September 15, 2006

"Singular they" mailbag

My post "'Singular they': God said it, I believe it, that settles it" has brought in quite a bit of email. I've added updates to the original post, citing the notes that added specifically to the content. But there were some more general comments as well, which I've posted below.

The incoming mail is in red, my comments are in black, and quotes from others are in blue.

You might want to re-read the original post before venturing past the jump -- and if the topic of "singular they" bores you, please feel free to return to our usual run of comic strips and astronomical observations.

This note comes from David Meadows:

You're citing a source that includes such spelling variants as "bee", "vaine", "minde", "esteeme", "then" (for "than") and "themselues", and claiming it as an authoritative proof of a point of English grammar?

Was your post intended sarcastically or am I missing some deeper linguistic point?

(This is a serious question. I am not a linguist by training and I usually find your Language Log articles extremely enlightening, but I'm struggling to understand how the KJV can possibly be relevant to your argument.)

Well, it was more of a joke than an argument. (For foreign readers, or those who don't get out much, the context of the joke includes a saying often seen these days on "bumper stickers and other Christian products".)

The argument was settled long ago: singular they has routinely been used throughout the history of English, by all the best writers, until certain subcases were artificially turned into "errors" by self-appointed experts. Successively less discriminating pseudo-authorities then generalized the proscription in successively sillier ways, although they have largely been ignored by the users of the language.

It's true that English spelling wasn't codified until the 18th and 19th centuries. And it's true that there's an analogy between the process of regularizing spelling and the process of deciding to forbid "singular they" and split infinitives and so on. But it's a false analogy.

Spelling is an essentially artificial intellectual construct, requiring many arbitrary choices, both large and small. Consistency is intrinsically valuable (though this value is often over-estimated) because it helps readers to cope with their difficult and artificial task; and communities of writers don't naturally converge on consistent practices. In terms of the large choices, we could write English using a system derived from Chinese logograms, or Greek letters, or Arabic letters, or Sanskritic devanagari characters, or the International Phonetic Alphabet, or something completely different. It's important for all of us to make more or less the same basic choice. Even then, there are lots of possibilities for variation, as Tudor and Elizabethan spelling practices demonstrate. This variation causes fewer problems for readers than the difference between cuneiform and the Latin alphabet would, but it's still plausible that it's easier for readers (though harder for writers) to agree on a consistent way to spell each wordform.

Thus the codification of English spelling was basically a Good Thing, and one which required the intervention of a small number of self-appointed authorities. Alas, they botched the task, leaving us with a mess that causes billions of dollars a year in on-going damages. But the result is now OUR botch, and it's too hard to change it, so we need to learn it, and we might as well try to enjoy it.

In contrast, the choice among he, she, it and they is part of a natural order, developed spontaneously by millions of speakers, hearers, writers and readers. Languages with no writing system also develop systems of this kind -- and with or without writing, no experts are needed in order to create the set of shared assumptions that foster communication. Eugene Volokh, following Friedrich Hayek, made a distinction that is useful here:

Language defined by changing usage is what some call a "grown order" -- a judgment formed by millions of people, based on their senses of what is convenient and comfortable for them. (Free market economic decisions are another classic example of something that's mostly a grown order.) Linguistic prescriptivism (dictionarymakers recording what they think should be the usage, not what is the usage), is a "made order" -- a judgment of a small group of people selected for the purpose of rendering their judgment. Made orders are sometimes useful, for instance in the setting of technical standards. But as to language, I think the grown order approach is far more likely to yield a language that is genuinely responsive to users' needs than the made order approach.

(Note that Eugene understands that dictionary-makers actually do try to record usage rather than dictating it -- he's responding to a correspondent who argued that "without the 'Language Police' standing athwart language liberalization, every usage would slip into the dictionary". Eugene answered that "every common usage should be in the dictionary", and went on to explain why.)

Glen Whitman responded to Eugene by arguing that self-appointed language experts should also be free to peddle their wares:

Yes, grown orders (also known as spontaneous orders) are generally more responsive than made orders. But does any form of linguistic prescriptivism necessarily fall in the latter category? Unless the prescriptivists actually attempt to enforce their standards on the rest of us in the manner of the French Academy, the power of the prescriptivists to influence the language is dependent on the willingness of other speakers to follow their lead. In other words, the prescriptivists' admonitions have their place within the grown order of language.

In the marketplace of usage, the prohibition of "singular they" has not exactly prospered. Like many similar prescriptions, it ekes out a meager and marginal existence as a sort of niche product, appealing to ... Well, that's enough, I don't want to be rude.

