W ORD C ONVERSION AND L EXICAL C ATEGORIES
Wylie Breckenridge
Draft Only
0. Introduction
According to traditional grammar, ‘cash’ in (1a) below and ‘cash’ in (1b) are distinct words from distinct lexical categories with distinct (but related) meanings:
(1) a. I needed some cash. b. I cashed a cheque.
‘Cash’ in (1a) is a noun that means cash, a kind of currency. ‘Cash’ in (1b) is a verb that means cash, a kind of event (in which something is exchanged for cash). The verb ‘cash’ and its meaning were derived from the noun ‘cash’ and its meaning via a process called word conversion .
1 2
Why do ‘cash’ in (1a) and ‘cash’ in (1b) belong to distinct lexical categories (noun and verb, respectively)? According to traditional grammar, it is because of what they mean – ‘cash’ in
(1a) stands for a person, place or thing, whereas ‘cash’ in (1b) is a ‘doing’ word – it stands for a kind of event or action. Words are categorised according to their semantic content. An adjective stands for a property or attribute. According to the way that contemporary linguists view the matter, it is because of their morphosyntactic behaviour – ‘cash’ in (1b), for example, can be inflected for tense (‘cashed’) and number agreement (‘cashes’), whereas ‘cash’ in (1a) cannot.
On either view, ‘cash’ in (1a) is a noun and ‘cash’ in (1b) is a verb. Only verbs take the suffixes
‘-ing’ (progressive), ‘-ed’ (past tense, participle), ‘-s’ (third person singular subject present tense), and ‘-en’ (past participle). They may be modified by adverbs but not by adjectives or demonstratives.
The two differ, because meaning does not reliably predict the functioning of a word. Common nouns normally take an article, proper nouns do not. Common nouns can be modified by determiners and adjectives, while proper nouns cannot (‘A tall girl’, ‘A tall Dana’). Common
(count) nouns can take plural inflection, while proper nouns cannot. Adverbs modify verbs, adjectives, adverbs. Nouns occur as the subject or direct object of a verb, as the object of a preposition, and with ‘be’ as a predicate nominal (‘My lawyer is Max’).
Word conversion is a common process in English. The underlined words in (2a) below are all verbs that have been converted from nouns, and those in (2b) are all nouns that have been converted from verbs. Other examples of each kind are easy to think of.
(2) a. i. Let’s bus the children home. ii. I’ll phone you later. iii. John bottled some beer.
1 The process is also called category change , functional shift , functional change , or zero derivation . It is distinct from the process of derivation , in which a new word is formed by the addition of a (non-null) suffix to an existing word (as in the creation of the adjective ‘lucky’ from the noun ‘luck’).
2 Why think that the verb ‘cash’ in (1b) has been derived from the noun ‘cash’ in (1a), rather than, say, the other way around? One reason: natural explanation of what it is to cash - to convert it into cash.
Natural explanation of what cash is - paper money. Note: not the result of cashing something - the outcome of a cashing action. Compare this with a noun-verb pair in which the noun is, intuitively, derived from the verb: ‘judge’ and ‘judge’. To judge is to decide on a verdict (roughly). A judge is someone who judges - the agent of a judging action. The idea here is that the denotation of one word in the pair is to be explained in terms of the denotation of the other, in which case the other is the one with the primary meaning.
- 1 -
b. i. Mary went for a walk. ii. The judge was very lenient. iii. We had some lovely experiences today.
Conversion is not restricted to just verbs and nouns, as the examples in (3) below show. The noun ‘red’ in (3a) has been converted from the adjective; the verb ‘up’ in (3b) has been converted from the preposition; the adverb ‘tall’ in (3c) has been converted from the adjective; and the verb ‘chink’ in (3d) has been converted from the onomatopoeic word.
(3) a. I have two reds in my hand. b. John tried to up the pace. c. Mary walked tall. d. They chinked their glasses.
There are even cases of conversion within the one lexical category. The count noun ‘diesel’ in
(4b) below, for example, has been converted from the mass noun ‘diesel’ in (4a):
(4) a. My car takes diesel. b. I own two diesels.
