Digitoneurolinguistic hacking

« previous post | next post »

The most recent xkcd takes on the scourge of Trochee Fixation:

(Click on the image for a larger version.)

The Huffman-coding crack in the mouse-over title ("If you Huffman-coded all the 'random' things everyone on the internet has said over the years, you'd wind up with, like, 30 or 40 bytes *tops*.") is a bit of an exaggeration, but the meaning is clear. I'm more in interested in the footnote in the sixth panel:

… we're modifying her vocabulary* to erase the words she's fixated on.

*Digitoneurolinguistic hacking! It's totally real! Ask Neal Stephenson.

This is probably a reference to the plot device in Snow Crash that the Wikipedia entry describes this way:

The book explores the controversial concept of neuro-linguistic programming and presents the Sumerian language as the firmware programming language for the brainstem, which is supposedly functioning as the BIOS for the human brain. According to characters in the book, the goddess Asherah is the personification of a linguistic virus, similar to a computer virus. The god Enki created a counter-program which he called a nam-shub that caused all of humanity to speak different languages as a protection against Asherah, supposedly giving rise to the biblical story of the Tower of Babel. […]

As Stephenson describes it, one goddess/semi-historical figure, Asherah, took it upon herself to create a dangerous biolinguistic virus and infect all peoples with it; this virus was stopped by Enki, who used his skills as a "neurolinguistic hacker" to create an inoculating "nam-shub" that would protect humanity by destroying its ability to use and respond to the Sumerian tongue. This forced the creation of "acquired languages" and gave rise to the Biblical story of the Tower of Babel. Unfortunately, Asherah's meta-virus did not disappear entirely, as the "Cult of Asherah" continued to spread it by means of cult prostitutes and infected women breast feeding orphaned infants …

But it might also be a reference to the echoes in The Big U of Julian Jaynes' The Origin of Consciousness in the Breakdown of the Bicameral Mind. However, I've mercifully forgotten the details, and my copy has vanished, perhaps taken by undercover operatives from the Sumerian Dictionary Project.

[I feel that the Wikipedia article's reference to "neurolinguistic programming"  in Snow Crash is somewhat misleading, not say a mistake. NLP is certainly controversial, not to say nuts, but Sumerian is not in the picture, as far as I know. And Snow Crash focuses on "neurolinguistic hacking", a term and concept that were perhaps inspired by NLP, but NLH (to coin an acronym) is anything but therapeutic in intent.]


  1. Titanium Carbide said,

    February 4, 2011 @ 10:00 am

    I couldn't resist fixing Wikipedia's mistake, but incidentally that outdates your link. If you want the previous version see http://en.wikipedia.org/w/index.php?title=Snow_Crash&oldid=409887177.

  2. greg said,

    February 4, 2011 @ 10:03 am

    And Snow Crash focuses on "neurolinguistic hacking", a term and concept that were probably inspired by NLP, but NLH (to coin an acronym) is anything but therapeutic in intent.

    Nonsense! Everyone knows that "hacking" is actually very well intentioned. It's the neurolinguistic crackers that are the the ones with bad intentions. It's the media and some misguided public interest groups which have confused the terms! Perhaps it's actually the nefarious design of one such NLCer who has attempted to spread the use of the inherently positive term in a negative way to mask the efforts of the truly antagonistic programmers.

    (Don't mind me, I'm just going to search the LL logs for discussion of white hats vs black hats.)

  3. Frank said,

    February 4, 2011 @ 10:05 am

    Holy crap! Thanks for that. Can you believe you're the only result on Google right now for "Digitoneurolinguistic hacking" "neal stephenson"?

  4. Leonardo Boiko said,

    February 4, 2011 @ 10:12 am

    Well, “neurolinguistic programming” books I’ve read… er… heard about… weren’t therapeutic at all. They were all about how to control other people by exploiting their astrological sign visual-auditory-kinesthetic affinities.

    [(myl) Your experience with "neurolinguistic programming" is obviously more extensive than mine.

