Words and Meanings
From Inside Language (1997) now out of print: this is the prepublication doc file). Several other pages relate to word meaning, see menu Words and my book It's All in a Word
To many people the most obvious fact about language is that it consists of words. Parents rejoice at the child’s first word, not at their first speech sound or their first sentence. Newspapers have columns about words, some aimed at improving people’s vocabulary, others making wry comments on current usage. Progress in a foreign language is counted by how many words the students have mastered. Without words there would indeed be little to say. But the earlier chapters have already made it clear that language is far more than words strung together. Without the systems of grammar, pronunciation and writing, words would convey little more than a series of labels.
This chapter looks at some of the properties of words and their meanings. The main emphasis is to demonstrate that the meanings of the words of a language form a system. Words are not isolated units with separate meanings of their own, to be learnt one at a time, but connect with each other in highly intricate relationships of meaning. This chapter first takes universals of human life, illustrated from colour words. Next it turns to the psychological approaches to the mental storage of meaning through the concepts of levels, components, and fields of meaning. It ends by looking at the process of derivation through which new words can be constructed. One of its threads is that meaning has to be looked at from many different angles and analysed in many different ways to begin to approach its complexity in human language.
1. Lexical entries and the dictionary in the mind
To understand speech or to produce it, speakers need to store the words of the language in their minds. Processing language fluently means having virtually instantaneous mental access to information about thousands of words. In this sense the mind has a dictionary that includes all the information a speaker needs about vocabulary. This is not to say that there is an actual list tucked away somewhere in the mind like that in the printed book. Nevertheless much of the information in the mind must overlap with that in the book; most of what is in a dictionary reflects what speakers already know about words. One of the uses of printed dictionaries is indeed to record the way that people use words, which in turn reflects the mental dictionary they have in their minds. (webtext: what are dictionaries for?)
Linguists often prefer the word ‘lexicon’ over ‘dictionary’, perhaps because
it provides a convenient adjective ‘lexical’ to use in phrases such as ‘lexical entry’. Printed dictionaries have entries;
mental lexicons have lexical entries that store parallel information even if in
a different way. To see the content of a lexical entry let us look at the noun
help. Speakers of English know that help is pronounced /help/: they know its
pronunciation and probably its spelling. When help is a noun, they know that
it is usually
‘uncountable’; that is to say, it is impossible to say he gave me two helps;
and they know that it can be followed by the preposition with, he
gave me some help with my car; in other words they know grammatical
information about how help may be used in sentences. They also know that, as a noun, help can be
preceded by the adjective great in some uses when it is countable,
Mark was a great help, but not by large, as in Mark was a large
help: they possess information about the words that can occur close to it,
technically called ‘collocation’. Help also enters into certain set
phrases such as There’s no help for the
wicked, and certain extended
expressions, to be of help, etc: speakers know the idioms of which
help forms part. Putting these elements together, the lexical entry in
the mind contains the information:
help: /help/ noun, uncountable, +with, +great, not large, no help for the wicked, to be of help ...
This entry is close in content, if not in form, to the skeleton of an entry in a standard printed dictionary. All of this information must be stored in the lexical entry in the mind in one way or another.
Detailed as the lexical entry for help already is, it has not mentioned anything about the meaning of the word in the conventional sense. The speaker’s knowledge of a word is not only its ‘meaning’ but also the way that it is pronounced, and the way it relates to the structure of the sentence. The meanings are only one aspect of what speakers know about words.
2. Colours and linguistic relativity
However much words in different languages seem to refer to the same ‘thing’, the meaning of a word is relative to the whole system of other words in the language. Languages do not just have different words for the same thing; they have different meanings for different things. Words do not connect directly to objects in the world; they connect indirectly via our minds. So they express the speakers’ attitudes towards the world they see and the way that they interpret it in their culture.
Hence anybody who learns another language is in for some surprises. Arabic speaking learners of English for example complain that they do not know how to express whether a cousin is male or female and whether an uncle comes from their mother’s or their father’s side of the family. English speakers have problems in mastering the degrees of respect with which Japanese relatives are referred to when talking to someone within the family and to someone from outside. Within the family the respectful forms are used with the ending -san. A mother might be formally addressed within the family as okaasan, more usually as okaachan; outside the family the word for ‘mother’ is usually haha.
Let us take a case in which there is little doubt that the objective facts of the world are the same for all human beings. The colours of objects seem to be a straightforward fact, expressible in precise physical terms of wavelength, luminosity, hue, etc, down to the 16 million colours claimed in advertisements for computer monitors. Colours are the same everywhere because they are part of the physical makeup of the world. So languages might be expected to have the same range of words for colours. This section tackles the thorny issue of the similarities and differences between colour words in different languages.
The Welsh colour term glas includes part of English green and grey as well as blue; Japanese traffic lights have a colour awo that is known as blue rather than green and is also used for the sky. Russian has two colours for blue, sinij (dark blue) and goluboy (sky blue). Hungarian has two colour words for red, piros and vörös, the difference, according to a Hungarian, being that piros is the red in the Hungarian flag, vörös the red in the Russian flag. Speakers of different languages do not see the world in a uniform way; some see colours where others do not.
Celebrated research by Berlin and Kay, however, revealed an underlying pattern to colour words across languages. The basic colour terms can be grouped on a scale, shown in Fig 4.1. All languages have the first two colour terms on the scale, namely ‘black’ and ‘white’. In Dani, spoken in New Guinea, the only colour terms are indeed mili (‘dark-cool’, i.e. black) and mola (‘light-warm’, i.e. white). Next on the scale comes the colour term ‘red’. Some languages, such as Tiv spoken in Nigeria, have a three-term system of ‘white’, ‘black’, and ‘red’. Next along the scale come ‘green’ and ‘yellow’, yielding the five-term colour system found in Navaho. The progression continues through languages that also have ‘blue’, ‘brown’, and a final group of ‘purple’, ‘pink, ‘orange’, and ‘grey’. The full scale of basic colours is shown in Fig, 4.1:
Each language exploits this scale by using all the colour terms to the left of a particular point but none to the right. Hununoo, spoken in the Philippines, goes as far as ‘green’ and ‘yellow’. Ancient Egyptian had eight terms, including ‘purple’ but excluding ‘brown’ in favour of ‘gold’. Hebrew has terms for all eleven.
