I’m told Carol Fisher Saller of the Chicago Manual of Style, in her new book The Subversive Copy Editor, recounts how she convinced an author that that of him who seeks should be that of he who seeks.
Tsk, tsk, tsk. Ms. Saller! You’ve clearly been staring at this stuff too long! You’ve simultaneously overthought and underthought this one. Overthought because you’re letting your ideas override your ear; underthought because you haven’t properly analyzed what’s going on here.
The mistake – and Saller’s not the only one making it – is the belief that the him (or he) is the subject of who seeks: he…seeks, yes?
No. The subject for the verb seeks is the relative pronoun who. It’s a dependent clause, who seeks (subject plus verb), modifying the word him, which is in the accusative case because it’s not the subject of anything; it’s the complement of the prepositional phrase of him. (You can’t make the complement of a pronoun phrase simultaneously the grammatical subject of a verb. Sentences are nested constructions, as we will see.)
Probably the most common instance of this confusion is with a sentence such as (often exactly as) Let him who is without sin cast the first stone. It’s so common, in fact, that Google hits for “let he who is without sin” outnumber those for “let him who is without sin” by nearly two to one. But those are all hypercorrections based on misanalysis. This isn’t a question of a nonstandard dialect versus a standard one, because this phrase only exists now in more formal English. It’s the very unfamiliarity of it in everyday use that leads people to screw it up.
Let’s take it apart:
The main clause is Let him cast the first stone. Note that the verb here is let; it is in the imperative. In English we don’t have a proper third-person imperative, so we use a second-person of let plus the third person in the accusative case. Thus we translate Qu’ils mangent du gâteau as Let them eat cake. Not let they – the eat, like the cast, is an infinitive. (Note, by the way, that there is no to on the infinitive. The to is not an intrinsic part of an English infinitive, and so interposing a word between it and the infinitive is not really splitting the infinitive – and assertions that it’s wrong are actually baseless in English, an invention on an inappropriate Latin model. The construction is inelegant at times, yes, but not grammatically wrong. But that’s a digression.)
And the who is without sin? It’s a relative clause modifying the him. Relative clauses do not change the syntactic structure of the main clause in which they are embedded; they can always be lifted out and leave the main clause intact. And the reason they can do so is that they have their own subject and predicate. The subject, as I mentioned, is the relative pronoun who.
Now, if the sentence were Let him that he has spoken to… then the modifier of him would be that he has spoken to and the that would be a conjugation introducing the clause with he as subject. (The person spoken to in this case is implied; linguists, when drawing out a tree structure, will indicate a trace of him after the to. If you invert it, Let him to whom he has spoken, you put the to whom in front and you don’t need the that.)
As best I can tell, people get confused on this because of the is following the him. It’s similar to the error where a person sees, say, My futon is arriving, and with it come two pillows, and think it should be comes because of the it before. But it’s the two pillows that are the subject of the verb; uninvert the sentence and you get …and two pillows come with it.
Now, if you have He who is without sin can cast the first stone, is that an error? No; lift out the relative clause and you’ll see He can cast the first stone. In this case, the he is the subject of the verb, and the dependent clause modifies the subject rather than the object.
It’s always good to be in the habit of taking sentences apart to see their structure. Whenever analyzing a sentence, take apart the various bits that make it up and identify the syntactic subject and verb of each clause. That’s syntactic, not thematic. The syntactical roles in a sentence don’t always go with the expected thematic roles. In Let him cast the first stone, the actor is him, but syntactically him is the object, and the cast is an infinitive rather than the inflected finite verb it would be if we just said He casts the first stone.
Sesquiotica 
Hmm. Maybe you’re overthinking for both of us. (How else would you know my thoughts without asking?) So here are my actual thoughts: I don’t believe that “he” is the subject of “seeks.” But I do believe that the case of “he” must be determined by the verb of the relative clause. Because I’m admittedly no grammarian, in this situation I chose to follow the example of the eminent old Fowler ‘s Modern English Usage (s.v. “whoever”).