    My impression is that it started with a psychotherapeutic focus, and later on did turn to what the Wikipedia article calls "interpersonal communications and persuasion". But I don't think "persuasion" extends to Sumerian brainstem viruses aimed at producing the neurological equivalent of "a system crash … at such a fundamental level that it frags the part of the computer that controls the electron beam in the monitor, making it spray wildly across the screen, turning the perfect gridwork of pixels into a gyrating blizzard".]

  5. John Roth said,

    February 4, 2011 @ 11:02 am

    Since I was peripherally involved in the early 80s (took classes, read books, met some of the participants in the 70's graduate linguistics seminar that started it all), I can say with some confidence that they were interested in the cognitive patterns used by outstanding performers in any field – that is, they had a very definite and explicit human potential orientation with the assumption that duplicating the patterns would duplicate the performance.

    They used investigations of several people with reputations as therapeutic wizards as their initial testbed, and then took those findings public. This is where the impression that it was theraputically oriented came from.

    The way the whole enterprise crashed and burned within a few years could make a great case study in good intentions leading to, well, junk.

    While I haven't been involved for a quarter century, I'm still amused by the way that current brain research is turning up some of the bases for some of their claims.

  6. David said,

    February 4, 2011 @ 11:05 am

    You really need to go read the XKCD Blag entry about this, not least because of what he found out about Google hit-counts, which you guys seem to use an awful lot.


    [(myl) Google's algorithms for estimating hit-counts work by extrapolating from the intersection of relatively small samples of stored indices for words or word sequences. The general approach, as of five or six years ago, was discussed for the case of boolean search here and here.

    The case of word-sequence search is basically similar: when you look for e.g. "xkcd whatever", there's an list that gives the top N hits for "xkcd", and another that gives the top N hits for "whatever", and the algorithm checks the intersection of these lists (if any) to look for instances of the sequence; and all of the resulting numbers are plugged into an equation that extrapolates to an estimate of how many sequence-hits they would find if they searched every page exhaustively. This algorithm, which they apparently change from time to time, seems rarely to be very accurate.

    However, the count of actual pages that are shown to you is clearly worse — it just reflects what actually emerges from the intersection of the top-N lists (where N is typicallysmall relative to the total number of relevant pages out there).

    The indices have gotten bigger, and the algorithms have no doubt changed somewhat, but the hit-count estimates are still not reliable. (There are half-a-dozen other LL discussions of these issues elsewhere, if you care about the issues.)]

  7. Stan said,

    February 4, 2011 @ 11:26 am

    Much as I enjoyed Snow Crash, Jaynes's book has stayed with me more vividly. I'd be curious to know what other L.Loggers make of it — especially since I recently read Daniel Smith's Muses, Madmen, and Prophets: Rethinking the History, Science, and Meaning of Auditory Hallucinations, which refers to Jaynes several times but doesn't commit itself to a particular interpretation of the phenomenon. I don't subscribe to Jaynes's thesis, but I find it very interesting at the same time. Of course, it helps that he wrote so well, and that his main proposals are ultimately both unprovable and unfalsifiable.

  8. Twitter Trackbacks for Language Log » Digitoneurolinguistic hacking [upenn.edu] on Topsy.com said,

    February 4, 2011 @ 11:49 am

    […] Language Log » Digitoneurolinguistic hacking languagelog.ldc.upenn.edu/nll/?p=2944 – view page – cached February 4, 2011 @ 9:46 am · Filed by Mark Liberman under Linguistics in the Tags […]

  9. Neal Goldfarb said,

    February 4, 2011 @ 11:59 am

    Snowclone + crash blossom = snow crash blossom cone (a garden-path newspaper headline that follows a productive structural pattern, which pattern has become a semi-meta-cliche).

  10. Neal Goldfarb said,

    February 4, 2011 @ 12:03 pm

    Alternatively: Snow crash blossom cone = A tasty treat made with Compound-N-chip ice cream.

  11. Aaron Binns said,

    February 4, 2011 @ 2:25 pm

    Inspired by this, I whipped up a similar chart for the same set of trochee phrases in the Internet Archive's Archive-It service. It's a hosted web archiving service of the Internet Archive (where I work) and contains just over one billion archived web pages from 2006-current. A mini-wayback machine (if you will) of curated web archives.


