peel Session version of New Puritan.

    Anyhoo, this reminded me of the other (possible) link between The Fall and Wittgenstein. In the song "In My Area" Smith sings "Understand time till I'm asked about it". He's either quoting Augustine or else he's quoting Wittgenstein quoting him in the PI. I've always wondered which it was. I suspect Smith would react rudely if I ever put the question to him.

    1. Thanks. I don't remember that about Wittgenstein shouting "What do you mean?" I was just thinking of The Fall (the single version, I think--if the video has a different version that's an accident). And I had never realized that "In My Area" had that line about time. I think it's much more likely that Smith read a bit of Wittgenstein than that he read Augustine, but you never know. As you say, you couldn't ask him and expect a helpful reply.

      I agree with you about interpretation. We only interpret when interpretation is called for, and it usually isn't. One time it is is when we have misunderstood, and then we might say we misinterpreted. But we might also say we misunderstood or just took it the wrong way. We usually just take things (one way or another, rather than interpreting them), I suppose, but perhaps not even that. "I took him to mean..." implies some interpretation or doubt about the meaning, and usually there is no doubt, just a kind of flow.

  2. I second Philip's skepticism about interpretation. My hypothesis is that in most casual - and even some not so casual - conversation we are simply responding in a highly automated way to a stimulus. But as Philip says, we don't simply respond, "We simply understand it" as well.

    But if we are simply responding automaton-like, what does it mean to "understand"? W gives us a clue in §6:

    "Don't you understand the call "Slab!" if you act upon it in such-and-such a way?"

    Ie, you understand a sentence if you respond as intended by the speaker. This is consistent with the tool metaphor applied to sentences (including elliptical ones): you use a tool to effect a result. And it is also consistent with W's insistence on context. As he says in the same paragraph in §6:

    "With different training the same ostensive teaching of these words would have effected a quite different understanding."

    Ie, a student might be trained to respond differently to the words, which suggests that the "meaning" of a sentence is the response intended by the speaker - and expected, owing to "training" (possibly implicit) of the speaker's audience.

    That view of "meaning" (which I'll call "intended response meaning", or ir-meaning) seems to address several issues raised in the post. Consider the troublesome issue of truth. "Slab!" is clearly neither T nor F but can have ir-meaning in a given context. Or imagine (if you can) a politician who is chasing votes by asserting an untrue proposition. There is no question that the speaker's assertion has ir-meaning - the intent is for hearers to respond in a very specific way. And this is also the case even if P is technically true but intended to mislead (eg, a quote taken out of context). Ie, truth is irrelevant to ir-meaning.

    And consider the problematic assertion about blades of grass. Its grammatical meaning is clear and it must have a truth value even if it's unknown. But without context (and even within likely contexts) it has no ir-meaning. As Duncan says:

    "There is a sense in which I know what "There is an even number of blades of grass on this lawn" means ... (in a Google-translate kind of way), but I would not know what someone who said it meant."

    Ie, we wouldn't understand the ir-meaning.

    Another example that I like is a casual "How are you?". Most of the time, a response much more detailed than "I'm fine, thanks." indicates that the ir-meaning of the question was not understood.

    BTW, I use "truth" in the sense of interpersonal agreement intended by Rorty's oft maligned quip "Truth is what your peers let you get away with saying."

  3. Thanks, Charles. This sounds about right. I wouldn't say that the meaning is the intended response (because then "How are you?" would mean "Fine," which it doesn't), but I take your point. The meaning depends on the intended response, is tied up with it. In that sense "How many blades of grass are there?" does NOT have to have a truth value. In a different sense it must do so, but that's to state a rule. And the rule is not very helpful in this case. It might even be harmful. Which is why I think Wittgenstein might have been something of a pragmatist in cases like this. His "No, no, of course not!" is not a grammatical remark so much as a pointer away from confusion toward a safer path.

  4. The problem to me with the leaves of grass sentence is that there is no hint of what kind of conversation is going on. Whom am I supposed to be informing about the number of leaves, and why? What hangs on my getting it right? What is the time frame? How big should a leaf be to count as a leaf? Etc, etc. (Counting leaves of might be the kind of task you’d be given in the army, but then the point would really be that the task is pointless – that it isn’t an actual task.) If I don’t know what the people are talking about, I don’t know what it would mean to call the things they’re saying “true” or “false”.

