<<

Chapter 17 Culture and Language

Each language sets certain limits to the spirit of those who speak it; it assumes a certain direction and, by doing so, excludes many others. — Wilhelm von Humboldt (1767–1835)

Learning a language represents training in the delusions of that language. — Frank Herbert, Whipping Star

There are about 6000 languages in the world, not counting the ones that have existed in the past but are dead now. I've said that before, haven't i? There are 6000 different languages in the world, and one of the ways in which they differ is in how they correlate with reality. Reality is represen- ted in different languages in different ways. But now, think about this: How do we talk about reality? How do we think about reality? Obviously, we can't talk about reality without using lan- guage, some language or other. So the way a given language represents reality obviously affects how we talk about reality — if we're using that particular language.

But let's bear in mind an important theme underlying this semester's work: Language is a social phenomenon. In some sense, a language belongs to the community of people that use it. If the fact that i am speaking English affects the way i talk and think about reality, then presumably it should have roughly the same affect on everybody else who uses English — the entire English- speaking (or at least English-using) community. And, given the differences between English and the Chinese languages, that may be a little bit different from the way a Chinese-speaking commu- nity like yours is accustomed to representing reality. What this means is that the way a given lan- guage represents reality ultimately becomes a reflection of how the community associated with that language represents reality, at least linguistically, and this is very much a matter of culture, from the anthropologist's view.

So language becomes a very important tool for investigating culture; the categories that are im- portant in a given language in some sense represent the categories that are important in the cul- ture of the people who use that language. Exactly what that sense is, how to use the evidence of language in analyzing culture and how reliable that evidence is, is a matter of some controversy and we'll be discussing that controversy to some extent in a little while. But there is no question that in some sense a community's language is a very important source of information about that community's culture. And when i say that, i don't mean merely that one can use language to find out about a community's culture by using language to talk about their culture, asking them ques- tions about their culture and so on, though of course that is, in one sense, using language to find out about their culture. I mean that an analysis of their language can itself lead to information about that community's cultural outlook on reality.

It is perhaps not surprising that such research, such analysis has been pursued most fruitfully in the area of contrastive linguistics. You look at two different languages and you notice the ways in which they are different, and then you consider the possibility that these differences are some- how symptomatic of cultural differences between the communities using these different langua- ges. Ways in which two languages are similar are of lesser interest in this kind of research; we expect languages to be similar to each other to some degree for the same reason we expect human cultures to be similar to each other to some degree: Because they're human, because they are ma-

313 nifestations of a common human nature. What the typical anthropologist, or anthropological lin- guist, or contrastive linguist is interested in is the extent to which human cultures and languages can differ from each other while still being jointly human.

In what ways do languages differ from each other? Well, remember what we said last semester: A language consists of a vocabulary (or lexicon) and a , and languages can differ in both areas. Obviously, all languages differ from each other in terms of vocabulary; that's obvious sim- ply from the fact that in my native language i say ‘left’ and ‘right’, ‘up’ and ‘down’ while you say ‘左’ and ‘右’, ‘上’ and ‘下’. But, at least to some extent, the concepts represented in our respective languages by ‘left’ and ‘左’, ‘right’ and ‘右’ and so on are exactly equivalent.1 What is much more interesting to the contrastive linguist, and therefore to the linguistic anthropologist, is the areas where the vocabularies of two different languages fail to match up. In theory, the most basic, most obvious version of such a difference would be for one language to have words for concepts that the other language doesn't have words for. More commonly, what happens is that one language will have fairly simple vocabulary, involving a relatively small number of morphemes, for certain concepts that another language will have to use much more complicated expressions to represent.

One common area of such diversity is what are known as kinship terms, the words we use to refer to important family relationships. The study of kinship terms is an important aspect of re- search in anthropology and ethnography, in large part because different cultures organize families differently and have different ideas about which relationships are important and why, and anthro- pologists generally agree that the simplest, and one of the most enlightening, ways of investigating a culture's family organization is to analyze the vocabulary of kinship terms that goes with it. In other words, the kinds of relationships that a particular culture calls by different words, and the kinds of relationships that it refers to by the same word, can tell us a lot about how that culture thinks about family relationships.

In (1) i've given a sample of some kinship terms in English and what i know of the equivalent terms in Chinese. Especially with regard to the Chinese, i've restricted myself to expressions made up of no more than 3 morphemes. It is immediately apparent that Chinese has at least 34 distinct terms for relationships that English subsumes under only 7. Now obviously an English- speaker could speak of hanns ‘older brother’ or ‘younger sister’ just as easily as a Mandarin- speaker could speak of hanns ‘哥哥’ or ‘妹妹’. But it seems to me that a Mandarin-speaker is to a certain extent compelled by the structure of Chinese vocabulary to make this distinction on every occasion.2 This is a rather significant matter in my case; my parents have only two child-

1Of course, it could be argued that there are subtle cultural differences between them having to do among other things with the fact that Western European languages like English have, for over 2000 years, been written exclusively from left to right while Chinese has traditionally preferred a right-to-left direction, only fairly recently adopting a left-to- right direction instead; or the use of ‘上’ and ‘下’ to mean ‘before’ and ‘after’ as well as ‘above’ and ‘below’, or the use of ‘left’ and ‘right’ as political labels. But we don't need to go into that. The main point is that, at the most basic defining level of which direction is which, ‘右’ means the same thing to you that ‘right’ means to me and contrariwise.

2This issue came up in discussion with one of my students who was concerned about how to represent in Chinese the precise content of the English sentence ‘He's my brother’, when it is impossible to deduce from context whether the ‘brother’ being referred to is older or younger than the speaker. We discussed this problem for some time, and i subsequently took the question up with several other Taiwanese friends, and there was general agreement that the closest equivalent in Chinese was the word ‘兄弟’, but that although this word is in fact occasionally used for this purpose — in translating from Western languages that do not automatically provide the information that Chinese

314 ren, and i'm the oldest of them. I normally think of my brother merely as ‘my brother’; i don't need to point out on every occasion that he's younger than i am. Whereas if i'm speaking Chi- nese, i have no choice but to refer to him as ‘我弟弟’. And as a Westerner and a native speaker of English, there's part of me that feels, Damn it, it isn't anybody's business that he's younger than i am!

(1) English Chinese brother 哥哥, 弟弟 sister 姊姊, 妹妹 grandfather 祖父, 外祖父 grandmother 祖母, 外祖母 uncle 伯父, 叔父, 姑夫, 舅父, 姨夫 aunt 伯母, 嬸母, 姑母, 舅母, 姨母 cousin 堂哥, 堂弟, 堂姊, 堂妹, 表哥, 表弟, 表姊, 表妹, (姑表哥, 姑表弟, 姑表姊, 姑表妹, 姨表哥, 姨表弟, 姨表姊, 姨表妹)

It is possible to distinguish in English between one's paternal grandparents and one's maternal grandparents, but it's not commonly done; normally, we would only do it in cases where there is a serious danger of being misunderstood. I don't know how routine it is in Chinese to use the clarificatory morpheme 外 in referring to one's grandparents.

