Linguistic commentary from a guy who takes things too literally

Humility, Nobility, Etc., Part II

Posted by Neal on March 19, 2008

So as I was saying, something always bothered me about the line

You have humility, nobility, and a sense of honor that are very rare indeed.

My reaction was always, “…that are very rare? Don’t they mean is?” This thought was always followed by, “Oh, wait, it’s not just the sense of honor that’s rare. The humility and nobility are, too.” And then: “OK, so why do I keep wanting it to be just the sense of honor that’s rare?” And ultimately, I’d give up.

This was several years before I even knew the term relative clause, even more before I learned the difference between restrictive (aka integrated) relative clauses and nonrestrictive (aka non-integrated or supplementary), illustrated here:

  1. Integrated relative: the UV light (that) we bought to help find the urine
  2. the UV light, which we bought to help find the urine

The integrated relative singles out just the UV light that we bought for urine-detection purposes from all the other UV lights that I might want to talk about, like the one for my fluorescent mineral collection or my home-tanning bed. Or, if there’s already only one UV light that I could reasonably be talking about, the integrated relative clause might just be a backgrounded reminder of why we bought it. (Cases like the latter are why some grammarians prefer the more general term integrated relative clause to restrictive relative clause, since the relative clause isn’t being used to restrict some set of possible referents. To make the point even clearer, consider the sun that is the source of all life.)

The supplementary relative, OTOH, has to refer to an object that has definitely already been identified — some particular UV light that we all know about. Unlike the integrated relative clause, a supplementary relative clause usually presents new information, or at least information that the speaker is
making more prominent than it would be in an integrated relative clause.

So that’s the semantic difference between integrated and supplementary relative clauses. The semantic difference corresponds to a syntactic difference. I’ll start with the integrated relative clause: the UV light that we bought is structured like this:


The relative clause forms a chunk (“constituent”) with the noun, and that entire constituent (labeled N’) is what gets attached to the determiner. (How do we know this? I’ll save that for another post, but if you’re impatient, see pp. 1061-1063 in CGEL for an overview.)

The supplementary relative clause, however, comes in higher up in the tree. Some have argued that it’s best represented like this, which will be good enough for our purposes:


An interesting piece of evidence for these different structures is the fact that proper nouns (which act as full NPs) can be referred to by supplemental relative clauses (e.g. Charlie Brown, who never wins), but not by integrated ones (*Charlie Brown who/that never wins). If you want to put an integrated relative clause with a proper noun, you have to turn it into a common noun and put it with a determiner: the Charlie Brown I know; every Linus I’ve ever met.

Coming back to the song lyric, that are very rare indeed has to be an integrated relative clause, because supplementary ones can’t start with that. Therefore, the farthest back it can go to form a constituent is sense of honor. Farther than that, and it takes in the determiner a, and it would have to be a supplementary relative clause. So our constituent is sense of honor that are very rare indeed, and we have a number disagreement between sense and are.

But wait: I’m assuming that the songwriters intended the relative clause to modify only sense of honor, and not humility and nobility. That is, I’m assuming they wanted to say that Charlie Brown has some nonspecific kind of humility and nobility (maybe because there aren’t different varieties of h. and n.), but that he doesn’t have just any old sense of honor: He has a very rare s. of h. What if this assumption is wrong? What if the songwriters wanted to say that Charlie Brown doesn’t have ordinary humility or nobility, but a special, rare variety of each? They probably did mean to say this; after all, they chose are rare instead of is rare.

Again, the problem is that the integrated relative clause that are very rare has to modify an N, not an entire NP. It can do this just fine with an abstract noun like humility. Just imagine a structure something like this:


Humility is ambiguous between being an N or a full NP. So is nobility. However, when you coordinate them with the NP a sense of honor, you’ve effectively declared them as NPs, and now they’re not eligible for an integrated relative clause anymore. If you want to talk not about humility, nobility, and a sense of honor in general, but about rare types of each, you just can’t do it the way the songwriters here tried to.

