Thought control to detect the misuse of language
« previous post | next post »
[This is a guest post by Mark Metcalf]
Recently read a short story by Chinese sci-fi author Ma Boyong (translated by Ken Liu) entitled "City of Silence" (Jìjìng zhī chéng 寂静之城) — a tale about a highly dystopian future in ("not") China. The story was referenced in an article in Wired.
Haven't been able to find an English translation online, so I got the Kindle version in a compilation – Invisible Planets. A thought-provoking story that describes a State in which the government controls people's thoughts by monitoring all of their communications in order to detect the "misuse" of language. The following excerpts from the book explain how the process evolved. Very disturbing, with echoes from recent history that are even more disturbing.
Initially, the State only shielded certain sensitive words, but it quickly discovered that it was essentially useless. Many users simply mixed in special characters or numbers or misspelled words to get around the inspection system. The appropriate authorities had to respond by trying to shield these variant spellings. But as everyone knew, the combinations of different characters to approximate the appearance of different words were virtually limitless. Provided you had some imagination, it was always possible to come up with a novel combination and get your meaning across. For example, the word politics could be written as "polit/cs," "政itics," "pol/itic$," and so on.
After the appropriate authorities finally caught on to the problem, they took a new tack. Since it was not possible to filter out all possible combinations of characters that might spell out a word, the solution was to forbid the use of anything except real dictionary words.
This procedure was initially very successful. The number of rule-violators went down significantly. But very soon, people discovered it was possible to use puns, homonyms, or rhyming slang to continue to express the same dangerous ideas. Even if the appropriate authorities filtered out all sensitive words and all possible puns and homonyms with those words, it was useless. Imaginative citizens gave their creativity free rein and used metaphor, metonymy, analogy, etymology, rhyming slang, and other rhetorical tricks to substitute non-sensitive words for sensitive ones. The human mind was far more creative than the computer. The computer might shield off one path, but the people had many more paths to choose from.
This contest, under the surface, seemed to go the way of the people. But then a man who could think outside the box appeared. It was unclear who he really was: some said he was the chief administrator at the appropriate authorities; others said he was a dangerous man who had been arrested for using too many sensitive words. He was the cause of the turn in the tide of the battle between the State and the people.
He suggested to the appropriate authorities that the regulations should no longer explain what was forbidden. Instead, the regulations should set forth what could be said, and how to say it. The appropriate authorities immediately took this advice and issued new regulations. The List of Sensitive Words was eliminated, and in its place came the List of Healthy Words.
This time, the people were on the losing side. In the past, they had delighted in playing cat-and-mouse games with the appropriate authorities on the Web and in daily life. But now the appropriate authorities had them by the throat, since the entire framework and building blocks of language were now under their control.
Nonetheless, the people refused to give up. They began to select words from the List of Healthy Words and use them in novel combinations to express illegal meanings. For example, writing stabilize twice in a row meant "topple," stabilize plus prosperity meant "shield." The appropriate authorities had to keep an eye on this sort of trend and, day after day, eliminate more and more words from the List of Healthy Words to prevent their use in these new roles…
…Life went on peacefully. No, to be precise, there was one bit of difference. That would be the List of Healthy Words: words disappeared from it at a faster and faster pace. Every hour, every minute, words vanished from it. As the pace of revision for the List quickened, e-mails and BBS forum posts became more and more vapid and banal. Since people had to use an extremely limited set of words to express and inexpressibly wide range of thoughts, everyone more and more preferred silence.
Selected readings
- "Only the Communist Party can save the earth" (2/12/19)
- "Ken Liu reinvents Chinese characters" (12/5/16)
- "Ball ball 你" (10/4/17) — third comment from bottom
- "Chinese sci-fi blockbuster The Wandering Earth battles horde of rampaging online pirates", South China Morning Post (2/11/19): China’s latest blockbuster The Wandering Earth is making megabucks at the box office but is fighting a bigger threat than a looming explosion of the sun – a voracious piracy industry that is eating into its ticket takings.