Obviously English has many more colour words than the eleven basic colour terms, such as burnt ochre, avocado, roan, lovat green, and so on. Basic terms have to be single words (thus excluding electric blue), must not be included in the meaning of other terms (scarlet is a kind of red), and are not used exclusively for a single object (thus excluding avocado but including orange). The most extensive range of colour terms is used by cosmetics manufacturers. Cayenne Red, Rare Orchid, and Innocent Rose are colours of lipstick for instance. 70% of these advertising terms do not figure in standard dictionaries. A clothing catalogue advertises a shirt in Cypress, Vintage Rose, Copenhagen Blue, Classic Navy, Russet Red, Black, Sunwashed Purple, Golden Sand, and Ivory. At the opposite extreme, the language of artists and art historians tends to be sparing in colour terms, restricting themselves to basic terms with some modification, for example warm red. Even Turner’s amazing colours are talked about by critics as blues, reds, and yellows.
The order to some extent reverses the scale of basic terms, with black and white being learnt rather late. By about the age of 4 years, children recognise about 73% of the colour terms tested, suggesting that there is no necessity to spend much time on basic colour terms in primary school.
One reason why colours may be structured into the set of basic terms in human languages is that certain ways of looking at the world are determined in part by how the human senses work. An experiment by Berlin and Kay asked speakers of different languages to look at a set of colour chips, similar to the paint charts available in DIY stores. They first had to point to all the chips that could be called a particular colour – which chips are blue? which red?—then to the ‘best’ example of each of the basic colour terms—which is the best red? the best blue? Whatever their native language, people by and large agreed on the best example of each colour, called the ‘focal colour’. The following figure shows how different shades of blue relate to one focal blue.
Oxford blue sky blue blue navy blue electric blue Fig. 4.2 Focal blue and other blues
Furthermore, even when there are no basic terms for a colour in a language, chips in focal colours are nonetheless easier to remember than those in other colours. Dani speakers with a two term colour system recognised the same focal colours as English speakers with an eleven term system. Three-year-old American children, whose colour system is not yet complete, preferred focal colours to the others. A central concept of ‘blue’ applies across languages; there is a universal colour ‘blue’ regardless of whether a language even a word has for it.
One explanation may be how the human eye works. Vision in the retina depends on rod cells, which are sensitive to dark versus light, and on three types of cone cells, which are sensitive to red, green and blue. The first five colour terms on the scale are then hardly surprisingly black, white, red, green, blue. Other colours are perceived by combinations of these cells. The mind processes the information from the cone cells in terms of two oppositions: red versus green, and yellow versus blue. This explains why it is possible to say greenish blue or yellowish red but not bluish yellow or greenish red. The same colour terms occur in different languages because human beings see the world in the same way, not because the outside world is structured in a particular way. Creatures whose eyes are differently constructed see different colours.
Colour words then show not only the similarities between languages in the meanings that they express because of the nature of human perception, but also the difference between languages in the meanings they choose to express, going from a minimal number of basic terms, such as Welsh, up to the maximum, as in Hebrew. This does not take into account other types of meaning that may be grafted on to a colour word: blue can be applied to a sad mood, a kind of musical note, or a risqué story.
3. Levels of meaning
The hunt has then been on to find universal meanings common to all languages. Noam Chomsky once suggested that, while concepts like ‘carburettor’ are hardly likely to be universal, concepts like ‘table’ or ‘person’ are common to all human minds. Another linguist, Anna Wierzbicka, pointed out that Polish has two words, stoä l‹ for a ‘small or low table’ such as a coffee table, and stolik for a ‘dining room table’. Poles do not have the single concept of ‘table’ which Chomsky took for granted but two separate concepts: they divide the world up in different ways. The Hebrew verb lidchot for example means both ‘postpone’, and ‘reject’, something which is hard for English speakers to grasp. Genuinely universal meanings true of all languages are hard to come up with. This section looks not so much at individual word meanings as at how the human mind thinks of the objects in the world in a particular way.
It is not the human eye that sees the world in a particular way so much as the human mind. The concept of ‘basic’ terms can be extended to other areas of vocabulary than colours. If English speakers are asked to name a bird, the odds are they will name a sparrow or a blackbird, not a great-crested grebe or a penguin. The birds that came to the minds of the Inside Language panel were, in order of frequency, sparrow, robin, crow, thrush, blackbird, and magpie (Q1). Like the colours, certain types of bird seem ‘birdier’ than others. A penguin swims; an ostrich runs; a dodo is extinct: they are not ‘proper’ birds because they differ from the prototype concept of ‘bird’ in our minds in one way or another, even if they are indeed called birds. A robin is a ‘proper’ bird; a penguin is not. a dodo ? emu? sparrow? blackbird? ostrich? penguin? robin?
The human mind has a central meaning for a concept and relates variations to this central meaning. Switching to other vocabulary areas, the words that come to mind for ‘tree’ are oak or apple rather than holm-oak or crab apple; for ‘clothing’, vest, or trousers, not belt or dinner jacket. Human beings have certain concepts they treat as archetypes, like oak; other concepts are seen in relationship to these central concepts.
The ‘prototype theory’ proposed by Eleanor Rosch develops this idea by claiming that the mind puts objects into three levels of vocabulary. The ‘basic’ level consists of those objects that strike people immediately when they look at the world around them, such as table or car. These basic level terms are the central prototypes already encountered: looking round a room we tend to see chairs, tables, lights, etc, the basic ways in which we organise our world. Above this basic level, the mind builds a second more general level at which objects are grouped together into ‘superordinate’ terms; the basic level terms tables and chairs are included within the superordinate term furniture; cars and trucks are grouped into vehicles. Superordinate terms are not directly available to our eyes; they are abstractions that place things in general categories.
