I think this really gets to the heart of what I’m calling into question, Jamie.

Perhaps it’s easiest to put it this way: when you say that “nobody ever gives any evidence that singular ‘they’ is ungrammatical,” what sort of thing would you even count as evidence?

You’ve already made clear, if I understand you correctly, that you would not count any number of classic books of grammar as evidence, on the grounds that, on your view, the question of whether to use ‘they’ for a singular antecedent is a matter of style, not grammar, and that anyone who writes against the practice is merely a stylist illegitimately addressing a grammatical question.

On the other hand, you imply here (in your snow in New Jersey case) that the fact that people speak in a certain way is very strong inductive evidence that they are doing so correctly. So perhaps your view is that the norms of grammar are fixed entirely by whatever the common usage happens to be in the time and place of writing or utterance.

I think there are many reasons to doubt that view, and I plan to explore them elsewhere at greater length. But even leaving those aside, I think that even on this strictly relativistic view, the best that could be said for singular ‘they’ is that some famous writers of yesteryear sometimes used ‘they’ to refer back to ‘anyone’, ‘someone’, or ‘everyone’, but that the usage of singular ‘they’ in a case like “My professor tried their best” is practically never found until recently.

Therefore, if the brute statistical facts of usage are the only marks of grammatical correctness, singular ‘they’ in the sense in which many people advocate using it today is not only historically wrong but quite possibly also wrong today (since I suspect that most English speakers would find it to be wrong).

If all you want to say is that singular ‘they’ is on its way to becoming correct as more and more young people are told that it is, and if you are willing to accept a radical relativistic view of grammatical norms, then I can see what you are saying, even if I don’t agree with your presupposition. In that case, though, there certainly would be a great deal of evidence against the correctness of singular ‘they’ in the professor example: I would estimate that the majority of English-speakers of all ages would flag that usage as incorrect.

Also, if the model of grammar we need to presuppose is the radically relativistic one according to which grammatical correctness is merely a matter of currently popular usage, then why invoke historical usage at all?

Hi, Jamie.

The Cambridge Grammar of the English Language describes itself, in the first sentence of the blurb, as “The Cambridge Grammar of the English Language is the first comprehensive descriptive grammar of English to appear for over fifteen years.”

That word ‘descriptive’ already gives the game away. It would be akin to a new work on ethics describing itself as “the first comprehensive descriptive ethics book” over a certain period of time. That already shows that the project it is engaged in is not the same as the one I’m interested in. Nor would it be the sort of thing people should turn to if they are interested in what the _correct_ thing to say is (unless they think it’s fairly straightforward to derive normative claims in grammar from merely descriptive claims).

The view I hold is that the descriptive changes in a natural language can be objectively better or worse, and that those who care about communication, etc. in a given language should therefore take an interest in the changes in the descriptive rules, and try to resist those that worsen the language.

It would be helpful to me if I could understand whether, and why, you disagree with that.

But what is it that, in your view, makes the rules of grammar what they are, constitutively? And what is it that, in your view, makes the rules of chess what they are, constitutively?

I’ve been thinking of this along the lines of chess rules also, and mentally composing the draft of my paper with that in mind. But I don’t think the conclusion to draw from it can be what you seem to think it is.

You seem to be saying that Huddleston and Pullum are able to speak authoritatively on the constitutive rules of English grammar because their research on the trends in _purely descriptive_ grammatical usage have led them to correct _descriptive_ conclusions, and (you seem to think) the normative facts of grammar follow straightforwardly from those descriptive facts.

In other words, as I read you — and please correct me if I’ve misunderstood — you are a radical relativist about grammar: that is, you believe that the normative rules about grammar can be derived immediately from the descriptive facts about the patterns of language use among the totality of speakers of the given language at that time, with no room whatsoever for any legitimate talk of right or wrong that cannot be reduced to talk of the frequency of a certain linguistic practice.

You seem to think that my own view is parallel to the view that Nigel Short’s stated conviction that a stalemate should be counted as a victory for the stalemating side is adequate for him to declare that he has beaten any player whom he has stalemated. But that is not remotely like what I am saying. In fact, that sounds like a kind of subjectivist view, in which any speaker’s opinion about grammar would automatically become authoritative, just because that speaker happens to have that believe about grammar. That radical grammatical subjectivism is perhaps even further from my view than radical grammatical relativism is.

What I am arguing for is not meant to be a form of radical relativism at all: in fact, that is just what I disagree with. My view is that some (limited) room must be left for the role of objective standards of better and worse in the rules of both grammar and chess.

