Grammar for education
was: Why education needs grammar
[all graphics removed to reduce file size]
Richard Hudson 2020
In a nutshell (2.2K)
This book[footnoteRef:1] is a defence of grammar teaching,
arguing that grammar should be one of the pillars of education,
alongside maths, science, literature, history and the rest. But I’m
a grammarian by trade, so I would say that, wouldn’t I? I love
grammar, so I’d like everyone to have the same amount of fun that
I’ve had. Fair enough, you may think, but not a strong argument in
itself; after all, every academic would like more followers. So I
need to persuade you. [1: I would like to thank Neil Sheldon for
statistics, Shahan Choudhery for pedagogy, John Walmsley for
history, …]
First, though, what do I mean by grammar? What I think should be
taught in schools is grammatical analysis – the ability to think
about the structure of any word or sentence, to see the patterns
and the intricate network of connections, to see them in any
sentence you write or read, to be able to look at a draft sentence
and spot the ambiguities, the potential challenges for a reader,
and the alternatives which could make it better.
For example, take this sentence from a few lines back:
(1) What I think should be taught in schools is grammatical
analysis.
I’d like you to be able to recognise that this is a cleft
sentence with the same meaning as (2).
(2) I think grammatical analysis should be taught in
schools.
And I’d like you to understand why I chose to use (1) rather
than (2): because it focuses attention on the words grammatical
analysis. In other words, I would like you to understand the
anatomy of this sentence – its structure – and how this relates to
its physiology – its function as an act of communication. I’d also
like you to know useful things about the words in this sentence:
for instance, that is and be are forms of the same verb (which
grammarians call BE), and that their different forms reflect the
way they are being used. In short, I would like you to be able to
apply the same kind of analytical thinking to your language that a
medical expert applies to your body.
Now I’ll try to persuade you that grammar in this sense –
understood as grammatical analysis – provides an essential
foundation for education. Here goes: Education is part process,
part content and the process is part socialisation, part
communication; so education is part socialisation, and part
communication of content. Now comes the critical bit: the content
of education is complicated, so the communication also needs to
convey complicated messages. Teachers and pupils need to be able to
talk about complicated ideas such as dinosaurs and spelling, novels
and photosynthesis. However hard a teacher may try, the
complexities just won’t go away, so pupils need to be able to
understand the teacher’s complex messages and to produce similarly
complex messages of their own.
The problem is that this level of complexity goes beyond what
children meet outside school, so they have to learn an extended
language – ‘academic English’. Many children pick up academic
English on their own – but many don’t, and need help. For example,
imagine a lesson where the teacher uses the word creativity. Some
children may know the word already, but what about those who don’t?
Some will be able to analyse it grammatically into creative + ity,
thereby linking it to words such as curiosity and maybe even
electricity, so they’ll work out that it’s a noun with something to
do with being creative. But others won’t be able to make this leap.
For them, its shape gives no clues to either its meaning or its
word class. These are the children who stand most to benefit from
learning grammatical analysis – not the academic superstars, but
the strugglers.
But that’s not all. Learning grammatical analysis opens up a
whole new world of possibilities, in much the same way as learning
arithmetic does. Not only does academic English suddenly become
transparent, but so does the everyday English that we all take so
much for granted. We notice that we form questions by moving the
first verb back before its subject, as in (3) and (4).
(3) Jo is her friend.
(4) Is Jo her friend?
But although this works fine with is, it fails completely with
most verbs, as witness (5) and (6).
(5) Jo invited her friend.
(6) *Invited Jo her friend?
(The * means that the sentence is ungrammatical.) Here, every
native speaker would agree that (6) is utterly impossible – bad
English. Instead, we have to add did (and remove the d from
invited).
(7) Did Jo invite her friend?
This example takes us into the heart of English grammar, where
we find a particularly neat combination of rules and patterns, with
a handful of interesting loose ends – something that any non-native
learner of English has to understand, but which few native speakers
are aware of. Surely this is at least as worthy of exploration as
anything else in the curriculum? What better place for a child to
recognise the rule-governed nature of their language than in their
very own ordinary speech?
The idea of teaching grammar in relation to the child’s own
language may sound obvious and even uncontentious, but it is
actually revolutionary. The fact is that the child’s own language
is likely to be a non-standard variety; so in England, it will be
some kind of Non-standard English involving forms such as we was
and them books. But we all believe (I guess) that one of the
school’s aims is, and should be, to teach Standard English; so how
does the study of Non-standard English mesh with that aim? Very
well, I would argue, because it separates two challenges that face
a child. The first challenge is to learn grammatical analysis – how
to look analytically at language. For that analysis, Non-standard
English is just as good as Standard English, even though the forms
concerned may be different. The second challenge is to add Standard
English as a second dialect alongside Non-standard English. The
issue here is the child’s emotional attachment to the local version
of Non-standard English. If you’ve grown up in a working-class area
of (say) Birmingham, that’s the identity that matters to you and
that you want to reflect in your speech; so studying the grammar of
Birmingham English helps you to value this identity, and when you
eventually learn about the grammar of Standard English, the
comparison will be a comparison of equals. In contrast, if
grammatical analsis is applied from the start only to Standard
English, the message for the child is obvious: Standard English has
grammar, but Birmingham English doesn’t. This puts Standard English
and Birmingham English into conflict, when they can easily coexist
in easy harmony.
My argument so far has brought us to the point where a child
knows something about their own internal grammar as well as about
the grammar of academic English. Not only do they have the
important skill of metalinguistic awareness (the ability to think
about language), but they also have a metalanguage (a language for
talking about language), plus awareness of rules constraining their
own speech. What a good start for learning a foreign language! When
presented with the French for ‘hello’ and ‘good evening’, bonjour
and bonsoir, they can immediately isolate bon and guess what it
means; and as their knowledge grows and becomes more complicated,
they can tame the complexities by talking about them.
And best of all, they can think about how different the foreign
language is from their own English, and marvel at it. Isn’t it
interesting that whereas English restricts the verbs that can be
moved in front of the subjects, French restricts the subjects that
allow this movement? So French forms questions by moving any verb,
such as mangeait ‘was eating’, but only if the subject is a pronoun
(elle, ‘she’) as in (9) but not in (11).
(8) Elle mangeait une glace. ‘She was eating an icecream.’
(9) Mangeait-elle une glace? ‘Was she eating an icecream?’
(10) Marie mangeait une glace. ‘Mary was eating an
ice-cream.’
(11) *Mangeait Marie une glace? ‘Was Mary eating an
ice-cream?’
Knowing something about grammar, and being able to think and
talk about it, turns the learning of a foreign language from mere
rote learning into an exciting journey of exploration.
So far, then, I’ve offered three reasons for teaching
grammatical analysis: as a support for learning academic English,
as a microscope for studying the child’s ordinary vernacular, and
as a tool for learning a foreign language. I have two more
arguments, to do with Standard English and with thinking.
The argument based on Standard English used to be the only
justification for teaching grammar, because it was assumed that
other kinds of English had no grammar. On this assumption, it was
easy to see that if children were to learn Standard English – e.g.
to say we were instead of we was, or those books instead of them
books – they needed a stiff dose of grammar. By the end of this
book, I hope you will at least be convinced that every kind of
English has its grammatical rules, and that most of the rules of
Standard English are shared with every other kind of English – as
indeed we have already seen to be the case for one such rule, the
one for forming questions as in (4) and (7). Nevertheless, there’s
still a place for grammatical analysis in the teaching of Standard
English to native speakers of non-Standard varieties. After all,
the differences are quite few in number, so they can easily be
dissected and displayed. For instance, Standard English uses those
as the plural of that, whereas the local dialect uses them (giving
that book, with two different plurals: those books or them books).
If children are used to exploring their own speech, such details
should be easy to learn and digest emotionally.
And finally we have the argument from thinking. We started with
the argument from communication: grammar allows us to communicate
complex ideas, so we should understand how it works. But before we
communicate complex ideas, we have to entertain these ideas and
make sure that they’re not just complex, but sensible – and maybe
even true. This of course is one of the main goals of education.
Children enter primary school thinking like infants, unable (or
unwilling) to see through the myth of Father Christmas, and leave
secondary school thinking like adults about adult issues, from
politics to climate change and careers. One of the key thinking
skills to be learned is analysis – recognising the full complexity
of an issue, including all its uncertainties and inner conflicts,
but being able to work through it to some kind of conclusion (such
as which way to vote in a referendum or election).
This is precisely the kind of thinking that’s encouraged and
developed in grammatical analysis. Grammar offers complexity in
spades, with a diversity of units (words, phrases, morphemes,
meanings) in a diversity of relations (with names such as ‘means’,
‘signals’, ‘is the subject’, ‘is a member’, ‘is a part’ and
‘follows’). It also offers uncertainties (e.g. is out of a single
preposition or a sequence of two words? And how about into?) and
inner conflicts (e.g. if is working is a form of the verb to work,
how come it looks as though it contains two separate verbs?).
