No detail too small
« previous post | next post »
John V. Burke wrote to draw my attention to a phrase in Walter Kaiser's "Saving the Magic City", NYRB, 12/3/2009 (emphasis added):
Roeck's book, for which he has done an impressive amount of research, tries to be a number of things at once: it is an account of the social and intellectual world of the expatriate community in fin-de-siècle Florence; it continues the biography of Aby Warburg he began with his earlier book; it is a history of late-nineteenth-century Florentine urban development; it is a cultural history; it addresses a wide variety of ancillary topics such as anti-Semitism, anarchism, labor conditions, and economic trends; and it discusses the various aesthetic theories being formulated at the turn of the century. No detail is too small to escape Roeck's net, not even the plans formed in 1898 to produce artificial ice commercially in Florence.
This echoes the classic example "No head injury is too trivial to ignore", discussed by Peter Wason and Shuli Reich, "A Verbal Illusion", The Quarterly Journal of Experimental Psychology 31(4):591-97, 1979.
Something interesting is going on here, as Wason and Reich observed, but it's not entirely clear what it is. The most striking aspect of the problem is that it's so hard for people to recognize that it exists.
Let's start with some real-world instances of the pattern No NOUN is too small to VERB:
No business is too small to play a role.
No donation is too small to make a difference.
No amount is too small to benefit from compounding interest.
No country is too small to make it as a sovereign nation.
No exposure is too small to intitiate cellular damage.
No news item is too small to be lampooned in the blogosphere.
No occasion is too small to warrant a gift.
You could translate these as something like All NOUNs, however small, [can or should] VERB. Thus the last one comes out "All occasions, however small, warrant a gift".
To understand how this works, consider the positive statement "This occasion is too small to warrant a gift". This presupposes that we can set a value on the scale of occasion size, such that larger occasions warrant a gift, while smaller ones don't; and it asserts that the designated occasion is below the threshold. Now if we say that "No occasion is too small to warrant a gift", we're asserting that the gift-warranting threshold is so low, in the context under discussion (which was Japanese culture), that no occasion worth mentioning is below it.
One additional wrinkle: In the pattern NOUN is too ADJECTIVE to VERB, the initial NOUN might either be the subject of the VERB ("The patient is too weak to eat") or the object ("The soup is too hot to eat"). All the examples that we've seen so far are subject-type cases; but object-type cases exist as well:
No service is too small to render to the King.
No problem is too small to obsess over.
But now consider these examples:
No pavement irregularity is too small to overlook.
No social injustice is too small to ignore.
No risk is ever too small to disregard.
No convenience is too small to forget.
No precaution is too small to be omitted.
No detail is too small for you to discard it as insignificant.
No expense is too small to escape the finance chief's attention
No worthy cause was too small to escape his generosity.
These can't be translated in the same way as the earlier batch. "All worthy causes, however small, escaped his generosity" — no, that's not what the writer meant at all.
And the corresponding positive assertions are already problematic in the same way, e.g. "This expense is too small to escape the finance chief's attention". To make them make sense, we want to flip something: "This expense was too large to escape the finance chief's attention"; "This expense was too small to attract the finance chief's attention".
But examples like "No detail is too small to overlook" are very common. You see them all the time, even in the work of careful writers like Walter Kaiser, in well-edited publications like the New York Review of Books. Legend has it that "No head injury is too trivial to ignore" was posted on the walls of London hospitals for decades. I suspect that this is untrue, but it's a good story — certainly versions of this same apparently incoherent phrase, attributed to Hippocrates, have been published many times in reputable books, journals, and other places, e.g. here, here and here.
So what's going on here? As discussed in many earlier LL posts, these puzzles tend to arise from interactions among negatives, scalar predicates, modals. And explanations include variants on "it's a mistake", "it's an ambiguity", and "it's an idiom".
There are many different types of what we've called misnegation (also overnegation and undernegation), and different explanations may be appropriate for different examples. In this case, though, I think that Wason and Reich's 1979 account has held up pretty well. I'll lay out their hypotheses and their experimental results in another post.
[Note — Francis Adams' translation of Hippocrates On Injuries of the Head has nothing that seems likely to be the source of the phrase "No head injury is too severe to despair of, nor too trivial to ignore". If you know where in the Hippocratic Corpus this comes from, please let us know in the comments.]
Robert Pollack said,
November 27, 2009 @ 4:46 pm
I wonder if "too" is being understood in these phrases more like "so."
[(myl) Not quite, since "No detail is so small to ignore", etc., are not exactly idiomatic. But if you fix up the rest of the phrase — "No detail is so small that it can safely be ignored" or whatever — then it works to express what the writer meant.]
Tim K said,
November 27, 2009 @ 4:47 pm
Fascinating. I read this whole article twice, and was about to post a comment asking you what you were on about, before I caught the error.
Interesting. So the author meant: "No detail is so small that it can escape Roeck's net."
But it took me several readings before I realized the author did not really say that. Neato.
