by Dave Maier
One starting point for any philosophical account of language is that the truth of a statement depends both on what it means and on how the world is. Handily for contemporary pragmatists of my stripe, this fits neatly with the post-Davidsonian project of overcoming the dualism of conceptual scheme and empirical content. All we need to do is show that the two factors that make up truth are not so detachable as contemporary dualists claim.
If it were as easy as that, though, we’d be done by now. Last time I said some things about semantic externalism, the idea that our meanings and other mental contents depend in some way on how things are in the world (as opposed, that is, to being transparently internal to the mind in the Cartesian manner). While not uncontroversial (there are a number of versions of this idea, some of which lead to serious problems), this thought is not generally regarded as scandalously radical or insane – possibly because when it goes bad, it does so in the direction of realism, contemporary philosophy’s default metaphysical assumption. The world, and the semantic content it determines, turns out to be too independent of our minds for us to know for sure what we are even saying. But again, for most contemporary philosophers, metaphysical realism, even of a problematic sort, has always seemed preferable to the unthinkable alternatives.
Things get dicier, or can easily seem to, if we consider the converse thought: that how things are in the world depend in some way on our meanings and beliefs. Stated so baldly, the only people who accept it are the most hard-bitten idealists. Not only does this thereby fall off to the forbidden side, it’s not at all clear how to state it in any more acceptably hirsute fashion. (I except the obvious cases, the subject of an entire book by contemporary realist John Searle (The Construction of Social Reality), such as the straightforwardly conventional, mind-created, but thereby no less real, truth that this sawbuck is more valuable than that fin – although inventive if also perverse counter-examples are available even for that one.)
I won’t be arguing for any particular doctrine today, let alone anything controversial, but instead simply batting about some examples, in the hope of a better understanding of a few important and interrelated things: first, how diverse our semantic options really are, and how little the dictionary really tells us about them; second, how essential to meaning are the creative and expressive aspects of language use; and third, the overlapping and indeed interconstitutive notions of a) knowing what a word means; b) knowing how to use a word appropriately; c) knowing the word’s referent; d) knowing what such a thing is. With any luck this may clear the way to discussing matters of meaning and truth without the threat of linguistic idealism seeming to hover over us at every turn.
This background of examples is important here because none of them work as a proof, as philosophers typically look to examples to provide, that actual linguistic practice entails this or that philosophical doctrine. I aim instead for a Wittgensteinian focus on “intermediate cases” where, as Wittgenstein definitely does not say, the rubber hits the road. This procedure requires a certain tolerance for not knowing what to say which makes some philosophers uncomfortable or even angry. However, it may be that the place in which you are pulled in various different directions is the best place to see what’s going on.
Take for example the third issue I just mentioned. We do distinguish, in particular cases, between knowing what “davenport” means and knowing what a davenport is. For example, in speaking to you, I may attribute the latter knowledge to a third party, based on their experience and understanding of those things which we call “davenports,” without thereby, or ever, attributing the former (as that person may call such things “sofas”). That is, he knows perfectly well what they are; he just doesn’t know that that’s what they’re (also) called.
In other cases, especially depending on how the question has come up, that distinction may make little sense. Maybe the context is one for which the term “davenport” is essential. Lady Bingham demands of her servants that the davenport be moved into the parlour. If they fail to understand, without her pointing at what is directly in front of them, what she means, there doesn’t seem to be any important difference between the one sort of knowledge and the other. Here, if they knew the one, they would thereby know the other in the relevant sense. Knowing all there is to know about sofas wouldn’t help.
The intent and context of an utterance is, perhaps not surprisingly, often the key, and meaning may have rather different ways of playing the same practical role. Let’s say (counterfactually) I’m a connoisseur of Japanese cuisine and sushi in particular. I am thus acutely aware of the difference between what I might call real wasabi, which is made from the wasabi plant (Eutrema japonicum, a.k.a “Japanese horseradish”), and the green paste typically offered in non-insanely-expensive Japanese restaurants in the USA, which is made of (regular) horseradish powder, water, and green food coloring, and which I may call “imitation wasabi,” to make clear my preference or simply to show off.
If I am asked the simple question – a question seemingly about the world, rather than about what we say – “Is that stuff wasabi or not?”, I will answer “certainly not!” But it is uncontroversial that we do not need to talk that way, and even that I myself might not consistently do so, even while maintaining to the last that the distinction between the two substances is an important one. You and I have eaten at a particular restaurant many times, one where only the poor substitute is available, and you are well aware of and indeed tired of hearing about my preference for the finer stuff. Yet faute de mieux I still put it on my tuna roll; and when I need it to be passed to me I ask not for the “imitation wasabi” – that’s all there is, after all – but simply for the wasabi. If instead of passing it you instead feigned ignorance of my meaning, protesting that there is unfortunately no wasabi here to be had, you would be being perverse.
So is the stuff wasabi or not? There is no real puzzle here, of course: if “wasabi” means the stuff made from E. japonicum, then no; if instead it means the green stuff, of whatever origin, yet still capable of blowing out one’s sinuses, that one puts on one’s sushi, then yes. Still, given the form of the question, which presents itself as being about the world rather than about what we say, we can easily let ourselves assume that the world itself, and not our intention, decides which meaning is the real one, and which merely contextual or something. The former meaning, for example, makes a real distinction between types of thing, while the latter does not.
Or does it? It’s very easy to let our examples go before we’re done with them; yet with a little invention we can easily construct variations in which “wasabi” plays all kinds of different roles. Even now, when I ask for it to be passed to me, I distinguish between it and the pickled ginger, which is already within my reach. That I would say the same were “real” wasabi to be had doesn’t mean I’m not making a contextually important distinction here as well.
Or how about this. If I am making a film, in which one scene takes place in a Japanese restaurant, the background prop need not be “wasabi” of any kind, but instead some green stuff in a jar. Yet I will demand of the prop guys, in these very words, that the “wasabi” be to the right of the “soy sauce” instead of to the left. Here the word literally refers to that green stuff, even though no one will be putting it on anything. Maybe I even wish to distinguish between it and what I would in contrast call “real wasabi” – here meaning either E. japonicum or mere horseradish paste, between which it would now be pointless to distinguish – because neither of the latter film well, while the former looks just right.
It seems to be a general principle that one may use a word literally to refer to something which is (“merely”) imitating the normal referent of the word, which in the following case doesn’t even exist in the first place. Five trick-or-treaters are at my door, but I don’t see the fifth, standing in back, and only distribute four treats. In this case, the natural objection could perfectly well be “you missed the ghost in back” — and it is no proper reply to this to say “that wasn’t a ghost — ghosts don’t exist.” That is, though the dictionary doesn’t say so (and why should it?), it is a perfectly cromulent use of “ghost” to mean “someone dressed up (say, on Halloween) as ghosts are typically thought of: that is, he’s wearing a white sheet with eyeholes or something.”
Similarly, again, “wasabi” can perfectly well mean “‘wasabi’’/“imitation wasabi” (or even, given the proper context, “imitation imitation wasabi,” etc.). Consider this actual sentence Google provides in response to the question “What is the difference between horseradish and wasabi?”: “Most wasabi sold in the United States is just horseradish.” We might say: use erodes scare quotes. Or, as Davidson notes, metaphors die and become literal speech; although I am reluctant to say even that the use of “ghost” here is not literal. (Incidentally, I live for the day when “cromulent” itself makes the dictionary.)
One more group of examples should do us for now. One key principle of Kripke’s (orthodox-externalist) semantics – one which provides both the real anti-internalist power of his view as well as its equally real difficulties – is that the reference of our terms depends essentially on an initial “baptism” in which the coiner of a term points to a thing and says something like “Xs are those things.” After that (here’s where the hardcore essentialist realism comes in) the world takes it from there. Much as I may want to, I may not use the term properly to refer to anything but things with the same metaphysical essence as the original thing.
I hear an echo of this attitude, if perhaps only that, in the idea that etymology in general (that is, not necessarily a “baptism” of the Kripkean sort) determines proper reference. I referred above to sushi as Japanese “cuisine,” but the origin of the word (French cuisine, Latin coquina) suggests that cuisine is something which is “cooked” (as sushi is not). (Note: I have never heard anyone make this specific objection; but see below.)
Or consider gratuities. This word for “tip” comes from the French word gratuit, meaning “free”: I am required by law to pay my food bill, but I may reward or stiff my waitstaff as I please. But some restaurants include a service fee in the bill itself. In some sense this is the same thing, and in others not. As a matter of semantics, must I register that difference by denying that this service charge is in fact a “gratuity”? As a native English speaker I say no, and it seems that even Merriam-Webster leaves it open: “gratuity” sends us to “tip” in sense #10: “a gift or small sum given for a service performed or anticipated.” (Not, again, that I think we should let the dictionary order us around.)
Returning to the kitchen: I once got into it (evidence at my old blog) with a philosopher who claimed that since “biscotti” means “twice cooked,” nothing counted as a “biscotto” until it emerged from the oven a second time. Similarly, on this view, nothing counts as a “cookie” unless it has been cooked: balls of uncooked dough do not count. And indeed, if I ask you to bring me a cookie from the kitchen, the default (but of course entirely context-dependent) assumption is that since I want a cookie because I intend to eat it, if you bring me instead a ball of uncooked dough, I will protest that such a thing is not a cookie at all.
But of course it can perfectly well go the other way. If I’m (in the middle of) baking cookies, I guess they’re not done yet, nor are they so when the cookies are in the oven. Even if I am approaching the oven with a tray of what are clearly balls of uncooked dough, I may not be able to answer the phone right now because, like I just said, I am in the middle of putting a tray of cookies into the oven. (Consider also snickerdoodles, which seem to have no etymological obligation to emerge from the oven before they attain full snickerdoodlehood; yet of course snickerdoodles are cookies, so if something is the former then it is the latter as well.) Again, as a native English speaker, I take these things to be flatly uncontroversial; but maybe we need to get the experimental philosophers on the case. Will the idea of putting a tray of cookies into the oven really provoke vox populi to protest “why are you putting cookies into the oven? They’re already cooked, or they wouldn’t be cookies!” I’ll believe that when I see it. (Or, I guess, we could be making biscotti.)
So let’s say we agree that we have two semantic options here: that we may say, of a tray of balls of dough ready for the oven either a) that it is a tray of cookies, or b) that it is not (yet) a tray of cookies. This does not mean that it is metaphysically indeterminate whether the thing is or is not a tray of cookies. It does not exist in a quasi-quantum superposition of states (cookies/non-cookies) until someone comes along and says one thing or another, thereby making it so. (That would be idealism.) It simply means that to ask whether something is a cookie is not the same as asking whether the term “cookie” applies to it.
But a natural response to this is that if something is a cookie the term does indeed apply, and not otherwise! So it may not seem that we have gotten anywhere. Indeed, it may seem that I have deliberately courted confusion. But as I noted at the beginning, this is where we want to be: not in confusionland itself, but rather in a place where we can see it up close, really feeling the pull in different directions rather than giving the “wrong” one perfunctory lip service. In any case, this hard-fought vantage point seems to me to be preferable, for pragmatist purposes, to the sort of clear-cut yet arbitrary distinctions out of which all too many philosophical doctrines are constructed. But the proof of this pudding is in the eating; and right now it’s still in the oven.