Thursday, March 27, 2014

Is There a Text in This Class?

A little late, I finally read Stanley Fish's famous essay. I don't have a lot to say about it, but I think it's interesting as an example of what ideas like Wittgenstein's look like when carried over into another realm. For what it's worth, here's a kind of summary plus criticism.

Fish's main point, as I see it, is that meaning depends on context. I agree.

He also says that:
one hears an utterance within, and not as preliminary to determining, a knowledge of its purposes and concerns, and that to so hear it is already to have assigned it a shape and given it a meaning.
This seems unfortunate to me, and might be one source of the idea that individuals make meanings, an idea that I take it Fish would reject. [Update: It turns out I was quite wrong about this.] One thing that's unfortunate about it is that one surely does not in any clear sense hear an utterance within a knowledge of its purposes and concerns. I don't (necessarily) already know the purpose of your utterance before you have finished uttering it. A few pages later Fish goes so far as to say of someone that, "her words will only be intelligible if [her audience] already has the knowledge they are supposed to convey." It's pretty obvious that Fish does not mean what he is saying. That is, he does not mean that if I ask you what time it is then I cannot understand what you say unless I already know what time it is. What he means is that I will not understand what you say unless I understand the context well enough to know what kind of thing you are going to say. If you say "six thirty" then in certain circumstances I will know that you mean breakfast time, in others I will know that you are giving me the score in a football game, and so on. If I had no idea what the context was then I could not tell what you meant, if anything.

But in understanding the relevant context I am not assigning a shape to someone's utterance, nor giving it a meaning. As Fish himself says, "To be in a situation is to see the words, these or any other, as already meaningful." These words in this situation already mean whatever they mean. There is no assigning or shaping that I as the speaker or the listener can do to them. The relevant purposes and concerns, as Fish says, are already there.

So we don't (Fish says, and I agree) always interpret utterances. Sometimes we do and sometimes we don't:
I am not saying that one is never in the position of having to self-consciously figure out what an utterance means.
This self-conscious figuring out is, I take it, the normal sense of 'interpreting.' Fish does not deny that we sometimes have to interpret utterances, but he points this out because he is so far from saying that we always do. He explicitly rejects the assumption that there is always:
a distance between one's receiving of an utterance and the determination of its meaning -- a kind of dead space when one has only the words and then faces the task of construing them. 
I think that what Fish means is perfectly right, he just doesn't always express himself perfectly (who does?). This is a problem with his use of the word 'assumption.' Words that are uttered, he says, "are immediately heard within a set of assumptions about the direction from which they could possibly be coming." Assumption is an act, though, and assumptions are things that we make. The phenomenon that Fish is describing is more passive than that. There is no dead space in which one faces the task of making the necessary assumptions before trying to understand, or immediately--the assumptions having been made--understanding, someone else's words. There is something like assuming going on here, and we can call it assuming if we like, but it is not consciously chosen. Fish fairly clearly recognizes this. At one point he refers to one individual's assumption of what the concerns motivating someone's question could possibly be, but elsewhere he talks about "the understood practices and assumptions of the institution" (not of the individual) and of people's being possessed by "a structure of assumptions, of practices," which I think is closer to what he means and should mean.

In short, I think I agree with Fish.

UPDATE: For more on Fish see here (where I still think I agree with Fish but explain why I'm not entirely happy with some things he says) and here (where I finally lose patience). In other words, "I agree with Fish" is by no means my last word on the subject. And if you're interested in what I've written here then you'll probably be interested in what I've written in those other two posts.


  1. I remember liking this take on Fish:

    1. Thanks, I should probably read that. I'm not sure about this though: "For Fish, people “understand” or are “persuaded” by a position because it fits into the structure of beliefs already in play, not because they have been swayed by the “reasonableness” of someone's argument;" Why so many scare quotes? Are we supposed to give up the word 'reasonable', and if so, why? But, as I say, I should read the book.

    2. part of the implications of something akin to forms of life (in a familial resemblance sort of way) for neo-pragmatists is that reasonableness isn't distinguishable from, other than, fits in web of beliefs/expectations, more like collage than logic tho. So not a problem for folk-psychology to talk in terms of being reasonable (acts as a kind of honorific) but not a Principle either.