    I tried to come close to the same color-coding, but I’m probably off a bit.

    [(myl) Neat! When I have a minute, I'll be curious to see what the correlation between the two sets of counts is. (Or with a little luck, someone will beat me to it…)

    OK, no luck, so I did (some of) it myself. A plot of the corresponding counts in the first five rows of the two tables (Randall's Google counts and your Internet Archive counts) is here, corresponding to a correlation of r=0.93. Leaving out "doctor doctor", the same plot is here, corresponding to a correlation of r=0.62.]

  12. Aaron Binns said,

    February 4, 2011 @ 5:34 pm

    I've added a brief HTML page with the chart along with links to the raw data, including counts by year: 2006-2010.


    BTW, IA is looking for ways to make our datasets more easily and readily available to researchers. If you'd like to find out more, please send me an email.

  13. Rod Johnson said,

    February 4, 2011 @ 7:57 pm

    Ignoring the (D)NL{P|H} issues, I have to say this is a pretty interesting insight about the whole monkey/ninja/robot/zombie/pirate phenomenon.

  14. Hermann Burchard said,

    February 4, 2011 @ 11:10 pm

    @Rod, Aaron: ninja phenomenon

    Brings to mind "ninja turtle" as an instance of ..

    The map is the territory..

  15. Aaron Binns said,

    February 5, 2011 @ 1:59 am

    I was a little hesitant in posting the Archive-It graph, as I knew it would be subject to statistical scrutiny and I'm unsure if I'll be able to provide additional data to satisfy. But, since I've been a longtime reader, I thought I'd go ahead and put it out there.

    Not to appear to be making too many excuses, the Archive-It dataset has changed a lot over time: in scope and size. In 2006, we had ~25 customers, almost entirely academic institutions and a few state libraries. In 2011 there are nearly 150 customers, including over a dozen schools ranging from elementary to high schools. The size, scope, depth and breadth of the collections have changed over the years. Not that this necessarily invalidates statistical analysis of the data, but it needs to be considered.

    If there is any more data I can provide, please let me know and I'll see what I can do.

  16. dl said,

    February 5, 2011 @ 8:48 am

    @ Hermann Burchard
    which brings to mind Alfred Korzybski who was an early user of the neurolinguistic meme.

    He refers here to the Korzybski method of direct neuro-linguistic …training.

  17. Rod Johnson said,

    February 5, 2011 @ 12:03 pm

    @Hermann: that's in the comic!

  18. Hermann Burchard said,

    February 5, 2011 @ 3:27 pm

    @Rod: Had wanted to give example of [word1] [word2] but can't get braces (instead of brackets [,]).

    Okay, I was reading the comic diagonally, but now i see "NINJA POWER MUTANT TURTLE" as well as "BANJO TURTLE", clear signs of right-brain prodding left-brain pathology, Julian Jaynes-style. — But, how does non-verbal right brain activate Wernicke and Broca centers in left brain?

    @dl: Korzybski said "the map is not the territory," but in linguistics the map (=language) is the territory. Is it Okay to admit that I read NLP literature in the eighties and later credit its influence in my '05 paper in Foundations of Science (along with Kant, Heidegger, etc)?

  19. Matt G said,

    February 5, 2011 @ 11:27 pm

    "The map is not the territory" is also the key utterance at the climax of the game of eschaton in Infinite Jest.

  20. bfgray said,

    February 6, 2011 @ 4:14 am


    thanks for a post full of interesting tangents. But I was disappointed you didn't go further into the core issue, the preponderance of trochees in internet memes. Are there any parallels to this that anyone knows of? Any other instances of the magic power of the trochee that spring to mind? The only one I can come up with is the witches' trochaic tetrameter in 'Macbeth': "Double, double, toil and trouble/Fire burn and cauldron bubble…". Are there other stress patterns with a similar status in other languages?