Philip Taylor said,
June 12, 2020 @ 5:51 am
It is wonderful. It is 1984 all over again, in a Chinese guise. But one thing struck me as I read the introduction — the "rules" for the capitalisation of Pinyin require, I believe, that a person's given names be elided and only the first capitalised. And (assuming that they were followed in this case) this momentarily led me to think that the author was /Ma·bɔɪ·ɒŋ/, whereas of course he is /Ma·bəʊ·jɒŋ/ (English approximations to Chinese phonology). I cannot help but feel that the "rules" should take effects such as this into account.
Philip Taylor said,
June 12, 2020 @ 6:43 am
The story is available online by courtesy of the Wayback Machine —
Part one
Part two
Cheryl Thornett said,
June 12, 2020 @ 6:49 am
That reminds me of a book with a similar theme from nearly 20 years ago, focusing on letters, Ella Minnow Pea. https://en.wikipedia.org/wiki/Ella_Minnow_Pea
ycx said,
June 12, 2020 @ 7:02 am
The entire excerpt reads like how 1984's Ingsoc developed Newspeak.
First, use censorship, when that fails, enforce mandated language, when that fails, enforce mandated language and thought patterns. It looks like the antagonists in Silent City are simply missing the final step.
Philip Taylor said,
June 12, 2020 @ 7:13 am
The Wikipedia entry on Ella Minnow Pea contains a fascinating fact — "The full title of the hardcover version is Ella Minnow Pea: a progressively lipogrammatic epistolary fable, while the paperback version is titled Ella Minnow Pea: A Novel in Letters".
If the paperback edition followed the hardcover (which I assume that it did), did this reflect a change of heart on the part of the publisher and/or author regarding the potential inaccessibility of the original title, or did it reflect the publisher's and/or author's view(s) of the probable demographics of the potential audience for each edition ?
And are there other examples of books with different titles for their hard-bound and paperback editions ?
Mark Metcalf said,
June 12, 2020 @ 7:57 am
Many thanks for the links, Philip.
jin defang said,
June 12, 2020 @ 8:29 am
brilliant reductio ad absurdam of a current trend whose implications are frightening.
Adam Field said,
June 12, 2020 @ 9:12 am
Obviously the dystopian part is completely hypothetical, but I like how the main thrust of it is… well, it isn't, I've seen it firsthand, in any video game that has any sort of language filter. Not to suppress dissidence, obviously, just to suppress edgy pre-teen trolls from yelling cusswords, or, let's face it, random ethnic slurs with no context, otherwise the chat would be covered in n-word this, f-word that. Of course, it is *anyway*, they just start calling people f@gs or f/\gs or £ags, knowing full well that our brains are very good at turning random symbols into words. And I've seen games where they use a whitelist instead of a blacklist, and sure enough, that exact same thing often starts happening. Of course, it's a lot easier filtering language when the language has to be sent through a computer system, but then again, with the lockdown right now, most of our communication *does* have to be sent through computer systems.
Dave said,
June 12, 2020 @ 1:44 pm
The https://esolangs.org/wiki/Whitespace language suggests that the … fine authorities may need to lock down ground as well as figure.
David Marjanović said,
June 12, 2020 @ 3:07 pm
In Pinyin, y exclusively occurs at the beginnings of syllables. Therefore, Boyong can only be parsed as bo-yong.
Philip Taylor said,
June 12, 2020 @ 3:25 pm
Yes, I agree that for those completely familiar with Pinyin, there is no problem at all. But for those less familiar, and who are more familiar with the English word "boy" than with the Chinese name "Bo", there may be momentary uncertainty, as there was for me. That is really all I was saying.
Stephen said,
June 13, 2020 @ 3:38 am
Gene Wolfe suggests that even when speech is limited to quoting approved texts, people will express themselves: see the language of the Ascians in the Book of the New Sun.
