Deep Thoughts XXXV

Meeting the minimum requirements is always good enough.

[This is a variant on XXXIV. There are times when we say things like "The minimum is not good enough." When we do that, what I think happens is that we have a context shift. "The minimum" is understood relative to one set of ends or norms while the "good enough" is understood relative to another. One kind of a case is where you're competing for a job. Meeting the minimum required qualifications is good enough for being hirable in principle (if it's not, the minimum requirements were incorrectly stated), but is not good enough for beating the competition. Another kind of case is where someone is being evaluated in a number of areas (or with respect to a multiplicity of assignments). In each area, there is a minimum requirement. Meeting that requirement is good enough for not failing according to that requirement. But there may be a second, meta requirement to exceed the minimum in most of the areas. (There cannot coherently be a requirement to exceed the minimum in all areas. For if there were such, then the "minimum" in each area would not be a minimum requirement but a maximum disqualifier.) In any case, when we keep the context constant (and as a rule in natural language context in short sentences stays constant), "The minimum is not good enough" is a self-contradiction.]

Context

It is usual to distinguish between a piece of language and its context. This distinction is, I think, bogus. Whether "flies" in "Fruit flies like butter" is a verb or a noun is determined by context (a "typical" context is still a context, just as a flat delivery is an intonation). Consider some ways one might try to make the distinction:

1. "Context is everything over and beyond the sentence." But this is useless unless we have an account of what "the sentence" as distinguished from the "context" contains.
2. "Context is everything over and beyond what is written or pronounced." But this is false—since a part of the context may be what was written or pronounced earlier. And if one amends to say "what is written or pronounced in the sentence", the the problem in (1) comes back. Besides, this definition would make everything in sign language count as context. There is an infinite variety of ways that linguistic expressions could be realized, and there is no hope of listing them all to make a definition like this work out.
3. "Everything that is linguistically relevant but which the speaker might not know about while responsibly uttering the sentence is context." But a speaker might not know what words she is using. For instance, the discovery that a linguistic sequence can be broken up into words is a genuine empirical discovery, and a speaker might not have it. It is normal, further, for the speaker of a language to fail to know exactly the sounds she is making. (In Russian and Polish, final voiced stops get devoiced; thus, a final written "d" is pronounced "t"; however, native speakers often do not know this, and falsely believe they are pronouncing a "d"!) Moreover, there will be aspects of what is traditionally called "context" which a speaker had better know about, because it determines the parsing of the sentence (e.g., "Fruit flies like butter").
4. "The sentence is what is intentionally produced by the speaker; everything else is context." This is dubious, because one can utter a written sentence simply by cutting out an apposite sentence from a newspaper story and mailing it to someone, we may suppose with a signature appended. In that case, one has asserted whatever was in that sentence, but one has not intentionally produced that sentence, unless "produced" includes exhibiting, taking up, making relevant, etc. But if "produced" is understood in such a wide way, then what are traditionally thought of as parts of the context may well be "produced", in that the time at which one speaks, the way one's face looks, etc. can often indicate which parts of the context are relevant.
5. "The sentence is what is chosen by the speaker's intention; everything else is context." But just a speaker can choose the precise wording, she can choose the precise context, either by producing it or by waiting for it. Of course, sometimes a speaker does not choose the context and may be even unaware of it. But likewise, sometimes a speaker does not choose the words and be even unaware of them (e.g., consider cases of misspeakings or cases where habit rather than choice determines the wording).
6. "Context is that which is not governed by conventional rules of grammar." This is either trivially true, if we take "grammar" to include only the rules that govern things other than context, or else is simply false—for there are rules that involve the interaction of wording and context.

"Now" and other indexicals (Language, Part IV)

In English, the pitch and rate at which one speaks typically do not affect the types of the tokens that one is using. Whether you say "home" quickly or slowly, with a low or a high pitch, your utterance is a token of one and the same type. But this need not be a universal truth. One can easily imagine a language where an utterance of "home" means one thing when spoken quickly and another when spoken slowly. In that case, there are two word types here, and which type one's token falls under will be partly determined by the sequence of phonemes and partly by the rate at which one speaks. (We might represent the two word types in writing as "home" and "h-o-m-e" if we wish.)

A language provides a mechanism for classifying linguistic tokens under linguistic types. There are very few restrictions on how the classification scheme can work, except contingent ones derived from our recognitional abilities. Imperceptible differences between tokens will not do the job.

