English grammar - Roger Berry 2012
Roger Berry
D3 Determiners: a class apart
Section D Extension
Roger Berry (1998) reprinted from English Today, 14/1, pp. 27-34.
This article is based on a book I wrote in 1997: Collins COBUILD English Guides 10: Determiners and Quantifiers. It describes the problems encountered during the writing process, including the main one: are determiners a distinct word class? It provides some more evidence on topics which are mentioned elsewhere, such as ’some’ vs ’any’ (A1) and the use of ’less’ with count nouns (C3).
The article has been shortened quite significantly.
[. . .]
Troubles
Determiners are a very problematic group of words in a number of different ways. I have divided these various troubles into three levels: for linguists, for applied linguists, and for learners/teachers. One might characterise these as being analysis, (pedagogic) description, and application respectively. The borderline between them is arbitrary to some extent, and a problem may apply to more than one of them. However, generally they represent stages in what may be called ’the process of pedagogic grammar’.
Troubles for linguists
Are determiners a valid word-class?
The first issue is whether determiners are distinguishable as a class from e.g. pronouns (with which they were conflated in earlier treatments of English). According to Hudson (1990) the answer is no: there is a generalisation that will be missed if we follow the normal practice and treat determiners and pronouns as two unrelated word-classes’ (1990:269). This generalisation is that most words that are determiners can also function as pronouns, for example that:
I don’t like that idea much.
That was many years ago.
The exceptions to this are a/an, the, no, every, plus my, your, her, our and their, which cannot be used as pronouns. Of these the possessives can easily be excluded since they have readily available pronoun counterparts (mine, etc). Likewise, no has a counterpart in none. Every is more problematic, unless one accepts that every one (not everyone, of course, which can only refer to humans) is a pronominal counterpart. This would then only leave the articles; these can be accounted for by saying that they are the basic determiners in that they add nothing to the meaning of a noun phrase apart from the basic idea of definiteness or indefiniteness.
After the above prestidigitation, the case for establishing determiners as a distinct class would seem to be rather thin. Similarly, in terms of meaning, there are few reasons for a distinction. In certain cases the use of a word as a pronoun implies human reference, as in the examples below, whereas there would be no such restriction of interpretation for a determiner.
He returned home a global figure, an almost godlike figure for some.
Those determined to kill can always find suitable opportunities.
In addition, there is often a degree of formality associated with certain pronominal uses, as the two examples show, which is absent from the determiner. (This is discussed further below.) But in the vast majority of cases the meaning of the determiner and the pronoun are the same.
What, then, is the reason for all the attention given to determiners? The answer would seem to lie in an increasing trend (in this structuralist-oriented century) to rely on syntagmatic criteria for establishing word-classes.
Membership problems
Even if one does accept the existence of determiners as a distinct class of words, there are several problems with delineating its membership. Beyond what may be called
the ’prototypical’ determiners alluded to above, there are cases of a more borderline nature. I will only mention a few briefly below:
□ open classes, e.g. cardinal and ordinal numbers and multipliers (twice, three times, etc.). Semantically, all that distinguishes them from the quantifiers is that they refer to specific amounts. However, they come later in the noun phrase, after postdeterminers (see below)
□ words like same, next, latter etc., which semantically would seem to be more like adjectives but which syntactically have something in common with postdeterminers (see below) in that they can precede numerals. There is an (unwritten) assumption in the analysis of determiners that they cannot be discontinuous; thus, if numerals are determiners, so is anything that can precede them.
They lived together for the next three years.
□ the Saxon genitive, which structurally and referentially does the same job as possessive determiners.
Mother’s cooking was horrible.
Zero determiner
One particular problem of membership concerns the so-called ’zero article’, which is said to ’occur’ in positions when neither definite or indefinite article nor other determiner is present. The concept of ’the presence of nothing’ as opposed to ’the absence of anything’ is a common one in linguistics, and in this case it allows linguists to say that all nouns have a determiner (if they do not have one, how can they be said to refer to anything?).
Most scholarly grammarians have advocated the acceptance of this as a third article (strictly speaking, it should be called ’zero determiner’). However, there is some disagreement about where it operates. For Christophersen (1939), for example, it operates in all situations where a noun phrase lacks a determiner. Quirk et al. (1985:246), on the other hand, make a distinction between the zero article, as in:
I like music.
and no article with proper nouns, which can be said to incorporate their own determiner:
I like Sid.
