Sunday, 10 October 2010

What was the first syntactic category?

I've been thinking lately about the origin of syntactic categories in language, and when things such as nouns and verbs could be said to exist.  In the literature, I've seen it implied more than a few times that nouns are the most fundamental syntactic category and likely arose first (see Luuk 2009 for a recent example).

Upon pondering what exactly a 'syntactic category' means, and what makes a noun a noun and a verb a verb etc, I came to the conclusion that it is only in relation to another word that you can say what syntactic category a word has.  We remember in school learning how a noun is a 'person, place or thing', while a verb is 'an action', but this is a simplification- relying on the semantics of a word is very unreliable in determining the category of a word in a sentence.

For example, the word 'hammer' is only a noun if it is used as a noun in a sentence: ex. "I have a hammer."  In isolation, it does not have a syntactic category- in English, it could quite possibly be a verb as well: ex. "Hammer the nail into the wall."  There are other languages, such as those in the Wakashan family, where most word roots can act as a noun, verb, or adjective in a sentence- fore example, 'wolf' can be a noun just as easily as a verb or ajective (I heard this discussed by Maggie Tallerman at the last Evolang conference in Utrecht, but I haven't seen it published).

My conclusion from this is that the first syntactic category must have arisen in contrast to another; and therefore, a minimum of two would have had to arise at the same time in order for there to have been any distinction between words and their roles in an utterance.

The only discussion of this idea I've seen occurs in Heine and Kuteva, A Genesis of Grammar (2007), where they cite Tallerman as well as Jim Hurford (as personal communication).

I've been trying to put my thoughts down about this into a coherent sort of paper, but it lies untouched in my folder not getting much attention.  I need a nudge!


  1. Good post!

    I think the idea of word classes comes later, and the first words were not really words at all. I think the two main points are the that the first utterances were simply monomorphemic, and also accidental. Here's my two pennies' worth:

    1. the most helpful way to see the first utterances is as of a single morpheme in length (duh, you might say, but I think that this is far less controversial than dealing with the actual word classes of the first utterances)

    2. As that lovely book by Burling (the best book on lang evo ever?) pointed out, language had to be comprehended, and this opens up the question of early pragmatics. A single morpheme like 'hammer', to take your example, can be used as a verb or a noun (at least), and early comprehenders would have to utilise an early form of pragmatics to decipher what the utterer meant. Was it 'hammer that!', 'bring me the hammer', 'bloody hammer!', etc. So again, word class is relegated. Single morpheme rules.

    3. most controversially...since it is practically impossible to imagine a reason why any individual humanoid might utter the first morpheme in the first place (being that there was no convention of morphemic communication anyway, at least of the verbal kind), it strikes me that the first uttered morphemes were comprehended accidentally. in other words, the utterer didn't know he/she was being comprehended at all. Burling's work logically leads to this conclusion. The final point here being, as with the example of the 'hammer', word classes in single-morpheme utterances were as open to subjectivity then as they are now.

    Is that of any use? I resisted the temptation to add 'MC Hammer' to my list.

  2. Take a look at the work of Alison Wray on the holistic protolanguage model, for example in a comment to Arbib "From monkey-like action recognition to human language". The idea is that the categories "coalesced" from an earlier utterance, demonstration or chant.

  3. Thank you Steve! I'm glad to have your readership :)

    I agree with you that the first utterances were likely holophrastic, like you say, sort of like 'one morepheme'. I've also been influenced by that lovely book by Burling (The Talking Ape, 2005) and think it's really important how he points out the role of the 'comprehend-er' in language evolution and how that could have driven the attribution of symbolic meaning to an utterance. So I think that monomorphemic utterances became compounded with another utterance, and it's when these two otterances together meant a different meaning to another, and then a whole community, the environment for the first syntactic categories appeared.

    This leads into Rick's comment (thanks Rick!) about Alison Wray and her theories on holistic utterances breaking down into components. I've always sided more with the idea (a la Maggie Tallerman) that language evolved synthetically, ie single words arose first and then combined to create new meanings. It is when two words were used and had a relationship to each other that you could attribute a syntactic role to each.

    The reason I side with this view instead of Alison Wray's scenario is that I see the process of using two utterances to create a new meaning to be more simple than breaking down one utterance into two meanings. Also, we can find similar examples of combinations of utterances in other species; Diana monkeys use two different calls together to mean something different from each of the calls individually.

    Thank you for commenting!


Related Posts Plugin for WordPress, Blogger...