The article, and many professional functionalist lingusts, confuse Chomsky's Universal Grammar theory and its later "parameters" revision with the much more general hypothesis, also advocated by Chomsky, that human language grammar is somehow computable in a logical sense, formed through the application of objective and universal (with a small "u") logical rules. They prefer to view language rather as some sort of nebulous and inherently unquantifiable expression of the human spirit that is intrinsically bound to the unique cultural attributes of the society that produces it, a precious snowflake different from all others, which cannot be related to other languages through universal logical principles.
They feel that the notion of universal (small U) principles of grammar somehow robs language of some essential human qualia by reducing it to a mere mechanism, and moreover many postmodernist linguists believe that the very idea of objectivity in language, to include supposed universal grammatical principles, is somehow part of a broader imperialistic plot by evil white heterosexual men to oppress and subjugate the glorious Other through malicious efforts of universal grammatical classification. (Yes, seriously. Your tax dollars support more than a few professors who believe this is true, in all seriousness, and teach it to impressionable young students every day.)
These fears do not, however, have any bearing whatsoever on whether human language is, as a matter of fact, computable, the result of some set of universal logical principles. It either is or it isn't, and wishful thinking about how we would like to conceive of language has no bearing on whether that is true or not. That is not how scientific knowledge works. (Of course, many denounce the scientific method itself as merely another tool used by evil white straight men to oppress the Other, and this despite the fact that it was invented by Arabs.)
The anti-computation camp frequently conflates Universal Grammar or parameters theory with the question of computability in general. When any evidence comes out that shows that a specific portion of UG is probably not true, they crow triumphantly about how language is really fundamentally something ineffable and pseudo-mystical, not able to be computed by logic.
UG and the computability question are not the same thing. UG is a specific proposed model of language computation. It has largely been falsified, shown to be inconsistent with the empirical evidence of how actual languages work. That is fine. This does not imply anything at all about whether language is universally computable according to some model.
It just tells us that language is not computable according to that model, not that it is not computable according to any model.
>> When any evidence comes out that shows that a specific portion of UG is probably not true, they crow triumphantly about how language is really fundamentally something ineffable and pseudo-mystical, not able to be computed by logic.
I think the overwhelming mass of work critical of Chomsky treats it as a computational object; the emphasis changes from global computability to local inference: what can a learner infer about the structural properties of a language? How can people parse complex sentence structures and ascribe meaning to them? Hence the rapid adoption of machine learning for explaining human behavior in psycholinguistics, developmental linguistics, and historical linguistics. Nothing pseudo-mystical there.
Sure, of course not all linguists who disagree with UG are anti-computationalists. (I don't know about "the overwhelming mass," but that is possibly the result of my exposure to a local concentration of anti-computationalists.) Many tentatively accept some sort of computability while rejecting the specific model of UG, as you correctly note.
My comment was directed at the author of the specific article, and at many functionalist and postmodernist linguists, who actively conflate the question of computability in general with the validity of UG as a specific model.
They feel that the notion of universal (small U) principles of grammar somehow robs language of some essential human qualia by reducing it to a mere mechanism, and moreover many postmodernist linguists believe that the very idea of objectivity in language, to include supposed universal grammatical principles, is somehow part of a broader imperialistic plot by evil white heterosexual men to oppress and subjugate the glorious Other through malicious efforts of universal grammatical classification. (Yes, seriously. Your tax dollars support more than a few professors who believe this is true, in all seriousness, and teach it to impressionable young students every day.)
These fears do not, however, have any bearing whatsoever on whether human language is, as a matter of fact, computable, the result of some set of universal logical principles. It either is or it isn't, and wishful thinking about how we would like to conceive of language has no bearing on whether that is true or not. That is not how scientific knowledge works. (Of course, many denounce the scientific method itself as merely another tool used by evil white straight men to oppress the Other, and this despite the fact that it was invented by Arabs.)
The anti-computation camp frequently conflates Universal Grammar or parameters theory with the question of computability in general. When any evidence comes out that shows that a specific portion of UG is probably not true, they crow triumphantly about how language is really fundamentally something ineffable and pseudo-mystical, not able to be computed by logic.
UG and the computability question are not the same thing. UG is a specific proposed model of language computation. It has largely been falsified, shown to be inconsistent with the empirical evidence of how actual languages work. That is fine. This does not imply anything at all about whether language is universally computable according to some model.
It just tells us that language is not computable according to that model, not that it is not computable according to any model.