Better still, school grammar is underpinned by a lively research
scene where these uncertainties and conflicts are debated, and
where resolutions are available.
It’s true that other subjects offer training in analytical
thinking, but grammar is very special because all the data needed
for an exercise in thinking is freely available in the children’s
own heads. They’re already experts in using grammar, and already
know a vast amount of fine detail (though they probably aren’t
aware of any of this knowledge); and this treasure-house of
information is vastly richer than any comparable body of
information that most of them already know, and probably richer
than anything else they will ever learn during their lifetime. This
puts grammatical analysis into a very special category on its own
as a training ground for sophisticated thinking – ahead of sciences
and mathematics as well as the humanities subjects.
Better still, all these patterns of grammar aren’t neutral and
‘out there’, like objects from the natural world, but (as already
mentioned) they’re the patterns that we use for communicating with.
It’s a short (but tricky) jump from communication to thinking, so
the patterns of grammar are closely related to patterns of
thinking. For example, English grammar forces us to decide on the
epistemological status of a thought: is it a fact, a mere
conjecture, a rumour, an impression, or what? The distinction
emerges clearly in examples like (12) to (16).
(12) They missed the bus.
(13) They must have missed the bus.
(14) They may have missed the bus.
(15) They seem to have missed the bus.
(16) It sounds as though they’ve missed the bus.
As an exercise in thinking, it would be hard to beat the
analysis of these grammatical contrasts.
We can now summarise the arguments for teaching grammatical
analysis:
· It helps children to learn academic language.
· It teaches them about the rules behind their own ordinary
spoken language.
· It makes foreign languages easier to learn and more
interesting.
· It helps them to learn the standard language.
· It develops important thinking skills – the ability to analyse
complex patterns and to think in sophisticated ways.
These arguments will be developed somewhat in the following
chapters, but most of the book will be concerned with
practicalities. What is the state of play in grammar teaching? (I
try to combine details about the UK with broad generalisations
about other countries.) How did we get to where we are now? What
kind of grammatical analysis should be taught? How should it be
taught? Who should teach it, and who should be taught? When should
the teaching happen? And finally, how should we get from where we
are now to where we (or at least I) would like us to be?
And just in case it’s not already obvious, I want to be
absolutely clear about what I am NOT proposing.
· I am not arguing for a return to a golden age of grammar
teaching; at least in the UK, grammatical analysis has never been
taught in the way I am proposing, though there are many positive
lessons to be learned from practice in the past and in other
countries.
· Even more emphatically, I am not suggesting that grammar
should be taught as a list of ‘common errors’ to be avoided. If
that’s what you want, you’ll find plenty of free websites offering
lists of the most important errors – but don’t be surprised when
you compare the lists and find that every one is different!
· Nor, perhaps more surprisingly, am I arguing that grammar
should be taught in order to improve writing. There is good
evidence that analytic grammar, well taught, can improve writing,
but that’s only one reason for teaching it. The benefits I offer go
well beyond writing.
Grammar
What do we know about grammar? This is a question for the
experts – people like me who spend their working lives studying
grammar. (There are more of us than you might think – certainly
thousands if you count all the grammarians in different countries.)
As I explain in chapter 0, it’s a very old research field, with
roots in ancient Babylon, and although we’re a long way from an
agreed final analysis, there’s a lot of agreement on the basics
which I lay out below, so you can take this as a fairly reliable
summary of what we all tell our undergraduate students.
Why ‘anatomy’ and ‘physiology’? I mentioned this analogy
earlier, but in case it wasn’t clear I should expand it a little.
In medicine, anatomy is the study of body parts and structures, so
if you were a novice learning about the human body, you would learn
about all the bones – their physical properties such as size, shape
and density, and how they fit together to make up a skeleton. In
contrast, physiology is the study of what these body parts do and
how they work together, so in relation to bones you would learn how
they move and what they do – which are used in walking, which in
lifting, and so on. This is a good analogy for grammar, because
grammar could be described as the anatomy and physiology of words,
studying their internal structures (anatomy) as well as how we use
them (physiology) to complete grammatical patterns and to convey
meanings. In grammatical terms, grammar is the study of both
structure (anatomy) and function (physiology); and as in medicine,
there is no clear boundary between the two approaches; in fact,
another useful analogy is a coin, whose two sides are different but
inseparable. Some people enjoy distinguishing ‘formal’ grammar from
‘functional’ grammar, but in fact we all pay attention to both
sides of the grammatical coin. The same is almost inevitable in
teaching about grammar, as will become very clear in the rest of
the book.
0. Texts and system
Grammarians distinguish sharply between texts – bits of language
in use, whether written or spoken – and the language system that
lies behind them. Texts can be of any length, from words (or even
word-parts) to a whole book, so every word in the previous sentence
is a text, as well as the entire sentence (and any other parts of
it that we may choose to identify). Spoken texts are purposeful
actions that have a speaker, an addressee (the person spoken to), a
place, a time and a purpose; they happen, and then they’re gone.
The same spoken text never happens twice, and until recently it was
gone for ever unless someone remembered it and repeated it. Written
texts are a bit more complex, because they too are the result of an
action by a writer, taking place at a particular time, but when we
think of written texts we think of the permanent marks on the
paper; so a written text is the permanent product of an impermanent
process. Thanks to tape recorders, the same is now true of spoken
texts, so an audio or video recording is also a permanent product
(the recording) of an impermanent process (the talk). Texts, then,
are essentially actions taking place in time, even if they have a
permanent product.
The language system that lies behind a text is quite different.
This is a body of knowledge that each speaker carries permanently
in their head, and which speakers learn from each other. When you
wake up each morning, your language system is still there, more or
less unchanged, even though the texts you produced the day before
have vanished into the mists of time. During the day your language
system may undergo tiny changes such as learning a new word or
spotting a new connection. The system is a vast collection of
patterns of many different kinds (which we explore below), from
tiny facts about the pronunciation of one word to broad
generalisations about how words can combine with one another. But
the system is very busy whenever you speak or listen, write or
read, because it is the system that allows you to make sense of
other people’s texts, and to produce comprehensible texts of your
own.
Like form and function, texts and the system are two sides of
the same coin, so although they’re very different, you can’t
separate them. If you want to analyse a text, you have to apply
what you know about the system; and if you want to analyse the
system, you do so by studying texts. This may sound dangerously
circular, but it’s what grammarians have been doing for thousands
of years, and it works.
Texts
Texts are much more concrete than the underlying system, so
they’re always the natural starting point for any teaching.
Sometimes you look at individual words; for instance, what do we
know about the grammar of the one-word text round? We might
consider this word out of context, noting its grammatical
versatility; or we might consider it as part of a larger text such
as the sentence He looked round the corner. In both cases, the word
round is a text – something you say, and maybe write on the board.
At other times, though, the text will be longer, and may be very
much longer. For instance, you could look at the use of round, or
of passive verbs, in a poem or a newspaper article. All these are
also texts, and they’re all material for grammatical analysis.
We can apply the analogy of anatomy and physiology to texts of
any length, so let’s start with a really short text: a single word,
the word texts at the start of the previous paragraph. Notice how I
use italics to distinguish a text that I’m quoting from the rest of
the text that I’m writing. This is standard practice among
grammarians, and a very useful convention for avoiding confusion
between form and meaning. For instance, the word texts consists of
five letters, but this isn’t generally true of texts
themselves.
Even a one-word text illustrates the complexities of the
form-function contrast mentioned earlier. A simple account
recognises just two things: a meaning and a pronunciation. In this
analysis, the word has a physiology, its function of conveying the
meaning, and an anatomy, in the form of its pronunciation. But it’s
actually more complicated than that because there’s no direct
connection between meaning and pronunciation; instead, this
relation is split into a number of shorter steps.
A fairly standard analysis recognises a number of levels of
analysis on an imaginary scale from most concrete at the bottom of
the scale to most abstract at the top. Starting at the top, here
are the levels that most grammarians recognise, all applied to the
word Texts:
· meaning: Texts means something like ‘a set of items each of
which is a text’, where the idea of a text awaits further analysis
in terms of language, action, authors, readers and so on. The units
of analysis at this level are concepts such as ‘set’ and
‘text’.
· syntax: the word Texts is the syntactic subject of are. Here
the units are words such as texts (or more helpfully, ‘TEXT:
plural’, i.e. the plural of TEXT).
· morphology: the word Texts consists of two parts, which by
convention we enclose in curly braces: {text} and {s}. The units of
morphology are morphemes, such as the stem {text} and the suffix
{s}.