It gets worse. Whereas Chinese manages to distinguish between five different kinds of ‘uncles’ and five different kinds of ‘aunts’ by means of labels consisting of only two morphemes apiece, it's impossible to make the same distinctions in English without some very complicated phrasing. All told i have had, counting the ones who have already died, four uncles and six aunts, and in English it is by far the simplest thing to simply refer to each one as simply an ‘uncle’ or an ‘aunt’.3 As for ‘cousins’, yes i can fairly easily distinguish between ‘male cousin’ and ‘female cousin’, and some European languages such as French do so automatically, but beyond that it is almost impossible to convey in English all the distinctions made by the Chinese terms at the end of (1). As it happens, all my cousins happen to fall into the categories referred to in Mandarin as ‘表哥’ and ‘表姊’, because they're all of them the offspring of my mother's various brothers, all of whom are older than she is and therefore their children are all older than i am. But in English it would probably take me about three sentences to explain what in Chinese would take only two mor- phemes.

OK, fine, English and Chinese have different vocabulary for describing family relationships. What of it? Now i don't know a lot about the details of this, but i would venture a guess, based on what i know about Chinese kinship terminology, that traditional Chinese culture regards each of the different relationships named on the right-hand side of (1) as involving distinct responsi-

inherently requires — it is nevertheless unsatisfactory since it carries connotations absent from the English word. This is a very good example of the ways in which choice of language can constrain the ways one both talks and thinks about reality. 3I once came across in an English novel a reference to the concept of ‘uncle-in-law’, meaning a man married to the sister of one's father or mother, as distinct from a blood-related uncle. But i wouldn't make too much of the fact. For one thing, i've only encountered this expression once in my life. For another, there were clear indications in the context that the expression was meant facetiously.

315 bilities. In particular, i note that in many cases Chinese makes a major distinction based on rela- tive age. I would assume, based on what i know about kinship systems in general, that the Chinese language makes these distinctions because in Chinese culture, at least traditionally, there are sig- nificant differences between one's relationship with one's 哥哥 and one's relationship with one's 弟弟, and likewise that one's relationships with blood-uncles and -aunts is different from one's relationships with ‘uncles’ and ‘aunts’ that have merely married into the family. In modern English-speaking culture, such differences are not the norm. A brother or a sister is simply a brother or a sister whether older or younger, one's uncles and aunts are the same whether they are one's blood kin or merely married into the family, and a grandparent is a grandparent whether on one's mother's side or one's father's. This is the way it is in English-speaking society, and it's been that way for so long that we don't have the kind of distinguishing vocabulary that Chinese has. In this respect, the linguistic differences clearly reflect actual differences in culture.

There can be further complications. Consider the expressions in (2). This is a list of seven dif- ferent possible geneological relationships; on the left-hand side i have given their most explicit possible labels in English, and on the right-hand side the most accurate versions i can come up with for these concepts in Chinese, bearing in mind that none of the labels on the left-hand side carries any information about the relative ages of siblings that Chinese requires to be expressed. Now i'm going to tell you that there are cultures, both in and in New Guinea, that have one single label for all seven of these; in those cultures, all the people listed in (2), all the indiv- iduals boxed in the geneological diagram in Fig. 17.1, are regarded as equivalent to one's father.4 Why this should be so is a very complicated matter, and i don't pretend to understand all of it. I'm merely pointing out that in this case, once again, a language's way of referring to relationships is a tip-off to how that language's speakers think of them, how their culture regards them.

(2) father 父 father's brother 伯父, 叔父 father's sister's son 表哥, 表弟 father's mother's sister's son 祖母姊的兒子, 祖母妹的兒子 father's sister's daughter's son 表姊的兒子, 表妹的兒子 father's father's brother's son 父的堂哥, 父的堂弟 father's father's sister's son's son 祖父姊的孫, 祖父妹的孫

♀ ♂ ♂ ♀ ♀

♂ ♂ ♂ ♂ ♀ ♂

♂ EGO/我 ♂ ♀

Fig. 17.1 — What ‘Father’ may Mean to Seminoles or Trobriand Islanders

4The cultures in question are the Seminole Indians of Oklahoma and Florida and the inhabitants of the Trobriand Islands east of New Guinea.

316 ↓↓↓隨便↓↓↓隨便↓↓↓隨便↓↓↓隨便↓↓↓隨便↓↓↓隨便↓↓↓隨便↓↓↓隨便↓↓↓隨便↓↓↓

We can understand, i think, that languages belonging to different cultures might have radically different structures for their kinship terminology. Different cultures place different responsibil- ities and expectations on certain personal relationships, and so it's reasonable that the vocabulary used to refer to those relationships should reflect these differences. But vocabulary mismatches between languages aren't by any means restricted to kinship terminology. It often happens, even in the case of languages like those of Western Europe that share an awful lot of cultural back- ground, that the vocabulary of two different languages carves up and organizes our experience of reality differently, so that it is literally impossible to establish a one-to-one correspondence be- tween the words of one language and those of another. In (3), i've got some examples of single words in English or French that have multiple equivalents in the other language.

(3) English French brown brun / marron / châtain monkey / ape singe chair chaise / fauteuil jug broc / cruche / pot carpet / rug tapis

The point here is that there is no one French word that covers the range of meanings covered in English by the word ‘brown’ — even if we restrict ourselves to ‘brown’ merely as the name of a colour. Some ‘browns’ are called ‘brun’ in French, some are called ‘marron’, and some are called ‘châtain’. Likewise, French makes a fundamental distinction between a ‘chaise’, which is a seat with a back, and a ‘fauteuil’, which also has arms. Both of these are called ‘chair’ in English — a word which is related historically to the French word ‘chaise’ but which covers the meaning of ‘fauteuil’ as well. Obviously, English can make the same distinction; what French calls ‘fauteuil’ is called ‘armchair’ in English. But that's part of the point. In English, an ‘armchair’ is regarded as a special kind of ‘chair’ — it's a chair with arms. In French, it's regarded as a different thing from a ‘chaise’.

On the other hand, whereas in English we distinguish between ‘monkeys’ and ‘apes’, French calls all non-human primates ‘singe’; and in French just about any floor-covering is a ‘tapis’, while in English we make a fundamental distinction between a ‘rug’ and a ‘carpet’, depending on whether it covers the whole floor or not.

Closer to home, a problem came up recently in my experience here in Taipei involving the English word ‘sign’. This word covers a lot of ground in English, and it was pointed out that Chinese has a variety of words that are not synonymous with each other but which all cover part of what in English is meant by the word ‘sign’. One dictionary gives me the ten words or expressions in (4):5 (4) English sign Chinese 記號, 符號, 信號, 標記, 標志, 告示牌, 指示牌, 招牌, 徵兆, 跡象

5Obviously, the morphemes 記, 號, 標, 示, and 牌 occur in more than one of these. But on the other hand, i know very well that at least 號 means other things besides ‘sign’. The point is that the different meanings covered by the English word ‘sign’ are conveyed in Chinese by at least 10 distinct expressions.

