Jonathan Robie wrote: ↑December 26th, 2017, 6:28 pm
I basically want a set of inflectional classes for each kind of word that can be inflected. Traditionally, we have talked about first, second, and third declension for nouns and adjectives, and many books seem to have ~7 categories for verbs.
From what we've discussed, I suspect you may have ideas for doing this better. What would you recommend? It might be worth starting a new thread for this ... how do you accommodate both lumpers and splitters, for instance?
Mounce's categories will get you a decent way there, especially for nouns. In all cases, Mounce's categories are hierarchical which goes a long way to accommodating both lumpers and splitters. It works a bit better for nouns (although Mounce does still say odd things in MBG like "this noun is in category X but inflects like it's in category Y". If so, why not put it in Y? Oh, because you're confusing whether to go by stem form or paradigm realisation). With verbs it's problematic more often because there isn't a single dimension along which to categorise verb inflection. So I think ultimately you need multiple categories—separating, for example, how personal endings work from how stem formation works. In this regard, I think certain inflection classes really belong on verb stems rather than entire lexemes.
In my SBL 2015 talk [1], I outlined my own ongoing efforts to produce a "better" inflectional class system. In [2] I talked about my 2014 project looking at noun inflectional classes just in terms of headword patterns in Danker's concise lexicon. I think the resultant system is a promising start to an alternative to Mounce. [3] outlines a parallel approach based on actual noun forms found in the GNT and using Mounce's categories as a point of departure. Both [2] and [3] are rigorous in that (unlike most grammatical descriptions of inflection) there is a clear, machine-actionable definition of what is meant by a particular class and the data can be evaluated against the definitions to catch any exceptions or problems with the rules.
For an extreme example of lumping versus splitting: I point out in [4] that while dative singulars are _always_ formed by adding –ι (so, in that sense there is only one type of dative singular formation), when viewed as a distinguisher in a paradigm (particularly versus the nominative singular and its morphophonological contortions) there are 28 (!) distinct forms the dative singular distinguisher can take.
For verbs, I've started a very lengthy series of blog posts on their classification from the ground up starting at [5]. Even though the series only covers the present so far, there is a lot said in those posts about differing notions of inflectional classes. But again, the approach is a very rigorous, data-driven one, which inevitably leads to me creating new categories where, say, Mounce, would just have a footnote pointing out an exception. The approach I'm taking in that series is to firstly classify within systems and then later I'll classify stem formation.
Adjectives are probably the category I've done the least work on. I think the way to think of adjectives (and this isn't novel by any means—most people do it) is as a choice of which inflectional class is used for each gender. My work on [2] and [3] touches on adjectives too.
If people would like me to produce a list of nouns and adjectives with some sort of inflectional class information, I think I can do that pretty quickly just by bringing together this 2014 / 2015 work. Most of it already there, I guess it just hasn't been noticed before
One final point on lumpers vs splitters, I presented a solution for lexemes in my SBL 2017 Annual meeting talk [6] that would be equally applicable to inflectional categories. This solution would allow, for example, my particular inflection classes to be linked to Mounce's or to Smyth's or whatever.
For those of you not familiar with the work going on for the new Perseus reading environment, I'm now expanding this morphological work from the 138k words of the GNT to the 21m words in the Greek corpus of the Perseus digital library. Once you have to deal with dialects and changes over time, it starts to get REALLY interesting
[1]
https://www.academia.edu/18816954/A_Mor ... ment_Greek
[2]
https://jktauber.com/2015/11/18/full-ci ... l-classes/
[3]
https://jktauber.com/2015/11/12/analyzi ... gy-part-1/
[4]
https://jktauber.com/2015/11/13/initial ... ning-form/
[5]
https://jktauber.com/2017/06/23/tour-gr ... gy-part-1/
[6]
https://jktauber.com/2017/11/22/sbl-papers-now-online/