My aim in this paper is to argue against this traditional picture. I agree that we use ‘cash’ in (1b) to mean cash, the event. But I argue that when we do we also use it to mean cash, the currency – the same thing that we use ‘cash’ in (1a) to mean. I suggest that, given this, we should take
‘cash’ in (1b) to be the same word as ‘cash’ in (1a). The conversion of ‘cash’ in (1a) to ‘cash’ in
(1b) was not a process whereby a new word was created with a new meaning, but a process whereby an already existing word acquired a new meaning. I propose that the same is true of every instance of word conversion: word conversion is not the creation of a new word, but the acquisition of a new meaning. I do not deny that there is such a thing as word conversion, but I do deny that it is a process whereby new words are created.
I finish by considering some consequences for the way we lexically categorise words. I propose that it is not words that we should be categorising as nouns, verbs, adjective, and so on, but the structural positions within a sentence in which we use them.
1. ‘Cash’ in (1b)
I think that we do indeed use ‘cash’ in (1b) to mean cash, the event. But I think that at the same time we use it to mean cash, the currency – the same thing that we use ‘cash’ to mean in (1a). If we do use ‘cash’ in (1b) to mean both cash, the event, and cash, the currency, then we use it to mean two distinct things with the one use. This might initially sound implausible, so I start with some general plausibility arguments. Then I state five facts about our use of ‘cash’ in (1b), and argue that they are best explained on the assumption that we use ‘cash’ in (1b) to mean cash, the currency.
Plausibility
It is a general fact about usage that we can use things to do two distinct things with the one use.
I can use a button to explode a bomb and thereby explode a dam, and if I were to do so I would use the button both to explode two distinct things with the one use (the bomb and the dam). I can use a match to light some paper and thereby light a bonfire, and if I were to do so I would use the match to light two distinct things with the one use (the paper and the bonfire). I can use a noise to frighten one person and thereby frighten the whole crowd, and if I were to do so I would use the noise to frighten two distinct things with the one use (the person and the crowd). I don’t see any relevant difference between using buttons to explode things, using matches to light things, using noises to frighten things, and using words to mean things. If there is no
- 2 -
relevant difference, then we ought not be surprised to find that we can use words to mean two distinct things with the one use.
It is widely recognised that we sometimes mean two distinct things by the one use of a sentence .
A speaker might use the sentence in (*) below to mean both that Mary has nice handwriting and that Mary is not a good philosopher:
(*) Mary has nice handwriting.
The sentence only means one of these two things (that Mary has nice handwriting), and because of this the speaker only says one of these two things (that Mary has nice handwriting). But the speaker means both that Mary has nice handwriting and (thereby) that Mary is not a good philosopher.
It is not as well recognised, but I think it is also true, that we sometimes (perhaps often) mean two distinct things by the one use of a sub-sentential expression. It would not be unusual for a speaker to use the sentence in (*) below to mean that the man on table six left without paying:
(*) Table six left without paying.
Were she to do so, by ‘table six’ she would mean both table six and the man on table six. Again, the expression ‘table six’ in (*) only means one of these two things (table six), and because of this the speaker only says one of these two things (table six) in using it. But she means both.
It might be that a speaker can only use the expression ‘table six’ to mean table six, and if there is anything by which she means the man on table six in uttering (*) then it is a more complex expression that contains ‘table six’ and other aphonic material (that is, unpronounced but semantically significant material). I have three things to say against this possibility. First, there is not, as far as I know, any good semantic reason to think that if a speaker means the man on table six by any part of (*) then it must contain aphonic material. Second, nor is there, as far as I know, any syntactic reason to think so either (good or bad). Third, no one thinks that a speaker can use ‘Mary has nice handwriting’ to mean that Mary is not a good philosopher only if it contains appropriate aphonic material. Since I don’t see any relevant difference between using the sentence (*) to mean that Mary is not a good philosopher, and using the sub-sentential expression ‘table six’ to mean the man on table six, I don’t see any reason to think that ‘table six’ must contain aphonic material if a speaker can mean the man on table six by it. All this being so, I think we should take seriously the idea that that a speaker might mean two distinct things by the one use of ‘table six’ in (*).
I think it is quite plausible, then, that we mean two distinct things by ‘cash’ in (1b). Next, I present five pieces of evidence that by ‘cash’ in (1b) we mean cash, the currency.
Understanding
Even if we know that an interpreter does not know that we use ‘cash’ in (1b) to mean cash, the event, if she knows that we use ‘cash’ in (1a) to mean cash, the currency, then we can reasonably expect her to understand that we use ‘cash’ in (1b) to mean some kind of event that involves cash, the currency.