Tim said,
November 27, 2009 @ 5:32 pm
Interesting that these are phrased in the active voice, even when the subject is really the object of the infinitive. That is, "The soup is too hot to eat" rather than "The soup is too hot to be eaten". The latter seems more "correct" when you think about it, but I don't think it's what most people would actually say.
Faldone said,
November 27, 2009 @ 7:01 pm
I think I've read most, if not all, of the examples of this error and I have to be convinced anew that it is an error every time I hear one.
Jerry Friedman said,
November 27, 2009 @ 8:49 pm
Thanks, Prof. Liberman. I try to write carefully and have succeeded to the extent of reading peevishly, but I imagine I've missed many of these. I hope I haven't missed too many in my writing.
Jerry Friedman said,
November 27, 2009 @ 8:55 pm
Also, if we're speculating on what causes this kind of error (there, I said it!), I'd explain it the same way as the misunderstanding of "This book fills a much-needed gap." "Fills a gap" and "much-needed" are often complimentary, so readers and quoters may simply take a combination of them to be complimentary, unless there's some obvious reason not to, such as a negation. Likewise "too small" and "escape someone's net" are often used to talk about how attentive to detail someone is, so the combination is taken that way.
Possibly compare "dethaw", "unravel", and "unloosen".
Karen said,
November 27, 2009 @ 9:21 pm
@Tim: That's a very common English usage, though, sort of like the classic "middle" or "medio-passive" (which is something like "she doesn't frighten easily" which means "it's hard to frighten her" rather than "it's hard for her to frighten others").
David Green said,
November 27, 2009 @ 11:22 pm
Here's a passage that seems semantically (though not structurally) similar, from comments on a news story about an elderly couple drowned when their small boat broke up in heavy seas:
"Anyone venturing into our waters hereabouts should have the sense to (1) check conditions and (2) be prepared. Lots of inexperienced and/or careless boatsmen as is evident with the numerous drowning deaths reported every year. It's deceptively treacherous out there…"
"Deceptively treacherous" … hmm….
rootlesscosmo said,
November 27, 2009 @ 11:34 pm
One way to clarify the meaning is to reverse the descriptor: "This detail is too large to escape the net" means the net catches details above a certain size, which this detail exceeds; "too small to escape" seems to imply that the net catches details below a certain size while larger ones escape, which is not how nets work.
slobone said,
November 28, 2009 @ 1:59 am
Reminds me of the celebrated Bushism "Our enemies are innovative and resourceful, and so are we. They never stop thinking about new ways to harm our country and our people, and neither do we."
Serge said,
November 28, 2009 @ 7:24 am
One way of explaining the frequency of cases like No head injury is too trivial to ignore is by considering the constituents like: No head injury is too trivial; No occasion is too small, which do not differ considerably in their meaning. They act as a semantic head, to which a clarification can be added: to ignore it, to warrant a gift.
Victor said,
November 28, 2009 @ 8:13 am
I'm still confused about this "error".
Let's analyze it by negating two negatives, which should produce an equivalent sentense. For example, "nothing left undone" can be written as "everything was done".
Applying this to
No detail is too small to escape Roeck's net.
we get
Every detail is too large to escape Roeck's net.
which is the intended meaning: details are large relative to Roeck's net, and they don't escape it.
On the other hand, in "no business is too small to play a role," this produces the incorrect "all businesses are too large to play a role."
What's going on here?
[(myl) Your method tells us that if no one is too short, then everyone is too tall. But negation doesn't work that way.]
mgh said,
November 28, 2009 @ 9:47 am
Victor, I think the opposite of "too small" is "large enough"
Victor said,
November 28, 2009 @ 10:33 am
Good point.
Maybe this naive negation is the reason why people see no error in "no detail is too small to escape Roeck's net" on cursory reading.
Spell Me Jeff said,
November 28, 2009 @ 2:08 pm
Is there a linguistic principle our there that says something like, "If an utterance seems to make sense, it will make sense, logic be damned"?
Conversations often have a small hitch where the speaker backs up to make a correction, and the listener says, "That's okay; I know what you meant."
Making sense out of things that don't stand up to logical analysis strikes me as a very context-bound process. To begin with, all the examples under discussion involve variations on a shopworn construction. Then they are embedded in a context where the intended sense is incipient. When the construction opens, the reader completes the intention, even when the words don't actually go there.
I should think that even AI could be trained to flip-flop a negation to make sense of something. Bushisms are pretty tricky in comparison.
Frans said,
November 28, 2009 @ 3:51 pm
It took me some rereading before I realized what was wrong, and, even now that I do, I almost can't see any other meaning than "No detail is so small that it can escape Roeck's net" upon reading the fragment.
[(myl) Yes, this illusion (if that's what it is) is very potent. See here for some further analysis.]
Flora Steele said,
November 29, 2009 @ 3:08 am
"This book fills a much-needed gap."
Was that a mistake or a review by Dorothy Parker?
ben said,
November 29, 2009 @ 4:26 pm
"No precaution is too small to be omitted."
