(I'm not sure about the title, but it's the best I can come up with).

On ELU somebody has asked this question about which form of the verb "to be" to use in the frame

She gives a blanket to me who ? cold.

Since neither am nor is sounds right to me, I am left to conclude that, in my idiolect at least, there is no rule that generates a grammatical sentence, and therefore the frame is ungrammatical.

What I mean by contingent grammaticality is that there is nothing intrinsically (Edit:) or intuitively (end edit) ungrammatical about the frame: with any other NP (object NP, if it marks case) in place of me, there is a choice of verb is or are that makes the sentence grammatical. But with this one particular pronoun there is (on my speculative analysis) no choice which makes it grammatical.

Has anybody come across this line of argument before?

  • You do actually come across sentences like this with is. It seems quite literary. It's quite possible it's gone out of fashion because it sounds wrong to most people. Or it could be that it no longer sounds right to many people due it having become rare after having fallen out of fashion. Or both ... Commented Mar 4, 2014 at 19:40

3 Answers 3


You might want to check Heck & Cuartero 2008 ("Long distance agreement in relative clauses", available from Heck's website) for a description of the pattern, references to previous works, and an analysis that you may or may not agree with.

  • Incidentally, I believe that ineffability is a better term for this pattern than "contingent grammaticality".
    – Koldito
    Commented Mar 3, 2014 at 22:50
  • 1
    I must say I disagree with your assessment: Heck & Cuartero seems to consider that She gives a blanket to me who is cold is grammatical (section 5.1) and then proceeds to analyze it whereas @ColinFine considers it ungrammatical. So they are at cross-purpose, so to speak. That said, thanks a lot for this reference, which suggests a nice explanation for a strange agreeing mistake that just about all French children make and which had been puzzling me for a while.
    – Olivier
    Commented Mar 4, 2014 at 9:57
  • That Heck & Cuartero paper has some pretty bizarre judgements in it. "She gives a blanket to me who is cold" also sounds horrific to me.
    – P Elliott
    Commented Mar 4, 2014 at 19:01
  • In fairness to Heck & Cueartero, they don't claim that "She gives a blanket to me who is cold" is grammatical; their example is "He had the nerve to say that to me, who has made him what he is today", which is subtly different. I'd be curious to know how native speakers rate this sentence. Its French counterpart is ungrammatical, contrary to what they claim (footnote 15 notwithstanding).
    – Olivier
    Commented Mar 5, 2014 at 9:28

There's a problem around constructions like that with most American English speakers.
(UK English speakers may claim to speak elsewhise, but pay attention to
what they actually do say, rather than what they claim they "always say").

The fact is that relative clauses virtually always have third person subjects.
This allows the existence of Whiz-Deletion, the rule that relates

  • a boy who is eleven years olda boy eleven years old
  • the man who is shovelling the snowthe man shovelling snow
  • a statue which is on the corner which is made of bronzea statue on the corner made of bronze

So many English speakers are uncomfortable with relative clauses with 1st or 2nd person subjects.
Most 1st- and 2nd-person verbs are am and are, and these are normally contracted to I'm and you're.
So "who am" sounds very very strange to most English speakers.

One way out -- there are quite a few -- is to insert a noun phrase like a person instead of who.
Since a person is third person, one can say is instead of fussing with other inflected forms.

  • 1
    I agree that these constructions are very rare with 2nd or 3rd person antecedents. But to you who are, while stilted, seems perfectly grammatical to my introspection. The same is true of to us who are, though there, as with to him who is, the presence of object pronouns shortly before the verb is a separate challenge to acceptability.
    – Colin Fine
    Commented Mar 4, 2014 at 19:02
  • @ColinFine, for "us who are", I am able to read that as an restrictive who, i.e. "us that are", but not to "me who am/is", not sure if that makes a difference. Maybe it's only non-restrictive relative clauses with 1st or 2nd person antecedents that feel ungrammatical.
    – dainichi
    Commented Mar 12, 2014 at 11:19
  • It's probly the isolated am. Am rarely appears in speech uncontracted, and is almost never stressed. That feels uncomfortable to many people.
    – jlawler
    Commented Mar 12, 2014 at 15:04

The general question asked

Has anybody come across this line of argument before?

hasn't really been considered in the two existing very interesting answers. If by that you mean the study of frames which are perfectly fine except for a few very specific grammatical categories, the answer is yes: contingent grammaticality in your sense is the source of many puzzling (and much studied) phenomena. To name just two of the most famous and pervasive instances, cliticization in Romance and differential object marking (in dozens of languages) are the source of many contingent grammaticality templates.

Elle a donné une couverture à X.
She gave a blanket to X.

is contingently ungrammatical (in your sense) if X is a pronoun and

Juan tocó X.
Juan touched X.

is contingently ungrammatical if X is animate and specific. English relatives are another much studied case.

X that I know.

is contingently ungrammatical if X lacks an article or conversely

The Colin Fine X.

is contingently grammatical if X is a relative.

  • Thanks for actually addressing the question I was asking. However, I don't think your answer is actually to the point either. Perhaps I wasn't clear. Your examples are cases where a frame excludes some word-classes or features. I was asking about a slightly different idea: my intuition is that there is nothing wrong with me as the antecedent of a relative clause, but that neither choice of verb is grammatical for me and therefore the whole construction must be ungrammatical.
    – Colin Fine
    Commented Mar 4, 2014 at 18:52
  • @ColinFine What is the difference between having the intuition that me is a fine antecedent and then realizing that there is no proper way to conjugate the verb in the relative and the intuition that moi is fine as a direct object (any French will say that's fine, just as any French will say their language is SVO without a second thought; strange as it may seem we usually don't think about clitics) and then realizing that it must cliticize?
    – Olivier
    Commented Mar 5, 2014 at 9:10
  • Anecdotally, it is very common in Spoken French that someone embarks in a long sentence and then haplessly stops because it is noticed that the coming oblique object is a pronoun and thus should have appeared way up in the sentence or otherwise the whole sentence will be ungrammatical (in which case usually the speaker just ends up producing an awkward "à lui"-type construction). Anyway, I'm sorry I haven't properly understood your question.
    – Olivier
    Commented Mar 5, 2014 at 9:12

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.