The which-hunting season is upon us again.

A minor spat about the use of which and that in relative clauses

Which-hunting, a play on words on witch-hunting (see link below), is a term used by linguists to refer to the criticism by some of the use of which for things, in defining (aka restrictive or identifying) relative clauses.

To follow this discussion you need to know the difference between defining and non-defining relative clauses. If you need reminding, see my post here.

The standard rule for relative pronouns as taught in EFL books

In defining relative clauses we can use who or that to talk about people, and or which or that to talk about things:

• For people:
• - The boy who is climbing the tree is my son.
• or The boy that is climbing the tree is my son.
• For things:
• - The car which is coming up the hill is my father's.
• or The car that is coming up the hill is my father's.

But in non-defining clauses, we can only who and which.

• - My son, who loves climbing trees, is the one in the blue jumper.
• - That red sports car, which can go incredibly fast, is my father's.

A little history - or how a suggestion became (for some) a cast-iron rule.

As I understand it, the freedom to use which or that for things and who or that for people in defining relative clauses, has been standard practice in English usage for the last few hundred years. And indeed many respected writers used them almost interchangeably.

But in A Dictionary of Modern English Usage (1926), one of the most influential style guides of all time, Henry W. Fowler (1858–1933) wrote (p713):

Let it be stated broadly, before coming to particular dangers, that; of which and that, which is appropriate to non-defining & that to defining clauses

A point he, together with his brother F.G. Fowler, had put forward in a previous book, The King's English (1906)

Although this advice appears to have been largely ignored in his native Britain, it was taken up with some enthusiasm by certain style guides in America, especially those used by newspapers, such as The Associated Press Style Guide.

It was also repeated in the edition of William Strunk Junior's Elements of Style (first published in 1918) revised by E.B.White in 1959, and so commonly known as Strunk and White, a book which has been very influential with several generations of American English teachers.

Writing in the Chronicle of Higher Education, linguistics professor Geoffrey K. Pullum noted:

The copy editor's old bugaboo about not using "which" to introduce a restrictive relative clause is also an instance of failure to look at the evidence. Elements as revised by White endorses that rule. But 19th-century authors whose prose was never forced through a 20th-century prescriptive copy-editing mill generally alternated between "which" and "that." (There seems to be a subtle distinction in meaning related to whether new information is being introduced.) There was never a period in the history of English when "which" at the beginning of a restrictive relative clause was an error.

This seemed to be a largely American discussion; elsewhere Pullum has written (my underlining):

The reform did not catch on among users of English. When President Roosevelt described December 7 as "a date which will live in infamy" he was using ordinary grammatical English, not making some kind of a slip. But, unfortunately, almost all copy editors in the USA are under instructions to enforce Fowler's stupid rule (British editors wouldn't dream of it)... .

In steps The Guardian

This doesn't apply to all British editors apparently. The Guardian (a UK newspaper) has just published an article 'That's the way to do it' in their Mind Your Language blog which regurgitates this non-rule, suggesting that the use of which in defining relative clauses, instead of their preferred that, is an example of 'hypercorrection which can make their prose sound pompous'.

A couple of observations:

• Most importantly, while using or omitting commas correctly in his examples, the author only mentions them very briefly rather than saying clearly that the use (or not) of commas is the main distinguishing feature (apart from context) between defining and non-defining clauses in writing, rather than the use of that or which.
• He doesn't address the question of who for persons. If it is so important to use different pronouns for things in defining and non-defining clauses to avoid ambiguity, why not for people. Perhaps we should also specify that for people in defining clauses. But the same commentators who insist on that for things, are often against the use of that for people. So their argument is not very consistent.

I was pleased to see that the writers of two linguistics blogs, Arrant Pedantry and Sentence First were among the first commentors, putting the other side of the story. They have also gone into a lot more detail on their blogs, and have been supported by other linguists, like Arnold Zwicky. (See below for links)

Reams have been written about this subject at the top linguistics site Language Log. If you are interested, see the link below.

The Guardian Style Guide

Most of the examples in that article are drawn from the Guardian Style Guide, which states baldly:

that or which? - This is quite easy, really: "that" defines, "which" gives extra information (often in a clause enclosed by commas)

And they give these examples:

• This is the house that Jack built; but this house, which John built, is falling down.
• The Guardian, which I read every day, is the paper that I admire above all others.
• I am very proud of the sunflowers that I grew from seed. (some of the sunflowers)
• I am very proud of the sunflowers, which I grew from seed. (all the sunflowers)

They then say:

Note that in such examples the sentence remains grammatical without "that" (the house Jack built, the paper I admire, the sunflowers I grew), but not without "which"

Apart from the insistance on this non-rule, I can see two problems here. Extra information (non-defining clauses) is always enclosed in commas, not simply often.

And that last sentence implies that the that in defining relative clauses can always be left out, making no mention of the fact that this is only true when it represents the object of the following verb. It would not be true of the sentence, 'This is the house that fell down.' for example.

How two contemporary style guides see it

New Fowler's Modern English (1996 edition) - to show the difference beween types of relative clause, the author gives two examples from Anita Brookner's A Family Romance (1993). Both the non-defining example and the defining one use the relative pronoun which. And although he refers with some approval to the ideas put forward in the original Fowler's, he admits 'it would be idle to pretend that it is the practice either of most or of the best writers'.

Merriam-Webster Dictionary of English Usage - Concluding quite a long entry, they say:

You can use which or that to introduce a restrictive clause - the ground for your choice should be stylistic - and which to introduce a nonrestrictive clause.

Links

• The Guardian - Mind Your Language
• The Guardian Style Guide
• Witch-hunt at Wikipedia

Links to linguistics blogs etc

These are quite long and detailed articles, although the language is relatively non-technical. I include them mainly for reference purposes.

• Arrant Pedantry - Click on his second diagram for a useful table of relative pronoun use. (Careful, the first diagram is the prescriptive approach)
• Sentence First - Very readable at the beginning, with a good introduction to the different type of relative clause. Goes into a lot of detail later.
• Arnold Zwicky
• Language Log - Geoffrey K. Pullum
• Arnold Zwicky - List of articles at Language Log
• Fowler - at Google Books
• Fowler - The King's English - at Bartleby Books