Freedom and flexibility

« previous post | next post »

Muriel Spark's memoir Curriculum Vitae antedates discourse-particle like to the early 1920s. And J.L. Austin, in his posthumous work Sense and Sensibilia, defends like as "the great adjuster-word, or, alternatively put, the main flexibility-device by whose aid, in spite of the limited scope of our vocabulary, we can always avoid being left completely speechless."

Muriel Spark describes her reactions, in her "pre-school infancy", to The Doorbell:

That ring at the door that I loved so much would bring, in the afternoon, my mother’s friends or, on rare occasions, my married aunts. In the evening a much more exciting variety of family friends rang the bell, many of them fairly eccentric, in whom I took a deep interest. […]

A pulley on the landing, or doorlifter as we called it, would open the street door for visitors who pulled that wonderful doorbell. In those days before I went to school, people were far more important to me than toys or nature. The beauty of walks over the hills and by the sea was beginning to seep into my consciousness by way of the sensations of smell and of sheer liberty and the lyrical suggestiveness of nature-verse, but it had not yet formed a positive delight in my mind such as people presented.

The magic pulley on the landing would often admit a voice first of all, calling up the stairs, for there was a curve in the staircase and one could not see immediately who the visitor was. Then on stage to us, as it seemed, came one of the following:

One of the regular visitors that she goes on to list was "a Miss Macdonald,"

. . . whose name was Margaret, as I gathered from a piece of conversation she reported. Miss Macdonald was dressed in navy blue with a white blouse. She was finer-bred than Miss Pride, but it was said she was not all there. I think my parents were sorry for her. All the time she spoke tears coursed down her cheek. They trickled down into her cup of tea. She couldn’t ever stop crying. She was bound up in a court case against someone who had wrongly accused her. Her brother, a lawyer, couldn’t do much more than he had already done. The word ‘like’ peppered her conversation. ‘My brother, like, wouldn’t go, like, any further with it, like . . .'

Wikipedia says that Muriel Spark, born in 1918, "was educated at James Gillespie's School for Girls (1923–35)", so Miss Macdonald's visits must have taken place in the early 1920s.

Miss Macdonald is presented as "not all there", and her excessive use of like is a symptom of eccentricity, like her constant crying. But both for her crying and for her use of like, what Muriel Spark considers worthy of note is the frequency and not the function.

And in Sense and Sensibilia, J.L. Austin offers a strong defense of like's function. He describes the

… large and important family of words that we may call adjuster-words — words, that is, by the use of which other words are adjusted to meet the innumerable and unforeseeable demands of the world upon language. The position, considerably oversimplified no doubt, is that at a given time our language contains words that enable us (more or less) to say what we want to say in most situations that (we think) are liable to turn up. But vocabularies are finite; and the variety of possible situations that may confront us is neither finite nor precisely foreseeable. So situations are practically bound to crop up sometimes with which our vocabulary is not already fitted to cope in any tidy, straightforward style.

He chooses pigs and pig-like animals as a case in point:

We have the word 'pig', for instance, and a pretty clear idea which animals, among those that we fairly commonly encounter, are and are not to be so called. But one day we come across a new kind of animal, which looks and behaves very much as pigs do, but not quite as pigs do; it is somehow different. Well, we might just keep silent, not knowing what to say; we don't want to say positively that it is a pig, or that it is not. Or we might, if for instance we expected to want to refer to these new creatures pretty often, invent a quite new word for them. But what we could do, and probably would do first of all, is to say, 'It's like a pig.' ('Like' is the great adjuster-word, or, alternatively put, the main flexibility-device by whose aid, in spite of the limited scope of our vocabulary, we can always avoid being left completely speechless.) And then, having said of this animal that it's like a pig, we may proceed with the remark, 'But it isn't a real pig'—or more specifically, and using a term that naturalists favour, 'not a true pig'. 