· phonology: This is a written text, so strictly speaking Texts
has no phonology; but thanks to our language system we know that it
would be pronounced /tɛksts/. Note the slants enclosing
pronunciations, and the use of symbols from the International
Phonetic Alphabet. The units of phonology are phonemes, classified
as vowels and consonants, as well as syllables and possibly larger
units for intonation.
· graphology: Texts is written as five characters: . The diamond
brackets are standard notation for written units. Notice that this
analysis distinguishes between upper-case and lower-case , a
distinction that has absolutely no counterpart in phonology.
Writing and pronunciation are different and need different
analyses. The units of graphology are graphemes (characters or
groups of characters), written words (separated by word spaces) and
punctuation marks.
This gives a four-level analysis in which the bottom level is
split into two: one for speech and the other for writing. The
analysis is starting to get complex, so you may find a diagram such
as Figure 1 helpful.
[graphic]
Figure 1: A four-level analysis of a text
If you like conceptual tidiness you may be wondering how grammar
fits into this analysis. I hope you’ll agree that it’s just a
matter of semantics, of what we want the word grammar to mean. If
we want to keep in step with our scholarly tradition, then we
certainly have to include both syntax and morphology; but these two
levels link upwards to semantics and downwards to phonology and
graphology, so it really is just a matter of taste as to whether we
call these neighbouring levels ‘grammar’ as well. Personally, I
don’t think it matters much, so long as we remember that the
neighbouring levels matter. If syntax and morphology are two sides
of the same coin, then so are syntax and semantics, and morphology
and phonology or graphology. At which point, of course, the analogy
of the coin collapses.
You may also wonder about the familiar distinction between
grammars and dictionaries, or between grammar and vocabulary. This
is an important issue which will receive a whole section (2.4) to
itself, but without spoiling the story too much I can reveal that
by the end of the discussion this distinction will be dead.
The language system
If a text can be analysed on four levels, what about the
language system that makes the text possible? Not surprisingly,
this has the same four levels, but it’s a bit more complicated
because these levels are interrelated and it’s the grammar that
decides how they can be related. For example, the syntactic
analysis treats Texts as a plural noun, while the semantic analysis
shows that its meaning is a set (a group with a number of members);
but these two facts are obviously interconnected, and it’s the
grammar that defines this relation by requiring a plural noun to
refer to a set.
The language system therefore defines two different kinds of
pattern:
· intralevel patterns, showing how units on the same level can
combine.
· interlevel patterns, showing how units on different levels can
combine.
The grammar tells us that a noun such as Texts can occur as the
subject of a verb such as are; this is an intralevel pattern
because the noun and the verb exist on the same level (syntax). But
the grammar also tells us that a plural noun such as Texts has a
set as its meaning – an interlevel pattern because the noun and the
set belong to different worlds: the world of syntax, and the world
of meanings.
Putting these two kind of information together, and multiplying
each of them by several thousand for each of four levels, the
result is one of the most complex structures in the human mind, and
possibly even in the universe. I’m not trying to impress or
frighten, but this is the background for this book about teaching
grammar at school. The good news is that school teachers have been
teaching grammar pretty successfully for thousands of years, so it
must be possible to boil the intellectual complexity down into a
teacher- and pupil-friendly form.
We can show the architecture of a language system by adding
interlevel links to the diagram for texts, giving Figure 2. Most of
these interlevels have well-established names – morphophonology for
the link between morphology and phonology, morphosyntax for that
between morphology and syntax, and semantics for that between
syntax and meaning. But there’s no established name for the link
between morphology and graphology, so I’ve suggested
‘morphographology’, with scare-quotes. The lack of a name for this
interlevel is hardly surprising, because hardly any linguists work
in this area – a great shame because English spelling is heavily
dependent on morphology.
For example, you may have noticed that both semantics (the study
of meaning) and syntax end in /ks/, but are spelt differently. Why?
Because they have different morphology. Semantics is based on the
adjective semantic (just as phonetics is based on phonetic), so the
and belong to different morphemes; but syntax isn’t matched by
syntac, and the shows, correctly, that there’s no morpheme boundary
between the /k/ and the /s/. Exactly the same is true of pairs like
tax (one morpheme) and tacks (two morphemes). Teachers and learners
deserve a thorough, systematic and accessible guide to the
morphographology of English; but as things stand currently, they’re
still waiting.
[graphic]
Figure 2: The organisation of a language
Language analysis therefore needs four levels and four
interlevels. Most of this is merely stating the obvious – that an
analysis of meaning is quite different from an analysis of
pronunciation, even though both are part of language. If you ask
what a fox is, you don’t want to hear that it’s a monosyllable, any
more than you would want information about animal classification in
asking how to pronounce the word fox. This is obvious. What is less
obvious is what a word is, because words, the units of both syntax
and morphology, combine meaning with a form which is just one step
from pure phonology. It’s all too easy to confuse texts and texts –
the word and its meaning. I once read an undergraduate essay which
reported that the fox, like many other common animals, is a
monosyllable. Foxes have tails and four legs, but the word fox has
three letters and a pronunciation; so it’s easy to imagine how this
example helps my argument that a stiff dose of grammatical analysis
is good for thinking.
Grammar and vocabulary
You won’t see vocabulary in either of my pictures of language,
so I need to explain how it fits in. We’ve all grown up in a world
in which publishers offer grammars and dictionaries: grammars for
the general stuff that applies to a lot of words, dictionaries for
the vocabulary, the individual words to which the grammar applies.
The grammar tells us about nouns and verbs in general, while the
dictionary lists individual nouns and verbs, thousands of them each
with its own ‘lemma’ (a little paragraph). Surely this is an
important distinction that a linguist ought to respect?
Well, no. It is true that, in any language, there are some facts
that are very general. For example, in English it is true that if
an adjective modifies a noun (as big modifies books in big books),
then the adjective almost always comes first. And conversely, there
are plenty of words which link meanings and pronunciations in a
completely arbitrary way, not predictable from any kind of general
rule – think of fox and text, for example. Faced with quick fox, it
makes total sense to ask why quick comes before fox, and the answer
consists in a very general rule; but it makes no sense to ask for a
general rule which explains why fox means ‘fox’.
But, and this is the crucial point, there is no natural boundary
between the general and the particular. Rules and facts can range
in generality from the very general to the very particular. Think
of English past-tense verbs. There is a very general rule: to form
the past tense of a verb, you take the verb and add {ed}, as in
{walk}{ed}, {work}{ed} and so on. (Remember: {…} for morphemes.)
But we also have several hundred irregular verbs like sink, think
and go. Where should we put them – in the grammar or in the
dictionary? As a matter of fact, they’re handled in both places:
both in the grammar (in lists of exceptions to the general rule)
and also in the lemma for the individual verb. Moreover, there are
useful sub-regularities; for instance, consider the following
verbs:
begin, swim, ring, sing, drink, shrink, sink, stink
What do these verbs have in common? Nothing in terms of meaning,
but they all have followed by or (both nasal consonants), and in
their past tense the changes to . Where does this factlet belong?
Grammar or dictionary? Since there is no good answer to this
question, many of us have come to the conclusion that it’s not
worth asking. This conclusion is somewhat controversial among
grammarians, but it makes such good sense that I’m now offering it
to you.
In short, there is no distinction between grammar and
dictionary, so the grammar – the syntax and morphology – includes
all the minutiae of the dictionary as well as the very general
rules of traditional grammar. There is no need to change the map of
language in Figure 1 and Figure 2, because they already cover all
of vocabulary. But of course publishers will continue to market
grammars and dictionaries as separate products, so we just need to
remember that the grammar (in the new and much broader sense) is
defined jointly by the publisher’s grammars and dictionaries.
This conclusion is important for teachers. For one thing, it
means that you can take a bottom-up approach in teaching grammar,
starting with individual words and their syntactic and
morphological properties. For instance you could cover a great deal
of the grammar of English by looking at the verb take. Its
morphology (take, takes, took, taken, taking) reflects the main
morphological distinctions of English, many of which are
fundamental to syntax; for instance, takes never occurs without a
subject noun (as in She takes …), whereas taken and taken can
easily be used without a subject (as in Once taken, this medicine
acts fast). Then there are the syntactic and semantic properties of
all forms of take; for instance, it can occur either with or
without an object noun:
(1) She took her tablets.
(2) She took to her bed.
And when it has an object, it may also have an object
complement, as in the rather academic (3):
(3) I take that to be a matter of opinion.
As for meaning, each of these syntactic options has a distinct
meaning, which is worth exploring even in the absence of a proper
technical language for doing so. All of these factlets about take
are shared with other verbs, so they invite classroom projects on
building lists of similar verbs. What other verbs have a past tense
where turns into , or have different meanings when they have an
object or no object, or allow an object plus an object
complement?
