1. Some people on twitter assume that I am someone who can't possibly know anything about 'grammar' because a) I write children's books and b) I'm opposed to young children learning much of the 'grammar', that is in the KS1 and KS2 tests.
2. I hate to do this but as several people have asked:
a) at primary school in the 1950s, we did parts of speech, a bit on relative clauses and comparatives (!) and not much more. They were much more anxious about maths, and verbal reasoning tests.
b) At Grammar school, in English we did 'box analysis' and 'clause analysis' - loads of it ie breaking up sentences into phrases and clauses and putting them back again. In French, Latin and German (which I did) we did conjugations, declensions, tenses, moods much of which we learned by rote.
c) At university I did historical linguistics along with learning Anglo-Saxon (or 'Old English') and then some Middle English.
d) On various occasions, I've voluntarily put myself through mini-courses in the various kinds of grammar available, especially M.A.K. Halliday. I've also read much of David Crystal's work.
e) I've also immersed myself in other ways of talking and writing and about language - mostly socio-linguistics of various kinds e.g. Labov, Dell Hymes, Trudgill.
f) Since 1998, I've presented about 20 half-hour programmes a year for BBC Radio 4 on 'language in use' ie a form of descriptive linguistics to describe how people use language. I'm still presenting that, now alongside linguist Laura Wright.
g) Though it's no formal qualification, my father was an applied linguist and we must have spent hundreds of hours talking about language in literature, education, daily life etc. and I have read most of his work.
f) I have an MA and a Ph.D. in children's literature much of which involved considering how it is we handle language in order to read and write: a combination of 'intertextuality' and 'rhetoric' - both branches of applied linguistics.
3. The result of all this leads me to think that there is absolutely no harm, and perhaps some advantage FOR PRIMARY SCHOOL CHILDREN in knowing the generally shared names for the basic parts of speech and a bit on subject-verb structures. Beyond that for this age of child I think that this particular kind of grammar involves too much work for too little return to spend time figuring out, learning and spotting such structures as 'determiners', 'fronted adverbials', 'embedded relative clauses', specific names for tenses e.g. 'present perfect progressive'.
4. The kind of 'grammar' on offer in these tests is a) not the only way of describing language and b) not necessarily the best 'knowledge about language' (KAL) to be taught in order to help children write well.
5. This kind of 'grammar' on offer to primary children attempts to describe patterns, conventions and 'rules' derived from 'meaning' and 'function' and then declares these as correct. It cannot and does not explain or explore changes in use, changes in meaning, variations in use, variations in meaning, or indeed use itself. Instead, it takes simplified, 'ideal' sentences or words, takes them out of any context of actual use and demands that children identify features of this non-real language. As a result, children are not invited to explore or understand language as it actually is.
6. This kind of 'grammar' on offer to primary children keeps trying to pin down a perfect or pure or totally correct way to describe a written language feature and give it a name. As a result. in my own lifetime, this terminology has changed many times over and gives rise to many arguments. In fact, more often than not, there isn't a more or less right way of talking about it, because the terminology is constantly being derived from the system itself, without referring outside and back to meaning, context and social function ie how and why is a particular word/phrase/clause/sentence being used at that particular moment by those particular people. The terms themselves are often opaque and don't refer to meaning and social function - though, mysteriously some of them do, like 'possessive' ie it indicates that we 'possess' things in real life 'out there' and therefore need words to describe and indicate 'possession'. 'Present perfect progressive' doesn't indicate anything of meaning and social function.
7. Let's not forget for a single second: this whole grammar apparatus was not introduced into schools because people in education thought it was a good idea. It was introduced because the Bew Report of 2011 that was set up by Michael Gove to make recommendations about 'assessment and accountability' said that 'Grammar, punctuation and spelling' were a good means by which teaching (not children!) could be assessed because grammar, punctuation and spelling questions have 'right and wrong answers'. In short, that is an untruth.
One exemplification of that: Schools Minister, Nick Gibb made what was described as a 'mistake' when asked to identify a word in a sentence. In fact, all that he did was give one of two alternatives for that word. It is only this kind of grammar test that calls it a mistake. It's the test that is the mistake, not the man who imposes it!
8. If we started again and wanted to think of how best to talk about language in primary schools we would start with investigation of forms of language in use. And we would do this alongside many forms of playing with and using language that children would enjoy and find interesting and useful for their writing.
9. The 'grammar' that is being taught is being applied artificially to the children's writing as a measure of what is supposedly good writing. In other words, the children are being asked to use things like 'embedded relative clauses' as the criterion. The only reason for this is that it then becomes a measurable, testable quantity. However, that has nothing to do with what makes writing good. Writing is being distorted to fit this particular kind of 'grammar'.