Kind regards,
Carol Saller
Ah, hello! Well, there’s the problem with sorting out why someone’s doing some particular language thing if you can’t ask them (and I thought it might be a bit importunate to email you out of the blue and ask): it’s a “black box” problem, like atom smashing — you can’t see the actual thing happening, so you have to interpret the bubble trail. But at least I got to cover the topic for everyone else.
I only have a copy of the New Fowler’s, and its entry on “whoever” goes nowhere near this topic, so — God bless Google Books — I’ve looked up the old one online; I assume that what I see at http://books.google.ca/books?id=Z4HI0RQIDK0C&dq=fowler%27s+modern+english+usage&printsec=frontcover&source=bl&ots=wbC3Xs1oHQ&sig=_G2cKsLtdAOTSLmkh8ZjLGI2YXw&hl=en&ei=HrGtSaDaE4nUNOPT4eEE&sa=X&oi=book_result&resnum=2&ct=result#PPA727,M1 (page 727) is a blotchier version of what you referred to. Do tell me if it’s not.
The relevant section covers this sentence: “For whoever was responsible for that deliberate lie there can be no forgiveness.” The person quoting it says that it should be “whomever”; Fowler maintains, and I agree with him, as apparently do you, that “whoever” is correct.
But that’s a different case from “that of him who seeks.” In Fowler’s sentence, which can be uninverted to “There can be no forgiveness for whoever was responsible for that deliberate lie,” there is no equivalent of the “him” in “that of him who seeks.”
As Fowler says, “_whoever_ is a relative that resembles _what_ in containing its antecedent in itself; as _what_ = that which, so _whoever_ = any person who; the _that_ & the _which_ of _what_ may or may not be in the same case, & similarly the _any person_ & the _who_ of _whoever_ are often in different cases; but the case of _whoever_ is that of the _who_, not that of the _any person_, that is, it is decided by the relative clause, not by the main sentence…”
The _him_ in “that of him who seeks” is like Fowler’s _any person_, the object of the main clause, while the _who_, like _whoever_, is the subject of the relative clause. When there’s a _whoever_ we don’t have an explicit pronoun object for the main clause; rather, we can say that the entire subordinate clause functions as the object. But when we have an explicit object that is not the grammatical subject of a subordinate clause, there’s no choice but to treat it as an object.
If I’m looking at the wrong bit of Fowler, though, do let me know!
Incidentally, if you go to page 231 of the Google Books copy of old Fowler, you will see a list of examples of the error of using _he_ where _him_ is required. It’s headed by “The bell will always be run by he who has the longest purse & the strongest arm.” It’s not exactly your sentence, but it’s close enough, I’d think, and it’s Fowler’s first example of the mistake. Pity he doesn’t take the space in that entry to explain why.
Yes, I see now exactly what you point out: the example I invoked as proof does not actually contain an equivalent of “him.” But this puts me in a terrible pickle! What am I going to do about the anecdote in the book? And what about that poor scholar who caved prematurely in the face of my confident declaration?
Do you think we can keep this just between the two of us?
I won’t tell him if you won’t.
Yes, “him” is the subject of the infinitive “cast,” which explains its case. Comparison to the French “qu’ils mangent,” where “mangent” has all the appearance of an inflected verb, obscures the point.
Pingback: I must disagree with whoever wrote that « Sesquiotica
It’s informative that ‘the New Fowler’s … entry on “whoever” goes nowhere near this topic, unlike the case in the ’eminent old Fowler’s Modern English Usage’. After reading the arguments for the necessity of this case or that in constructions where there does seem genuine ambiguity, I’m convinced that they’re driven largely by the felt need to defend cherished practices as being governed by logical and consistent rules, rather than the more valid need to question whether such rules should be replaced because they are inherently ILlogical and INconsistent. Adding more and more involved qualifications and sub-analyses does not appear to be the best way forward.
I’m not sure I feel happy with ‘…that of him who seeks…’ (which I admit is accepted as grammatically traditionally ‘correct’ – along with ‘This is he’), never mind ‘…that of he who seeks…’. ‘…that of the person who seeks’ sounds somewhat more acceptable, if a little starchy – of course, this neatly dodges the ‘should it be nominative or accusative’ issue.