He suggests that like has special value in this respect:

But, one might ask, do we have to have 'like' to serve this purpose ? We have, after all, other flexibility-devices. For instance, I might say that animals of this new species are 'piggish'; I might perhaps call them 'quasi-pigs', or describe them (in the style of vendors of peculiar wines) as 'pig-type' creatures. But these devices, excellent no doubt in their way, can't be regarded as substitutes for 'like', for this reason: they equip us simply with new expressions on the same level as, functioning in the same way as, the word 'pig' itself; and thus, though they may perhaps help us out of our immediate difficulty, they themselves may land us in exactly the same kind of difficulty at any time. We have this kind of wine, not real port, but a tolerably close approximation to port, and we call it 'port type'. But then someone produces a new kind of wine, not port exactly, but also not quite the same as what we now call 'port type'. So what are we to say? Is it port-type type ? It would be tedious to have to say so, and besides there would clearly be no future in it. But as it is we can say that it is like porttype wine (and for that matter rather like port, too); and in saying this we don't saddle ourselves with a new word, whose application may itself prove problematic if the vintners spring yet another surprise on us. The word 'like' equips us generally to handle the unforeseen, in a way in which new words invented ad hoc don't, and can't.

All of Austin's examples involve adjusting the reference of noun phrases. But Miss Macdonald used like in a syntactically much freer (and thus more useful) way, which Austin would no doubt have applauded, since this freedom equips us to handle an even wider range of referential surprises. For the same reason, like-users during the past century have increasingly followed her lead.

Muffy Siegel ("Like: The Discourse Particle and Semantics", J. of Semantics 2002) notes that

. . . like can, like other discourse particles, appear […] pretty much before any constituent and have scope over that constituent, indicating the lack of a fixed grammatical role.

As I observed in "Divine ambiguity" (1/4/2004), it's easy to see how more syntactically constrained uses of like could morph into more syntactically protean uses — and I suspect that we'll be able to find examples back into the 19th century and beyond of this process.

As for the modern meaning, Muffy Siegel quotes the definition from Lawrence Schourup's 1982 dissertation "Common discourse particles in English conversation":

like is used to express a possible unspecified minor nonequivalence of what is said and what is meant.

This is very much like what Austin wrote:

If we think of words as being shot like arrows at the world, the function of these adjuster-words is to free us from the disability of being able to shoot only straight ahead; by their use on occasion, such words as 'pig' can be, so to speak, brought into connexion with targets lying slightly off the simple, straightforward line on which they are ordinarily aimed. And in this way we gain, besides flexibility, precision; for if I can say, 'Not a real pig, but like a pig', I don't have to tamper with the meaning of 'pig' itself.

Among the few of you still reading, some are probably thinking "But surely it's a problem that the Miss Macdonalds of the world sprinkle like so vigorously into their speech — literally every third word, in Muriel Spark's example".  And some may recall the example from Muffy Siegel's paper that I cited in a long-ago post, where 4 likes are added to 22 non-like words:

She isn’t, like, really crazy or anything, but her and her, like, five buddies did, like, paint their hair a really fake-looking, like, purple color.

In response, let me suggest Muriel Spark was indulging in memoiristic hyperbole, and Muffy chose to lead with an especially striking example, in which her informant was trying to transmit some ambiguous information in a scrupulous way.

If we look at extended conversation, even Lena Dunham only uses like about once in 81 words:

But what about the clear rate change over time? Are Americans getting less and less willing to shoot their referential arrows straight ahead? It's possible. But rather, I suspect, the adjusting force of like has just been getting gradually bleached out, as commonly happens to such words, so that the threshold of target-deviation for using it has gradually shrunk.

[h/t to Barton Swain, "Managing the Decline of, Like, a Great Language", WSJ 4/19/2015 ("Interdiction has failed. The grammatical version of legalizing marijuana may be in order.") ]


  1. Rodger C said,

    November 28, 2015 @ 12:35 pm

    Bit what is it like to, like, be a, like, bat?

  2. Rodger C said,

    November 28, 2015 @ 12:36 pm


  3. Rubrick said,

    November 28, 2015 @ 3:02 pm

    In light of some recent scholarly reanalysis of punctuation in the First Folio, I've antedated it to Hamlet, Act 3 scene 2:

    HAMLET: Methinks it is, like, a weasel.

  4. Jon W said,

    November 28, 2015 @ 3:18 pm

    Or Exodus 16:31, in the KJV: "And the house of Israel called the name thereof Manna: and it was, like, coriander seed, white; and the taste of it was, like, wafers made with honey."

    [(myl) More divine ambiguity.]

