Evidentiality represents the abstract grammatical system encoding the source of information (Aikhenvald, 2004), and one hypothesis is that evidential systems' development in languages could be explained through the action of cognitive biases on language learning (Fedzechkina et al., 2012). Consequently, this article tests the learnability of different two-category (Aikhenvald, 2004) evidential systems, in a semi-artificial language learning experiment with Bulgarian speakers, and hypothesizes to reflect the typological pattern of those systems' distribution in the world languages. The alternative hypothesis predicts that the learnability of the evidential systems would depend on the languages that the participants speak and would be influenced by a morphosyntactic transfer from participants' native language - the grammatically evidential “Non-firsthand versus ‘everything else’“ (Aikhenvald, 2004) Bulgarian. Despite the fact that the mixed-model analyses performed on accuracy and response time did not display any significant effect or interaction (p > 0.1), the small trends showed by the results possibly reflect a salient cognitive and semantic distinction between direct and indirect evidential information that is easier for acquisition, compared to marking which violates this categorisation.
|Title of host publication||Proceedings of Linguistic Evidence 2018|
|Subtitle of host publication||Experimental Data Drives Linguistic Theory|
|Place of Publication||Germany|
|Publication status||Published - 2019|
- L2 acquisition
- L1 transfer