It is quite possible, then, to have a language where the classification system is time-dependent. Thus, "home" at odd-numbered hours of the day means a tank, and at even-numbered hours it means an F-16. There are, thus, two word-types with the same phonemes, and to distinguish between them you have to check what time it is. Such a language might well be useful for confusing an evesdropping enemy.[note 1]

Imagine now a language L1 where the sound "chow" when uttered at a time t denotes the time equal to t+7minutes, when t is during an even-numbered hour, and denotes the time equal to t-7minutes, when t is during an odd-numbered hour, and where the type of the word is identified by both the sound "chow" and the time of utterance. This language can be understood as containing continuum many word-types, identified partly by the time at which the word-type is tokened and partly by the phonemes. This is a very odd language, but a possible one.

Observe that in L1 no utterance of "chow" is the utterance of an indexical. What an indexical refers to depends on both the token's type and on the context of utterance. But what "chow" refers to does not depend on the context of utterance, but only on the token's type. The token's type depends on the time of utterance, but that is a different matter.

Consider now a language L2 where the sound "fow" when uttered at a time t denotes the time t, and where the type of the word is identified by both the sound "fow" and the time of utterance. Just as no utterance of "chow" in L1 was an indexical, so no utterance of "fow" in L2 is an indexical. Rather, it is the utterance of a fine, upstanding, context-free referring term.

But now a question: How do we know that utterances of "now" in English are utterances of an indexical? Why not analyze utterances of "now" in English precisely the way utterances of "fow" are to be analyzed in L2? There are, I submit, no facts of linguistic practice (normative or not) that allow us to distinguish between English's "now" and L2's "fow". If linguistic facts supervene on facts of linguistic practice, there is no fact of the matter whether an utterance of "now" should be read as an indexical whose type is identified by the phonemes or as a non-indexical whose type is partly identified by the phonemes and partly by the time of utterance.

If we understand "now" along the lines of "fow", then any argument for the A-Theory of time based on our use of "now" is likely going to fail. For "fow" is perfectly at home in the eternalist world of the B-Theory. And what I said about "now" goes for tenses as well.

This strategy is closely parallel to the old failed B-theoretic attempt to translate "now" into the time of utterance. That attempt failed because when one translated "It is now 11:56 am", it translated into "It is 11:56 am at 11:55 am", and hence a sentence that one could reasonably be wrong about got translated into one that no one could be reasonably wrong about, which is absurd. On the present strategy, an utterance of "now" at 11:56 am does refer to the 11:56 am, indeed is rather like a proper name for it. In a sense "It is now 11:56 am" may be a tautology, but it is not a trivial tautology. Rather, it is like "Cicero is Tully" or "London is Londres."

If all this is right, then no deep facts about language hang on the distinction between indexicals and non-indexicals. There may be more than one way of classifying bits of utterances into types, and for any way of classifying that makes a bit of utterance into a token of an indexical, there is a way of classifying that makes that bit of utterance into a token of a non-indexical identified in some non-phonemic way. Each classification should give rise to the same proposition as expressed by the utterance as a whole.

Taking up into language (Language, Part III)

We can take items in the world up into language. An easy, though very unlikely, example would be if I came across a bunch of rocks randomly deposited by an avalanche, and noticed that they were arranged into the shape of the word "here". I then took more rocks and added to the left of the ones already there "There was an avalanche" and then to the right of them I put a period. The resulting production would count as my inscription of the sentence "There was an avalanche here." That one of the words in the sentence was not shaped by me is irrelevant. I took something in nature which wasn't a word though it was shaped like a word, the word "here", and made a token of the word "here" out of it, without actually touching it. The bunch of rocks thus got taken up into a linguistic item, much like a stump can be taken up into being a chair just by my treating it as a chair--which can happen without my even touching it (e.g., I can treat it as a chair by trying to sit on it, but tripping forward accidentally).

Likewise, word tokens from one language can be taken up into linguistic items in another language. I can write a note by gluing in words cut out from newspapers, including foreign newspapers, and I can do so with no regard for what meaning the snippets had in their original language. Thus, I can take a German newspaper story about a poison, and cut out the word "GIFT" (which means "poison") from the headline, and then take some words from an English-language newspaper, and put together the note "YOU HAVE BEEN A GREAT GIFT IN MY LIFE." And, no, I wouldn't be saying that you've been a poison! The snippet would thus have been taken up into a new linguistic item. The original meaning of the snippets I use is irrelevant. New linguistic items are made of them.

We could say that the rocks in my initial example already were a token of the English "here" even before I got to it, and the token "GIFT" in the German headline was already both a token of the German "Gift" and the English "gift". I suspect this is mistaken. It makes sense to ask of a sentence token in isolation what language it is in, even if the orthographically or phonologically same sentence could figure in more than one. So that won't do. My suggestion is that we are mistaken in identifying the token with just the inscription. The token includes the inscription, but includes the context--and hence the intentions of the speaker--into which that inscription was taken up. If this is right, then once again linguistic tokens are more than just inscriptions and sounds, but include context, as in my earlier posts on language.

