Langage is actually a technical term, but sadly langugae isn't. Obviously I meant langwich.
No, quite the opposite! Basically languages vary very widely, but not in a completely unbounded way, and it's not clear how the brain encodes this information. If it's encoded in a complex way, then this explains the weird patterns of variation seen, but involves the postulation of lots of implausible innate structure. So it seems more plausible that it's encoded in a simple way, but this doesn't predict the patterns of variation, and in general makes acquisition a lot harder. Howver, if acquisition is driven by general cognitive categorisation processes, then you can explain the patterns of variation while also encoding them in a simple way.