    3. That sounds fair enough. I guess that's just how I understand the idea of reasonableness. Maybe others think of it differently.

  2. "her words will only be intelligible if [her audience] already has the knowledge they are supposed to convey."

    I agree with you that this doesn't make sense and would add that I think it suggests a highly suspect concept of the purpose of speech, viz, to convey information. For what purpose? If I say to you "I was born in Olney, Texas", to what use might you put that information? OTOH, if I say to you "I recently saw movie X and consider it to be excellent" (in some way that I know is important to you), the purpose is quite likely to be that I intend to create in you a behavioral disposition to see the movie (assuming I don't know that you have already seen it). Ie, although speech often conveys information, it seems likely that the ultimate intent is to cause the hearer to take some action, either immediately or in the future. As you note, exactly what action depends on the context, a multidimensional entity.

    In that view of the process, I enthusiastically agree with you that "one surely does not in any clear sense hear an utterance within a knowledge of its purposes and concerns". Only I would go further than you may be willing to go and suggest that the purposes and concerns of the speaker are in general totally transparent to a hearer. Here's why.

    The business about "shape", suggests something along the lines of W's PI comments about a place being prepared for a word in a language game. A speaker's intent in emitting an utterance will depend on a hearer's previous experiences with that utterance. If the speaker knows that in contexts similar to the present context the hearer has typically reacted a certain way, then the speaker's intent in making the utterance presumably is to evoke a similar reaction. Ie, a place has been prepared in the hearer's neuronal networks for that context-dependent utterance.

    And that seems a promising way to think of "understanding" an utterance: it is "understood" if the hearer reacts as intended by the speaker. But then a given hearer may "understand" an utterance in an unconventional way, ie, may react in a way different from the way many or most other hearer's would react. (Which introduces the possibility of a speaker's conning a hearer.)

    "{Fish] explicitly rejects the assumption that there is always... a distance between one's receiving of an utterance and the determination of its meaning".

    I agree with this but find it an awkward way of expressing what I take to be the underlying mechanism. We presumably learn to react to specific utterances previously heard in specific contexts in specific context-dependent ways. Heard in a new context, how to react to the utterance must be determined ad hoc. One can call the mechanism by which the reaction is effected "interpretation", but then it is necessary to explain in turn what one has in mind when using that word. As one with a background in pattern-recognition, I have in mind something along the line of nearest-neighbor processing in which a pattern recognizer reacts to a new pattern (in this case a familiar utterance in an unfamiliar context) as it would to the previously experienced pattern that is in some metric "closest" to the new one. My guess is that this idea can be given a (currently speculative) neurological implementation, although - alas - not by me. Rightly or wrongly, this is how I would translate Fish's vague "assumptions" (as described by you - I haven't read the paper) into something more tangible.

    1. Thanks. I mostly agree, but:

      1. it seems likely that the ultimate intent is to cause the hearer to take some action Yes, this is certainly plausible. But is it so likely that we can assume it without finding out whether this is the case? It's not something I would want simply to assume. Maybe you wouldn't either.

      2. If the speaker knows that in contexts similar to the present context the hearer has typically reacted a certain way, then the speaker's intent in making the utterance presumably is to evoke a similar reaction. On the one hand this is probably often true, on the other it sounds like a joke about how conversations in unhappy relationships go.

      3. Heard in a new context, how to react to the utterance must be determined ad hoc. Yes, although couldn't the use of a familiar utterance in a new context be a joke, say? That is, a use of language that prompts a spontaneous reaction, one not involving anything that would normally be called "determining how to react"?

    2. 1. I don't think "assume" is the right word here. As far as I know, all current models for meaning/understanding are hypotheses, in which case the task is to find instances of speech that fit the model and instances that don't. There are many instances that obviously fit a speaker-intended-action/hearer-actual-action model: imperatives, interrogatives, instructions, advertising, politicking, wooing, et al. As I suggested in my comment, even the information transfer model raises the question "what action is the recipient supposed to take in response to obtaining the information?". So, it's incumbent on one questioning the model to suggest instances that clearly don't fit it. For several years, I've tried to apply it consistently, and it seems to work - although sometimes requiring perhaps a little shoehorning to make it fit.

