Does 'he dived' take longer than 'he dove'? An experimental approach to iconicity in past tense morphology

In the Germanic languages, two types of inflection exist. Some verbs take the strong inflection, where ablaut is used to form past tense and past participle (e.g. drive-drove-driven). Other verbs take the weak inflection, where a dental suffix is used (e.g. play-played-played). There are also verbs that vary in their preterite and past participle form (e.g. dive-dived or dive-dove). In a diachronic corpus study, De Smet & Van de Velde (2020) show that in Dutch this variation can be exapted to express aspect in an iconic manner. Their results indicate that the longer weak preterites (e.g. schuilde ‘hid’) are used more often in durative contexts, while the shorter strong variants (e.g. school ‘hid’) are used more often in punctual contexts. For the past participles, this image is reversed: the longer strong variants (e.g. gescholen ‘hidden’) are used more often in durative contexts, while the shorter weak variants (e.g. geschuild ‘hidden’) are used more often in punctual contexts. In this paper, we seek experimental validation for these results. Participants (N=664) were presented with a forced choice task where they had to choose between weak or strong preterites and past participles of nonce verbs in sentences suggesting either a durative or a punctual context. We worked with three different between-subject conditions, namely preterite singular, preterite plural and past participle. Each survey consisted of 20 target items and 10 filler items. The target items were 20 nonce verbs of the five most productive ablaut subclasses in Dutch (cf. Knooihuizen & Strik 2014). Every verb only appeared once in each survey to prevent priming effects. Half of the target items were presented in a durative context (which was suggested through the use of adverbials), the other half in a punctual context. Hence aspect was manipulated within subject. Results were analysed using a generalized linear mixed effects model with random effects for subject and item and random slopes for aspect by subject and by item. Though no overall effect of aspect on verb inflection was found, results indicate an iconic trend for verbs of one specific ablaut subclass that supports the corpus results from De Smet & Van de Velde (2020). Because this ablaut class shows the most variation in real language use (of the five subclasses selected for the experiment), it could be that language users need to be familiar with a certain amount of variation for a specific class in order to become routinized in exapting the variation to express aspect. Furthermore, the durative-punctual distinction was also found to be portrayed in yet another iconic manner: verb forms with vowels that are sound symbolically associated with long slow movements were used more often in durative contexts, while verb forms with vowels that are associated with quick, short movements were used more often in punctual contexts. References – De Smet, Isabeau & Freek Van de Velde. 2020. Semantic differences between strong and weak verb forms in Dutch. Cognitive Linguistics 31(3). 393-416. Knooihuizen, Remco & Oscar Strik. 2014. Relative productivity potentials of Dutch verbal inflection patterns. Folia Linguistica Historica 35. 173-200.