And for another thing, merging grammar and vocabulary helps with
the PR. Everyone understands, and accepts, that children’s
vocabulary increases, and that schools play an important part in
this growth; but grammar is different. Many people think of grammar
as a static, non-growing, body of rules; and indeed some linguists
have even suggested that the child’s grammar is to all intents and
purposes complete by the age of five. One of the themes of this
book is that schools can and should help children’s grammar to grow
as well, so if grammar and vocabulary are actually one and the same
thing, this argument is much easier to make.
Grammar and meaning
The reason why grammar is so important to humanity is that it
allows us to express complex meanings, so semantics (the link
between grammar and meaning) is at the very centre of the study of
language and of grammar in particular.
… grammar should never be taught divorced from the meaning that
the sentence pattens convey. To teach grammar without reference to
meaning is that strategy that has given grammar a bad name. Anyone
who has ever felt that grammar is boring, dull, pointless and
irrelevant has almost certainly been taught it in this way.
(Crystal 2017a: 118–9)
The complexity comes from the general rules for modifying
ordinary word meanings.
For instance, suppose you had a vocabulary of (say) 20,000 words
(which, incidentally, is a reasonable guestimate for many people).
If each word expresses one meaning, that gives you 20,000 meanings
you could express. But add on morphology, and the vocabulary
suddenly grows. Every verb has four or five forms even in English
(which is relatively poor in this respect compared to many other
familiar languages), and every noun has a singular and a plural;
moreover, new words can be created, as driver is created from drive
– a new noun created out of a verb. Each of these differences gives
a different word, differing either in meaning or in syntax from all
the existing words.
Then there’s syntax, which enriches the range of possible
meanings even more. If you know just two words, you have two
meanings; but if you allow them to combine with each other, one
modifying the meaning of the other, you get a third meaning. And if
you have ten adjectives (pretty, big, ugly, …) and ten nouns (girl,
book, idea, …), they combine to give 100 new meanings. Some of the
meanings may not be very useful, such as pretty book, but at least
they are available if the need should arise. Syntax is the
powerhouse of meaning creation, because there’s literally no limit
to its creativity. There’s no such thing as the longest possible
sentence, so you can always add an extra word or two to create a
slightly different meaning.
All this is fundamental to teaching, where meanings are central.
Teaching can be imagined as a system for transferring information
from the teacher’s head into the pupils’ heads; but how does this
magical operation work? One way is to do it directly, by
transmissive teaching: the teacher first converts an idea into a
meaning, expressed in words; and (all being well) the pupils
recognise these words, identify the meaning intended and convert it
into the same idea in their own heads. Of course the operation
rarely goes as smoothly as that, but at least that’s the aim of
transmissive teaching; and transmissive teaching is surely part of
every teacher’s repertoire, even if it sits alongside a host of
less direct methods. And it all depends on the infinitely subtle
meanings that grammar allows us to create. Take away grammar, and
the whole of education collapses.
You may wonder how we can represent meanings without complete
circularity. Using italics for words and quotation marks for
concepts, we could say that book means ‘book’, but what is ‘book’,
apart from being the meaning of book? This problem faces anyone who
writes monolingual dictionaries, and the practical answer is to
provide an expanded description of the concept to be defined. Here
is the definition of book given in one fairly small dictionary (the
Collins Cobuild Student’s Dictionary):
A book consists of a number of pieces of paper, usually with
words printed on them, which are fastened together and fixed inside
a cover of stronger paper or cardboard.
A definition like this shows the way forward in defining
meanings: the essential trick is to show how the concept being
defined is related to other concepts – in this case, concepts like
‘paper’, ‘words’, ‘printed’, ‘inside’ and ‘cover’. This kind of
analysis invites a ‘mind map’ or ‘semantic network’ as its visual
manifestation, and could play an important part in school teaching
about how knowledge is interconnected. The mindmap in Figure 3
shows how the concepts linked to ‘book’ can in turn be linked to
other concepts in a rapidly expanding network that would ultimately
contain the whole of human knowledge – if only we had enough time
and paper. Notice how the map includes the words that carry some of
these concepts as meanings: the synonyms attach and fix, and the
word book.
[graphic]
Figure 3: A mindmap for the concept 'book'
Exploring meanings in this way is a really good classroom
activity because of its analytical focus on the students’ own
existing knowledge and understanding. This particular example is
all about words, which (according to my expanded definition of
‘grammar’) are part of grammar; but the same kind of treatment is
possible for more traditional grammatical items such as tense and
number. For instance, what is the time structure of a sentence such
as (4), containing a ‘past perfect’ verb (here had chatted)?
(4) I met him at 5.00 although we had exchanged messages at
4.00.
The analysis in Figure 4 shows the time structure of this
sentence, with the message exchange before the meeting, which was
before now (the time of speaking). Notice how this diagram makes
the mindmap slightly more sophisticated by using an arrow to point
from earlier to later (so 4.00 is before 5.00).
[graphic]
Figure 4: The time structure of a past perfect
To summarise, grammarians are well aware of meaning, and have
developed various ways of analysing it so as to be able to bring
meaning and grammar together; mindmaps are very crude
approximations to the kind of analysis that some grammarians use,
but they are a step in the right direction and have the advantage
of being easy to implement in a classroom. Mindmaps aren’t the only
tools on offer, but they are undoubtedly the most accessible.
Grammar and communication
It’s tempting to think that meaning and communication are the
same thing. After all, what is communication but the transmission
of meaning from one mind to another? However, a moment’s thought
shows that they are very different. As any teacher knows all too
well, being able to put a thought into words does not guarantee
that it will reach the mind of someone listening or reading. Our
ancestors were also aware of this challenge, and responded by
moulding our language into a reasonably serviceable tool for
communication. One of the goals of school education is to make sure
that school leavers are aware of this functionality in the language
and can use it well.
You may remember that in my nutshell summary I included this
example, which I called a cleft sentence:
(5) What I think should be taught in schools is grammatical
analysis.
This example illustrates one of our communication options
because it conveys exactly the same meaning as its simpler version
(6):
(6) I think grammatical analysis should be taught in
schools.
As, indeed, do all the following sentences:
(7) Where I think grammatical analysis should be taught is in
schools.
(8) What I think should happen to grammatical analysis is to be
taught in schools.
(9) The one who thinks grammatical analysis should be taught in
schools is me.
(10) In schools, I think grammatical analysis should be
taught.
(11) ?Schools I think grammatical analysis should be taught
in.
(12) ?? Taught in schools I think grammatical analysis should
be.
(13) Grammatical analysis I think should be taught in
schools.
(14) Grammatical analysis should be taught in schools, I
think.
(15) I think they should teach grammatical analysis in
schools.
Examples (7) to (9) illustrate ‘clefting’, which splits the
sentence into two parts linked by is, thereby focusing attention on
the bit after is by presenting it as new information. Examples (10)
to (13) do the reverse, by ‘fronting’ one part of the sentence,
i.e. putting it at the front of the sentence, thereby marking it as
relatively familiar information, or the ‘starting point’ of the
sentence; but you may wonder whether (11) and (12) really are
allowed by English. (As a grammarian I can assure you that such
sentences are quite common.) The point of example (14) is to show
that the main clause I think can be turned into a mere tag, an
afterthought. And (15), of course, shows that the original sentence
(6) was already a departure from the most basic grammatical form,
thanks to the passive be taught. In converting it back to an active
version I have had to supply a new subject, the vague they. Apart
from this addition, all these sentences have the same meaning, but
very different communicative value.
When you consider the various devices made available in our
grammar, they all seem to be geared to making life easier for the
receiver (the reader or listener); so you could take the structure
of English grammar as evidence for our ultimate altruism, as
speakers accepting responsibility for increased effort on our part
in order to benefit others. Alternatively, you could also take it
as evidence for hard-headed rationality, showing how our ancestors
recognised that there’s no point in talking if the other person
can’t understand. Either way, we find that the grammar is concerned
with two related questions:
· What does the receiver know already?
· What mental resources does the receiver have for processing
incoming words?
The current knowledge of the receiver is relevant to the first
three sentences, the clefted ones where one word is removed from
its normal position and located at the end of the sentence while
the rest of the basic sentence is turned into a relative clause.
The diagram in Figure 5 shows how this operation applies to produce
example (5): What I think should be taught in schools is
grammatical analysis. It shows that the cleft version includes two
extra words: what and is, which change a relatively simple sentence
into a much more complicated structure – a complete reorganisation
of the syntax.
