There is a new posting on Babel’s Dawn that is very interesting. Bolles is outlining a new higher level analysis of language: phonetics, vocabulary, syntax, and he adds sociality. Roughly sociality is a set of limits on language allowing speaker and listener to cope. He starts with the notion that a sentence can have a topic and a sub-topic can be added, but a sub-sub-topic is something that we cannot cope with.
It is a nice post and I recommend reading it. ( a fourth component of language )
This sociality idea touches on ideas that I have been playing with.
Dealing with written language leaves out something very important. Oral language is limited in its complexity by the number of things that can be held in working memory (quoted as 2-7, average about 4). This is not literally 4 words because there can be chunking, where a number of words form one ‘thing’ in the mind. So: understanding an oral sentence requires that the words are in an order that allows chunking and that the total number of resultant ‘things’ is not greater than the working memory limit of the listener. In other words, the limit of coping may be a firm physical limit.
The sociality level of analysis is really about communication. If language is being used to communicate then the principles that allow language to achieve communication are the tools of sociality analysis. Oral communication is the primary function of language and therefore the highest level of analysis should properly be how oral communication is possible.
Once we get to the sociality analysis, I think we can dispense with the sentence as THE unit of language. Phrases on the one hand and multi-sentence groups on the other can be analyzed at this level. For example, the use of ‘so’, not just in its usually senses, but to signal a change of speaker or of the overall topic of conversation, has to be thought of as a meta-communication. In a parsing diagram it would have to appear outside the sentence.