    I very much agree with Philip’s point about interpretation. But I have some hesitation concerning Charles’s use of the word “automaton-like”. That makes it sound as if we either deliberate actively or function like automata. And – unless by “automatic” one simply *means* “without active deliberation” – that sounds like a way of denigrating the response: “one is simply running on routine, zombielike, one is not really engaged in the conversation”. But I should like to say: the response may be as smart, apposite, imaginative as you like, and still not be preceded by an active process of interpretation.

    As for Charles’s suggestion that “you understand a sentence if you respond as intended by the speaker”, that can’t be quite right. If I fail to do what you expected me to do, my defence may be that you misspoke, spoke carelessly, etc. If you told me to water the orchid, how was I supposed to know you had more than one? I responded correctly to what you said - *to the words as used in that particular context* - although I didn’t quite do what you intended me to do.

  5. Thanks, Lars.

    Yes, the idea that the statement about the leaves of grass has a meaning that is "perfectly clear and can be grasped straightforwardly" is, I think, just wrong without some context to help make sense of it. The sentence might be true or false (and someone can insist that it must be one or the other if they want to do so, although this might be an expression of commitment to a certain principle more than it is a statement of fact), but there's no knowing whether the sentence has a meaning, nor what that meaning is, unless, well, unless we know what it means. Its meaning just is not obvious.

    I agree with Charles that, on the one hand, we simply respond in normal cases but that, on the other, we don't simply respond, we also understand. That is, I would say, we simply respond, but not as an automaton might because we, unlike an automaton, understand. This understanding is not a process, though, that precedes the response. We respond with understanding, in a way that is characterized by understanding, not in a way that is, say, caused (or preceded) by understanding. Most of the time, anyway.

    If you respond as the speaker intended then you probably understand the sentence, but you might understand and yet respond differently. Examples: I intend to make you drop what you are carrying by shouting nonsense at you. If you drop it, have you understood what I shouted? I tell you to water the orchid. Instead of putting three ice cubes on it, as I intend, you pour water, drenching the soil in the pot. Did you understand the sentence "water the orchid"? I would say Yes and No. You clearly heard the sentence and responded in an understandable way. But watering an orchid is not like watering other plants. With a newly planted tree you dump a bucket of water on it every day for a month. That would kill an orchid. What it mean to water a plant correctly depends on the plant. And although 'to water' does not mean 'to water correctly' (otherwise watering incorrectly would make no sense), it does carry some implication along those lines. If I ask someone to clean my cell phone and he puts it in the washing machine he has not understood what I meant by 'clean', what kind of action I was asking for. (Or perhaps he has understood but wishes to destroy my property and get away with it on a technicality.) But the speaker doesn't get to dictate the meaning of words by an act of intention. It is quite possible to misspeak, speak carelessly, etc.

  6. The grammatical meaning of the sentence "How are you?" in the assumed "casual" context (ie, when one doesn't really want to know) is indeed not the word "fine", but its ir-meaning is the utterance "fine" (or any equally terse response). The two meanings often are highly correlated but can be quite different. Suppose B has been trained to bring a slab whenever B shouts the sentence "No need for a slab!" That B consistently delivers a slab in response would suggest that B understands the sentence's ir-meaning but would indicate nothing about B's understanding of its grammatical meaning. Which is to say that although speakers don't get to dictate the grammatical meaning of sentences, they can dictate the ir-meaning either explicitly by prior agreement with hearers or implicitly by playing on hearers' learned responsive dispositions. The latter is precisely what advertisers and politicians routinely do when they create grammatically correct slogans the ir-meanings of which are to activate learned behaviors.

    Addressing Duncan's and Lars' s objections to ir-meaning unpersuasive revealed a possible deficiency in ir-meaning (and other concepts of meaning as well). Consider a student who has demonstrated "knowledge" of a certain fact prior to an exam but freezes during the exam and has no confidence in the answer given. If the answer is correct, does that show that the student actually did "know" the correct answer at the time of the exam? My strict Sellarsian answer has been "no" since the examinee was unable to justify the answer at the time of the exam. Eg, in a multiple choice exam, there's typically a significant probability of guessing the correct answer, so it's clear that any given correct answer has no justification value. But if the exam is fill-in-the-blank, the justification may be considered implicit in a negligible probability of guessing the answer. Correct essay answers are explicitly justified. Therefore, I now think that whether the student "knew" the correct answer may depend on the test structure.