317 On the other hand, just as a French word may have more than one, mutually exclusive English equivalent, so can a Chinese word. I offer the examples in (5). (5a) offers several different English translations for a single Chinese morpheme; (5b) offers two different English translations for a bimorphemic word, or two-word expression, in Chinese. Part of my point is that there is no one word in English that conveys both of the senses in (5b), nor is there any one word or morpheme in English that covers all the meanings in (5a). (5) a. Chinese 掛 English hang; caught; ring up; hang up; concerned about; register b. Chinese: 伸腿 English: stretch one's legs; get one's foot in the door; kick the bucket

↑↑↑隨便↑↑↑隨便↑↑↑隨便↑↑↑隨便↑↑↑隨便↑↑↑隨便↑↑↑隨便↑↑↑隨便↑↑↑隨便↑↑↑

A similar problem arises with colour vocabulary. I've just given you one example of a vocabu- lary mismatch between English and French with regard to colour vocabulary. I'm going to men- tion a couple more here. In (6), i've given you a (somewhat oversimplified) diagram of how the visible spectrum is divided up in English into six different colour-terms, and for purposes of how the same spectrum is divided up in two African languages.6

(6) English: purple | blue | green | yellow | orange | red

Shona: cipswuka | citerna | cicena | cipswuka

Bassa: hui | ziza

Note first of all that both Shona and Bassa use fewer words to describe the same range of colours. Note that Bassa, in particular, uses only two, one (ziza) to describe the ‘warmer’ colours toward the long, red end of the spectrum and another (hui) to describe the ‘cooler’ colours toward the short, blue end. Note further that, although at first glance it looks as though Shona uses four words, there are really only three; cipswuka occurs at both ends of the spectrum, describing both the long- wavelength colours orange and red and the short-wavelength colour purple.7 The other two words, citerna and cicena, refer respectively to blue and yellow — and each of them also refers to half of the range that is covered in English by the word ‘green’.8 In short, some colours that in English would be called ‘green’ would be called ‘citerna’ (= ‘blue’) in Shona while others would be called ‘cicena’ (= ‘yellow’). Furthermore, some colours that would be called ‘blue’ or ‘yellow’ in English would in Shona be called ‘cipswuka’, not ‘citerna’ or ‘cicena’.

(7) English: green | blue | grey | brown

Welsh: gwyrdd | glas | llwyd

6Information drawn from Gleason, H. A. Introduction to Descriptive Linguistics, Orlando, FL: Holt, Rinehart and Winston (1961), pp. 4–5. 7Presumably, the Shona arrangement would make more sense if displayed in relation not to the linear spectrum which is how physicists tend to think of colour but to the colour-wheel that is commonly used by painters and decorators. 8I don't know whether there is any significance to the fact that all three of these words start with ci-, but as it stands i am inclined to suspect that this ci- may be a separate morpheme.

318 Another example of this sort of thing is shown in (7). In Classical Literary Welsh, there are three basic terms that cover roughly the same semantic area as four basic colour words in English. At least partly as a result of major influence from English, Modern Welsh has adapted to the point that it divides up the spectrum about the same way English does — with one important exception. While in Modern Welsh ‘glas’ is the equivalent of English ‘blue’, it is still used, as it was in the Classical Literary language, to refer also to the fresh green of growing plants; as far as the Welsh are concerned, the word ‘gwyrdd’ just won't do for that.

I need to make it very clear first of all that i am not saying anything about the visual experience of these different people. All evidence indicates that the Welsh, the French, the Africans, what- ever, see the same colours that English-speakers see, have the same visual experiences with re- gard to colour. The issue is, how do they talk about it? How is the colour-experience organized with reference to these different languages?

I also need to make it very clear that in this discussion we're not referring to any language's total vocabulary for describing colour; most languages have a rather wide variety of ways for doing this. I'm concentrating here on what Brent Berlin and Paul Kay call ‘basic colour terms’, which are defined as in (8):9 (a) Only single-morpheme words can be ‘basic colour terms’; ‘green’, for instance, qualifies, but not ‘greenish’ or ‘pea-green’. (b) The colours referred to by basic colour terms are clearly distinct from each other. ‘Red’ is a basic colour word, distinct from ‘green’, ‘blue’, etc.; ‘scarlet’ and ‘crimson’ are not because they refer merely to specific shades of ‘red’. (c) A basic colour term can be used in describing the colour of anything, unlike ‘blond’, which can only be used in talking about hair and perhaps a few other objects. () A basic colour is com- mon throughout the culture and recognized by all the speakers of the language in question. All fluent speakers of English know what is meant by ‘yellow’, but a lot of them are not familiar with the word ‘saffron’. (8) To be ‘basic’, a colour term must: a. consist of one single morpheme b. not refer merely to a more specific variety of the colour referred to by some other term c. not be restricted to a small number or range of objects d. be commonly used & understood throughout the relevant culture.

Berlin & Kay's research indicates that there is a hierarchy, a fundamental ranking and ordering, of colour words, which is listed in (9). It is possible for a language to have only two basic col- our words. In such cases, those words refer simply to ‘light’ and ‘dark’ or to ‘white’ and ‘black’. If a language has three colour words, one of them always refers to ‘red’; apart from ‘white’ and ‘black’, which from the physicist's point of view aren't really colours at all, ‘red’ is the most basic, most important colour for humans. (9) a. white, black b. red c. green, yellow d. blue e. brown f. purple, pink, orange, grey

9Basic Color Terms: their Universality and Evolution, Berkeley: Press (1969).

319 Languages with four or five basic colour words, in addition to having a word for ‘red’, have a word for either ‘green’ or ‘yellow’ or both. ‘Blue’ comes next, and ‘brown’ comes after ‘blue’. The last four basic colour concepts in Berlin & Kay's list are in (9f); there doesn't seem to be any strong ranking between them. The basic rule for a hierarchy like this is that, if a language has a basic colour word for any of the colours in this list, it also has basic colour words for all the col- ours above that one.

Note that Berlin & Kay's list is not meant to be exhaustive. It includes 11 colour categories; these happen to be the 11 recognized by the vocabulary of English. There are languages with 12 basic colour terms; as shown in (10), Russian distinguishes between two different shades of blue while Hungarian distinguishes between two different shades of red. In (11) is what i've been told by my students in previous years about the basic colour-vocabulary of Chinese.10 (10) Russian Hungarian sinyi ‘deep blue’ piros ‘bright red’ goluboi ‘light blue, sky blue’ vörös ‘deep, dark red’ (11) Basic colour terms in Chinese: 白 white, 黑 black, 紅 red, 綠 green, 黃 yellow, 藍 blue, 棕 brown, 紫 purple, 橙 orange, 灰 grey

Let's look at another example of differences in how different languages divide up experience. Both English and German have two different basic to describe the concept of eating, but they're not equivalent. In English, the difference between ‘eat’ and ‘devour’ has to do with the way it's done; ‘eating’ is just ‘eating’, but if you do it with great speed, essentially shoveling food down your throat as though you were afraid it was going to be taken away from you if you didn't hurry, and are intent upon consuming everything available, then the ‘devour’ is used. In German, on the other hand, the difference between ‘essen’ and ‘fressen’ depends on who's doing it; ‘essen’ is used if it's a human being, ‘fressen’ if it's an animal that's eating. (12) German essen vs. fressen depends on who's doing the eating English eat vs. devour depends on how it's being done Language and Thought Languages organize human experience differently. So far we've been looking at differences in vocabulary between different languages. But such differences don't occur only with regard to vocabulary; you can also find them in the of different languages. The grammars of different languages induce different ways of representing reality, and this can lead to different ways of perceiving reality.