[graphic]
Figure 5: The construction of a cleft sentence
Why did I bother to make life more complicated for both of us –
for me as writer and for you as reader? Because I thought it would
improve communication by focusing your attention on the phrase
grammatical analysis. By the time I was composing this sentence I
had already told you that I think schools should teach grammar, and
that I was just about to explain what I meant by grammar, so I
calculated that the focus on grammatical analysis was what you
needed. Figure 5 hints at the complex mental operations that may
have been involved in producing my sentence. The structure is
complicated, but the point of the example is that you don’t have to
be a professional grammarian to cope with it. You probably read it
without any difficulty, and every English speaker produces such
sentences on a regular basis, and as needed by the communicative
demands of the immediate situation. Indeed, they are even more
common in casual conversation than in formal writing (such as this
book) – one research project reports approximately one such example
in every thousand words (Biber et al. 1999: 961).
Cleft sentences, like all the other grammatical structures
illustrated in (7) to (15), are a way of managing and shaping the
information contained in a sentence. When used successfully (as
they usually are) they mould the message to fit the hearer’s
existing knowledge. But this isn’t the only way in which grammar
contributes to communication, as you’ll see if you compare the next
two examples:
(16) That we’re going to miss our train if we don’t walk a bit
faster is very clear.
(17) It’s very clear that we’re going to miss our train if we
don’t walk a bit faster.
I’m sure you’ll agree that, although both sentences carry the
same message, (16) is a lot harder to read than (17). The general
principle explaining why is the principle of ‘end weight’: the
heavier parts of the sentence (such as the long subordinate clause
that … faster) should stand at the end of the sentence. Such
sentences have received attention from cognitive psychologists, who
explain the difference in terms of working memory: (16) demands far
more mental resources because of the very long subject That we’re
going to miss our train if we don’t walk a bit faster. While you
(the reader) are processing this subordinate clause, you have to
remember that it needs to be linked somehow to a main verb; so this
need for a main verb has to be held in memory while you’re
processing all the 15 words of the clause. In contrast, (17) is
easy because the two tasks are separated: the extra word it stands
in for the subordinate clause and shows that it’s the subject of is
clear, so by the time you reach the word that, which introduces the
subordinate clause, you already know how it’s going to fit into the
meaning of the whole sentence. Once again, the sentences carry the
same meanings but communicate it in very different ways; and,
thanks to the extra word it, easy communication comes at the price
of more complex grammatical structures.
How many English teachers are aware of the principle of end
weight? (Crystal 2017a: 106) Another principle of clear writing is
‘order-of-mention’, a preference for mentioning events in the same
order as they occurred, which deserves a great deal of attention in
teaching (Crystal 2017a: 109–116).
Grammar and emotion
It may come as something of a surprise to see grammar linked to
emotion. Meanings, yes, even communication, yes, but surely not
feelings? But why not? After all, if we think of ordinary grammar
as a tool for interacting socially, it would be surprising if it
was incapable of expressing our feelings. And indeed, once you
start looking there’s plenty of feeling in language.
Take exclamations, which enjoy special grammatical patterns in
English:
(18) Oh what a wonderful morning!
(19) How very kind of you!
(20) Isn’t that awful!
All these examples are nothing if not emotional, and the emotion
comes from the grammar, not from the individual words. They also
express a meaning, of course, but they present the speaker’s total
emotional commitment to that meaning, rather than as a neutral
matter of fact.
Or take interrogatives such as who, what, when, where, why, how.
Emotions in this case attach to the question itself, as in (21) and
(22).
(21) Who on earth did you ask?
(22) Why in heaven’s name did you do it?
Every interrogative allows a following expression such as on
earth or in heaven’s name, which are not found anywhere else. They
add nothing to the meaning, but they tell us a great deal about the
speaker’s state of mind. The expressions all indicate the same
emotion of exasperation or surprise at the possible answer to the
question.
Admittedly naked emotion only exists in odd little pockets of
grammar, but these pockets are real and are thoroughly integrated
into the heart of grammar. They are important evidence against two
views of grammar: that it is all about cold logic, and that it is
dry and remote from real flesh-and-blood people.
Grammar, pronunciation and spelling
Grammar is a bridge between communicated meaning and the sounds
or marks that communicate it. As the Duchess said to Alice (with
careful allusion to the saying Take care of the pence, and the
pounds will take care of themselves), "Take care of the sense, and
the sounds will take care of themselves." Grammar (including
vocabulary, of course) is the mechanism that guarantees this
connection. If we think of words as the basic building blocks of
language, then each word faces in two directions at the same time:
upwards towards inaudible and invisible meanings, and downwards
towards the audible or visible signals of pronunciation and
spelling.
The area of grammar that links downwards is morphology – the
study of word-shapes. Morphology reveals a complex mixture of the
arbitrary and the regular. Take the word reveals, for example. This
is arbitrary in that there is no generalisation to explain why the
morpheme {reveal} should mean the opposite of conceal or hide.
That’s just how English is – and likewise through the tens of
thousands of words that make up our vocabularies. It’s true that we
can trace the history of the word reveal back to Latin revelare,
meaning ‘un-veil’ and linking to velum meaning ‘veil’, but sooner
or later you hit an impenetrable wall of arbitrariness – in this
case the connection between the meaning ‘veil’ to velum; and in any
case the history of reveal is irrelevant until we check an
etymological dictionary. So the connection between reveals and its
meaning is basically an arbitrary social convention that we
inherit, more or less unthinkingly, from many generations back.
On the other hand, the {s} on the end of reveals is anything but
arbitrary: it is completely predictable from the rule for making
present-tense verbs agree with their subjects (whereby morphology
reveals, but morphologists reveal). Morphology spans the whole
gamut from total arbitrariness to fully motivated and
rule-governed, with a considerable middle ground where there is
some motivation, but not much; for example, the re- of reveals is
also found, with much the same ‘reversive’ meaning, in other verbs
such as reverse, remove, replace, and reinstate, but how many of us
are aware of this similarity?
Morphology, then, studies the relations between words and their
shapes; but of course what we mean by a word’s shape depends
heavily on whether we are thinking of spoken or written language.
In English, morphemes are where sounds and spellings meet, which is
part of the reason why English spelling is so ‘un-phonetic’ – why a
word’s spelling often can’t be predicted from its pronunciation, or
its pronunciation from its spelling. Consider the following
examples:
(23) The tax is unfair.
(24) The tacks are sharp.
The point of the examples is, of course, that the words tax and
tacks have exactly the same pronunciation: /taks/. If English
spelling had been phonetically regular, they would have shared the
same spelling, which would probably have been . The actual spelling
has the advantage of highlighting the morphological difference
between the monomorphe {tax} and the bimorpheme {tack}{s}. In a
nutshell, the final /s/ has two completely different reasons for
being there: as part of the root morpheme {tax} and as the morpheme
that distinguishes plural nouns from singulars. By writing in tacks
we allow it to be a separate morpheme. Moreover, this spelling also
allows this morpheme to be the same as the one at the end of (say)
toes or horses, in spite of the fact these have a different
pronunciation for this morpheme: /z/ in toes and /ɪz/ in
horses.
It’s quite uncontroversial among grammarians to claim that
English spelling is oriented strongly towards morphology, and the
claim is also accepted by educationalists (Nunes & Bryant
2009). But if it’s true, it has serious implications for the ways
in which we teach spelling in schools, where UK government policy
is to promote phonics – the teaching of ‘grapheme-phoneme
correspondences’ (e.g. that between the letter and the phoneme
/s/), and little more. Instead, schools need to teach morphological
analysis in a systematic way so that children can understand the
pressures from morphology. As we shall see in later chapters, this
argument based on spelling provides a strong argument for teaching
grammatical analysis.
The aim of this section has been to expand the bottom part of
the diagram in Figure 2, giving the structure in Figure 6. This
diagram focuses on the ‘interfaces’ – the links between different
kinds of structure:
· morphosyntax relates the words of syntax to the morphemes of
morphology (e.g. ‘the plural of tack’ is {tack} {s})
· morphophonology relates morphemes to their phonology (e.g.
{tack} is pronounced /tak/)
· morphographology relates morphemes to their spelling (e.g.
{tack} is spelt )
· grapheme-phoneme correspondences show how graphemes are paired
with phonemes (e.g. is pronounced /s/).
[graphic]
Figure 6: From words downwards
You may be wondering whether morphemes deserve their important
place in this picture. You probably accept that words are different
from phonemes and graphemes, and that phonemes and graphemes are
different from each other; so you’re happy to distinguish the word
tack from the phoneme /t/ and the grapheme . But is the word tack
really different (as I’m claiming) from the morpheme {tack}? After
all, they’re the same size, so why not simplify the diagram by
relating the word tack directly to its phonemes and graphemes? And
it would be very easy to write rules for changing words without
invoking morphemes at all – for example, to make a noun plural,
just add /s/ at the end. If you are thinking along these lines,
then you’re in he very respectable company of some distinguished
linguists, so your objection deserves serious attention.
