You are currently browsing the category archive for the ‘disagreements’ category.

If you haven’t already heard, Texas’s state senate was engaged in a political thriller last night. I was able to catch the last 40 or so minutes of its livestream, riding that delightful roller-coaster of modern politics that trundles from dim optimism that something good will happen to the crushing realization that the deck has been stacked against it so thoroughly that you can only hope to minimize your losses.

Well, thankfully, there was a substantial group of state senators who managed to lack such pessimism and who stood up to do right by their constituents. Literally, in the case of Sen. Wendy Davis, who began a filibuster of an anti-abortion bill 13 hours before the special session was set to end at midnight. By rule, she had to keep talking throughout, stopping only to take questions. She was not allowed to sit, lean against a desk, leave the floor, or eat. None of this “silent” filibuster garbage of the federal senate: no, this is “Mr. Smith Goes to Washington”, wreck-your-body-because-you-believe-in-what’s-right dedication.

What I hope becomes an iconic picture of this event.

Sen. Tommy Williams couldn’t have looked more like a political caricature while arguing for a second strike against Davis’s filibuster because another member helped her adjust her back brace. [Photo: Patrick Michels, Texas Observer]

Davis made it ten hours before getting three strikes on her filibuster. Other senators took over, stalling until 11:45, when the Republican majority managed to trigger a vote to trigger a vote on the contemptible SB 5. With this first vote passed and the bill’s passage imminent, Sen. Leticia Van de Putte challenged the first vote on the basis of her having an outstanding inquiry at the time of the vote, a challenge the Senate president overruled by noting that he had not recognized her inquiry.

With calm fury reminiscent of Joe Welch, she asked, “At what point must a female senator raise her hand or her voice to be recognized over the male colleagues in the room?” The gallery erupted in cheers that continued for the last fifteen minutes of the day, a cacophony that prevented the final roll-call vote from happening until after the midnight deadline. It was powerful stuff, and my eyes welled more than once throughout it.

But this is a grammar blog, and I wouldn’t bring this up unless there was a grammatical discussion to be had. Let’s go back to the second of the three strikes against Davis’s filibuster, the objection I’ve included Patrick Michels’s excellent picture of above. Claire Cardona wrote at the Dallas Morning News’s filibuster liveblog:

“Sen. Tommy Williams, R-The Woodlands, raised a point of order on the filibuster because Davis had help from Sen. Ellis to readjust her back brace. [...] but Sen. Judith Zaffirini, D-Laredo, brings up a part in the rules that would permit Davis to sit. [...] Zaffirini notes that the rules said ‘may not lean on his desk, his chair, and that note doesnt apply to Sen. Davis.'”

The rule in question, which I think is Senate Rule 4.01 [PDF, p. 8-9], reads:

“When a member has been recognized and is speaking on a motion to re-refer a bill, he must stand upright at his desk and may not lean thereon (61 S.J. Reg. 1760, 1762 (1969)).
When a member has the floor and is speaking on a bill or resolution, he must stand upright at his desk and may not lean or sit on his desk or chair (61 S.J. Reg. 1059 (1969)).”

Of course, we all know what is meant there, that the intended interpretation is gender-neutral he. We see such usages regularly, probably think little of it, and move on. To claim that the use of he in this rule means it doesn’t apply to a woman is crazy, surely.

Or is it? During a filibuster, pedantry is everywhere. I learned but one thing from the time I volunteered at a Model Congress convention: intent and common usage mean little as the parliamentarian sifts through Robert’s Rules of Order to decide which of the fourteen simultaneous objections takes precedence over the rest. Grammatical pedantry, so often out of place, is only fair in this situation.

[Rt. Hon. W.L. Mackenzie King and guests unveiling a plaque commemorating the five Alberta women whose efforts resulted in the Persons Case]

The Canadian “Persons Case”, in which women were suing to be granted the right to be appointed to the Senate, depended in part on whether he included women. The legal gender-neutrality of he had in fact been specified in 1889’s Interpretation Act.

I’m no legal expert, and I’m certainly not an expert at what terminological choices are taken as given in the Texas Senate rulebook, so take this discussion with a grain of salt. But Zaffirini’s argument isn’t without merit. I found no declaration at any point in the Senate Rules that he is to be assumed as gender-neutral. In fact, there are eight instances of “his or her” in the rules — one occurring in the first sentence of Rule 4.01, the very rule being debated. If he is understood to be gender-neutral in the third and fourth sentences of Rule 4.01, why is he insufficient in the first?*

I wrote last year about how such uncertainty of the interpretation of gender-neutrality on the male pronoun in governmental settings was not a new concern: even the first U.S. Congress, in 1790, used “he or she” in some of their bills. It’s all hair-splitting, of course, but it’s a hair that may need to be split. When the issue at hand is so entwined with gender, maybe it’s a good time to examine our assumptions, starting with a little pronoun.


