I recently posted about the “is is” phenomenon, which has been much discussed (see this Language Log post, for example); people may find it more or less acceptable, but one could hardly be surprised by its existence. Now, however, Logger Mark Liberman has written about a construction that I find so improbable I’m amazed it’s as common as it apparently is, the doubling of that around an adverb. Examples:
“There are statements that obviously that she has made that the president doesn’t agree with…”
“… knowing that in most cases that they will mess up the drawings.”
“…it seems that apparently that they just wanna be me…”
I assume, perhaps too hastily, that no one will find this grammatical, but is anyone familiar with it? Do any of you find yourself occasionally producing a sentence like this? It may be an up-and-coming phenomenon that I just haven’t noticed, geezer that I am.


  1. I don’t *know* that I’ve produced such a sentence, but it wouldn’t startle me. They’re “buying time” phrases, which you introduce because you don’t know what you’re going to say yet. You put your surface mind on autopilot and you forget that the “that” was already inserted at the beginning of the meaningless noise (“apparently,” “in most cases,” “obviously”, whatever) — so when your brain finally engages and you’re ready to generate meaningful language again, you throw it in, just in case. The “that” is so weak semantically — it’s functioning as a verbal colon or dash, basically, that — that — you see? — that you just keep throwing it in.
    I’m *the* native speaker to consult for “slow on your feet” phenomena 🙂

  2. To me this just seems like hesitation (performance error)… you start your sentence, get to “that” (a natural breaking point), then restart at “that”. If you left out the second “that”, it could sound odd or even ambiguous, especially since you can’t use commas in speech: “… statements that [pause] obviously [pause] they will mess up the drawings.”
    It also seems significant to me that many examples are from spokespeople such as Gibbs; spokespeople have to be especially careful with indirect discourse, since their statements are likely to be heavily edited. Leaving an extra “that” helps to ensure that these third-party statements are not misconstrued or misrepresented as something that the spokesperson is saying in his/her own voice.
    At any rate, I’m somewhat confuzzled by the idea that this has anything to do with the underlying grammar; it seems to me like an obvious performance issue.

  3. Some of the examples at Language Log are written, not spoken. Therefore “slow on your feet” and “hesitation” do not seem to be the whole story.

  4. I haven’t encountered this in writing, and would find it strange there. Listening, I probably wouldn’t even notice it; it sounds very much like the sort of thing I hear on Sunday-morning shows and just mentally edit out.

  5. I regularly produce misplaced “that”-s after adverbial clauses (e.g. “he said (that) if he was running late, that he would call to let us know”, with or without the first “that”), and while I consciously know that it would make better sense to drop them (or in examples of the type I just gave, to replace them with apodosis-marking “then”-s), I just can’t bring myself to mind; but paired “that”-s around a one-word adverb? That’s shocking.

  6. I’ve noticed that, under the right circumstances, if the intervening clauses are long enough and numerous enough — or simply just wordy enough — for the initial “that” to lose some of the semantic glue necessary to bind it to the following corresponding clause, that the word “that” can get repeated without it sounding too unnatural. 😉
    I’ve never seen or heard it sandwiching a word, though. That’s just weird.

  7. I’ve done that, in both writing and speaking, and it always makes me and my friends laugh. We always marked it as one of those signs that my English skills were getting messed up after living in Japan for too long…

  8. Yes, I produce sentences like that, usually after changing the sentence around too many times, but hopefully I catch all of them before I hit “post”. If you take the second example,
    “… knowing that in most cases that they will mess up the drawings”
    it can be written in two ways:
    “…knowing that in most cases they will mess up the drawings”
    “..knowing in most cases they will mess up the drawings.”
    If you rewrite the sentence, the second time around you may put “that” in a different location, forgetting it is already in the sentence. It’s good argument for proofreading (or at least using the “preview” function), or at least.

  9. Yeah, I left it in intentionally. The first time I wrote “It’s a good argument for using the preview function or at least proofreading.” Then I decided proofreading was the salient point, and the preview thingy was too meta. After cutting and pasting and second guessing my initial word choice, I ended up with the prepositional phrase “or at least” in the wrong spot, or rather in two spots. Same kind of error.

  10. Rather, the two possibilities are
    “…knowing that in most cases they will mess up the drawings.”
    “…knowing in most cases that they will mess up the drawings.”
    Besides the obvious advice about proofreading, there’s probably also a cautionary tale in there somewhere about being your own editor.

  11. I think that, usually, the second “that” is, as Sam says well, a ‘re-start’ – more a stuttered return than a hesitation, it seems to me. The adverb / adverbial phrase would be a self-interruption, from which the speaker would return to the ‘clausal’ flow by repeating an open-bracket function.

    On the other hand, sometimes, the adverb / adverbial phrase could be substituting for an actual ‘clause introducer’:
    “I think that, obviously, it’s the case that glaciers are Wicked Witches of the West.”
    Now, whether it’s “the case” which is obvious, or that “it’s the case”, or that “I think” it’s the case — whatever it is that’s “obvious” is, in my de-glaciation example, systematically ambiguous.
    Which would compel the question of whether unclear expression is, in this case, a mirror of unclear thought . . .