Let’s start with the objection based on size: tack is the same
size as {tack}, so they must be the same thing. Would you raise the
same objection for very short words such as the pronoun I? Would
you argue that since the word I is the same length as the single
(but internally complex) phoneme /ai/ and the single letter they
must be the same thing? Surely not. The point here is that words
and phonemes or letters have very different kinds of properties.
The word I is a pronoun, and more precisely (in grammar-speak) it’s
the first-person singular personal pronoun in subjective form. None
of this information has anything to do with pronunciation or
spelling. In contrast, the phoneme /ai/ is a long vowel which is
made up of two parts that form a diphthong, and the letter is an
uppercase vowel letter in the Roman alphabet; and this time the
information has nothing to do with the word’s grammar. So in
grammar size really doesn’t matter: two things can be different
even if they have the same length. Returning to tack and {tack},
the same argument applies: even though they have the same length,
they have very different properties. The word tack is a common noun
with the meaning ‘tack’, whereas the morpheme {tack} is a root
morpheme which isn’t tied to this one word. It is found not only in
the plural tacks but also in the verb to tack (as in You need to
tack it down at the edge) and all its forms: tacks (this time
singular!), tacked and tacking, not to mention any words derived
from the noun or the verb such as tacker and tacky. We can also
identify it in homophones such as the noun tack (meaning harness
for horses) and may even recognise it (wrongly) in the word tackle.
In short, the word tack has a very different status and set of
relations from the morpheme {tack}.
That leaves the second objection, that rules can easily be given
without mentioning morphemes. Yes, it’s true that we could say that
plural nouns add , but the question is whether this is just a
common-or-gardent . The morpheme analysis claims that it’s a
special , different from those in or . What’s special about it is
its pronunciation, its grapheme-phoneme correspondences: as pointed
out earlier, it has different pronunciations according to how the
previous sound is pronounced. In cats, dogs and horses it is
pronounced /s/, /z/ and /ɪz/. Of course it would be possible to
formulate the rules so as to give the right answer: plural nouns
add , which is pronounced /s/ after a sound like /t/, and so on.
But this would miss a crucial fact: exactly the same differences
are found in the which makes a verb singular (as in eats, begs and
forces). If we connect words directly to their spelling and
pronunciation we have to ignore this similarity, but we can
recognise it by by introducing the morpheme {s} between the words
and their pronunciations or spellings. The two analyses are shown
in Figure 7 and Figure 8. Personally I think it’s obvious which
analysis is better – but it is only possible if we distinguish
words and morphemes.
[graphic]
Figure 7: Plural nouns and singular verbs without {s}
[graphic]
Figure 8: Plural nouns and singular verbs with {s}
Grammar, intonation and punctuation
I said in the previous section that morphemes are the
meeting-point for pronunciation and spelling, but there’s another
place where pronunciation meets writing, namely wherever intonation
– an important part of pronunciation – meets punctuation. This
place is somewhere in the realm of syntax, meaning and
communication, which is a clumsy way of saying that it’s
complicated! But the important point that I’d like to make here is
that intonation and punctuation are functionally similar: whatever
syntactic, semantic or communicative pressures apply to intonation
also apply to punctuation.
Most people are much more aware of punctuation than of
intonation, but we all use intonation all the time – and use it
with great expertise; though comparison is very difficult, most of
us almost certainly use intonation more accurately and more
effectively than punctuation. What I would like to do is to make
you more aware of this skill that you have (if you speak English as
well as reading it); but of course this is difficult in a book
because there’s no standard way of representing intonation. Let me
start with a little activity for you, based on one of the glories
of English grammar: the question tag. This is the little question
tagged onto the main clause in examples like these:
(25) He can swim, can’t he?
(26) He can swim, can he?
Question tags are glorious for three reasons: First, they’re
complicated because the tag varies with the main clause: can’t he
after he can, isn’t she after she is, and so on, including didn’t
it after examples like it broke. (The younger generation have
recently made things even more complicated by introducing an
invariant innit as in I missed the bus, innit? Since I don’t
pretend to understand this innovation I shall ignore it.) Second,
they carry very subtle contrasts concerned with communication; and
third, the intonation is rather rigidly limited. There are
basically two options for the intonation: rising or falling, which
are easily distinguished visually and provide an easy way into the
study of intonation. (For simplicity I’ll drop the question mark
from the examples to leave room for the intonation mark.)
(27) He can swim, can’t he ↗
This starts from an assumption (that he can swim), and asks for
confirmation; the speaker isn’t sure, but thinks the hearer knows
for certain. As a communication, this is very different from
(28):
(28) He can swim, can’t he ↘
Here, the speaker is making a judgement or an exclamation, so it
could be introduced (at least in this example) by Wow! As a
personal comment it expresses the speaker’s state of mind, or even
feeling, so the only uncertainty is about whether the hearer feels
the same. That’s what the tag is about.
These two examples both have the same words, in which a positive
main clause takes a negative tag (can swim … can’t …) – positive
plus negative. Another possibility is positive plus positive, with
rising intonation:
(29) He can swim, can he ↗
In this case the idea that he can swim comes from someone else –
probably the hearer – and the speaker is self-distancing. The
larger context might Oh, so he can swim, can he ↗ In that case, why
is he so reluctant to get in? Once again, the dynamics of the
communication are quite different. And finally, to the fourth
logical possibility (positive + positive with falling tone):
(30) He can swim, can he ↘
Try saying this to yourself, and I predict that you’ll find it
very hard. If I’m right, it’s because this combination isn’t
permitted by the grammar. The grammar allows two patterns of words:
positive + negative, or positive + positive (it actually also
allows combination where the first element is negative, but I’m
ignoring those for simplicity); and it allows two intonation
patterns: falling or rising. But falling intonation is only allowed
after the negative tag. Why? Because falling intonation on a tag
means not only certainty but also emotional commitment, while the
positive tag means self-distancing and doubt; and commitment and
doubt are contradictory emotions.
The point of the example is that intonation is closely connected
with grammar, while also preserving a functional life of its own
because it carries a distinct range of meanings or communicative
functions which combine with those carried by the words. Much the
same is true of punctuation, which is in many respects a much
cruder and less efficient substitute for intonation.
Just like intonation, punctuation is also related to grammar, so
skill in punctuation indicates some skill in syntactic analysis;
but as I’ll explain below, punctuation (like intonation) also has
direct links to communication. Since communication and syntax can
pull in opposite directions, this produces tensions within the
punctuation system which skillful writers can exploit well but
which baffle novices. But let’s start with syntax, and specifically
the syntax of sentences. What is a sentence? Since some punctuation
marks (full stop, question mark and exclamation mark) are called
‘sentence punctuation’, we may expect them to mark the boundaries
of the sentences that are defined by syntax; and by and large they
do. But they’re the result of a mental analysis by the writer, and
don’t in themselves explain the analysis. How do I decide where to
put a sentence mark? This takes us back to my first question: What
is a sentence? How do I recognise sentence boundaries when I’m
writing?
For a grammarian, the answer is easy: a sentence boundary is
where syntax runs out. Words inside a sentence are connected by
syntactic links, but words in different sentences are not. Take the
sentence that I’ve just written: Words inside … are not. Every word
in this sentence has a syntactic link to at least one other word in
the sentence; for instance, the first word, Words, is linked
directly to are (Words … are …) and also to inside, which in turn
is linked to a, which is linked to sentence. This analysis is shown
in Figure 9; the details are controversial but the main point about
syntactic relations holding a sentence together is not.
[graphic]
Figure 9: A sentence held together by syntax.
Every arc (curved arrow) in this diagram represents a relation
defined by at least one syntactic rule which allows one word to be
subordinate to another, while but indicates an equal relation
between the two parts of the sentence which, in grammar-speak, are
therefore coordinate. Syntax is all about subordination and
coordination, so wherever you can find one of these relations
between two words, they are glued together by syntax and are
therefore part of the same sentence, and if you’re writing rather
than speaking, you need a capital letter at the beginning and a
sentence mark at the end. Incidentally, using arrows to show
syntactic relations may be useful in teaching grammar at school;
it’s the only diagramming system that I’ve seen in a modern grammar
textbook for schools (Stone 2010).
Syntax is also relevant to punctuation within the sentence.
