Further to the post below about the relationship between language and thought, a friend has just remarked to me that the absolute (East-West-North-South) spatial reference system in the language of the Kuuk Thaayorre is in fact a relative system, relative to the magnetic poles or (since they are unlikely to have known about the poles) relative to the movements of the sun.  Accordingly, such a language would have been unlikely to have developed in parts of the world with continuous cloud cover.   Which observation brought to mind a famous article by French mathematician and physicist Henri Poincare, where he considered what type of mathematical physics humans may have  developed if the earth had been always covered in cloud:  no theory aiming to predict the return of meteors, no models of planetary motion, not much study of ellipses and related number theory, and perhaps a theory of gravitation long delayed.  (Thanks:  DW).