Another way of looking at proper nouns is to say that they do not need determiners to refer.
Ordering problems
A noun phrase can have more than one determiner, up to an (intuitive) maximum of four: ’all the many such possibilities’ (although this is not attested in the Bank of English). The usual way of explaining the ordering of these combinations is in terms of three positions:
□ predeterminers, which include all, both, half, and multipliers, such, many, exclamative what, rather and quite.
□ central determiners, which include the most common determiners, such as articles (the and a/an), demonstratives (this, etc.) and possessives (my, etc.). Words such as each, which cannot be combined, are usually included here.
□ postdeterminers, whose membership is somewhat uncertain (see above) but it includes little, few, many, several and numbers.
Of course, many combinations are not possible. Thus, half can occur before both the and a, while all cannot precede a and such cannot precede the.
Half the building was in flames.
I ordered half a pint of lager.
All the ironing is done.
Mother made such a fuss about it
Some of these restrictions can be explained away easily on logical grounds, but others cannot.
Beyond this, there are a number of exceptions to this three-position approach to determiner ordering, because some words can have different positions.
□ Every can come before few but also after possessives, which makes it both a central determiner and a postdeterminer; there is, however, a difference in meaning:
Every few days there seemed to be another setback.
Television cameras would be monitoring his every step.
□ Such can come after many (a postdeterminer) but can be a predeterminer before a.
Is this the last of many such occasions?
Mother made such a fuss about it.
□ Many can be a predeterminer before a and a postdeterminer after definite determiners.
Many a successful store has paid its rent cheerfully.
None of her many lovers seemed to want to marry her.
Many a and such a are probably best accounted for as single units, in the same way that a few and a little generally are (because they do not behave as combinations; see Huddleston 1984: 234-5).
Definite and indefinite determiners
This distinction is somewhat problematic. First, these two groups are really rather different in referential terms. While the definite determiners (typically the, possessives and demonstratives) can clearly be said to ’determine’ the noun group by referring to something familiar, established or identifiable in the listener/reader’s consciousness, the indefinite determiners (typically a/an and the so-called ’quantifiers’) only do so in a vague way, as Ducrot and Todorov point out (1981:253). This explains why Jespersen (1933) calls the indefinite article the ’article of incomplete determination’.
Second, while the distinction between definiteness and indefiniteness can help to understand the difference in meaning between certain pairs of words, such as the and a/an, and which and what, the borderline is not always clear. Both, for example is usually grouped with indefinite determiners, but it undoubtedly has an idea of definiteness; it refers to something in the text or the environment that is familiar, or established in the mind of the reader or listener. That is why both and both the mean the same thing.
Troubles for describers
This section looks at problems for the applied linguist/pedagogic grammarian when trying to interpret this analysis according to a particular purpose (most typically language teaching). There are some features that are not particularly problematic (at least for description), e.g. the association with type of noun (countable or not) which explains the (basic) difference between many and much, and between (a) few and (a) little, but others are.
The variety of patterns
One initial problem with some determiners is that they can occur in a number of patterns. To take all as an example, the following patterns are possible:
’all boys’, ’all the boys’, ’all of the boys’.
The structural aspect is not too hard to describe, but what about the meaning? The first is clearly different in meaning from the others (due to the absence of the), but is there any difference between the last two? The reader is invited to struggle with this one, as many people have before.
In addition, three determiners, all, each and both, can undergo the phenomenon known as ’quantifier float’ or ’placement’ where they occur after the noun they determine, typically in the so-called middle position, but also elsewhere. Here they have something in common with adverbs, as Giusti (1990) notes:
Wed all like to make easy money.
Liver and eggs are both good sources of natural iron.
The sergeants each carried one.
And although quantifier float is possible with nouns that follow the verb, there are restrictions. For example, each cannot come after a noun used as a direct object (you cannot say I saw them each) but can after an indirect object.
He handed them each a cup of tea.
Both and all are possible after direct objects (I saw them all, I saw them both).
Again, while the structural aspect can be described (if not fully explained), what of the meaning? Is there an extra emphasis to this use, as its label in the Collins COBUILD English Dictionary (1995) - ’emphatic pronoun’ - would seem to suggest?
Assertion
This concept, that is, whether you are asserting the existence of a thing or not (see Quirk et al. 1985:83-84), is involved with a number of determiners. The case of some and any is well-known now in the literature (e.g. Lakoff 1969; see also Lewis 1986:33-37 for a clear account of why some and any are not suppletive forms). These two examples show how they can be in opposition:
I haven’t found some of the money.
