Thank heavens for Michael Rosen

It's good to see that sometimes, an article about 'grammar' can appear in the media and it's not a complete load of codswallop. Michael Rosen, who I already knew to be an intelligent and sensible chap, has performed this feat in the Guardian. Sadly, the sub-ed has added a headline which rather over-simplifies his point (I suppose that's the sub-ed's job) and this has led to numpties in the comments who've missed the point. 

The article is called 'Sorry, there's no such thing as correct grammar', and responds to the fact that some bloke called Martin Gwynne is giving 'grammar lessons' in Selfridges, reported in the Telegraph. The Telegraph commenters are even more of a bunch of mud-for-brains, immediately leaping on spelling and punctuation errors of their fellow commenters:

How can I criticize the grammar in this blog post? Only on minor points ... there's no need for a hyphen between an adverb and an adjective when there's no possible ambiguity (easily-accessible; privately-published). 

"How can I criticize the grammar in this blog post?" By going to America. Here in the UK, old bean, it's "criticise".

 Anyhoo, back to the Rosen piece. What is refreshing is that he says this:

Whereas linguists are agreed that language has grammar, what they can't agree on is how to describe it. So, while there is a minimum agreement that language is a system with parts that function in relation to each other, there is no universal agreement on how the parts and the functions should be analysed and described, nor indeed if they should be described as some kind of self-sealed system or whether they should always be described in terms of the users, ie those who "utter" the language, and those who "receive" it (speakers and listeners, writers and readers etc).

It would be nice if he made us sound a little bit less like an argumentative rabble who can't sort out what their own discipline is about, but he's broadly right. His main point is his last one, and he seems to be firmly of the opinion that language should always be described in terms of its users (he and I differ on this point, but we won't fall out over it).


He also says, quite rightly, that

In fact, we would neither be able to speak nor understand if we didn't know [grammar]. 

He also notes (again, quite accurately) that there are two meanings of 'learning grammar': 

whether we get to know grammar in order to be "correct", or in order to describe what people say and write.

Of course, what this Gwynne person is teaching (though the Telegraph doesn't say) will certainly be the former. As Rosen says,

People attending his classes will feel immensely pleased that they have been told what's right and will probably spend a good deal of time telling other people they meet or read where and how they are wrong.

And he is very down on this, arguing quite vociferously that it's not right to allow 'grammar' to become something that belongs to an elite, leaving the others to feel that they are doing it wrong. He makes what seems to be a proposal for teaching everyone how to write and speak standard English:

If we are serious about enabling those who want to acquire what we have called standard English then first we should be honest about change and its lack of encoded rules. Then, together with them, we should look closely at how such people's speech and writing diverges from the kind of English that they would like to acquire. There will always be social reasons for this and knowing these helps people take on the dialects they don't fully speak or write.

I think he's saying that rather than tell people 'you're wrong, this is the right way', we should instead accept the differences but ensure people know that they are non-standard. He is arguing against a 'grammar police' and suggesting that language should be more open, that the reasons for it varying should be known, and that variation should be embraced. I think he's saying that everyone should study linguistics. Well done that man.


He does also say this, however, which seems to contradict his point, which is that we can teach people standard English, and also be basically wrong (emphasis mine):

many people... imagine that because it is called standard, it is run by rules and that these rules are fixed. I've always understood rules to be regulations that are drawn up in some agreed list. They are fixed (until such time as they are amended) and they are enforceable. In fact, there is no agreed list, a good deal of what we say and write keeps changing and nothing is enforceable. Instead, language is owned and controlled by everybody and what we do with it seems to be governed by various kinds of consent, operating through the social groups of our lives. Social groups in society don't swim about in some kind of harmonious melting pot. We rub against each other from very different and opposing positions, so why we should agree about language use and the means of describing it is beyond me.

Well. No agreed list, all right. No one sat down and said 'let's put the subject first, and let's have objects following the verb. We'll have a negation particle, I think, and let's form questions by changing the word order'. But as I've said repeatedly, language is run by rules. It simply is a set of rules - that's pretty much its definition. Not in the sense of rules that have to be learnt and followed, but more like instructions. Like a computer programme: it comprises a set of instructions, which it follows in order to run. If the rules are wrong (altered in some way, like missing the final bracket off an html command), the programme crashes. We actually use the word 'crash' in syntax to describe what happens if some linguistic rule is not satisfied.