"It's me" vs "It is I"

     [This is a fast-access FAQ excerpt.]
   Fowler says:  "me is technically wrong in It wasn't me etc.;
but the phrase being of its very nature colloquial, such a lapse is
of no importance".
   The rule for what he and others consider technically right is
*not* (as is commonly misstated) that the nominative should *always*
be used after "to be".  Rather, it is that "to be" should link two
noun phrases of the same case, whether this be nominative or
accusative:
    I believe that he is I.  Who do you believe that he is?
    I believe him to be me.  Whom do you believe him to be?
According to the traditional grammar being used here, "to be" is not
a transitive verb, but a *copulative* verb.  When you say that A is
B, you don't imply that A, by being B, is doing something to B.
(After all, B is also doing it to A.)  Other verbs considered
copulative are "to become", "to remain", "to seem", and "to look".
   Sometimes in English, though, "to be" does seem to have the
force of a transitive verb; e.g., in Gelett Burgess's:
    I never saw a Purple Cow,
      I never hope to see one;
    But I can tell you, anyhow,
      I'd rather see than be one.
The occurrence of "It's me", etc., is no doubt partly due to this
perceived transitive force.  In the French C'est moi, often cited
as analogous, moi is not in the accusative, but a special form
known as the "disjunctive", used for emphasis.  If etre were a
transitive verb in French, C'est moi would be Ce m'est.
   In languages such as German and Latin that inflect between the
nominative and the accusative, B in "A is B" is nominative just like
A.  In English, no nouns and only a few personal pronouns ("I",
"we", "thou", "he", "she", "they" and "who") inflect between the
nominative and the accusative.  In other words, we've gotten out of
the habit, for the most part.
   Also, in English we derive meaning from word position, far more
than one would in Latin, somewhat more than in German, even.  In
those languages, one can rearrange sentences drastically for
rhetorical or other purposes without confusion (heh) because
inflections (endings, etc.) tell you how the words relate to one
another.  In English, "The dog ate the cat" and "The cat ate the
dog" are utterly different in meaning, and if we wish to have the
former meaning with "cat" prior to "dog" in the sentence, we have to
say "The cat was eaten by the dog" (change of voice) or "It is the
cat that the dog ate."  In German, one can reverse the meaning by
inflecting the word (or its article):  Der Hund frass die Katze
and Den Hund frass die Katze reverse the meaning of who ate whom.
In Latin, things are even more flexible: almost any word order will
do:
    Feles edit canem
    Feles canem edit
    Canem edit feles
    Canem feles edit
    Edit canem feles
    Edit feles canem
all mean the same, the choice of word order being made perhaps for
rhetorical or poetic purpose.
   English is pretty much the opposite of that:  hardly any
inflection, great emphasis on order.  As a result, things have
gotten a little irregular with the personal pronouns.  And there's
uncertainty as to how to use them; the usual rules aren't there,
because the usual word needs no rules, being the same for nominative
and accusative.
   The final factor is the traditional use of Latin grammatical
concepts to teach English grammar.   This historical quirk dates to
the 17th century, and has never quite left us.  From this we get the
Latin-derived rule, which Fowler still acknowledges.  And we *do*
follow that rule to some extent: "Who are they?" (not "Who are
them?" or "Whom are they?")  "We are they!" (in response to the
preceding)  "It is I who am at fault."  "That's the man who
he is."
   But not always.  "It is me" is attested since the 16th Century.
(Speakers who would substitute "me" for "I" in the "It is I who am
at fault" example would also sacrifice the agreement of person, and
substitute "is" for "am".)