    [(myl) Jokes aside, I'm not entirely sure that it's true that there's a statistically unexpected preponderance of trochees in the names of superheroes, internet memes, TV shows, etc. We expect quite a few trochees, because initial stress is much commoner than final stress in English two-syllable words, and two-syllable words are pretty common, and noun compounds made up of two monosyllabic nouns will also have a trochaic stress pattern.. So the question is whether there are more fully- or partly-trochaic names than we expect by chance.

    Frankly, I'd tend to guess that the answer is "no". I guess this could be the topic for a breakfast experiment, given a list of words with their prosodic shapes and frequencies, and a list of names to check. I considered this and rejected it, on the grounds that it was unfair to comics. But given Randall's dedication to science, and the evident interest of LL readers in this issue, maybe I'll reconsider.

    For a zeroth-order check, taking the start of a list of male superheroes, I find that out of the first 24 (batman, wolverine, iron man, spider-man, superman, flash, hulk, silver surfer, captain america, green lantern, nightcrawler, daredevil, gambit, green arrow, human torch, iceman, cyclops, deadpool, nightwing, aquaman, thor, magneto, x-man, blade) only 8 are fully trochaic (batman, silver surfer, gambit, iceman, cyclops, deadpool, nightwing, x-man), while some others are partly-trochaic (captain america, green lantern, etc.). I get the same score (8 fully-trochaic out of the first 24) for a list of female superhero names. So that's why I'm guessing that a more rigorous test would have trouble rejecting the null hypothesis that memetic names are metrically random.

    It's true that "pop" meters in English are generally falling (featuring trochees or dactyls) whereas "art" meters tend to be iambic, presumably because words tend to cross foot boundaries more often, creating greater possibilities for metrical tension. ]

  21. army1987 said,

    February 6, 2011 @ 6:30 pm

    Dunno, but in Italian probably a majority of content words are two syllables with the stress on the first.

  22. Faldone said,

    February 7, 2011 @ 1:14 pm

    I think trochees are quite common in English, too. All three two-syllable words in my first sentence are trochees

  23. Charly said,

    February 7, 2011 @ 8:08 pm


    The line from Macbeth is one of many, many Shakespearean iambic pentameter lines. Of course, this poetic structure can take many forms:

    "shall I comPARE thee TO a SUMmer's DAY?"

    'if YOU had KNOWN the VIRtue of the RING"

    A similar pattern, alexandrins, was very popular in French poetry/drama. Molière was a fan. Those are generally 12 syllables long, with 6 strong.

    Also, trochee NAMES are very very common in English. So much so that the song "Happy Birthday" is very awkward if the name is otherwise.

    "Happy Birthday dear Charly/Matthew/Gunnar/Lizzie" sounds a whole lot better than "Anastasia/Bo-o-ob/Jennifer,' to my ear.

    [(myl) Actually, while you're right that most of metrical Shakespeare is iambic pentameter, bfgray was correct that the witches' incantation from the first scene of act 4 in Macbeth is trochaic tetrameter catalectic, from the beginning:

    Thrice the brinded Cat hath mew'd.
    Thrice, and once the Hedge-Pigge whin'd.
    Harpier cries, 'tis time, 'tis time.

    to the end:

    By the pricking of my Thumbes,
    Something wicked this way comes:
    Open Lockes, who euer knockes.

  24. baylink said,

    February 7, 2011 @ 8:27 pm

    Yes, but that's syllable count, not stree.

    How many English given names have second syllable stress?

  25. baylink said,

    February 7, 2011 @ 8:28 pm

    Excuse me: "two syllable English given names"

  26. bfgray said,

    February 8, 2011 @ 2:40 am


    thanks for getting back to this topic – I was worried I'd missed the boat – and for your mini-(pre?)-breakfast experiment. I would be delighted if you chose to delve further into this 'issue' – but first, I'd better make my query clearer.

    Of your 8 trochaic male superhero names, only one (Silver Surfer) is a double trochee (I'm sure there's a more technical term, but I'm leaving Wikipedia out of this). One trochee all by itself doesn't have any peculiar impact. What I'm wondering is whether there really is anything 'special' (a word I'll have to leave undefined) about a whole string of trochees, as in 'Robot Monkey Super Jesus'.