As well as building upwards from the basic level, the mind also build downwards into more specific categories. At the third ‘subordinate’ level objects break up into more specialised terms; the basic level term car is now sports-car, the basic red is pillar-box red, and so on. The mind sees the world in three levels of abstraction going from the most general to the most specific, with the basic level coming in the middle as the most useful everyday term. Returning to colours, the abstract superordinate level is the word colour itself, which relates to the more concrete basic level terms blue, red and so on, which in turn relate to the subordinate terms sky-blue, pillar-box red, and so on. The superordinate category vehicles in a sense consists of basic terms cars and lorries etc, which consist of subordinate terms sports-cars, four-wheel drives, etc.
superordinate level colour
basic level red, blue, green, etc
subordinate level pillar-box red, sky-blue, etc
This three-way division in people’s mind has been confirmed to exist through several experiments. In word association tests, people usually think of basic level terms such as hammer and drum before they think of superordinate terms like tools and instruments, or subordinate terms like claw hammer or tom-tom. In reaction tests, people react to the basic level terms quicker than the others; they recognise pictures of objects called by basic level terms more easily. There are sometimes grammatical differences between the three levels of terms. In English, subordinate terms are often uncountable in grammatical terms, furniture, rather than countable, a furniture, while basic level terms are countable, a chair rather than chair. In German, basic level terms may have both masculine and feminine gender, der Hammer (‘the hammer’, masculine) and die Säge (‘the saw’, feminine); superordinate level terms, however, chiefly have neuter gender, das Werkzeug (‘the tool’, neuter).
Levels of meanings in English
The growth of advertising this century introduced consumers to subordinate levels that they had never previously dreamt of. Butter and tea were what people bought in the grocer’s, not Kerrygold or PG Tips. Only sugar remains a basic term with hardly any brandname differentiation. It is the kiss-of-death for a product that becomes a basic level term in its own right, such as Hoover, Xerox, and aspirin, since it suggests that other makes exist. Conversely, a company that tries to divide its product into subordinate terms may be asking for trouble, as manufacturers such as Coca Cola and Persil found when they tried to launch sub-varieties of product.
Though some generalisations about these levels can be made that are true for a whole language, the division into three levels varies slightly according to the individual. Rosch found that aircraft mechanics treated wing as a basic level term with many subordinates, where to the lay-person it has none. At some level everyone experiences the same world, yet the things they see depend in part on the ways they have been influenced by their society and their personal experiences. I go for a walk and see something I call trees; other more expert people may see holm-oaks, or Scottish pines.
The relationship between these levels is technically called ‘hyponymy’. The words are related through of a tree structure like those for grammar where one superordinate word includes a group of hyponyms. The test for whether two words are hyponyms is whether you can say ‘An X is a kind of Y’. So a dog is a kind of animal; a car is a type of vehicle; Glenmorangie is a brand of whisky. Hyponomy is only one of the types of meaning relationships between words. Words can also be ‘synonym’ if they have the same meaning, truthful/ honest, munch/chew, powerful/strong, though it is hard to find ‘absolute’ synonyms that always have the same meaning in all contexts That car is very powerful/ strong. Words can be antonyms if they have opposite meanings, as seen in the box below. Words are then linked to each other in networks of meaning relations.
This section has again demonstrated the diverse ways in which languages use words based on the same underlying properties of the human mind. At some level all human beings perceive things in the same way. The way that they organise vocabulary will be similar even if the words and meanings differ.
One of the most complex areas of meaning is opposites, technically called ‘antonyms’. Here are some of the categories used by D.A. Cruse.
A Complementaries: These divide up an area into two mutually
exclusive areas: A door is either shut or open
interactives: These have a stimulus response relationship: If you command someone, they obey
satisfactives: One is an attempt, the other is successful: If at first you don’t succeed, try, try again
B Antonyms: a pair of words with a gradation between two extremes He was blowing hot and cold.
Directional:. These show motion in opposite direction: He ascended the Jungfraujoch over the glacier and descended by train.
Antipodal: These indicate two extreme ends of a dimension: They searched the house from top to bottom.
Counterparts: Two objects with reverse dimensions: concave, convex
Reversives: Motion in opposite directions: rise, fall
Converses: Relationship of one direction to another: above, below
4. Language and thinking [web page]
The discussion of categorisation leads in to the much-vexed question of how language relates to thinking. On the one hand differences in the way people think might lead to differences in their language: I can’t see blue so I don’t have a word for it. On the other, differences in language might lead to differences in thinking: I have a word for ‘blue’ so I recognise it when I see it.
The Whorf-Sapir hypothesis
This section explores the area called the Whorf-Sapir hypothesis after the two researchers who first stated it clearly, Benjamin Lee Whorf and Edward Sapir. When Whorf worked as an insurance inspector, he was struck by the problems caused by language. For example a pile of drums might be classified as ‘empty gasoline drums’ by a garage and treated with no special care. Yet such drums are extremely dangerous as they contain explosive vapour. The way the workers labelled the situation was at fault, that is to say, their language. Whorf was mostly concerned with the ways that language restricted thinking in American Indian languages. Hopi has one word pahe for ‘drinking water’ and another word keyi for ‘natural water in lakes’ where English has one, as if they were two different substances. Hopi, he claimed, also has no indication of time, whether past, present or future; the mental organisation of time in its speakers has a different basis to that in what Whorf called ‘Standard Average European’. Whorf’’s most famous example is the seven words that he alleges Eskimos have for ‘snow’, compared to the single word snow in English, showing the speakers of the two languages perceive snow differently.
Considerable controversy has raged over the substance of these claims. For example Hopi speakers are perfectly capable of expressing time concepts. So far as ‘snow’ is concerned, there is a dispute on the one hand over how many words Eskimos really have for ‘snow’, on the other over whether English vocabulary is so deficient, having at least sleet, slush and hail. An interesting commentary on perception of snow is found in Peter Høeg’s novel Miss Smilla’s Feeling for Snow, whose eponymous heroine is a Greenland Inuit who is indeed the world’s authority at distinguishing types of snow.