It is occasionally claimed that English has two verbs, ‘love’ and ‘like’, while French has only one, ‘aimer’, that has to do the work of both.11 This is not quite true. It is true that French has

10I acknowledge the existence in addition of the word 青, which according to my dictionary can mean ‘green’, ‘black’, ‘blue’, ‘dark grey’. I do not know what the exact status of this word is in contemporary Chinese vocabulary (i'm fa- miliar with it mostly because it's the name of a friend!); Chinese is a language associated with a culture with several thousand years of history behind it, and for all i know 青 may, in its colour-meanings, be a relatively archaic word, or it may be limited (like ‘blond’) to certain relatively narrow -areas. There is no question that modern Chinese includes the basic words 綠, 黑, 藍, and 灰 to distinguish between these colours. 11This claim is often made by parents or teachers trying to discourage children from overusing the verb ‘love’, as in ‘I just love apples’ as opposed to ‘I really like apples’. The (linguistically, if not nationally, rather chauvinistic) im-

320 no verb that is precisely equivalent to the English verb ‘like’ — in the sense that it 1 expresses a feeling/emotion similar to but weaker than ‘love’ and 2 is syntactically transitive. And it is true that French typically uses ‘aimer’ where English would more normally use ‘like’ rather than ‘love’, as in (13b). But in addition, French has a verb, ‘plaire’, that expresses almost exactly the same semantic content as the English ‘like’ but has a very different grammar. While in English ‘like’ is a , the subject referring to the person who experiences the feeling/emotion and the whatever arouses that feeling, French ‘plaire’ isn't a transitive verb at all; its sub- ject is that which arouses the feeling in question, and the person experiencing that feeling is referred to by its indirect object (間接賓語);12 cf. (13c). (13) a. English: I love my wife French: J'aime ma femme. b. English: I (really) like apples French: J'aime (bien) les pommes. c. English: I like this colour. French: Ce couleur me plaît. (lit. ‘This colour pleases me.’)

The upshot of all this is that in English ‘liking’ is something that a person does to something or someone else; if i say ‘I like apples’, by using a transitive verb like ‘like’ i'm at least implying that this is something i'm doing to the apples. Whereas if i say in French ‘Les pommes me plaisent’, i'm implying that instead the apples are doing something to me, that i'm reacting to them in a certain way.

A similar sort of difference, at both the lexical and grammatical level, has to do with how differ- ent languages express possession. The like English and the Romance lan- guages like French have a special verb — ‘have’ in English, ‘avoir’ in French — that expresses possession. This verb is transitive in all these languages, at least superficially; in both English and French one can say the sentences in (14a). However, this verb doesn't act quite like other transitive verbs at least with regard to the passive ; while passive forms of this verb exist in both English and French, they have very idiosyncratic meanings; the two sentences in (14b) both mean ‘I've been used, I've been cheated, I've been taken advantage of’. But my main point here is that, in these languages, there is a specific verb used to express possession.13 (14) a. English: We have a nice house. French: Nous avons une chic maison. b. English: I've been had! French: J'ai été eu!

But many other languages don't have any such verb and express possession in very different ways. At least some of the , such as Russian, and Semitic languages like Hebrew ex- press possession by constructions that mean literally ‘to/for me there is …’, as in (15b).14 And South Asian languages like Sanskrit or Hindi and at least some of the like Hungarian express it by saying, roughly, ‘my X exists’ as in (15c).

plication is ‘Be grateful you speak English and not a lexically-impoverished language like French; your language allows you the option of two different verbs; use them.’

12This distinction isn't obvious in (13c) since the English verb ‘please’, which is related to it but is used rather more rarely, is also a transitive verb. But the French verb never takes a direct object, only an indirect one; the me in (13c) is specifically an indirect-object form. A more literal translation would be ‘This colour pleases to me’ or ‘This colour is pleasing to me’. 13The status of the Chinese verb 有 in relation to this whole discussion is something i'm not entirely sure about. 14Cf. the difference between 有我的馬 and 我有馬.

321 (15) a. English: I have a horse. French: J'ai un cheval. b. Russian: U menya loshad'. Hebrew: Sūs lī. (lit. ‘To me [is] horse’) (lit. ‘Horse [is] for-me’) c. Sanskrit: Mama aśvo'sti. Hungarian: Van lovam (lit. ‘My horse is’) (lit. ‘Is horse-my’)

↓↓↓隨便↓↓↓隨便↓↓↓隨便↓↓↓隨便↓↓↓隨便↓↓↓隨便↓↓↓隨便↓↓↓隨便↓↓↓隨便↓↓↓

When we have a thought and we want to express it in words, how do we go about doing this? Do we merely ‘translate’ our thoughts into words one after another? Are our thoughts naturally couched in some internal language (often referred to as ‘Mentalese’) which can be translated into an overt language so that we can communicate them to the people around us? This notion, common as it is, is incorrect and oversimplistic. This can be demonstrated first of all through a study of certain kinds of speech errors, errors that clearly indicate that our minds are not working merely on the level of individual words but rather on several levels simultaneously. Consider the examples in (16). In (16a) we have a phoneme being copied from one word to another word earlier in the phrase. In (16b) we have essentially a Spoonerism, the interchange of two phonemes. In (16c) we have the interchange, not of two pho- nemes but of two features — the voicing of /b/ and the voicelessness of /f/ — between two words. In (16d) we have the interchange of two -stems, the inflexional being unaffected. And in (16e) we have the interchange of two whole NPs. Clearly, since human beings perform all these different types of errors, human minds must be juggling linguistic elements — phrases, words, morphemes, phonemes, phonological features — at several different levels at once. Human linguistic behaviour is not simply linear, and there's no reason to think that human thought is, either. (16) What Was Said What Was Meant a. the firing of minority faculty the hiring of minority faculty b. odd hack [Šd hæk] ad hoc [æd hŠk] (專門的) c. pig and vat big and fat d. There are many churches There are many ministers in our minister. in our church. e. He sliced the knife with a salami. He sliced the salami with a knife.

Speech errors like those in (16) are constrained in ways that seem to have little to do with our thought processes. For instance, phonological errors like those in (16a–c) typically only affect content words, not function words; this is another example of the importance of this distinction in linguistics — but since function words don't seem to correspond directly to thought at all, it's difficult to understand why they should be peculiarly immune from phonological errors, which are presumably unconnected to conscious thought. The fact that can be interchanged in- dependently of their morphological marking, as in (16d), suggests that inflexional morphemes are added to them at a stage in the generation of the sentence subsequent to their being inserted into the sentence, instead of all the morphemes relevant to a particular word or phrase going in at once.

