Meaning: Semantics – Semantiek

There are two major ways to look at semantics. One connects language to the real world, the other connects language to mental representations. According to the first, >referential (‘real world’) theories of meaning, we use language to describe things in the world. So when we say:

Koningin Beatrix gaat volgende week naar Den Haag.
Queen Beatrix is going to The Hague next week.

the meaning of ‘Koningin Beatrix’ is the actual person, you could touch her (if you can get close enough), and the meaning of ‘Den Haag’ is the real city, the ‘object’, in the west of the Netherlands. This seems reasonable enough. After all, ‘Den Haag’ does not mean ‘a city in the east of Japan’. However, this theory becomes problematic for words like ‘volgende week’ (next week) because there is no thing in the real world that these words refer to, yet we cannot say that these words do not mean anything. Similar problems arise with a sentence as:

Willem-Alexander, kroonprins der Nederlanden, is de man van Máxima.
Willem-Alexander, crown prince of the Netherlands, is Máxima’s husband.

The words ‘Willem-Alexander’, ‘kroonprins der Nederlanden’ and ‘de man van Máxima’ all refer to the same person, yet we can’t really say that they mean the same.

In the 16th century the ‘kroonprins’ was Philip II, but in a text about the 16th century the word ‘kroonprins’ doesn’t mean something else than in a newspaper article about Willem-Alexander (>>see Plakkaat van Verlatinge study pack). The meaning of ‘kroonprins’, then, seems independent of an actual person in the world, so it can not have the same meaning as ‘de man van Máxima’. If you have trouble understanding this, compare the following two sentences:

Willem-Alexander is de man van Máxima.
Willem-Alexander is Máxima’s husband.

Willem-Alexander is Willem-Alexander.
Willem-Alexander is Willem-Alexander.

The first sentence is informative whereas the second sentence seems meaningless. If ‘de man van Máxima’ meant the same as ‘Willem-Alexander’ the two sentences would mean the same. They do not, and so we conclude that referential theories of meaning are problematic because they do not cover all aspects of the meaning of words and sentences.

>Representational semantics form an alternative theory of meaning. They connect the meaning of a word to a >mental representation or >concept in the mind. These concepts, in turn, can be related to the real world and vice versa:

LANGUAGE MIND/THOUGHTS WORLD

So when you use the word ‘spinazie’ (spinach) this is connected to a concept in your mind, which in its turn is connected to other concepts. You know for example that ‘spinazie’ is a vegetable, it makes Popeye strong, it’s green, etc. These ‘thoughts’ you have about spinach are connected to the real object ‘spinach’ in the world. Linking words to their mental representations enables us to explain links in meaning between words; they’re not all separate unrelated things in the world, but they share properties that we are aware of. We know that ‘spinach’ and ‘carrots’ are vegetables. In this respect they are related, even though they are two separate things in the world.

Word meanings can be related in several different ways. Carrots and spinach are both vegetables. So the meaning of ‘carrot’ and ‘spinach’ is related in that they are both part of the group of vegetables (they’re sisters in the family tree of vegetables). ‘Spinach’ and ‘vegetable’ are then mother and daughter. These relations in meaning are called >sense relations.

Sense relations – word meaning

Above you have seen that there is a difference between sentence meaning and utterance meaning. However, this is not to say that you need a whole sentence to convey meaning. Words carry meaning by themselves as well, this is called >word meaning. You won’t be surprised that linguists have found a way to structure the meaning of words so that they can be linked together in several ways with the use of sense relations. In this paragraph we will discuss 5 ways in which word meanings can be related.

Before you read on, it is important to realise that linguists assume that words are stored in a sort of ‘dictionary’ in your head called the >lexicon. The lexicon stores words, affixes and irregularities of the language. Remember that regular aspects of language are predictable and therefore assumed to be the result of grammatical rules. For example, the Dutch past tense is often indicated by the suffix –te ik werk – ik werkte (I work – I worked). This past tense is the result of a grammatical rule and so the past tense of ‘werk’ does not need to be stored in the lexicon. But the past tense of lopen (to walk) is irregular ik loop – ik liep (I walk – I walked). The past tense in this case is irregular and so the past tense needs to be stored, as it is not predictable. Because our brain size is limited the assumption is that only those things that are not predictable, not the result of rules, are stored in the lexicon. Items stored in the lexicon are called >lexical entries.

