I Request You to Read This Post

Several weeks ago, I tweeted about a weird construction that I see frequently at work thanks to our project management system. Whenever someone assigns me to a project, I get an email like the one below:

I said that the construction sounded ungrammatical to me-you can ask someone to do something or request that they do it, but not request them to do it. Several people agreed with me, while others said that it makes sense to them if you stress you-they requested me to work on it, not someone else. Honestly, I'm not sure that stress changes anything, since the question is about what kind of complementation the verb request allows. Changing the stress doesn't change the syntax.

However, Jesse Sheidlower, a former editor for The Oxford English Dictionary, quickly pointed out that the first sense of request in the OED is "to ask (a person), esp. in a polite or formal manner, to do something." There are citations from around 1485 down to the present illustrating the construction request [someone] to [verb]. (Sense 3 is the request that [someone] [verb] construction, which has been around from 1554 to the present.) Jordan Smith, a linguistics PhD student at Iowa State, also pointed out that The Longman Grammar says that request is attested in the pattern [verb + NP + to-clause], just like ask. He agreed that it sounds odd, though.

So obviously the construction has been around for a while, and it's apparently still around, but that didn't explain why it sounds weird to me. I decided to do a little digging in the BYU corpora, and what I found was a little surprising.

The Corpus of Historical American English (COHA) shows a slow decline in the request [someone] to [verb] construction, from 13.71 hits per million words in the 1820s to just .2 per million words in the first decade of the 2000s.

And it isn't just that we're using the verb request a lot less now than we were two hundred years ago. Though it has seen a moderate decline, it doesn't match the curve for that particular construction.

Even if the construction hasn't vanished entirely, it's pretty close to nonexistent in modern published writing-at least in some parts of the world. The Corpus of Global Web-Based English (GLoWbE) shows that while it's mostly gone in nations where English is the most widely spoken first language (the US, Canada, the UK, Ireland, Australia, and New Zealand), it's alive and well in South Asia (the taller bars in the middle are India, Sri Lanka, Pakistan, and Bangladesh). (Interestingly, the only OED citation for this construction in the last fifty years comes from a book called World Food: India.) To a lesser extent, it also survives in some parts of Africa and Southeast Asia (the two smallish bars at the right are Kenya and Tanzania).

It's not clear why my work's project management system uses a construction that is all but extinct in most varieties of English but is still alive and well in South Asia. The company is based in Utah, but it's possible that they employ people from South Asia or that whoever wrote that text just happens to be among the few speakers of American English who still use it.

Whatever the reason, it's an interesting example of language change in action. Peter Sokolowski, an editor for Merriam-Webster, likes to say, "Most English speakers accept the fact that the language changes over time, but don't accept the changes made in their own time." With apologies to Peter, I don't think this is quite right. The changes we don't accept are generally the ones made in our own time, but most changes happen without us really noticing. Constructions like request that [someone] [verb] fade out of use, and no one bemoans their loss. Other changes, like the shift from infinitives to gerunds and the others listed in this article by Arika Okrent, creep in without anyone getting worked up about them. It's only the tip of the iceberg that we occasionally gripe about, while the vast bulk of language change slips by unnoticed.

This is important because we often conflate change and error-that is, we think that language changes begin as errors that gradually become accepted. For example, Bryan Garner's entire Language Change Index is predicated on the notion that change is synonymous with error. But many things that are often considered wrong-towards, less with count nouns, which used as a restrictive relative pronoun-are quite old, while the rules forbidding their use are in fact the innovations. It's perverse to call these changes that are creeping in when they're really old features that are being pushed out. Indeed, the whole purpose of the index isn't to tell you where a particular use falls on a scale of change, but to tell you how accepted that use is-that is, how much of an error it is.

So the next time you assume that a certain form must be a recent change because it's disfavored, I request you to reexamine your assumptions. Language change is much more subtle and much more complex than you may think.