Michael Watts said,
June 13, 2020 @ 3:51 pm
This is an English approximation to the Chinese spelling, not to the Chinese phonology. "Bo" = b- + wo in the same way that "dui" = d- + wei. English can accommodate a bw- onset without difficulty.
Michael Watts said,
June 13, 2020 @ 3:55 pm
By the way, pursuant to the discussion a while ago of certain words appearing to be "a syllable and a half", I was struck by something that happens in this song: https://www.youtube.com/watch?v=OUWOMJE2nqM
One of the half-choruses goes like so:
What I find interesting is that — to my ear — you're forced to consider "power" and "hour" as single syllables, while "shower" is two syllables. But of course "power", "hour", and "shower" all rhyme with each other according to the rhyme scheme of the chorus.
Philip Taylor said,
June 14, 2020 @ 4:38 am
Michael — '"Bo" = b- + wo in the same way that "dui" = d- + wei. English can accommodate a bw- onset without difficulty' . Interesting. I had three teachers of Mandarin teachers, not one of whom made any mention of this. Now off to consult San Duanmu and Yen-Hwei Lin to see what they have to say on the subject …
Philip Taylor said,
June 14, 2020 @ 5:02 am
P.S. Assuming that you are correct (and I have no reason al all to believe that you are not), I wonder why Pinyin does not make this explicit by including a medial "u"-glide, as in "buo" rather than "bo". I don't have a comprehensive pinyin dictionary to hand, but none of the less comprehensive word lists that I have include words commencing "buo-", so it is not as if the use of "buo-" to indicate what you describe as 'b- + wo' would conflict with a pre-existing use of "buo-". I note, however, that "Chao Yuen Ren pointed out that the *w or "u glide after labial initials in the Karlgren and Wang Li columns are not phonemic, since there are no syllables without this glide" [Routledge Handbook of Early Chinese History, Paul R Goldin (ed)].
Philip Taylor said,
June 14, 2020 @ 6:41 am
One further comment on b[u]o (etc), plus a question.
Comment — I now note that Julian K Wheatley (Learning Chinese, elementary level) states "[…] bo, po, mo, fo rhyme with duo, tuo, nuo, and luo. […] In other words, o by itself always equals uo (and never ou). Apparently, the creators of pinyin felt that, following labial initials, it was not necessary to indicate the labial onset with 'u'". My stress/italics on "necessary" — it may not have been necessary , but it would most certainly have been helpful !
Question : did the 2006 English language edition of Yuen Ren Chao's Linguistic Essays ever make it to the west ? I have been trying to source a copy, but Abebooks seem not to have heard of it, and the nearest library copy I can find is in Cambridge, some 500 miles east of here.
Michael Watts said,
June 14, 2020 @ 8:59 am
I assume this is for the same reason that — drawing from another example we discussed recently — no one has felt the need to respell "tree" / "train" / "tropics" etc. as "chree" / "chrain" / "chropics" etc.
I did have a teacher who seemed to believe that the /w/ glide in bo/po/mo/fo was a necessary feature of the phonetics of the syllable, something that would be impossible not to pronounce. I can believe that this is true of sole Mandarin speakers. (The case is a little harder to make for "fo".)
Rodger C said,
June 14, 2020 @ 9:23 am
@Michael Watts: I'm bemused by your apparently unshakable assumption that pronunciations like "chree" are universal in at least American English. Does anyone have a map of where this occurs? It strikes my inner ear as a Philadelphianism.
Michael Watts said,
June 14, 2020 @ 10:20 am
Can you provide an example of it failing to occur in American English? News broadcast / radio / entertainment / surreptitious recording?
For what it's worth, [tʃɹi] is the only pronunciation of "tree" acknowledged by Merriam-Webster. (Transcribed "\trē\", but the [tʃ] is quite clear in the supplied audio pronunciation.)
