brain-gear

A guest post by Stefan Hartmann over at A Replicated Typo from back in May, and the ensuing comments on Reddit, inspired me to weigh in on the issue of Universal Grammar – or rather, the issue of discussing Universal Grammar.  I’d like to highlight something that too many people involved in this discussion overlook and which leads to a lot of unnecessary finger-pointing and eye-rolling.  That is to simply to define what one means by “Universal Grammar” in the context in which you are using it.  In particular, making clear whether one views the human language faculty as having principles that are domain general, domain specific, or some combination of these, would go a long way to help communication between “formalist” and “functionalist” camps in linguistics, something I think is very much worthwhile.

Universal Grammar in its most basic conceptualization is simply the observation that our brains have the capacity to learn and use language and that this capacity appears to be in-born and uniquely human.  One might recall Chomsky’s well-known comments regarding rocks and kittens which just amount to the observation that humans learn language and other things don’t.  I am unaware of any linguist, or really any person, who could disagree with this observation, unless they have a rather unconventional working definition of “language”.

So, everyone should be in agreement, right?  As anyone with some experience in the field knows, that is not the case.  When what one means by “Universal Grammar” is not made clear from the outset of a discussion, arguments that ensue are too often fruitless.  Many working within generative theoretical frameworks appear baffled at commentary from non-generative linguists who claim there is not sufficient evidence for the existence of UG.  Of course, those non-generativists almost certainly do not take issue with the observation that humans have some cognitively-grounded mechanism(s) that allows for language acquisition and use, though they are often not convinced by the Argument from the Poverty of the Stimulus, which is usually taken to necessitate domain-specific Universal Grammar.  Where the non-generativist exclaims, “There is no solid evidence for UG!,” they usually aren’t denying humans’ innate ability to acquire and use language, but rather they question the assumption that there are absolute and in-born constraints that apply exclusively in the cognitive domain of language.  In other words, they doubt the validity of domain-specific Universal Grammar.  Similarly, when a generativist exclaims, “How can you deny the existence of UG!”, they are in that instance often referring to the “rocks and kittens” version of the UG argument, which probably no one disagrees with.

Non-generativist perspectives on grammar often maintain an outlook in line with domain-general Universal Grammar, although this terminology might be avoided.  This is not to say that everyone working within generativist frameworks is disinterested in domain-general approaches to grammatical constraints, but it is hard to deny the role that the assumption of domain-specific Universal Grammar has had in shaping generative models of grammar.  Non-generative theorists may take the position that there is no reason to assume domain-specific constraints on grammar (though many leave open the possibility that such constraints may be discovered), and that language structure arises as a result of constraints which are not necessarily exclusive to the domain of language.

Arguments surrounding the validity of the Poverty of the Stimulus, and whether domain-specific Universal Grammar is necessary, is another matter; one I plan to take up in future posts.  This, it seems, should be the focus of discussions regarding UG.