Can you have a comma before because?

« previous post | next post »

I got a message from a former teacher who said her friend had sent her my article about Strunk and White and it had stimulated her to ask me the following question:

For 31 years, this is the rule I taught to all of my elementary school students: do not put a comma before "because." Since I noticed that you did so at least twice in your article, I am wondering if I taught the students incorrectly (I hope not) or rather if Scots follow another rule (I hope so). I'd really like to know.

Oh, dear. The problem was not how to answer the question; the problem was how to do so kindly and gently. I did not do well enough

I did at least quell my first instincts, stifling all sorts of imaginary openings ("What the hell is wrong with you…"; "I pity the kids whose lives you have blighted…"; "Nobody in their senses could believe . . ."; "I'm not a Scot, and this isn't about ethnicity, you stupid . . ."). But even after I expunged all those from my head, the reply I sent nonetheless began: "The rule is bunk."

I did explain something about the evidence after that; but I added, "I'm afraid you taught those kids a fictive rule that is of no real use to man or beast, and could do actual harm to people's writing". I added that "correcting" sentences by removing commas before because would be likely to make the writing worse, not better.

I could have done better than telling her straight out that the rule she taught was bunk. It is bunk, of course; but heck, couldn't I sugar the pill to help the medicine go down? Apparently not.

I want to be a better person. (I want to be the sort of person Barbara was, actually, with that amazing combination of critical eye and gentle heart, though that is setting the bar rather high.) I'd like to explain things kindly to the hidebound victims of grammatical fascism nice well-meaning people who write to me out of the blue, and show them what the evidence is, and do it all in a way that is charitable and kind and helpful. So let me start again. Here's how I should have put it.

Dear reader:

Thank you so much for asking me this intelligent and interesting question. It is rare to find people who have the intellectual fearlessness to reflect seriously on the question of whether they might be wrong in some deeply held belief — still rarer when they taught it for many years. You must have been a really good teacher.

The answer to your question can be very easily discovered by doing a few simple experiments. Anyone who has a Macintosh computer running OS X and can pull together a few plain text files of prose that they consider well written can do experiments of the right kind. A command like this to the prompt in the Terminal program:

      grep ', because' textfile

will produce as output all the lines in textfile that have a comma followed by a space followed by because. And if you leave the comma and the space out, it will give you all the lines that have the word because. In fact you can do better than that. A tiny change, adding "-c", will give you a command line that actually counts the lines that have because preceded by a comma, without showing them all to you:

     grep -c ', because' textfile

That is just about all you need to get started. Using simple commands just like these I did a quick-and-dirty search on 44 million words from The Wall Street Journal from 1987 to 1989 that I happen to have on my laptop (it's a widely and cheaply available set of files that were made accessible to linguists for this sort of purpose, under an initiative spearheaded by Mark Liberman, back in 1993). I searched for instances of because that had a preceding comma, and found 6,354 of them. It is true that the cases that aren't preceded by a comma outnumber these (there are over 41,000 comma-free because-phrases), but the ones with preceding comma constitute some 13 percent of the total. That's far too many for us to imagine that each one was a slip-up that the writer didn't spot and the subeditors at the WSJ didn't catch. The notion that a comma before because is a mistake begins to look just too implausible.

When I looked at a very small collection of fine novels and plays from about a century ago that I happened to have knocking around, the figures were a lot closer — more of the "because" occurrences had commas: there were 246 with no comma, and 118 with a comma. So the cases with the comma are a huge percentage of the total — almost a third.

I suggest, therefore, that there is no comma-forbidding rule at all, and students should never be taught that there is. Instead, there is a well-known distinction in educated written Standard English between heavy and light punctuation styles. A great deal about placement of commas is left to the writer's discretion. (Not everything: the rule that you do not put a comma between subject and verb is fairly strict.)

I tend toward heavy use of commas myself. If I were writing a sentence that said this:

If anyone thinks eggs bacon or sausages are bad for your health they should in my opinion look at the epidemiological evidence because many people like my father have reached their late eighties having breakfasted on such foods every day or just about every day.

I'd probably use quite a few commas in it:

If anyone thinks eggs, bacon, or sausages are bad for your health, they should, in my opinion, look at the epidemiological evidence, because many people, like my father, have reached their late eighties having breakfasted on such foods every day, or just about every day.

