I think we can’t, because grammar and spelling aren’t equivalent to chemical bonding.
But language does expose the flaw in you and your compatriot’s claims - no matter how much variation is possible by changing the letters and words in English sentences to form new ones, you’ll never see that there is an unknown and unreachable hill elsewhere in the fitness landscape where German sentences are found. And another for Russian, another for Greek and even more remote ones for Chinese, Arabic and Sanskrit. So no matter how well you estimate the number of English sentences in the search space, your estimate will always be woefully short of the actual answer - just like your estimate of the number of protein that can perform a given function may also be woefully short of the actual answer.