In contrast to this, if we know that an interpreter does not know that we use ‘eat’ in (*) below to mean eat, a kind of event, then we cannot reasonably expect her to understand that we use it to mean some kind of event that involves food, even if we know that she knows that we use ‘food’ in (*a) to mean food.
(*) a. I need some food.
- 3 -
b. I will eat a hamburger.
Why the difference? Because we use ‘cash’ in (1b) to mean what we use ‘cash’ in (1a) to mean, but we do not use ‘eat’ in (*b) to mean what we use ‘food’ in (*a) to mean.
Response: we do not use ‘cash’ in (1b) to mean cash, the currency, but there is a word ‘cash’ that we do use to mean cash, the currency.
Suppose we introduce a new word, ‘bloff’, and use it to mean cash, the event. If we know that an interpreter does not know that we use ‘bloff’ in (*) below to mean cash, the event, then we can not reasonably expect her to understand that we use it to mean some kind of event that involves cash, the currency. That is because we do not use ‘bloff’ in (*) to mean cash, the currency.
(*) I bloffed a cheque.
For suppose that I do not use ‘cash’ in (1b) to mean cash, the currency, but only use it to mean cash, the event. If my interpreter does not know that ‘cash’ in (1b) is used to mean cash, the event, and if that is all I mean by ‘cash’ in (1b), then my interpreter would no less understand what we mean by ‘cash’ in (1b) if we were to use a completely different word instead:
It might be helpful to consider a less familiar case. A friend of mine, Shannon, recently uttered the following sentence while looking down at her jeans:
(*) I’ve misbelted myself.
I did not know at the time that ‘belt’ in (*) is used to mean a certain kind of event. In fact, it was false at the time that ‘belt’ in (*) is used to mean a certain kind of event (we can assume that this was the first time ‘belt’ had been used like this). Nevertheless, Shannon expected me to understand that by ‘belt’ she meant something about a belt (which I did). This only makes sense if she used ‘belt’ to mean belt.
Ellipsis
One possible interpretation of what a speaker means by the second sentence in (*) below, is that she needed some cash, the currency:
(*) I cashed a cheque. I needed some.
This is because, I propose, cash, the currency, is made salient by the speaker’s use of the first sentence, because in that sentence she uses ‘cash’ to mean cash, the currency. It follows that by
‘cash’ in the first sentence she means cash, the currency.
One might respond to this as follows. Yes, it is a possible interpretation of what the speaker means by the second sentence that she needed some cash, the currency. But that is not because she means cash, the currency, by ‘cash’ in the first sentence. Rather, it is because by ‘cash’ in the first sentence she means a certain kind of event, and that kind of event involves cash.
Suppose that is right. Since eating events involve food (at least a vast majority of them), it ought to be that one possible interpretation of what a speaker means by the second sentence in (*) below is that she needed some food.
(*) I ate a sandwich. I needed some.
- 4 -
That is indeed a possible interpretation, but only if food is made salient by other features of the context, perhaps by preceding parts of an extended conversation as follows:
(*) A: What food did you eat? Why did you eat food?
B. I ate a sandwich. I needed some.
This is unlike the case of (*) above, however. In (*) above, there is no need for cash, the currency, to be made salient by other features of the context – it is made salient by the speaker’s use of ‘cash’ in the first sentence. This can only be so if she uses ‘cash’ in the first sentence to mean cash, the currency.
One might respond to this, in turn. Let’s grant that cash, the currency, is made salient by the speaker’s utterance of ‘cash’ in the first sentence. That need not be because she uses ‘cash’ in that sentence to mean cash, the currency – it may simply be because there is some word ‘cash’ that means cash, the currency.
Suppose that is right. Since there is a word ‘hard’ that means hard, as in difficult, it ought to be that one possible interpretation of what a speaker means by the second sentence in (*) below is that the questions were hard, as in difficult.
(*) The chair was hard. The questions were too.
That is indeed a possible interpretation, but only if ‘hard’ is zeugmatically reinterpreted in the first sentence. This is because ‘hard’ in the first sentence is not used to mean hard, as in difficult.