      3. What I mean by "determined ad hoc" is just that a hearer's response to a specific utterance-context combination won't necessarily have been learned previously in which case it must be synthesized (I suggest one possibility in my comment). In the case of a common utterance used in a context that suggests that it may be intended to be a joke, if that synthesis results in a laugh, the utterance was "understood" notwithstanding that the utterance in that context doesn't have the "meaning" it has in other contexts.

      The general idea isn't original to me, of course. Among other places, it makes several appearances in PI (eg, in the builder-helper example in the first few pages). I may have taken it a bit further than some in thinking about how the synthesizing might be effected non-computationally.

    3. I don't know whether I can find any instances that clearly don't fit, but what about saying "I love you"? Sometimes this is said with a very specific outcome in mind: "Honey, I know you picked up the kids from school yesterday so it's my turn, but could you get them again today please? I love you." In other cases, though, "I love you" is more like "goodbye" (or "hello"). What's the intended action in that case? What action do I try to bring about by saying goodbye, for that matter (when it isn't a way of telling someone to go)? I suppose words like these function as a kind of social glue, but I'm not (usually) thinking of that when I use them. If saying words made no difference then we wouldn't say them (generally), but that doesn't mean that we intend the difference in question. We aren't that calculating or aware.

      The same kind of thing goes for your point about determining the meaning of an utterance in a new context. It's not that I disagree, but you're using the word 'determining' in a non-standard way, it seems to me. For technical purposes that is fine. But I don't take my purposes to be technical. I'm not trying to get at underlying mechanisms, for instance. So I might be talking past your concerns.

    4. Well, this will be a reprise of past exchanges with Philip. et al. While I appreciate what I take to be the Wittgensteinian approach to language, speculations about how language works at a higher level need to be consistent with current understanding of relevant brain processes. Speculations based only on behavior should generally be fine. But once one goes "inside the head" ala some of the Fish quotes, they become fair game for challenge if they are based on suspect assumptions about brain functionality. Given Fish's rather literary style, it's hard for me to tell when his quotes cross that red line, but it appears to me that some do.

      I don't know what the "standard way" of using the word 'determining' is, but if it suggests some sort of "conscious" process of weighing this or that possible meaning and then in some sense "calculating" a response, I consider that use debatable. IMO, just as throwaway utterances like "hello", "goodbye", "how are you", "give me a call", etc, often are reflexive and have at most the intent of terminating an exchange, so are many of the more complex responses that I describe as "synthesized". In short, I take a rather deterministic view of our linguistic behavior.

      "that doesn't mean that we intend the difference in question. We aren't that calculating or aware."

      Good point. A recurring problem I detect in phil of mind is that the its vocabulary includes many words "stolen" from the psychological vocabulary, or worse yet from the quotidian one. In an attempt to avoid confusion, I often describe the neurological configuration to which a psychological term like "intent" presumably is meant to refer as being "a context-dependent behavioral disposition". But then that phrase typically requires further explanation since it isn't common, never mind standard. Sigh.

    5. Yes, it's tricky, and I'm sure you've had conversations along these lines before. Suspect speculations are fair game, I agree, but I'm trying to avoid any speculation at all. And I think it's possible to read Fish as being in line with my kind of project (so far as I have one), which is based entirely on familiarity with the language we know and therefore don't have to speculate about at all. But some people want to take this kind of ordinary, harmless thing and make an "exciting" theory out of it. If science can shut that tendency down then good luck to it, but I doubt anything can stop it.

  3. I had a belated moment of clarity this morning about our discussion of Fish and a particular context of his essay. Fish is responding directly to a supposed distinction between "objective" literal meanings and "subjective" interpretations. That specific hermeneutic pairing shows up in the student's question at the beginning of the essay: "Is there a text in this class, or is it just us?"

    Seen from this perspective, Fish is challenging the existence of "literal" meanings (a category that is not necessarily coextensive with "clear" language). Fish argues that "literal" meanings are interpretations whose character as interpretation is concealed by convention. When someone is sufficiently possessed by "a structure of assumptions, of practices," the conventions become transparent and convention-dependent meanings appear as if they are literal (fixed, objective, independent, self-supporting). The fish can't see the fishbowl (pardon the pun).

