February 26, 2004

Why are negations so easy to fail to miss?

Over the past month or so, a series of posts here have sketched an interesting psycholinguistic problem, and also hinted at a new method for investigating it. The problem is that people often get confused about negation. More exactly, the problem is to define when and how and why people get confused about negation, not only in intepreting sentences but also in creating them. The method is "Google psycholinguistics": the analysis of internet text as a corpus, as a supplement to more traditional methods like picture description, reaction time measurements or eye tracking.

This all started with could care less. It's clear that this phrase has become an idiom, meaning "don't care", even if it's not clear exactly how the not disappeared from the apparent source cliché couldn't care less. In this Language Log post from last month, Chris Potts discusses a range of other examples where the presence or absence of negation seems to leave the meaning (in some sense) unchanged. For example: "That'll teach you (not) to tease the alligators."

Followups in our pages and elsewhere (here, here, here, here, here) discussed many cases of developments of a different kind, where extra negations create an interpretation at odds with what the writer or speaker meant. An antique and canonical example (cited by Kai von Fintel) is "No head injury is too trivial to ignore." The literal meaning is the opposite of what the author wants it to be, but this is not irony or sarcasm -- the author is just confused. The extra negations are sometimes explicit negative words (like not and no) and sometimes implicit parts of words with negative meanings (like refute, fail, avoid and ignore). Generally the result has at least two negatives, and often a scalar limit, conditional, hypothetical, or other irrealis construction as well.

In fact, this description is predictive -- if you think of a construction that meets these conditions, and check with Google or Altavista, you will generally find lots of examples whose literal meaning is clearly the opposite of what the writer intended.

The obvious hypothesis is that it's hard for people to calculate the meaning of phrases with several negatives (perhaps especially in combination with things like scalar limits and hypotheticals). The implicit negation in words like fail and ignore may be especially difficult to untangle. This explains why the errors are not detected and corrected: we accept an interpretation that is a priori the plausible one, even though it's incompatible with the sentence as written or spoken, because it's too hard to work out the semantic details.

However, this may not provide an adequate explanation for why the errors are so commonly made in the first place. The pattern is predictive of errors, but it doesn't predict how common the errors will be, either in themselves or by comparision to "correct" interpretations of the same pattern.

In this post, Geoff Pullum mentions the particular case of "fail to miss" used to mean simply "miss." A little internet search shows that this sequence is moderately common (around 2,400 ghits for "fail/failed/failing to miss", or one per 1.8 million pages), and that when it occurs, it is almost always used in the "wrong" meaning:

Miss Goodhandy doesn't fail to miss an opportunity to humiliate Steve, and gives him a few good swats with the jockstrap's thick elastic waistband.

Although his attendance at school was still very poor, Stanley never failed to miss a movie at the local theaters.

Canceling a few flights here and there seems like a good trade-off because the results of failing to miss a real threat are so severe.

This is sure to be a killer tournament, don't fail to miss it!

It seems to me that there are several different psycholinguistic questions here: why do most people not even notice the problem in sentences like this? why do people stick in the extra "fail to" in the first place, given that the sentences mean what their authors intend if they just leave it out? why are uses of "fail to miss" so often accompanied by an additional negative ("doesn't fail to miss", "never failed to miss", etc.)? and why do people hardly ever use "fail to miss" to mean "fail to miss"?

In fact, almost the only internet examples of "correct" usage of fail to miss are copies of this famous passage:

This is what The Hitchhiker’s Guide to the Galaxy has to say on the subject of flying: There is an art, or, rather, a knack to flying. The knack lies in learning how to throw yourself at the ground and miss. Pick a nice day and try it. All it requires is simply the ability to throw yourself forward with all your weight, and the willingness not to mind that it's going to hurt.

That is, it's going to hurt if you fail to miss the ground. Most people fail to miss the ground, and if they are really trying properly, the likelihood is that they will fail to miss it fairly hard. Clearly, it is the second part, the missing, which presents the difficulties.

