What does 'literally' literally mean?

Glen Newey

As the story goes, an Oxford philosopher is asked by his wife to watch the soup while she’s out of the room. She returns to find him staring fixedly at the broth, as it bubbles all over the hob. The joke is on the prof, at once other-worldly and too literal-minded to follow a simple instruction (as Bernard Williams remarked, what would he have done if his wife had asked him to keep an eye on the soup?). But what does it say about the word literally, if a simple task like soup-watching can be cocked up by taking it too literally?

Literal-mindedness in analytical philosophers can sometimes amount to a vocation. In my current Philo-101 class I tell my Belgian students about conversational implicatures, and how they differ from entailments. I recount the tale of Nobby, who tells his mum that he stayed at Olga’s till 10 p.m., when in fact he stayed all night; for Nobby has read his Paul Grice, and knows that while ‘I stayed at Olga’s till ten’ conversationally implies that he left at ten, it doesn’t entail it. For if he stayed all night, then he stayed at least till ten, but it can’t follow from that that Nobby left Olga’s at ten, since that makes the claim that he stayed all night false.

So far so Gricean. I tell students that implicature is a form of philosophical purgatory to be gone through before the harps can twang. But someone – Nobby’s mum, for example – might say: ‘Cut the crap.’ Some of my francophone students do say this ('coupez la merde'), in so many words, and side with his mum against Nobby, resting their case on their greatly greater knowledge than mine of literal meanings in French. If someone’s literally keeping his eye on the soup, is he really dunking his eyeball in the bouillabaisse? Or can you literally keep your eye on something just by watching it? Are Magritte’s pipe pictures banal, because obviously (another nag from the same stable) a painting isn’t literally a pipe – for instance, you can’t stuff it?

People are upset that new lexicons, shaped by web usage, have allowed that fine old English word ‘literally’ to be debased. These complaints have a venerable lineage. They worry that ‘literally’ no longer means only ‘literally’ but signals hyperbole, as in: ‘The restroom attendant was like, literally, the size of a whale.’ We infer the guy was fat, or big. Then the L-word lies to hand, for the smartass reply: ‘Really, literally? I mean, how did he get into the restroom?’ But what does ‘literally’ literally mean? In the just-quoted sentence, you might think the ‘like’ in fact says it were as if he were that big, in which case it’s presumably figurative. Anyway, some whales aren’t that big – baby ones, say. ‘Ah, but he didn’t say a baby whale. He meant one that was colossal.’ One that was colossal in relation to its own kind? If so, maybe that’s what he literally meant about the restroom guy.

It turns out to be pretty awkward to specify how the literal and its implied antonym – the metaphorical, usually, or figurative – line up against each other. In 'On Truth and Lies in a Non-Moral Sense', Nietzsche glosses the idea of truth via metaphors: as a coin whose face has been worn away through use, or as ‘a mobile army of metaphors’. People talk of literal translations. But it’s not obvious what this is. Is it word for word? Luther reportedly said at the Diet of Worms: ‘Hier stehe ich, ich kann nicht anders.’ Would a literal rendering be ‘Here stand I, I can not otherwise’? But this isn’t even English: it’s Googlish. And ‘I can not’ doesn’t mean what ‘I cannot’ means, which is what he’s usually taken to have meant.

As Humpty says to Alice in wonderland, deciding on meanings means deciding who’s boss. Surely, for instance, the meaning’s plain when the government sends out signboard vans telling foreigners to ‘go home’ – or, speaking metaphorically, to fuck off (‘You literally want me to go away and copulate at the same time?’). We all know what is meant – for instance, by ‘home’. Except for those of us who, literally, don’t speak English. And those who, because they left poverty, or forced marriage, or other kinds of servitude, or came to join loved ones, had no home there, but came to find one here. Literally.


  • 15 August 2013 at 12:54pm
    semitone says:
    There was an infamous exam for the Economic Geography course at the Australian National University where the questions stayed the same every year, but the answers changed. Back in the days when Australia had seven banks, "Does Australia have five banks?" was quite a tricky question, and impossible to revise for.