Syntax and context (Language, Part II)

To resume my attempt to erase the distinction between utterance and context, I shall argue that to judge whether a sentence is syntactically correct can require information about what looks like "context".

First, note that the fact that a given uttered sequence of sounds is in language or dialect L₁ rather than in some other language L₂ is surely very much a contextual feature, and one that has to be known in order to judge syntactic correctness. It may be that the speaker announced which language she was speaking, or that she is assumed to be speaking the same language as her interlocutor previously was speaking, or that she is speaking the majority language in the culture. Often one can guess from the words said, or the accent with which they are said, what language is being spoken. Often, but not always. Consider a case where we have two closely cognate languages. Then, one will sometimes have a case that the same set of sounds could be parsed either as a correct sentence of one language or as a somewhat mispronounced sentence of the other language. Which is the right interpretation depends on which language was contextually established as the one in which this utterance is being, and this in turn will answer the question whether the utterance was syntactically correct. So, if we recognize such a thing as context at all, we should likewise recognize as part of the context the fact that a givien language was in play, and hence we should conclude that context is relevant to syntax.

Second, we can come up with some somewhat odd-ball cases. You say: "Jones was walking" and I add "under the bridge". Whether what I said was syntactically correct depends on what you had said--had you said nothing, my addition would have been nonsense. So, what you say helps determine syntactic correctness.

Third, it seems to me that in gendered languages to use the wrong gender in words that refer to the speaker is to make a syntactic mistake--but then whether a sentence is syntactically correct will depend on whether the speaker is male or female, an apparently contextual feature. We can imagine even more radical versions of this--we can imagine a language where, say, completely different word order is to be used by men and by women. Similarly, it seems a syntactic mistake for a collective to speak in the first person singular. However, I am aware that there are other ways of interpreting the gender/number case (one might say that all the sentences in question are syntactically correct regardless of who says them, but there is some other kind of error in them).

So what? Well, if we need to know what we normally think of as context to determine syntactic correctness, then it seems that the choice of the setting in which we utter a set of noises is just as much a linguistic choice as the choice of what noises to make, because the setting and the noises interact to produce a syntactically correct or incorrect sentence.

Frankfurt and functionalism

We learn from Frankfurt counterexamples to the Principle of Alternate Possibility that we should not use counterfactuals and nomic modality to characterize intrinsic features of things. What Frankfurt cases show is that it is frequently possible to modify counterfactual and nomic modal properties of a thing, event or process without actually causally affecting the thing, event or process in itself.

This is a useful lesson. Here is one interesting application: standard functionalist theories of mind cannot be right. I will give the roughest sketch of the argument. Start with two observations:

1. Functional characterizations of a system depend crucially on the system's counterfactual properties. (An and-gate is a system of which it is true that it would produce a 0 given two 0's or one 0 and one 1, and would produce a 1 given two 1's, even if in the course of the system's functioning it is actually only fed one of these options.)
2. That a system exemplifies a functioning mind is an intrinsic property of the system.

Now, what we learn from Frankfurt cases is that we can radically alter the counterfactual properties of a system in just about any way we wish without actually causally interacting with the system, and hence without altering its intrinsic properties. It follows from this observation and (1) and (2) that whether a system exemplifies a functioning mind cannot be a matter of the system's functional characterization. (We can, for instance, make sure that a given logic gate could never have produced a different output from what it actually produced, because someone watching the system would have intervened and forced it to produce the output it actually produced, had this watcher seen the inputs being different. But if this were done, then this would not be an and-gate, as it would not have the counterfactual properties of the relevant kind of logic gate.)

I should say that there is a way out of this argument, and it is to embrace an Aristotelian functionalism that instead of characterizing functions counterfactually, characterizes them teleologically. But that is not what "functionalism" means in the context of the theory of mind. (One might also try to do the Aristotelian move and define the functionalism evolutionarily. But not hard to see that this fails [e.g., see this, or this, or just extend the argument here].)

Or maybe we could try to get out of the argument by supposing that we can cut a system away from its environmental context and characterize the functioning of the system by looking at how the system would function apart from its environment, thereby isolating the system from the kind of purely counterfactual interference that a Frankfurt-style watcher would impose on it. But that simply doesn't work. How does the human body work absent its environmental context? It simply dies, pretty quickly, when not supplied with air, food and water.

Of course one might try to define a notion of an appropriate environment and then calculate the counterfactuals relative to that. But if we are allowed to be completely free in choosing what environment counts as appropriate, then just about everything is up for grabs. And if there is an objectively appropriate environment for the system, then again we get some version of Aristotelian teleology, which is not what functionalism wants.