    This isn't a tremendous insight into Fish's argument, but I do like it as a thumbnail version of his position.

    1. Thanks. You know Fish's essay better than I do, but here's what I think is going on (correct me if I'm wrong). As you say, Fish is responding to a supposed distinction between "objective" literal meanings and "subjective" interpretations. Seen from this perspective, I would say, Fish is challenging the existence of "literal" meanings but not challenging the existence of literal meanings. That is, he rejects the literal-interpreted dichotomy, or rather one version of it (text vs just us). What he rejects, it seems to me, is one kind of understanding of literalness (a kind that I don't understand, but that he seems familiar with). I don't read him as rejecting all senses of the word 'literal,' so that we would have to stop using the word in order to be good Fish-ites.

      Rejecting one form of the literal-interpreted dichotomy does not mean embracing the idea that everything is interpreted (assuming an ordinary use of the word 'interpreted,' of course). I've tried to explain above why I think it's misleading to characterize Fish's position as involving the view that "literal" meanings are interpretations--he comes so close to denying that anything is an interpretation that he feels the need to say explicitly that he accepts the possibility of interpretation (i.e. of self-conscious figuring out what an utterance means). So it would be odd to say that he thinks all understanding is interpretation.

      I tried unsuccessfully to link to a passage from O.K. Bouwsma's book on Wittgenstein that I think is helpful and similar to the kind of thing Fish is saying. Bouwsma is recording a conversation in which people are discussing whether the sentence "I am here" makes sense. Of course the sentence has a use, but if it were just shouted out in any circumstances whatsoever would it have a use? Max Black says that there would be no point in shouting it out (except in unusual circumstances) but that if we had to choose between calling it True and calling it False we would clearly say it was True. According to Bouwsma, Wittgenstein said: "No! No! Of course not, etc. Context determines use." A sentence has no meaning at all, and is therefore neither true nor false, outside of a context in which it has a use. And what use or meaning it has depends in part on the context. I think that what you are trying to get at with the word 'interpretation' is something like this point. But I don't like that way of making the point because I think it's misleading, suggesting a kind of conscious, voluntary act that conveys meaning.

    2. [continued]

      Here's an example. Some people like flying Confederate flags. They often talk as though the idea that this is racist is just an interpretation, and that they don't mean it that way. One way to try to make this claim is to embrace relativism: all meaning is interpretation and my interpretation is just as legitimate as yours. Another way would be to make a literal-interpreted distinction and insist that the literal meaning of the flag is non-racist (just colors and shapes, perhaps, or a historical-or-political-but-not-at-all-racist symbol) while the idea that it is racist is a mere interpretation, a projection. I think Fish wants to reject both of these moves. (I certainly think he ought to reject them, but I also think he really does reject them.) Not every interpretation is equally good, and the objective or literal meanings of things are not immune to, for instance, the parts of history we might want to ignore.

      This might not be a great example, partly because I'm not sure that a flag can have a literal meaning. But I hope it helps make the point. Here's another example: the n word. Its literal meaning is unproblematic (I take it) but it's actual meaning is obviously far from benign. And this is not something that I would want to call an interpretation. The word is racist. Am I violating Wittgenstein's point about context in saying that? If it helps, let's juts say that the word would be racist in any context I can think of where I am the speaker. Is this a matter of convention? Of course. But is it a matter of interpretation? No! No! Of course not, etc. It is an objective and independent fact that it would be racist of me to use that word to describe or refer to another human being.
      I feel as though I ought to say something about Biblical and Constitutional interpretation now, but I want to go home. So instead I'll just say what I think Fish, Wittgenstein, and I would agree on. There are interpretations. There are literal meanings. But not all meaning is one or the other. Here I'll just speak for myself: most of what we say is not meant literally and is understood without interpretation. This is indeed because of structures of assumptions, practices, etc., but it is true all the same.

    3. Re-reading what you and I say here I realize that I'm not contradicting you at all. Which is good. We can reject "literal" meanings without rejecting literal meanings. And we probably should.

    4. I think that you are right: we seem to be saying very similar things but saying them in different ways. Even your reference to Biblical and Constitutional interpretation is in the same range of concerns that frame the conversation for me (it is interesting that the question of literalness pushes us both into the realm of foundational documents that order specific communities of practice, whether legal or religious).

