Grammaticality without grammar, iteration without forms
I want to show how possible Harrisian approaches to two issues in linguistics are inter-related in their underlying conceptualisation, which then becomes more fully available for inspection.
It is an interpretation of what Kant is saying in the third of his Critiques, The Critique of Judgement that there can be true judgements of beauty, of the form ' Such-and-such is beautiful' , without there being rules of beauty, stateable in advance of any particular exercise of judgement. The interpretation is persuasively developed by Mary Mothersill in her book Beauty Restored. If she is right, then what Kant is saying is one very good reason for assigning him to Romanticism rather than Classicism. Judgement is not the application of rules, but the exercise of intuition. After the event, we may be able to reflect and point to features of the object which prompted us to our judgement - indeed, to be able to do so is necessary to the project of making our judgement intersubjectively available - but in making our judgement we worked with no checklist of features necessary and sufficient to form our judgement. Kant calls this kind of judgement reflective judgement and says that it is judgement without a concept, in contrast to determinant judgement, judgement with the aid of a concept (checklist, rules, etc).An analogy may help in understanding Mothersill's Kant. Scientific revolutions are not predictable; they involve human creativity. It is only after the event that an exercise of human creativity can be recognised as successful in advancing our science of the world. So much is common ground out of Popper, Lakatos, Kuhn, Feyerabend and others.
Could it be the same with grammaticality? Someone produces a novel form - let's say (borrowing the example of the colour spectrum from P>H> Matthews), they use the form 'magentaness' - and, surprisingly, in its context, it fits. It isn't rebarbative, we understand it, but better than if it were a neologism coined by someone learning to speak English. We endorse its use, and judge it grammatical. But, in advance, there was no rule which unambiguously specified the circumstances in which a colour word could or could not take the ending '-ness'. A core grammar, in a Chomskyan sense, would allow all colour words to take '-ness'. A periphery to this core could do little more than list the forms already allowed. It would not be able to specify the conditions under which a new form would be allowed. It is the essence of creativity in language to be able to find a context for novel forms such that they will not simply be allowed to pass (as 'acceptable in the context') but will stand repetition in another context (as 'grammatical'). And the intuition of grammaticality is the intuition that this form could pass muster again in another context.
Already, my language is heading me towards an engagement with another vexed issue, that of iteration, a topic of debate between Roy Harris and myself in the past.
We might say that the essence of iteration is the ability to hear one acoustic blast as a repetition of the form of another, without it being the case that necessary and sufficient conditions for repetition can be stated in advance. The degree of permissible acoustic variation is not set out of context, and contexts cannot be defined in advance. It is just the same with Witgensteinian 'seeing as'. Whether you can see my rabbit drawing as a duck drawing depends a bit on you and a bit on me. You may be more charitable or I may be less proficient, and vice versa. There are no necessary and sufficient conditions which a rabbit drawing must possess in order to be seen as a duck drawing. It all depends. Yet once seen as a duck drawing features can be pointed out in my drawing which make it available to be seen as a duck drawing. Only in this way can what I see be made available for you to see. None of this is to say that any old drawing can be seen as a duck drawing. That is not the claim being made. There is a difference between drawings and Rorschach ink blots, and there is a difference between blasts and linguistic blasts - a world of difference.
But what difference? The cognitivist will want to enter the fray at this point and say that - of course - we have all given up the search for necessary and sufficient conditions for concept application, whether because we have read Wittgenstein on family resemblance or philosophers of science on the failure of the logical empiricist project of reducing theoretical sentences to observation sentences. But that this is the case does not mean that there is not something which guides us in judgements of beauty or assignment of sound blast to phonemic category. The features which we think we pick out after the event were actually there as part of some schema, script, frame, template in our heads. It's certainly not as neat and tidy as a set of necessary and sufficient conditions, but it is definitely something. And without it our intuitions of beauty and our assignments of sound to form would be wholly mysterious. We would be unable to give any account of them after the event any more than before.
This is not a bad response, but it does not really come to grips with the creative character of Kant's reflective judgement or some categorisations of linguistic form. The whole idea is that what we see is something we have not seen before, and have not been prepared for. We did not in any sense know that a painting which looks like that (demonstrative dthat) could be beautiful, or that a sound blast like [zaaaaa] could be heard as the utterance of How's that?
The cognitivist will want to respond, however, that something went on between our ears nonetheless and that the job of cognitive science is to make sense of what did go on. Perhaps a huge amount of general world knowledge was drawn upon, perhaps some very general analogic thinking mechanisms. In other words, perhaps the kind of thinking involved was not modularised and encapsulated. For example, we were able to hear [zaaaaa] as How's that? not least because we were watching a game of cricket, saw the ball hit the batsman's leg, saw the speaker throw his arms in the air while uttering [zaaaaa] and so on. It is absolutely true that had the sound blast [zaaaaa] been tape recorded and played to us out of context we would not have been able to assign [zaaaaa] to the form / How's that?/. But, and this is the catch, there are perfectly good realisations of the form /How's that?/ which could have been recorded and recognised out of context. Some cricketers have better articulation than others.
What seems unavoidable is the postulation of the form /How's that?/ existing somewhere. For it is this form that a cricketer intends to realise by [zaaaaa], and it is our knowledge that the cricketer intended something by the utterance of [zaaaaa] that motivates our processing of the acoustic blast as a linguistic blast. If we thought that [zaaaaa] was an index, a symptom, of the cricketer's being in pain after having been hit by the ball, we would handle the sound quite otherwise, and certainly not as the intentional iteration of a form. We might seek to classify the cry of pain - as a shriek, a moan, a yelp, and so on - but the cricketer would not have been seeking to produce a shriek, a moan or a yelp, though an actor in a play about cricketing mishaps might well intend to do so.
1994; Unfinished draft material for a contribution to a Roy Harris Festschrift. The published contribution 'Language, Art and Kant' can be found in G Wolf and N Love. Eds., Linguistics Inside Out. Roy Harris and His Critics. Amsterdam: John Benjamins 1997, pages 226-28