[Lingtyp] comparative concepts

[William Croft]

Let me jump in again before some of the points made today get lost.

1. David Gil argues that syntactic categories "based exclusively on distributional criteria, and...blind to the semantics" can be defined as comparative concepts. But distributional criteria are criteria defined by a construction or constructions in a specific language, and so by definition are not comparative: the constructions are constructions of particular languages. This is what makes grammatical categories language-specific, and in fact construction-specific. 

I am aware of David's and others' skepticism regarding the crosslinguistic validity of semantic categories. That is why I included a chapter on semantics [ch. 3] in "Radical Construction Grammar", which is otherwise a book about syntactic structure; see also the CogniTextes paper I referred to earlier.

2. Jan's distinction between functional and semantic categories appears to correspond to my factors of propositional act function and semantic class in my 1991 analysis of parts of speech (and the 1990 paper he cites). I expand this twofold approach to all "meanings"/"functions" in my morphosyntax textbook (in preparation), using the terms 'semantics' and 'information packaging'. Both Jan and I are trying to develop a general functional framework, However, I think it's impossible to be comprehensive -- after all, the framework somehow has to describe everything that everyone would want to say in a human language. We can only hope to present (almost) all parts of the skeleton, and put some flesh on the bones.

3. David is right to say that using an arbitrary term like 'banana' for a language-specific category "would have made the discussion virtually unreadable". But I disagree with David's assertion that calling it "Language-A-Subject" presupposes that it is an instantiation of a comparative concept 'subject'. That's the whole point of capitalization, and prefacing it with the language name: to make it into a proper name. To update an example in "Radical Construction Grammar", just because Donald Trump and Donald Duck are both named Donald doesn't mean that they have anything in common. (Well, maybe I need a clearer example.) 

So why is it easier to understand when you call the Language A category '[Language A] Subject' instead of '[Language A] Banana'? I wrote earlier that comparative concepts are typically points or narrow regions in conceptual space, and the hybrid comparative concept of a construction is any form in any language that encodes a particular point or region in conceptual space. So it makes sense to name the language-specific category, defined by a construction that encodes among other things that point in conceptual space, by the name of the comparative concept -- as long as you know the difference, which can be cued by the presence/absence of capitalization. (See Appendix A of chapter 1 of my morphosyntax textbook, at my website, for some suggested rules of thumb for labeling language-specific categories in practically useful ways.)

David is worried that using the same label, even capitalized vs. lower-case, will lead to confusion. But I think that the discussion we've had here suggests that the confusion, or perhaps disagreement, is theoretical/conceptual. (However, I'm not sure Wheaton College's suspension of Larycia Hawkins turns solely on misunderstanding regarding a comparative vs. religion-specific definition of 'God'.) Once most or all typologists and documentary linguists come to agree that language-specific categories and comparative concepts are different things defined in different ways and serving different purposes, though both vital in their own ways, then the terminological concern will go away.

Bill