Making a Point

by David Crystal

  In lists of three or more items, a comma should be consistently omitted or included before the final ‘and’: red, white and blue; red, white, and blue.

  And you will thus find the serial comma throughout all my later writing for that Press including the series of Cambridge general encyclopedias. But everything still depends on the publisher. The house style used by Profile Books omits them. In the introduction to my Spell It Out (2012), we read:

  we encounter a host of anomalies, variations and exceptions.

  In my manuscript, it appears as:

  we encounter a host of anomalies, variations, and exceptions.

  In the margin of the online copy-edited typescript, in bold blue under ‘track changes’, I see the message:

  Jane 23/2/12 16:38

  Deleted:,

  The death of my comma, precisely timed. (I was allowed to keep my serial comma in the present book, given its subject-matter!)

  Every writer on punctuation in the twentieth century recognized the diversity of practice. Here’s G V Carey in Mind the Stop (1939, Chapter 3) arguing that usage here is ‘a matter of individual choice’. He belongs to what he calls the ‘final comma school’, as do I, and he gives a reason:

  because the ‘no final comma’ principle breaks down now and again through ambiguity, whilst the ‘final comma’ principle can be followed consistently with less risk of it, I personally vote for the latter.

  Ernest Gowers belongs to the ‘no final comma school’. After giving an example, he says:

  commas are always put after each item in the series up to the last but one

  though he immediately acknowledges that ‘practice varies’. Interestingly, he gets the reason wrong:

  Those who favour a comma there (a minority, but gaining ground) …

  In fact, it was the other way round. Everyone from Murray to Fowler had used the serial comma. It was the omitters who were gaining ground.

  Whatever the school you belonged to, everyone agreed that there were exceptions, and that avoidance of ambiguity must be the primary rule. If it’s ambiguous to omit the comma, don’t omit it. If it’s ambiguous to insert the comma, don’t insert it. We see both cases here:

  Adding the comma helps: In the first part of the evening, the choir will sing two hymns, Old Man River and Shenandoah.

  How many items are being sung? Two or four? If you’re not sure whether the named pieces count as hymns, you have no idea. Adding a comma before the and would show that they don’t.

  Omitting the comma helps: I’m inviting my brother, a playwright, and three actors to the party.

  How many people are coming to the party? If playwright is in apposition to brother (that is, the brother is the playwright), there are five. If it’s a list, there are four. Omitting the comma before the and would push the sense towards the ‘five’ interpretation.

  Any textbook on punctuation will draw your attention to examples like these. Their value is that they show total consistency to be impossible, when it comes to the serial comma, because of the complexity of English syntax. And we can generalize this point to the use of the comma in other syntactic domains. Even quite basic rules might have exceptions. Take the one I described in the previous chapter: no comma between subject and verb or between verb and object. That’s a fairly strict rule today, but there are still exceptions. If a clause uses two instances of the same word, we see writers using a comma to help remove a source of possible confusion:

  Whatever his name is, is of no concern to me.

  And when commas are used in direct speech, there may be a comma between the verb and an object:

  She said, ‘I’m in the front room.’

  The only rule that never has any exceptions is the one forbidding a comma within the elements of a phrase that has no serial content. We never find:

  the, car

  her beautiful, dress

  I will, go

  in, the garden

  The only possible way to interrupt a phrase is if we include another unit within it, and then we need a different convention involving pairs of marks:

  This is her beautiful (I do hope you agree) dress.

  There are also some strictly controlled specialized settings where individual conventions apply, such as in bibliographies and indexes, where the comma separates surnames and first names (Crystal, D.). In mathematics it separates thousands, as in 14,236 (and in some places, decimals, see p. 136). Even here, though, we may see variation, such as the replacement of the comma by a space.

  Exceptions and ambiguities exist, but we mustn’t exaggerate their importance as a guide to general practice. With the serial comma, for every one case of possible ambiguity, there are ninety-nine where there’s no ambiguity at all, and we are faced with a straight pragmatic choice. As Ernest Gowers says:

  The correct use of the comma – if there is such a thing as ‘correct’ use – can only be acquired by common sense, observation, and taste.

  And, I would add, a linguistic perspective.

  Interlude: Pun-ctuation

  Punch could never resist a pun, and punctuation was no exception. In No. 16 for 1849 we read of a new proposal under the following heading:

  STREET PUNCTUATION

  We understand that an attempt is to be made to introduce a new system of Punctuation, on the principle of street stoppers, or street stops.

  This will enable the publishers of School-books to bring out an Illustrated Work on Punctuation, in which a comma may be represented by an unac-comma-dating cabman, who, by refusing to move on, occasions a slight pause in the progress of traffic; while a coal-waggon at a stand-still, would be very fairly emblematical of a co(a)-lon or semi-co(a)-lon, as the case may be. An advertising van, would convey a good idea of a point, a dead stand-still, or full-stop. Notes of admiration could easily be shown by the astonished foot-passengers: notes of exclamation by the Omnibus-drivers in full cry at the impediment: and notes of interrogation, by the policemen inquiring why the drivers do not move on.

  The traffic jams in Fleet Street (where Punch had its office) were a daily source of complaint.

  27

  Hy-phens

  Any book that set out to cover every eventuality in punctuation would be unpickupable. I’ve therefore focused on the general principles behind punctuation and how to deal with the problems that feature frequently in its use. It has taken me three chapters to explore just the main uses of the comma. If I were to cover all the ways in which the comma can be used, I would end up having to write an entire grammar of English from a punctuational perspective, as every construction has to be punctuated, and commas operate at all levels of syntactic organization. Similarly, if I were to cover all variations in the use of the hyphen, I would have to write an entire dictionary, because each compound word has its own story. It is the most unpredictable of marks. Henry Fowler sums it up well in the opening sentence of his entry on hyphens in his Dictionary of Modern English Usage: ‘chaos’.

  The hyphen has an ancient history (the name is from an Ancient Greek word meaning ‘together’), and the mark that it describes can be seen throughout the manuscript era in several functions, but usually to draw a reader’s attention to the way words, or parts of words, separated through some accident (such as being unable to fit at the end of a line), needed to be brought together. In the sixteenth century, John Hart emphasized this linking role in his term for the hyphen: a ‘joiner’. Hyphens became increasingly common in print from the 1570s, even though the term itself isn’t recorded in English until the early 1600s.

  We then see the hyphen becoming a standard feature of English orthography, with the same two broad functions that it has today: as a divider, to show a word-break at the end of a line, and as a linker, to mark the unity of a word containing distinct grammatical elements. But it took a while for a consistent practice to develop. It’s clear from the way it is sprinkled around the publications of the Elizabethan period that there was a great deal of uncertainty over its use. For examp
le, in Shakespeare’s First Folio we see hyphens that we would never tolerate today, such as for-sake, a-gaine, and yon-der. And any word sequence that was felt to have a tight semantic link might be hyphenated, such as red-plague and for her wealths-sake.

  With just two main functions to perform, we might expect hyphenation to be a straightforward matter. But repeatedly we see the usage pundits keeping their distance. Here is Ernest Gowers’s elegant disclaimer: ‘If I attempted to lay down any rules I should certainly go astray, and give advice not seemly to be followed.’ He is alluding to a style book written for Oxford University Press by John Benbow in which we read: ‘If you take hyphens seriously you will surely go mad.’

  Nonetheless – and allowing for this possibility – we do need to give hyphens some serious consideration, especially as both functions give rise to variation and change. The hyphen that identifies a word-break at the end of a line is the easier matter to deal with. It presents few problems once we recognize that it’s located at a syllable boundary and not after a syllable fragment. Monosyllabic words are straightforward: they are never broken. We don’t see str-ing or go-ne. (The convention isn’t self-evident, as is shown by the way children make such breaks when writing their first stories. It has to be taught.) The only usage issue relates to differences between British and American practice over polysyllabic words.