The relationships between language and thinking have chiefly been examined through vocabulary rather than grammar. As well as the multiple Inuit words for ‘snow’, Bedouins have many words for ‘camel’, aboriginals many words for ‘kangaroo’, and, according to Peter Mühlhäusler, ‘Many New Guinea languages ... make dozens of distinctions between different types of cordilyne leaves, according to whether such leaves are used for dressmaking, decoration, magic, or other purposes’.
Languages also have particular lacks. English for example has no everyday gender-free word for children of the same parents, apart from the academic term sibling: it must be I have two brothers and a sister, not I have three Xs. German, however, has a plural noun Geschwister meaning ‘sisters, brothers, brother and sister, brothers and sisters’, as does Bahasa Melayu with suadara. Another unsurprising gap is that Quechua-speaking Indians living in the Andes have no word for ‘flat’.
If one language has more words for a specific area than another, does this mean its speakers see the world differently, or just talk about it differently? In terms of colours, Dani speakers should see a very different world divided into black and white, compared to Tiv speakers who see red, white and black, Hebrew speakers with eleven colours, and so on. Yet, as we have seen earlier, people everywhere nevertheless distinguish the same focal colours regardless of how many colour words they know. The limitations or richness of their basic colour vocabulary have not affected their perception of colour. As with the levels theory mentioned earlier, it may be a question of how ‘expert’ one becomes on a topic, whether aircraft wings, colour or snow; English-speaking skiers can doubtless recognise a range of types of snow going from powder to avalanche snow even if they do have a single word to label them by.
What about other senses? An experiment was carried out with speakers of Bahasa Melayu, one of the languages of Malaysia, which reflects Malaysian cooking in recognising several degrees of saltiness such as masin kitchup ‘salty like soy sauce’, masin ayer laut ‘salty like sea water’, masin garam ‘salty like salt’, and masin maung ‘horribly salty’. Sure enough, Malaysians were able to make much finer distinctions between solutions with different amounts of salt than English speakers. In this case their experience, whether of cooking or language, had clearly influenced their perception of taste.
Leaving vocabulary for a moment, two lines of experiments have tested whether grammar too affects thinking. People were asked to sort out objects consisting of red squares, blue triangles, etc. Speakers of languages with adjective + noun word order, as in English (black cat), would sort the objects out first by colour. Speakers of languages with noun +adjective order, as in Spanish (gato negro), would first sort them out by shape. Language influenced the way they put objects into categories.
The second line of experiments explored the grammar of conditional "if" clauses. English and many other languages have a type of conditional clause for describing things that are untrue: If pigs had wings, they would fly is one example where, even if everyone knows that pigs do not have wings, they can nevertheless speculate about what might happen if they did. Chinese, however, is said not to have such clauses; at best it can express their meaning in a long-winded fashion. The prediction is therefore that Chinese speakers would have difficulty in conceiving of such hypothetical situations. Indeed when they were given a long chain of such hypothetical statements, 98% of Americans responded with contrary-to-the-facts answers, but only 7% of Chinese. However, later experiments contradicted this finding by showing that, when one hypothetical statement was involved rather than several, bilingual Chinese scored the same as English speakers; if there is in fact any difference, it is how they deal with a chain of logical consequences, not a singe sentence. These lines of research then suggest that, while the grammar of a particular language may make certain ways of thinking more difficult, presumably because they are less usual, it is does not prevent them from happening altogether.
Metaphors of meaning
George Lakoff and Mark Johnson have claimed that a human being looks at the world through particular metaphors. A typical metaphor concerns emotions. We may be on an emotional high or feeling very low; on cloud Lucky Seven or in the depths of depression; over the moon or down in the dumps; our spirits rise or sink. In other words, happiness is metaphorically ‘up’, sadness is ‘down’. Indeed ‘up’ is the metaphor for good things in general, ‘down’ for bad things: everything’s looking up, everything’s going downhill rapidly, going up in the world, looking up to someone/looking down on someone. Or indeed high-minded versus low-spirited. Though there is no intrinsic reason why ‘up’ should be better than ‘down’, it is a universal metaphor among human beings.
Other common metaphors suggested by Lakoff and Johnson are:
‘Time means money’. You save time; you spend it; you waste it; and you use it.
‘Love is war’. You make a conquest, capture someone’s heart or surrender.
‘An argument is a building’. It hangs together, is well-constructed, and is supported on a foundation, so that it doesn’t fall apart.
‘More is better’; a big man, large-hearted, megabucks
‘Illness is a fight’; conquer disease, fight against cancer, the war against diphtheria, battle for life, attacked by malaria, gave up the fight, a new weapon in the armoury against cancer. Children who bravely fight against illness are given prizes for courage.
In some ways the debate over the relationship of language and thinking it is like the chicken and the egg; human beings think through language and they use language to express their thoughts: who can separate the dancer from the dance?
5. Components of Meaning
This section and the next describe two complementary ways of looking at the meanings of words. This section analyses their meaning into smaller ‘components’ of meaning, rather like breaking a molecule up into its separate atoms.
One example is the area of vocabulary for drinks. Anything that one drinks is normally liquid; anything that one eats is normally non-liquid; hence English speakers dispute whether soup is eaten or drunk and have to say liquid food for a liquid with the nourishing properties of food. One component of the meaning of any word for drink is ‘liquid’. So the noun drink has a component [+liquid], shown in square brackets; the noun food has a component [- liquid], meaning non-liquid.
Drinks may also be hot or cold. The meaning components are [+hot] for tea or soup, [- hot] for cola or Pimms. Again, any breach of the convention is usually remarked upon; cold soup and cold tea are unusual enough to need specifying; hot cola would be extraordinary. Adding the two components together, part of the meaning of tea is then [+liquid, +hot], of cola [+liquid, - hot].
