However, if it were the case, you don't need to explain it, only observe it. For example, I think we can say from observation that there needs to be a minimal complexity. n-neurons cannot be self aware and n+1 can. I'm sure it's not just the number, but you get the point. The gradient of self-awareness need not start at the gradient of intelligence, nor need it have the same slope.
I don't believe in non-physical explanations. We don't know the mechanism though, so we do actually do need to explain why some permutations of matter, as you put it, (i.e. the permutation in brains) has this sense, and others (matter in rocks, matter in computers?) does not. Although I'd think process, rather than just arrangement.
I can recommend the book „How language began“ by Daniel Everett, where he demonstrates convincingly that our path to language, culture and consciousness was not a single switch, but evolutionary process. Recent research of cetaceans shows that their language and culture may overlap in sophistication with our ancestors.