Loading ...
Sorry, an error occurred while loading the content.

Re: Creating a metaconlang; anyone want to join?

Expand Messages
  • Sai Emrys
    ... I presume they already have *something* like this for cases where spies have to meet in hostile territory under observation... And no, it s not at all far
    Message 1 of 18 , Nov 30, 2007
    • 0 Attachment
      On Nov 30, 2007 8:33 PM, <li_sasxsek@...> wrote:
      > In today's ultra-paranoid government, I wouldn't be one bit
      > surprised if they took notice and started looking for something like
      > this. It's not too far from the idea of steganography.

      I presume they already have *something* like this for cases where
      spies have to meet in hostile territory under observation...

      And no, it's not at all far from steganography.

      > We do still use tone, we just don't use it phonemically. I'm not
      > sure using it is an option. Clicks are certainly an option, and
      > maybe even sighs or other cues could be used. Do they all have to
      > be audible, or are we going to mix in some body language?

      Certainly!

      In fact, that's a desirable feature; anything you can do in person is
      a potentially exploitable substrate channel. (E.g. suppose how far you
      stand from people, or at what angle, were "phonemic"?)

      For the paranoid types, it'd ensure that people just listening in on
      your conversation only get the 'plain' version... would beat the "wink
      and mouth, then write something with the music turned up" version of
      avoiding it given on TV. ;)

      (This isn't particularly my goal as such, so much as just a cute
      bonus. The goal is more simply, to have an extra information channel
      riding on top of normal conversations, that can be done in real time,
      to make people who are 'in' on it have better and possibly secret
      communication.)


      One thing that Alex points out, correctly, is that a-c are
      significantly different tasks.

      A = create a conlang on a very limited channel
      B = find features unmarked (or marked inefficiently) in carrier
      language, bind them to unused 'phoneme's (plus a bit)
      C = combo of B&D, such that substrate is not arbitrary to the 'true' message
      D = encode A using the erstwhile (substrate) content as a sort of one-time-pad

      B is the version I'm most interesting and think is most plausibly
      implementable as a usable, real-time metalanguage.

      - Sai

      P.S. Check your mail setup. Your reply-to is set as your personal
      address rather than the list address...
    • caeruleancentaur
      ... I like the term semantic space. I am reminded of the secret hand gestures used by the Bene Gesserit in the Dune stories. Charlie
      Message 2 of 18 , Dec 1, 2007
      • 0 Attachment
        >Sai Emrys <sai@...> wrote:

        >* use semantic space* that isn't denotatively used by the language
        >itself (e.g. English doesn't have tone, clicks, facial gestures,
        >signing, so we could use any of that)

        I like the term "semantic space."

        I am reminded of the secret hand gestures used by the Bene Gesserit in
        the "Dune" stories.

        Charlie
      • Joe Schelin
        ... I was going to say that this concept reminds me of the scene in Dune where there is polite conversation going on, with double meanings to some of the
        Message 3 of 18 , Dec 1, 2007
        • 0 Attachment
          > >Sai Emrys <sai@...> wrote:
          >
          >>* use semantic space* that isn't denotatively used by the language
          >>itself (e.g. English doesn't have tone, clicks, facial gestures,
          >>signing, so we could use any of that)
          >
          > I like the term "semantic space."
          >
          > I am reminded of the secret hand gestures used by the Bene Gesserit in
          > the "Dune" stories.
          >
          > Charlie
          >
          I was going to say that this concept reminds me of the scene in
          'Dune' where there is polite conversation going on, with double meanings to
          some of the words, so there is an extra layer of meaning floating through
          the conversation.

          Oh, yeah, I'm new(ish?) I've been lurking for a while, just now
          decided to post :)

          Joe
        • Jim Henry
          ... My understanding is that the code-talkers just translated the English messages into Navaho and talked normally to each other, relying on the obscurity
          Message 4 of 18 , Dec 1, 2007
          • 0 Attachment
            On Nov 30, 2007 7:11 PM, <li_sasxsek@...> wrote:

            > Seriously though, maybe a look at Navaho code talk could give some
            > ideas.