*: Possibly because the third and fourth sentences are based on 1969 formulations and the first sentence isn’t. But are Senate rules to be interpreted diachronically or synchronically?

This blog was linked to a while ago in a Reddit discussion of uninterested and disinterested. (My opinion on them is that uninterested is restricted to the “unconcerned” meaning, while disinterested can mean either “unconcerned” or “impartial”, and that’s an opinion based on both historical and modern usage. In fact, despite the dire cries that people are causing the two words to smear together, it actually looks like the distinction between them is growing over time.)

The reason I bring this up again is that one of the Redditors was proposing that having a strong distinction could make sense, because:

“Some people draw a distinction between disinterested and uninterested. There is nothing to lose and perhaps subtlety to be gained by using that distinction yourself. Therefore observing the distinction should always be recommended.”

But I’ve already asked my question about this in the title: is there really nothing to lose? Is there no cost to maintaining a strict distinction between words? Or, more generally, is there no cost to maintaining a grammar rule?

Well, in a myopic sense, no, there’s nothing much to lose by having the rule. In the case of uninterested and disinterested, it would be hard to argue that not being able to use disinterested to mean “unconcerned” is a substantial loss. It can be done, though: I, for instance, am a great lover of alliteration, and as a result, I like to have synonyms with as many different initial letters as possible. There’s a cost, small though it may be, to not having disinterested available as I’m constructing sentences. But that’s a triviality.

A more substantial consequence is that it introduces a discontinuity in the historical record. If we decide that from now on disinterested only means “impartial”, then historical and current uses of the “unconcerned” sense will be opaque to people taught the hard-and-fast rule. That’s problematic because, despite the belief of some people that this is an illiterate usage, it’s actually common even for good writers to use. This, again, isn’t a big problem; we regularly understand misused words, especially ones whose intended meanings are very close to their actual meanings. Saying that we can’t have a rule of grammar because sometimes it isn’t followed is the sort of whateverism that people accuse descriptivists of, not a reasonable concern.*

No, the true cost is a higher-level cost: the overhead of having another distinction. This might also seem trivial. After all, we have tons and tons of usage rules and distinctions, and a lexical distinction like this is really little more than remembering a definition. But let me illustrate my point with an example I recently saw on Tumblr (sorry for the illegibility):

The distinction here is well-established: affect is almost always the verb, effect almost always the noun.** Yet here we see that it is costly to maintain the distinction. First, it’s costly to remember which homophone goes in which role. Second, it’s costly to make an error, as people may mock you for it. Third, it’s very easy to get it wrong, as the replier did here.

If there were really no downside to adding an additional rule, we’d expect to see every possibly useful distinction be made. We’d expect, for instance, to have a clear singular/plural second-person distinction in English (instead of just you). I’d expect to see an inclusive/exclusive first-person plural distinction as well, as I sometimes want to establish whether I’m saying we to include the person I’m speaking to or not. The subjunctive wouldn’t be disappearing, nor would whom.

But all distinctions are not made. In fact, relatively few of the possible distinctions we could make at the word level are made. And that suggests that even if the reasons I’ve listed for not maintaining a lot of distinctions aren’t valid, there must be something that keeps us from making all the distinctions we could make.

So next time someone says “there oughta be a rule”, think about why there isn’t. Rules aren’t free, and only the ones whose benefits outweigh their costs are going to be created and maintained. The costs and benefits change over time, and that’s part of why languages are forever changing.


*: Of course, if the distinction is regularly violated, then it’s hardly whateverist to say that it doesn’t exist.