  • 15 August 2013 at 10:21pm
    Alex Abramovich says:
    Yes, but baby whales are pretty fucking big.

    • 17 August 2013 at 6:26pm
      Baby blues are 25 feet long!

  • 16 August 2013 at 9:51am
    Jake Bharier says:
    Am I expected to interpret this literally? I assume a baby whale is pretty in the eyes of its parent. But what about the rest of the sentence?

  • 16 August 2013 at 11:01am
    aisia says:
    Er, what crap is supposed to be being cut here?

    Is it the claim that it is not part of the literal content of 'I stayed till ten' that Nobby left at, or immediately after, ten? Perhaps this claim isn't as assured as the Gricean believes, but it's hard to see that it's just crap. Is the phrase 'only until' strictly pleonastic? If Noddy was murdered a ten, and a lawyer asks Nobby whether he was with Olga until ten, is he lying if he says yes? If so, it's not obvious. And if it's not obvious, it seems rather intemperate to tell the Gricean to cut his crap.

    Or is the crap Nobby's utterance 'I stayed until ten', made after he had stayed all night. Of course Nobby's mum, or anyone else, is going to consider that crap, but they can cite Grice while they do so, i.e. 'cut the crap that is your misleading conversational implicature'.

    Perhaps the the crap to be cut is the whole business of making distinctions so fine as those between entailments and implicatures. But again, why is that business crap? Sure, there may be many people to whom it is of little interest. But it's all in the service of the wider question of what 'literal meaning' really amounts to, a question many people, including myself, Friedrich Nietzsche, and apparently Glen Newey, who after all took the trouble to write this blog post devoted to it, find rather interesting.

  • 17 August 2013 at 9:19pm
    Glen Newey says:
    Interesting remarks.To take the final comment first: I wasn't trying to suggest that making linguistic distinctions is crap. Like aisia, I find language endlessly interesting. A large part of my working life involves doing just that, or doing things that rely on it.

    As regards the respect in which Nobby's mum justifiably says 'Cut the crap', I think it's what aisia indicates in paragraph two. One consequence of the Gricean insistence on strict entailment is that those apprised of the distinctions - lawyers and politicians prominent among them – can hide behind formulations that are sophistical so as to be able to rebut charges of mendacity. This trades short-term gain for long-run depreciation in the currency of political debate, and you find politicians saying things like, 'It all depends on what the meaning of "is" is.'

    I don't think 'only until' is strictly pleonastic, for the reason given in the post itself, namely that 'until m' is consistent with 'until n' for n>m. Lying seems to me to raise further questions (if you want my extended views on the speech-act of lying, see my 'Political Lying: a Defense', Public Affairs Quarterly 1997). Would Nobby would be lying in answering 'Yes' to the question whether he stayed at Olga's till ten? No; but I reject the approach, common among analytical philosophers, of defining a set of conditions on assertoric speech-acts, and then deriving the wrongness of lying from that account – basically because it seems to me that deceptive intent is what matters morally, and clearly that intent can be present despite the absence of mendacity, as in the Nobby example, which (as often with politicians) is concocted precisely to dodge that charge. But a lot of moralists in the casuistical and homiletic literature on lying, Augustine and Kant among them, have thought otherwise.

    One final point. The post really takes aim at the idea that there is some ultimate semantic bedrock where 'literal' meaning (couched say in terms of truth-conditions) lies. That ignores the multifarious jobs that language does, which include, but aren't exhausted by, contractual legalese or computer programs. Try having a conversation with someone who insists on taking your every utterance au pied de la lettre. Philosophers of course, particularly the anals, love tripping people up in this way, and it can be fun.

    But – one of St Paul's better lines – the letter killeth, too. Or it can. A good example is Colin McGinn's recent departure from U Miami, which resulted from his repeated use in e-mails of the term 'handjob' supposedly to mean (literally?) 'a job done by hand' – giving him ludic deniability against accusations of sexual harassment made by the female graduate student to whom they were addressed. Here, as with eyeball-dunking in the 'keep one's eye on the soup' example, we're entitled to ask whether these forms of literalism are what the locutions in questions literally mean.

