Extreme Illusion of Understanding.

Mark Liberman at the Log posts about Lau, Geipel, Wu, & Keysar, “The extreme illusion of understanding” (Journal of Experimental Psychology: General, 2022), whose abstract reads:

Though speakers and listeners monitor communication success, they systematically overestimate it. We report an extreme illusion of understanding that exists even without shared language. Native Mandarin Chinese speakers overestimated how well native English-speaking Americans understood what they said in Chinese, even when they were informed that the listeners knew no Chinese. These listeners also believed they understood the intentions of the Chinese speakers much more than they actually did. This extreme illusion impacts theories of speech monitoring and may be consequential in real-life, where miscommunication is costly.

Mark says:

In the first phase of the study, 240 native speakers of Mandarin Chinese were paired, and given 12 pragmatically ambiguous phrases […] Both speakers and listeners tended to overestimate the success of the verbal disambiguation […].

In the second phrase of the experiment,

We recruited 120 native English-speaking Americans as listeners. Each American listener was yoked to a Chinese speaker and was presented with an English version of the phrases and meanings. The procedure for the American listeners was identical to that of the Chinese listeners, except that they heard the speakers via audio recordings.

A similar overestimation of understanding persisted:

Next, we report the most surprising finding: the illusion of understanding persists even when the listener doesn’t know the language.


On average, American listeners who did not know Chinese identified the intended meanings 35% of the time, which was better than chance (25%) […] Though American listeners were less accurate than Chinese listeners, […] they still overestimated their success by 30pp, believing that they succeeded 65% of the time […] The Chinese speakers overestimated here as well. While Chinese speakers indicated that the American listeners would understand less (50%) than the Chinese listeners (70%),[…] they still overestimated the American listeners’ understanding by 15pp […].

I’m surprised at how surprised I am that people would think they could understand so much of a language they don’t know; I thought I was more cynical.


  1. David Eddyshaw says

    I wasn’t surprised at all; however, I’m not sure that this is due to superHat levels of cynicism on my part, exactly.

    The authors seem to see this straightforwardly as a Bad Thing, but I’m not so sure. I think that it may reflect a default human presumption of understanding, going with our status as an unusually cooperative species. Occasional misunderstandings in a very abnormal experimental setting don’t make me feel cynical about it.

  2. I think that it may reflect a default human presumption of understanding, going with our status as an unusually cooperative species.

    Oh, sure; it would just be nice to think we could correct better for it given such an obvious cue as “This is Chinese, and I don’t know Chinese.”

  3. David Eddyshaw says

    This is Chinese, and I don’t know Chinese

    Obviously people are just naturally Chomskyan. Or, at least, the sort of people who get recruited into academic psychology experiments are …

    The study should be repeated with Martian instead of Mandarin.

  4. Stu Clayton says

    The study should be repeated with Martian instead of Mandarin.

    Next you’ll be invoking Wittgenstein.

    I’m surprised at how surprised I am that people would think they could understand so much of a language they don’t know; I thought I was more cynical.

    I agree with David E here: “I think that it may reflect a default human presumption of understanding, going with our status as an unusually cooperative species.” Some mornings I’ve dealt with as many six presumptuous persons before lunch. The presumptuousness usually starts with an attempt to palm off their manipulative intentions as cooperative ones.

  5. David Marjanović says

    Judging from “an English version of the phrases and meanings”, it seems people were trying to guess which of a limited number of known things the other person was saying. And indeed they did better than chance (35% instead of 25%), likely because of such things as intonation. How well these work is easy to overestimate, and indeed the participants did.

  6. Could this be another instance of the known fact that people simply are bad at putting numbers on their intuitions, similar to how people are normally bad at estimating probabilities?

  7. David Marjanović says

    Sure, but that alone doesn’t explain why the errors are all in the same direction.

  8. Heh. The other day in a Mongolian chemist’s shop (pharmacy) I encountered an Indian woman trying to buy some medical products. The shop assistant was talking to her in Russian and actually believed the Indian lady could understand what she was saying, even though the Indian lady gave no indication that she knew Russian and was, in fact, speaking English all along.

    I helped interpret for them. The Indian lady bought one product but not the other, which she didn’t feel she understood well enough. She had decided to go to a chemist’s where they could explain it to her.

  9. I definitely recognize the “speaker” side of this phenomenon — it takes a conscious mental effort to remain aware that someone else might not understand a language that I do understand.

