Making a Point

by David Crystal

  Commas separate sentences? Not normally. Teachers would throw their hands up in horror at the thought. Any child who wrote:

  I went to the beach at the weekend, I had an ice-cream …

  would be likely to have the comma corrected. On the other hand, what would a teacher do if a (precocious) child came in the next day and said: ‘Please, miss, why did you correct my comma when in my history book about Julius Caesar it says this?’

  I came, I saw, I conquered.

  Read this chapter, would be my answer.

  Punctuation has a differentiating function: it developed in order to allow writers to express differences of meaning. It would be pointless (sic) to have a punctuation mark that did exactly the same job as some other punctuation mark. One or the other would eventually die out, or would come to be associated with different kinds of users (such as British vs American). There will always be a few cases where there’s an overlap in function – where users have a free choice between mark X and mark Y – but most of the uses will be distinct, and juxtaposing examples is the best way of seeing what the distinctions are.

  That is why commas aren’t usually used to join sentences. There are already marks which do that job very well – the period, the semicolon, and the colon, which allow us to make the broad distinctions of meaning I’ve discussed in previous chapters. Let me illustrate using dear old Smith and Brown of semicolon fame. We don’t normally join two sentences like this:

  Smith is going to speak about cars, Brown is going to speak about bikes.

  This is making the comma do a job it was not designed to do. And in marking it wrong, in a learner’s writing, it’s important to explain this. The comma has evolved to convey a set of meanings that is different from those conveyed by other marks. To use it in a place where one of those other marks should go is to mislead your reader.

  Once again, we’re talking semantics. Discussing the use of the comma in solely grammatical terms is not the solution. We have to dig deeper. This is why the traditional grammars were of limited help. If we know how to talk about grammar, the description of this next sentence is easy to make:

  Smith is going to speak about cars; and Brown is going to speak about bikes.

  The semicolon is linking two independent clauses, the second one introduced by a coordinating conjunction. Now let’s turn to this version:

  Smith is going to speak about cars, and Brown is going to speak about bikes.

  The description is the same: the comma is linking two independent clauses, the second one introduced by a coordinating conjunction. Grammar is thus no help in showing us when to use the comma and when the semicolon. We have to talk about the kind of meaning that the marks are conveying. Semantics again.

  What commas do is they allow us to show a closer semantic association between the two clauses than if we used the semicolon. Let’s recapitulate the point made about semicolons in Chapter 22:

  Smith is going to speak about cars; Brown is going to speak about bikes.

  The semicolon shows the clauses are linked, more than they would be if they were separated by a period; but the link is not very tight. Each clause could be used as an independent sentence without the other:

  Smith is going to speak about cars.

  Brown is going to speak about bikes.

  The semantic link remains the same even if we reinforce the semicolon by a coordinating conjunction:

  Smith is going to speak about cars; and Brown is going to speak about bikes.

  The value of a coordinator is that it makes explicit the nature of the semantic relationship between the two clauses, and this can vary. We can replace the and by one of the other words that have a coordinating function, such as but, or, and so.

  Now look what happens when we use a comma. Immediately the semantic link between the two clauses is tighter, and this is usually reflected by other signs of tightness in the second clause, such as:

  Smith is going to speak about cars, and Brown will follow him by speaking about bikes.

  Smith is going to speak about cars, and we’re really looking forward to his talk on this subject.

  Here, words like him, his, and this refer directly back to Smith (grammarians call this effect anaphora), and bind the two clauses tightly together. It’s no longer possible to use the second clause as an independent sentence:

  We’re really looking forward to his talk on this subject.

  The comma, in other words, mustn’t be studied in isolation. It forms one of the ways in which we make a close semantic connection between clauses, and the closer that connection (shown by such devices as anaphora) the less likely we are to use a semicolon:

  Smith is going to speak about cars; and we’re really looking forward to his talk on this subject.

  The punctuation disrupts the flow. The other cues in the second clause (his, this) are signalling that the writer sees it closely related to the first. The semicolon is signalling that the relationship is more distant. The writer can’t have it both ways.

  It’s the tightness of that link which explains why, in a few instances, we can link independent sentences by a comma. This happens when the tightness is signalled by other features, usually by the sentences being completely parallel in construction and the rhetorical effect of the whole being greater than the sum of the parts. That’s why we see it in relation to Julius Caesar – and also here, in this fragment of Winston Churchill’s speech to the House of Commons on 4 June 1940:

  … we shall fight on the beaches, we shall fight on the landing grounds, we shall fight in the fields and in the streets, we shall fight in the hills …

  This shows how a rhetorical motive – he is in the last, climactic paragraph of his discourse, and the end of the speech is just seconds away – can use the comma to link any number of sentences. That’s why I said ‘don’t normally join’ above. But you have to be an experienced language user to use this strategy effectively.