Philip Taylor said,
June 14, 2020 @ 10:33 am
Rodger C has already queried the "chree" issue, so I will pass over that, but there is another, more fundamental point, on which I would welcome the views of Michael and others. Leaving aside whether or not anyone actually says "chree", the orthography under discussion is that of the English language, which whilst not set in tablets of stone, nonetheless evolves only very slowly — in my lifetime, for example, I have witnessed the virtual demise of "shew"/"shewn", the loss of the æ ligature digraph in words such as mediæval, and the loss of the œ ligature digraph in words such as fœtus, but few if any other changes to spelling of which I am aware.
Pinyin, however, (unless I am very much mistaken) is an artificial orthography, created ex nihilo by Zhou Youguang and others, and subsequently revised. Now give that Zhou Youguang was working from a tabula rasa, there was (IMHO) no reason why he should not have created an orthography that was as phonetic as possible, given the set of characters which he elected to use. So he could have spelled (e.g.,) bō "buō", just as he spelled "guò", but he did not. And my question, which I really think is a very simple one, is "why ?".
Michael Watts said,
June 14, 2020 @ 10:59 am
I spoke directly to this before:
If something has to be there, indicating it is unnecessary. This is true in general; compare how common it is for spoken (and written) Mandarin to have a 没 without actually realizing the 有 which nevertheless follows it.
This is a speculative theory, but I think it's a reasonable one. The -wo rime is distinct from the -wei and -wen rimes in that those latter two are always abbreviated inappropriately following a nonzero onset, while -wo is only abbreviated following labials. (And, um, f-.) There is in fact an obvious relationship between labial stops and a labial glide.
Another oddity in pinyin from an English perspective is that yan is spelled with an "a" despite being pronounced with what an English speaker would recognize as the DRESS vowel. You could explain this by the fact that the pinyin "e" represents an entirely different sound… except that in one case, ye, "e" is used to represent the DRESS vowel.
(If it were up to me, I would probably spell the more normal "e" of ge / le / ben / etc. differently. But really, if it were up to me, I'd say to use zhuyin and abandon the idea of getting a Roman alphabet to correspond to Chinese phonology while also trying to match its use in a variety of other languages.)
Michael Watts said,
June 14, 2020 @ 11:07 am
I should note that of course English orthography is no less artificial than Pinyin orthography, and was created ex nihilo just as Pinyin was, and subsequently revised. The difference is in how long ago that happened.
Philip Taylor said,
June 14, 2020 @ 11:37 am
OK. I thknk the point that Michael is making ('though he does not make this explicit) is that Pinyin was created for use by native Chinese speakers (and therefore not created as a means to enable non-native speakers to learn to read and understand Chinese). In which case, Michael's point "[i]f something has to be there, indicating it is unnecessary" is perfectly valid — if the target audience for Pinyin would automatically "know" that /b/ (and /p/, /m/ & /f/) must be followed by /w/ when they precede what San Duanmu terms "the mid vowel", or as we might roughly transliterate it (following Wikipedia), /o/.
It is, therefore, singularly unfortunate that we non-native speakers have to learn this rule explicitly, just as we have to learn that the vowel in "uan" is the DRESS vowel (roughly speaking), not the TRAP vowel (also an approximation, of course). When I first set out to learn Mandarin Chinese, I asked each of my three teachers to help me locate an English-Chinese dictionary which used the IPA for Chinese phonetics; sadly none was able to locate or identify one, even in the best bookshop in Shanghai.
Rodger C said,
June 14, 2020 @ 12:08 pm
Can you provide an example of it failing to occur in American English? News broadcast / radio / entertainment / surreptitious recording?
My own usage. West Virginia, born 1948.
Michael Watts said,
June 14, 2020 @ 1:20 pm
Minor note: -uan indicates the (roughly) DRESS vowel only following x- / j- / q- / y-, where it is the combining (or independent) form of the syllable yuan. In other cases, it is the combining form of the syllable wan, and is more like TRAP or PALM than it is like DRESS.