Next drinks can be divided into alcoholic and non-alcoholic: beer is [+alcohol], lemonade [- alcohol], cola [- alcohol], whisky [+alcohol], etc. Any breach needs to be spelled out; non-alcoholic beer is a possibility even if non-alcoholic whisky is unthinkable. There has been considerable controversy over whether UK manufacturers should sell alcoholic lemonade. The choice of alcoholic or non-alcoholic is usually confined to cold drinks with the exception of Irish coffee and mulled wine: any deviation has to be mentioned, as in hot rum and hot whiskey. So a main part of the meaning of whisky is the components [+liquid] [+cold] [+alcohol].
More details can be added according to the level of expertise of the drinker. Tea may be [± China], [± Indian], or [± herbal]. Of course not everybody knows all of the components of meaning of each word. It is not every speaker of English that knows the difference between Chinese tea and Indian tea, or between an Island whisky and a Lowland whisky. But all speakers use the components to distinguish one word from another at their own level of knowledge.
Family relations [web page]
The meanings of words for family relations can also be partially captured through components. Taking English as the example, [+male] represents part of the meaning of father, son and uncle; [+female] is a component in the meaning of mother, daughter and aunt; [+previous generation] is part of mother and uncle; [+same generation] is included in the meaning of sister, brother, cousin, and so on. Components in the meaning of brother are then [+male] [+same generation], of uncle [+male] [+previous generation]. A speaker of Japanese has an additional component [±respectful].
Needless to say, such components do not sum up all the meaning of the word; there is more to the meaning of brother than ‘a male relative of the same generation’. Nevertheless a sizeable proportion of the meaning of a word is contributed by the components it shares with other words. Sometimes the difference between two words indeed comes down to a single component; fox versus cub is a difference of [± mature], just as the difference between two words can be a single phoneme, box and fox.
It is not always obvious how to describe these components best. The question of sex has been treated here neutrally as a matter of two components, [± male] and [± female] rather than as a single component of maleness [± male]. Some of the dispute over sexist terms concerns the nature of these components and the priority between them. A person who believes that the word man is always [+male] will take exception to general remarks about the human species phrased in terms of man, say The earliest man lived in Africa (particularly when a crucial skeleton happens to be that of a woman, Lucy).
Those who defend such remarks are essentially claiming that the component [+male] is not present in the meaning of the word man when it is used generally, only the component [+human]. The same with the discriminatory overtones of girl compared to woman. If woman has the component [+mature], girl the component [–mature], as no-one would dispute for man versus boy or adult versus child, an adult female should be called a woman not a girl. The derogatory implications of girl were stunningly revealed in a television discussion in which a noted feminist writer was addressed as my dear girl by the eminent linguist present, who had just been claiming that her allegations of discrimination through language were totally unjustified.
Cookery terms [YouTube]
English cookery terms have particularly noteworthy components of meaning, described by Adrienne Lehrer. What are the differences between baking and roasting, between boiling and frying, or between stewing and simmering?
Is water used or not? Boiling, simmering, and poaching imply the use of water [+liquid +water]; frying and sautéing imply the use of fat or oil [+liquid +oil]; while baking and grilling imply no liquid at all [- liquid]. So it is not possible to poach or simmer in oil, nor to boil in oil (at least for culinary purposes), even if boiling and deep-frying are the same activity apart from the liquid concerned.
How is the heat applied? In grilling, barbecuing, and toasting, the food is exposed directly to the heat, i.e. [+direct heat]. In baking the heat is indirect [- direct heat], and in boiling heat is transmitted via the liquid.
How vigorous is the cooking action? boil and reduce are [+vigorous]; simmer is [- vigorous]. So it is possible to say It was boiling violently but not It was simmering violently and, in reverse, I simmered it slowly but not I boiled it slowly.
What utensils are used to contain it? Frying takes place in a frying pan; boiling of water takes place in a kettle in England, though in much of the rest of Europe a saucepan suffices; wokking takes place in woks. The meaning of various processes may be marked [+frying pan] or [+wok].
The meaning of each cooking term is made up of a certain number of components. Woe betide the cook who does not know that simmer and stew are [- vigorous] compared to boil. As always, there are areas of doubt and confusion. Roast has two senses, in one of which it means ‘to expose to a direct flame’ (roast an ox on a spit), in the other ‘to bake in an oven’ (Sunday roast). What the British call grill, the Americans call broil, and so on. Cooking vocabulary relies on components whose differences of meaning are crucial to the tradition of cooking in English.
The same is true of other languages. A major component of meaning in Japanese cuisine is [+uncooked] for food such as sashimi. Japanese has many words for rice, including ina when it is growing, kome when it is uncooked, gohan when it is cooked and eaten as part of a Japanese meal, and raisu when it is eaten with Western dishes.
Components are then a useful way of describing the meaning of words, which can be related to children’s acquisition of vocabulary, as we see in Chapter Seven. However, they do not cover all aspects of meaning of all words. Family relations and cookery are often used as examples precisely because such limited areas of vocabulary easily show components. But the search for universal components that apply to large numbers of words has not been very successful, as seen in later chapters; nor have many vocabulary areas yet been described in this way.
Special Tasting Vocabularies
Wine-tasting: words used by wine tasters
high frequency dry
medium frequency acidic aromatic clean fruity light refreshing smooth
lower frequency astringent balanced bitter bland bouquet common delicate developed flowery fragrant full-bodied gentle lovely ordinary perfumed pungent scented soft sour sweet tangy tart young
Apples: the most frequent adjectives used by professional tasters tasting Cox's Orange Pippins (with peel) in order of frequency:
juicy crisp chewy acidic tough astringent floury sweet green sharp stringy bitter Cox-like estery fruity sugary musty alcoholic scented core-like pear-like pineapple-like phenolic spicy fatty sulphurous
adjectives (a sample taken from two issues of the
Scotch Malt Whisky Society’s Bottlings list)
assertive astringent biscuity bitter bronze buttery chewy clean complex crisp damp dry equinoctial filling fine fruity full gentle heavy intense large malty mellow medicinal mild mouldy oleaginous pale peaty peppery phenolic pungent powerful red-gold rewarding rich robust salty sharp sherried shiny simple smokey smooth sour spicy spirituous stalwart straw-coloured strong sulphurous sweet tart uncomplicated
Sources: Lehrer, 1975: Williams et al
7. Words and the Mind
This section now raises the central issue of how words are stored in the speakers’ minds and how they are integrated into the stream of speech during speaking or listening. The mental lexicon does not take the form of a straight alphabetical list of words, each with a separate meaning, like the printed dictionary. Instead the words are tied together in intricate networks; blue is related to red, father to son, boil to bake. The meanings are furthermore linked to the ways that speakers see the world, each language existing in a world of its own whether of family relations or ways of cooking.