**: Affect is a noun in psychology, effect a verb meaning “to cause” that is largely reviled by prescriptivists.

Let me talk about something that I feel like I’ve been circling around for some time, but never quite directly addressed. It’s a common thing in grammar grousers: playing up other people’s questionable usages as symptomatic of a larger disease while playing down one’s own as a clever subversion of stodgy English. Whereas the complainant’s usages are all justified by improving the language or enlivening the prose or just plain sounding right, the scorned writer’s usages are utterly unjustified — not because the complainant has considered possible justifications and found none of them sufficient, but rather because it is simply self-evident that an error is an error.*

Thus we see Salon’s Mary Elizbeth Williams writing a screed against sentential hopefully, but then absolving herself for using stabby and rapey. I find both of those to be worse than the targets of her ire — especially rapey, the jokey tone of which I find borderline offensive. Crucially, though, even as I reject her words, I can see why she likes them; it’s just that for me, their benefits don’t outweigh their downsides. Williams, on the other hand, seems to ignore any potential upsides to the usages she dislikes. When she says rapey, she sees it as the considered usage of a professional writer, an improvement on the language. When you write sentential hopefully, it’s because you can’t be bothered to think about your usage and the effects it could have on the language.

Similarly, I got into a short Twitter war with a follower who tweeted that she wanted to send copies of education majors’ grammatical errors to future employers. I pointed out that the follower (whose Twitter name is “Grammar Nazi”, about which the less said the better) had questionable usages in her bio:

“A soon to graduate English major whose biggest turn on is good grammar.”

In my grammar, there’re three missing hyphens, but she responded to me noting this with “I’m sure you’re aware compounding is a grey area. Rules may be generally agreed upon, but no official guidelines exist.” Such “generally agreed-upon” rules were probably settled enough for the tweeter to treat as errors had others broken them, but because she’s doing it, it’s okay. Her choice to go against the standards is justified, because she sees the justification. The education majors’, with their justifications left implicit, probably wouldn’t be.**

This forgiveness extends, of course, to include other people whose viewpoint the writer is sympathetic to. Kyle Wiens, who wrote that Harvard Business Review piece on his intolerance for grammar errors in his hiring practices, had a couple of questionable usages in the piece — nothing too bad, but things that would violate a true Zero Tolerance stance. Another blogger quoted some of the piece and added:

“Ignoring the one or two grammatical glitches within the quoted text (they may be the result of a message that was delivered orally, rather than in written form), the message [...] should be taken to heart. If you write poorly, you tell your reader: I haven’t changed. My education hasn’t made me better, it hasn’t touched my core. [...] I’m certainly not looking to have excellence be part of my personal brand – it’s too hard and too time consuming.”

The blogger seeks out an explanation for Wiens’s errors that diminishes the errors, but then chooses an explanation for everyone else’s that diminishes the writers.

We all do this to some extent. The most prominent example for me is when I come home from work and find a pile of dishes in the sink from my roommates. “C’mon guys, you can’t be bothered to do the dishes?” I wonder to myself and to anyone I talk to over the next few days. Yet I’ve just realized that I forgot to finish the dishes this morning before going to campus. Somehow I can’t muster the same indignation at myself as I have toward my roommates, because I had an excuse. (And I’ll tell you it as soon as I figure it out.)

Sure, it’s fair to give known-good writers more leeway than known-bad ones. But every error has a cause, and every usage a rationale. Don’t decide ahead of time that someone can’t be wrong or can’t be right.

*: This isn’t unique to grammar by any means; half of politics is explaining away your side’s missteps while playing up the other side’s.

*: By the way, you may wonder if I’m not doing exactly what I oppose here by complaining about a minor error that some people do not see as an error. On that, two points. One, hyphenating phrases that are used as adjectives (especially more-than-two-word phrases) is about as standard a rule of punctuation as one can find. Similarly with hyphenating a phrasal verb in its nominal form. Two, not that she needs to justify herself to me, but she doesn’t explain any reason why she’s breaking the rule, so as far as I can tell, she’s breaking the rule just to break it — hardly appropriate behavior for an otherwise hard-liner.

All right, it’s time for the second grammar review section; last week’s looked at contractions and their homophones, and today I’ll look at who and whom.

The simplest advice I can give about using whom is not to. Contemporary English doesn’t require whom in any situation other than exceedingly formal writing. Just use who all the time.

Before you think that I’m just some lazy linguistic anarchist for suggesting this, let me point out that I am only agreeing with John McIntyre, former president of the American Copy Editors Society and an editor at the Baltimore Sun, who writes:

“There is a problem that even educated writers have with figuring out whether a subordinate clause should begin with who or whom. If you have that difficulty, you can, except in the most formal circumstances, just use who.”

But perhaps you have a reason to use whom, whether because you need to write very formally, or because you have a stodgy teacher/superior who insists upon its use, or because you’re just good old-fashioned curious about the niceties of English grammar. In that case, here’s my advice on how it’s used.

In short, who and whom are the same word with different case markings. Who is in the nominative (or subjective) case, and whom is in the accusative (or objective) case. That’s the only difference — not that one is more formal than the other or anything like that.*

So knowing how to use whom is simply a matter of knowing when each case should be marked. Unfortunately, English rarely marks case, so it’s not something that we, as native English speakers, are used to thinking about. In fact, aside from who(m), the only other sort of case marking in English is on personal pronouns, and even then only on a few of them.** I and me, for instance, are nominative and accusative versions of each other, as are he and him, she and her, we and us, and they and them.

The first guideline, then, is to use whom wherever it replaces an accusative pronoun (me, him, her, us, them). So:

(1a) Who saw you? (She saw me.)
(1b) Whom did you see? (I saw her.)
(1c) Whom did you give the gift to? (I gave it to her.)

Your intuitions with personal pronouns’ cases are probably pretty accurate, so when you can rephrase the sentence, you’ll do well. The trouble is that you can’t always easily replace who(m) with a personal pronoun. For instance:

(2a) The fellow who(m) I saw at the bus stop
(2b) I don’t care who(m) did it.
(2c) Who(m)ever the werewolf stalks is in trouble.

Since there’s no question to answer here, you need to get a bit cleverer and look at the syntactic structure of the sentence. In these examples, who(m) is filling for a missing noun phrase somewhere in the sentence; linguists refer to the missing noun phrase as a “gap”, and who(m) as its “filler”. Even though the filler is usually in a different position from the gap, structurally the filler and gap are linked. Whatever case would be assigned to the gap manifests itself on the filler.

When the gap is an object, whom is appropriate. (2a) can use whom, because it’s filling a gap in the object of the verb saw.

When the gap is a subject, whom is inappropriate. (2b) can’t use whom, because the gap is the subject of the verb did.

(2c) gets tricky, because we seem to have two conflicting case assignments. Who(m)ever is the object of stalks, so you’d expect accusative case, but it also looks like the subject of is, so what do we do? In general, only the closest case assignment matters, and case doesn’t trickle down within a phrase. Since the subject of is is actually the whole phrase who(m)ever the werewolf stalks, not just who(m)ever, its case assignment doesn’t manifest. Only the assignment within the smaller phrase who(m)ever the werewolf stalks matters, and that’s accusative case from stalks. Thus whomever is appropriate here.

This is a bit subtle, and I don’t think I’ve done a great job of explaining it. (A newspaper columnist and I got into a fight about such a case assignment three years ago, and we still haven’t settled it.) This is the sort of situation where you’re probably best to just go with who; even if you have spent the time to prove to yourself that whom is correct, there’s a pretty good chance that someone else will insist that you’re hypercorrecting.

Summary: In contemporary American English, whom is necessary only in certain situations within very formal writing, so you can get by just fine without using it. If you choose to use it, remember that it is not the formal variant of who but rather the accusative-case variant of who. If who is replacing a subject of a sentence, it should never be whom. Whom is reserved for objects of verbs and objects of prepositions.

The Back-to-School Reviews so far:
I: Confusing contractions (your, you’re and the lot) [09/04/12]
II: Who and whom [09/10/12]


*: Whom seems more formal because it’s mostly used in formal writing. In informal writing, who is the form for both nominative and accusative cases.

**: Technically speaking, the apostrophe-s on possessives is a way of marking genitive case, but that’s a topic for another time.

Post Categories

The Monthly Archives

About The Blog

A lot of people make claims about what "good English" is. Much of what they say is flim-flam, and this blog aims to set the record straight. Its goal is to explain the motivations behind the real grammar of English and to debunk ill-founded claims about what is grammatical and what isn't. Somehow, this was enough to garner a favorable mention in the Wall Street Journal.

About Me

I'm Gabe Doyle, a graduate student/doctoral candidate in Linguistics at UC San Diego. I have a Bachelor's in math from Princeton and a Master's in linguistics from UCSD.

In my research, I look at how humans manage one of their greatest learning achievements: the acquisition of language. I build computational models of how people can learn language with cognitively-general processes and as few presuppositions as possible.

I focus on learning problems that have traditionally been viewed as difficult, such as combining multiple information sources or learning without negative data or ungrammatical examples. My dissertation models how children can use multiple cues to segment words from child-directed speech, and how phonological constraints can be inferred based on what children do and don't hear adults say.



@MGrammar on twitter

Recent Tweets

If you like email and you like grammar, feel free to subscribe to Motivated Grammar by email. Enter your address below.

Join 749 other followers

Top Rated

Follow

Get every new post delivered to your Inbox.

Join 749 other followers

%d bloggers like this: