Is 'grammar' a sealed system with its own protocols created by and for itself, or is it a system that is part of a wider process of making meaning…which itself is part of a wider system of human behaviour?
The SPaG is based on the former – that it is a sealed system with its own protocols. These protocols, under this theory, can be ‘spotted’ and, by implication, explained, by observations of grammar itself in tiny units of written language. Under the second theory, grammar can only be fully demonstrated and, by implication, explained, on the basis of making meaning in much longer passages of language in use (including speech), which themselves are dependent on what producers of language and their receivers (writers and readers, speakers and listeners) want to say and mean.
If you reduce language to the little sentences of a SPaG test and then ask children to produce passages of writing which include features from the test for little sentences, you get crap writing. How do I know? I have a ten year old. His homework last night could well have been an instruction on how to produce crap writing about fireworks night. He had to include certain grammatical features. This is precisely what I’m referring to as the fallacy of the ‘sealed system’. It’s assumed that sticking the grammatical feature into some writing will lead to satisfactory meaning-making.
The further problem of this way of teaching and testing language is that grammar can’t take the strain. The harder people try to make grammar fit the straitjacket of ‘right and wrong’ answers, and the sealed system of thinking, the less able is the grammar to take it. The test becomes littered with false distinctions, (see subordinating conjunctions and prepositions as an example), highly disputed categories (see subjunctive), and questions which are supposed to have one answer but which have more than one right answer. Last year, children were not allowed to write ‘The sun shone bright’ because, the word ‘bright’ was ‘not an adverb’. David Crystal told the examiners they were talking cock. He was overruled. What does he know about language? The devisers of the KS1 draft SPaG couldn’t even sort out how to describe our means of asking questions. They got their auxiliaries and main verbs in a muddle. In trying to make the micro-grammar ‘right’, they end up being ‘wrong’. We all know how to ask questions. We do it every day all day. Struggling to find right and wrong ways of describing this, in sufficiently simple terms for seven year olds ended up with them being ‘wrong’ by their own criteria – not mine.
There is yet another problem which none of us has commented on: to make the sealed system approach work, you have to exclude everyday speech. So, the language that the children are actually using themselves everyday and all day is excluded from this classification system. There is a powerful body of theory that investigates speech – from sociolinguistics to psycholinguistics to pragmatics. To take one tiny obvious example: two of the most used words (or non-standard dialect versions of them) are ‘yes’ and ‘no’. Will ‘yes’ and ‘no’ and their many subtle ways of being used be in the SPaG test? I very much doubt it. That’s because it is much harder to squeeze the ‘grammar’ of ‘yes’ and ‘no’ into the sealed system model. They are so clearly part of ‘meaning-making’ and human behaviour and not part of rules that can be derived from writing (other than drama).
To return to the issue of the possibility of a ‘comprehension-grammar’ this then raises a pedagogical point: how might we best teach ‘grammar-for-meaning’? I would suggest that this is where the compare-contrast principles come in,and it’s where we might pose problems for children in terms of ‘how might we best write ‘x’ kind of writing for ‘y’ kind of purpose?’ not, ‘how might we stick this grammatical feature’ into any old bit of writing?’!!!!
That’s the alternative way to teach grammar whilst paying respect to how grammar works in language: humans invented it in order to make meaning within our human needs, desires and activities.