Chapter 2

LANGUAGE AND CATEGORIES

Introduction

In this chapter, I want to back up a little and talk about the relationship between language and thought and critical thinking and the study of language — which a study of grammar ideally should be. Language is, of course, a very complex phenomenon, but I would like to generally talk about the way that language shapes our own perception of reality. And I am not simply referring to vocabulary items.

As a way of beginning, let me say that I believe that language determines our understanding of the world, most often without our being consciously aware of its power. Hence, understanding the nature and structure of language is really a form of self-knowledge. For it stands to reason that if you are not aware of how and why you use language in the way you do, you are not fully aware of who you are as a person. On the other hand, a study of the relationship between language and thought is also a study of our limitations. For the limits of our ability to use and understand language are really the limits of our world.

I don't expect the study of traditional grammar to take us too far into what are essentially linguistic and philosophical issues. But I would like to present this discussion to you here so that you can get a sense of why I think any formal study of language is important, particularly as we enter the age of sophisticated interactive telecommunications.

Whorf/Sapir Hypothesis

Let’s turn to the idea that language is the essential shaper of our understanding of reality. This notion is not unique to Western culture, but it was first given its most persuasive linguistic expression in the writings of Benjamin Lee Whorf. It has a long history beginning with Plato, but it is often referred to as the Whorf/Sapir hypothesis since Whorf is seen as the scholar who first gave substance to the seminal ideas of linguist, Edward Sapir. I want to discuss this theory with you not necessarily because I believe it is true, but because it will help me to model certain assumptions upon which our approach to the study of traditional grammar in 110J is based.

First, we must recognize how powerful a force language is. We constantly hear talk referring to language as the tool of human societies; it is supposed to be the primary instrument we use to communicate one to another. In this view, we are ‘in control of’ language in the same way that we might be in control of a soup spoon or a computer. But suppose we say instead that language is the tool of Nature; it is Nature’s way of shaping and controlling human cultures. For however much we can point to the way in which languages have evolved, I do not think we can show that this evolution is the result of human design and power. It is something that seems inherent in the nature of language itself.

Basically, Whorf claims that our ‘picture’ of reality is shaped by the grammatical categories of the language we use. His analysis is primarily concerned with vocabulary items, with words, but with the power of a broad range of underlying concepts and categories that we are usually not even aware of. So what I’d like to do in the time remaining is give you an idea as to what is meant by category and suggest some of the ways that can examine language on a categorical level.

Concepts and Categories

We need to talk about concepts for it is at the conceptual level that Whorf's theories are relevant. Concepts are theoretical constructs we use when trying to model the relationship between thoughts (brain states) and actions (behaviors). (For a fuller discussion of models and concepts, see Appendix 1 and 2.) Concepts are what allow us to interpret and make sense of the world we live in. Hence, concepts literally create reality for us.

Researchers believe that some of our most basic concepts are genetically derived. They are part of the codes which define what we are, biologically, as human. For example, the fact that all natural languages have essentially the same structure is explained in terms of the codes each of us is born with. On the other hand, other concepts and their interrelationships are learned. They are not natural, through recent research is pointing to the possibility of secondary or cultural coding. On the other hand, at a purely behavioral level, it seems that the origin of our concepts, i.e. whether they are learned or inherited, does not determine the degree of power they have over our thoughts and actions.

Now we want to say that, in our model, concepts always come in sets or systems which are defined by a common set of functions or features. Let’s say that these groups can be referred to as categories. For example, red is a concept but it has meaning only when it is related to the category of con

concepts we call color. Let's also say it is categories and sets of categories as opposed to single, isolated concepts that are most important to understanding the underlying structure of language. In addition, these categories can be analyzed to form different levels of generality or hierarchies. Modern science is based upon the assumption that it is possible to model natural phenomena in terms of categorical hierarchies. (See Appendix 3 for a further discussion of scientific theories.)