Here’s a case from chess that works in parallel with my actual view: the rule of how many moves must pass under certain circumstances before one side may declare a draw. How was that rule decided on? Not by just looking at what the majority of players happened to be doing in their games — it is unlikely that there was any clear majority of players around the world who independently came to any views on the matter. And also not by making room for any random expert on chess to make his or her own declaration on the subject: that would lead to chaos, where what we need is clarity on what the rules are.

Instead, a number of experts look into the matter objectively and try to come up with a rule that harmonizes existing practice with the objective constraints on the rules, given the ends the rules are made to serve. It should be neither too easy nor too difficult for a player to claim a draw by this rule, and that stands as an objective constraint on the rule it would be reasonable to put in place.

If the committee put in charge of the matter by FIDE wisely comes to the conclusion that the rule should be that either side may claim a draw after both sides have made fifty moves without any pawn moves or captures, and this is confirmed after there has been room for fair discussion, then that decision ought to be held as authoritative by players everywhere — NOT because someone happened to declare that that’s the rule, but because everyone is better off with some sort of rule on the matter, and a good faith (and successful) effort has been made to come up with a rule that harmonizes existing practice with an objective consideration of what is reasonable and helpful, given the aims of the game of chess.

Now, let’s say that some YouTuber who knows almost nothing about chess manages to make chess a far more popular game than it currently is, and that this YouTuber ignorantly claims that any chess game that lasts for more than thirty moves is an automatic draw, regardless of what the players want it to be. As a result, let’s imagine, the number of chess players triples in 2025, and two thirds of them, having learned everything from the YouTuber, believe that all games lasting more than thirty moves are automatically drawn. The serious players of the world try to explain that this is not how the rules of chess work. But along come some researchers on chess from Cambridge, who make a quick survey of the beliefs of chess players in 2025. They discover that the majority of them do in fact hold that a chess game is a draw if neither player resigns or gets checkmated in the first thirty moves. On your view, does it follow from this that the FIDE rules are incorrect, just because of the influence of the ignorant YouTuber?

What if the ignorant YouTuber had falsely claimed that every chess game is a draw if nobody captures a piece within three moves? Would the sheer numbers of people who think this be sufficient, on your view, to change the rules of chess to make that the case, thus spoiling the game by making it quite silly?

The radically relativistic view could at least make room for the idea that the rules of chess before 2025, so I suppose the people who want to keep playing the game more seriously could say even then that they are playing the pre-2025 rules of the game. But now suppose that people start flooding the internet with evidence that many people — like some guy down at the bar in town — believed in a twenty-move rule for draws even back in the 1970s or perhaps even the 1670s. They rack up these stories and love to reel them off, one after the other, and then use them as a basis for jeering at anyone who says that there was a pre-2025 understanding that the game is only drawn after fifty moves with no pawn moves or captures. When defenders of the pre-2025 rules point to clearly-stated rules under FIDE, the US Chess Federation, etc. showing that the fifty-move rule was generally understood, the radical relativists scoff and say that that proves nothing, since FIDE and the USCF are just associations of self-important sticklers who failed to understand that the rules of chess can be derived immediately from the practices of the time, and that the practices were enormously flexible among, say, the sum total of players at different times (including all the kindergarteners, drunks, etc. who ever tried to play the game).

None of that would come close to establishing that the fifty-move rule should not be followed, or that it as a historic falsehood before 2025, I hope you will agree. What I’m saying is that something similar is going on with grammar. We cannot determine the rights or wrongs of grammar _merely_ by looking at what people happen to be saying and writing: we have to make some room for what is reasonable, or good or bad, given the aims of the language.

But in saying that we should work out good rules of grammar that do their best to coordinate both the ends of language and existing practice, I am certainly not saying that any random person, or even any single expert on language, should be taken as authoritative on the matter.

Compare: we are all better off if we have good traffic laws. Good traffic laws are to some extent arbitrary, but they must be guided by the needs of drivers and also the needs of pedestrians. An experienced driver’s opinion on the speed limit in school zones, say, should not be authoritative just because the driver expresses the opinion publicly. There ought to be a body of traffic laws that everyone in a locality accepts as binding, and a system with checks and balances should be put in place to determine those laws.

But if the body that tries to determine what the traffic laws ought to be simply does extensive research into what rules people already happen to be following, then they have simply failed to do what we need them to do. We are not looking for the rules people actually follow, but rather for the rules that people ought to follow. A descriptive grammar is something of interest, just as a descriptive summary of what chess players happen to be doing; but these things should not be mistaken for the normative guidelines that chess players, drivers, and speakers of English who aim for correctness are looking for.

