[This is part of a series of defectivity case studies.]

Sims (2006: ch. 4, henceforth just S) documents gaps in the genitive plural (gen.pl.) forms of modern Greek nouns. These gaps appear to be reasonably well known and quite common: S’s primary data comes from two Greek dictionaries published in the 1990s. The LNEG dictionary contains over 1,600 nouns defective in the gen.pl.; the LKN over 1,300.¹ If this is even remotely close to correct, this is still probably the most extensive  case of defectivity, in terms of affected lexemes, documented in any language. Interestingly, S notes that there is “surprisingly little agreement between the two dictionaries about which lexemes are defective in the genitive plural” (S: 81, fn. 54), with the two dictionaries agreeing on 470 nouns.

Modern Greek declines nouns for four cases and two numbers. The primary gen.pl. marker, written -ων (-on) may or many not trigger rightward shift of primary stress, depending on the noun. Pappús ‘grandfather’ has no gen.pl. stress shift (pappúdon), whereas náftis ‘sailor’ and ónoma ‘name’ do (naftón, onomáton).² There does not seem to be much—except etymology—which predicts whether or not a noun will shift. S suggests that there is an important distinction between nouns which already have final stress in the nom.sg., and others which have stress on earlier syllables in the nom.sg., as the former, represented by pappús and referred to by S as “type 1” and “columnar stress” nouns, never have stress shift in the gen.pl., and do not exhibit gen.pl. gaps.³ Assuming this is correct, all that remains is to understand why there is no productive default in the non-“type 1” declensional classes.

There are several other things going on here that are not yet well understood. First, modern Greek is in some sense diglossic, and more specifically pluricentric, with vernacular and learned registers (known as dimotiki and katharevousa, respectively) existing alongside various divergent dialects. One might be tempted to suggest that this would result in heightened awareness of lexical variation in gen.pl. stress shift. Second, S notes that the genitive itself is in competition with a periphrastic alternative that uses a preposition governing the accusative case (S: 86f.). According to S, the periphrastic accusative is associated with dimotiki and the genitive with katharevousa. Since katharevousa is itself reasonably artificial, colloquial use of the genitive may be increasingly moribund. Third, the LNK dictionary also lists quite a few nouns are defective in other plural forms as well. These have received little attention; some are probably  singularia tantum (see, e.g., Sims 2015: 150) but perhaps something else is going on too.

Most of the points made by S in her 2006 dissertation are echoed in her 2015 monograph. I have little to say about her idea that defectivity is encoded via implicational hierarchies in the morphological paradigms, and I didn’t find the experimental results informative.

Endnotes

1. See S (ibid.) for these citations. I have briefly consulted both works in print, just to spot check S’s coding. Since I found no major issues, I am instead working from tables in S’s appendices. Thanks to Lucas Ashby for help digitizing these.
2. I have taken the liberty of standardizing S’s transliterations somewhat.
3. It is hard to tell whether S thinks this is a synchronic explanation for defectivity. Elsewhere, S (ch. 5) claims that defectivity is grammaticalized (or perhaps lexicalized) by essentially having the morphological module return some kind of “null” for defective combinations, so that there is no need for further synchronic motivation. S seems to consider this a desirable conclusion.

References

Sims, A. 2006. Minding the gap: Inflectional defectiveness in a paradigmatic theory. Doctoral dissertation, Ohio State University.
Sims, A. 2015. Inflectional Defectiveness. Cambridge University Press.