Revisiting regularity: a comprehensive corpus study of A-Umlaut in weak Old High German participles

Since the 19th century, the debate on sound change within linguistics has centred around two competing options: the regularity hypothesis, first proposed by Osthoff and Brugmann’s (1890: XII) claim that sound change occurs following exceptionless rules and demonstrated by Bloomfield’s (1984: 351) assertion that “phonemes change”, and lexical diffusion, which suggests that sound changes gradually through the lexicon (Wang 1969), and potentially that there is a correlation between frequency of word and the extent and rate at which they undergo sound change (Bybee 2002). Whilst Labov (1981) originally suggested that both mechanisms were at play, he recently concluded that “the principle of regular change can be applied with good confidence that you are in touch with linguistic reality” (Labov 2020). In this paper, I focus on the phenomenon of A-Umlaut (also known as Brechung) (Schmidt, Langner & Wolf 2007: 223; Braune 2018: 85), which is the anticipatory lowering of Proto-Germanic /i/ and /u/ to /e/ and /o/ except when blocked by a consonant cluster containing a nasal (Salmons 2012: 126). Traditional grammars, however, acknowledge that this change did not occur regularly: Braune (2018: 52) claims that no regular rule can be formulated and Schmidt (2007: 223) specifically notes that this change was grammatically conditioned, with all preterite participles in the first ablaut-series retaining the /i/. This grammatical condition is irreconcilable with mechanical regular sound change.

This paper quantitatively examines A-Umlaut in Old High German (OHG) regular past participles, as attested in the Deutsch Diachron Digital Referenzkorpus Altdeutsch (Zeige, Lars Erik et al. 2022; initially queried using Krause & Zeldes 2016; also directly analysed using pandas: The pandas development team 2022; McKinney 2010). Initial analysis of the high frequency participle gitan is used to identify three variant outcomes for participles which should display A-Umlaut in OHG. This analysis is then quantitatively extended to all perfect participles in the corpus. Finally, the effect of frequency is investigated by comparing the percentage of outcomes which show A-Umlaut for each lemma against the frequency of the lemma, and of the participle forms of the lemma. This analysis is performed across the whole corpus, and within the Alemannic and Bavarian dialects. I find that A-Umlaut was not complete in the preliterary period (contra Schmidt, Langner & Wolf 2007: 223), but is attested until the 11th century. Additionally, of the 2434 regularly formed past participles, 772 (31%) do not display A-Umlaut. I further find that there is no statistically significant correlation between frequency of either lemma or participle and degree of application of A-Umlaut to the participle, either across the whole corpus, or in either of the dialects examined.

Historical linguistics aims to “arrive at models of language that are quantitatively driven from evidence” (Jenset & McGillivray 2017: 44). I argue that, as this irregularity cannot be dismissed as fossilised forms within an otherwise regular system, or be explained by diffusion and lemma frequency, that these two models are insufficient to explain A-Umlaut. This result highlights the value of quantitative methods in historical linguistics. (498 words)

References

ANNIS4. https://corpus-tools.org/annis/download.html

Bloomfield, Leonard (1984): Language. Chicago, IL: University of Chicago Press.

Braune, Wilhelm (2018): Althochdeutsche Grammatik I. Althochdeutsche Grammatik I. De Gruyter. doi:10.1515/9783110515114.

Bybee, Joan (2002): Word frequency and context of use in the lexical diffusion of phonetically conditioned sound change. Language Variation and Change. Cambridge University Press. 14(3). 261–290. doi:10.1017/S0954394502143018.

Jenset, Gard B. & Barbara McGillivray (2017): Quantitative Historical Linguistics. Bd. 1. Oxford University Press. doi:10.1093/oso/9780198718178.001.0001.

Krause, Thomas & Amir Zeldes (2016): ANNIS3: A new architecture for generic corpus query and visualization. Digital Scholarship in the Humanities 31(1). 118–139. doi:10.1093/llc/fqu057.

Labov, William (1981): Resolving the Neogrammarian Controversy. Language. Linguistic Society of America. 57(2). 267–308. doi:10.2307/413692.

Labov, William (2020): The regularity of regular sound change. Language (Baltimore). Linguistic Society of America. 96(1). 42–59. doi:10.1353/lan.2020.0001.

McKinney, Wes (2010): Data Structures for Statistical Computing in Python. 56–61. Austin, Texas doi:10.25080/Majora-92bf1922-00a.

Osthoff, Hermann & Karl Brugmann (1890): Morphologische Untersuchungen auf dem Gebiete der indogermanischen Sprachen. Leipzig ; Hirzel.

Salmons, Joe (2012): A history of German: what the past reveals about today’s language (Oxford Linguistics). 1st ed. Oxford, U.K: Oxford University Press.

Schmidt, Wilhelm, Helmut Langner & Norbert Richard Wolf (Hrsg.) (2007): Geschichte der deutschen Sprache: ein Lehrbuch für das germanistische Studium. 10., verb.erw. Aufl. Stuttgart: Hirzel.

The pandas development team (2022): pandas-dev/pandas: Pandas. Zenodo. doi:10.5281/zenodo.7344967.

Wang, William S-Y. (1969): Competing Changes as a Cause of Residue. Language. Linguistic Society of America. 45(1). 9–25. doi:10.2307/411748.

Zeige, Lars Erik, Schnelle, Gohar, Klotz, Martin, Donhauser, Karin, Gippert, Jost & Lühr, Rosemarie (2022): Deutsch Diachron Digital - Referenzkorpus Altdeutsch. Humboldt-Universität zu Berlin. doi:10.34644/LAUDATIO-DEV-MIXVDNMB7CARCQ9CABMW.