This is unlike the case of (*) above. In (*) above, there is no need for ‘cash’ in the first sentence to be reinterpreted in a zeugmatic manner. This is because it is used and interpreted the first time as meaning cash, the currency. Admittedly, there is something a bit awkward about it, but that might be due to the fact that ‘cash’ is being used in verb position – it is not an instance of zeugma.
Response: not obviously zeugmatic, because not as great a difference between cashing events and cash, as there is between hardness of chairs and hardness of questions.
I conclude that the best explanation of why a possible interpretation of what a speaker means by
…
Anaphora
There is similar evidence from anaphora. One possible interpretation of what a speaker means by ‘that stuff’ in the second sentence in (*) below is cash, the currency – that which she means by ‘cash’ in the first sentence. It follows that by ‘cash’ in the first sentence she means cash, the currency.
(*) I cashed a cheque. I love that stuff.
Again, one might respond that yes, on one possible interpretation, what the speaker means by
‘that stuff’ is cash, the currency, but she is not using it anaphorically upon ‘cash’ in the first sentence. Rather, she is using it to refer to something that is involved in events of the kind she means by ‘cash’ in the first sentence.
But if that is right, then it ought to be that one possible interpretation of what a speaker means by ‘that stuff’ in the second sentence in (*) below is that she means food:
- 5 -
(*) I ate a sandwich. I love that stuff.
Again, this is a possible interpretation, but only if food is made salient by other features of the context:
(*) A: What food did you eat? Why did you eat food?
B. I ate a sandwich. I love that stuff.
This is unlike the case of (*) above. In (*) above, there is no need for cash, the currency, to be made salient by other features of the context – it is made salient by the speaker’s use of ‘cash’ in the first sentence. This can only be so if she uses ‘cash’ in the first sentence to mean cash, the currency.
Again, one might respond to this, in turn. Let’s grant that cash, the currency, is made salient by the speaker’s utterance of ‘cash’ in the first sentence. That need not be because she uses ‘cash’ in that sentence to mean cash, the currency – it may simply be because there is a word ‘cash’ that is used to mean cash, the currency.
Suppose that is right. Since there is a word ‘hard’ that means hard, as in difficult, it ought to be that one possible interpretation of what a speaker means by the second sentence in (*) below is that the questions were hard, as in difficult.
(*) The chair was hard. The questions were that way too.
I think that is indeed a possible interpretation, but it requires a zeugmatic reinterpretation of
‘hard’ in the first sentence. This is because ‘hard’ in the first sentence is not used to mean hard, as in difficult.
This is unlike the case of (*) above. In (*) above, there is no need for ‘cash’ in the first sentence to be reinterpreted in a zeugmatic manner. This is because it is used and interpreted the first time as meaning cash, the currency. Admittedly, there is something a bit awkward about it, but that might be due to the fact that ‘cash’ is being used in verb position – it is not zeugma.
Modifying ‘cash’
It is possible to interpret ‘counterfeit’ in (*) below as making more specific what kind of cash, the currency, is meant by ‘cash’. that what a speaker means by (*) is that she exchanged a cheque for counterfeit cash. This is only possible if ‘counterfeit’ is interpreted as modifying the currency meaning of ‘cash’.
(*) I counterfeit cashed a cheque.
For suppose that it is interpreted as modifying the action meaning of ‘cash’. Then it ought to be possible to interpret that what a speaker means by (*) below is that she
(*) I tasty ate a sandwich.
(*) The chair was A+ hard.
When interpreted in this way, ‘counterfeit’ is not understood as modifying the event that is meant by ‘cash’. If it were, then I ought to be able to use the sentence in (*c) just as well. But I can’t. The most salient interpretation of this sentence is that I broke a cheque while being wrinkly.
(*) a. I counterfeit banked a cheque.
- 6 -
Rather, it is understood as modifying the kind denoted by ‘cash’ in its ordinary use. On the special-meaning theory, someone who understands (*a) would have to interpret ‘cash’ with two meanings on the one use - one on which it denotes a kind action, the other on which it denotes a kind of currency. But then it ought to be zeugmatic.
Being more specific
A speaker might follow her utterance of the first sentence in (*) below with an utterance of the second, in order to specify what kind of cash, the currency, she means by ‘cash’ in the first sentence. It follows that by ‘cash’ in the first sentence she means cash, the currency.
(*) I cashed a cheque. Counterfeit cash, unfortunately.
