We are still a long way off from being able to explain away all the constraints, let alone the fact that different languages have markedly different kinds of constraints. Some of these differences may be due to whether a language is case-marked, as German is (this was explained earlier). Other differences might be due to whether the verb appears early in the sentence, or late. Characterizing these differences, and identifying the commonalities (a prerequisite for ultimately understanding the reasons underlying the various constraints on the different sentence forms) is the remit of linguistics. And even if the goals of psycholinguistics and linguistics differ (see Chapter 1), the two are not easily separated, and nor should they be. None the less, since the early days of contemporary linguistics (as started by Noam Chomsky), there has been a tension between the two. Partly it has to do with the distinction between what is grammatical and what is processable. A sentence like `What did you throw the tray that you brought the book that I didn't want to be read to out of up on away for?' is grammatical, but pretty hard to process. Similarly a sentence like `The wedding ring the woman the man the cousin loved loved loved tarnished' is grammatical, but is also, to all intents and purposes, pure gibberish. Why it is gibberish is still a matter of speculation. Our ability to keep track of who did what to whom just cannot keep up with what grammatical convention allows. And whereas in these cases, the human limitations do not appear to have constrained the evolution of those grammatical conventions, only future research will determine the extent to which such limitations may nonetheless have influenced aspects of the grammars that we each use every day.
Neither this chapter nor the previous one have really addressed the most fundamental issue of all-what does a sentence mean? How is this meaning integrated into the meaning of the surrounding sentences? And what, exactly, is meaning? What these last two chapters have done is move us some little way closer towards understanding how a meaning is arrived at. And there is no magic behind this, only the seemingly mundane task of identifying who or what is being referred to, and what roles they should have. Along the way, we have encountered some enduring puzzles, such as, on the one hand, the surprising degree of ambiguity that languages tolerate, and on the other, the surprising constraints on what kinds of ambiguity are not tolerated (these were the him/himself facts). But none of this tells us what meaning really is. In principle, we could write a computer program that could do much of what has been described here, and such programs do exist. But would such a computer really understand? In the next chapter, we take a closer look at what meaning might be.
On the meaning of meaning
Within the Babel metaphor, the tower's summit touches the very essence of language-the ability to convey meaning. Language itself is merely a tool, a tool of the trade in meaning. But like any trade, there is a difference between the trade itself-how it should be conducted, the conventions used, and so on-and the thing that is traded. Knowing something about how we convey meaning is quite far-removed from knowing what this thing is that we are conveying. Getting to the summit is quite different from being able to look around once we have got there.
The distinction between the meaning conveyed and the conventions used to convey it is perhaps most famously captured in Lewis Carroll's poem Jabberwocky.
'Twas brillig, and the slithy toves
Did gyre and gimble in the wabe:
All mimsy were the borogoves,
And the mome raths outgrabe.
Anyone reading this verse would have an idea of what it ought to mean if only they knew the meaning of all the words. It is as if the grammatical conventions have been kept, and the individual meanings thrown out. Fortunately, a translation was provided, through the looking-glass, by Humpty Dumpty, who knew a thing or two about meaning:
"When I use a word", Humpty Dumpty said, in a rather scornful tone, "it means just what I choose it to mean-neither more nor less."
He then went on to discuss the temper that some words have, and the extra wages he would pay them if he made them mean a great deal. Sadly, on our side of the looking-glass, words do not quite work in this way. On the other hand, Humpty Dumpty's assertion, about his choosing what a word should mean, is not as absurd as Lewis Carroll might have thought-we do use words to mean just what we choose them to mean. The challenge is to understand why everyone appears to have made the same choices.
This challenge is all the more intriguing given that speakers can invent new meanings for words, or use words in new ways, and still be understood. An example from British English makes use of the fact that the `anorak', a kind of coat, is the much favoured garment of the `nerd' and the train-spotter. A novel use of the word is slowly entering some circles: ,to anorak' as in `You're not still anoraking away at your computer, are you?'. Many readers will recognize the intended meaning, even if they have never heard `anorak' used in this way before.
So what exactly is meaning, and what does it take to understand an utterance? What do we need other than a few dictionary definitions and a knowledge of the language's grammatical conventions? After all, many a child has suffered too many years at school learning another language in exactly this way, using a grammar book and a dictionary-hardly an enjoyable way to learn a language, and, by all accounts, hardly a successful way either. But the trouble with the dictionary approach to meaning is that we do need more; definitions are not enough. If each dictionary definition were expressed simply in terms of other words in the same dictionary, it would be impossible to know what any single word meant. The word `speech', for instance, is defined in the OED (second edition), using the term `word', whilst `word' is defined using the word `speech'. It is because of this inherent circularity that a dictionary is not a repository of meaning. Those English children would not get very far if they were given a dictionary written entirely in French, or German, or whatever other language they were learning. And never mind about learning a second language, what about learning the first? Children could not possibly learn the meanings of new words in terms of words they had already learned, because in the early stages of word learning, they quite simply would not know enough words to put together the definition of a new one. And yet from the ages of 18 months to around six years, they none the less manage to learn upwards of nine new words each day. That is a lot of words. According to the OED, which also defines `speech' as the natural exercise of the vocal organs', it is also a lot of exercise!