Subordination is a complex affair, as it involves a range of
different relations which school grammars managed to distinguish
reasonably efficiently by the end of the nineteenth century (though
very few even try these days). Since each such relation involves
two words, we have to start by distinguishing the subordinate from
the word to which it is subordinate, called respectively the
dependent and the head; so in the pair words … are, the head is are
and words is the dependent; but in the pair words in, words is head
and in is dependent. Ultimately subordination is a matter of
meaning: the head provides a general meaning which the dependent
makes more precise; so words inside a sentence are a particular
kind of word, not a kind of sentence. But identifying heads and
dependents is complicated and the answer isn’t always obvious
(though in most cases grammarians agree on the answer). This
distinction isn’t in itself relevant to punctuation, but once we
start distinguishing different kinds of dependent the distinctions
are highly relevant to punctuation. For instance, consider these
sentences:
(31) They do _ whatever I say.
(32) They misbehave _ whatever I say.
Which of the marked blanks would accept a comma? I hope you
agree that a comma would be possible after misbehave, but quite
wrong after do. Why the difference? Because although whatever is a
dependent of the verb in both sentences, it is a complement of do
but an adjunct of misbehave. The distinction between complements
and adjuncts is a central part of syntactic theory, and is a modern
version of the old-fashioned idea of ‘a complete idea’: a
complement ‘completes’ the verb do whereas misbehave is already
complete without the adjunct, which is simply added on. This isn’t
the place to explore the classification of dependents, but my point
is simply that your use of punctuation is already sensitive to this
particular distinction. Syntactic theory can explain your
punctuation practices.
You won’t be surprised to learn that your punctuation is
actually even more complicated than this, because as I mentioned
earlier it tries to do two different, and conflicting, jobs at the
same time. On the one hand it reflects syntactic patterns such as
sentences and the complement/adjunct distinction; but at the same
time it tries to help communication by ‘chunking’ the written text
into manageable units, and by indicating communicative functions.
Chunking allows us in effect to use full stops anywhere we want the
reader to pause. Everyone. But everyone. Knows. It. Works. But only
so long as you don’t overdo it.
Communication and syntax place different and potentially
conflicting demands on the punctuation system. Take the question
mark. As a sentence-marker this signals the end of a sentence, but
as a tool of communication it signals a question. But what if the
question is only part of a sentence, as in the very everyday
example (33)?
(33) Is this your kettle _ because it’s boiling_
I’ve omitted punctuation, so how would you punctuate it if you
wrote it down? Since it’s partly a question, you’ll want a question
mark, but where would you put it?
(34) Is this your kettle, because it’s boiling?
(35) Is this your kettle? because it’s boiling.
You can see the problem: (34) respects the question mark’s
syntactic function, but wrongly implies that ‘cos it’s boiling is a
question; but (35) respects its communicative function at the
expense of the syntactic function – and what about the next word,
because? Should it have a capital letter?
I hope to have shown that both intonation and punctuation have
strong links to syntax, just as strong as those from pronunciation
and spelling to morphology. Fortunately we don’t on the whole need
instruction in pronouncing words or in doing intonation when
speaking, but both spelling and punctuation need a great deal of
explicit teaching; in fact, they take up most of the time devoted
in our primary schools to ‘literacy’ teaching, and between them
they constitute the ‘transcription’ skills which writing experts
contrast with the higher-level skills of ‘composing’. But since
both spelling and punctuation build on grammatical foundations, it
makes intellectual sense to see spelling, punctuation and grammar
as specially linked, as in the ‘Spelling, Punctuation and Grammar’
tests which every child takes on leaving one of England’s primary
schools.
Grammar and identity
One slogan which has emerged in the recent UK debate about
grammar teaching is “Tools not rules”: grammar is more like a
tool-kit than a rule-book. If a tool is an artefact dedicated to
serving some function, a hammer is a tool for hitting, then grammar
is an archetypical tool except that it serves a number of different
functions. One function is to allow us to express complex meanings,
and another is to help us with complex communications; as I
explained in section 2.6, these functions are distinct. A third
function, however, is to announce our social identity. Grammar
provides alternative ways of expressing the same meaning which are
geared either to communication or to social identity. What we say
reveals not only what we’re thinking but also who we are.
For example, take the two ways of saying the same thing in (36)
and (37).
(36) I didn’t say nothing to nobody.
(37) I didn’t say anything to anybody.
As I explain in chapter 4.1, these two sentences have precisely
the same meaning, but they project very different identities. Those
who want to sound educated tend to use the second, but most people,
most of the time, use the first – just as Shakespeare and others
did before them. It all depends on what kind of social identity you
want to project. And of course since our social identity can shift
from situation to situation, so can our grammar: someone who used
(36) with friends or family might switch to (37) in a committee
meeting or a lecture. This is one example among many which
distinguish the grammar of Standard English from other dialects;
others include they were/was, those/them books, the book that/what
I bought and come quickly/quick. If you find these choices
familiar, then it must be because you know both forms, so in
principle you could also use both. But you also know that each
alternative form comes with social baggage attached, which you may
or may not value.
Any grammar allows some choices based on social rather than
semantic criteria – in other words, different ways of saying the
same thing which are appropriate to different social situations or
to different types of people. One particular manifestation of this
principle is in language change, where the relevant social variable
is age: old people speaking differently from youngsters. This
contrast is very familiar in slang, where this year’s exciting
innovation is next year’s old-fashioned quaintness, but it also
applies to other areas of language, including grammar. At any given
moment some details of your language are gradually changing, and if
you know where to look, you can see (or hear) them changing. Take
the verb to have, as in (38) and (39)Error! Reference source not
found..
(38) They had finished.
(39) They had a good time.
Although they look the same, these verbs are very different
syntactically. The contrast is clearest between (38) and (39), as
you’ll see if we try to make them negative or to turn them into
questions:
(40) They hadn’t finished.
(41) Had they finished?
This had is an auxiliary verb, which take {n’t} in the negative
(had > hadn’t) and form questions by moving up before their
subject (they had > had they). In contrast, the one in (39) is a
‘lexical verb’ – i.e. not an auxiliary. Here the rules are very
different.
(42) They didn’t have a good time.
(43) Did they have a good time?
In both cases, we have to add did: had > didn’t have, they
had > did they have. It’s simply wrong to treat this had like
the auxiliary, giving *They hadn’t a good time and *Had they a good
time? (Remember: * marks an example as ungrammatical.)
This major distinction between auxiliary and lexical verbs is
actually an innovation in English grammar which as come in since
the time of Shakespeare. It’s happened very gradually over the
centuries, as each generation pushed it a step further; and we’re
still pushing, sorting out the consequences. One locus of ongoing
change at the moment is the possessive have, as in (44).
(44) She has brown eyes.
I predict that my readers will now divide according to age (and
geography) into three groups: ultra-conservatives, conservatives
and innovators. The ultra-conservatives may be happy with (44) but
mere conservatives will prefer (45) (where, please note, got does
not have the usual meaning of to get, meaning ‘come to have’).
(45) She’s got brown eyes.
When we try the negation and question tests, the differences
become more interesting. Ultra conservatives (who tend either to be
elderly or to live in the north of Britain) will accept (46) and
(47). They’re still treating the possessive have like an ordinary
auxiliary verb – in spite of the fact that there’s no other verb in
sight, and in fact this verb behaves in other respects just like a
transitive lexical verb such as possess.
(46) She hasn’t brown eyes.
(47) Has she brown eyes?
Mere conservatives, on the other hand, prefer the has got
alternative, so for them negatives and questions are
straightforward, as in (48) and (49). This is straightforward
because have is a straightforward auxiliary verb, supporting the
(rather meaningless) got.
(48) She hasn’t got brown eyes.
(49) Has she got brown eyes?
Finally, the innovators (presumably led by young Americans) have
turned possessive have into a lexical verb like the one in have a
good time. For them, the negative and question based on (44) are
(50) and (51).
(50) She doesn’t have brown eyes.
(51) Does she have brown eyes?
The direction of change is clear: from the ultra-conservative
pattern either to the mere conservative one or to the innovative
one. The whole process is driven partly by a desire to tidy up this
little corner of English grammar, but partly too by a desire to
identify with ones region and age-group.
Another consequence of the link between grammar and identity
which is particularly relevant to education is the way in which
‘academic language’ varies from subject to subject. Every school
subject has its own special grammar, which experts use
automatically and probably without even being aware of it but which
novices are expected to master. Teachers have a responsibility for
helping novices to build this special grammar on top of their
existing everyday grammar. Chapter 4.1 gives an extended example of
this process in mathematics but let’s briefly consider another
example: tense in the English lesson. Imagine a student writing
about Romeo and Juliet, knowing that Shakespeare lived centuries
ago and that Romeo and Juliet were supposed to have lived even
earlier. In ordinary English we would expect the past tense,
whether we’re talking about Shakespeare or about Romeo:
(52) Shakespeare presented Romeo as madly in love.
(53) Romeo was madly in love.
But by convention English expects the present tense: presents,
is. This convention presumably stems from the idea that both
Shakespeare and Romeo are both eternally present through the
written medium, which makes good sense to the experts, but may need
to be explained to novices who want to write like experts.