The next note is from John Brewer:

I appreciated the posts on this, but I wonder if Prof. Liberman could offer any insight on the issue that Mr. Leman notes but then disposes of perhaps too quickly.  Unlike Prof. Liberman's prior Language Log example of the KJV rendering of Deut. 17:5, which seems like a "pure" singular they (in that presumably only one individual at a time is being stoned to death in the most common application of the provision), most of Mr. Leman's examples fit that category only if one accepts that "every one," "everyone," and "every man" are indeed semantically as well as syntactically singular in the relevant contexts.  It's not clear to me that this is necessarily the case.  Rather, they seem at least loosely analogous to those collective nouns like Congress, Parliament, Israel & Judah in the Scriptures (when referring to the nations rather then their namesake patriarchs), etc., since with "every one" etc. there is necessarily a plurality of individuals doing the same thing at the same time.  While right now the U.S. preference seems to be to treat Congress etc. as syntactically singular for purposes of verb agreement while the British are more willing to treat them as syntactically plural (Congress is considering legislation, but Parliament are considering legislation), it seems widely accepted in the U.S. to use they/them with such nouns as the antecedent, even in the same sentence in which a singular verb is used.  "Congress is going to raise our taxes again unless we tell them to knock it off."   "If Citibank calls back, tell them we need to stop payment on that check."  Note that the use of they/them in this latter context is not driven by contemporary feminism, egalitarianism, or (more cynically) fear of giving offense to the thin-skinned, since the potential alternative pronoun in modern usage would usually be "it" rather than generic he/him.

I agree with Mr. Brewer's assessment of current American practice. But the prescription against "singular their" has come to apply to all the various sub-cases -- simple indefinite nouns, nouns universally quantified with each or every, collective nouns like committee, and so on. (See "The SAT fails a grammar test" for an example where committee ... their was characterized as an error by the Educational Testing Service.)

Mr. Leman's examples with "each" also seem potentially susceptible to being understood as semantically plural.  As with the "every" examples, they could generally be paraphrased using a construction such as "all of the children of Israel" which would be syntactically plural.  It's much harder, for me at least, to get a semantically plural reading of "anyone" in the TNIV rendering of Rev. 3:20, so I'm not convinced that the KJV etc. precedents he cites directly justify that, even leaving aside the fact that the TNIV translators have rejected the KJV as an authoritative or even persuasive model in any number of other ways.  But anyone who is willing to accept the general TNIV approach to translating texts raising gender/sex issues but then gets hung up on that sort of "singular they" seems to me to be straining at a gnat after swallowing the camel.

I'm certainly not suggesting that we should follow the linguistic norms of the KJV -- among other things, they're 400 years out of date. My point was only to add to the list of examples showing that (variable but common) use of "singular they", in all its forms, has been part of standard English for a long time. (Well, that and to make a little joke about divine sanction, leading to a hypothetical bumper sticker or at least a coffee cup.)

Finally, there's an interesting KJV text I noticed after reading Prof. Liberman's earlier post on Deut 17:5 in which both singular and plural pronouns are used for the same antecedent in the same sentence.  It's Numbers 24:2:  "And Balaam lifted up his eyes, and he saw Israel abiding in his tents according to their tribes; and the spirit of God came upon him."  (I assume Israel isn't/aren't abiding in Balaam's tents!) 

Wow. Yes, it's clear that those aren't Balaam's tents:

[1] And when Balaam sawe that it pleased the Lord to blesse Israel, hee went not, as at other times to seeke for inchantments, but hee set his face toward the wildernesse.
[2] And Balaam lift vp his eyes, and he saw Israel abiding in his tents, according to their Tribes: and the Spirit of God came vpon him.
[3]And he tooke vp his parable, and said, Balaam the sonne of Beor hath said, and the man whose eyes are open hath said:
[4] Hee hath said, which heard the words of God, which saw the vision of the Almightie, falling into a trance, but hauing his eyes open:
[5] How goodly are thy tents, O Iacob, and thy Tabernacles, O Israel!
[6] As the valleyes are they spread forth, as gardens by the riuer side, as the trees of Lign-Aloes which the Lord hath planted, and as Cedar trees beside the waters.
[7] He shall powre the water out of his buckets, and his seed shall be in many waters, and his King shall be higher then Agag, and his Kingdome shall be exalted.

I wonder what the original pronouns are?

And last in chronological order, but by no means least in interest, is a note from David Russinoff:

Following a link from Sally Thomason's recent post on "singular 'they'", I was led to an earlier one by you ("All Lockers ..."), which in turn refers to two other articles on the subject, with the suggestion that they represent two sides of the issue. In fact, both articles embrace the usage in question, one more enthusiastically than the other. I suspect that you are aware that there actually is an opposing view but are convinced that no expression of it could be worthy of consideration. In case I'm wrong about this, I refer you to the entry on "their, etc." at <>.

The link takes us to a page on which Mr. Russinoff explains that

Occasionally ... in the spirit of Kingsley Amis's dictum, "The defence of the language is too large a matter to be left to the properly qualified", I offer my own observations. It is my hope that arrogance, pedantry, and dogmatism can compensate for what I lack in other credentials.

A link in the left margin then leads us to a spirited attack on "the use of a plural third person pronoun with an indefinite singular antecedent such as anyone, everybody, no one, a person, or each party".

I invite you to read it for yourself. To my mind, the best part is the end:

I cannot deny a certain admiration for [Churchyard's] scholarship, but ... I was reminded of Fowler's reaction to a similar argument posed by Otto Jespersen:

I confess to attaching more importance to my instinctive repugnance for [such nonsense] than to Professor Jesperson's demonstration that it has been said by more respectable authors than I had supposed.

Bravo, Mr. Russinoff! It's inspiring to see a man stand up for his convictions, even in the face of the Word of God. The "grown order" of language emerges from the competing and cooperating instincts of its users, embodied in their speech and writing, not from the pronouncements of religious or cultural authorities. So keep it up, and may the best repugnance win!

Posted by Mark Liberman at September 15, 2006 08:50 AM