Trying to understand the nature of meaning is complicated by the fact that, just like the tower at Babel, there are different levels of meaning, each supporting, and supported by, the other levels. For instance, the gist of a conversation is a kind of meaning that differs considerably from the meaning of an individual utterance that may have been a part of that conversation. In turn, the meaning of that utterance will be quite different from the meaning of any one of its individual words. And what about the meaning conveyed by the inflections -ing, -ed, -s, -able, re-, un-, and so on? Aspects of the meaning of an utterance may not even be conveyed using words at all; the same words uttered with a different intonation can signify quite different things (see Chapter 7).
Evidently, meaning is a complex concept-intonation, words, utterances, and whole conversations all convey meaning, but different kinds of meaning. Meaning can even be conveyed without any sound at all; gestures, such as a shake of the head, or a shrug, can augment an utterance, or even replace it. Meaning can even be conveyed without words or gestures-by a well-placed silence, for instance. And to complicate matters further, the intended meaning of an utterance might not be its actual meaning; it could be the exact opposite. Someone eating an ice cream with particular enthusiasm might respond to `You like that, do you?' with `No, I can't stand it, but I'm on a strict diet'. Irony and sarcasm are not the only times that the literal and intended meanings may differ: `Could you pass the relish?', asked during a meal, is rarely responded to with a simple `Yes'.
There is no one thing that we can call `meaning'. Instead, there are many different kinds. I
f ever there was a whole that was greater than the sum of its parts, meaning is it. But although there is no one thing called `meaning', the different kinds can be divided into two broad varieties: the meanings of individual words and the meanings of combinations of words.
On the meaning of words
At its simplest, the meaning of a word is the knowledge that one has of the situations or contexts in which it would be appropriate to use that word. But this is a far cry from practical reality, which is that meaning is what happens when some particular neural circuitry somewhere in the brain is stimulated into activation. That is hardly a helpful description, but it does have its uses. Imagine that we could picture the activation of the individual neurons in someone's brain, and that by chance, we happened to picture the neurons whose pattern of activity corresponded to the meaning of, for instance, the word `linguist'. How would the brain itself know that the activation of these cells corresponded to `linguist' and not to `fish', or to the sensation of an itch on the arm? It sounds like an impossible question to answer. But there is an answer. And the answer requires us to imagine that we can also picture the activation of the neurons in someone else's brain whose activation is also equated with the meaning of `linguist'. The chances are that these neurons will not be exactly the same as the ones in the first person's brain. They might have slightly different connections with other neighbouring neurons, they might be bigger or smaller, there might be more of them, they might be in a slightly different part of the brain, and so on. But if the two sets of neurons (one in each brain) are so different, how come their activation can nonetheless be equated with the same thing-the meaning of `linguist'?
The answer to this last question is simple: it does not matter that the neurons differ from person to person, so long as they would become activated in just the same situations. All that really matters is that one pattern of neural activity arises whenever the word `linguist' is heard, and another pattern arises whenever `fish' is heard. So long as the two sets of neurons only become active on those respective occasions, the brain will distinguish between the two words on the basis of the distinct patterns of activity they evoke. And if the brain responds differently, the brain's owner (for want of a better description) will also respond differently.
But something is missing. It is all very well to say that the brain can distinguish between the words `linguist' and `fish', but is that enough? Anyone could distinguish between two unknown words without knowing what they meant, or between two mystery objects without understanding what they were. Once again, what is understanding? What gives it its essence?
There has been a long (and to some people, odd) tradition of scientists and philosophers who believe that the essence of meaning, and indeed the essence of `thought', does not reside in the physical stuff of the brain at all. For instance, in the seventeenth century, Rene Descartes (of `I think, therefore I am' fame) believed that the mind is quite separate from the brain, with some kind of spiritual link between the two. Even in the twentieth century, various philosophers have believed that there is something magical in the brain, that we can neither see nor measure, but which makes the brain more than just a machine. But as we move into the twenty-first century, let us assume that there is no magic, and that meaning and understanding are somehow related to whatever is going on in the brain. Let us also assume, as we did at the start of this section, that the meaning of a word is really just the knowledge-which in turn is just the accumulated experience-of the situations in which it would be appropriate to produce that word. The next question is: what is experience to a bunch of nerve cells?
