I'd argue in the first instance that few civilisations are understood on the sole basis of their writings, in the majority of cases there's at least a limited amount of material archaeology, although I'd agree if you said that ancient writings can be hugely important to a broader understanding of any given culture.

To the writings: grammatical structure is a fundamental part of any given language and evolves over that language's "life" just as its vocabulary does. Right now I'm writing in English using grammar that I hope is broadly correct, but this structure and my vocabulary could seem alien to an English reader in the 1600s. Different readers in the 1600 might each translate my sentences into very different sentences depending on their own locality, dialect and personal idiosyncrasies. Equally, if I translated every word to its German or French counterpart the grammatical structure would immediately be incorrect, even if every word was translated properly.

I'm saying all that to illustrate that our "grammatically correct" may not be correct in another language at any given time. Where we only have a few examples of a language we can only be informed about grammar, structure, vocabulary and so on by those few examples. Where those few examples differ in age by a few hundred years, or originate a few hundred miles apart, we don't have a base reference to tell us what dialectic, geographic or evolutionary changes or idiosyncracies we might be seeing across the different texts.

To answer your question: we can't tell, and there may well be never be a "correct" benchmark to assess it against.