It makes a certain sort of linguistically-informed sense that /ɑ/ would be raised into [ɛ] when following a high front vowel (here, /i/ and /y/). You could make the argument that the vowel is still phonemically /ɑ/, for instance.
But in my mind this is undermined by the fact that yan would appear to be just ye with syllable-final -n, and ye and ya both exist and contrast with each other.
I have a vague suspicion that some pinyin orthographic choices were made with an eye toward accommodating different regional pronunciations of the same syllable, but I can't support that suspicion at all.
Rodger C said,
June 14, 2020 @ 1:35 pm
My last post about "chree": I find that anyone puzzled about his matter need only google /bunched r vs retroflex r/.
Michael Watts said,
June 14, 2020 @ 1:43 pm
This is more of an assertion than an example. But I went to listen to the samples of West Virginian speech at the speech accent archive ( http://accent.gmu.edu/browse_language.php?function=find&language=english ). There are four. english384 is a 57-year-old male, and english388 is a 70-year-old female. (I don't know in what year they were 57 and 70.) english137 is a 26-year-old male, with the advantage that an official "phonetic" transcription is provided.
I claim that all of them exhibit initial [tʃ] in pronouncing "train". english137 is transcribed [tʰɹ], but this is a matter of convention rather than accuracy. Compare english6 (Brooklyn), also transcribed as [tɹ].
Philip Taylor said,
June 14, 2020 @ 3:29 pm
I will pass on the phonetic analysis, leaving that to others more skilled in the task. But I am intrigued to know how speakers were asked to identify themselves. Speaker "english2, female, birmingham, uk" has an accent totally atypical of Birmingham. It is, to my British ear, pure "home counties". English40 (Dudley, UK) is far closer to what I would expect to hear from a native Brummie, although the Dudley and Birmingham accents are not identical, merely similar.
RP said,
June 14, 2020 @ 4:53 pm
Philip Taylor,
On changes to (British) orthography in your lifetime, what about "fulness", "dulness" and "connexion" almost entirely dropping out of use, hyphens becoming much less common ("cooperate" largely taking over from "co-operate", "tomorrow" replacing "to-morrow", and "girlfriend" entirely displacing"girl-friend"), and the full stops in many initialisms ("UN", "UK") and in titles like "Mr" almost entirely disappearing?
Philip Taylor said,
June 15, 2020 @ 1:58 am
An interesting list, RP, with some surprises and some with which I would disagree. "fulness" and "dulness" are unknown to me, so I will check; "connexion" I always associate with the GPO — "strips connexion" were ubiquitous when I was a GPO apprentice (that would be BT these days, of course); "co-operate" I still regard as the norm in <Br.E>, but perhaps I am the last of a dying breed; "to-morrow" I do not recall seeing, while "girl-friend" / "girlfriend" both read correctly to me. Period after "Mr" in British English is simply wrong — an abbreviation takes a period if and only if the final letter of the abbreviation is not the as the final letter of the word (this differs from American practice). Periods in "UK", "UN", etc, rather like "girl-friend" / "girlfriend" — I am simply ambivalent. But an interesting list, and I shall need to research "fulness", etc. Incidentally, I asked a retired headmistress (now in her mid-80s) for her list, and she could not think of any words that had become re-spelled in her lifetime, but asked for more time to think about it. She, however, had never been exposed to "shew" / "shewn", unlike myself.
Philip Taylor said,
June 15, 2020 @ 2:31 am
Just a note on "fulness" — the New English Dictionary (1898) says "the spelling 'fullness', though less common (except in the U.S.) than 'fulness', is here adopted as more in accordance with analogy". The OED (most up-to-date edition) says — as well as noting the above — "The spelling 'fullness' is now by far the most common one, although 'fulness' is still frequent. Compare similar variation, and etymological note, at 'dullness', n".
So, despite the fact that I have never spotted "fulness" in the wild, it would seem that it is still alive and well in Britain. One lives and learns !
Andrew Usher said,
June 15, 2020 @ 6:16 am
I have seen 'fulness', though I don't remember where, presumably in an old-enough book. 'Shew' and its inflections are of course common in the same type of source,
To the subject of 'chree' and so on:
I quite agree with Rodger C. and am pleased to see someone elese recognising the connection to the bunched r. Note that I don't say, or don't like to say, 'retroflex' for the other option because I'm not convinced many Americans actually retroflex – better 'alveolar', 'apical', or just 'tongue-tip r'.
Michael Watts's three West Virginia examples he picks out from that page of samples, besides sounding like three completely different accents, don't all agree on 'tr', and I'd say all three are intermediate between the true [tr] of english351 and the true [tʃr] of english292, both from Missouri. I do agree that the transcriptions (when present) should be given no real weight and are not consistent, however they don't matter here.
The best argument that /tr/ is really [tr] for anyone is that it is possibly consciously to produce [tʃr] substantially different from one's normal /tr/, but not so for [tr] (This is essentially what John Wells said in his blog comment on the matter). There are some for which the reverse would be true – but according to all I know, certainly a minority.
k_over_hbarc at yahoo.com
Rodger C said,
June 15, 2020 @ 10:27 am
I wonder if some of us, including perhaps myself when I'm not paying attention, aren't really saying [ʈʂɹ]. Indeed, when I first saw "chree" I thought it must represent this. I still don't know why Michael Watts is so insistent on asserting himself as the American in the room.
Daniel said,
June 15, 2020 @ 2:53 pm
Andrew, the problem with asserting that someone would pronounce "chr" differently is that that combination (as pronounced [tʃr]) does not occur in any English words. They are not contrastive. This is what allows "tr" to blend from [tr] to [tʃr].
As an aside, my six-year-old daughter noticed this difference in the American pronunciation of "tr" compared with British speech. She initially insisted that British people say "crain" for "train".
The idea of having noncontrastive pronunciations relates to pinyin bo, po, mo, fo. After other initials, there is a contrast between "e" and "uo", but after b/p/m/f there is no such contrast. The syllables "be", "pe", and "fe" don't exist, and "me" is a recent phenomenon as a very reduced part of the common word "shenme" (meaning "what") where the second syllable's full pronunciation was historically "mo". The letter "o" is chosen because the vowel is rounded. I suppose the labial nature of the preceding consonant causes the rounding ("e" by itself is a back vowel in mandarin Chinese).
Unlike Michael, My Chinese teacher did not emphasize the "u" onglide in bo/po/mo/fo, but I have heard it in other's speech to varying degrees. In any case, the vowel is "o", and using a pure "e" vowel after b/p/m/f would sound very weird
Andrew Usher said,
June 15, 2020 @ 6:36 pm
Yes, /tʃr/ doesn't exist in English words as such, but it's merely an accidental gap and not a feature of the phonology; similarly /sr/ doesn't exist, yet the very first time I saw the name 'Sri Lanka' I read it with /sr/ (I'm now not sure that's actually correct) and had no trouble doing so.
That is why John Wells used the example of 'century' compressed to two syllables, as that would unambiguously have /tʃr/ and would normally be distinguished from 'sentry' which differs only in lacking the /ʃ/, and further anyone that does distinguish them is going to pronounce 'tree' the second way, not the first. Only those that really have the merger and for whom century/sentry differ only in syllabicity (is Michael Watts among these?) could correctly use 'chree', 'chrain', etc. as phonetic spellings.
The cluster /tʃr/ occurs across word boundaries, too, and is certainly normally distinct from /tr/ there. Again I'd expect those with a true merger to at least some of the time pronounce 'flat ring' as 'flatch ring', which I never do.
In your Chinese example, the non-contrastive pronunciation is the presence of the glide before 'o' – not the e/o distinction, which is phonemic. At least that's my take from your commentary.