Φωνήεντα Chapter 3

by C Ryan Moniz

original research· spring 2010 - harvest 2016
updated & published· spring 2022

philology

“Morphology has to do with the units out of which words are put together (for example, whether there is a suffix between the root and the desinence [i.e. inflexional ending]), while morphophonology cuts across different morphological configurations to account for the place of the accent, and for the application of ablaut [...] Folding the outcomes of all these independent processes together into the four types... yields static templates which don’t generalize from one case to the next [...] In contrast, separating word-formation and inflection from accent and ablaut rules and constraints, as the compositional approach does, integrates each morphological configuration into the system, and makes intrinsic predictions about the accent and ablaut of a word regardless of [the] morphemes it happens to have.”^❦1 Kiparsky 2010 p 7

Table-12 — OT constraints governing accent in rank order, Frazier 2007

• Max(Accent)_root • do not delete underlying root accent • _FAITH
• ¬OP-Dep(Accent) • “depend on non-optimal paradigms”: realize dominant endings by asserting an accent into the stem (in contrast to the realization of recessive endings) • _ANTIFAITH
• Max(Accent)_deriv • do not delete underlying derivational suffix accent • _FAITH
• Max(Accent) • do not delete underlying accent anywhere • _FAITH
• Align=Post-Accent • accent the morpheme following a postaccenting morpheme • _MARKEDNESS
• AlignLeft • accent the leftmost morpheme of a prosodic word • _MARKEDNESS

Max(A)_root » ¬OP-Dep(A) » Max(A)_deriv » Align=PoA » AlignL

Hysterodynamic nouns can be explained as nouns with a postaccenting root and an unaccented derivational suffix. Only the nominative zero-ending is dominant, contrasting with a recessive zero-ending for the endingless locative; no pre-accenting ending (or preaccenting alignment constraint) is required to explain the form of the endingless locative.

Proterodynamic nouns can be accounted for by inherently accented derivational suffixes. Note that the surface forms are predicted correctly regardless of whether the root is postaccenting or not; this could mean that all unaccented roots were also postaccenting, and the only differentiation between animate and inanimate mobile roots is the accentuation of the derivational suffix, i.e. the suffixes of animates are unaccented, while the suffixes of inanimates are accented.

The static nouns exhibit root accent regardless of case or accent of the derivational suffix, because of the high ranking of Max(A)_root. Frazier’s constraints can still be used to explain the accent of nouns fitting the model proposed by Kloekhorst (2013), assuming the following distribution of accentual properties for morphemes which comprise nominal stems:

Together with the recessive (non-nominative) and dominant (nominative) endings, this accent distribution and Frazier’s constraints provide a comprehensive compositional explanation for the surface accent and ablaut grade of Kloekhorst’s (2013) revision of the Leiden model (which also assumes one vowel phoneme *e)

Other attempts have been made to explain PIE nominal ablaut without the use of antifaithfulness constraints. Keydana (2012) attempts to explain the hysterokinetic nouns using only markedness constraints: prosodic constraints which strive to keep the prosodic word binary, i.e. containing two syllables (FtBin), constraints which parse morphemes into syllables (Parse-σ, Parse_Head, Parse_Non-Head), and a constraint which reduces full-grade before an accented syllable (reduce). Keydana also posits a markedness constraint which disallows the cluster *-rs at the end of the prosodic word (*rs]_ω); forms such as *ph₂·tér-s, with the regular animate nominative ending *-s, are typically proposed as the nominative for animate nouns with stem-final *r, even though the reflexes predict forms with a long vowel, i.e. *ph₂·tḗr. Szemerényi’s law, which deletes final fricatives after sonorants and results in compensatory lengthening, is used to explain this change, so the underlying final *-rs is still assumed; Keydana instead uses the constraint *rs]_ω to predict this allomorphy in nouns with stem-final *r.^❦12 For a detailed OT explanation for the vowel length predicted by Szemerényi’s law, see Sandell & Byrd 2015

Keydana’s constraints explain the hysterokinetic nouns by means of inherently accented derivational suffixes, and no accent in the stem. Unaccented morphemes can contain either an inherent vowel, in which case they are considered non-heads and can be parsed as syllables, or no inherent vowel, in which case they cannot be syllabified. Thus, the strong case endings can be differentiated from the weak endings by means of assuming non-head status for the weak endings on the one hand, and vowel-less strong endings on the other. However, Keydana’s constraints do not seem to predict the surface accent for any type other than the hysterokinetic nouns. For example, regardless of whether proterokinetic nouns are assumed to have non-head roots or zero-grade roots, syllabification constraints cannot correctly predict the strong (nominative) form of proterokinetic roots with root-accent:

A higher-ranked antifaithfulness constraint, such as Frazier’s ¬OP-Dep(A), would resolve this discrepancy (i.e. it would render *[[gʷém]_σ[·tu-s]_σ]_ω the most optimal form for the nominative), but such a constraint would then require both dominant and recessive endings, and the analysis would no longer rely solely on markedness constraints.

Compositional approaches to PIE nominal accent-ablaut which use OT constraints push the necessity for the observed allomorphy of athematic nouns to be analyzed as morphologically governed. The different types of athematic nouns can be explained by the combination of morphemes with inherent accentual features, and their allomorphy can be explained by means of competing morphophonological constraints. However, they alone do not explain the vowel quality allomorphy between *e and *o unless they also follow the Kortlandt-Beekes chronology that posits the complementary distribution of *o with *e. Frazier’s (2006, 2007) constraints provide an accurate and comprehensive means of explaining the accent distribution reconstructed for the (Kloekhorst 2013) period during which there was only one vowel phoneme *e.

Φωνήεντα·

1 · 2 « 3 » 4 · 5

references

bibliography

• Alderete, JD 1999. “Morphologically governed accent in Optimality Theory.” U Massachussets
• Frazier, M 2006. “Accent in Proto-Indo-European athematic nouns: Antifaithfulness in Inflectional Paradigms [Thesis].” U of NC Chapel Hill
• —— 2007. “Accent in athematic nouns in Vedic Sanskrit and its development from PIE.” Proceedings of the 18th Annual UCLA Indo-European Conference
• Halle, M 1997. “On stress and accent in Indo-European.” Language 73, p 275-313
• Keydana, G 2012. “Ablaut in indogermanischen Primärnomina: Die hystorkinetischen Stämme.” Das Nomen in Indogermanischen. Morphologie, Substantiv versus Adjektiv, Kollektivum. Akten der Arbeitstagung der Indogermanischen Gesellschaft 14,16, p 113-128
• Kim, R 2002. “The continuation of Proto-Indo-European lexical accent in ancient Greek: preservation and reanalysis.” Current Work in Linguistics (Working Papers) 7,2, p 59-93
• Kiparsky, P 1973. “The inflectional accent in Indo-European.” Language 49, p 794-849
• —— 2010. “Compositional vs. paradigmatic approaches to accent and ablaut.” Proceedings of the 21st Annual UCLA Indo-European Conference
• Prince, A & Smolensky, P 1993. Optimality Theory: Constraint Interaction in Generative Grammar. U of Colorado
• Sandell, R 2014. “Perspectives on the reconstruction of the Proto-Indo-European accentual system.” UCLA Indo-European Workshop
• Sandell, R & Byrd, AM 2015. “Extrametricality and non-local compensatory lengthening: the case of Szemerényi’s law [Lecture Handout].” The LSA 2015 Annual Meeting

philology