On Finnic and Khanty vowel harmony: domains, slopes and their role in typology

• Published in: Suomalais-Ugrilaisen Seuran Aikakauskirja
• Main Author: Fejes, László
• Format: Article in Journal/Newspaper
• Language: Finnish
• Published: Suomalais-Ugrilainen Seura 2021
• Subjects: vowel harmony; vacillation; typology; neutrality; khanty
• Online Access: https://journal.fi/susa/article/view/95344; https://doi.org/10.33340/susa.95344

The main purpose of this paper is to draw attention to the inaccuracies in the article by Paul Kiparsky and Karl Pajusalu (2003). Their influential study suggests a typological classification of vowel harmony based on mostly Finnic languages and their dialects which is claimed to be supported by a particular OT analysis. However, in many cases, their example languages and dialects are hardly classifiable into the category suggested by them. Moreover, their OT analysis completely ignores cases of variation (including vacillation) in the example languages. The present study deals with the data available for Estonian (Mulgi and Northeastern dialects), Veps and Khanty, and demonstrates – mostly based on the same sources referred to by Kiparsky and Pajusalu – that the actual patterns of harmony are considerably different from the patterns described in their article. It is emphasized that many patterns do not arise from the behavior of neutral vowels but instead, they can be results of the length and the nature of the harmonic domain. The paper also introduces the notion of harmonicity slope, according to which the domain has no strict borders and in which the tendency for the vowels to harmonize gradually diminishes. The main purpose of this paper is to examine Finnic and Khanty vowel harmony. These languages play a central role in an article by Paul Kiparsky and Karl Pajusalu (2003). Their influential study suggests a typological classification of vowel harmony, which is claimed to be supported by a particular analysis in the Optimality Theory framework. In this paper a closer look will be taken at the linguistic data and the literature, and it will be argued that in many cases example languages and dialects are hardly classifiable into the suggested category. Based on data from the sources referred to by Kiparsky and Pajusalu on Estonian (Mulgi and Northeastern dialects), Veps and Khanty, it will be shown that these (varieties of) languages exhibit different patterns of (dis)harmony from what could be presumed based on their classification. Moreover, the suggested analysis ignores cases of variation (especially vacillation), typical for the example languages. In the present study, it is emphasized that many patterns do not arise from the behavior of neutral vowels, but instead, they can be results of the length and the nature of the harmonic domain. The paper also introduces the notion of harmonicity slope, according to which the domain has no strict borders, and in which the tendency for the vowels to harmonize gradually diminishes.