I think that statements like “My professor tried their best” would be identified as incorrect by most English speakers alive today, and certainly by most speakers alive twenty or thirty years ago, so it seems that that sort of talk is incorrect both on the radically relativistic view and on the slightly objectivist view I hold. But even if that were not so, it seems to me that the radically relativistic view (as well as the radically subjectivist view of your Nigel Short case) are deeply mistaken.

Thanks, Jamie.

I think that something you (and nearly everyone else with mainstream beliefs today) take for granted as a fundamental fact is something that I just don’t accept. I reject the fundamental claim that Vox lusorum, vox dei in chess, and I think the same goes for grammar and usage. I can imagine many cases in which the majority of language-speakers speak and write ungrammatically in that language, and in which a language can be improved by the agreement to follow the conventions set out by a well-informed scholar, or body of scholars, who writes on that topic, even if most speakers of that language do not follow the rules set down for it. I take it that that sounds ridiculous to you and many others, but I think it is true nonetheless and I also think that I have good arguments for it.

I’ve been hesitant to write up the article I have in mind because I wasn’t sure I correctly understood the view that nearly all philosophers today seem to take for granted. This discussion has been very helpful in confirming that. Thank you for clarifying what seems so obvious to you. This is useful in getting me started.

Take care!

Well, you can of course keep *asserting* that “it’s a confusion to think” that any work to clarify and standardize the rules of a language should be taken as authoritative. But this is merely a constant restatement of the very claim that I dispute — that is, the claim that grammatical rules can be inferred directly from the practices of the majority of speakers of that language (give or take the small wrinkle you mention at the end of this comment).

If you have some _argument_ to present for this view, I’d certainly be interested to hear what it is. But if you think it’s just too obvious a thing to argue for, I’m quite happy to leave the discussion where it stands.

(By the way, I have a shorter example of the sort of thing you mention at the end, from my old professor Charles Morgan: “No head injury is too small to ignore.”)

That’s not quite it, Jamie, but it’s moving in the right direction.

What I think is that, in law, chess, and grammar, what fixes the norms cannot be reduced to a mere observation of actual practice. At certain points, we all benefit from the ability to defer to some authority on the matter, who can set rules to co-ordinate things between us. That way, we can avoid difficulties in advance by following the rules; and if there is a dispute about the correct rule, we can refer to what the authority says on the matter.

However, the authority is constrained to a certain extent by current and past practice, and by the purposes of the institution in question (language, chess, driving on roads, etc.). The authority should also be receptive to reasoned criticism, but should not be overturned by the pressure of superior numbers of people who lack the background knowledge to come to decisions. (For instance, if some YouTuber persuades a large number of activists that raising the speed limit on residential streets to 120 miles per hour, and the activists trust the YouTuber on the basis of mere trust in the YouTuber’s authority, the speed limit should not be changed merely because of the number of people who would like it to be).

You ask, “So, to be clear, you think that when Henry Fowler, or William Strunk, published some claims about English grammar, they (a) were better than the actual grammar that people spoke and wrote, and (b) thereby became the rules of English grammar?”

I think the rules of grammar have changed a number of times in this sort of way. For instance, consider the adoption of apostrophes and quotation marks in recent centuries. I take those to be improvements in the precision and clarity of the language, and genuine changes in grammar. But it seems very unlikely to me that we could ever have reached the agreement we take for granted today on those matters if there had been no texts outlining their rules that students could all be taught from and others that older readers and writers could refer to as authoritative reference books. It is, I hold, the writing of these books, and our agreement to take them as authoritative, that has allowed the English language to become clearer and more precise.

There may have been all sorts of other ways of differentiating possessives from plurals, or from indicating when people are expressing their own thoughts and when they are simply quoting others. But if we just left it up to different people to try to popularize their different ways in their different localities and social groups, we English speakers would soon lose the ability to understand each other. And there are many people who are simply not careful in how they speak or think at all. So, yes, I do hold that many writers of dictionaries, style guides, books on grammar and usage, etc. have made improvements on how many people spoke before, and have played an indispensable role in improving grammar and usage.

Sometimes, these books clarified or reminded us of rules that already existed when they were written, but at other times they may establish rules for the first time, by standardizing things.

I also hold that it is possible for the mass of people to become more stupid and illiterate. For instance, if space aliens hit the planet with a ray that reduced everyone’s speech to simple grunts, making communication possible, but the next generation of people were able to speak properly and revived speaking and writing, it would be wrong for someone to complain that the dictionaries and grammar books people turn to are out of date on the grounds that the only word left in the English language is now “ugh.” People who point to the majority of people in English-speaking countries and said that what most of them are speaking must be English would be mistaken. What happened is that people lost the ability to speak properly for a generation. Radical relativism about language rules — a position I reject — leaves no room for the possibility that the majority of people are getting things wrong. I find that implausible as well as harmful.