  British practice is to follow the way a word divides grammatically or etymologically; American practice is to follow the way the word sounds. So if we see bio-graphy and philosophy, we can guess we are reading a British book following Hart’s original rules; if biog-raphy and philos-ophy, it must be American (as in Webster’s dictionaries) – though during the twentieth century some British dictionaries began to follow US practice, especially those aimed at a foreign learners’ market. The issue was never serious, as most words divide in the same way (both principles would show pun-ish, for example) – and in any case it’s largely of historic interest now, as we hardly ever see hyphenated line-breaks today, thanks to sophisticated typesetting on paper and software-governed typesetting on screen.

  Word division at line-endings never bothered the usage pundits much, except when it gave rise to a facetious miscue (such as the-rapists). They left such matters to the printers. Henry Fowler devotes five pages in his Dictionary to examples of correct and incorrect hyphenation, and doesn’t mention line-breaks at all. He, as everyone else, was entirely focused on the way hyphens are to be used, or not used, in word identification, where there are two central issues. Do we use a hyphen to separate the elements in a compound word? And do we use one to separate a prefix from the rest of the word?

  There are three options if we decide to bring two (or more) elements together to make a compound word. We can write flower pot (spaced), flower-pot (hyphenated), or flowerpot (solid). These are the choices, bringing together two levels in the punctuation hierarchy: the word space and the hyphen. They are now in competition with each other. To space or not to space? That is the question. And if we decide on the latter: to hyphenate or not to hyphenate?

  Most writers have no idea. They may have an instinctive preference or a long-standing habit for writing a word in a particular way, but they know that their choice may not be shared by other writers. And instinct never guarantees consistency. That’s one of the reasons publishers employ copy-editors: they will check that a writer has hyphenated consistently and in accordance with the publishing-house style. And because publishers make different decisions over which words should be hyphenated, authors – especially those who have written for a variety of houses – are not going to have a clear intuition about what to do, and will generally be happy to leave that to the editors.

  Even if you’ve spent the whole of your authorial or journalistic life with a single publisher, your intuition may let you down. This is because there are two factors influencing publishing practice: tradition, which fosters a firm’s identity, and language change, which affects everyone. The concern for identity that we saw in the case of the serial comma is seen again here. British publishers opt for the hyphen much more than American houses do. Canadian and Australian publishers sometimes follow British practice, sometimes American. And within a country, there may be various policies. Oxford and Cambridge style guides make different decisions, for example. I normally write no-one. When in 2008 I put this in a book for CUP, the copy-editor changed it to no one. When in the same year I put no one in a book for OUP the copy-editor changed it to no-one. And changing language fashions can create havoc with style guides. In a later book for OUP my no-one was changed to no one.

  Changes in fashion are the main reason why the obvious solution to any question about hyphenation – look it up in a dictionary! – won’t always help. Look up flower-pot in the online Oxford English Dictionary and you’ll see the heading: flower-pot | flowerpot, without further comment. It’s left up to you. And if you were in the habit of using the Shorter Oxford Dictionary, and had internalized its recommendations, you would have had a real shock in 2007, when the sixth edition was published and you saw that around 16,000 items had had their hyphens removed. Most of the changes had the hyphen replaced by a solid setting (pigeon-hole > pigeonhole, cry-baby > crybaby, bumble-bee > bumblebee), but quite a few ended up spaced (test-tube > test tube, ice-cream > ice cream, hobby-horse > hobby horse). Reactions ranged from the hysterical to the bemused. Some observers called it ‘hyphengate’.

  Why did the editors do this? They are partly reflecting changes in fashion. The chief editor, Angus Stevenson, commented at the time that the dictionary is reflecting the current dislike of the hyphen on the part of designers: ‘The hyphen is seen as messy looking and old-fashioned.’ And indeed, his view is reflected across the dictionary world. Certainly, most people would agree that there’s something clean and modern about the first of the following alternatives (6th edition) compared with the conservative appearance of the second (5th edition):

  If you buy too much ice cream with your pin money you’ll end up with a pot belly.

  If you buy too much ice-cream with your pin-money you’ll end up with a pot-belly.

  Apart from anything else, it’s far easier to type the first version in the fast-moving Internet world. The position of the hyphen on a keyboard doesn’t help.

  But there’s more to it than the pragmatic factors of fashion, aesthetics, and ergonomics. We still have to explain why some words have ended up spaced and some have ended up solid. Why is it ice cream and not icecream? And why bumblebee and not bumble bee? We might imagine it would be even faster to leave the space out. The reasons are all to do with semantics – the ease with which we visually perceive the meaning and the familiarity we have with the concept.

  Legibility is affected when unfamiliar sequences of letters come before the eye, or the letter-sequence distracts us by making us think of an irrelevant word. So, for example, we are unlikely to find arrowworm or lookingglass because of the unfamiliar juxtaposition of the middle consonants, and both of these are indeed spaced in the new Shorter Oxford. Similarly, pairs of vowels can cause momentary uncertainty, so we are less likely to encounter deicer, moreish, goahead, or toing and froing, and never antiinflation, belllike, or freeenterprise. Irrelevant associations arise if we were to write weeknight, wall-eyed, and tearoom. Length can also be a factor. Readers don’t like short elements (less than three letters), so when email first arrived it was spelled e-mail – never e mail. Similarly we see U-turns and H-bombs. Compounds that become lengthy (around ten letters or more) also prompt separation – so commonsense, goodlooking, and backprojection are often avoided. And we see a hyphen routinely when the typography is unusual, such as when the base of the word begins with a capital letter (pro-English), a numeral (post-2000), or a punctuation mark (un-‘hip’).

  But note the qualifications – ‘usually’, ‘less likely’, ‘unlikely’, ‘often’, ‘routinely’. These are tendencies, not rules. And any of these tendencies can be overruled by the factor of familiarity, which
explains a great deal of the variation in the history of hyphenation. A compound word is a semantic unit as well as a grammatical one: the meaning of the whole is different from the sum of the parts. But it takes time for the meaning of a new compound to become familiar, and this is reflected in changes in orthographic practice, as shown by looking at the historical citations in the OED. Typically, the first recorded examples are spaced: the two elements retain a trace of their separate identities. Then, as people get used to the new concept, we find the compound hyphenated as well as spaced. And eventually, once the original meanings have been lost sight of, we see it written solid. So, for example, we find pigeon holes in the sixteenth century, pigeon-holes predominating in the seventeenth, and pigeonholes growing in frequency in the twentieth. ‘For sale: a type writing machine’, says an ad in 1881, and soon after we get type-writing and then typewriting. As email became more familiar, it began to drop its hyphen, and email is now the most common form. Again, the underlying principle is a trend, not a hard-and-fast rule; but it makes sense. And it explains why it’s necessary, every now and then, for a dictionary to change its recommendations. It is simply trying to reflect the new ways in which people are thinking of the words.

  Here’s a more detailed example of the way practices change. It’s standard now to spell today, tomorrow, and tonight without a space or hyphen. But when the words first arrived in Old and Middle English they were seen as a combination of preposition to followed by a separate word (dæg, morwen, niht), so they were spaced. This usage was reinforced by Dr Johnson, who listed them as to day etc in his Dictionary (1755). But people began to think differently in the nineteenth century, and we see the big new dictionaries (such as Worcester’s and Webster’s) hyphenating the words. People began to get fed-up with this in the twentieth century. Henry Fowler came out against it in his Dictionary of Modern English Usage (1926):