    For example, I will sometimes find myself discussing some issue in Chinese with person A, in the presence of person B (who does not speak or understand Chinese), and turn to person B and ask “so what do you think?”. (Oddly enough, I don’t think I would ever pose that question in Chinese — my brain realizes that I can’t actually speak to person B in Chinese, even if it doesn’t realize that the preceding conversation has been meaningless to them.)

  10. I’m really scratching my head over this. The study makes it look like it’s a really common phenomenon, in complete contradiction to common experience. I mean, all the people in The Murders in the Rue Morgue who claimed to recognize languages they didn’t understand, none of them claimed to have figured out what was being said…

    I wonder how many of the participants were monolingual (or at least accustomed to conversing in only one language), and whether people accustomed to hearing and speaking many languages would have had more realistic judgments.

  11. @Y: It’s not hard for me to envision that if a listener knows that they are being told one of four things, they could erroneously conclude that the intonation they heard was only consistent with one of the possibilities.

  12. David Eddyshaw says

    Yes, intonation is probably the key thing: those without particular linguistic training are very likely not to realise how language-dependent intonation really is, especially once you leave Europe (e.g., in Kusaal, as in many other African languages, questions end in a falling intonation.)

  13. questions end in a falling intonation.

    Yes-no questions or Wh-questions? Or both?

  14. David Eddyshaw says

    In Kusaal, both.

    The intonation interacts with the tone system (naturally); the mid toneme, which is always realised as a level tone, is systematically replaced word-finally by the high or low tonemes, which are realised as falling tones before pause; then the final fall is “stretched”, to be greater than in statements or commands.

  15. The intonation interacts with the tone system

    Which is one reason for the proliferation of final particles in Cantonese and Chinese in general.

  16. Brett: I missed this. It was a multiple choice test? That seems like really poor experiment design, telling people that they have a chance of guessing what the other is saying.

  17. questions end in a falling intonation

    We have that in English, too.

    “What did you do that for?” — usually falling, different nuance if rising.

  18. “What did you do that for?” — usually falling, different nuance if rising.

    I disagree with your “usually”. Falling intonation means that’s an exclamation (usually critical/exasperation) — and might equally be followed by an exclamation mark. gloss: you’ve done something stupid, I can see no good reason for it, explain yourself!

    As a question seeking the reason, yes it has rising intonation.

  19. David Eddyshaw says

    It’s still a question, though.

    I think this is a matter of the intonation which marks English focus partly overriding the intonation that marks a question: it still has the rising intonation which marks English questions, but the rise has been shifted to the focus instead of being sentence-final:

    What did you do that for?

    I suspect that CGEL in its wonderfulness actually addresses this somewhere, but I’m currently about a thousand miles away from my own copy and can’t check.

    I think that Kusaal can’t do this particular combination at all: at any rate, the VP focus particle is not permitted within content questions (though it’s fine in yes/no questions.)

  20. PlasticPaddy says

    French is even weirder; rising tone is optional if the sentence must be a question, mandatory if it could be a statement, esp. the case where a statement is turned into a q by using rising tone only.

  21. xoxoxoBruce says

    If the English speaker thinks he has a handle on what is being said, right or wrong, his face, expression, movements like almost imperceptible nodding, will be interpreted by the Chinese speaker and be convinced he’s getting through.

    The baby is smiling, nope just gas.

  22. It’s still a question, though.

    they [the non-Mandarin speakers] heard the [Mandarin] speakers via audio recordings.

    So the fact that the sentence is a question by syntactic structure is utterly irrelevant. The most the hearers could go by is (supra-segmental) intonation.

    If the English speaker thinks he has a handle on what is being said, right or wrong, his face, expression, movements like almost imperceptible nodding, will be interpreted by the Chinese speaker and be convinced he’s getting through.

    The whole point of this experiment was that the hearers couldn’t see the face of the speakers, nor vice versa. (And the survey was about what the hearers interpreted; not whether the speakers interpreted they were “getting through”.)

  23. David Marjanović says

    We have that in English, too.

    “What did you do that for?” — usually falling, different nuance if rising.

    The difference is that, in Russian, every question gets this intonation as far as I’ve noticed.

  24. Of course, Russian has the advantage of the question marker ли.

  25. Does it?

    You can ask a question without ли and you can insert ли into a declarative sentence. ли is a marker of uncertainty, not question. They are not disconnected, question is a request to resolve an uncertinty, after all.

  26. You can ask a question without ли and you can insert ли into a declarative sentence. ли is a marker of uncertainty, not question.

