{"id":2018,"date":"2024-09-25T16:06:14","date_gmt":"2024-09-25T20:06:14","guid":{"rendered":"https:\/\/www.wellformedness.com\/blog\/?p=2018"},"modified":"2025-07-29T02:47:50","modified_gmt":"2025-07-29T06:47:50","slug":"kisseberth-zonneveld-exceptionality","status":"publish","type":"post","link":"https:\/\/www.wellformedness.com\/blog\/kisseberth-zonneveld-exceptionality\/","title":{"rendered":"Kisseberth & Zonneveld on exceptionality"},"content":{"rendered":"

[This post is part of a series on theories of lexical exceptionality<\/a>.]<\/p>\n

In a paper entitled simply “The treatment of exceptions”,\u00a0Kisseberth (1970)<\/a> proposes an interesting revision to the theory of exceptionality. Many readers may be familiar with the summary of this work given by Kenstowicz & Kisseberth 1977:\u00a72.3<\/a> (henceforth K&K). Others may know it from the critique by\u00a0Zonneveld (1978: ch. 3)<\/a> or Zonneveld’s (1979)<\/a> review of K&K’s book. I will discuss all of these in this post.<\/p>\n

Kisseberth (1970)<\/h2>\n

A quick sidebar: Kisseberth’s paper is a fascinating scholarly artifact in that it probably could not be published in its current form today. (To be fair it was<\/em> published in an otherwise-obscure journal, Papers in Linguistics.<\/em><\/span>) For one, a<\/span>ll the data is drawn from Matteson’s (1965) grammar of Piro;1<\/sup> the only other referenced work is <\/span>SPE<\/em>. Kisseberth (henceforth K) gives no <\/span>page, section, or example numbers for the forms h<\/span>e<\/span> cites. <\/span>I have tried to track down some of the examples in Matteson’s book, and it is extremely difficult to find them. K gives no derivations, only a few URs, and the entire study hinges around a single rule. But it’s provocative stuff all the same.<\/span><\/p>\n

K observes that Piro has a rule which syncopates certain stem-final vowels. He gives the following formulation: <\/span><\/p>\n

(1) Vowel Drop:\u2028\u2028V -> \u2205 \/ VC __ + CV\u2028 <\/span><\/p>\n

For example, [kama<\/span>] ‘to make, form’ has a nominalization [kamlu] ‘handicraft’ with nominalizing suffix \/-lu\/<\/em>, and [xipalu<\/span>] ‘sweet potato’ has a possessed form \/n-xipa-lu<\/span>-ne\/ [nxipalne] ‘my sweet potato’.2<\/sup><\/span> One might think that (1) is intended to be applied simultaneously, as this is the convention for rule application in SPE<\/em>, but this would predict *[nxiplne], with a medial triconsonantal cluster. Left-to-right application gives *[nxiplune]; the only way to get the observed [nxipalne] is via right-to-left (RTL) application, which I’ll assume henceforth. As far as I know, the directionality issue has not been noticed in prior work.<\/p>\n

Of course, there are exceptions of several types. (I am drawing additional data from the unpublished paper by CUNY graduate student H\u00e9ctor Gonz\u00e1lez, henceforth G. I will not make any attempt to make Gonz\u00e1lez’s transcriptions or glosses comparable to those used by K, but doing so should be straightforward.)<\/p>\n

One type is exemplified by \/nama\/ ‘mouth of’, which does not undergo Vowel Drop, as in \/hi-nama-ya\/ [hinamaya] ‘3sgmpssr-mouth.of-Obl.’ (G 5a); under RTL application we would expect *[hinmaya]. This is handled easily in the SPE <\/em>exceptionality theory<\/a> I reviewed a few weeks ago by marking \/nama\/ as [-Vowel Drop].<\/p>\n

However, other apparent instances of exceptionality are not so easily handled. Consider two forms involving the verb root \/nika\/ ‘eat’. In \u00a0\/n-ni<\/span>ka<\/span>-nan\u0268-m-ta\/ [hni<\/span>ka<\/span>nan\u0268mta] ‘1sg-eat-Extns-Nondur-Vcl’ (G 5b) both vowels of the root satisfy (1) but do not undergo syncope. One might be tempted to mark this root as [-Vowel Drop], but it does<\/strong> undergo deletion in other derivations, such as in \/nika<\/span>-ya-pi\/ [nikyapi] ‘eat-Appl-Instr.Nom’ (G 4d). Rather, it seems to be that the following \/-nan\u0268\/ fails to trigger <\/strong>deletion. This is not easily handled in the SPE\u00a0<\/em>approach. K gives a number of similar examples involving the verbal theme suffixes \/-ta\/ and \/-wa\/, which also do not trigger syncope. If a morphemes vary in whether or not they undergo\u00a0and<\/strong> whether or not they trigger Vowel Drop, one can imagine that these properties might cross-classify:<\/p>\n