    The analogous issue with respect to meaning can be stated as: what can one conclude about a hearer's knowledge of the meaning of a speaker's sentence based on any single response by a hearer? And again the answer would seem to be that strictly speaking, the hearer can be judged to "understand" (ie, "know") the meaning only if the hearer not only responds as intended but also can justify doing so. But here too, justification may be implicit if the possible responsive actions are sufficiently numerous to make the chance of a lucky guess sufficiently low or if a hearer's responses to numerous utterances of the sentence are consistently correct. Applying this augmentation seems to resolve ambiguities identified in some of Duncan's and Lars's examples. Eg, in Duncan's example of nonsensical shouting, the fact that the hearer dropped something provides no information about whether the shouting was understood since we can neither be sure that the hearer wasn't just startled nor estimate the number of viable alternative responses.

    Other examples raised the issue: can one conclude from a response that is not the one intended by the speaker that the hearer did not understand the speaker's ir-meaning? And the answer seems to me clearly "no". The subject may understand a speaker's ir-meaning, for some reason not want to be seen as understanding it, and therefore purposefully and consistently not respond as intended. In general, I don't see any way to get from observation of that behavior to a conclusion that the hearer has understood the speaker's ir-meaning.

    Finally, the ir-meaning of a sentence is obviously a feature of that sentence, so I don't see that "misspeaking" - which I assume means uttering some other sentence - is relevant.

  7. Thanks, Charles. Adding in the part about justification (not just responding the right way but being able to explain why) helps avoid some problems. But can we always give a justification or explanation? I'm not sure that we can (although I'm struggling to come up with a good example, so I don't blame you if you aren't convinced).

    The ir-meaning is meaning in the following sense: the "meaning" of a sentence is the response intended by the speaker - and expected, owing to "training" (possibly implicit) of the speaker's audience. You also say that this is a feature of the sentence. I take Lars's point to be something like this: Imagine I have several orchids and that one of them needs to be watered. I ask you to water it. But I mis-speak in the sense that I don't think to specify which one it is that needs to be watered. Perhaps to me it's obvious. Or perhaps I have forgotten that there are other orchids in the same room. Or perhaps I accidentally say "orchids" (plural) when I mean "orchid" (singular). If you then water all the orchids you don't seem to have misunderstood the actual sentence that I uttered. On the contrary, you have done exactly what I asked you to do. Similarly, if there are several orchids and I ask you to water "the orchid" then you have not misunderstood the sentence if you do your best to water the one I mean but in fact get it wrong. The fault lies with me and what I said. That is, you understood what I said but I failed to say clearly what I meant.

    If ir-meaning is a matter of what I mean then I think I understand the idea, but if it's a feature of a sentence then I'm not sure I understand. Unless ir-meaning applies to sentences-as-meant.

  8. Duncan -

    You are quite right - in saying "feature of a sentence" I, uh, "misspoke"!! ir-meaning is a feature of a speaker's use of a sentence in a specific context.

    If I use a sentence with the intent that you respond a certain way and you don't, you have not understood my ir-meaning. But I think instead of calling that misspeaking, I'd just say that I used an ineffective sentence. Suppose I intend that you water one of three orchids and say "Please water the orchid". If you water the wrong one, you clearly didn't understand my ir-meaning. But as I argue above, I don't think it can be assumed that you understood even if you do water the intended one! That would be equivalent to a correct answer in a multiple choice quiz. And if ask you to justify your choice, you can't. I chose a bad sentence for evoking the desired response and just got lucky.

    OTOH, suppose that I ask you to water the two driest of several orchids and that I know which two are driest. If subsequently I find that you watered the right ones, the unlikelihood of doing that by chance may be sufficient to justify the conclusion that you understood my ir-meaning.

    As you can tell, I don't see the requirements for justification as necessarily being either precise or stringent. But it seems like some sort of justification is needed.

    As you can also tell, I don't think understanding grammatical meaning has a role here. "Please water the orchid" has a clear grammatical meaning, but is a bad choice in the multiple orchid context. But I'm not sure why that's relevant. If for some reason I thought saying "Please kill the two driest orchids" would evoke the response that you water them, again one might argue that you have understood my ir-meaning if in response you do water them. But the grammatical meaning of my sentence seems to play no role.

