@wallhackio i feel like the whole field (a) suffers from epistemological failings in approach and (b) has been eaten by A·I

@Lady could you elaborate on a

@wallhackio the goal of computational linguistics is to transform human language into a machine-processable model which you can then apply technological processes to. the core questions of the field are “what does this model look like” and “how do you do those transformations”. unfortunately, the model tends to be something which looks like a Western, Enlightenment, Structuralist approach to semantics, which i think is ultimately incorrect. firstly, I’m not convinced that English and Japanese (for example) DO model meaning in the same way; i think the question is open whether there even IS a model which all languages can be mapped to unproblematically. if there isn’t, you have to choose a model, which has implications for how fully different languages can be supported. the trend of course is to fully support English and let other languages fall where they may. secondly, i think the Structuralist approach is inadequate even for English; the idea that there is a comprehensible ontology of concepts and language is simply a means of expressing relationships between those concepts has been, i think, discredited since the 1970s. one of the reasons why A·I is so effective is because it doesn’t bother with this notion; it has no need for a backing ontology. but of course this is also why A·I sometimes produces nonsense