Linguists assume there are various categories in language at all levels of analysis. Generally speaking, linguistics as a discipline is roughly divided into a study of phonetics (sounds), syntax (relationships) and semantics (meanings). And though they are interrelated, we usually maintain a distinction between the sets. Hence, we say that there is a categorical distinction between a syntactic problem and a phonetic one. Grammar has traditionally been restricted to a study of the relationship between words in sentences as opposed to sounds and meanings. For example, even in a traditional curriculum, when we offer a traditional grammar class we don't work on pronunciation and elocution (there are speech classes for that) nor do we focus on the meanings of individual words (our department used to have a class exclusively devoted to words and their meanings). These divisions are based on how we categorize language studies.

Another way to think about categories is to think about names. We can have specific names for individuals and then we have names for abstract categories. The word woman does not name or refer to a specific thing or object. It is the name of a subcategory of persons; man is the name of the other subcategory. Human being is the name of a larger category which includes men and women. Note that there is a sense in which categories only exist in and through language. The Honda Civic sitting outside my window is an object with four wheels, black paint, a V-Tech engine and so on. It is an object that can be described in minute physical detail. However, it belongs to the general category of manufactured products, motorized vehicle (cars, trucks, motorcycles). It also belongs to the subcategory, passenger vehicle (car). However, I cannot point to a specific item in Civic that these category names refer to though each has a set of conditions that has to be met. Still, the meanings of these terms lies in the way we use language.

Categorical Distinctions

I have said that categories, like concepts, come in clusters or sets. Although a Thunderbird, a Harley-Davidson motorcycle, a Oldsmobile Pro Stock dragster, an M-60 tank and a Mack truck don’t have much in common from a purely physical point of view, they all belong to the category of motorized vehicles. Also, category descriptions are relative in that they respond to specific human needs. Which means that sometimes we want to say that two elements belong to the same category and sometimes we want to say that they don’t. For example, in elementary school, students are often taught how to categorize by doing ‘same and different’ exercises. In the text I am looking at now, there is a picture of a bird, a cow and a house. The student is asked to cross out the picture which is ‘different’ from the others, which, in this case, would be the house. The next problem includes a picture of a bird, a cow and a horse. Here the student is expected to cross out the picture of the bird.

These problems may make category distinctions seem like simple matters of convention and usage. This is sometimes the case and sometimes not. It all depends on a lot of different variables. As we pointed out, some categories we recognize are always ‘there’ because they are part of the genetic codes which structure human languages, e.g., the categories of time an d space or causality. Others may be based on cultural conventions and, therefore, subject to human revision, e.g. the categories of woman and man seem to be undergoing significant changes in our culture.

races based on color is a deep distinction. However, we don't seem to have a problem with Harry-Sally marriages, i.e. marriages between men named ‘Harry’ and women named ‘Sally.’ This would not be a deep distinction, in our sense of the term.

We have category rules in grammar and so we can have deep and ‘shallow’ distinctions in grammatical analysis as well. Let's look at the following two examples:

1) The duck wanted to vacation in Florida but didn't.

2) Home was wented by the duck.

In (1), we can say that the predicate ‘wanted to vacation in Florida’ has to have a human subject. Since ducks are in the category of creatures that can't have long-range wants such as vacationing in Florida, then a duck can't be the subject of a sentence unless the duck were given human capabilities, e.g. in a Disney cartoon. In (2), the sentence is supposed to be a passive of

3) The duck went home.

But ‘to go’ is not in the category of transitive verbs, so it cannot be passivized. In both cases, we would say the sentences are interesting because they point to fairly deep categorical distinctions.

Now what happens when grammatical rules such as these are continually being broken? Well, generally speaking, we usually feel that the speaker has either not mastered the language or is otherwise defective in a major way. In the case of children, we usually make allowances by saying that they are in the process of learning the categories they will hopefully master as adults. So when they make a category mistake as in the case of personifying inanimate agents, we say that they really do believe that inanimate objects can ‘act’ as human agents but we attribute this belief to an ‘immature’ or undeveloped mind.