The Free Dictionary translates ‘honi soit qui mal y pense’
(Government, Politics & Diplomacy): ‘shamed be he who thinks evil of it’ or ‘shame upon him who thinks evil upon it’ (the motto of the Order of the Garter).
WikiAnswers has: ‘It has two translations: the literal translation is “Shame be to him who thinks evil of it” but it has taken the meaning of “Evil comes to he who thinks it.”
Wikipedia contains the appalling ‘Evil to he Whom Evil Thinks’.
The semi-original Old French is preferable.
Orwell’s Sixth Law does however seem to drive modern usage – ‘Did you see who they were speaking to?’ is becoming an acceptable or even the required version in all registers, and there is no traditionally correct version of the accepted ‘It’s us!’
With regard to the issue of the original constructions, perhaps it’s better to say that there is an apparent dual role being demanded of him/he and whoever/whomever, without delving into nesting rules and the fascinating law ‘When there’s a _whoever_ we don’t have an explicit pronoun object for the main clause; rather, we can say that the entire subordinate clause functions as the object. But when we have an explicit object that is not the grammatical subject of a subordinate clause, there’s no choice but to treat it as an object’. If we are prepared to accept that there is in these examples a genuine demand made for a word to fill a simultaneous nominative/accusative role, rather than argue fine points about how to bend the laws to explain away such awkward and contentious (as you must agree!) occurrences, we could possibly eliminate such problems.
The obvious solutions are to use a noun where possible, or to rewrite using two sentences, say:
Let the man who is without sin cast the first stone.
Someone was responsible for that deliberate lie. There can be no forgiveness for that person.
Admittedly, some of the antique flavour of the originals is lost, but then, so is some of the antiquated flavour.
Edwin Ashworth
Well, linguistic analysis doesn’t really focus on supply and demand – it’s mainly about what people are actually currently doing. And when we look at ordinary everyday usage in sentences that are less convoluted than some of the ones that are used in more formal print registers, we can see that where there’s a whoever as the subject of a subordinate clause the entire subordinate clause functions as the object. (I can’t tell you, for instance, how many times I’ve heard, in a store or fast-food establishment, “Can I help whoever’s next?”) It’s only when people come to more involved sentences and a more formal register that they start doubting themselves; observing (fairly accurately) that whom is reserved for more formal use, but no longer having a natural feel for its use, they try to apply it according to what appears right to them, but their analysis is often faulty, because they are used to simply using syntax intuitively, and once they start looking at it, they allow their intuitive sense to be overruled by judgments that are often not well-formed. This is where those of us who have studied syntax more formally can apply our analyses and see where artificial rules are being imposed that are at odds with what is natural. My determinations are aimed making the uncommon more convoluted sentences consistent with the common less convoluted ones – with making English syntax like it already is the rest of the time, as people know intuitively.
And it is just that point that comes to bear on Orwell’s Sixth Law – for those who don’t know it, it’s from Politics and the English Language, and it comes after he has made five prescriptions aimed at his ideal of clear, plain, strong, effective usage:
i. Never use a metaphor, simile, or other figure of speech which you are used to seeing in print.
ii. Never use a long word where a short one will do.
iii. If it is possible to cut a word out, always cut it out.
iv. Never use the passive where you can use the active.
v. Never use a foreign phrase, a scientific word, or a jargon word if you can think of an everyday English equivalent.
vi. Break any of these rules sooner than say anything outright barbarous.
I’ll leave aside for the moment debates about his first five rules, which would take us a bit far from the immediate topic (but read, for instance, Orwell’s Liar if you want to start in on it), but the sixth rule’s appeal to avoiding what is “outright barbarous” assumes we know what is outright barbarous, which of course we do by our intuitive sense of the syntax we use all day, every day, except when we’re trying to be more formal then we’re comfortable being. This supports my point – about the grammatical basis for my analysis of whoever – but also supports your point, which is a grand old rule among editors: sometimes it’s best just to avoid the issue by rewriting. Clearly users are uncomfortable with either whoever or whomever here, because their surface analysis conflicts with their intuitive sense, so let’s just avoid it by using a form that doesn’t inflect differently for nominative and accusative.