    Sure, but it provides an option English doesn’t have. Не так ли?

  27. Stu Clayton says

    English has up-так for that.

  28. Yes, I think you are right. It always marks question in the main clause and can be a marker of uncertainty in a subordinate clause.

  29. David Eddyshaw says

    Regarding question formation cross-linguistically:

    The Smolensky/Dupoux paper responding to the Evans/Levinson paper on linguistic diversity that Bathrobe linked here


    explains that languages that don’t have wh-movement actually do have wh-movement, but it can only be seen by Chomskyan initiates. (Yes, really – pretty much, anyway.)

    We heathen are just too sinful to perceive it.

    (Basically this comes from the grammatical school whose practitioners will refer to German as a SOV language, casually in passing, as an accepted fact unworthy of even a footnote.)

    E & L’s response is good. I also like the passages where they temperately point out the sheer impossibility that Optimality Theory could conceivably reflect any sort of psychological reality at all.

  30. David Marjanović says

    They cite this paper, which claims to prove that OT is NP-hard. That would tank it; trust me, you cannot solve NP-hard problems in your head. But although it was published in 2006, it’s only been cited 56 times, which is pathetic. The reason seems to be this paper from 2009, which says a different mathematical representation of OT is not NP-hard at all.

  31. Is this the Smolensky/Dupoux paper?

    Chinese questions do not locate wh-expressions in a different superficial position than the corresponding declarative sentence (….) is a counter-example to a wh-movement des-universal but, famously, generative syntax has revealed that Chinese behaves like English with respect to syntactically determined restrictions on possible interpretations of questions; this follows if questions in both languages involve the same dependency between the same two syntactic positions, one of them “fronted.” In English, the fronted position is occupied by the wh- phrase and the other is empty, whereas in Chinese the reverse holds (Huang 1998; Legendre et al. 1998). It is the syntactic relation between these positions, not the superficial location of the wh-phrase, that restricts possible interpretations. Such a hypothesized cog-universal can only be falsified by engaging the full apparatus of the formal theory. It establishes nothing to point to the superficial fact that wh-expressions in Chinese are not fronted.

  32. John Cowan says

    Or, at least, the sort of people who get recruited into academic psychology experiments are …

    Yes, we are all WEIRDos on this bus.

    John Wells told me (in a place Dr. Google, like Satan, cannot find) that AmE intonation was structurally identical to BrE intonation, but that the actual realization was almost (but not quite) completely different. I have my doubts about the first point.

  33. We heathen are just too sinful to perceive it.

    Is it saying that the use of interrogative pronouns is universal in human language; where they go in the sentence is a superficial matter?

  34. John Cowan says

    Yes, and also that there are only two places for them, fronted or in situ.

  35. David Eddyshaw says


    I think what’s gone wrong in the stuff referenced by Smolensky/Dupoux is part of the all too familiar typically Chomskyan thing of taking what are really pragmatic restrictions to be syntactic constraints. Evans and Levinson’s reply further down basically says as much – or at least makes the point that whatever the “restrictions on semantic interpretation” may be that “Chinese” content questions supposedly share with languages with wh-movement, they can be explained much more naturally and straightforwardly than by such generative syntax fantasies.

    In any case, for Smolensky and Dupoux to deploy this sort of thing as a counterargument to the proposition ‘not all languages have wh-movement” tells you a great deal about how astonishingly abstract their notion of grammar is. This kind of intellectual self-abuse is of no value whatsoever to anyone genuinely interested in the actual grammatical structure of a real human language.

  36. @DE, I do not understand how do we make the distinction betwen “syntax” and “pragmatics”:(
    Word order is word order, neither s. nor p.

    The question of placement of interrogative pronouns is a question fo word order. But I do not understand what exactly is meant by “it is a question of word order syntax” or “it is a question of word order pragmatics” (unless we have a theory of how “syntax” and “pragmatics” exist in a speaker’s head). What would be a test that allows to distinguish between these two?

  37. As a linguist friend once told me, the problem with generative grammar is not its answers, but that it is asking the wrong questions.

    The central question it aims to solve is how to represent syntactic relationships by means of a formal framework (akin to formal logic). There is no reason why such a framework should be a good representation of language, just as a representation of tax law by ranked constraints, or a representation of music by binary features will never be useful to anyone, and it’s not a matter of figuring out the right constraints or features.

  38. the problem with generative grammar is not its answers, but that it is asking the wrong questions.

    That’s the impression that I get.