If an adult were to use expressions that violated the same categorical distinctions, we would not take them literally but as attempts to be ‘poetic’ or ‘creative.’ On the other hand, if an adult were serious in thinking such sentences to be well-formed, then we would have major reservations regarding that individual’s intelligence, brain development and so on. In other words, we tend to view radical categorical ‘mistakes’ as reflections of extremely grave neurological or developmental problems.

However, we have to distinguish between deep categorical infelicities and relatively simple surface phenomena related to dialect differences and the like. Consider the following:

In grammatical analysis, we are interested in language categories that have a major effect on our thoughts and perceptions. We call these deep categories. I have said that the origin of a category is not as important as the degree of difficulty involved in changing it in relationship to the range of its influence. We want to say that deep categories are those that are both fundamentally essential and difficult to revise, modify or change. (I know this explanation sounds pretty lame, but it's the best I can do.)

I often like to think of the way categories operate along the lines suggested by the Bohr model of the atom. We are told to imagine electrons circling a nucleus composed of protons and neutrons. In this model, the energy required to keep an electron circling in its orbit is much less than that required to cause an electron to jump into another orbit. Likewise with linguistic categories. It is possible for an idea to move within its own category rather easily, but it is extremely difficult for it to be recategorized, to enter into new relationships with radically different concepts. For example, in our language community, many people are having a great deal of trouble accepting the idea of same-sex marriages because they blue the categorical distinctions defining clear roles for women and men whereas different-color marriages may cause problems because they confront a set of categorical distinctions based on race and color.

In both cases, we want to say that, in our culture, the distinctions between women and men is a deep distinction just as the distinction between

5. He be coming tomorrow.

6. I gave him book on table.

7. Ivan helped all the mans.

These are illustrations of what we might call ‘surface’ rules rather than the deeper conceptual ones. From the logical perspective we are taking here, they mean little. Of course, it is not always such an easy task for ordinary people to distinguish between surface and category differences. For example, someone might argue that (2) is the same kind of example in that the speaker is showing that he or she knows the passive but is just confused about ‘to go.’ I would not argue too hard against this position, though I could make a better case for:

8. The duck goed home. (go + past)

The point is that we all know of people who are regarded as ‘stupid’ or ‘mentally retarded’ due to very simple differences in regional or nonstandard ethnic dialects such as these. So we need to be very careful in the way we analyze grammatical features.

Culture, Language and Category Differences

The point I've just made is particularly apt when we turn to cultures other than our own, since we cannot always assume that categorical differences are ‘mistakes’ in logic or understanding. I once had a student from Uganda who told me that his grandfather had no need to count beyond three. Instead, he gave things names. For example, he named all of his cows. So when he traded them, he would not simply say that he had traded four cows. He would describe the transaction something like: ‘Today I traded Bob, Jim, Tony and Mary for many sacks of grain.’

It would be easy to label this man and his culture as ‘primitive,’ implying that they have yet to develop a number system such as the one we use which would allow them to define the world ‘as it is.’ But if we think about what the category of number does in English, then I believe we cannot be so quick to judge. For this man and everyone else in his culture, cows do not ‘exist’ in an abstract, impersonal universe. When we say there are nine cows in the pasture, we are looking at the cows in very general categorical terms as objects which can be counted. In the old man’s culture, this point of view would not be possible. Instead, we might say that he and his people experience a relationship to their environment where cows are like persons and cannot be categorized as abstract entities.

So here we see how a very simple matter like the ability to count objects according to an abstract system points to deep categorical features

of a culture. Even a ‘technologically advanced’ society like the Japanese has a counting system very different from that of the West. For example, to say, ‘three people,’ one would use the expression sannin or mitsu. For ‘three birds,’ one would say sambikki. For ‘two cups of water,’ nippon. I think this reflects the fact that the Japanese mind and language do not clearly distinguish between the number of certain types of objects and their category features. In both the case of the Ugandans and the Japanese, we can see how categorical differences in language point to differences in interpreting and experiencing the world.