Response: what she is clarifying is what kind of cash was involved in the event she meant by
‘cash’. But:
(*) I ate a sandwich. Tasty food, in fact.
Ok, so maybe she was specifying what kind of cash she might have meant by ‘cash’ in the first sentence. But:
(*) The chair was hard. As hard as the questions, in fact.
2. Only one word
I agree that by ‘cash’ in (1b) we mean cash, the event. But I have just argued that we also use it to mean cash, the currency – the same thing that we mean by ‘cash’ in (1a).
If this is right then it is difficult to maintain, I think, that ‘cash’ in (1b) is a distinct word from
‘cash’ in (1a). I think we should question this traditional claim, and take seriously the idea that
‘cash’ in (1b) is the identically same word as ‘cash’ in (1a).
TW: does this mean that (1a) changed its meaning because ‘cash’ did when the verb use of
‘cash’ started?
If ‘cash’ in (1a) and ‘cash’ in (1b) are the same word, then what should we say about word conversion, if it is not a matter of one word being converted into a new distinct word? I suggest the following: rather than a new word being created, an existing word acquires a new meaning.
‘Word conversion’, then, is bad name for this phenomenon.
Does this mean that (1a) also acquires a new meaning? I think not.
But not like ambiguity. When it is used with the new meaning it also used with an existing meaning – used with both meanings. Analogy with a tool: starting gun.
I conjecture that all cases of word conversion are like this. Describe a case: ‘misbelt’.
3. Lexical categories
What is a noun? According to both traditional grammar and contemporary linguistics, it is a type of word. According to the former, it is a word that has a certain kind of meaning; according to the latter, it is a word that exhibits a certain kind of morphosyntactic behaviour.
- 7 -
So is ‘cash’ a noun or a verb? It exhibits both the morphosyntactic of nouns and verbs, so seems should say that it’s both. Could say it is neither. Could say it is one and not the other.
But I suggest that even this is the wrong way to describe the situation. I suggest this: the word
‘cash’ is neither a noun or a verb. So what should we say? Might try this: sometimes used as a noun , sometimes used as a verb . But that assumes that some words are nouns, verbs, etc.
Better: it can be used in noun position or used in verb position . I suggest that to be a noun or a verb is not to be a certain kind of word; rather, it is to be a certain structural position within a sentence. Should not talk of words being nouns, verbs, adjectives, etc. Should talk of noun positions, verb positions, etc., which words may be used in. Part of what is involved in understanding a sentence is determining the syntactic position in which a word occurs – it depends upon the sentence that is being used, not the word. That does not affect its meaning, but it affects the interpretation of what the speaker means. We do not need to know, when interpreting a sentence, whether or not a word is a noun or a verb, etc. What we do need to know is whether it is in noun position or verb position, etc. By ‘the noun in this sentence’ we mean: the word in noun position. By ‘every sentence must have a verb’ we mean: every sentence must have a word in noun position.
The kind of thing that a word denotes affects how effectively it can be used in different positions. The word ‘eat’ means a kind of event (let us say). So, if it is used in verb position it places a tighter constraint on interpretation than does the word ‘belt’, which stands for a kind of object rather than a kind of event. So the kind of thing that a word denotes will have an effect on the syntactic position that we typically use it in. Thus the intuitive idea that to be a noun is to stand for an object, to be a verb is to stand for an event, to be an adjective is to stand for a property, and so on. If we use words in verb position to mean a kind of event, then we would expect to typically use in verb positions words that mean events (e.g. ‘run’). This explains the intuition that the lexical category of a word is determined by its meaning.
The linguistic view is not surprising if nouns are syntactic positions. If the property of being a verb is a property of structural positions in a sentence, then it is not surprising that linguists classify words as verbs just in case they appear in certain structural positions. Because our interpretive abilities are so good, we do not need to. We can use ‘belt’ in verb position to mean a kind of event, even though ‘belt’ does not mean that kind of event. This explains why word conversion is so common (at least in English).
The one word can be converted into many categories:
(*) a. John wants to up the volume. b. We’ve had many ups this year. c. The switch is up. d. Mary jumped up.
References
Bach, ‘Conversational Impliciture’.
Recanati, ‘Unarticulated Constituents’.
Gamut semantics textbook, vol. 2, ‘type raising’.
- 8 -