There's another, more cross-linguistic aspect to this whole question of the relation between lan- guage and thought that is of interest to me as a typologist. You'll remember, back when we talked about syntactic typology in Chapter 8, we talked about how different languages order grammatical elements differently. You'll remember that some languages, like English and French and German, have prepositions, while other languages, like Japanese and Hindi, have postpositions; i've given

322 some examples of these in (17). You'll remember in particular that there are logically six differ- ent ways the major sentence constituents — the subject, object, and verb — can be ordered with respect to each other, and for every one of these 6 possibilities there are at least a few languages known that prefer that particular order, although cross-linguistically some orders are definitely preferred over others. (17) Prepositions Postpositions English: from London Hindi: Dilli se Delhi from on the table ‘from Delhi’ French: de Paris mez par from Paris table on ‘from Paris’ ‘on the table’ sur la table Japanese: Tōkyō kara on the table Tokyo from ‘on the table’ ‘from Tokyo’ German: von/aus München bō de from Munich stick with ‘from Munich’ ‘with a stick’ am Tisch on-the table ‘on the table’

But if you've spent your whole life — furthermore, if not only you but the culture you live in, everyone around you, has spent their whole lives — speaking one particular language that happens to have a very strong preference for one order over all 5 of the other possibilities, you're very liable to take that order for granted. For instance, if most or all of your experience of human language is with sentences built on an SVO plan, with the verb in the middle preceded by the subject and followed by the object like the ones in (18), you're very liable to think that's somehow ‘normal’, that's the ‘obvious’ way for sentences to be built. This is even more likely to be the case if you're acquainted with more than one language, but for whatever reason all the languages you're acquainted with happen to favour the same order. For instance, all of you are fluent in Chinese, and all of you have at least some facility with English. But these two languages, how- ever different they may be in other ways, both happen to be SVO languages. And if you also happen to be somewhat familiar with French or Spanish or Italian or any one of a large number of other languages, all of which happen to be SVO, it's understandable you're going to be inclined to take that ordering pretty much for granted. You're very liable to think that sentences in these languages are built this way simply because that's the ‘obvious’, ‘logical’ way for sentences to be built. (18) a. My daughter wants to read this book (English) b. Ma fille veut lire ce livre. (French) c. Mia figlia vo leggere questo libro. (Italian) d. 我女兒要讀這本書。 (Chinese)

So far, i've been speaking hypothetically; i can't really speak for you people. But i can tell you that, especially within French culture, this attitude is quite common. Traditionally, the French are encouraged by their culture, their educational system, etc. to believe that their language is

323 somehow particularly logical.15 Traditionally, French schoolchildren are explicitly taught that in French the subject goes before the verb and the object, if any, goes after because that is the log- ical way to do it — before one can have an action, one must have an actor, someone to do the action, and only then can one have not only an action but something acted upon. Take my word for it; as a child, i spent four years in French schools, and that's what i was taught. Of course, being a native speaker of English, which is also an SVO language, this did not particularly both- er me at the time.

But i was forceably reminded of this many years later when i began studying Breton. Breton is a Celtic language, and like all favours VSO order; cf. the examples in (19). (19) a. Harzal a ra ar c'hi. barks the dog ‘The dog barks.’ b. E vezo aes al labour ganimp. will be easy work for us ‘Work will be easy for us.’ c. Emañ Ronan en e di gant e dud. is Ronan at home with his parents ‘Ronan is at home with his parents.’ d. E lenn Arzur e gentel en e levr. reads Arthur his assignment in his book ‘Arthur reads his assignment in his book.’ e. Deut ez eus c'hoaz tud d'az kerc'hat. come there are already people for you looking ‘Some people have already come looking for you.’

But Breton is spoken in France and most of the traditional descriptions of are written in French for French-speaking people. And on the very first page of a 1918 grammar of Breton of which i own a copy,16 written in French, it says, ‘Instead of following the logical order (subject, verb, object), the Breton sentence follows the order of thought: it's the most striking word that comes first.’ Note that the author distinguishes between ‘logic’ and ‘thought’; note also that, like the traditionally-educated Frenchman he is, he automatically assumes that ‘logical’ order is the same as that enshrined in the typical French sentence: subject, verb, object. Now, by the time i read this sentence, i had had a little too much experience with formal logic to readily accept this assumption; some electronic calculators make use of a form of mathematical notation that is some- times called ‘reverse-Polish logic’, according to which you first enter the operation — +, -, *, whatever — and only afterwards the ‘operands’, the numbers being acted upon. There are some mathematical implementations for which this is the most ‘logical’, or at least the most efficient way to do things — and it corresponds very closely to Celtic VSO sentence structure. So from this point of view the sentence structure in a Celtic language like Breton is much more ‘logical’ than the SVO order of a language like French or English.

15In fact, this is why i suspect that the words 法文 and 法國, by which Chinese refers to the and the French nation, would be rather flattering to the French; given that one of the meanings of 法 is ‘law, rational method’, it suits their national self-image fairly well. 16F. Vallée, La langue bretonne en 40 leçons (6e éd.) Librairie Saint-Guillaume, 1918.

324 But again, if you've spent your whole life using SVO languages, isn't that going to affect your thought processes? A certain amount of our more complicated thought processes tend to take place in a very linguistic form. I know, when i'm working through a difficult problem, a lot of my thinking is in the form of phrases and sentences. In some cases, i may actually pronounce these phrases and sentences with my mouth; but even when i am thinking completely silently, these phrases and sentences are forming themselves in my brain. And i would guess that a lot of you have had similar experiences. The point is, if these phrases and sentences are being formed inside your brain, How are they structured? and What language are they in? The answer to the first question is pretty much dependent on the second, right? For me, they're usually in English; for you, i would guess they're normally in Chinese. And that means that in both cases they tend to be SVO structures because both these languages are SVO languages. And if the thinking you are most aware of tends to be modelled on your native language and the grammat-ical structures characteristic of that language, isn't it reasonable to imagine that all your thinking, even the think- ing you're not so aware of, is likewise modelled on the structures characteristic of that language?

But if that's so, then you have to wonder about the Japanese people, whose language is verb-final, or the Arabs, whose language is verb-initial. Do these people even think like we do? And if not, can we ever possibly communicate with them?

This is one of the many reasons why i, as a typologist, find linguistic typology so very important: It makes us aware of how different languages are from each other, while at the same time remin- ding us that we are all human beings, and all this linguistic diversity is an expression of a com- mon humanity. The connection between our thoughts and the linguistic expressions and struc- tures we use to express them is not simple and straightforward; we don't translate our thoughts into words one after another. Instead, we convert thoughts into words and words into thoughts in terms of substantially larger units, including whole noun phrases and verb phrases.

