There’s an old joke: St. Peter hears a knock at the Pearly Gates. He says, “Who goes there?” A voice replies, “It is I.” St. Peter says, “Go away! We don’t need any more English teachers.”
For who other than a hard-core grammatical prescriptivist would say “It is I?” And would even the driest English teacher (not that that many are that dry anymore), arriving with others (I was about to type “friends,” but it’s hard to think that such a person could have any left), say “It is we”? Or, on the other side, answering the door, say “It is they”? I have seen “It is he,” it’s true, but…
But no one in normal English speaks that way. Not even the well-respected, highly educated people. So we’re all wrong, then? What’s with this, anyway?
This “rule” is obviously not organic to English, since it seems so awkward to pretty much every native English speaker (except the ones who have had “It is I” drummed into them and so accept it – a linguistic perversion that can be accomplished with any irregular usage if you can get people to think it’s more formal, polite, and correct, since English is capricious that way; see An historic(al) usage trend: a historical usage trend (part 1)). The idea behind it is that the is there is a copula: it equates two things. A=B. Identity means identity, so both must be the subjects: “I am he.” (If you recognize that as the first three words of “I Am the Walrus,” remember that the next four are “as you are me.” It’s not a grammar lesson from The Beatles.)
There are some problems with this reasoning. First of all, when you draw up the rules for a language, it helps if they actually describe what the language actually does, as opposed to enforcing practices that are quite different from what established usage is. If you get an idea about language and make a theory and it turns out not to be an accurate description, you shouldn’t bend the subject, you should change the theory. Otherwise you have linguistic phlogiston, a mumpsimus. And something unfortunately all too common.
Second, language is not math. Or, more precisely (since one may construct a mathematical language), English is not math. Why this isn’t incredibly obvious I don’t even know. Try performing a mathematical operation on a sentence. Give me the square root of “To be or not to be.” Language is waaaaay less tidy than math, but it’s a lot of fun. You don’t get to derive new equations and results, but linguists are discovering a lot of really fascinating weirdness. Grammatical prescriptivists, on the other hand, if they applied their thinking to the realm of math, would insist on only using certain equations in certain ways and would argue that some solutions are unacceptable because they involved, for instance, irrational numbers. They would be like the lawmakers who legislated the value of pi to be exactly 3.
And incidentally, even in math, if you establish that in this instance of an equation a=3 and b=3, you don’t necessarily change all b to a. But anyway, syntax is sequence and form; identity is semantics. Two different areas of grammar.
Third, English is not Latin. Many of prescriptivists’ ideas, such as this one, are derived from and/or supported by appeals to Latin grammar. You might as well use a barbecue to bake a cake, or dress patterns to make shoes. Each language has its own set of rules, its own parameters, its own ways of handling this and that. French is descended from Latin but you could never say “C’est je” in French, so why would we insist that English use “It is I” just because Latin, which English is not based on, does similarly?
The real ace in all of this is that “It is I” is supposedly equating “It” and “I”. OK, what’s the “It” here? If I say “I am he,” then there’s a “he” we were talking about who turns out to be me. But where’s this “it”? There’s no object I’m claiming is me. The it is actually empty. The only reason it’s there is because in English we require every finite verb to have something in the subject position. Not every language does. In Chinese you can say you shu, “have book”, to mean “There’s a book”; you can say shi wo, “is I/me”, to mean “It’s me” (or “It is I” if you’re one of those people). But we have to put in these empty its and theres in English for it to be a complete sentence. (We may say, casually, Got it, but even casually we don’t say Is me instead of It’s me.)
So it’s is really an existential predicate. But it’s bootless to argue that since there’s only one real thing there (me), it must be the subject. The point is precisely that it’s not the subject because that’s not how English syntax works. A thing can’t be both subject and predicate. We can’t say I am to mean It’s me, because it means something else, so we have an existential verb and an empty subject, and make me the predicate.
Which leads us to another fact of English syntax: the case filter. Put simply, English nouns and pronouns are by default in the objective (accusative). For each finite (conjugated) verb, there has to be one subject, which means one noun phrase in the subject (nominative) case, and that noun phrase is the one that is specifying the verb – it’s in the “subject” position. We don’t do this with non-finite verbs: I want him to go, I want to see him going. Those hims are the subjects of an infinitive and a participle, but they’re still objective. But if the verb is finite, one noun phrase and one only is treated as its subject: I desire that he go. The one you want is him. (Note that there can be inversions: What fools are we! Sam I am!)
And that is a real rule of English. One that we all use all the time without having someone tell us, one that guides our comprehension and usage. Not phlogiston. There is no cake batter dripping from the grill. So if someone at your door says “It is I,” you’re fully enfranchised to say “Go to hell!” (You probably don’t want them at your party anyway.)