This is an interesting question, but part of answering it may involve rethinking the assumptions behind it.
The main issue is how we could gain any evidence about it. Obviously, no writing could be preserved from pre-linguistic (or even preliterate) societies. Cave paintings and pictographs and other art forms exist from preliterate societies but we can't really know if those societies had verbal languages or not; if they communicated by visual signals or gesture, that would constitute a sort of "language" (just as ASL does).
As every society that now exists has some form of language, and even animals use limited sounds or gestures to communicate, the closest we can come to investigating this is by looking at feral children, who seem to have missed a key window for language development. Even in those cases, though, the only way we can investigate their inner thought processes is by teaching them language and talking with them.
Wittgenstein famously stated that the limits of my language are the limits of my world. Although we can imagine systems of symbolic thought quite different from languages as we know them, in so far as such systems could be mediums for thought, we would call them languages.