Robert Vago CUNY Queens College *VV IN HUNGARIAN In Hungarian (phonological) words, suffixation derives robust instances of heteromorphemic vowel sequences (V1+V2). In this presentation, I will have two goals: (a) to report on a thorough investigation of the facts (as culled from traditional grammars, Kenesei et al. 1998, Siptár & Törkenczy 2000, inter alia) to bring together a conspiratorial web of mechanisms to respect a *VV constraint (V1 deletion; V2 deletion; allomorphy); and (b) to provide analyses for the data within the framework of Optimality Theory (OT; Prince & Smolensky 2004 [1993] et seq.), distinct from other treatments of related or overlapping topics within OT (e.g., Stiebels & Wunderlich 1997; Siptár 2008). I will advocate Gouskova’s (2013) approach to exceptionality, whereby exceptionality is the property of morphemes, which can be indexed for reranking constraints in the hierarchy. In this space, I will cite but a few exemplars of each mechanism, though in reality the patterns are reinforced by a multitude of cases, some more so than others. For the sake of clarity, I will use the symbol “[ ]” to indicate the position where a V has been deleted. Hungarian has the following seven short and seven long contrastive vowels, each identified both orthographically and phonetically: SHORT: i [i], ü [y], u [u], ö [ø], o [o], e []ܭ, a [ ]ܧ/ [LJ] LONG: t>L@ޝ, Ħ>\@ޝ, ~>X@ޝ, p>H@ޝ, Ę>¡@ޝ, y>R@ޝ, i>D@ޝ For ease of exposition, I will cite Hungarian data tokens orthographically. V2 deletion. If a V-final stem is followed by a V-initial suffix, whether derivational or inflectional, V2 deletes. Cf. hat-os ‘the number six’ vs. millió-[ ]s ‘the number million’; szag-ok ‘odors’ vs. kapu-[ ]k ‘gates.’ The following constraints will be proposed, ranked as given: *VV: Vowel sequences are disallowed. (ignoring morpheme internal cases) MAX V]ST: Do not delete a morpheme-final V in stems. MAX +V: Do not delete a morpheme-initial V in suffixes. V2 deletion also takes place between suffixes. Cf. hoz-om ‘I bring it’ vs. hoz-ná-[ ]m ‘I would bring it.’ For these cases, let’s add to the inventory: MAX V+: Do not delete a morpheme-final V in suffixes. V2 deletion is by far the most common pattern in the language. The others are less robust and are to be treated in terms of exceptionality. V1 deletion. In the case of a handful of suffixes (around 5), VV clusters are avoided by deleting V1. This is conditioned only by specific derivational suffixes, which attach directly to bare stems (roots). Cf. zongora ‘piano,’ zongor[ ]-ista ‘pianist.’ These suffixes are indexed to reorder the constraint ranking from the one given above as *VV » MAX V]ST » MAX +V to: *VV » MAX +V » MAX V]ST. It is interesting that the suffix –ista (and its ilk), if followed by a typical V-initial suffix, is exceptional with respect to its initial vowel (it induces V1 deletion) but regular with respect to its final vowel (it allows for V2 deletion). If we take into account the above motivated MAX V+ constraint, we arrive at the constraint rankings for V1 deletion cases: (DEP-C (no C epenthesis) is omitted, as it is a very low ranking constraint in Hungarian.) Robert Vago CUNY Queens College /zongora+ista+unk/ ‘our pianist’ zongora+ista+unk zongor[ ]+ist[ ]+unk ܐzongor[ ]+ista+[ ]nk (=zongor+istá+nk) zongora+[ ]sta+unk zongora+[ ]st[ ]+unk *VV MAX V+ MAX +V MAX V]ST *!* *! *! *! * * * * Allomorphy. Another leg of the *VV conspiracy leans on allomorphy. In a number of cases a V-initial suffix has a CV-initial allomorph following a morpheme ending in V. E.g.: magas rangú ‘high-ranking’ vs. jó formá-jú ‘well-formed.’ Conversely, the infinitive suffix has the allomorph -n before V, -ni otherwise (Siptár 2009): e.g. vár-ni ‘to wait’ vs. vár-n-om kell ‘I have to wait.’ Clearly, the allomorphic variation finds its motivation in the avoidance of *VV sequences. No deletion. In some cases, stems or suffixes prevent V2 deletion in a following suffix or a subset of suffixes, as in szomor(-)ú ‘sad,’ szomor(-)ú-an ‘sadly.’ In other cases, certain suffixes systematically do not undergo and do not serve as context to V deletion: e.g. eszperantó-ul (beszél) ‘(speaks) Esperanto.’ Once again, constraint reranking plays an important part in the analysis. We have seen the three basic components of the implementation of V2 deletion: *VV » MAX V]ST » MAX +V. This ranking was reordered to *VV » MAX +V » MAX V]ST in order to accommodate V1 deletion. Not breaking up VV sequences is the manifestation of still another reranking scenario: MAX V]ST » MAX +V » *VV. Finally, as is well-known, the OT framework is particularly well suited to provide principled explanations for conspiracies. References Gouskova, M. 2013). Unexceptional Segments. Natural Language and Linguistic Theory 30. 79-133. Kenesei, I., R. Vago, and A. Fenyvesi (1998). Hungarian. London: Routledge. Prince, A. & P. Smolensky (2004 [1993]). Optimality Theory: Constraint Interaction in Generative Grammar. Oxford: Blackwell. Siptár, P. (2009). “Morphology or Phonology?: The Case of Hungarian –ni.” In den Dikken, M & R.M. Vago (eds.), Approaches to Hungarian, vol. 11. Amsterdam: John Benjamins. Siptár, P. & M. Törkenczy (2000). The Phonology of Hungarian. Oxford: Oxford University Press. Stiebels, B, & D. Wunderlich (1997). Second Stems in Hungarian Nouns. Theorie des Lexicons. Düsseldorf: Heinrich Heine Universität.
© Copyright 2026 Paperzz