↑↑↑隨便↑↑↑隨便↑↑↑隨便↑↑↑隨便↑↑↑隨便↑↑↑隨便↑↑↑隨便↑↑↑隨便↑↑↑隨便↑↑↑

Sapir-Whorf Hypothesis

When she thought without the subtle intention to speak, there often were no words at all. A quick, almost holo- graphic perception of the idea shot through her. — Gregory Benford & David Brin, Heart of the Comet (Bantam, 1986), p. 189

There is a hypothesis that has been around in linguistics for close to a century, and i suspect has probably been around in popular culture for rather longer and will probably be around in popular culture for quite a while yet. I say that partly because this hypothesis is easy to credit because it makes intuitively obvious, or at least attractive, statements about the relationship between human language and human culture; however, at least in the way it's commonly understood, this hypothe- sis is rather misleading.

This hypothesis is usually referred to as the Sapir-Whorf Hypothesis, connecting it with two important figures in the history of American linguistics in the first half of this century. When i say ‘American linguistics’ in this case, i mean three things. 1 Both (1884–1939) and Benjamin Lee Whorf (1897–1941) were Americans; 2 both were involved in the development of a distinctly American approach to the study of language, an approach that was self-consciously distinct from the tradition of linguistic scholarship that had developed in Europe; 3 both are known among other things for their research on the native languages of North America, and it is primarily in connection with this research that the Sapir-Whorf Hypothesis was developed.

325 The Sapir-Whorf Hypothesis, sometimes known more generally as the Linguistic Relativity Hypothesis, claims basically that there is a strong causal interaction between one's thought patterns and the language one speaks. In (20) i've given you a somewhat edited version of a 1940 statement of this hypothesis by Whorf.

(20) The grammar of each language is … itself the shaper of ideas, the program and guide for the individual's mental activity …. Formulation of ideas is not an inde- pendent process … but is part of a particular grammar, and differs, from slightly to greatly, between different grammars. We dissect nature along lines laid down by our native language. The categories and types that we isolate from the world of phenomena we do not find there because they stare every observer in the face; on the contrary, the world is presented in a kaleidoscopic flux of impressions which has to be organised by our minds — and this means largely by the linguistic sys- tems in our minds. We cut nature up and organize it into concepts … largely be- cause we are parties to an agreement to organize it in this way — an agreement that holds throughout our speech community and is codified in the patterns of our language. The agreement is, of course, an implicit and unstated one, BUT ITS TERMS ARE ABSOLUTELY OBLIGATORY; we cannot talk at all except by subscribing to the organisation and classification of data which the agreement decrees.

The Sapir-Whorf Hypothesis makes some definite claims about the interaction between a commu- nity's language and the way that community perceives reality, some of which can be tested, and we'll be talking about such tests shortly. But the Sapir-Whorf Hypothesis is also subject to much misunderstanding by the general public.17

Popular discussions of the Sapir-Whorf Hypothesis are often taken up with stories about the sup- posedly amazing vocabularies of exotic peoples and about what they presumably say about those exotic cultures' ways of viewing the world. For instance, one frequently comes across claims that the Inuit or Eskimos have dozens, or even hundreds, of words for different kinds of snow. And, within the context of the popular understanding of the Sapir-Whorf Hypothesis, the conclusion is typically drawn that the large snow-vocabulary that the Inuit language has reflects the great importance of snow — and the different types of snow — in Inuit culture.

Unfortunately, the whole factual basis for this discussion is missing.18 The argument is based on the claim that the Inuit language has a large number of words for different kinds of snow, but in fact, according to one of the most reliable sources, the Inuit language has only two basic words for snow (cf. (21a)): One refers to snow while it's falling through the air or blowing in the air, the other to snow lying on the ground. That's it. English has more words for snow than that; cf. (21b)

17One correspondent of mine has pointed out that a lot of what's really at issue in the Sapir-Whorf Hypothesis — at least, what Whorf himself had in mind in statements like that in (20) — is that how you look at the world, at reality, is conditioned by the language that molds your thoughts, and that in particular modern quantum physics simply makes a lot more sense if you can think about it in a Native American language than if you use a Modern Western European language; unfortunately, most of us, including most physicists, are more familiar with Modern Western European languages then we are with Native American languages.

18For a fuller discussion of the issues raised here, see Geoffrey K. Pullum's often hilarious essay ‘The Great Eskimo Vocabulary Hoax’, pp. 159–71 of his equally hilarious collection of essays under the same title: The Great Eskimo Vocabulary Hoax and Other Irreverent Essays on the Study of Language (University of Chicago Press, 1991).

326 (21) a. qanik ‘snow in the air’ aput ‘snow on the ground’ b. snow slush sleet flurry blizzard crust powder drift

The claim that Inuit has some amazingly huge number of words for snow is what we Americans call an ‘urban legend’, a claim that has no real basis in fact but has been repeated often enough that a lot of people believe it and, furthermore, has gotten gradually exaggerated in the course of being repeated. The original claim was that there were three words for snow (one of which later turned out to be a misunderstanding); Whorf reproduced this claim in one of his popular articles, but in the process changed the number to ‘at least seven’. Subsequent retellings inflated the num- ber into two digits and even three digits without ever referring to any decent evidence or reliable research; a few years ago there was an offhand remark in referring to Inuit's ‘100 words for snow’, and a Cleveland weather-forecaster is on record as referring to its ‘200 words for snow’. The details of this irresponsible propaganda and inflation are chronicled by the research described in the Pullum essay mentioned in note 18, which suggests that the dura- bility of this claim, and its attendant exaggeration, is due to subliminal racist attitudes towards those extremely exotic (or simply weird) people living in the extreme north, who as is well-known (though never documented) have so many other weird cultural practices.

Pullum further undermines the conclusion of the popular discussion of Inuit snow-vocabulary — the hypothesis that the putatively large number of Inuit words for snow reflects the great impor- tance of this concept in Inuit culture — when he points out:

When you come to think of it, Eskimos aren't really that likely to be interested in snow. Snow in the traditional Eskimo hunter's life must be a kind of con- stantly assumed background, like sand on the beach. And even beach bums have only one word for sand. But there you are: the more you think about the Eskimo vocabulary hoax, the more stupid it gets. (p. 166)

Another personal favourite of mine is the claim that Saame, the language spoken in extreme nor- thern Scandinavia by the people commonly called ‘Lapps’, has five words for ‘antlerlessness’, the condition of being without antlers; these five words refer to the five reasons why a reindeer may not have antlers. I haven't been able to find the words themselves in my reference sources, but i do remember what they mean. As shown in (22), reindeer may lack antlers because of being the wrong sex, since only male deer carry antlers; only male deer have any use for antlers. A reindeer may lack antlers because of being the wrong age, since only adult male deer carry antlers; only adult males need them. A reindeer may lack antlers because of its being the wrong time of year, since stags only need their antlers during the fall; they lose them in the winter and gradually grow them back in the spring. There are various kinds of disease that can prevent a stag from growing antlers. And finally, it occasionally happens that human beings artificially remove a stag's antlers for some reason. (22) Five reasons why a reindeer may lack antlers: a. wrong sex b. wrong age c. wrong time of year d. disease e. artificial removal

That's all very well but there are problems with this aspect of the Sapir-Whorf Hypothesis. First of all, the claim that Saame has five words for antlerlessness isn't really a claim about the Saame language itself so much as it is about the economy of the Saame or Lapp people. And when you come right down to it, it's not so much that Saame has five words for antlerlessness; it's that rein- deer herders have five words in their professional vocabulary for antlerlessness, and because

327 most reindeer herders in Scandinavia happen to be Lapps, they speak Saame, and so their profes- sional vocabulary is part of the Saame vocabulary.

I have never seen a reliable claim that the Saame language actually lacks a word for the general condition of being without antlers, without specifying why the antlers are absent. If this were the case, if a Lapp couldn't possibly say that a reindeer didn't have antlers without at the same time explaining why it didn't, the way it's virtually impossible in Chinese to mention your brother without specifying whether he's your older or younger brother, that might be interesting. But as far as i know, it isn't true of Saame. In the absence of any such evidence, i assume that Lapps who are not professionally involved in taking care of reindeer, when they want for whatever rea- son to say that a reindeer has no antlers, have a perfectly good generic expression in their language that conveys that meaning and no more detail and they use that. In which case the fact that Saame has five words for the five possible reasons for antlerlessness is merely a fact about professional concerns of reindeer herders and the fact that a lot of them happen to be Lapps and to speak Saame. It doesn't prove that the five possible reasons for antlerlessness have some sort of deep signifi- cance for Lappish culture, more than antlerlessness itself.

But the Sapir-Whorf Hypothesis doesn't refer simply to vocabulary; it also makes claims about grammar. Whorf noted that the has no grammatical category corresponding to what in Western European languages is called ‘tense’, and argued that a culture that uses such a language must necessarily have a completely different concept of time from that characteristic of Western European civilization, if it has a concept of time at all. Let's just note in passing that Chinese doesn't have any grammatical category ‘tense’ either, and yet manages very nicely to express the thoughts that category is designed to express in Western European languages; obvi- ously, the Chinese have no difficulty understanding time in the same way as Western Europeans, even if they don't use quite the same methods of expressing it.

At this point it is a good idea to distinguish between what we would call a ‘strong’ version of the Sapir-Whorf Hypothesis and a ‘weak’ version, as in (23), because as we shall see while one version may be false the other may well be true. (23) a. The Strong Version of the Sapir-Whorf Hypothesis: It is impossible to conceive a thought which cannot (easily) be expressed in your language (or in some language in which you happen to be fluent). b. The Weak Version of the Sapir-Whorf Hypothesis: The language you are accus- tomed to using channels your thinking by providing appropriate categories, so that it is easier for you to conceive thoughts that can be (easily) expressed in that lan- guage than thoughts that cannot be.

The ‘strong’ version of the Sapir-Whorf Hypothesis, the claim that a human being is literally unable to conceive or formulate any thought that cannot be expressed comfortably in hanns lan- guage, is relatively easy to disprove. For one thing, although as mentioned earlier language is often an important part of our internal thinking processes, every one of us from time to time thinks thoughts without the help of language. This is made all the clearer if you think of all the times you've had trouble finding the right words to express a thought you have, even if you're speaking your own language, a language in which you are fluent; we've all had this experience, and it proves that some thinking — some very conscious thinking — does go on without language being involved at all. But further evidence against the ‘strong’ version of the Sapir-Whorf Hypo- thesis comes from the fact that we are able to step back from the purely language-supported thought patterns that are normal to us and think about the very fact that, on the one hand, we normally use language in thinking but, on the other hand, we can step back from it; obviously, the thinking we're doing when we do that must to some extent be independent of language. Furthermore, the

328 fact that we can step outside our own culture and think about how other cultures understand the world, even cultures whose languages we do not perfectly understand — the very kind of thinking that led to the Sapir-Whorf Hypothesis in the first place — proves that we can think independently of our language; human language does not control human thought to the absolute degree that Whorf apparently believed it did.

Nevertheless, there is experimental evidence supporting a much weaker version of the Sapir- Whorf Hypothesis, one that would claim that, although our language may not rigidly and abso- lutely control our thought-patterns, it does channel them in the sense of making some thought- patterns much ‘easier’ than others on the basis of what categories a particular grammar treats as of greatest importance. I'm going to tell you about two sets of experiments done in comparing the perceptions of different communities distinguished by language; in each case, one group of experimental subjects were native speakers of , the other of one or another of two Native American languages. In each case, the activity that the experiment focussed on had nothing directly to do with language itself, but in both cases it was found that the subjects tackled the activity differently in ways that reflected differences between their languages.

The language, spoken by about 200,000 people in the Southwestern United States, has an interesting feature in its grammar. Any verb referring to an act of holding or grasping must be marked morphologically to agree with the shape of the object being held. Thus, such verbs take different forms depending on whether the object being held is long and flexible like a piece of string, or long and rigid like a stick, or flat and flexible like a piece of paper, and so on. In a very broad sense, this notion should not be strange to you; after all, a classifier language like Chi- nese has different classifiers for long, flexible things (條), long, slender (though not necessarily rigid) things (枝), and flat, flexible things (張). What's important to understand about Navajo, in considering the story i'm about to tell, is that in that language the verbs themselves are marked morphologically to agree with their objects on the basis of shape.

Back in the 1950's a research project was done to compare the sorting behaviour of four groups of children.19 The basic experiment was done like this: Each child was shown two objects that were different from each other and a third that was like each of the other two in different ways; it might be the same size as one of them but the same colour as the other, for instance. And the child was asked to place the third object with the object it most resembled. The basic point of this kind of test is to see what features the child focusses on, which characteristics the child judges are most important: For instance, if the child tends to focus on size instead of colour, hann will tend to group objects that are the same size together whatever their colours may be, but if hann tends to focus on colour instead of size hann will do the opposite.

This kind of test had been used in the past to study how different concepts develop in the mind of a ‘typical’ child, and experiments done both in America and in Europe had led psychologists to conclude that it takes longer for a child to develop a concept of shape than a concept of size or colour, so that young children typically ignored shape entirely when sorting things in this way. However, Carroll & Casagrande suspected that they might get different results if they tried this test out on children growing up in a Navajo-speaking community. In this community, as far as language was concerned children as young as 3 were using the shape-agreement morphemes cor- rectly on verbs of holding and grasping, and so the Sapir-Whorf Hypothesis would predict that if

19The relevant research was reported in J. Carroll & J. Casagrande, ‘The Function of Language Classification in Behavior’, published in 1958 in Readings in Social Psychology (Holt, Rinehart & Winston), ed. by E. Mccoby, T. Newcomb, and E. Hartley.

329 such children were confronted with the sorting test they would do a lot better at sorting by shape than children of the same age growing up in English-speaking communities. This is the prediction that Carroll & Casagrande set out to test.

In their initial experiments, Carroll & Casagrande tested two groups of children, both of which were living on the same Navajo reservation and were about the same age and in general were very much alike, except that one group was made up of children whose families normally used the while the other was made up of children whose families had pretty much abandoned use of that language and were using English most of the time. Thus, although they were in most respects similar the two groups of children differed from each other in terms of their native languages, Navajo vs. English. On the basis of the Sapir-Whorf Hypothesis, Carroll & Casagrande predicted that children who were growing up speaking Navajo would be much more aware of shapes than their agemates who spoke English instead, and that this would be reflected in their performance on the sorting test. And indeed, this was the case; the Navajo-speaking children were more likely to sort objects by shape — and correspondingly less likely to sort them by colour or size — then their English-speaking agemates, indicating a much greater awareness of differences in shape. There could be no doubt that the Navajo language had contributed to this sensitivity.

But things subsequently got more complicated (don't they always?). Bear in mind that in this first set of experiments Carroll & Casagrande had compared two groups of children both native to the Navajo reservation and in general pretty much alike in terms of age and socio-economic status; the only difference was language. Later, they extended their research to look at two other groups of children, neither of which had had any contact with the Navajo reservation, the Navajo language, or Navajo culture. They selected a group of children from a middle-class neighbourhood in Boston and another from a working-class neighbourhood in Harlem, New York. Both of these groups were the same average age as the two groups that had been studied on the Navajo reserva- tion. And, of course, both were English-speaking.

Rather surprisingly, it turned out that the Boston children did approximately as well at sorting by shapes as the Navajo-speaking children on the reservation, although the children from Harlem did not (the results of all these experiments are summed up in (24)). This was quite noteworthy, since as i've already mentioned it was generally understood among psychologists, on the basis of many years of performing such tests on children from various communities in both North America and Europe, that children at this age did not normally think in terms of shape in sorting objects. That the Navajo-speaking children should do remarkably well in this respect was predicted by the Sapir-Whorf Hypothesis; but why should the English-speaking children from Boston do so much better than expected? (24) Comparison of Shape-Awareness of Young Children Speaking Navajo (a Shape-Aware Language) and English (a Shape-Unaware Language) Navajo-speaking children on Navajo reservation aware of shapes English-speaking children on Navajo reservation not aware of shapes English-speaking children in middle-class Boston aware of shapes English-speaking children in working-class Harlem not aware of shapes

In interpreting this experiment, it should be noted that it was done in the 1950's, relatively early in the great economic boom that enveloped all of Western civilization during the period follow- ing World War II. As a result of these economic circumstances, middle-class American fami- lies had a lot more money than families of the same class would have had before the war, which meant that their children would have had access to a much larger variety of toys than their par- ents had had — INCLUDING TOYS SUCH AS BUILDING BLOCKS AND JIGSAW PUZZLES THAT DEPEND

330 CRUCIALLY ON SHAPES. I can tell you that i was a child in the appropriate age-bracket at the time we're talking about, and i sure had access to toys like that. However, the children of the poorer, working-class families in Harlem would have had much less access to such toys because their families could not afford them; and the same would have been true of the children on the Navajo reservation, whether their families spoke Navajo or English.

So basically, the conclusion was that 1 3-year-old children from middle-class families in the 1950's would have had access to toys and games that depend crucially on shape, and therefore would tend to be more aware of shape than their parents would have been at the same age 2 At the same time, poor children of the same age generally did not have access to such toys and there- fore would tend not to be as aware of shapes as they were of sizes and colours 3 However, child- ren being brought up speaking Navajo would have been more aware of shapes than children brought up speaking English, no matter how rich or poor their families were, because requires that awareness. As one commentator put it, ‘the effects of a middle-class upbringing are similar to those of learning to speak Navajo’, at least in this respect.

The second experiment i want to tell you about was carried out during the 1980's and compared the behaviour of two groups of adults. One group was a bunch of students at the University of Chicago, all of them American citizens with English as their native language; the other was a group of Mexican Mayans, Mexican citizens whose primary language was Maya, a language na- tive to and some of the other countries of Central America. The critical linguistic differ- ence in this case is that both languages have ways of marking plural number on nouns, but in Maya these tend to be optional while in English they are obligatory. Furthermore, while plural markers in Maya are optional, they are much more likely to be used on some nouns than on others. In particular, Mayans are much more likely to use plural markers on nouns referring to people and animals than on nouns referring to inanimate objects such as rocks and bowls. So in general — at least according to the Sapir-Whorf Hypothesis — English-speakers are forced by their lan- guage to be aware of the distinction between ‘one’ and ‘more than one’ to an extent that Mayans are not, and are therefore likely to be more conscious of this distinction than Mayans are, especi- ally with regard to inanimate objects — even when not actually using language to talk about them.

In this experiment,20 the two groups of subjects, the English-speaking group from Chicago and the Maya-speaking group from Mexico, were presented with a variety of pictures that basically showed the same scene but were different in terms of the number of various entities in the scene: there might be one or two dogs, one or two chickens, one or two bowls, one or two babies, one or two bricks lying on the ground, etc. (i've given samples in Fig. 17.2 on the next page). The subjects were asked to do a variety of things with these pictures. Sometimes they were to describe the pictures while holding and looking at them; sometimes they were shown a picture, and then asked to describe it from memory after it had been taken away; sometimes they were invited to match one picture up with a variety of similar pictures. The Sapir-Whorf Hypothesis predicts that in performing all these tasks the English-speakers would display much more awareness of the number of babies, bowls, bricks, chickens, and dogs then the Mayans, and furthermore that while the English-speakers would display roughly equal awareness of the number of all these dif- ferent things the Mayans would pay more attention to the number of babies, chickens, and dogs then to the number of bowls or bricks, and that this would be true whether the subjects were asked to perform a linguistic task such as describing a picture or a non-linguistic one such as matching it with another one. This prediction of the Sapir-Whorf Hypothesis was confirmed by Lucy's ex-

20reported in J. Lucy, Grammatical Categories and Cognition: a Case-Study of the Linguistic Relativity Hypothesis, published 1992 by Cambridge University Press.

331 periments; there definitely is a difference between the way Mayans and English-speaking Ameri- cans look at the world, and that difference definitely corresponds to a difference between the grammars of their respective languages.

Fig. 17.2. — How Many of What, and How Important is it, Anyway?

The conclusion is that grammar does influence our thinking, but it's not the only factor in our cultural environment that does so. Language is certainly not a dictator that forces us to think in only those ways that are compatible with its structure, as the ‘strong’ version of the Sapir-Whorf Hypothesis would imply; it is manifest that human beings can think thoughts without the inter- vention of their respective languages, and similarly human cultures can have habits and commu- nal structures that are independent of their languages.

But on the other hand, language, both vocabulary and grammar, definitely does influence our thinking, and especially since most of human culture is passed on from one generation to the next by means of language it is at least to some extent molded by language. As witness the rela- tively early stage at which Navajo-speaking children become sensitive to shapes, because of its deep interaction with other aspects of human culture language is also a very useful tool for anthro- pological research, often tipping us off to aspects of culture that we might otherwise not have noticed.

332