Incidentally, “This is he” (as opposed to “This is he who…”) is a separate question. Insistence on it is founded largely on Latin conceptions and does not match standard native use; it has always been something of an imposition, demanding as it does an exception to the standard English case filter, which assigns nominative case only to the subject of a verb. It requires a special rule for the copula; rather than making usage more elegant, it actually makes it less so, from a perspective of clarity and consistency of rules. (Similarly, the idea that one must not end a sentence with a preposition, an invention of the 18th century and always an imposition, founded on Latin and French usage, actually requires an addition rule in syntax, whereby a noun phrase when raised must take with it any dominating prepositional phrase – if we see PP and NP as forming together an inflected noun, we can understand the idea, but in English they really don’t, and it’s not natural, as has been well demonstrated.)
Thank you for the very prompt and cogent reply. And admittedly, one man’s meat is another man’s ‘outright barbarous’. The trouble really starts when these two men opt to teach English. I wish they would iron out their differences in their inter-varsity forums and then set the curriculum – without, of course, stifling creativity.
After that, they could tackle world peace. Ah well, it’s an aspiration.
“Probably the most common instance of this confusion is with a sentence such as (often exactly as) Let him who is without sin cast the first stone. It’s so common, in fact, that Google hits for “let he who is without sin” outnumber those for “let him who is without sin” by nearly two to one. But those are all hypercorrections based on misanalysis. …”The main clause is Let him cast the first stone. Note that the verb here is let; it is in the imperative. In English we don’t have a proper third-person imperative, so we use a second-person of let plus the third person in the accusative case. ..”
“The main clause is ‘Let him cast the first stone.’ Note that the verb here is let; it is in the imperative. In English we don’t have a proper third-person imperative, so we use a second-person of let plus the third person in the accusative case.”
This is correct grammar. Struggles with saying it is awkward or not what one hears only lead to debasement of our language.
Although I agree with your analysis, I have a series of questions for you that might justify an alternative interpretation. Consider the use of “that” in a clause that acts as the object of a verb, such as “The teacher suggested that he do his homework.” In this case “he” is the subject of the clause, right? Although I can tell by ear that “do” is the correct form of the verb, I don’t fully understand why. Is it an implied infinitive, like the “cast” you discuss above? In any event, can we correctly say (awkwardly) “Let that he cast the first stone?” If so, can we then elide “that” to get “Let he cast the first stone?”
In “The teacher suggested that he do his homework” (which can be “The teacher suggested he do his homework”), the verb “do” is in the subjunctive. It’s what linguists call an irrealis: it’s referring to a something that is not in actual reality but is merely suggested (or, in other instances, posited or desired). It’s like “It is the order of this court that you be hanged by the neck” – not “that you are hanged by the neck.” This future subjunctive (because it’s referring to a posited future action) happens to be the same in form as the infinitive (in present subjunctive it’s the same as the past-tense “you” form: “If I were a rich man”).
The relative “that” in “The teacher suggested that he do his homework” is deletable because it is only a relative and is not also a pronoun. You can’t delete the “who” in “let her who is without error” or “that of him who seeks” because it’s not just a relative, it’s a pronoun, and it’s the subject of the verb. “Let that he cast the first stone” is unidiomatic because “let” requires a direct noun object with an infinitive verb, not a full finite clause as an object (you could perhaps say “Permit that he cast the first stone” but that wouldn’t mean the same thing; the “let him” construction is a stand-in for a third-person imperative, which English formally lacks, and is not really directing the addressee to allow the action).
I trust that your ear tells you that “Let he cast the first stone” is weird and awkward. This is why: “Let” in this case is an imperative, the only verb form that allows a non-explicit subject, but it requires an explicit object and, fully extended, an infinitive verb phrase. That’s why it’s “You let him! Let him do it!” Not “Let he do it!” or “Let he does it!” And “The teacher suggested he cast the first stone” is altogether different; it’s indicative, the main subject is “The teacher,” and the relative clause is a full finite clause in the subjunctive. It just happens to have a superficial resemblance because of the parsimony of English inflections.