I like all of those commas. Hardly any are strictly required (only the first is mandatory, I think), and none are forbidden by the rules of grammar. A good writer will think about the structure and rhythm of a written sentence, and make decisions about where to place the very slight pause for thought or breath that a comma intuitively signals, or to separate words or phrases that ought not to be run together.

Sometimes a comma before because is almost mandatory, underlining the fact that it certainly should not be forbidden. The comma can remove a nasty ambiguity. Take these examples:

[1]    I didn't marry Bob because I wanted a stable home life.

[2]    I didn't marry Bob, because I wanted a stable home life.

Version [1] is best if the meaning is that you did marry Bob, despite knowing it was going to be a wild, unstable, and sometimes tempestuous relationship. Version [2] is best if you didn't marry Bob, and the reason was precisely that you wanted a stable home life.

The bottom line is that (I'm sorry to tell you) you were teaching your students a rule that is clearly not valid. Obeying it blindly would do actual harm to good writing (so let's hope your students all took you with a pinch of salt). To "correct" [2] to make it the same as [1], for example, would be a grave mistake.

Note carefully that I am NOT taking the view that if you can find any instances of something then it's grammatically OK. Many people write quite badly. In 44 million words there will be quite a few typographical errors, and some grammatical slips as well. But what I'm saying is that when you find, say, tens of thousands of relevant instances and an eighth of them have some property P, that's way too many for anyone to suggest that property P signals ungrammatical status. Journalists do commit occasional slips; but not thousands and thousands and thousands of systematically similar and easily detectable ones, in a mere 44 million words.

The plausibility of calling something an error is high if you just find one case, or the author was inexpert, or nobody else does the same thing, or the people who read it couldn't make head or tail of it, or the author himself didn't like it and later changed it. But when large numbers of expert professional writers write huge numbers of sentences with property P, and their editors let them through, and readers don't find themselves balking as they read... Under those conditions, the plausibility of saying it's a mistake to have property P drops away to zero. You have to be sensitive to the possibility that sporadic occurrences in print might be tricking you, yes. But there is no serious possibility of that here.

Let's take a look at the first sentence in Alice's Adventures in Wonderland that has a because preceded by a comma, whatever it is (and I'm literally picking a novel at random and suggesting we look at the first case; I could just as easily have suggested the seventh or the fifteenth):

First, because I'm on the same side of the door as you are; secondly, because they're making such a noise inside, no one could possibly hear you.'

Here we actually have TWO cases for the price of one (which indicates that my figures above are only approximately correct: whenever there were two cases on one line, it would only have been counted as one). Now, do you really feel inclined to say Lewis Carroll couldn't write? Does anyone really feel that removing his commas was a change that some editor should have applied to his manuscript with a red pencil? No. And that should come close to settling the matter. Or if not, a hundred similar cases should.

(You might like to note one other thing in passing, by the way: there's no main verb in Carroll's sentence! I wonder how many times English teachers tell their students that every sentence must have a tensed main verb? Most do; but not all. Not this one, for example. The rules are subtle!)

The take-home lesson that I wish we could send to all your past students is not just that there are areas of grammar where the rules don't define a simple cut-and-dried answer that holds in every case; it's that although grammar is thought of as being like religious teaching on morality — where you get told by experts about the content of the eternal rules that you should obey — it is in fact more like a science discipline like chemistry or biology. The grammatical world is complex, and not to be oversimplified or approached through a haze of prejudice or dogma, but you can do experiments to find out what the world is like. That's the key point that linguists are concerned to make about studying grammar. Without (I hope) throwing away the sensibilities of the humanities, they apply to the study of grammar some of the methods of science.


Geoff Pullum

That's how I should have done it. Of course, it would have taken an hour or so, and I get several such messages a day. And there is of course a place that a serious student of the grammar of written English could go for a thorough treatment of the function and distribution of commas: chapter 20 of The Cambridge Grammar of the English Language. I can't spend all day providing a full-time unpaid grammar-question answering service in addition. Though I would if I could.

[Comments are closed, because . . . There, you see the comma? I did it again!]

Comments are closed.