Communication boards and indexicals (Language, Part I)

This is the first in what may be a series of posts developing a view of language that erases the distinction between language and context. The view may self-destruct before it's fully developed. I'm having fun here. No originality is claimed.

Some disabled people communicate with a communication board. A communication board is a board with printed pictures, some representing objects like shoes and chairs, some representing verbs like sitting, and some representing emotions like happy or sad. One communicates with a communication board by pointing to a sequence of pictures. High-tech communication boards will say the word, but it's important to my argument that I be talking of a low-tech one, which is just a pre-printed board. If a communication board is sufficiently large and extensive, and there is sufficient syntactic structure in the order in which one points to the pictures, this will be a language.

Let's say that the pointing is done with a finger. Now, what are the words or other linguistic units in this language?

Here is a really bad suggestion: The words are constituted by pointings with a finger and two token pointings count as of the same type if and only if the finger directions in the two pointings have the same relationship to the natural axes of the speaker's body (within some measure of precision; I will use the word "speaker" regardless of whether a language is spoken or not). The pictures on the board, on this suggestion, are simply context.

What's wrong with this suggestion? Well, for one, it means that if the speaker points to parent, fruit and happy, expressing (let's suppose) the proposition that the parent is happy with the fruit, and I shift the board over by an inch, and the speaker again points to parent, fruit and happy, then the speaker has used different word types, because her pointings are now in different directions. That is absurd--surely the speaker has said the same sentence.

Another way to see the absurdity of this view is that it will be impossible to give a story about the syntax of the language in terms of the arrangement of word types, since whether a given sequence of finger pointings, identified by direction relative to the speaker's body, is syntactically correct depends crucially on what the pictures pointed to are, and not just on the angles (again, think of a case where the board gets shifted over). But we don't want context to be the primary determiner of syntax!

One might think that the mistake in this story is that it is not the angles relative to the speaker body that matter for identifying the word type, but rather the direction of the finger as measured in some natural coordinate system based on the configuration of the board. (E.g., run the x-axis along the horizontal side of the board, the y-axis along the vertical side, the z-axis upward from the board, and then specify the cartesian coordinates of the tip of the finger and the finger's big joint.) But that's silly, too. Suppose that the speaker's board gets upgraded by getting a few new pictures, and with existing pictures moved a bit to accommodate the new ones. The speaker's language, thus, becomes extended. But now if we identified word types in terms of the coordinates of the finger relative to the board, the same sequence of finger positions as before would now be expressing something completely different. More seriously, previously syntactically correct sequences of word types would no longer be syntactically correct. In other words, we have a completely new language. But that is surely a hamfisted way of describing what happened in the board upgrade. There is something that is obviously wrong with the previous two accounts. The crucial thing to note is that the pictures that are pointed out are not mere context. They are crucial for the syntax: whether a sequence of three pointings is syntactically correct depends precisely on what parts of speech the pictures represent. Clearly, the thing to do is to either identify word-types with the pictures that are pointed out (more precisely: picture-types, in order to allow for upgrades of the board), or with pointings-at-x, where x ranges over the pictures (or picture-types) on the board.

Hypothesis: What happens with the communication board is also what happens with demonstratives. The thing pointed to is not context: it either is a part of the sentence (much as some folks think that items referred to de re are parts of the proposition) or a pointing at (de re) it is a part of the sentence. And something like this happens with all indexicals. At this point I am offering no argument, except the suggestive analogy of the communication board language.

Apparent disanalogy: In the communication board language, it is not the the picture tokens that function as word types (or, equivalently, it is not the pointings-at-picture-tokens), but the types of pictures (or the pointings at a type of picture). But in true demonstratives, there is no similar type/token distinction on the side of the things pointed out. One simply points at Alexander, not at something of the sort of Alexander.

This disanalogy is due to the fact that in a typical communication board, none of the pictures refer to the picture-token there, and that in typical demonstratives, we are trying to refer to particulars. But I submit that these are mere accidents. We could imagine that some of the pictures indicate particulars, like George, Socrates, etc. And there would be nothing absurd about a picture that indicates the picture-token that it is. (Maybe it's a very a beautiful and emotionally significant picture, so it's worth talking about as an individual. When a board is upgraded, it gets scraped off the old board and pasted on the new one.)

Moreover, we do in fact have cases where demonstratives point to a type, it's just that we don't use them quite as much as ones where we point to particulars. We've learned this from Kripke. Point to water and say: "We will call this 'water'." The "this" refers to the natural kind, not the particular bunch of water.