Now let us take a look at 5 different kinds of sense relation:

1. >Synonymy: words that sound different but mean the same.

True *synonyms, words that mean exactly the same, are hard to find. Often when two words mean the same only one of them will be appropriate in a certain context:

Generally these two words still count as synonyms. Here’s a Dutch example:

2. >Homonymy: words that sound the same, but mean something different.

Homonyms are a lot easier to find than synonyms:

He works at an international bank.
The building is at the south bank of the Thames.

He stepped in a hole in the ground.
The whole company is curious about the results.

Notice that spelling doesn’t matter for homonymy, it’s pronunciation that counts. Dutch examples of homonyms are:

As you can see homonyms in one language don’t always translate as homonyms in another language. The word ‘bank’ has two meanings in both English and Dutch, but they are not the same two meanings. The Dutch word ‘meer’ translates into two separate English words, ‘more’ and ‘lake’. This is a clear sign that we are actually dealing with two words that somehow in the course of history have ended up sounding the same in one language.

3. >Polysemy: one word used in different situations.

The difference between homonymy and polysemy is not always clear. Polysemy is extending the meaning of a word to other uses:

Your lips are part of your mouth.
Water flows through the mouth of the river.

She dropped the bowling ball on my foot.
They waited for him at the foot of the mountain.

And in Dutch:

So where homonymy deals with two different words that happen to sound the same, polysemy is about one word that can be used in different situations. You can probably think of the similarities between a human mouth, a mouth of a river and a mouth of a bottle, although it is quite hard to list exactly what properties they share. With polysemes it is possible to generalise certain properties of a word so it can be used in different circumstances.

The assumption is that homonyms are listed as separate lexical entries, whereas a polyseme is one single lexical entry (whose properties can be generalised to different uses). In other words: homonyms are two separate words, but a polyseme is a word with different meanings when used in a different context.

4. >Antonymy: words with opposite meanings.

Antonyms are word pairs that contrast in meaning, such as ‘big-small’ and ‘warm-cold’. There are two groups of antonyms. The first consists of >gradable antonyms (bigger than…) which have relative meaning. For example, a big mouse is still a lot smaller than a small elephant, so the meaning of a word like ‘big’ is not constant but relative to the context in which it is used. Here are some examples of that first group:

It’s cold on the North Pole.
It is very cold for this time of year.

Our sun is a young star.
He is the oldest member of his tribe.

It should be clear that a young star is still a lot older than an old man, so the meaning of these words is flexible in that way. Here are some Dutch antonyms:

>Non-gradable antonyms cannot usually be used in comparative constructions. For example, someone cannot be more born or more dead than someone else. Another difference between gradable and non-gradable antonyms is that the negative of one of the members of a pair of gradable antonyms does not necessarily imply the other member. That is to say, a man who is not big is not necessarily small. However, a man who is not dead is definitely alive.

5. >Hyponymy: the meaning of one word is included in the meaning of another word.

Hyponymy is a very important part of structural relations in word meaning. It reflects the hierarchy of relations between word meanings and shows how we can understand that two words or concepts that are very different on the surface are somehow related in meaning. For example a snake is very different from a dog (or more to the point here: the concept for snake is very different from that for dog), yet we know both are animals and so the concepts share a part of their meaning. Again we can use a tree structure to show the hierarchal relation of hyponymy (Diagram 1):

A cottage is a type of house, a house is a type of building (and a cottage is a type of building too, of course). We say that cottage is a >hyponym of house, and house is a >hypernym of cottage (this is what we called a mother-daughter relation above). In order for us to understand that ‘cottage’ and ‘house’ are related in meaning part of the meaning of ‘cottage’ must be included in the meaning of ‘house’, and vice versa. Hyponymy is a relation of >inclusion.

Sentence meaning

We know that words have meaning, and if you put words together in a sentence then of course the sentence has meaning too. This sentence meaning is context-independent. You can almost say the sentence doesn’t really need ‘the real world’ to mean something. Taken in isolation from context the sentence...