  12. michael farris says:

    I’m suspicious anytime people start trying to explain real phenomena away as performance errors.
    Just as in traditional descriptive linguistics ‘euphony’ is invoked anytime the analyst can’t figure out just what some abiguous polysemic morpheme ‘really’ means, performance error is invoked anytime people are confronted with usage that doesn’t conform to what they want to exist.
    Yeah, performance errors exist, I see them all the time in my own comments since I stubbornly refuse to use the godddamned ‘preview’ option (previewing is for wimps!)
    All that said, I think it’s pretty simple; ‘that’ introduces subordinate clauses, and in real spontaneous usage (as opposed to usage that’s been fussed over by copy editors and other arbiters of correctness) it can occur very, very frequently without being perceived as obtrusive and in fact aids in realtime parsing.
    I didn’t even notice anything weird about the title of the post until after I read what it’s about. The brain perceives ‘that’ hardly consciously noticing what’s going on but still grateful for the help in keeping the sintax strait.
    (nb. a comment on another thread about spelling reform has re-activated that part of my brain so expect lots of ad hoc spelling experimentacion on my part for the time being)

  13. I frequently find myself tempted to use double thats, but I edit them out unless I’m not editing. I suspect that if someone looked at a number of instances they’d find sentence-types where a single that, even though grammatical, doesn’t seem to be enough.

  14. I could see myself repeating “that” as a buttress midway through a long complicated noun clause. But I can’t construct a plausible example just now. It would be an inelegant construction at best.
    Certainly, two “that”s separated by one or very few words reads wrong to me.

  15. OK, it’s just my geezerdom, then!
    I produce sentences like that, usually after changing the sentence around too many times
    Yeah, I understand that phenomenon, it was the ones that didn’t seem to be the product of excessive rewriting that puzzled me.

  16. The “is is” construction I have heard, but not “that adverb that”. If someone could come up with some audio, it would be easier to tell if it was an intentional construction or some sort of vocalized pause–or inattention. Also, at least one of the examples at The Log was African American.

  17. If we’re working with spoken language here, I can handle this the same was as “um” in the US or “uh” in France, but written, no. When I taught English before switching over to a stupid career, I always taught that the word “that” should be pruned from a sentence just about any time (that) it could be without obscuring the meaning of the sentence.

  18. I’m suspicious anytime people start trying to explain real phenomena away as performance errors.
    “Away”? Like: “can occur very, very frequently without being perceived as obtrusive”?
    ‘I think that clearly that it’s weird.’: No frustrated anticipations of conformist usage — if the emphasized “that” is grammatical, what is it doing? if it’s not grammatical, what does its presence – surely, if this ‘not’ is the case, a “performance error” (??) – indicate? (The latter having been language hat’s question, as I understand the blogicle.)

  19. Bill Walderman says:

    From time to time I find myself doubling up thats both in speaking and in writing. It’s not grammatical in my idiolect, however. As someone else remarked, in writing it’s simply a result of rewriting a sentence too much, and in speaking it’s just a performance error on my part.
    But I don’t agree with deleting “that” wherever you can get away with it. “That” is a useful for clarifying the structure of a complex sentence.

  20. “That” is a useful [technique] for clarifying the structure of a complex sentence.
    I agree, Bill; a grammatically doubled that can be more clear than leaving one out (and remaining grammatical) or leaving both out ungrammatically. I’m thinking of something like: “He said that that dog was barking.”
    You could leave the first “that” out, but then the other person might ask, “I heard him say that one of the dogs was barking! Which f%#&ing dog?!” You could add a “not this dog”, meaning the barker is somehow farther away or ‘posterior’, or you could use a different kind of linguisticality (like a pointing finger).
    But I’m pretty sure, in the first place, that I’d use both “that”s, to be clear that a) this is what he claims, and b) somehow, I’m mirroring in my expression the differentiation “he” made between the barker and the mute whatever.

  21. This construction is very common among my students in spoken usage. I try to discourage it in written usage because I feel like it’s not well-established as common practice and may look like an error to some readers, but obviously it is no impediment to communication, so I wouldn’t call it wrong. That said, it “feels” wrong to me (because both “thats” are doing the same thing, hence it feels reduplicative) and I would never use it in my own speech.

  22. ToussianMuso says:

    The doubling of “that” seems to me to have the same function as the doubling of “is”: the word seems in the speaker’s mind to belong to both phrases, and so one covers the bases with redundancy, for example, “The thing is, is people talk this way.” It’s like a little internal psychological debate: does the “is” REALLY belong with the preceding noun phrase or the following clause? Well, both, darn it!
    The growing habit of inserting “that” in every possible grammatical place (“The other thing is, is that, oddly enough, that people attempt to explain it.”) would seem to warrant a similar explanation.
    Such phenomena may be acceptable in speech, but I can’t imagine them being grammatical in writing. I’m not sure if this is in spite of my best efforts to be a good descriptive linguist or because of them, as I have never seen such constructions in written text, except in the examples given here which are presumably citations of spoken comments.

  23. ToussianMuso says:

    Basically, the function is purely psychological rather than grammatical.

Speak Your Mind