I don’t think that vocabulary is a part of grammar. I also am not moving from “We would benefit from X existing” to “X exists.” What I am saying is that, since we are all better off agreeing to an organized set of rules, we ought to agree to follow a coherent set of such rules and treat them as helping to determine correct and incorrect speaking and writing. It would certainly be surprising if that entailed neoPlatonism, but again you provide no argument whatsoever: you just scoff at me for being so far wrong and misguided that there can be no point discussing the matter.

Once again, you rely on mere assertion and re-assertion of your claim that there aren’t any authorities of the relevant kind for grammar. I know you think this, and I’ve explained why I think it’s incorrect. Why would you think that saying it yet again, after I’ve asked so many times whether you have any argument for this position, would be more persuasive?

I am still grateful for your participation in this exchange, since I reckon that you have probably thought more about this sort of question than the great majority of contemporary academics, all of whom seem to agree with you. But I have never heard any arguments for the position you all hold. I feel pretty confident that, if you had any arguments for the view, you would have provided them by now rather than continuing to just confidently state the view I take to be doubtful, over and over again.

This actually does tell me what I wanted to know, so thanks again.

Grymes writes:

“JD’s Nigel Short thing looks like an argument to me:
1) we ought to agree to follow a coherent set of grammatical rules (including that the singular “they” is ungrammatical, or whatever) and treat them as helping to determine correct and incorrect speaking and writing.
2) analogously, in chess, we ought to agree that stalemating is better than being stalemated, and treat that as helping to determine which (diachronic) configurations of pieces on the board count as a win.
3) but this normative truth (2) does not entail that stalemating is actually a win in chess games that aren’t operating according to Short’s house rules.
4) by analogy, normative truth (1) should not be taken to entail that the singular “they” is ungrammatical (or whatever) in language games that aren’t operating according to Kalef’s house rules.”

I’m not sure that this is Jamie’s argument, since it seemed to me that he was relying on there being a significant difference between chess rules and grammatical rules. But whether or not this is what he means, I’m grateful to you for providing a clear argument to respond to.

It seems that the generalized conclusion of this argument is this: the mere fact that it would be better if some normative rule N were accepted as the correct rule is not sufficient for making N the correct rule.

I certainly agree with that conclusion, but — if it is indeed Jamie’s argument — it does not get to the heart of the issue at all. For again, I do not hold that just anyone’s assertion that some rule would be an improvement is by itself sufficient to make that rule a normatively binding rule. I do not even think that this is true in cases where the rule would be a clear and objective improvement over the existing rule. There is often good reason to make an improvement of some kind to the existing set of rules to achieve a certain end, and yet there could be hundreds or thousands of rules that would do the trick. It certainly does not follow from this that all those possible rules are normatively binding rules!

What I think does follow is that it would be good, in such cases, for one such rule to be accepted as normatively binding. But for this to work, there has to be widespread agreement on the new rule. For the reasons I already mentioned in this discussion, this means for one thing that players of the game, or speakers of the language, should work toward agreements on a set of rules that makes chess, or communication in English, clear, precise, efficient, etc. Where there is an overall need for some sort of change in the rules, they should help along such a change. But where a change is being made in practice that worsens the language overall, due to sloppiness or bad ideas or what have you, they should try to resist the change in the rules while it is still possible to do so.

The first — but certainly not the only! — question to ask in the Nigel Short case is whether the new rule he prefers (stalemate should be counted as a win for the stalemater) is an improvement. My own view is not, and that such a rule would make chess worse rather than better. There were discussions on the topic when Short made his proposal, and as I recall most people argued that there are good reasons for not changing the rule. But suppose instead that the game would in fact have been improved, overall, by the adoption of Short’s rule, and moreover that this could be shown clearly. In that case, chess players would have had a good reason to help promote a change in the chess rules. I don’t see any implicit appeal to Neoplatonism there.

In the case of ‘they’ being used in sentences like “My professor tried their best to learn our names” where (let’s say to make it simpler) there’s no reason to think that the professor wishes to be referred to by gender-nonbinary pronouns, I hold that there already is a rule that such usage of ‘they’ is incorrect, and I also hold that it would make the language worse overall if the rules were changed so that ‘they’ became acceptable (let alone preferred) as correct grammar. I have a number of reasons for thinking that. I therefore hold that there are good reasons for resisting the change of the rule toward allowing or requiring singular ‘they’ in that sort of context.