How speakers carry out tasks with vocabulary provides some hints how it is treated in the mind. The fact that speakers can select words to put in their sentences at a rate of 150 or so per minute shows the speed with which they can access their lexical knowledge. Chapter Two showed that content and grammatical words have different access times, as shown by pauses: a grammatical word like the is found more readily than a content word like desk. The mind is organised in a way that makes function words easier to get at than content words.
Word length supplies a further clue to mental processes. Experiments have required people to remember lists of countries such as Chad, Burma, Greece, Cuba, Malta, and Czechoslovakia, Somaliland, Nicaragua, Afghanistan, Yugoslavia. The monosyllabic list is far easier than the polysyllabic list, even if it is not said aloud. On average, people can remember 4.17 out of 5 countries for the monosyllabic list, 2.80 for the polysyllabic one. The length of the word affects how it is processed in the mind, unlike a book dictionary in which word-length is more or less irrelevant.
Other clues come from the well-known feeling of having a word on the tip of the tongue. [WEB TOTT test] For some reason I have a problem with the word for the type of novel written in the eighteenth century such as Fielding’s Tom Jones whose structure derives from the travels of the hero. I know the word begins with "p", I know it is polysyllabic and that it is something like picturesque. Eventually I retrieve the word picaresque. This phenomenon is called the tip-of-the-tongue experience.
An influential experiment by Brown and McNeil induced the tip of the tongue
sensation by getting people to find words to match definitions such as:
- a navigational instrument used in measuring angular distances, especially the altitude of sun, moon, and stars at sea.
If this example did not suffice to induce a tip of the tongue experience, try
these definitions from various dictionaries; the answers are in the next box.
All the words come from the original experiment but the definitions have been
taken from different dictionaries: [try ityourself on [WEB TOTT test]
- semi-circular or many sided recess, with an arched or domed roof
- an underground conduit for drainage, a common sewer
- a fragrant drug, that melts almost like wax, commonly of a greyish or ash colour, used both as a perfume and a cordial
- a small boat with oars that is used especially in China
In a tip-of-the-tongue experience, people typically remember the first letter of the word, say "s" in the case of the navigational instrument, then how many syllables it has, say two, and perhaps that it ends in "t". They may come up with a word, say secant, that seems to fit, and they may be satisfied with this incorrect solution. Two of the Inside Language panel thought of sexton for example (Q3). Or they may go on to think of the right word itself, sextant, as did 37.5% of the Inside Language panel. Sometimes people may hit on the right word but find it still does not sound right; I am never very satisfied with picaresque but have a nagging feeling that a better word is lurking somewhere.
The storage of vocabulary in the mind therefore relies on the basic shape of a word. There is a pattern in the speakers’ minds for the word ‘s - - - - - t’, the beginning being most readily available, the end less so, and the middle hardly at all. Even in spelling only 7% of mistakes occur on the first letter of the word. Five of the Inside Language panel came up with alcove for apse, getting the first letter right. It is comparatively easy to name some words that begin in "gh", slightly more difficult to name words ending in "gh", and harder still to name words that have "gh" somewhere in the middle, as crossword puzzle addicts will confirm. The Speaker of the House’s confusion of Paddington Bear with Postman Pat mentioned in the last chapter shows the classic signs of a tip-of-the-tongue mistake. Both are characters in children’s books consisting of two words beginning with bilabial plosives /p/ and /b/. This will be discussed further in Chapter Ten. The form of the lexical entry in the mind is then beginning to emerge.
Tip of the Tongue Results
Here are the answers the Inside Language Panel (48 people) gave to the tip-of the tongue definitions (Q3). The correct answer is in bold.
Definition Answer supplied
instrument sextant (18), compass (14),
used in measuring angular telescope (20), sexton (2), protractor (2)
distances, especially the radar, pergonometer
altitude of sun, moon, and
stars at sea
alcove (15), apse (8), igloo,
sided recess, with an dome, portico, auditorium, niche,
arched or domed roof cavern, cathedral, church, cloister
(Advanced Learners) conservatory, cupola, grotto, nave
conduit drain (s) (20), pipe (5), cesspit (2),
for drainage, a common culvert (3), sewer (2), main (2),
sewer (Oxford) tunnel, University of Essex,
water table, viaduct, cloaca (0)
a fragrant drug, that
melts incense (4), opium (3), quinine (3),
almost like wax, commonly frankincense (3), peppermint (2),
of a greyish or ash colour, ambergris, scented candle,
used both as a perfume and musk, soap, resin, hashish
a cordial (Dr Johnson)
a small boat with
junk (15), sampan (11), canoe
that is used especially coracle (3), gondola (2), dhow (2),
in China (COBUILD) rowboat, dinghy, kayak
8. The structure of content words
The last chapter saw how words are made up of morphemes, for example how a word such as books consists of a content morpheme, book, and an inflectional ending for the plural morpheme ‘-s’. The present section sees how content words may be built up through the complementary process called ‘derivation’. The starting point can be a much-quoted sentence from the O.J. Simpson trial, which illustrates many of the ways in which English forms words:
We the jury in the above-entitled action find the defendant Orenthal James Simpson not guilty of the crime of murder in violation of the Penal Code section 187A, a felony, upon Nicole Brown Simpson, a human being.