Experience as the essence of meaning
There is considerable evidence to support the idea that the pattern of activity across the neurons associated, for instance, with the concept of fish gradually changes each time a fish is encountered, becoming each time more specific to just those contexts in which fish are encountered. This gradual change has much in common with the way Pavlov's dogs learned the relationship between the bell and their dinner. The association between the two progressively strengthened with each pairing of the bell and dinner-in effect, the neural activity associated with the bell progressively changed to reflect the context in which the bell occurred (i.e. dinner). In the fish case, the association between `fish' and the contexts in which fish can occur would also strengthen, with the neural activity associated with `fish' gradually changing to reflect these contexts (we return later to why these changes come about).
So patterns of neural activity change as a result of the accumulated experience of things, including both their physical characteristics and the contexts in which they have occurred. In effect, these patterns are a reflection of those experiences. And naturally, objects that are similar to one another will tend to occur in similar contexts, to behave in similar ways, and to lead to similar experiences. Consequently, the neural activity associated with them will also be similar. Although the pattern of activity that is evoked in response to a shark will be different from the pattern evoked in response to a goldfish, components of those patterns will be similar, because the two fish share certain attributes. It is this similarity in their patterns of neural activity that corresponds to the concept of fish.
Of course, the neural activity that is evoked when we see a fish may not be the same as the activity that is evoked when we think about a fish, or when we hear the word `fish', or when we eat a fish. In all likelihood, these patterns do differ. But because they may each occur in similar contexts (we might hear the word `fish' when we see one, or eat one), there will be elements of the corresponding patterns which are similar. Whether we call one pattern a concept, and another pattern a meaning is unimportant. What is important is that the meaning of `fish' is simply a pattern that has become associated with the contexts appropriate to the things that `fish' refers to-and it is this that is the essence of its meaning.
But what about all those words that do not refer to objects? In fact, they could all work on pretty much the same principle-the concept associated with something is the accumulated experience of that something, whether it is an object, an event (e.g. running, flying), a property (e.g. yellow, fast, high), or whatever. And in each case, the meaning of the word is simply a pattern of neural activity that reflects that accumulated experience.
Intermission: if neural activity is just complicated chemistry, could we not fill a test-tube with meaning?
The last few pages have been based mainly on conjecture. We cannot observe the patterns of neural activity that we have been talking about, or watch those neural structures change as they accumulate experience. But much work in neurobiology has directly observed how neural systems react and change as a function of experience. Admittedly, much of this work has been carried out, for both ethical and practical reasons, on slugs and the like. But if slugs can do it, then presumably we can too-only better.
One feature of this approach which may make some people uncomfortable is that meaning has been reduced, in effect, to just some very complicated chemistry.
No-we could not fill a test-tube with meaning. But imagine that a computer could do something that was exactly equivalent to what we suppose to be happening in the brain. Would such a computer, through its own (perhaps limited) experience, accumulate meaning? Would it understand? Would it, by virtue of any such understanding, have ethical and moral rights? Would we dare switch it off?
As we delve deeper into the brain's workings, such questions necessarily come closer to the surface of our understanding. Fortunately, questions like these, and their answers, are beyond the remit of our current ascent of Babel. But that will not make the questions go away, or the answers any easier to accept, whatever those answers happen to be.
Beyond the meaning of individual words
The meaning of a multi-word sentence is a very different beast from the meaning of an individual word-whether in terms of neural activity or in the terms that we are probably more used to. So it is as well to take a step back from the brain
, and to just stick for the moment with our everyday vocabulary.
To understand a sentence generally requires knowledge not just of the meanings of the words making up that sentence, but also, and amongst other things, knowledge of the world within which specific things, and events, occur. For instance, imagine a non-fluent speaker of English having trouble with something they were reading and asking `What does the man she saw last night mean?' It is unclear how one might best go about answering this question. But now imagine a slightly different version of this same question, whispered in the context of two people gossiping: `Whom did she mean by the man she saw last night ?' In this case, it is extremely clear what would need to be known in order to answer the question.' _link_ Here, the sequence `the man she saw last night' is used to refer to someone quite specific who inhabits the world that the speaker and hearer, or writer and reader, share. To know who was being referred to requires knowledge about this world; who was in it, what they were doing, and who they were doing it with (not to mention how often and for how long).
The meanings of individual words are the building blocks which allow the hearer or reader to (re)construct some aspect of the speaker or writer's world. But what is it that is constructed? In what form can one person's world exist in another person's mind? Indeed, in another person's brain?
The Ascent of Babel: An Exploration of Language, Mind, and Understanding Page 15