  The tightness of the semantic link shown by a comma is also why we can use it in a much wider range of constructions. Semicolons reinforce the link between clauses that coordinate ideas. Commas do this too, as we’ve just seen, but they also reinforce clauses where one idea is subordinated to another, as shown by the use of subordinating conjunctions:

  Smith is going to speak about cars, because it’s a subject he loves.

  Smith is going to speak about cars, whether we want him to or not.

  Again, we wouldn’t expect to see these clauses joined by a semicolon:

  Smith is going to speak about cars; because it’s a subject he loves.

  Smith is going to speak about cars; whether we want him to or not.

  Mixed messages again.

  The focus in these examples is on the second clause. That’s always how it is. Imagine Hilary writing to her friends about the forthcoming conference. She reaches the end of the first clause:

  Smith is going to speak about cars

  She now has to decide how much she wants her next thought to be linked to this one. If she wants to make a totally separate remark, she will use a period. If she wants to show a balanced contrast, she will use a semicolon. If she wants to introduce a list, she will use a colon. And if she wants to immediately say more about Smith, his speaking, or his cars, she will use a comma – or no punctuation at all.

  We must consider this last option, before moving on to the other. ‘No punctuation’ is the ultimate marker of semantic tightness. With only word spaces left, the words between clauses now look the same as the words within the clauses.

  Smith is going to speak about cars and Brown is going to speak about bikes.

  If the clauses are short, this presents no problem to the reader, who can take in the structure of the sentence at a glance. But the usage has to be motivated, of course. Why would a writer want to bind the two clauses together so tightly? Well, imagine a scenario where Hilary rushes into a room in a state of great excitement, because she has just learned the exciting news. There’s no pause. And if
this excitement were to be represented in writing, there’s no time for a comma.

  Comma omission is most likely when the semantic link is one of ‘and’. It’s a reflection of the normal way in which people narrate – a strategy that is found in young children from the age of three, when they first learn to coordinate sentences:

  Mikey did go in the garden and – and – he did go on the swing and – and – and – he wented high up …

  It’s also the commonest way of linking sentences in the early stories that children write. And it continues into adult spoken narrative style: any story will contain more instances of and than any other conjunction. But if we start to be more subtle in our connecting, making a contrast using but or and yet, then writers usually insert a comma:

  Smith is going to speak about cars, but Brown is going to speak about bikes.

  It’s less easy now to imagine someone rushing in and telling this story (and thus, omitting the comma). The focus is now more meditative, on the contrast between the two events. The comma reinforces the meaning conveyed by but and draws our attention to it. Why go to the trouble of drawing the readers’ attention to a semantic contrast if you’re then going to make it more difficult for them to see the contrast (by omitting the comma)? Mixed messages again.

  Any grammatical construction that binds two short clauses tightly together will motivate the omission of the comma. Look what happens when we leave out parts of the second clause, as here:

  Smith is going to speak about cars and wants a room with a projector.

  A comma is less likely, as we need to read in the subject of the first clause (Smith) and this forces the two parts together. We might find a comma if the writer saw the second clause as some kind of afterthought, in which case a dash would be a more effective way of conveying this meaning:

  Smith is going to speak about cars – and wants a room with a projector.

  Still, if the meaning warrants it, a comma can nonetheless have a reinforcing function:

  Smith waited for a few seconds, and then began to speak.

  We can imagine an actor reading this novel-fragment aloud and pausing dramatically after seconds.

  I’ve used the word ‘short’ several times in the last few paragraphs. The question of length. This is the factor that has disturbed everyone from Lindley Murray to Henry Fowler – and which caused all of them to throw in the towel when it came to formulating rules to explain the use of the comma. The longer the clauses being connected – and especially the longer the opening one in a sequence – the more we’re likely to introduce a comma to help the reader see the overall structure of what we’re saying.

  Smith is going to speak about cars using the same pictures that he had when he gave the talk in Edinburgh last year, and Brown is going to speak about bikes.

  The function of the comma is now not so much grammatical as psycholinguistic. It gives the reader time to assimilate, time to mentally breathe.

  Length would have been one of the factors that lay behind the worry of Cobbett (and others) about taste. It’s at the heart of comma uncertainty, and we’ll see it surfacing again in the next chapter. The reason that it’s so central is because it unites the two approaches to punctuation that have dominated thinking in this field: the elocutional and the grammatical. The longer a construction, the more difficult it is to speak aloud; and the longer a construction, the more content it contains, and thus the more difficult it is to assimilate. But what one person might feel is ‘long’ another might feel is ‘not so long’. That’s where individual differences (‘taste’) enter in. One person says, ‘I need a comma to make the meaning of this sentence clear’; another finds the same sentence perfectly understandable without a comma. It’s because they have different processing abilities.