There are five main methods for forming new words in human languages:
One method is to add a morpheme to the end of an existing word—a ‘suffix’. Most languages use suffixes both for grammatical inflections and for forming new words. A word such as defendant is derived from a base form defend by the addition of the suffix ‘-ant’. The same base defend can lead to different words according to the suffix that is added: defender by adding an ‘-er’ suffix, defence by adding ‘-ce’ (and omitting the "d"); defensible by adding ‘-ible’, and so on. Similarly felon leads to felony and felonious; violate to violation and violator; guilt to guilty. These new words can in due course combine with the usual grammatical inflections: defendants, violations, felonies, and so on. The different suffixes have particular meanings. The ‘-ant’ ending shows a person carrying out the action: someone who defends is a defendant, someone who assists an assistant, someone who immigrates an immigrant, etc. The ‘-tion’ ending shows that the word is an abstract noun of a general kind, violation, operation, affliction, revolution; the ‘-y’ ending shows an adjective meaning ‘full of the quality’, guilty, windy, spicy. None of these outcomes are guaranteed since many suffixes have several meanings—‘-ant’ in important hardly means a person who imports—and may also form part of the base of some nouns—city has a "y" though not an adjective. A word may have more than one suffix: organisationally for instance is built up from organise + ation + al + ly.
In English these suffixes often change the class of the word concerned from noun to verb, verb to adjective, and so on. That is to say, ‘-ant’ and ‘-tion’ change the verbs defend and violate into the nouns defendant and violation; ‘-y’ changes the nouns guilt and milk into the adjectives guilty and milky. A single base form of a word can play many roles in the grammar by using these additional morphemes. Typical suffixes for turning verbs into nouns are ‘-er’ and ‘-ation’, as in reader and information; for turning adjectives into nouns, ‘-ness’ and ‘-ity’, blackness and reality. However, these endings cannot necessarily be applied to every word: the result has to be blackness not blackity and reality not realness. Native speakers do not always agree about which new words can be derived; not everyone was happy with John Major’s term additionality.
A second method for forming new words is to add something to the beginning of an existing word—a ‘prefix’, as in entitle with ‘en-’ or asleep with ‘a-’. Many prefixes came originally from Latin, such as pro or anti, and are still active in other Latin-influenced languages as well as English. Often English prefixes seem to have more definable meanings than suffixes. The prefix ‘anti-’ and ‘pro-’ in the sense of ‘against’ and ‘in favour’ can be added to almost any noun—anti-war, Anti-Christ, anti-washing up, anti-linguistics, pro-life, pro-vegetarianism. Similarly ‘re-’ in the sense of ‘do again’ can be found added to verbs in repay, replay, retake, refer, reconstitutionalise. In many cases the prefix has become so much part of the word that the original form has ceased to exist; apart from a few witticisms, no-one uses gruntled and shevelled from disgruntled and dishevelled or kempt and couth from unkempt and uncouth.
A third method of word-derivation is to create a word out of two or more other words, a process called ‘compounding’. [Web page] The examples in the OJ sentence are above-entitled and human being, both of which are made up from other words, as are tea-time, motorway, and disc-jockey. Considerable uncertainty still surrounds whether particular compound lexical items should be written as one word, two words, or linked with a hyphen. The publishers of my last book wanted to know whether I preferred babytalk, baby talk, or baby-talk; Chapter Seven will reveal what the decision was for this book. Compounding is a common way of forming words in human languages; in Maori for example the act of buying a lolly becomes a new verb ‘to lolly-buy’ hoko rare. In German the ‘Danube steamship travel company’ is Donaudampfschiffahrtsgesellschaft.
The fourth method of deriving words is to change their word class without
altering anything else, called ‘conversion’. So head is usually a noun
but it becomes a verb in He headed the ball and She headed the
committee. Say is a verb, but in He had his say it is
converted into a noun. Up is a preposition, but in a student slogan of
the 1970s Up the grant it became a verb; in Life has its ups and
downs it is converted into a noun. Recent examples of computer jargon
converting nouns into verbs with "-ed" include He was
The fifth method of derivation used in some languages is to add or change material in the middle of the word, called ’infixes’. [web page] In English the base form of the word is seldom broken up, abso-blooming-lutely, and kanga-bloody-roo are rare exceptions. Chapter Two showed that infixes are a crucial feature in the grammar of Arabic; the base of the word is filled in with different vowels to yield different meanings. Grammatical morphemes can be shown by changing vowels in the consonant skeleton of the word; the verb drs ‘study’ has the past tense form daras ‘he studied’. But new Arabic words can also be derived in the same way: madrasa means ‘school’ and dars means ‘lesson’, both variants of drs.
Some of the these processes are effectively dead in English, some are still productive. The ‘-ard’ ending of bastard, coward, and niggard hardly leads to new words nowadays. Suffixes are often still available. ‘-ism’ can be meaningfully added to any thing or any person that has an identifiable idea attached to it, Thatcherism, linguisticism, anti-Europeism, etc, as can ‘-wise’, language-wise, situation-wise, politics-wise, etc. Each decade finds itself new productive morphemes. The 1990s seems fond of ‘mega-’, as in megabucks, Megadog, megadegeneracy, Megalab, megazine, mega-elated, of ‘Brit-’ as in Britprop and Britmusic, and of ‘-gate’ as in Watergate, Irangate, Squidgygate, to cite forms observed recently. It is hard to predict which morphemes can be used productively. Until one heard the Stevie Wonder song Yesterme, Yesteryou, Yesterday, one would hardly have anticipated that ‘yester-’ was a productive affix of English. Part of the knowledge of words is then how to build them into other words, and which prefixes and suffixes can be used.