Back to Whorf

These are the sorts of observations Benjamin Whorf made when he compared the Hopi language to those of the West. Even though there are categorical differences between, say, English, French and German, Whorf felt they were relatively insignificant when these languages were compared to Hopi. Hence, he grouped them together and gave them the generic name, Standard Average European, SAE. He argued that SAE and Hopi had entirely different grammatical categories and hence they interpreted the world in entirely different ways. For example, he compared the SAE concept of causality, agents making things/events occur, with the Hopi concept of preparation, the invitation or preparing of an event’s occurrence. For the Hopi, events happen because they want to happen.

According to Whorf, the Hopi concept of human agency allows human beings to act in such a way that they might increase the desire of events to occur. It is obvious that this view is rather incompatible with the SAE understanding of causality. If we’d like it to rain, we make it rain by seeding clouds. Or, if we pray, we pray to God (a superhuman agent) to make it rain on our behalf. The Hopi seem to believe that the rain has a will of its own. Hence, one has to invite the rain to fall. The power of the rain dance is not a matter of exerting casual pressure on the rain but in demonstrating the sincerity or good faith of those who need rain.

It is interesting to note a study done in Switzerland comparing American Indian children with a group of children from France in respect to the habit of giving human characteristics to inanimate phenomena and objects. The study showed that both the French and the Native American children began with the same personifying tendencies, but as the French children grew older, they moved towards a more conventional SAE interpretation. The Native American children, on the other hand, became even more ‘extreme’ as they became adults.

A Model for Conceptual Analysis

Of course, we don’t often consider categories like space, time, causality and agency as we go about our daily affairs. We take them for granted, for they are the conceptual ground upon which we stand. And it is for this reason that they are difficult to analyze or discuss even when we want to. Just how do we talk about causality or personal identity in ordinary language? How do we see the eyes through which we see?

The approach that I feel is most useful for this course is usually called conceptual analysis. (Here we don't care about the distinction between categories and concepts.) It involves the use of language to look at a behavior, analyze it, generalize it into abstract terms and then describe it in ordinary language. The focus of the method is dialogue. The investigator must ask real questions and get real answers, since the purpose of conceptual analysis is to break a concept down into its component parts to better understand how it works. If the questions and answers aren't honest, then the analysis fails. (This method is explained in more detail in Appendix 2.)

So let me conclude this chapter by giving you two examples of conceptual analysis in traditional grammar. Both take the form of dialogues and both are aimed at helping others to see that we all ‘know’ a lot more about the grammar of English than any set of rules we might find in a book.

The linguist, D. Terence Langendoen, once used the simple English tag-question to illustrate through conceptual analysis that traditional teaching grammars make poor usage guides because they: a) cover points that are so ‘obvious’ that few have problems with them, b) are usually dependent on a single, fairly narrow dialect as the standard for correctness, c) seldom account for the fact that all languages and their dialects change and evolve over time, so what may be correct for at one time may not be correct at another, even for the same dialect, d) usually end up making language instruction appear much simpler than we actually know it to be.

Langendoen invited his students to come up with a clear and consistent usage rule that would tell, say, a non-native speaker of English how to form correct English tag-questions. I shall summarize his main points with the following examples of my own. First, we start out with the easy cases:

1) They can go, can't they? (Tag = can't they?)

2) They shoot horses, don't they? (Tag = don't they?)

3) *They can go, can't he?

4) *They shoot horses, can't they?

These are illustrations of correct and incorrect tag questions. The following are not tag questions at all.

5) They shoot horses, don't you think?

6) So he likes beer, hunh?

7) They can go, I guess.

Langendoen claimed that (1) and (2) might be easy to explain with a single rule. But then he challenged his students to use the same rule to cover the following examples:

8) He has to go, doesn't he?

9) He has a coat, hasn't he?

10) He has a coat, doesn't he?

11) She has a cold, doesn't she?

12) She has eaten, hasn't she?

13) She has to eat, doesn't she?

14) She must stop eating, shouldn't she?

15) Everyone has to eat, don't they?

16) He ought to stop smoking, shouldn't he?