    There is no reason why such a framework should be a good representation of language

    Or from my point of view, why do they think those grotesque gangly trees with their invisible elements and elements skipping from place to place (subject to ‘constraints’) actually represent the reality of language — i.e. ‘universal grammar’ and the unique human ability to language?

  39. drasvi, actually Russian illustrates syntactic/pragmatic distinction in word order quite nicely. As is well known, Russian grammar tolerates any order of words in a simple SVO sentence. So it is syntactically possible (not ungrammatical) to say “Любит Таня Машу”, but it cannot be used for conveying a simple fact, there must be a lot going on to make it meaningful.

  40. The term ‘movement’, sneakily, makes generativists of us. A movement has a source and a destination. When one says that “What did you eat?” is related by ‘movement’ to “*You ate what?” (paralleling “You ate [something]”), one is implicitly admitting a privileged word order in a deep structure of some kind, transformed into another word order for external consumption.

  41. Agree with Y. And the old idea of some element being “understood” is adopted in generative grammar as an “invisible element” or “trace”. Very hard to fault them on that. It’s the final contorted formalism (virtually for the sake of formalism) that I find distasteful.

  42. David Eddyshaw says


    Evans and Levinson reply to Smolensky and Dupoux’s remarks on wh-movement on p482, in R6.8: ‘Subjacency and “invisible Wh-movement.”’ The S&D proposals in fact abstract away from such mundane features as word order altogether: essentially they claim that “Chinese” shows similar semantic limitations to English in content questions, and as these limitations in English have been “explained” by Chomskyans to their own satisfaction by “constraints on movement”, this means that “Chinese” must also have movement, but that it’s not “overt.”

    As E&L say (p474)

    The complex representational structures look undermotivated, and covert processes proliferate where alternative models deftly avoid them

    Well, yes. Just a tad undermotivated …

    The alternative and much more plausible analyses turn on the notion of focus, which is interesting: Kusaal actually has a real constraint against marking focus distinctions in content questions, and we’ve just been discussing how focus intonation in English interacts with interrogative intonation (i.e. real linguistic issues to do with content questions, with actual overt expression, instead of let’s-pretend invisible movement.)

  43. David Eddyshaw says

    Apropos of nothimg, apart from my aversion to Optimality Theory, I’ve discovered two consonant-assimilation rules in Cahill’s account of Konni morphophonology which are sensitive, not to whether the word in which they occur is a noun or a verb (there are actually quite a few precedents for that) but to which noun class a noun belongs. There is nothing whatsoever in the actual segmental input to explain why identical clusters of consonants are treated differently in different noun classes.

    Actually, much as I would like to blame this on Cahill’s having fallen among Optimality Theorists, it looks like what has actually gone wrong is that he has taken the singular citation form of nouns as basic. He’s succeeded at least in showing that this can’t really work (it doesn’t work in Kusaal, either, and it’s pretty easy to come up with other languages where the same is true.) It does mean that a number of his other morphophonological rules are poorly motivated too: they turn out to be needed only to repair the complications caused by picking the wrong form as basic.

    But the real point is that phonological rules are not real, and the only justification for picking one set over another that gives identical results is descriptive clarity. OT models fail on this criterion epically.

    (I don’t mean to diss Cahill. He explains what he’s about very clearly, and provides all the data you need to come up with your own preferred analysis instead. Respect.)

  44. noun class
    Celtic speakers are not in the position to complain:-E

  45. David Eddyshaw says

    Well, we’ve only got two (unlike you Russians); and nouns of different classes don’t have different rules for internal sandhi. But then neither do they in Konni, in reality.

    Konni has five noun classes (which represents a fair bit of merging since Proto-Oti-Volta; even Kusaal retains seven.) Like Kusaal and most of the rest of Western Oti-Volta, though, Konni has lost grammatical agreement by noun class, at least if you analyse it properly. It’s just a matter of noun and adjective morphology now. Konni’s closest relative, Buli, has also reduced the number of noun classes, but it still does the grammatical-gender-based-on-class thing according to the Way of the Ancestors (and the Bantu Cousins.)

  46. David Marjanović says

    Is this the Smolensky/Dupoux paper?

    You forgot to put the URL in.

  47. I did but it appears to have become garbled. Here it is again:


  48. OK, I added that to the earlier comment, so people won’t have to scroll through the whole thread to find it.

  49. DE, are you saying that Konni has marked single nouns, aka singulatives?

    (Apropos of, Paul Newman’s book A History of the Hausa Language: Reconstruction and Pathways to the Present came out about two weeks ago. I imagine you have bought it and read it, or will have done so by tomorrow.)

  50. David Eddyshaw says

    No, the Konni system is more interesting.

    Konni has, like Buli, a system where definite articles are suffixed to the noun and partly fused with them.

    What’s odd about the Konni system is that the singular forms with the postposed article look straightforwardy cognate to the ordinary sg forms seen elsewhere in Oti-Volta. On the other hand, the “indefinite” (and citation) form drops the sg noun class suffix and adds -ŋ in all noun classes except the equivalent of the Kusaal human-only a/ba class, where the indefinite sg ending is zero.

    This seems to be unique in Oti-Volta, and I’m not at all sure what to make of it as yet. The ŋ could represent the ending which turns up in Kusaal as the -m noun class suffix of a noun class which consists of mass nouns and abstract nouns; in Kusaal you can very often transfer a stem to this class to make a corresponding abstract noun, e.g. gbanya’a “lazy person”, gbanya’am “laziness”; Oti-Volta languages often exploit the meanings partly associated with noun classes to do this kind of derivation (e.g. Kʋsaa “Kusaasi person”, Kʋsaal “Kusaasi language”, Kʋsaʋg “Kusaasiland”, where the stems are all identical but the noun classes differ.) So the Konni “indefinite” meaning perhaps developed from an abstract sense. Konni itself has merged the old -m class with an originally distinct class with the class suffix -bu.

    Be that as it may, the indefinite singular form can be predicted from the definite singular, but not vice versa: not even given the noun class membership.

    Didn’t know about the Newman book. Ta!

  51. David Eddyshaw says

    Actual singulatives don’t figure much in Oti-Volta languages; however, it looks as if the class with sg gu/ku and pl di/ti, which is securely reconstructable to the protolanguage, was particularly favoured for nouns where the sg is effectively singulative, like “leaf” and “grass”, but along with a lot of others that don’t fit that pattern. Interestingly, Yom actually splits this noun class in two, with different plural suffixes for the singulative/collective types from the straightforward sg/pl sort. It’s tempting to take this as inherited from the protolanguage, but no other Oti-Volta language does the same AFAIK. On the other hand, Kusaal and other languages have a fair number of plural-only forms in -di with mass meanings, like siind “honey”, zuod “friendship”; although formal plurals from other noun classes also turn up as mass nouns, di-plurals are especially prone to do so. So that would fit. And there seems to be something unstable about -di as a count plural suffix in Western Oti-Volta: “irregular” pairings of sg and pl noun class suffixes often involve nouns with sg -gu unexpectedly taking a pl suffix other than the regular -di.

  52. It seems highly adaptive, that we assume understanding. It keeps us listening and engaging. If we really understood how much of our communication was missing its mark, would we even try?

  53. Stu Clayton says

    It seems highly adaptive, that we assume understanding.

    Do you mean assume that you understand what is being said to you ? Or assume that what you say is being understood by those to whom you are saying it ?

    In interactional situations there is no occasion for such assuming. Here you have immediate feedback, in the form of aggreement, disagreement or evasiveness. This here thread is full of such.

    That’s what keeps people listening and engaged – for a while, anyway. Anyone can get bored and go home to read. Or write a book that will really keep ’em guessin’.

    What you might wonder about in an interactional situation, from time to time, is whether someone is pulling your leg, say, or lying to you. Without a concrete reason to consider whether that’s true, and how to verify it, you’re not “assuming” it’s false. If the idea never enters your head that you might be abducted by aliens at any time, you’re not “assuming” it won’t happen, nor that it won’t.

    Reading The Concept of Mind at an early age was a lucky break.

  54. “Reading The Concept of Mind at an early age was a lucky break.” : never heard of it, but sounds interesting.

  55. Given that human interaction is very recursive (both in the sense “if I say it, she will think that I think that she thinks that…” and in the sense “I say ‘….’, and she answers ‘…’ and I say ‘…’) I can’t guarantee that false assumptions can’t help even for two people.
    E.g. in seduction assuming that everyone wants you does seem to help.

    And then it is much more complicated with groups.

  56. Oh — I meant both, Stu, and specifically in the context of striking up conversation without a common language.

  57. dravsi: stop ranting, it’s creepy.

  58. V: Stop assuming the worst of others and giving them orders; drasvi is not being creepy.

  59. Noted. Apologies, dravsi. I’m sorry, but you sound like a Russian troll; I’m sure you’re not, but you sound like one. I would like to dissuade you of the toxic opinions Russian state media has fed you, but it’s your opinion, in the end.

  60. in seduction assuming that everyone wants you does seem to help

    Thanks for the hint.

  61. I can replace “that everyone wants you” with “your attractivness”. Seduction is a good word. I must be creepy, but I do not think I am ranting.

    It is not necessary a ‘false’ assumption, but I think assuming that someone likes you when this someone is not sure (or likes you but is not going to do anything about that) increaces your chances, and assuming that someone does not like you when this someone is desperately in love with you decreaces your chances. Also many people say they find confidence per se attractive (some say they find shyness attractive).

    Same must be true for many other collaborative activities…

  62. Ах, обмануть меня не трудно!..
    Я сам обманываться рад!

  63. David Eddyshaw says

    Paul Newman’s book A History of the Hausa Language

    Courtesy of Y’s alert, I have now read this …

    It’s pretty good, but a bit disappointing, not so much because of anything wrong with Newman (the blurb on the back describes him as “the world’s foremost authority on the Hausa language”, which strikes me as a bit over the top for a jacket blurb, but is, on the other hand, true …)

    It’s just that it’s almost all internal reconstruction, which is inevitable given the fairly primitive state of actual rigorous comparative work even for fairly low-level parts of Chadic, and the absence of written sources dating back long enough for substantial changes in Hausa itself to have been documented. Nothing wrong with that, either, but it’s mostly fairly familiar stuff from other sources. Nice to have a lot of it together in one place, though.

    I was narked that he refers to Greenberg’s classic demonstration of how it came about that Hausa nouns almost all end in long vowels but does not actually quote from it, explain the reasoning behind it, or cite the paper in the references. It struck me as a surprising omission given the importance of the paper.

    In general, the book is what the SF people call a “fix-up.”


    Again, not unusual for the genre.

  64. I’d never heard of a “fix-up” before. I’ve recently read some of Chandler’s short stories which were later united into novels, but I didn’t know there was a name for it.

  65. John Cowan says

    Le Guin used the somewhat more elegant term story suite for fixups without glue: several of her collections are like this.

  66. Y: a fix-up is a book composed of several short stories or novellas, implied to be in the same setting.

  67. @V: I think that quite a bit more is required than that, actually. A fix-up is, in its final form, a novel, and the constituents need to reflect that. Not only do they need to share a setting, but they need to tell a common story. So I would not even consider, for example, The Dying Earth by Jack Vance to be a fix-up, but rather a story collection. There is a common setting, and there are some common plot threads, but each of the seven sections has a different main character, and the continuing plot disappears entirely after the fourth story. The initial four could be considered a short fix-up of their own, but they cover less than half the book. A more conventional fix-up is the second book in the series, The Eyes of the Overworld. It collects a number of previously published adventures of its villain protagonist, Cugel the Clever, which are edited to fit together better, as well as a small amount of completely new material.

  68. Alexia P. Keaton says

    Stu Clayton:
    Do you mean assume that you understand what is being said to you ? Or assume that what you say is being understood by those to whom you are saying it ?

    In interactional situations there is no occasion for such assuming. Here you have immediate feedback, in the form of aggreement, disagreement or evasiveness.

    Do you truly “have” the other person’s feedback, agreement, disagreement, and evasiveness?
    Or do you possess yet another set of assumptions of your own understanding of the other’s (dis)agreement?

    Why would your interpretation of feedback be immune to the same potential assumptions in your interpretation of the content, as presented by the hypothesis?

    Have you never had the experience of sitting in a movie theater with a friend/date when some part of the movie causes you to begin to cry, your companion feels your thoracic shaking or quickened shallow breaths, and reaches their hand over to hold yours to provide emotional support… and you realize they’ve misread your physical feedback as an expression of distress/sadness, while in fact you are crying because the events or ideas on the screen were so beautiful it made you cry out of release and awe and happiness?

    In the dark cave of the theater, feedback is incomplete. Take it a step further, and perhaps even your understanding of your companion’s misunderstanding is also a misunderstanding. Perhaps it is their feedback of reaching for your hand which was actually signifying, “Oh good, I’m glad you also find this scene exultantly beautiful. I don’t express it via crying, but I wish to hold your hand and press flesh to seal our simultaneously affected Affects”. Or maybe even as direct as, “I’ve been sitting here the whole time wanting to have an excuse to reach out and touch you because I’m attracted to you, and now there’s a plausible entry point to do what I wanted anyway”.