This one could also be written "no precaution is too small to omit", and in general the examples in the second class seem to able to be phrased passively or actively indifferently (except for the ones involving "escape"), with "No detail is too small for you to discard it as insignificant." becoming simply "no detail is to small to be discarded as insignificant" (since no one would include "by you" there). & (again aside from the ones involving "escape") in the second set the agent of the verb is something other than whatever is too adjective, while in the first the agent is the thing which is too adjective.
speedwell said,
November 30, 2009 @ 11:46 am
I can only make sense of the examples by taking them as "[something] is [too A to B]", reading "too A to B" first, deciding whether that part makes sense, and then (if it does make sense) thinking whether the "no/all something" meets or does not meet the condition of "too A to B."
"No donation is too small to make a difference" would then become, in my mind, "Can something be too small to make a difference? Yes, that is possible. May a donation be too small to make a difference? Possibly, but the statement argues that this is not the case. Valid sentence."
"No social injustice is too small to ignore" would become, "Can something be too small to ignore? Well, the newest miniaturized cell phone, the interval of time between connecting flights, or the only child soldier in the lineup, might be too small to ignore. Can a social injustice be something that is too small to ignore? No, it is not that sort of thing. Invalid sentence."
"No head injury is too trivial to ignore" would become, "Can something be too trivial to ignore? No, therein lies a contradiction, as if I had said 'too feline to be a cat'. Invalid sentence."
Leo Petr said,
November 30, 2009 @ 5:33 pm
I am not a linguist, but I would interpret the phrase as a short form of:
[There is] No detail [that] is too small to escape Roeck's net.
Both "there is" and "that" are frequently omitted in colloquial English.
Ellen said,
November 30, 2009 @ 6:31 pm
Leo Petr, I think that's a given. It also doesn't change anything.
The problem, here, is with "too small to escape Roeck's net". (See the excellent comment above from Speedwell.) It's small things that can escape a literal net, by escaping through the holes. "Too small to escape a net" doesn't make sense with real nets. A figurative net like Roeck's net would be the same. Big things will be caught, but small things can escape. The scale is backwards.
Rick S said,
November 30, 2009 @ 7:48 pm
Here's a hypothesis that I think could explain these expressions. Statements of the form "No NOUN is too ADJECTIVE to VERB" anticipate and are meant to counter a proposition that the speaker expects a hearer to assume, or to be about to state (perhaps as an exception, an objection, or an excuse). These propositions can generally take one of two equivalent forms: something like either "It is too ADJECTIVE to consider" or "It is ADJECTIVE enough to ignore". Since the proposition is unstated, both forms carry equal weight in the speaker's mind, and in formulating a rebuttal to both forms simultaneously, the speaker blends them in the process of negating them.
As a possible test of this hypothesis, we might look for blends where the order of the proposition forms is reversed. For example, "No NOUN is too small to ignore" is attested, but is "No NOUN is small enough to consider"? If so, what (if anything) does their relative frequency suggest?
flug said,
January 6, 2010 @ 3:19 am
"No detail is too small to escape Roeck's net"
vs: "Roeck has woven a net of such exquisite fineness that it sweeps up every detail, no matter how small."
Thinking as a writer, part of the problem is that you really don't want to get into the mechanics of the metaphor (the net) here–it just doesn't add anything.
What you're really interested in doing is emphasizing the smallness of the details that Roeck is picking up.
So you put together a phrase emphasizing smallness ("No detail is too small") with word showing that he is rounding things up ("net") and don't sweat the details too much.
We're using "net" in a way that leverages one particular aspect of its meaning (it gathers things up) without worrying about–or even thinking about–how actual nets work in reality.
In short, a metaphorical/illustrative/suggestive approach to language function rather than strictly logical.
It is a little like how advertising copy or product packaging will sprinkle little bits of suggestive words or phrases here and there without even pretending to make whole or even partial sentences. "New, improved!" "Adds sparkle" "Pet friendly!" "Just add water" "Mmm-mmm good!" "100% natural" "Freezes in seconds!" "Nontoxic"
The human mind is very good at filling in the details and assembling these suggestive little bits of information into a narrative that is as coherent as we need.
When these fragments are (even superficially) packaged in the form of complete sentences, it only helps us in assembling our internal narrative that seems perfectly complete and sensible.
That fact that sentence makes no sense when logically parsed out, is really irrelevant.
flug said,
January 6, 2010 @ 3:20 am
Another thing that strikes me in thinking about the phrase is how the two different ways of understanding it (the one apparently intended by the author, which I'll call "metaphorical" and the one you get by carefully parsing the words, which I'll call "logical") are like those trick paintings which you can see as a young woman or an old lady's face–but you can't really see both at once. At best, you can sort of quickly flicker between the two different ways of perceiving the painting.
If I just accept the metaphorical meaning, all is well. But when I try to force myself to think about the phrase in its strictly logical meaning it starts to flicker back and forth between that and the metaphorical meaning. This makes it very uncomfortable/confusing to think about.
One reason for that, I suppose, is that in the larger context of the passage the logical meaning of this phrase is completely nonsensical, whereas the metaphorical meaning makes perfect sense in context.
So the logical meaning of the phrase is actual illogical or meaningless in context, making it hard to think about.