One of the points Langendoen was trying to make is that after we get past a few clear cases, it becomes increasingly more difficult to cover all of the variations of the English tag-question using one simple usage rule. Yet, this simplicity is precisely what many students seem to demand in a grammar book; consequently, they become increasingly angry and frustrated when the rules they are given either don't fit or don't help them with difficult constructions.

Again, this frustration is based upon a failure to understand that natural languages are not like computer programs or video games where someone makes up the rules that someone else can memorize. Natural languages are extremely complex systems of communication, and the fact of the matter is that no linguist has yet developed a convincing scientific theory to fully account for even basic phenomena like predication or negation.

Another important lesson that comes out in this exercise is that we really know a lot more about the grammar of English than we think we do. Perhaps we are not able to spell out the rules explicitly and know all the jargon but we are able to deal with very complex sentence structures. Our second example of conceptual analysis will bear hopefully this out.

Q: OK, give me some examples of the passive.

A: ‘The apple was eaten by John,’ ‘The car was stolen by the boy.’

Q: Then what you mean by the active would be sentences like ‘John ate the apple’ and ‘The boy stole the car?’

A: Yes, they’re the active forms.

Q: What about ‘The car was stolen’ and ‘The apple was eaten?’ Are they

A: ‘Derived from?’

Q: Oh, sorry. Just tell me if you think it’s passive.

A: No, I don’t think so. Maybe it’s ambiguous.

Q: What about ‘He remembered that he came to dinner six times’? Can that be passivized?

A: ‘Passivized?’

Q: Sorry again. Does that sentence have a passive?

A: ‘That he came to dinner six times was remembered by him.’ Yes, that sounds really awkward but I think that would be the passive.

Q: What about ‘He is ten feet tall?’

A: No, that doesn’t have a passive form.

Q: What about ‘He ran to town?’

A: No, you can’t say, ‘Town was run to by him.’ You need to have a noun following the verb.

Q: Oh, like ‘He ran home?’

A: ‘Home was run by him.’ No, that won’t work.

Q: Why?

A: I don’t know. It’s just funny, unless you say, ‘The home was run by him.’ But then, ‘the home’ means something different does ‘run’ — you know, like ‘operate’ or ’manage.’

Q: Then what about ‘He resembled his brother?’

A: ‘He was resembled by his brother.’ No, that doesn’t sound right either.

Q: So can we generalize and say that only action verbs as opposed intransitives and linking verbs have passive forms?

A: Yes, I guess that sounds pretty much right. Action verbs that take direct objects. Someone or something does something to someone or something else.

Q: But what about ‘The cake was eaten?’ How is that passive again? We don’t have an agent or doer. We don’t know who performed the action.

A: Oh, that’s just understood.

Q: Understood?

A: Yes, you just assume that someone ate the apple. Maybe the context will tell you so you don’t have to repeat it. But someone had to do the eating.

Q: What about ‘He was angry?’ Is that passive?

A: No, because no one did anything.

Q: What about ‘He was angry at me.’

A: No.

Q: OK, try ‘I angered him.’

A: Yes, that’s all right. ‘He was angered by me.’

Q: Are you sure now?

A: Sure. ‘He was angered by Sue’s going out with Tom.’ ‘He was pleased by Sue’s going.’ ‘He was upset by Sue.’ We say things like that all the time.

Q: Tell me then. What is the difference in meaning between ‘He was upset

at Sue’ and ‘He was upset by Sue?’

A: Well, um ... in the first case, Sue isn’t trying to upset him. In the second case, she is.

Q: So ‘Sue upset him’ always means that Sue tried to upset him?

A: Oh, I don’t know. Maybe.

Q: OK. What would you do with ‘Sue gave her cat to John?’

A: Well, I could give you several passives: ‘Her cat was given to John by Sue’ or ‘John was given her cat by Sue.’

Q: Are you sure those are all right?

A: Yeah, sure. They sound OK to me.

I think this example should give you an idea as to how the method works. When we turn it around, we can also use the method to understand why a person isn't learning a new concept. At any rate, we are now ready to start to do some comparisons of how professional grammarians approach English grammar and how their analyses differ based upon the conceptual assumptions they have.