  5. Rubrick said,

    November 28, 2015 @ 4:15 pm

    And while we're there, John 14:4, American King James:

    "And where I go, you know? And the way, you know?"

  6. Michael Watts said,

    November 28, 2015 @ 4:48 pm

    I don't disagree with this analysis of discourse like, but I feel like it doesn't fully explain the striking phenomenon I observed (and mentioned in the comments to the earlier like-post) of someone using like at very high frequency while attempting to speak a foreign language. On the one hand, that context does seem to be particularly likely to feature "possible minor nonequivalence between what is said and what is meant"; on the other hand, my personal instincts say that if you're using it while trying not to speak English, your own mental model of discourse like must not see it as a "word" on the same level as "goat", and possibly not even on the same level as "ouch!"

  7. Charles Antaki said,

    November 28, 2015 @ 5:44 pm

    I'm a bit suspicious of the graph – it might be that the speakers are doing a differently weighted portfolio of things in their conversation. If Lena Dunham was, say, making more assertions that wanted hedging than was Willie Nelson, you might expect more use of "like" as a qualifier. (Leaving aside the irrelevant-for-these-purposes use of like in the non-– discourse marker sense, which I'm sure Mark Liberman has discounted).

    [(myl) This is a good point — but of course it's the subjective opinion of the speaker that determines whether a given assertions "wants hedging", so it's not easy to see how to investigate the matter. But the basic transcripts are on line — Willie Nelson here and Lena Dunham here — so you can look into it.]

  8. peter said,

    November 28, 2015 @ 11:06 pm

    Is it possible that a widespread linguistic phenomenon began, six or seven decades before, as the idiolect of a single, eccentric speaker?

  9. Ken Miner said,

    November 28, 2015 @ 11:32 pm

    but what is it like to, like, be a, like, bat?

    Hypothesis: 'like' cannot interrupt a noun phrase. Has to be "what is it like to, like, be a bat?". If it's going to be in the language, we may as well posit some rules for it.

  10. Bob Ladd said,

    November 29, 2015 @ 5:13 am

    I remember noticing the vernacular usage parodied in the excerpt from Muriel Spark when I first moved to Scotland (mid 1980s). Somehow it seems a little different from the 1950s American beat-generation usage (which Mark cited in another recent post and which he and I are both old enough to remember), which I take to be the source of the modern usage.

    Here's a hunch about what the difference is, though it would hard to prove even with Big Data: to me, the Scottish vernacular version seems to put like AFTER a constituent, while the modern (American?) version puts it BEFORE a constituent. (This analysis of the American version is consistent with what Muffy Siegel said in her 2002 paper.) In the quote from Muriel Spark, the first like feels to me like a hesitation after My brother, and the final one is literally at the end of the excerpt from the utterance attributed to Miss Macdonald. Spark is a sensitive enough writer (and a caustic observer of subtle class-related behavioural differences) that if Miss Macdonald had been hedging something that followed that last like, she would have included it in the quote.

    I suppose the easiest way to operationalise this hypothesis in a Big Data trawl would be to look for the prevalence of turn-final like in vernacular Scottish speakers and people like Lena Dunham.

    [(myl) Interesting idea. Four reactions:
    (1) I think that "trailing like" does exist in American usage, though it's rarer than "specifier like". I'll try to find some examples.
    (2) Could it be that discourse-particle like, like phrase-final rises, also came to America with Scots-Irish immigrants? Neal Cassady, after all…
    (3) Looking for corpus evidence would probably work given phrasing information, or maybe just timing information.
    (4) Does "trailing like" happen in the Map Task corpus? I don't remember noticing it, but I might just have missed it.

  11. Graeme said,

    November 29, 2015 @ 6:55 am

    "She isn’t, like, really crazy or anything"

    It is driving me, like, crazy. But 'like' as a fuzzying device in a positive statement ('like, her five buddies') seems straightforward. In a way it may not be in a negative statement, of the sort I've quoted above.

    I presume what's meant is 'she is somewhat crazy'. Yet 'she isn't really crazy or anything' would do that work. Does the 'like' here strengthen the level of craziness attributed – a politeness gesture? Or does it signal the speaker is only inventing mental illness metaphorically?

