Monthly Archives: October 2012

A case study in self-prescriptivism

Caution: Those of you who are not me might not find this interesting.

I’ve been thinking about the that-rule. Last year, some wise bloggers convinced me that it was perfectly fine to use ‘which’, as well as ‘that’, to introduce a restrictive clause, and since then I’ve been happily getting by without the rule. There’s nothing wrong or unclear about ‘this is the house which Jack built’. I still usually prefer ‘that’, but only on aesthetic grounds, and I have no qualms with anyone whose tastes differ.

I only found out about the rule in my mid-20s, when I was getting into editorial work and wanted to make my linguistic instincts more systematic and explicit. In the course of my reading, I came across the that-rule (among others). It seemed to make sense, so I made it part of my arsenal.

Silly boy. Oh well.

But I’ve been wondering: why did it seem to make sense to me?

To find out, I looked at my undergraduate dissertation: 15,000 words of my 22-year-old self’s cleverest writing. How did pre-editorial Tom handle his restrictive clauses?

Excluding the ones that were introduced by a preposition (these are cases in which you unarguably need to use ‘which’), I found 54 restrictive clauses: 42 used ‘that’ and 12 used ‘which’.

Then I looked a little more closely and found something more interesting. When the subject of the restrictive clause’s main verb was plural, I was more likely to use ‘which’. For example: “There are nonrepresentational aspects of experience which are not introspectible.” (It was philosophy of mind. Please don’t ask.)

95% of the singulars used ‘that’ but 63% of the plurals used ‘which’.

So, while I had clearly developed (without realising it) a strong preference for restrictive ‘that’, I’d also developed a moderately strong exception to it for plurals.

Maybe this had something to do with my knowledge of the demonstrative role of ‘that’ – you can say ‘look at that dog’ but not ‘look at that dogs’ – but I’m just guessing. I wonder whether this distinction is common in general usage, or whether it was just my own oddity. Does anyone know?

Anyway, when I came across the that-rule, it must have seemed right partly because it chimed with my majority practice. But by adopting it consciously, I let it override the exception: since then, I’ve used ‘that’ for singular and plural alike. (I just searched a few months’ worth of blog posts from my late 20s: not a single unprepositioned restrictive ‘which’.)

And even though I’ve now junked the rule, the exception hasn’t survived. Following the rule for several years has shifted my aesthetic sense of what feels right – even though I no longer believe that there is a right and wrong here.

You could strike a wistful note, about how a pointless rule diminished the complexity of my language. But I don’t think this particular nuance was one of the ones that add any real richness.

Free reign or free rein? My kingdom for a horse!

This sort of thing is quite common:

Recruiters who continued to meet their quotas were given almost free reign, they say.

It should be ‘free rein’. The metaphor is horse riding, not monarchy. If you give a horse a free rein, you let it wander where it likes – the opposite of keeping a tight rein on it (see also ‘rein in’ and ‘hand over the reins’).

But ‘free reign’ also has a logic to it: to reign freely is, obviously, to have more power and independence than to reign with checks and balances. However, ‘free reign’ doesn’t have a record of being literally used to describe an unconstrained monarch. This logic is just rationalising a homophone.

In fact, the earliest uses recorded by the OED, from the 17th and 18th centuries, are in the plural form, ‘free reins’. This makes plenty of sense for a horse, while ‘free reigns’ would make very little sense for a monarch. The Google Books Ngram shows the singular form being roughly twice as common as the plural for much of the 19th century, then shooting ahead from about 1875.

‘Free reign’ popped up occasionally during the 19th century but it only really edged into noteworthiness around 1930. Since then, it’s grown more and more common:


In 1950, ‘free rein’ outnumbered ‘free reign’ 25 to 1; by 2000, its lead was just 5 to 1.

But ‘free reign’ is still generally regarded as nonstandard.

Robert Burchfield says that it is a mistake. Bryan Garner rates it a stage 2 usage (‘widely shunned’). And Merriam-Webster, despite its limp conclusion that “If you are not sure whether you want rein or reign, your dictionary will help you,” says that ‘free reign’ is the result of a confusion. It blames the decline of horse riding for the misunderstanding of rein-based expressions.

So it’s about horses, not monarchs. There’s a good mnemonic that can help here, thoughtfully provided by William Shakespeare:

A horse, a horse, my kingdom for a horse!