      Your discussion of "relativists" and "literalists" w/r/t the meaning of the Confederate flag made me think of Lewis Caroll's Humpty-Dumpty. He assures us that "when I use a word, it means just what I choose it to mean--neither more nor less." Humpty doesn't fall neatly into either category: he is aware that his words could mean something else, but when HE uses them, they mean only and exactly what he intends. The joke is that we always say both more and less than we mean.

      On a side note, most Confederate flag apologists that I know--from my family on out--make an argument like Humpty-Dumpty's: when I fly the flag, it means precisely what I mean, neither more nor less ("heritage not hate," etc.). They do not grant that the symbol exceeds their intention, and they want to avoid responsibility for the pain and offense caused by that excess by blaming it on "reading too much in" or willful misinterpretation. I'm not entirely certain why, but that seems to me to be a variation of the literalist position (I don't think I've ever met a relativist in the sense you use it). Something like a "presence-literalist": this is the literal meaning so long as I am here to guarantee it.

    5. Yes, it is a Humpty Dumpty issue. We don't get to dictate what our words and flags mean. They mean what they mean. That meaning is dictated by the context, which is shaped by people, but individual people don't just get to choose their own meanings.

      (By the way, if you ever want to go in on a business selling "Heritage of hate, not hate" bumper stickers let me know.)

    6. I think Humpty is being treated a bit dismissively here. I agree in general with the assertion that "meaning is dictated by the context". But that applies to the words "meaning" and "context" as well, so I don't think the matter is quite as clean-cut as the assertion suggests.

      In a context in which some terms are formally defined, the "meaning" of such a term is it's definition and therefore is unambiguous; eg, "group" in mathematics. And even in informal contexts, the conventional meaning of many words is unambiguous; eg, "house" (although "home" might be not be).

      But if one takes a speaker's "meaning" to be the intended hearer reaction (as described in my first comment), then "context" includes the psychological state of the speaker, including the speaker's assumptions about the psychological state of the hearer. In which case Humpty is right: an utterance of his "means just what I choose it to mean". Even so, I'm still inclined to agree with "meaning is dictated by the context". But only in the sense that if "context" is sufficiently inclusive, some concept of determinism applies, in which case Humpty isn't really "choosing". But I rather doubt that's what you had in mind.

    7. I agree that "meaning is dictated by context" is not the whole story. We can choose to define words as we like, as when we give a technical meaning to a term like 'group' or 'interpret'. But then the rules of the game say that we have to make such definitions explicit and upfront. Even then there are limits. I can't, for instance, avoid racism by claiming that by some racist word I mean something innocent and then going ahead and using the racist word.

  4. In the present context, describing meaning as "convention[al]" and "institution[al]" can be taken as equivalent in that both words suggest establishment by community consensus, which is also a way of interpreting "objective". Although since I despise and eschew "subjective", in the interest of equity I try to avoid "objective" as well. Perhaps I should add "literal" to the list.

    1. You're right about functional equivalence of "conventional" and "institutional" in this context, and I think that Fish would agree with "community consensus" as the underlying principle (much of his work argues explicitly that interpretation is always produced within and limited by the horizon of one "interpretive community" or another).

      I have deeply ambivalent feelings about "subjective" and "objective." They are fraught with conceptual baggage, but they can be useful for thinking about poles of experience and modes of inquiry. I have no such ambivalence about "literal." It's a self-deceptive and dangerous term, and it is most often used as a cudgel.

    2. Really? "Literal-minded" is a insult, but I don't think that's the kind of cudgel you mean. Is it bad to tell the following joke?

      It's hard to explain puns to kleptomaniacs because they always take things literally.

      Words are (mostly) OK. It's people that are the problem.

    3. It might be bad to tell that joke, but for different reasons.

      I suppose we should say "words don't cudgel people, people cudgel people." It has, however, been my experience that people frequently cudgel each other by insisting on "literal readings" of certain texts (and concepts).

      Ironically, the most common use of "literally" these days is to signal they mean something figuratively. "My head literally exploded." "That must have been messy."

    4. Yes, I started drafting a post about similar thoughts this morning. Maybe I should post it.