    [(myl) People often express themselves redundantly. Consider all the lawyers' idioms like "goods and chattels", "indemnify and hold harmless", "books and records", "costs and expenses", …]

  12. Meg Wilson said,

    November 29, 2015 @ 12:18 pm

    We should be at least a little skeptical about Spark's account, written ~70 years after the fact. (If she drew on a diary entry or some such, that would put it closer to the event, but still presumably years after.) She could be conflating multiple memories of multiple individuals, or inserting a different particle than the person actually used.

  13. Bob Ladd said,

    November 29, 2015 @ 12:42 pm

    Reply to two of MYL's four points:

    (2) It's possible that like was imported to North America along with final rises, but it's worth pointing out that when I "hear" the Scottish trailing like in my head, it seems to have mostly falling – again, phrase-final? – pitch.

    (4) I haven't spent enough time with any but specific excerpts from the HCRC Map Task corpus to know, but I think it's at least possible that the speakers (almost all Glasgow University undergraduates) were too middle-class to use trailing like very much. But it's an empirical question.

    [(myl) I didn't mean that trailing like actually carried the final rises, just that the two features of discourse-particle like and rising declaratives might both have been influenced historically by Scots-Irish varieties of English.

    Also, Muriel Spark's memoir describes Miss Macdonald as "finer-bred than Miss Pride", so for her, trailing like doesn't seem to have been beneath the station of Glasgow University undergrads.]

  14. Terry Hunt said,

    November 30, 2015 @ 11:38 am

    The reported speech of Muriel Spark's acquaintance Miss Macdonald:

    ‘My brother, like, wouldn’t go, like, any further with it, like . . .

    uses 'like' no more densely than does a current friend of mine's. However, he (aged around 30 and a native of Portsmouth, Hants, if it's relevant) also interjects 'well . . .', 'um . . .', y'know' and similar terms so prolifically that together, such matter can form well over half of his conversation.

    Regarding your own speculation that:
    'like and rising declaratives might both have been influenced historically by Scots-Irish varieties of English.

    my own purely subjective feeling as a (mostly southern) Brit rising 60 is that both features seem most familiar as a feature of the working-class Liverpool (i.e. Scouse) register/accent – old episodes of The Liver Birds may offer corroboration: though scripted, I believe the writer was also Liverpudlian and was aiming for authenticity. Scouse is of course strongly (Northern and Southern) Irish- as well as Welsh-influenced, as you will know better than me.

  15. Lesley Graham said,

    November 30, 2015 @ 3:26 pm

    My children do exactly the same thing in French: "Nous étions genre à la cantine et il a fait genre comme s'il allait nous parler mais à la fin il est allé genre voir les autres filles."

    Re. Muriel Spark's example, there is an entry in the Scottish National Dictionary that documents this sort of use in general Scots:
    "II. adv. 1. (1) Used parenthetically with a modifying, depreciatory or apologetic force = so to speak, as it were, to let you understand, to be more precise, like being inserted either before or after the words so qualified. Gen.Sc. Only dial. or colloq. in Eng."
    The earliest quotation is from 1815.

    [(myl) Wonderful! Here's the entry — and along with the examples of "trailing like" we find:

    Sc. 1827  C. I. Johnstone Eliz. de Bruce III. x.: “If I could get like a down-bed, or a claught o' a silver tea-pot,” thought Mrs Haliburton, whose sensible mind ran more on plunder than fire.

    Sc. 1885  Stevenson Letters to Baxter (1956) 157: Tak off what I'll like be owin' you.

    Sc. 1846 Edb. Tales (Johnstone) II. 85:
    But I am wae to hear ye need country quarters for health, sir. What is like the matter?

    Sc. 1756  M. Calderwood Journey (M.C.) 213: The room-rents are cheaper than at Moffat, like about seven shillings a week for the bedrooms, and less for the smaller rooms.

    Sc. 1886  Stevenson Kidnapped xxix.: What's like wrong with ye? … Just tell me what like ye'll be wanting.

    It seems more and more plausible to me that the 20th-century American usage really did come from Scots-Irish immigrants.]

