Language and information

Lecture 2. Sublanguages

2.2. Subject-matter sublanguages

Audio recording—Opens a separate, blank window. Return to this window to read the transcript.]

Now, the base sublanguage, and the reduced-sentence sublanguage, and the metalanguage (also a sublanguage) have all been defined in respect to language structure. However, the metalanguage also has another property, namely, it has a subject-matter. It has a subject-matter which is specific to it, in contrast to the very general area of natural language, which has no limitations that we know of, virtually, in subject-matter. It is also possible, it turns out, to define sublanguages based on particular subject-matters without going to a particular definition in respect to language structure, but purely on the merits of the subject-matter itself. The reason is that if one talks about a reasonably structured part of the universe, of the world, or what is perceived that way, and speaks in such a way that only relevant things are said, or what seems to be relevant, only, is said, we have, as a result of this, strong limitations upon the use of words, because some words are not used because they are not part of what is perceived as what one is talking about, other words are not used because they are not relevant to the subject, and so forth. There are strong limitations on the use of words. These limitations are strong enough to constitute constraints on word occurrence.

Now, since the theory as presented yesterday did not deal with fixed, stated classes of words, like noun and verb, but with classes that were determined by their occurrence relative to each other—and this is how we found this dependence, and dependence on dependence, and so forth—if the same method is applied in a body of texts—meaning, let's say, anything in a closed subject area, or in a closed science—by the same method one gets different word-classes, different sentence types, and indeed a different grammar.

I want to say some things about this difference, and then to see what happens when we consider this situation in respect to sciences.

First of all, we find out that in general we have not dependence on dependence in the way that was described yesterday, but we find specific sets of arguments which occur only under particular sets of operators. To take a simple example, in the whole of English the sentence polypeptides were washed in hydrochloric acid and hydrochloric acid was washed in polypeptides. Both seem to be grammatical sentences. One may or may not know that the second has a dubious range of possible activities, but for all intents and purposes those are sentences of English. In a laboratory report, the second sentence simply would not occur. It's not that it is wrong, but it is not said, that's all. One may find a sentence, as I found, let's say, the polypeptides were washed in hydrochloric acid, one will not find the other sentence. So if one considers sentences which have been limited in this way, that they actually occur in this material, you see that certain things will not occur, you see that what you get is not nouns and verbs, what you get is is washed in and a number of other 'verbs', let's say, of this type, and words for cells, or molecules, or tissue, and words for acids, for liquids—water—and you'll find that the first set, cells or tissue is washed in the second set, the liquids or acids, and so forth, but not the other way around.

Now, there are a number of other differences, one or two of which are very important as soon as we deal with a sublanguage at all, not necessarily a science sublanguage. The metalanguage of a sublanguage is not in the sublanguage itself. If we took even, let's say heraldry—it's a reasonably closed subject—if one took sentences about heraldry, and then wanted to describe those sentences, one would describe them in English, in a number of sentences. Those would not be sentences of heraldry, they would be sentences, actually, of grammar, particular sentences of grammar: they would say which words occurred as verbs on which nouns, and so on. So the metalanguage of a sublanguage is, in general, not a subset of it. It is another subset of English. This is very important, because it frees the structure of the sublanguage from having to satisfy the limitations required for language. In language as a whole we can only figure out—in principle, at least, or procedurally—we can only figure out the grammar of the language from seeing which combinations occur and which combinations don't occur, and only on that basis. We cannot by fiat say that this and this is a noun, or something like that, because in what language did we say that sentence, and how did we arrive at that language?

But here, we can decide, for example, that phrases, whole phrases, are members of a word-class. In English, if a person says antibodies are found in a certain cell, this has to be taken as the passive of some other sentence, because that's the grammar, and you cannot decide when to stop a grammatical analysis. In immunology, in immunology papers, which I will present in a moment, one can say antibodies appear in a cell and you can say antibodies are found in a cell. These two 'words' occur completely one in the place of the other, there are no differences in their neighborhoods, and we would want them to be in the same class. We cannot do that in English. But in a sublanguage, we can say—we give a list of what is the vocabulary of the sublanguage, we say the verbs that occur between antibody and cell include appears in, is present in, was found in, and so forth, is contained in; a whole number of things.

Now, though the sentences of a sublanguage are a subset of the sentences of the language, the grammar of a sublanguage is not a subgrammar of the grammar of English. There are reasons for this. It's not just part of the grammar. First of all, there are certain constraints, obviously, in the sublanguage which don't to English sentences if they're not satisfied by English sentences. Obviously, the constraint about molecules or cells being washed in liquids is not satisfied by English sentences because you can say it the other way around. If in sublanguages there are special word—noun-classes, let us say, which occur only with special operator classes, then this is a constraint which is not satisfied, this is in the grammar of the sublanguage, not in the grammar of English.

On the other hand, there are constraints in the grammar of the whole language which are not in the grammar of the sublanguage. Now, since the sentences of the sublanguage are also sentences of the language, they cannot violate the grammar of the whole language, but they can avoid parts of the grammar. Some things in the grammar simply may not be exercised in them, in the way that does not make them ungrammatical. An example is a very major difference between sublanguage and language, namely, that sublanguages—in general, naturally, I have not scouted all sublanguages in the world—sublanguages to a large extent apparently do not have likelihood differences. The likelihood differences, which are graded more or less, and which give you the meaning-differences between words—an extremely important thing in language—is in abeyance in sublanguages. They either have less of it, or they have none of it. They have something instead: they have word subsets, but they do not have a gradation of one antigen being more likely to be injected than another antigen, or something like this.

Language, I also want to say, is not really an envelope of sublanguages, as one might think, one might phrase that question. It's not really an envelope of sublanguages, and its grammar, the grammar of language is not really a washing-out of the differences between various sublanguage grammars, because there are constructions and properties—for example, the lack of an external metalanguage—which apply to language and to none of its sublanguages. So these are really not quite identical systems.