Who’s afraid of homophones? Or how bringing contemporary participants to the lab can provide new insights to historical cases of language variation and change

Homophony avoidance has often been claimed to be a mechanism of language change. Throughout the history of Dutch, several instances of morphophonological change have been argued to have occurred or been inhibited to avoid homophony between present and past tense (Van Bree 1987; Van Loon 2014). For example, the schwa apocope in Early Modern Dutch that affected nearly all Dutch words did not take place in weak preterites (hoopte PST-3SG ‘hope’), which would render them indistinguishable from the present (hoopt PRS-3SG). It did take place however in irregular preterites (kocht < kochte PST-3SG ‘buy’) which remained distinguishable from their present counterparts (koop PRS-3SG) without schwa.
Yet, the question remains how plausible this homophony avoidance explanation is. While the topic is not uncontroversial (King 1967, Lass 1987; 1997a; 1997b: 355-361, Sampson 2013), bringing contemporary language users to the lab can provide some answers. Iinitial experimental research indeed suggests homophony avoidance to be a cognitively plausible explanation for certain patterns in language change (Kaplan 2011; Kaplan & Muratani 2015; Yin & White 2018). In this paper, we seek further experimental substantiation of homophony avoidance as a cognitive factor in language change. As no speakers of Early Modern Dutch exist today, we use a contemporary case of homophony between present and past tense to test the plausibility of this mechanism. We focus on Dutch non-verbs ending in a dental which, when used in the preterite plural, become homophonous with the present plural (compare vatten PST-3PL to vatten PRS-3PL ‘understand’). Language users can avoid this homophony by switching to a perfect (ze hebben gevat PERF-3PL) which has become semantically interchangeable with the preterite in many contexts in present day Dutch.
In a Wug-type testexperimental task, participants (n=232) were presented with a forced-choice-task requiring them to choose between the perfect or preterite of non-verbs to fill a blank in a stimulus sentence. The non-verbs either showed a double dental, causing the preterite plural to be both homophonous and homonymous with the present plural (compare vatten PST-3PL to vatten PST-3PL), a single dental, causing the preterite plural to be homophonous, but not homonymous with the present plural (compare duldden PST-3PL to dulden PST-3PL ‘tolerate’) or no dental, causing no homophony. Two between-subject conditions were created: one with only singular verbs and one with only plural verbs. Finally, to back up these experimental data, we also carry out a corpus study on the same case study in contemporary Dutch.
Results from both the experimental data and corpus data show that language users indeed tend to use the perfect more often for double dental verbs (and to a lesser extent single dental verbs) in the plural, suggesting homophony avoidance can be a plausible mechanism in language variation. More generally, this paper shows how experimental research can be an important asset in historical linguistics by providing a deeper understanding of cognitive aspects of language variation and change as well.

Kaplan, Abby. 2011. How much homophony is normal? Journal of Linguistics 47(3). 631–671.
Kaplan, Abby & Y. Muratani. 2015. Categorical and gradient homophony avoidance: Evidence from Japanese. Laboratory Phonology 6(2). 167–195.
King, Robert D. 1967. Functional load and sound change. Language 43(4). 831-852
Lass, Roger. 1987. The shape of English. Londen: Dent.
Lass, Roger. 1997a. Arse longa, vita brevis: last words on ‘harmful homophony’. Studia Anglica Posnaniensia XXXII. 21-31.
Lass, Roger. 1997b. Historical linguistics and language change. Cambridge: Cambridge University Press.
Sampson, Geoffrey. 2013. A counterexample to homophony avoidance. Diachronica 30(4). 579-591.
Van Bree, Cor. 1987. Historische grammatica van het Nederlands. Dordrecht: Foris.
Van Loon, Jozef. 2014. Historische fonologie van het Nederlands. Schoten: Universitas.
Yin, Sora Heng & James White 2018. Neutralization and homophony avoidance in phonological learning. Cognition 179. 89–101.