Douglas Adams offers us a clue here, I think: you can fail to do something only if you first intended to do it. It's relatively rare for people to intend to miss something, but missing things is generally easy to do, so when you try to miss something, you usually succeed (and you might describe what you did as avoiding rather than missing, anyhow). Therefore, failing to miss things just doesn't come up very often. Perhaps this hole in the semantic paradigm leaves a sort of vacuum that bad fail to miss rushes to fill?

We can test this idea with "fail to ignore", because ignoring things is often both desirable and hard to do, and failing to ignore things is therefore an event that we often may want to comment on. There are certainly plenty of "wrong" interpretations of fail to ignore:

The Judge Institute is a building that no-one in Cambridge can fail to ignore. Much has been written about its jelly-baby hues, its pyramid-like proportions, and its metamorphosis from the husk of Old Addenbrooke's. (context)

Progressive thinkers and activists need to consider the practical implications of these principles. Good people of the world cannot fail to ignore them.

In New York, state Sen. Michael Balboni (R-Mineola) is circulating a proposal based on the original California bill, and plans to introduce the measure in the next 10 days. "Various industries in New York are looking at our legislation," said Balboni legislative assistant Tom Condon. "We have to ask, are all these lawsuits beneficial to our economy? And we can't fail to ignore possible negligent conduct from these manufacturers. It's a difficult issue."

but there are plenty of "correct" interpretations as well:

[T]he chapter points out the pitfalls that are likely when making decisions: ignoring opportunity costs, failing to ignore sunk costs, and focusing only on some of the relevant costs.

He managed somehow to answer their questions, trying and failing to ignore the addictive joy of a kindred spirit touching his.

The story of a black lawyer who tried and failed to ignore his race.

And "fail to ignore" is also less common (in both right and wrong interpretations) than "fail to miss" (about 1200 ghits to 2400 -- though the verb miss is also about twice as common as the verb ignore). In any case, the counts are large enough (tens of millions for the basic words such as fail, miss and ignore, and thousands for phrases such as fail to miss, fail to ignore) that one could imagine fitting some simple statistical models for the generation process that would permit testing different answers to some of the questions asked above.

As another example, consider the counts in the table below

to underestimate
to overestimate

Nearly all the "to underestimate" cases are logically mistaken substitutes for "to overestimate":

It is impossible to underestimate the long-term impact of Phoebe Muzzy’s ’74 longstanding role as an Annual Fund volunteer.

It is almost impossible to underestimate the importance of rugby to the South African nation in terms of its self-esteem on the world stage.

It's impossible to underestimate Lucille Ball's importance to the new communications medium.

It's impossible to underestimate the value of early diagnosis of breast cancer. (BBC)

Why are these mistakes so common? Why are correctly-interpreted uses of "impossible/hard/difficult to underestimate" so rare -- except in discussions of the mistaken ones? Is there a connection between these two facts?

Google psycholinguistics may point the way to the answers, despite its obvious and severe practical and theoretical difficulties as a methodology.

[Update: Fernando Pereira observes that "[f]or those of us skiers who spend a considerable time in the trees, the chance of 'failing to miss' is why we wear helmets". However, the single result of searching for |"fail to miss" wilderness ski| failed to produce any other correct uses:

The clay like soil of the Adirondacks makes it difficult for water to run off and creates these mud holes that can cause you to sink up over your knees if you fail to miss a rock or log when crossing.

This may be sampling error, but apparently it's not enough to do something where missing things can be both difficult and desirable. Seriously, I think in this situation people are more likely to use the word avoid. I couldn't find any examples involving skiers and trees, but there are plenty of cases in a slightly generalized frame, e.g.

Low-hangng [sic] branches and limbs can be a problem for boaters who fail to avoid getting caught in them.

Finally, while checking all this out, I found this amusing piece (entitled "Do not fail to avoid neglecting this post") about the difficulties of calculating the polarity of summaries of SCOTUS decisions. ]

Posted by Mark Liberman at February 26, 2004 09:18 AM