Why grammatical theory is so damn hard

[sacundim]

It's all the insanely fine distinctions one has to make in order to get the seemingly simplest-looking stuff right. Like, say, the distinction between so-called "present participles" and "gerunds" in English:

• The next showing of the movie is at 8 PM. (Participle)
• The enemy destroying the city was an unexpected development. (Gerund)

What must one say to properly describe these constructions? Most people's initial attempt is to say that it's simple: in both, you've taken a verb and made it into a noun. They're wrong, of course.
First, one must understand the notion of a predicate and its arguments. Verbs are predicates, and their subjects and objects are arguments. But nouns can also be predicates, and have their own arguments. Verbs in English combine with their arguments directly, with no need for additional words or construction; nouns, however, need their arguments to either be possessive or to appear with the preposition of. Thus:

• The enemy destroyed the city.
• the enemy's destruction of the city

In this regard, participles behave like nouns, but gerunds behave like verbs. It's reasonable to say that participles are a case of a noun being derived from a verb; but gerunds are cases of a noun phrase being derived from a verb phrase.
But wait! We're not done! We also need to state the rule(s) whereby given a verb like show, we can form a noun showing. Most people, again, are inclined to think this is very simple: you just stick -ing at the end of the verb. These people are wrong again, of course.
Why? Because they've failed to take into account the word form that's usually rendered as showin' (and popularly referred to by the completely inaccurate description "dropping the g"; there's no g in the pronunciation, and nothing is actually dropped). Most people, again, unwilling to accept the possibility that this might be complicated stuff, will just tell you now that there's just a rule that changes -ing to -in'. They'll give you a really blank stare when you ask them to explain, then, how come one can't apply that rule to sing and get sin.
How does one state the rules involved, then? I'm going to pull an Arnold Zwicky on you guys*: I'm going to call the abstract thing that encompasses both showing and showin' by an utterly meaningless, seemingly impractical name: I'm going to call it paradigm cell #324. (What's a "paradigm cell"? Well, as I mentioned in an entry a couple of weeks back, it's one of the boxes in a conjugation table, as opposed to the actual word inside the box.)
Now, here's the rules:

1. The inflectional paradigms of English verbs contain, among others, paradigm cell #324. (That is, conjugation tables in English have a specific "shape.")
2. Various constructions in English grammar call for paradigm cell #324. These include, but are not necessarily limited to, the progressive (I am running), the rule for deriving participial nouns (the next showing of the movie), rules for forming certain forms of non-finite clauses (the enemy destroying the city), and possibly, rules for forming noun phrases out of said non-finite clauses (gerunds). (And be grateful that I'm not calling them "construction #58," "construction #97," "construction #125" and "construction #444"...)
3. Given a verb stem (which is, in fact, just whatever thing happens to be inside of one particular paradigm cell for the verb in question), there are two distinct rules that can be applied to determine the pronunciation of the word inside paradigm cell #324 for the verb in question. One of them is to append -ing to the stem; the other is to append -in' to the stem.
4. There are no phonological, morphological or syntactic conditions that call exclusively for either of these rules (I can't recall if there are statistical effects here; there could be). The choice of rule is mostly (if not completely) a matter of dialect, style and sociolect.

Do note how paradigm cell #324 is a really abstract thing, that does not itself actually have either a meaning or a pronunciation. All that it has is rules that say how it may be pronounced, and rules that say in what larger phrases it may occur (and it is only these larger phrases that, in turn, have rules that tie them to meanings). And there's more than one rule in either "side" of paradigm form #324; there's more than one way to pronounce it, and there's more than one function that it's called to fulfill. This is, of course, an instance of one of the two biggest insights that you can come to acquire as a grammatical theorist: The relation between sound and meaning in language is many-to-many, and very indirect.
(The other big insight is that grammars are organized as systems of rules partially ordered in terms of specificity, and that when both a specific rule and a general rule apply, the specific rule trumps the general one. Hell, the ancient Sanskrit grammarians clearly understood this very, very clearly, 2,500 years ago.)
* For the uninitiated, here's a perfect example of Arnold Zwicky pulling a himself.
PS note that I actually did not manage to get around to stating the rule(s) whereby given a verb like show, we can form a noun showing...