He gave her that nice thing.

...means something, even if you don’t know who he or she is and what exactly the nice thing is. The words in this sentence carry information: there is a 3rd person singular male that gives an object with good qualities to a 3rd person singular female. The sentence is almost a >code for this information. All you need to do is to >decode the sentence to get to the meaning of the sentence. In short the context independent sentence meaning can be found by a process of >coding and decoding.

Sentence meaning may be context-independent (as opposed to utterance meaning, which we will talk about below), but that does not imply that this meaning is always clear. Sentences (and utterances) can be >ambiguous, i.e. they can have several meanings. We have already seen a form of ambiguity in word meaning. Homonyms are ambiguous. The word ‘bank’ can either mean ‘a river bank’ or ‘a financial institution’. Similarly, a sentence can mean either one thing or another. There are two types of ambiguity. >Lexical ambiguity is mainly due to using homonyms, a famous example is:

The painting was found by the tree.

Here the word ‘by’ causes the confusion. Of course the intended meaning is ‘in the proximity of’ but the sentence also allows the reading that the tree found the painting. Newspaper headlines are a famous source of ambiguity, here are a few examples for your amusement:

Iraqi head seeks arms
Kids Make Nutritious Snacks
Miners Refuse to Work after Death
Astronaut Takes Blame for Gas in Spacecraft

Another type of ambiguity is >structural ambiguity:

Complaints About Referees Growing Ugly
are the complaints or the referees ugly?

Peter decided to read this page on Monday.
did he decide or read on Monday?

John likes to paint his models nude.
who’s nude?

Visiting relatives can be boring.
are the relatives visiting you or are you visiting them?

The ambiguity in these sentences is caused by the fact that a phrase can be interpreted in different ways. For example, ‘visiting relatives’ can be a verb phrase, but it can also be a noun phrase (the subject of the sentence).

If we ignore the possible ambiguities sentence meaning is very clear. In the sentence:

Hij vind haar leuk - He likes her

There is a 3rd person singular male that is fond of a 3rd person singular female. Getting to the meaning of the sentence simply requires decoding of the words. However, as noted above, the meaning is different in different situations. If you see Willem-Alexander with Máxima, and you say the sentence above, ‘he’ means W.A. and ‘she’ means Máxima. If you say it when you see the milkman and the woman that lives at number 5, ‘he’ and ‘she’ mean the milkman and the woman at number 5.

Words like ‘he’ and ‘she’ (and ‘you’, ‘it’, ‘here’, ‘there’) are called >indexicals. They do have a sort of set meaning, like ‘3rd person singular’ or ‘the area in my immediate surroundings’, but they are always used to >refer to something that can only be deduced from the context.

With indexicals the semantic meaning, which is context independent, falls short of >pragmatic meaning, which does depend on context. In other words ‘3rd person singular male’ is a lot less specific than ‘Willem-Alexander’. Semantic meaning often underdetermines pragmatic meaning. The examples we saw in the previous chapter about saying one thing, but meaning another (‘It’s cold here’ meaning ‘close the window please’) are another example. So for language use the semantic meaning is not enough, we need context dependent pragmatic meaning. This opens up a whole new field of meaning to study and we will talk about it in the next chapter about >pragmatics.

Conclusion

We have seen that there is a lot to say about words and sentences and their meaning. There are different ways to think about meaning and the main ways presented in this chapter seem more open to debate than the rules we introduced in the other chapters about syntax and phonology. This is partly because theories about meaning rely a bit more on intuitions and philosophy, and partly because the other areas of linguistics are presented in a more black and white style because they are partial to testable rules. Or perhaps it is because those areas have been studied more extensively because studying meaning can get very complicated.

However, we were able to see that yet again it is possible to analyse meaning in a way that reveals structure and a form of predictability that we have seen in the other fields as well. It is possible to look into sentences and words and say something meaningful about their meaning. However, as soon as sentences stop being dead objects and become utterances there is a whole new field of meaning to be studied. This is what >pragmatics is about and we will talk about in the next chapter.

next
(pragmatics)