Identity and grammar, then, are closely intertwined. Any
particular bit of grammar may be linked to some kind of social
identity, just as it may be linked to a semantic or communicational
function. Grammar is not monolithic – always the same under all
circumstances. Nor is it modular, a box of grammar disconnected
from the rest of life and knowledge. Why should it be? Given the
way we learn all our grammar through some kind of interaction with
other people, intimately interconnected with the whole network of
social relations and goings on, it would be astonishing if grammar
remained untouched by these other influences.
Grammar and research
You may wonder where all these ideas and facts about grammar
come from, so it’s time I said a bit more about grammatical
research. As I explain more fully in chapter 0, grammar teaching is
very ancient, and the same is true of research, even if research
may seem too grand a name for people sitting around and thinking
hard about their language. Grammar teaching and research have had
ups and downs during their four-thousand year history, but at the
moment – as of 2020 – research is on an up. In fact, it’s thriving,
though the same can’t yet be said of grammar teaching. Most
universities have a department of linguistics, and every such
department contains at least one grammarian, and maybe several; so
the academic calendar is well supplied with international
conferences on grammar. For example, the Linguist List, an
international collecting point for information about linguistics
world wide, lists conferences about linguistics. For June 2020 it
lists 17 such conferences, including three that are specifically
about grammar, and most of the others are likely to include
numerous presentations with a focus on grammar.
What does a grammarian do when researching? At one time
grammatical research meant pouring over old books studying a dead
language such as Latin or Greek. This stereotype was so familiar
that it even provided the hero for a famous poem published in 1855
by Robert Browning, The grammarian’s funeral which includes the
following lines.
This [man], throws himself on God, and unperplexed
Seeking
shall find him.
So, with the throttling hands of death at strife,
Ground he
at grammar;
Still, thro' the rattle, parts of speech were rife:
While he
could stammer
He settled Hoti's business let it be!
(Hoti is a Greek conjunction meaning ‘that’ or ‘because’.) Maybe
some modern grammarians still work like this, but most of us do
things differently. Not that we’re short of passionate commitment
to grammar, and in particular to understanding grammar, but we have
different languages, different data, different methods and maybe
different goals. The classical languages are no longer centre
stage, and have been replaced by English, which attracts more
attention than any other language; for example, we now have no
fewer than five recent ‘blockbuster’ grammars of English each
containing between 1,000 and 2,000 pages (Quirk et al. 1972; Quirk
et al. 1985; Biber et al. 1999; Huddleston & Pullum 2002;
Carter & McCarthy 2006), not to mention thousands of scholarly
jornal articles about English. But grammarians have also studied
thousands of other languages from across the world; for example, my
PhD was a study of the grammar of Beja, a language spoken by about
a million people in the north-eastern part of The Sudan. This
research was what we call ‘field work’, where I sought out native
speakers of the language; but since then most of my work has been
on English, where I am my own native speaker. Fieldwork on an
exotic language is a wonderful preparation for looking at one’s
native language (which turns out to be just as exotic in its own
way).
Throughout its long history, grammatical research has combined
descriptive work – exploring and describing the factual details of
particular languages – with theoretical work. Description and
theory are actually two sides of the same coin, which we might call
‘grammatical analysis’. Any description presupposes some kind of
theoretical framework (even if it’s only the idea that language
consists of basic units such as words), and every theory
presupposes evidence in the form of descriptions. One of the
reasons why the study of grammar is in such good health nowadays is
that we have really good theory to support really good description.
In fact, we have a lot of competing theories – which, of course, is
not something to boast about because we can be sure that they’re
not all right, and there’s a disturbing possiblity that they are
all wrong. The most famous theory, and probably the one with the
most followers, is called ‘Minimalism’ and was invented by Noam
Chomsky, who is certainly the most famous linguist (and
grammarian); but there are many more, including one that I invented
called ‘Word Grammar’ (Hudson 1984; Hudson 1990; Hudson 2007;
Hudson 2010; Gisborne 2010; Duran Eppler 2011; Traugott &
Trousdale 2013).
Let me put flesh on these rather abstract bones about
description and theory. Consider the very elementary English data
in Table 1.
subject
present
present + not
reduced + not
present + n’t
they
they are
they are not
they’re not
they aren’t
you
you are
you are not
you’re not
you aren’t
we
we are
we are not
we’re not
we aren’t
he/she/it
she is
she is not
she’s not
she isn’t
I
I am
I am not
I’m not
????
Table 1: The amn't gap
Assuming that you know English pretty well, my question for you
is what you would put into the bottom right-hand cell. Would you
say I amn’t or I aren’t? Or something else? As an extra data-point
I might point out that you probably do have a ready-made form for
the question in (54).
(54) Aren’t I your friend?
If you come from Scotland or Ireland I predict that your answer
for the statement is I amn’t, but otherwise I predict that you
don’t have any answer at all: you don’t like either I amn’t or I
aren’t. Why? This question moves us straight into grammatical
theory. What kind of answer does your theory allow? Here are some
possibilities that you might consider:
· We don’t like amn’t because it’s too hard to pronounce. This
can’t be right because the Scots and Irish use it very happily; and
in any case, we can easily argue on theoretical grounds against
this kind of explanation for the simple reason that phonology – the
bit of language that handles pronunciation matters – exists
precisely to make sure that every word is pronounceable. So if we
didn’t like saying amn’t we could easily change it into a more
pronounceable form such as /amənt/ (where ə is schwa, the vowel at
the start of about).
· We don’t like amn’t because we’ve never heard it. This can’t
be right either because we’re good at generalising, and the
relevant generalisation in the table is very simple and obvious. If
is not gives isn’t, and are not gives aren’t, what does am not
give? This is a very simple analogy indeed, and yet we all resist
it. In any case, this explanation begs the question: why have we
never heard amn’t? The answer, of course, is that we’ve never heard
it because nobody else uses it, but this takes us back to the first
question: why don’t any of us use amn’t?
· We don’t like amn’t because it’s somehow incoherent or
contradictory. This is the explanation that I’ve proposed (Hudson
2000), but it takes us right into the heart of grammatical theory
because we need to know precisely what the mechanisms of grammar
are. My suggestion is that generalisations in grammar (as in the
rest of our thinking) apply by default, so they apply unless
they’re blocked by an exceptional subcase. For example, by default
a noun becomes plural by adding {s}, but exceptionally the plural
of mouse is mice, so the default rule doesn’t apply. This is
unproblematic because it’s clear that mouse is a subcase of ‘noun’.
But if two incompatible generalisations apply, and neither is a
subcase of the other, we reach a stalemate where we simply don’t
know what to do. In the real world this is a familiar situation;
for example, if one door is for females and the other is for
students, which one should a female student use? This situation is
parallelled in the case of *amn’t, which is simultaneously ‘first
person’ (notice that be is the only verb in English that has a
special form for the first person, i.e. for use with I) and
‘negative’, neither of which is a subcase of the other. And the
problem is that these two features demand incompatible forms: am
for ‘first person’, and aren’t for ‘negative’.
This explanation strikes me as much more plausible than the
other two, but of course there may be a fundamental fault in it,
and someone else may be able to come up with an even better
explanation. But whatever the merits of my explanation, the point
is that the problem demands an explanation formulated in terms of a
clear and coherent theory of grammar.
I hope to have given some sense of why so many people (like me)
find grammar an exciting and challenging area of research. As
chapter 0 will show, the grammar taught in schools has always been
validated ultimately by research, but the links to research are all
too easy to forget or ignore, leaving the teacher with nothing but
half-understood dogma, to be taught (in the worst cases) by
corporal punishment and rote learning, rather than as a body of
research-based learning which throws light on this important area
of our intellectual life.
Grammar and correctness
Finally we come to the idea that many discussions of grammar
treat as their starting point: that grammar is about correctness.
This view presents grammar as primarily a means for avoiding error.
I’ve deliberately left this view to the end because I wanted to
persuade you that grammar is actually much, much more than
that.
Correctness, in this view, is to be achieved by learning a list
of common errors, and then learning to avoid these errors. Just to
be clear, this particular interpretation of error-avoidance is
absolute anathema to me, and I reject it totally. As a grammarian I
think it is total nonsense, and in education I believe that its
effects are all, and always, negative. In case you don’t know what
I’m talking about, here’s a list of errors from one website that I
found by googling for “common grammar mistakes”:
· They're vs. Their vs. There
· Your vs. You're
· Its vs. It's
· Incomplete Comparisons
· Passive Voice
The lists starts well with some homophones where spelling needs
morphology (section 2.8), but then we hit incomplete comparisons.
What on earth are they? Apparently they’re sentences like this
one:
(55) Our car model is faster, bigger, stronger.
Faster, bigger an