But just because a language doesn't actively use something doesn't mean the "universal" rule governing it is meaningless.

Well, no, but it does suggest that what we think is hardware is really software and learned behaviour.

To use a crude analogy - just because your processor does floating point maths doesn't mean that your software has to use the floating point feature.

The gotcha is the 'universal' concept. It seems to be a bit of a blunt instrument for describing a complex set of related phenomena.

In this case there's not enough information to decide if there genetic differences, which would be interesting enough in itself, or whether the differences are learned.

Perhaps a more interesting question is - whose perceptual grammar is more complex? Ours, or theirs?