  We’re now entering the world of psycholinguistics. We know that there are limits governing how much we can hold in working memory at a time. Psycholinguist George Miller once defined it as ‘the magic number seven, plus or minus two’. The formula should be seen as no more than a guideline, as it is affected by many factors; but it’s a valuable mnemonic. And it shows how different people are.

  There’s an experiment anyone can do. Ask someone to repeat after you a list of items, such as a string of numbers between one and ten, increasing the length of the task by one each time. Say the items steadily, with a slight pause between each one:

  You

  Listener

  Four

  Four

  Six, two

  Six, two

  Eight, one, three

  Eight, one, three

  The aim is to note when the listener finds it difficult to repeat the string confidently and correctly. Many people start to have trouble at around five elements, and seven seems a top limit for most, but some can handle up to nine with ease.

  Now take the task to a more advanced level. If your listeners have failed at repeating six or seven items in a row, ask them if they would like to repeat eight items accurately. They will say yes, but wonder how. What you do is group the eight items into two sets of four, and speak them without pausing between the digits, a bit like reading out a long telephone number:

  eight one three four / two nine six one

  Most people can handle that. The rhythm and intonation help them keep such long strings in their head. And if we were to show this in writing, we could do so like this:

  eight one three four, two nine six one

  If we now look at sentences where there is variation over the use of the comma, we’ll find that the point of greatest uncertainty comes when the first clause approaches five semantic units (content words, along with any grammatical modifiers, such as prepositions or articles). There’s no problem if the opening clause contains two semantic units (I’ll underline each unit in the first one):

  It would be unusual to see a comma here. Similarly with three semantic units:

  or four:

  But uncertainty sets in when we get to five:

  And if the first sentence gets longer than this, most people will feel the need of a comma.

  Note that we can increase the length of the first sentence in other ways, such as by adding adjectives or other forms of modification.

  As we reach the ‘magic number five’, the need to breathe sets in again.

  These are adult intuitions. With children, it takes time for their working memories to develop in order to handle such complex sentences. They will therefore be more likely to introduce a comma in places where an adult wouldn’t feel it necessary. Teachers who see ‘too many commas’ in young writers thus need to be aware that the extra commas may actually be helping them express themselves, as they attempt more complex constructions, and that it takes time to develop a mature comma-using style.

  Keep all this in mind as you read the next chapter. Because everything that I’ve said here in relation to the linking of sentences applies equally when we consider the way commas are used to link constructions lower down the grammatical scale: phrases and words.

  25

  Commas, the small picture

  Grammar always offers us choices. People often ask why grammar is so complex. The answer is simple: because we want to express complex thoughts. If all we wanted to say was ‘Me Tarzan. You Jane’, it wouldn’t be complex at all. But we need to do much more than this.

  The choices are particularly apparent in a narrative where we string sentences together. We like to vary the way we say things. Only in the simplest early readers do we see repeated constructions:

  Peter can see a pig.

  Jane can see a horse.

  Mummy can see a cow. …

  The primary options for change occur at sentence ends. So, to return to the example in the previous chapter, when Hilary begins her report about Smith and decides to go for a tightly linked second part, she will use a comma, but what follows can be several grammatical constructions, such as:

  Smith is going to speak about cars, and he’ll need an hour.


  Smith is going to speak about cars, which will need an hour.

  Smith is going to speak about cars, needing an hour.

  As long as the second construction continues to act as a clause – containing a (finite or non-finite) verb – we need the comma. If we omit it, there’s a real chance of ambiguity. In the last sentence above, it’s Smith that needs the hour. But in the following version it’s the car:

  Smith is going to speak about cars needing an hour to warm up before they can be driven at speed.

  This is one of the cases where the presence or absence of a comma is nothing at all to do with taste. It depends on what you mean. Do you want to say something more about cars or not? If you don’t, you’ll keep the comma in. If you do, you’ll omit it, to show you’re thinking about cars in a more specific way.

  This ability to be specific or not is a major feature of English grammar, and shows up both in the way we speak (through our intonation) and in the way we write (through punctuation). It can make all the difference in the world. How many sisters does Mary have in this unpunctuated sentence?

  my sister who lives in China has sent me a letter

  It depends. If we punctuate it like this, she has just one sister:

  My sister, who lives in China, has sent me a letter.

  If we punctuate it like this, she has an indeterminate number of sisters: