Sometimes I worry that I’m not properly using this blog as a chance to get the word out about linguistics, so to rest my troubled mind, let’s talk a little about a component of syntactic theory: case. (Please stop clicking on the nearest link in an attempt to escape.) If you’ve never heard of case, here’s a quick overview. Syntactic theory dictates that all noun phrases must be assigned a case in order to be grammatical. This requirement is called the Case Filter. The Case Filter explains why (1a) is grammatical but (1b) is not; an inflected verb like ate can supply case to its subject (the man), but the uninflected infinitive to eat can’t supply case to its subject.  As a result, the man is not assigned case in (1b), breaking the Case Filter and resulting in an ill-formed sentence.  We can remedy this situation, as in (1c), where told supplies case for the man, satisfying the Case Filter and fixing the sentence.

(1a) The man ate seven sour cream donuts.
(1b) *The man to eat seven sour cream donuts.
(1c) I told the man to eat seven sour cream donuts.

In an attempt to build suspense, I’ve so far avoided saying what exactly this “case” thing is that’s being assigned. Roughly, case is a marker on a noun phrase (NP) that indicates what the NP’s role is in the sentence. English only has three morphological cases: the nominative (or subjective), accusative (or objective), and genitive (or possessive) cases. I’m going to overlook possessive case in this post, because it’s not relevant to the final point and is rather different from the other two cases. Nominative case is used to mark a subject, while accusative marks an object. These two cases are only apparent in pronouns. The nominative forms of the personal pronouns are I, you, he, she, it, we, they, while the accusative forms are me, you, him, her, it, us, them. (2a) is correct because the subject is in the nominative case and the object is in the accusative. (2b) is incorrect because the cases are swapped.

(2a) I saw him.
(2b) *Me saw he.

(2c) The octopus ate the cuttlefish.
(2d) The cuttlefish ate the octopus.

Note that (2c) & (2d) are both correct because non-pronominal NPs are “zero-marked” in English. That means that they don’t exhibit any outward marking of their cases; a nominative non-pronominal NP looks the same as the accusative non-pronominal NP. Zero-marking appears in some other places in English: the plural of sheep is sheep, and the conjugated verb eat in I eat fish is indistinguishable from the uninflected infinitival form in I want to eat fish.

Because only the pronouns have apparent case markings, English is often said to have an “impoverished” case system. Compared to a Finno-Ugric language like Estonian or Hungarian, which has tons of cases with exotic names like the inessive, superessive, ablative, translative, and exessive, English seems as poor as a pauper on payday. And what’s worse is that English has been steadily reducing its case markings. Back in Old English, not only were all nouns marked for nominative and accusative cases, but also dative and instrumental cases.

Why has English shed so much of its case system? Well, quite simply, it got outsourced. Prepositional phrases took over the roles of morphological case marking for most of the oblique cases, like dative and instrumental. That’s why we now say that Vikings lived by the sword instead of lifde sweorde — the instrumental changed from an -e suffix on the noun to the prepositions with and by. The English equivalents of those exotic Finno-Ugric cases are mostly recreated through humble prepositions. As for the structural cases (nominative and accusative), the modern (relatively) fixed word order has rendered them obsolete. Outside of poetic writing and certain syntactic alternations like topicalization, the word order of Modern English is Subject-Verb-Object. As a result, the structural cases are redundant and their markings have just fallen out of the language.

But not entirely! Pronouns have zealously retained their case markings through hell and high water. Except, of course, for whom, which has been losing its grip on that accusative -m for some time. And this, at long last, brings us to the whole point of this post.  Reading through a column about hypercorrection by Paul Mulshine, I was struck by one supposed example of hypercorrection, the use of whomever for whoever:

(3) Whomever the Republicans nominate should assume he must replace Iowa’s seven electoral votes …

(3) comes from the pen of George Will, who Mulshine claims has engaged in a spot of hypercorrection; according to Mulshine, whomever ought to be whoever. Mulshine asserts that “‘Whomever’ may sound more impressive to the unlettered, but it cannot serve as the subject of a sentence.” But whomever isn’t the subject of (3). The subject is the whole phrase whomever the Republicans nominate. And this is where case assignment gets really tricky, because we start to get a conflict.

Obviously, subjects get assigned nominative case; that’s why I love you is sweet and Me love you is stupid. But the subject in this sentence is actually a whole clause — a “sentential subject”, as it’s known in the biz. Semantically, who(m)ever is the head of the sentential subject, so you might well expect the nominative case to manifest itself on who(m)ever, yielding an m-less whoever. I imagine this is Mulshine’s train of thought here.  But within the sentential subject, who(m)ever is the object!  It has moved from the object position to the front of the clause, but if the embedded question were answered, we’d have “The Republicans nominate McCain”, with McCain, the object, replacing who(m)ever.*   Therefore, who(m)ever ought to get accusative case, yielding whomever. And so it seems that who(m)ever in (3) needs a Schrodinger’s m, an m that simultaneously exists to satisfy the accusative case and does not exist to satisfy the nominative case.

This dilemma ends up being resolved by arguing that the nominative case on the sentential subject doesn’t ever have to visibly manifest itself on an NP; it is an abstract case assigned to the whole sentential subject. That means that the nominative case never gets assigned to who(m)ever, and the problem clears right up, with George Will being technically correct to write whomever. But note that it took substantial analysis to realize that — certainly more than us case-deprived English speakers are prepared to do in fluent speech. Furthermore, note that this is such a weird situation that even native English speakers such as Mulshine will make mistakes on it. But more than anything else, note that it doesn’t matter. There is no ambiguity in the meaning of the sentence, no effect of omitting or including the m. That’s why case is falling out of the language; it just doesn’t do anything for us, and it can get really difficult to apply accurately. In my own speech and writing, I have an alternation between who and whom for the accusative case. I use whom in situations where it pleases my ears and I am confident in the accusative case assignment, and everywhere else I go with who. There are probably a lot of children who are being taught that who is the standard accusative form. I say good for them, and good for the language. It’s moving on.

*: If you’re having trouble with this part of the argument, think of the sentential subject as a question on its own: Whom did the Republicans nominate? Clearly in this question the Republicans is the subject and Whom is the object.