Unfortunately (as I see it), many people and bodies with the power to influence the way people speak are caving in to pressure to promote ‘they’ as grammatical in the sort of context I mentioned. And, since what determines the normatively binding rules of grammar is in part (but only in part) the practices of the speakers and writers in the language, and the way it is taught and the way the reference books are written, these people are at risk of worsening the language, as I see it. I therefore hold that I and others have good reason for resisting those changes before it is too late to preserve the norm against using ‘they’ in that kind of case, and many others.

One way in which the advocates of (to my mind bad) change on singular ‘they’ operate is, as I see it, to cherry pick instances throughout history of people using ‘they’ in the singular, and to use this as grounds for asserting that the taboo against singular ‘they’ is based on a historical fiction. But, I argue, these cherry-picked instances clearly do not represent a general practice among literate English speakers to use ‘they’ in the ways we are asked to permit it now. ‘The professor tried their best to learn our names’ would have been flagged as ungrammatical by a significant majority of literate speakers, writers, and teachers of English for hundreds of years. This fact is obscured by the dishonest habit of pointing to a well-worn set of carefully selected instances of Jane Austen, etc. using ‘they’ to refer back to ‘anyone’, ‘everyone’, etc., and a smattering of other singular ‘they’ cases that I hold to be extremely rare, and trying to pass this off as standard practice.

The other move that is being made at the same time, and this makes the matter more difficult to deal with, is the insistence that grammar and usage are either purely descriptive or else normative only in a sense that allows one to move immediately from a statement about majority practice to a statement about binding norms. Since ‘My professor tried their best…’ has never (I hold) been accepted as correct English by the majority of literate English speakers, this move actually tells against the permissibility of ‘they’ in that context for now. But the popularity of that view among the intelligentsia makes it harder to get anyone to speak up any longer about resisting the degradation of language, or even to say in public that language can be worsened. With the opposition hamstrung in this way, it is much easier to spread the confusing and multiply-ambiguous ‘they’ as standard practice with very little resistance. If this sort of pressure persists for long enough in the places it is being exerted, the practices of English speakers will eventually be worsened along this and many other lines, and that would be a bad thing. Therefore, it seems to be a good thing to do what one can to resist the pressure.

None of that is parallel with the Nigel Short case. Nigel Short knew that the rule he wanted to introduce would have been a new rule, and there were many good reasons for not adopting the rule he suggested. What I am advocating is not a new rule, but the preservation of an accepted rule. If Nigel Short were to try a sneaky move of pretending that stalemate has always been accepted as a win throughout history, and that it is only a few sticklers at the USDF and FIDE who managed to pull the wool over everyone’s eyes and make a stalemate count as a draw, then he would be doing what I see people as doing when they present “The professor tried their best” as the sort of sentence most English speakers over the past few centuries would accept as perfectly correct. I am far from endorsing that kind of move.

Thanks for your reply, grymes.

I am saying that the great majority of literate speakers of English have, for centuries, seen sentences like “My professor tried their best to learn our names” as grammatically incorrect. One hears such things more often now, but I think that the great majority of English speakers today would still see it as wrong.

Moreover, I hold that for nearly all that time, singular ‘they’ (especially in such cases) has been taught as wrong in the reference books on correct English writing, and has been taught as wrong in classrooms and school textbooks. ‘They’ has been understood as a third person plural pronoun, while the third person singular pronouns are he, she, and it. This is how the language has been taught for centuries.

I know that many people feel that we should take a more permissive approach toward the use of ‘they’ as a non-gender-specific singular pronoun in all cases, and that some even feel that we should actively push for more people to change the language. I disagree: I feel that, on balance, there are stronger reasons for continuing to see singular ‘they’, in most if not all contexts, as incorrect. But I can see their reasons for arguing for the change.

But to say that singular ‘they’, especially in contexts like ‘My professor tried their best’, has been standard practice throughout the last few centuries of English is simply empirically false.

I’ll put money on this. We’ll have someone pick, at random, a hundred books or letters or articles written between 1700 and 2000, say, and find the first discussion of a specific person in each one (‘Mr. Francis’ or ‘my friend’ or ‘my cousin’ or what have you). Find the first instance of a personal pronoun used to describe each of these people. If most of the pronouns are ‘they’, ‘them’, or ‘their’, I lose the bet. If 99 or 100 use ‘he’, ‘him’, his’, ‘she’, ‘her,’ or ‘hers’, then I win.

Any takers?