    • 20 August 2013 at 9:56am
      Phil Edwards says: @ Glen Newey
      McGinn is plagued by people wanting to read his words too literally, or not literally enough. He recently complained on his blog of being quoted in an interview as saying "I'm the most enlightened person in the world." He doesn't deny having said, in the interview, "I'm the most enlightened person in the world"; however, he objects to being quoted as saying "I'm the most enlightened person in the world", as this gives the impression that he's the kind of person who would say "I'm the most enlightened person in the world."

      Monty Python's retort to Oscar Wilde comes to mind - "You did, Colin, you did."

  • 20 August 2013 at 9:07am
    aisia says:
    Thanks for the reply. I'm with you through the first three paragraphs. But I rather feel that, through both post and comment, two distinct issues have been run together: the first issue of what literal meaning amounts to, and a second issue of whether and when literal meanings really matter. You seem to be saying that language is so messy that, contrary to expectations, the literal meaning of 'keep your eye on the soup' is, or could fairly be treated as, what the philosopher's wife originally intended by the phrase, rather than how her hapless husband interpreted it.
    My response would just be to insist that of course the literal meaning of 'keep your eye on the soup' involves unpleasant eye-dunking. Literal meanings can get problematic, but I think the best way to deal with that is not to expand our conception of literal meaning but diminish the importance we attribute to it. Let us press politicians (and sleazy philosophers) on their implicatures. As for clear figures of speech like 'keep your eye on the soup', I would be happy to grant that the content - i.e., what governs the truth conditions - of the phrase is something other than its literal meaning. Don't fetishise literal meaning, I think, is the important lesson to be drawn from the considerations you adduce, but we can stop fetishing it without abandoning out intuitions about what literally means what.

    • 20 August 2013 at 9:33pm
      Glen Newey says: @ aisia
      It seems to me the literal meaning in the soup case is that in terms of which the truth- (or fulfilment-) conditions are to be grasped. Had the woman said to him 'Keep your eye on the soup' and he had gone about the task with attentiveness, but failed to immerse his eyeball, I do not think she would have had a complaint ('But you didn't stick your eye in it!'). 'To keep an eye on X' literally means something like 'to attend to X for a contextual purpose', and someone who does that has literally done what was asked. Different conceptions of literal meaning do duty in different circumstances. And one of the risks with the sort of faux-boneheaded literalism affected by McGinn is that it risks forsaking the wherewithal to distinguish itself from the real sort.

  • 22 August 2013 at 2:53pm
    Mark Dowson says:
    There is an extensive (literally) discussion of this on Language Log at

  • 22 August 2013 at 10:13pm
    Timothy Rogers says:
    Well, the French-speaking lads and lassies seem to be on the right trail here when they advise the prof. to stop shoveling the dung. Any sentence containing “implicatures” and “entailments” pushes the discussion in this disrespectful direction because, while these terms do have their technical uses in a detailed discussion of language or logic, they come off as professorial posh or just plain old pretentious jargon in the example used. In other words, the ideas conveyed could be equally well conveyed in simpler language (good old “imply” and “infer”, “literal” vs. “figurative”, “intention” vs. “perception”, and the commonplace figurative uses – usually hyperbolic in nature – of “literal(ly)”. When it comes to pulling off a good lie or misrepresentation the asymmetry of imply and infer is interesting: you don’t have to imply a false conclusion in order for your listener to infer one, so the speaker has the upper hand (figuratively, of course) unless he or she is talking to a professional logician. I think that native speakers of English understand all of these matters on the basis of long experience in using their own language, and they might be shocked, even outraged, to discover that they had been engaging in implicature, like the man astonished to learn he’d been speaking prose all his life. However, I’m not sure what these observations entail.

  • 30 August 2013 at 2:16am
    gugone says:
    One of my mother's favorite stories was "the recipe said to put the mixture over high heat and beat it for ten minutes, so I did, but when I came back it was burnt".

Read more