The second talk in the #futureoflinguistics webinar is by @drculbertson: The future of linguistics is multidisciplinary, multimodal, multicultural
What are elements of a theory of language?
- features of the linguistics system
- features of the cognitive system
- cognition-external forces
Different approaches have weighted these differently, but they are all relevant for a theory of language #futureoflinguistics
- features of the linguistics system
- features of the cognitive system
- cognition-external forces
Different approaches have weighted these differently, but they are all relevant for a theory of language #futureoflinguistics
Traditional approaches have relied on introspection, grammars, fieldwork, typological samples, small corposal, but mostly focussed on spoken European languages and cultures - there are however, important limitations to these data sources #futureoflinguistics
So the #futureoflinguistics is also related to the future of evidence in linguistics - which needs to integrate multidisciplinary methods, needs to adopt a multimodal approach, and a multiculutral approach going beyond European languages and WEIRD populations
For example, new sources of evidence are behavioural experiments, bayesian & neural network models and bayesian phylogenetics, largescale corpora, emerging languages, cross-linguistic & cross-cultural studies - which should all play a bigger role than b4 #futureoflinguistics
Culbertson next talks about Artifical Language Learning experiments, which can uncover (domain-general) biases in learning - e.g. a preference for harmonic patterns which holds for speakers of both harmonic & non-harmonic languages #futureoflinguistics https://www.sciencedirect.com/science/article/pii/S0010027720302110
Preferences for harmonic patterns are especially expressed in child learners. e.g., English-, French-, & Hebrew-speaking children all prefer harmonic patterns and even generalise harmonic patterns in artifical language learning experiments #futureoflinguistics
Although harmonic orders are most common, there are many languages without such harmonic patterns - experiments and computational models show this is a weak (deafeasible) bias, likely amplified over time via cultural transmission #futureoflinguistics
Children learn by observing objects in the world &build a representation of conceptual structure. This structure is mapped to a syntactic order& filtered through a simplicity bias leading to different population-level typologies via cult. transmission #futureoflinguistics
Behavioural experiments indeed show orders corresponding to a particular underlying conceptual structure emerge in improvised gestures - namely they exhibit homomorphy, which itself is based on properties of what the world is like #futureoflinguistics
https://muse.jhu.edu/article/764695#fig01
https://muse.jhu.edu/article/764695#fig01
Generally, these results show that we need to look beyond the traditional toolbox of methods, and integrate linguistic, cognitive, and non-cognitive mechanisms - this also means (interdisciplinary) collaboration is key for the #futureoflinguistics