    Phenomenologically, there is – these strings seem to have a peculiar power all their own. I would suggest that the special role of trochaic tetrameter in Shakespeare – where it's used almost exclusively for the words of witches ('Macbeth'), fairies (e.g. 'Midsummer Night's Dream'), and madmen (Poor Tom/Edgar in 'King Lear') – confirms my intuition. Other poems that make extensive use of trochaic metres, such as 'Hiawatha', have a very particular feel – perhaps only because this metre is rare in English, but perhaps not.

    I have to admit that, shortly after writing my original post, I did have a short crisis of faith in which I thought I was an idiot – after all, most English common nouns have stress on the first syllable; but strings of common nouns, on the other hand (such as 'Wizard Pirate Narwhal Raptor') are *not* common. I remember being told at school that iambic pentameter was favoured by Shakespeare and other poets because it approximates the natural rhythm of English speech. Is that entirely untrue? After all, because of unstressed articles and clitic (?) pronouns, most English sentences begin on an unstressed syllable, don't they? So, in that case, trochaic rhythms *might* have a special status in English, which *might* go *some way* towards explaining the popularity of the kind of internet memes Munroe is referring to… no?

    (Footnote: I've just noticed that your comment about 'art' and 'pop' metres [BTW, do you have any references for that?] would seem to contradict my old English teacher's claim about iambic pentameter, and thinking about it, I guess most sentences uttered in conversation would end on a weak syllable, wouldn't they? Hmm.)

  27. chris said,

    February 8, 2011 @ 10:13 am

    I'm inclined to think that the whole phenomenon is a snowclone of "Teenage Mutant Ninja Turtles", unless it can be firmly antedated. TMNT themselves are fairly well known in the relevant subculture (I think), a string of trochees, and also the same kind of conceptual mashup, and anything else that you might want to substitute into the formula will fit better if it is another trochee. Since lots of words *are* trochees, the formula is very productive and there's often a trochaic synonym for some other word you might have trouble fitting in.

    And if you happened to have seen the TV show as a kid, then you might remember the song, which reinforces the meter.

  28. J. W. Brewer said,

    February 8, 2011 @ 3:05 pm

    @baylink, I can think of iambic first names, but my quick scan of the 100 most common names for both baby boys and baby girls given in the U.S. in 2009 saw zero iambs but lots of trochees. The iambic "Christine" has plummeted since its heyday 40 years ago (when it was as high as #14 in popularity) and is now down to #623 (probably not because of prosody, since its trochaic relative "Kristin" is in similarly steep decline). The iambic Laverne (of . . . "and Shirley" fame) has not been in the top 1000 since '71 and was never in the top 100.

  29. ENKI-][ said,

    February 12, 2011 @ 6:37 pm

    I feel the need to note that NLP is used by 'pickup artists' — and that group can be argued to be fairly unconcerned with human potential in the abstract. Regardless of the circumstances of its nativity (itself fairly confused: Korzybski was worried about clarity of communication, not therapy per-se) it has an extremely pragmatic (if dubious) bent.

  30. Kevin said,

    February 13, 2011 @ 12:27 pm

    That NLP/NLH stuff in Snow Crash just totally put me off the book. Controlling minds by uttering Sumerian syllables… ugh.

  31. Digitoneurolinguistic Hacking « The Connectome said,

    February 16, 2011 @ 3:41 pm

    […] another note, here's a short blog post that explains a little more about digitoneurolinguistic hacking. It's a term coined by Neal Stephenson, one of my very favorite authors, in his novel Snow […]

  32. Trochee-trochee, Grelling-Nelson « Arnold Zwicky's Blog said,

    April 17, 2011 @ 1:26 pm

    […] of an xkcd cartoon on the plague of trochees, especially double trochees (Language Log discussion here), to friends, I reflected, not for the first time, that trochee is indeed a trochee (´ˇ), but so […]

RSS feed for comments on this post