One more angle to consider: Maybe the songwriters were not restricting themselves to just the rare kinds of humility, nobility, and senses of honor. Maybe they wanted to say that these qualities in any form are rare. In that case, what they wanted was a supplementary relative clause, which can associated with full NPs without a problem. In that case, the are would have been fine, but they should have used a which instead of a that. (And, as at least one commenter noted, you could also just say which is very rare indeed — in other words, it’s rare for anyone to have these qualities, which amounts to the same thing as saying these qualities are rare.)

I think it’s really cool that theoretical syntax can shine a light on something that was a mystery to me for years, and pinpoint the source of the offense. Though I haven’t pointed out where my analysis agrees and disagrees with those of the commenters, I trust that each of them can take the relevant parts of this post as my response. Thanks, TootsNYC, Ran, Kip, and Ellen K. for your thoughts!


8 Responses to “Humility, Nobility, Etc., Part II”

  1. Ran said

    Wait, so to you ?”You have humility that’s very rare indeed” is perfectly acceptable? Because to me it sounds flat-out wrong. (I can’t actually put my finger on what’s wrong with it — something like “He wishes he had humility that he doesn’t” is maybe a bit awkward, but grammatically fine — but my ear can’t accept that example.)

  2. Neal said

    Right: humility that’s rare is OK for me. I don’t generally think of humility as coming in more and less common varieties, but I can accommodate my views to accept such a thought if someone expresses it. Also, it would be smoother to say it as rare humility, but since that’s ambiguous between a restrictive and non-restrictive use of the adjective, I could see humility that’s rare as a choice to reduce the ambiguity.

  3. Ran said

    I too can accommodate my views to accept such a thought, but (as I said previously), only if someone expresses it by using “humility” as a count noun designating a variety of humility: “You have a humility that’s very rare indeed.” (Similarly, for me ?”You have rare humility” sounds only borderline-grammatical, while “You have a rare humility” sounds quite fine.)

  4. Neal said

    Is this true of all mass nouns for you? For example, can you say information that’s hard to get or water that you can’t drink? If so, is it a semantic restriction, such that just abstract mass nouns can’t be modified by an integrated RC?

  5. Neal Goldfarb said


    I think your 5th-grade intuition is explained better by psycholinguistics than theoretical syntax.

    The psycholinguistic explanation is that attaching the relative clause to “sense of honor” is preferred under Late Closure and Minimal Attachment.

    The problem I have with the theoretical-syntax explanation is that it depends too much on theoretical notions of constituency whose psycholinguistic significance in this context may be relatively minor compared to the effect of the parsing issues. Note that under your explanation, [NP [DET NULL] [N'[N humility] [N nobility] [CONJ and] [N honesty]] [RelC that are very rare]] would be ok; it’s the fact that “sense of honor” is preceded by a determiner that (in theory) renders a similar structure for the actual lyrics ungrammatical. But as you’re reading or hearing the line, there’s no ungrammaticality (even in theory) until you get through “a sense of honor” and hit “that are very rare.” And by that point, the double-whammy of Late Closure combined with Minimal Attachment push you toward attaching the RC to “a sense of honor.” As a result, the parse that would lead to the constituency problem never even arises.

  6. Neal Goldfarb said

    Where’s the edit function when you need it?

    The first paragraph of my post should be “I think your 5th-grade intuition is explained better by psycholinguistics than BY theoretical syntax.”

    I wouldn’t want you to think I meant to say that your intuition is explained better by psycholinguistics than theoretical syntax is.

  7. Neal Goldfarb said

    One more thing. Does “you have humility, nobility, and honesty that are very rare” sounds just as odd to you as “you have humility, nobility, and a sense of that are very rare”? If so, I think that would weigh in favor of the psycholinguistic explanation.

  8. Neal said

    Neal G:
    Your psycholinguistic hypothesis makes sense, but at least in my grammar, it is not supported. For me humility, nobility, and honesty that are rare is fine.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s

%d bloggers like this: