I am a descriptivist, as are most people with training in linguistics but few editors, who tend to mistake house style guides for commandments from above rather than the prejudgements of difficult-to-negotiate words that they really are. However, descriptivists might be permissive, but they still have a notion of right and wrong. "Correctness" in usage surely means -- if it means anything -- broadness of use. If most people mean a certain thing when they say a word, that's what that word means. It's possible to frame notions such as agreement in number by this metric (although it's a lot easier simply to say that it's mandatory in English without discussing why). Plural nouns agree with plural verbs in English because most people make them agree. Language is pretty much "democratic" in this sense: if you are in a minority, you are wrong, and the smaller the minority, the wronger you are. It's clear, or should be, that there will be a spectrum of "wrongness" (or spectra, because what is wrong in one context or for one group is often correct for another: so it is "wrong" to write "color" in English but correct to do so in American English). There are reasons to weight the "votes", of course, so that if the usages favoured by the better educated, or newspapers, or similar sources that use language in particular ways, are not more correct, they are felt to be by most speakers. An example of this spectrum: using "thus" to mean "because of this" is only slightly wrong (probably a minority of writers use the "correct" usage and it's only the weightedness that pushes it into "correctness"); but using "the dog are barking" is as wrong as you can get in standard English, spelling errors aside.
I don't think that an extreme descriptivism works. In
this post, a descriptivist misanalyses a speech act. Let's deal first with the misanalysis.
If you look to the right, Treasure Island's having their show right now.
"their" is not used because "Treasure Island" has indeterminate gender but because collective entities are often used with a plural verb by English speakers. This happens even in sentences that have already displayed correct agreement. It's a simple outcome of confusion over whether entities that are aggregations of people should be treated as plurals or singulars. (I noted this in a previous post, which I can't find, but "Treasure Island" can be compared with "the crowd" or "the committee".)
Even if this analysis were correct though, I do not see how finding one example of a usage makes that usage correct. If every other speaker of a language denies it as a correct usage, how can one person's usage be elevated to the status of the other billion's?
In any case, for inanimates, there is a readymade alternative to he and she where gender is not known, and the speaker would, by this analysis, be considered wrong by nearly all speakers of English not to have used it. It's it.
I use the singular they, and in my view it's the best candidate for the nongender-specific pronoun. But is it a good substitute for "it"? No, I don't think so.