  16. J. W. Brewer said,

    November 30, 2015 @ 5:38 pm

    While I myself have a decided preference for "Scotch-Irish" over "Scots-Irish," at least when talking about the relevant subset of my own ancestors, I am rather pleased to learn from the google books n-gram viewer that the two variants have remained comparatively close to each other in frequency for the last two centuries, sometimes swapping the lead but without either driving the other out of the market.

    Scotch-Irish settlers were more heavily concentrated in some parts of the U.S. than others (although then their eventual descendants spread out further). Is there any learning among dialectologists of parts of the U.S. where lexical items or syntantic idiosyncracies believed to be of Scotch-Irish origin cluster? If so, does it match up at all with whatever regional pattern may have been spotted in the early AmEng evolution of the relevant senses of "like"?

    (One internet source FWIW says that Neal Cassady's paternal-line ancestors were Irish Quakers who came to New Jersey in the early 18th century, which would put them outside what I would think of as the "mainstream" of Scotch-Irish immigration w/o making them mainstream Irish-Americans either – Neal's youthful stint as an altar boy before moving on to stealing cars etc. was supposedly the result of the influence of his Catholic mother, whose maiden name suggests German ethnicity. In any event his family background was so unstable it seems unlikely to have had much influence on his idiolect and instead you would want to know about dialect features in the juvenile-delinquent demimonde of Denver in the late 1930's.)

  17. Gunnar H said,

    November 30, 2015 @ 6:05 pm

    @Lesley Graham:

    You get the same in Swedish with the word typ ("type", equivalent to French genre):

    I suppose it is natural to draft words with this sense into use as vagueifying fillers: English uses "kind of" and "sort of" in a similar way. I wonder if they developed independently in each language or are calques of each other.

    The comparable filler in Norwegian is liksom (lit. "just as", "like as"). This word seems different to me in certain ways, but I can't, like, articulate them.

  18. Rod Johnson said,

    November 30, 2015 @ 8:18 pm

    Contra Austin, does like necessarily express nonequivalence?

    "Mishkin wondered what a spaceship looked like. What could you compare a spaceship to? Itself? 'The spaceship looked utterly like itself.' Jane shook her head. Mishkin's father shook his head. Mishkin tried to play the flute. His skin itched. He wished he could think of something a spaceship looked like. Not itself. He decided to buy a toy spaceship and describe that." (From Options by Robert Sheckley, somewhat sketchy link here.)

  19. Ran Ari-Gur said,

    December 2, 2015 @ 2:04 am

    @Rod Johnson: Or Antony's description of a crocodile in Antony and Cleopatra II.7 ("It is shaped, sir, like itself, and […]"). I think that the semantics of like are such that it's technically correct to describe something as being like itself (either in general, or in any specific respect) — otherwise it's hard to explain why "it is like nothing else" expresses uniqueness while "it is like nothing" does not — but it's rare for anyone to actually describe something as being like itself, because there's almost no situation where it's cooperative to do so.

  20. Terry Collmann said,

    December 3, 2015 @ 4:04 am

    Trailing 'like' seems VERY Irish to me: Muriel Spark's Miss Macdonald sounds just like my Dublin nieces and nephews.

  21. Rod Johnson said,

    December 4, 2015 @ 12:10 am

    @Ran Ari-Gur: I concur, and thanks for the Shakespeare reference, which I had totally forgotten.

  22. Chris said,

    December 4, 2015 @ 5:01 pm

    I must admit, I keep tripping over the reference to _Sense and Sensibilia_ by J.L. Austin — as opposed, of course, to _Sense and Sensibility_ by Jane Austen.

    There must be a relationship but I have no idea what.

  23. Charles Antaki said,

    December 5, 2015 @ 5:01 am

    I had a look at the transcripts via links that Mark Liberman kindly made available, with the idea of checking to see if a heavy “like” user simply had more opportunities (or requirements) to use a hedge (if that’s what “like” is) than a light one.

    But as Mark foresaw, identifying occasions where hedging is allowable (let alone required, an odd possibility anyway) is hard, and probably not sensible. Pretty well anything declarative or interrogative that one can say (or, anyway, that Dunham and Nelson say in their interviews) might have the odd “like” inserted somewhere. So without a more careful analysis of the data, the graph (at the top of the entry) probably stands. Whether the operative variable really is age of speaker or some other is another matter, though.

RSS feed for comments on this post