Writers and Derivation
Here are some sample new words produced by English writers, using the standard processes of derivation, mostly for comic, witty, or poetic effect.
adding a suffix
obscurism, lessness, me-ism, Brazilification Douglas Coupland
Omnianism, exquisitor, scalbie Terry Pratchett
swimmy eyes, her eyes became a bit soup-platey P.G. Wodehouse
gorgeosity, heighth, howly, rabbiter Anthony Burgess
adding a prefix
teleparablising, ethnomagnetism Douglas Coupland
ambi-sinister, cumulo-dynamic Terry Pratchett
resuffered Dylan Thomas
ultra-violence Anthony Burgess
underdogging, café minimalism, downnesting Douglas Coupland
stoning-of-suspected-adulteresses rotas, black-on-black eyes Terry Pratchett
star-gestured, year-hedged, star-flanked Dylan Thomas
dream-dimmed eyes, that gong-tormented sea W.B. Yeats
horrowshow kick-boots, flatblock (n) Anthony Burgess
changing word class
upping with the lark, handerchiefing my upper slopes, he righthoed, he trousered the key P.G. Wodehouse
the one with the snake-pit hairdo Terry Pratchett
fellowed, do you not sister me? Dylan Thomas
we upped in the lift, I fisted him Anthony Burgess
adding an infix?
emallgration, survivulousness Douglas Coupland
To bring this chapter together, the lexical entry in the mind contains a variety of information about a word:
how it is pronounced and spelled, strongest at its beginning and end, weakest in the middle: sextant
how it is used in the sentence in terms of phrase structure etc: John fainted him
how it relates to other words through systems of meaning: grandfather, grandson
how its meaning shares features and divides up fields: blue, red; grill, barbecue
how it may be built up or decomposed into other words: defendant, yesterme.
Comparatively little of the meaning of a word can be stated as a straightforward link to a single object in the world.
The conclusion is that words do not so much have individual meanings of their own as enter into complex systems of meaning shared with other words, whether in terms of components or levels or the relationships of synonyms, antonyms, etc. While this point might seem obvious, many experiments ignore it by testing the learning of pairs of new and old words as if there were a single constant link between words and things. Language students too have to learn vocabulary lists of pairs of words matched with their translations in the second language, a very small aspect of their meaning in the light of the discussion here. Such pairings are untypical of language, where it is rare to find an identical match in meaning between two words both within one language and across different languages. The Dane Otto Jespersen provided the example of the word for ‘bat’ having slightly different overtones in each language. In French chauvesouris emphasises the bat’s baldness and its resemblance to a mouse; in Latin vespertilio suggests the evening when the bat is out, and in Danish flagermus suggests its flapping. A word contrasts with all the other words in a language rather than having a single definite meaning of its own; the meaning of words can be broken up into components, levels, and fields, as well as many other aspects, none of which give a full account of their meaning by themselves. Yet the languages of the world make use of the same resources for creating meaning, and the same ways of making new words from old, another case of similarity in diversity.
Sources and further reading
A general book on meaning is: Lyons, J. (1995), Linguistic Semantics, CUP
1. Lexical entries and the dictionary in the mind
A popular account of the mental lexicon is given in: J. Aitchison (1987), Words in the Mind, Blackwell
2. Colours and linguistic relativity
The original book is Berlin, B. and Kay, P. (1969), Basic Colour Terms: Their Universality and Evolution, Berkeley, University of California Press. A book which provides much information is: Wyler, S. (1992), Colour and Language, Gunter Narr: Tubingen. The stages of child development come from Johnson, E.G. (1987), ‘The development of colour knowledge in pre-school children’, Child Development, 48
3. Levels of meanings
The original source is Rosch, E. (1977), ‘Human categorisation’, in N. Warren (ed.), Studies in Cross-Cultural Psychology, New York, Academic Press. An account of the whole area is: Taylor, J.R. (1989) Linguistic Categorisation, Oxford Clarendon Press. The categories of opposites are based on: Cruse, D. (1986), Lexical Semantics, CUP
4. Language and thinking
Whorf’’s ideas on language and thinking can be found in the readable collection: J. Carroll (ed. (1956), Language Thought and Reality: Selected Writings of Benjamin Lee Whorf, MIT Press. New Guinea languages and much else are described in: Mühlhausler, P. (1994), ‘Babel revisited’, UNESCO Courier, February, 1994. Malaysian terms for tasting were tested in: O'Mahoney, M., and Muhiudeen, H. (1977), ‘A preliminary study of alternative taste languages using qualitative description of sodium chloride solutions: Malay versus English’, British Journal of Psychology, 68, 275-278. The experiment on word order and setting is: Hooton, A.B., and Hooton, C. (1977), ‘The influence of syntax on visual perception’, Anthropological Linguistics, 19/8. 355-357. The Chinese hypotheticals are discussed in Cromer, R. F. (1991), Language and Thought in Normal and Handicapped Children, Blackwell. The main work on metaphors and thinking is: Lakoff G. and Johnson, M (1980), Metaphors We Live By, University of Chicago Press
5. Components of Meaning
The original work on features was: Clark, E. (1971), ‘On the acquisition of "before" and "after"’, JVLVB, 10, 266 275. Her more recent ideas are in Clark, E. (1993), The Lexicon in Acquisition, CUP
Cookery vocabulary is analysed in Lehrer, A. (1969), ‘Semantic cuisine’, Journal of Linguistics, 5, 39-55; her later article applying this to wine is: Lehrer, A., (1975), ‘Talking about wine’, Language, 51, 4, 901-923
6. Words and the Mind
The tip-of the-tongue phenomenon is described in Brown, R. and McNeill, D. (1966), ‘The "tip of the tongue" phenomenon’, JVLVB, 5, 325-337
7. The structure of content words
The processes of derivation are described in books on morphology such as: Spencer, A. (1991), Morphological Theory: An Introduction to Word Structure in Generative Grammar, Blackwell, Oxford. The sources of word-formation are: D. Coupland, (1992); Generation X; P.G. Wodehouse (passim); T. Pratchett (1992), Small Gods, Victor Gollancz; Dylan Thomas (1952), Collected Poems, Dent; A. Burgess (1962), A Clockwork Orange, Heinemann