Much less/Or even

« previous post | next post »

Dick Margulis writes:

An NPR reporter this morning, talking about people in Libya: "…have never spoken to a Western reporter, much less seen one."

I hear this frequently (although I don't recall reading it). It is a reversal of what was intended: "have never seen a Western reporter, much less spoken to one."

This occurs with both "much less" and "let alone."

I wouldn't begin to know how to do a corpus search to detect the frequency with which people reverse the arguments of the expression in speech. It occurs to me, though, that the production error seems to be akin to the misnegation phenomenon that you've posted about more than once.

"NEGATIVE X, much less Y" presupposes that X and Y can be placed on some scale of accessibility, such that X is easier/commoner and Y is harder/rarer.  Normally, in fact, you need to pass through X in order to get to Y. A typical example from the NYT:

For all his achievements, Rex never got their attention, much less their rightful applause.

If X and Y are in the opposite order, with X being further out or less accessible, then the normal idiom is "NEGATIVE X or even Y":

Now over thirty years old, he had never dated or even associated with women.

Note that in such examples, X and Y can be inverted by switching idioms:

For all his achievements, Rex never got their rightful applause, or even their attention.
Now over thirty years old, he had never even associated with women, much less dated them.

And Dick is right that people sometimes get their scales and idioms confused. Here are some apparently-backwards web examples with "much less":

I've never tried Trimmit, much less heard of it.
We can read virtual books that we've never bought or borrowed, much less held in our hands.
Before using JavaBuilder, I had never used YAML — much less heard of it.

And here's one with "or even"

It's crazy because I saw Hackers a few times as a kid and never noticed that shit or even thought about it.

And Dick is right that these reversals have something in common with what we've called misnegation, namely confusion about the interaction of negation, semantic scales, and modality (here in the sense of gradation of counterfactual situations).

Individual backwards examples might represent a local semantic confusion — like a word-interchange error, but on a more abstract level — or might reflect a systematic re-interpretation of the meaning of the idiom.  In the case of the NPR report, my money would be on the speech-error interpretation.

But let's note that this use of much less is indeed an idiom, rather than a regular combination of the words much and less.  One source of evidence for this is the fact that attempts to fill out the paradigm of combinations in other ways generally don't work: "X, much more Y", "X, a little less Y", and so on. (Though "still less" does work…) Another piece of evidence is the fact that (what the OED identifies as) the original way "to characterize a statement or suggestion as still more unacceptable than one that has been already denied", namely "less" without any "much", is now distinctly odd:

a1637 B. Jonson Magnetick Lady iii. iv. 72 in Wks. (1640) III,   You never fought with any; lesse, slew any.
1663 B. Gerbier Counsel to Builders sig. g4v,   Dimensions and Formes, which are not to be mended, lesse contradicted.

And the earliest citation for "much less" is from 1671:

1671 Milton Paradise Regain'd iii. 236 The world thou hast not seen, much less her glory.

A Google Books search antedates this to 1655 or 1645 or 1643 (or perhaps beyond), but it still seems to be a 17th-century innovation.


  1. Dick Margulis said,

    February 24, 2011 @ 9:49 am

    Thanks for your comments. I agree that the NPR example, as well as all other examples I recall hearing, seemed like simple speech errors rather than reinterpretation of the idiom. This seems like just another example of our not keeping as strict logical tabs on what comes out of our mouths (sometimes fingers) as we like to flatter ourselves that we do.

  2. Murray Smith said,

    February 24, 2011 @ 10:04 am

    In the age of the telephone, it is often easier to speak to someone than to see them. In that framework, the NPR example is not reversed.

    [(myl) That's certainly another possible explanation. If we had a link to the original story (and most NPR reports are now available as audio files and even as transcripts) we might be able to tell from context whether this was a reversal or not.]

  3. GeorgeW said,

    February 24, 2011 @ 10:32 am

    Murray Smith: That was my initial interpretation as well, particularly in the Libyan revolt context. As an example, Kristof's NYT Op-ed today involved phone calls to people in Libya.

  4. GeorgeW said,

    February 24, 2011 @ 10:56 am

    P.S. Murray and my interpretation support myl's interesting analysis. We were trying to make sense of the X and Y.

  5. John C. said,

    February 24, 2011 @ 11:00 am

    I like thinking of "much less" as an idiom. I hear it used all the time and have never had a problem with it until I read this article.

  6. KevinM said,

    February 24, 2011 @ 11:14 am

    Related: "Each one was more [adjective] than the next." The speaker means that the phenomenon was growing, but is saying that it was decreasing.

  7. Greg B said,

    February 24, 2011 @ 11:21 am

    Thanks for this post. I've noticed the inversion more and more lately (recency illusion alert) with 'let alone': "NEGATIVE Y, let alone X" where Y is harder/rarer.

    For example: "I've never been his friend, let alone speak to him" or "I've never even -been- to Portugal, let alone seen a picture of it".

    My interpretation is that the speaker wants to emphasize the difficulty/rarity of the hypothetical: "I've never done -this- big thing, let alone that smaller irrelevant thing".

    [(myl) It's also possible that some people have lexicalized the "let alone" idiom with the scalar presupposition inverted.]

  8. Spell Me Jeff said,

    February 24, 2011 @ 12:22 pm

    I was going to suggest that an idiom as long and multi-sectioned as this one is almost destined to get reanalyzed in a way that defies surface logic.

    And then I remembered "I could care less." Length, apparently, has little to do with it. We just play fast and loose with negativity in general.

    And still the world turns.

  9. Leo said,

    February 24, 2011 @ 12:22 pm

    I will be amazed – and enlightened – if Murray Smith and GeorgeW's explanation turns out to be correct. Is it really true that Libyans are more likely to have spoken to a Western reporter, but not seen him, than the other way round?

  10. John Lawler said,

    February 24, 2011 @ 12:36 pm

    The classic work on the use, conditions, and meanings of let alone is
    Fillmore, C.J. / P. Kay / M. O’Connor (1988), "Regularity and idiomaticity in grammatical constructions: the case of let alone". Language, 64, 3, 501–538.

  11. Paulus said,

    February 24, 2011 @ 12:41 pm

    This reminds me of the fact that, as a non-native speaker, I have long been puzzled by the idiom "if not". It seems to vacillate between two different senses, as in these examples:

    1. A spokesman apologized for the timing, if not the subtance. [even (though) if he didn't apologize for the subtance.]

    2. It's one of the biggest lobbying firms in Washington, if not the biggest. [or even (possibly) the biggest.]

    When context doesn't make it very obvious, I often have a hard time figuring out which of the two (opposed) senses is meant. Am I missing any clues that make the meaning clear?

  12. richard howland-bolton said,

    February 24, 2011 @ 1:02 pm

    Obviously a form of hysteron proteron, since normally (telephones and avoidance of violence apart) you'd see someone before you'd get close enough to speak to them.
    Those NPR folks are really into their rhetoric!

  13. Brett said,

    February 24, 2011 @ 1:08 pm

    The use of "less" without the "much" in this construction is indeed distinctly odd, but I'm sure that I've seen the analogous (and at least as odd) construction using "more" in old documents at least once or twice. There is also an quite current construction involving "more," which serves a similar function to "much less," although using a completely different syntactic structure. For example, I can say:

    He climbed Mount Everest. What is more, he climbed it from the Tibetan side.

    Although I assume it's much less common, the "what is more" can be shortened just to "more." (In that form, it's something I will sometimes say, but I would probably never write.)

  14. Pflaumbaum said,

    February 24, 2011 @ 4:00 pm

    @ Paulus

    Am I missing any clues that make the meaning clear?

    I guess one clue that works for certain examples like the ones you give above is the use of the emphatic form of the ([ðiː]) for the 'or even' usage.

  15. Sybil said,

    February 24, 2011 @ 4:26 pm

    @Paulus: thanks for your comment, which has made me re-examine why I do not find those locutions problematic. I'll post my thoughts (for what introspection is worth) if I have any.

    This is why I love Language Log. Well, one of the reasons. One is provoked to think.

  16. Eric P Smith said,

    February 24, 2011 @ 6:08 pm

    @ Paulus: Your point about “if not” is interesting. As Sybil suggests in reply, the point gives little difficulty to native English speakers.

    As a native English (UK) speaker, I reckon that the semantic meaning of “A, if not B” is close to “I assert A, but I am deliberately saying nothing about B”. Thus “A spokesman apologised for the timing, if not the substance” means “A spokesman apologised for the timing, but I say nothing about whether he apologised for the substance.” Equally “It’s one of the biggest lobbying firms in Washington, if not the biggest” means “It’s one of the biggest lobbying firms in Washington, and I leave open the question of whether it is the biggest.” In both these instances of “A, if not B”, and I believe in all instances, there is no entailment that B is true, and no entailment that B is false.

    It is when we come to pragmatics that things get interesting. The speaker would not have bothered saying “if not B” if there were no implicature of some sort. Depending on context, the implicature may be anything from “B is true” to “B is false”, with a full spectrum of modal possibilities in between. In your first example, the implicature is “B is probably false”: “he probably didn't apologise for the substance”. In your second example, the implicature is “B may possibly be true”: “it may possibly be the biggest”.

    I’m afraid I can’t suggest any rule of thumb to judge the implicature in any case. A full grasp of the pragmatics of any language is extremely complex, and it is not surprising that native speakers have an advantage.

    [(myl) Are you sure that there isn't a scalar direction implied? It seems to me that "Nine hundred thousand, if not a million" is fine, but "a million, if not nine hundred thousand" is problematic.]

  17. Chris said,

    February 24, 2011 @ 7:52 pm

    As a native Australian English speaker, I definitely see the ambiguity Paulus is referring to. "A, if not B" for me can mean either "A, and maybe even B" or "A, though admittedly not B".

    On the above example, "A spokesman apologised for the timing, if not the substance”- I don't think that can be said truthfully if the speaker is aware that the spokesman did apologise for the substance.

  18. Eric P Smith said,

    February 24, 2011 @ 8:13 pm

    In reply to (myl)’s editorial comment on my comment above: Yes, in a sentence of the form “A, if not B” there is usually a scalar direction presupposed, in that B is stronger assertion than A. But consider “I’ll see you on Saturday if not before”. That can mean “I’ll see you on Saturday, except that if I see you before Saturday then I may cancel our appointment for Saturday”. In that case it is not an instance of the idiom that Paulus addressed. But to my ears it can also mean “I’ll see you on Saturday, whether or not I see you before”. In that latter case, it is an example of the “A if not B” idiom that Paulus addressed, but there is no scalar direction presupposed.

    In reply to Chris: I should not say “A spokesman apologised for the timing, if not the substance” if I knew that the spokesman did apologise for the substance, but that is a matter of pragmatics, not truth values. I might very well say it if I were unsure of whether the spokesman apologised for the substance, and its truth would not be compromised if it later transpired that the spokesman did apologise for the substance.

  19. Hermann Burchard said,

    February 24, 2011 @ 10:14 pm

    @Paulus, Sybil, Eric, Chris:
    As a practitioner of logic (more than of linguistics) I cannot follow Paulus' examples very well at all. "A if not B" as a formula from the Propositional Calculus should be read as "if not B, then A" which is tautological to "A or B," with the inclusive "or", so perhaps clearer "A and/or B." His two examples then become (with some judicious shortening):
    1. "Spokesman apologized for the timing and/or for the substance."
    2. "It's one of the biggest and/or the biggest lobbying among firms in Washington."

    In neither example is truth of "A & B" excluded.

    By contrast, Paulus brings in modal operators such a "possibly." For these there is no explicit indication, so perhaps this discusssion falls under a topic: "Unspoken semantic functions implicit in sequencing of parts of speech."

  20. Hermann Burchard said,

    February 24, 2011 @ 10:42 pm

    As Eric's "stronger assertion" and myl's example indicate, the doubts arise from the possibility that "A implies B" so that the truthvalues are

    "A or B" & "not A or B"

    which is tautological to B.

    This works for the lobbying firm but is not strictly applicable to the spokesman example because "substance" is stronger in a way than "timing" without any outright inference.

    Could this be the source of Paulus dichotomy . .

  21. Hermann Burchard said,

    February 24, 2011 @ 10:52 pm

    May need to switch A B, sorry, for agreement among comments . .

  22. Nelson said,

    February 24, 2011 @ 11:00 pm

    Possible difference:
    "A, if not B." (even though not B)
    In this form, I usually see A as a statement using a verb other than "to be".
    For example, "The reform *reduced* the corruption, if not the inefficiency."
    B is another full statement.
    "It is A, if not B." (it could be B, but if it is not B, it is A."
    In this form, "to be" is usually used and A and B are adjectives or nouns of some sort.

  23. Jack Nelson said,

    February 25, 2011 @ 12:14 am

    There's an old academic joke that relies on this reversal to make the point that teaching forces you to learn: "Have I read the book? I haven't even taught it!"

  24. Jonathan D said,

    February 25, 2011 @ 1:20 am

    I think the interpretation of "A, in not B" has to be all about context. The pragmatics of not saying "if not the substance" when you know he has are a bit of sidetrack – the real question is how we tell the difference between the implications "let's address this much, because we don't know if there was more" and "let's address this much, even though we could talk about what hasn't happened". Can't see anything more than context there.

    I think there is definitely a scalar direction in "A, if not B" – "see you Saturday, if not afterwards" would sound reasonable even less often than "[as low as] 1m, if not 900k".

  25. Jerome Rainey said,

    February 25, 2011 @ 2:45 am

    A charming variation on this idiom by a (fictional) non-native speaker of English:

    "…your sister wanted some new sneakers. She drove me crazy night and day, she wanted sneakers, she wanted sneakers. Anyhow, we couldn’t afford to make any ends, no less start in with sneakers!"

    From the novel "How the García Girls Lost Their Accents" by Julia Alvarez

  26. ChrisD said,

    February 25, 2011 @ 2:53 am

    "to say nothing of" functions rather similarly to "let alone", although it's a bit more restricted in its distribution. It also reminds me of the German equivalent for this which is "geschweige denn" (really sloppily glossed as as a derived form of the word for silence silence plus a logical connective), which has to do with "saying nothing". What do other languages have for this lexical item?

  27. Paulus said,

    February 25, 2011 @ 7:49 am

    Running through COCA examples of "A, if not B", type 1 ("A, but not B") seems much less common than type 2 ("A or even B"). As Mark notes, many type 2 examples ("weeks if not a year", "difficult if not impossible") are clearly based on a scale, often a metric scale. Expressions like "one of the biggest if not the biggest" are clues that the "if not" is a case of type 2. Type 1 cases, on the other hand, are not always as obviously related to a scalar direction: "Reclaiming the love of his life, if not his memory", "[Maoism] is abandoned, effectively, if not in name", "There is an argument that a certain degree of regulatory forebearance may have exacerbated, if not directly caused, this crisis".

    Regarding Eric's point that the two types are equivalent with respect to truth conditions, I think I agree with Chris that "The holiday shopping season seems to be off to a solid, if not spectacular start" is not consistent with "For all I know, it could be spectatcular". Here, as in other type 1 cases, that "not B" seems to be a full semantic implication, rather than a mere pargmatic implicature. I think you couldn't truthfully utter the Maoism sentence if you were unsure whether Maoism was abandoned in name. It might true, though, that type 2 cases can be explained in terms of implicature.

    Perhaps Jonathan D is right that all we have to go on (in cases that aren't obviously type 2 ("if not the biggest")) is context. In that case, you can correctly intepret "an Iran that, if not more friendly to the United States, is more integrated to the international community" only if you already know that Iran isn't cozying up to the U.S. anytime soon.

    @Nelson: I'm not sure the verb "to be" is special in this regard. There are plenyt of Instances of type 2 with other verbs ("Hundres if not thousands of employees could lose their jobs").

  28. George said,

    February 25, 2011 @ 9:04 am

    OT perhaps but still broadly in the "which of the two is the less likely?" realm is the difference between 'as well as' in English and 'aussi bien que' in French. So "in France as well as in Germany" becomes "en Allemagne aussi bien qu'en France". I had big problems with that early on in my acquisition of French.

  29. Hermann Burchard said,

    February 25, 2011 @ 9:34 am

    all about context .. which is to say, all about semantics. Still looking for a good reference on who first explained this.

  30. baylink said,

    February 27, 2011 @ 10:54 am

    I'm generally hearing an implicit "if not A, then *at least* B", which does for me.

RSS feed for comments on this post