  9. Thanks, Charles. I don't know whether I disagree or not, but it all sounds a bit automaton-like. With a multiple choice quiz all you can go by is the answer, which might be a guess. But couldn't there be all kinds of signs of understanding in an act of plant-watering other than the act itself? There seems to be more to life than stimulus-response, although I suppose if you define the terms broadly enough there might not be.

  10. I don't really mean to suggest that S-R is all there is. I'm just exploring how far you can go modeling mental activities that way, mostly in an attempt to eschew the "mentalese" vocabulary, many words of which strike me as doing more harm than good.

  11. Do we really need an expression like “ir meaning”, over and above the expression “what the speaker meant the listener to do”? By the way, “ir meaning” doesn’t capture all aspects of what a speaker means by her words in a particular situation of speaking. If I say, “What a warm day!”, I don’t envisage any specific response, or necessarily any response at all, from my audience. Yet my words may have a straightforward meaning as uttered in that particular context (uttered, say, as I wipe my brow).

  12. One response to misuses of language is to introduce new terms, or new ways of using old terms. And sometimes these catch on. But another approach is to show that the existing terms are perfectly all right when used properly. That's the Wittgensteinian approach, although there's no need for everyone to agree with him, of course.

  13. Duncan -

    My concern with mentalese isn't "misuse" in a sense that could be corrected within a language game by using words properly. Rather it's using those words in language games for which they're inappropriate. Mentalese is appropriate - and preferred - in casual conversation. It would be ridiculous to say in casual conversation "Given the present weather context I have a behavioral disposition to take steps to avoid getting wet". But some of us consider that the neurological underpinnings of the concept "belief" are indeed context-dependent behavioral dispositions and that in formal conversation it might be better to speak in those terms instead of saying "I believe it's raining".

    Lars -

    The intent behind "ir-meaning" wasn't to create a neologism, just shorthand for the purposes of this discussion. Never used it before, perhaps never will again.

    Your utterance clearly has grammatical meaning, but if no response is expected, what's its purpose? Wiping your brow achieves much the same result. You might see it as merely a move in a social protocol, but even so there may be an expected response: "Yup, it's a scorcher."

    Which reminds me of a point I haven't yet made about "responses" - they can be latent rather than immediate. Someone telling an audience that GDP is expected to increase 2% in 2Q13 doesn't expect everyone to immediately take some overt action but must expect the statement to have some effect that will become manifest in the future. Otherwise, why say it? Similarly, a social protocol has to end, so the expected "response" to the last utterance may indeed be only to turn or walk away.

  14. Thanks, but what's a formal conversation? Do you mean something like technical work in psychology? I can imagine working on animal behavior, say, and wanting to avoid talking about beliefs rather than observable forms of behavior. And we might do this in anthropology too. It could be useful. But what's appropriate or inappropriate surely depends on what one's purpose is. I can't imagine a context in which it would be appropriate to say: "Given the present weather context I have a behavioral disposition to take steps to avoid getting wet." (Or: the only such context I can imagine would be one in which I was making fun of behaviorism.) Maybe if I was learning to talk that way in order to practice my anthropological technique. Is that the kind of thing you mean?

  15. By "formal conversation" I mean one with participants who are mostly specialists in a discipline relevant to the topic, eg, your blog. Another example is the comment section for this post in which I tried to analyze Eric's six cases in terms of behavioral dispositions. Unsuccessfully, to all appearances; as here, other commenters were polite but clearly unconvinced.

  16. Got it, thanks. I suppose I'm more concerned to understand ordinary language in terms of ordinary language than to use a specialized language to understand reality. Perhaps if it were clear to me that my choice of words was causing problems then I might prefer to use other words, but of course I think I'm right, not confused.

  17. Perhaps an application closer to home would help - PI §51.

    To simplify (and hopefully clarify), define a "color name" as being any of the four symbols "R", "G", "B", "W", and define a "color paradigm" as any surface which when illuminated with "standard" (according to some definition) lighting emits light with one of four readily distinguishable spectral densities, each bearing one of the four color names. An "element" of the 3X3 matrix defined in §48 is then a particular color paradigm.

    One issue raised in §51 is the nature of the correspondence between a color name and an element "in the practice of the language" of §48. One answer posited is that a student learns via ostensive teaching to produce (using any of multiple media: uttering, writing, pointing , etc) a different color name when visually attending to an example of each color paradigm. The student would later learn to produce sequences of nine color names in the prescribed order when visually attending to a matrix of elements.

    Having successfully completed that training, the student could be described as having acquired a family of behavioral dispositions to produce either a color name when attending to any single element in a matrix or a sequence of color names when attending to a complete matrix, all in the context of the language game.

    §51 also alludes to a second answer re the nature of the correspondence, viz, that participants have been trained to respond to color names by producing a color element. §51 posits that a student might do this by having a mental image of the element (or perhaps some other color paradigm) "come before [the] mind". (What observable action might accompany this is not addressed.) §53 describes an alternative in which a student is trained to respond to a color name by pointing to a sample element that is paired with that color name in a chart. Any observable responses would constitute another family of learned context-dependent behavioral dispositions.

    It's reasonable to ask what might be gained by expressing the situation in that vocabulary. For one thing, it (again, hopefully) eliminates terminological ambiguity. In particular, it avoids the questions what and where "color" is. The word is used only as a qualifying adjective, not as a noun. Nor does it use "seeing", another word that is ambiguous when you "look at what really happens in detail, as it were from close up". We usually take "seeing" a color to be having a certain phenomenal experience - mental image - and take "recognizing" the color to be the result of comparing the mental image to something analogous that is recalled from memory. It's a controversial position, but there is reason to suspect that mental imagery is at most epiphenomenal. If so, then "seeing" in that sense is irrelevant to the issues at hand because it can't cause responsive action. Expressing the language game activities in terms of behavioral dispositions avoids such complications since dispositions can be envisioned as being selective neural structures, each of which is excited by neural activity consequent to a specific to sensory stimulus and responds by issuing specific motor neuron excitation. In that view, mental imagery plays no role.

    Wanting "to understand ordinary language in terms of ordinary language" is fine as long as one sticks to ordinary conversation. But discussions of phil of mind seem to me far from ordinary, and ordinary language doesn't seem up to the task. Everyone knows what "to see a color" means in ordinary language, but in a technical forum much of what everyone "knows" is arguably wrong. And I don't see how a "technically correct" (ie, arguably more accurate) description can be fashioned out of the vocabulary of ordinary discourse.

  18. This might be why I work more on ethics than the philosophy of mind. But it also has to do with ideas about the nature of philosophy and philosophical problems. I don't think of philosophy as dealing with technical issues. Not that there's anything wrong with technical issues, and not that there's anything wrong with introducing technical vocabulary when dealing with them. But philosophy done in the manner of the later Wittgenstein is not like this. At least not usually.

    What if we forget about Wittgenstein and just try to deal with problems the best we can? I don't mind technical terms, but when to use them depends on when they will help. And if they are said to be helpful because they will allow us to avoid ambiguities, I would want to know which ambiguities are causing problems, and what those problems are. If I am alleged to be confused, I would normally take some convincing of this. If the person talking to me says that he gets confused and therefore prefers to use a certain vocabulary then I might go along. It's all going to be context-dependent.

  19. Well, I can't opine on philosophy in general since I'm at least 99% ignorant. My focus is on "phil of mind", which perhaps should be called something else since I see many current issues addressed under that label as inherently "technical" since they at least skirt brain physiology.

    I didn't mean to suggest that anyone else is confused, just that I find much of what I read confusing because the terms are often poorly defined at best. Eg, what exactly is a "belief"? Davidson (as I recall) says it's holding P true. OK, but what does "holding P true" mean? I can easily think in terms of a neurologically implemented behavioral disposition as evidence for ascribing something like what we seem to have in mind when using "belief", but I have no idea how to think about an attitude towards the truth value of a proposition.

    Anyway, I really appreciate exchanges like these. Many thanks for your patience!

  20. Sorry, I missed this comment somehow. I agree that it is not crystal clear what it means to hold something true, and that important terms are often poorly defined in philosophy. Better, more precise definitions can often help in such cases. I just prefer to avoid technicalities unless I need them.

    Thanks for your comments, by the way. I like these exchanges too.