I haven’t found any of the money.
A far less tractable case is that of much and many. As Quirk et al. point out (1985:384): ’Much and, to a lesser extent; many have acquired some non-assertive force’. The problem is: how much? Much would seem to be rare in assertive circumstances in modern English unless it is being used with an intensifier (so much, etc), or is rather formal (’There is much confusion.’)
Many is more debatable; again the reader is asked to ponder how formal many sounds in the examples below and whether there is a difference:
We have concluded many agreements.
We haven’t concluded many agreements.
Coverage
One issue for the describer is how far to go in covering different uses of the same word. Given the similarity between the pronoun and determiner uses of these words pointed out above, it would seem to make sense to group them. However, many can also be adverbs, modifying verbs, adjectives or other adverbs. Generally their meaning is similar, although some are informal.
My back feels all achy.
Lili opened her eyes a little wider.
If you feel confident you will be less anxious.
Landlords say they will not wait any longer.
I can’t believe he was that good an actor.
With some of these, such as more, less, the adverb use is in fact the most common. In some cases, the adverbial meaning is not derivable from the pronoun/determiner.
In 1850 some fifty thousand women worked in such places.
In addition, there are copious idioms in which determiners play a prominent part, where their individual meaning is not apparent.
It scarcely mentions women at all.
The decision on what to include will depend on the underlying rationale of the work, not to mention the amount of space available.
Formality of some determiner and pronominal uses
Two problems involving formality were alluded to above: that of certain pronoun uses and that of much and many in assertive contexts. There are others. In particular, few and little are regarded as formal, whether as pronouns or determiners:
There is little hope.
Few countries have accepted this claim.
’Not much hope’ and ’not many countries’ would be non-formal alternatives here.
The problem with formality is, first, in identifying it - what is formal for one person may sound normal for another - and, second, in expressing it; its effects on use are so hard to quantify, how can advice be given? Beyond this, there is the issue of whether we have an adequate pedagogic metalanguage to describe it. It is not enough to talk only of formal vs informal, as my use of ’non-formal’ above indicates.
Acceptability
There are three issues of acceptability I would like to raise. The first is a well-known one: the use of less with plural count nouns (instead of fewer). This use is widely attested and has apparently been a feature of English for a long time:
I did expect more food and less people.
Against this we have claims that this usage is ungrammatical or incorrect. Even more frequent is the use of less followed by than:
It affects less than 70 children a year.
When writing the book I was urged to indicate no restriction for less. I felt this would have been wrong; the writer has a position of responsibility towards the learner and to suggest both forms are interchangeable could be seen to be misleading. Also, it is my impression that most learners prefer a conservative approach to acceptability. Thus I chose the following formulation (Berry 1997a: 128):
’In informal English it is quite common to use “less” with count nouns in the plural. For some people, though, this is not acceptable.’
I am aware that this formulation will not suit everyone, but other writers have taken the same approach. Swan (1995) has a similar hedge.
Second, there has been an equally long dispute about which personal pronoun to recommend to refer back to an indefinite noun phrase (determined by e.g. each, every or any) where the gender is in doubt. Here I had no qualms in recommending they or their since the alternatives are awkward or problematic and since this usage is of some antiquity.
Each individual person thinks their case is justified.
The third issue is the position of the floating quantifier. Most pedagogic accounts give this as incorrect:
They all are just interested in making money.
However, analysis showed this to be fairly common, especially in spoken and American English, so it is at least worth noting.
[. . .]
Questions, suggestions and issues to consider
1. In what way might scientific and pedagogic accounts differ with respect to the notion of the ’zero article’ (or ’zero determiner’)?
2. Do you agree with the claim that determiners are not a separate class? Or do you feel that the ’prestidigitation’ involved (making inconvenient exceptions disappear) is not valid?
3. What position should teachers adopt on the above-mentioned issues of acceptability? Namely:
□ less with count nouns
□ the use of their to refer back to indefinite noun phrases
□ floating quantifiers placed in front of auxiliaries (they all are interested rather than they are all interested)
Should they be descriptive or prescriptive on these matters (see A1, Approaches to Grammar)? Note that the use of their to refer back to indefinite noun phrases reinforces what was said in B2 about the similar use of they.