    When do we ever exit the theatre and have perfect light with which to perceive/translate each other?

    We are always selecting among interpretations, and our selection process is filled with prior expectations and habits. Whether it happens with the first exchange or after several, eventually something internal to us motivates us to adopt one of the possible summations as correct, and thereafter we are further motivated to reinforce this sense of correctness. Hence the comments about this as a possible adaptation. A creature which constantly checked and re-checked for feedback verification at all times will quickly be outcompeted and displaced by organisms whose adaptations include “the subtle art of not giving a f**k”. Opportunistic misunderstandings of the degree to which two people are in sync, are the mechanisms by which many of us were conceived.

  69. It occurred to me to look and see whether the OED had the specific sense of fix-up discussed above. However, the entries for fix seem to be serious need of an overhaul, and the lack of the publishing sense is just one of the issues they have.

    It seems to be universally agreed that the term was coined by Canadian science fiction author A. E. van Vogt, although it appears to be less clear exactly when he started using the term. Interestingly, it is also unclear at what point in the 1940s van Vogt started writing stories that he intended to be concatenated into fix-up novels. His first science fiction story, “Black Destroyer”* was pretty clearly intended as a standalone narrative, but he quickly started writing sequels to it. However, the original versions of some of those short stories differed quite a bit from the fix-up form in which they are better known today. The Voyage of the Space Beagle, normally considered van Vogt’s best book, contains several completely new chapters, and—even more remarkably—a new main character, Elliot Grosvenor, who takes over the heroic role mostly played by the mission commander Morton in the first two original short stories. The constituent stories that made up The Voyage of the Space Beagle were published in 1939, 1939, 1943, and 1950; only in the last, “War of Nerves,” which was published only a few months before the book version, does Grosvenor play the full role that he does in the revised novel.

    Between 1939 and 1950, van Vogt published some other novels—a not-necessarily-well=defined mixture of unitary but originally serialized works (Slan, his first full-length book, and The World of Ā seem to me to have this character) and fix-ups from stories that were originally written separately. Teasing out how and with what plans for ultimate publication The Weapon Makers and The Weapon Shops of Isher were written could be a complicated undertaking. After several short stories, he the author seems to have decided at some point that he was actually authoring a serialized novel that was a sequel to the first few stories, although exactly when he came to that realization remains fairly murky. Moreover the earlier stories were later published as part of a fix-up of their own.** And some of van Vogt’s original short story versions continued to be published in short story anthologies even after the novel versions—including, rather strangely it seems to me, an anthology of just van Vogt’s stories, M33 in Andromeda, which contains the title story and “Discord in Scarlet” from The Voyage of the Space Beagle in (I believe) their original 1940s versions.

    * “Black Destroyer” (1939) is sometimes identified as the first story in the Golden Age of Science Fiction. Although that is obviously an arbitrary milepost, I find it interesting that van Vogt’s very first contribution to the genre could be so influential.

    ** These two sentences also seemingly describe Moorcock’s development of the first seven installments in the Elric Saga, up through “Doomed Lord’s Passing.”

  70. As I was editing a bad link in the above comment, it occurred to me that it might conceivably be interesting to mention something I do when I edit most comments that include links. For external links (not to earlier Languagehat posts), the software automatically adds the

    rel=”nofollow ugc”

    attributes to my links. Their meanings are explained pretty well at this search engine optimization how-to site; nofollow tells Google not to count the link at all in the PageRank algorithm, while ugc is a lower-level warning that the link is “user-generated content.” I know that my links are not spam (although they are user-generated content), so I don’t think the attributes are necessary. However, it occurred to me that, since I am allowed to edit those tags out of the final displayed HTML, some very well-written spambots could presumably do likewise at sites like Languagehat that use that comment-editing add-on.

  71. David Marjanović says

    It was added back in, though.

  72. Brett: huh, I didn’t know van Vogt was Canadian. That’s about the time Science fiction went en vogue in Bulgaria also (’30s), although famous writers in other genres had dabbled in it before that also — but they were not fix-ups.

  73. I’m sorry if my comments have not been the most coherent in the last eight months, but my chronic pain has left me with an inability to sleep.

  74. @David Marjanović: Ah, so the system is smarter than I am! You’re right, of course. If I edit the ‘rel=”nofollow ugc”’ out of a comment, it doesn’t reappear if I go to edit the same comment again; however, they do nonetheless still appear in the actual HTML page source.

    @V: There’s nothing distinctively Canadian about most of van Vogt’s writing, and he lived more than half his life in California (where he got involved with L. Ron Hubbard—during the pre-Scientology, Dianetics-based phase in the 1950s).

  75. an inability to sleep

    I know how that feels. (I guess a lot of us have experienced that to some extent over the past couple of years.)

    I also chime (I suspect) with your reaction to (how do I put this diplomatically?) one particular correspondent’s rather inarticulately-expressed opinions — that I probably would react against more strongly if I were more sure what they were expressing. I’m hampered because I’m not fluent in the language they’re clearly thinking in. (I also was rebuked by our host some time ago.)

    You might reflect it’s more important for you to guard your sleep patterns than write anything here/on any social media. And/or if what you have to say is worth saying, it’ll still be of value after you write it privately and re-read it critically a couple of hours later, before posting publicly.

    Posterity will not take your silence as implicating you by association with whatever howlers (in your view) appear here. I rather suspect posterity just won’t care either way.

    So I’ve trained myself to restrict my expostulations to be purely verbal — as in yelling at the screen.

  76. I also chime (I suspect) with your reaction to (how do I put this diplomatically?) one particular correspondent’s rather inarticulately-expressed opinions — that I probably would react against more strongly if I were more sure what they were expressing.

    I’m pretty sure you’re wrong, and I’m not quite sure why some people are reading his comments so uncharitably unless it’s a simple matter of Russophobia, which I hope is not the case.

  77. unless it’s … Russophobia

    No not -phobia. I of course haven’t turned against the Russian people: they’re as much caught in the middle of Putin’s powerplays. (It might be I plain don’t understand the Russian mindset, not having the language.) And I don’t specifically mean the comments wrt the current conflict.

    Anyhoo, I don’t find I learn anything or am stimulated to any novel ideas. My advice was more along the lines: it’s not worth losing sleep over.

  78. I don’t find I learn anything – sorry for this.

  79. PlasticPaddy says

    Please do not give up. Your comments may be misinterpreted or understood as the opposite of what you are trying to say (or you are trying to make a very subtle point or ask for more evidence for a judgment that you question, or to point out facts that you think are relevant in a way that may not be obvious to less original thinkers). It is like the joke about the Russian woman whose post-coital statement is “Ivan, my soul you will never understand!” 😊

  80. David Eddyshaw says


    drasvi’s English is some orders of magnitude better than my Russian …

    (Nor would it have occurred to me that the English word “neighbour” doesn’t always map neatly into other languages; in hindsight, not surprising, but it’s a neat illustration of the interaction between language and culture. Who is my neighbour, indeed?)

  81. David Marjanović says

    Liebe deinen Nächsten wie dich selbst. Nominalized “next”, which does not occur in any other context. Someone prominent in Austria’s xenophobic party actually tried to use it as a justification for xenophobia a year ago or so.

    I’m not sure why Nachbarn wasn’t chosen, but as things stand today it would carry the weird implication that you only have one.

  82. Lars Mathiesen says

    Elsk din næste som dig selv. Danish does exactly the same thing, nabo would not mean the same in that context.

  83. @Hat I’m pretty sure …

    Then I have no idea what it is you’re sure about. have you even been reading [your own] comments?

    @DE drasvi’s English is some orders of magnitude better than my Russian …

    Not evidence for anything.

    Peter Ustinov — who could be witty and apposite in at least six languages (attrib, and I’m paraphrasing): I’ve met professional translators who speak a dozen languages fluently: they can say nothing in any of them.

  84. Then I have no idea what it is you’re sure about. have you even been reading [your own] comments?

    Yes, I didn’t understand what he was saying. Then he explained. I, unlike you, give people the benefit of the doubt.

  85. @drasvi sorry for this

    You’ve no need to apologise. I own my reactions; and how I handle them: I’m now in the habit of skipping over your postings.

    Perhaps your postings would be more informative and cogent if you restricted yourself to Russian?

    My counsel and concern was for @V’s well-being.

  86. AntC: I haven’t read upthread, but I’m okay. My brother’s family is visiting for Easter, and my little nephew is a delight. He demands that I read fairy tales to him, and is fascinated by the operation of the coffee machine — I think he likes the rumbling noise it produces when a button is pressed, and likes to be the one that presess it. He can almost form complete sentences now.

  87. John Cowan says

    Well said, V.

Speak Your Mind