            My understanding is that the "code-talkers" just
            translated the English messages into Navaho and talked
            normally to each other, relying on the obscurity
            and difficulty of Navaho to stymie the Japanese
            (or Germans?) who were listening in. Any obscure
            natlang or a priori conlang could work the same way;
            preferably some language about which nothing has
            been published in your enemy's language.

            --
            Jim Henry
            http://www.pobox.com/~jimhenry
          • Jim Henry
            ... English does use tone, though not phonemically as tonal languages do. I suspect using suprasegmental tone as your extra channel will make English sound
            Message 5 of 18 , Dec 1, 2007
            • 0 Attachment
              On Nov 30, 2007 6:26 PM, Sai Emrys <sai@...> wrote:

              > Recently I've been giving more serious thought to creating a
              > metaconlang. That is, a sort of grammar that can ride on top of an
              > existing language (English as a default, given its status as lingua
              > franca). It'd have a few features:
              >
              > * use semantic space* that isn't denotatively used by the language
              > itself (e.g. English doesn't have tone, clicks, facial gestures,
              > signing, so we could use any of that)

              English does use tone, though not phonemically as
              "tonal" languages do. I suspect using suprasegmental
              tone as your extra channel will make English sound
              not just

              > * be anywhere from almost undetectable to unobtrusive (on the level of
              > an odd mannerism / speech impediment)

              but _really weird_. Relatively quiet clicks, and
              subtle alterations of facial expression, gesture,
              and posture are probably better things to explore.

              > I see (b) as the most interesting variant, though (c) could also be useful...

              I agree. Evidentiality, validationality, distinction of
              exclusivity for we/us/our and number for you/your;
              perhaps number marking for nouns that are
              invariant in English (deer, sheep, fish)...
              Things like that could be useful.

              --
              Jim Henry
              http://www.pobox.com/~jimhenry/review/log.htm
            • li_sasxsek@NUTTER.NET
              ... some ... There s much more than that to it. It was sort of a code on top of a code, with the use of a little-known language like Navaho only making it
              Message 6 of 18 , Dec 1, 2007
              • 0 Attachment
                > [mailto:CONLANG@...] On Behalf Of Jim Henry

                > > Seriously though, maybe a look at Navaho code talk could give
                some
                > > ideas.
                >
                > My understanding is that the "code-talkers" just
                > translated the English messages into Navaho and talked
                > normally to each other, relying on the obscurity
                > and difficulty of Navaho to stymie the Japanese
                > (or Germans?) who were listening in. Any obscure
                > natlang or a priori conlang could work the same way;
                > preferably some language about which nothing has
                > been published in your enemy's language.

                There's much more than that to it. It was sort of a code on top of
                a code, with the use of a little-known language like Navaho only
                making it more difficult for the enemy to decipher.

                > The Navajo Code Talker's Dictionary

                When a Navajo code talker received a message, what he heard
                was a string of
                seemingly unrelated Navajo words. The code talker first had
                to translate
                each Navajo word into its English equivalent. Then he used
                only the first
                letter of the English equivalent in spelling an English
                word. Thus, the
                Navajo words "wol-la-chee" (ant), "be-la-sana" (apple) and
                "tse-nill" (axe)
                all stood for the letter "a." One way to say the word "Navy"
                in Navajo code
                would be "tsah (needle) wol-la-chee (ant) ah-keh-di- glini
                (victor) tsah-
                ah-dzoh (yucca)."

                Most letters had more than one Navajo word representing
                them. Not all words
                had to be spelled out letter by letter. The developers of
                the original code
                assigned Navajo words to represent about 450 frequently used
                military terms
                that did not exist in the Navajo language. Several examples:
                "besh- lo"
                (iron fish) meant "submarine," "dah-he- tih-hi"
                (hummingbird) meant
                "fighter plane" and "debeh-li-zine" (black street) meant
                "squad." <

                http://www.history.navy.mil/faqs/faq61-2.htm
              • li_sasxsek@nutter.net
                ... Yes, and we could do other things if English is the language we are working with. Maybe we could slightly undo some of the vowel reduction, or even
                Message 7 of 18 , Dec 1, 2007
                • 0 Attachment
                  > [mailto:CONLANG@...] On Behalf Of Jim Henry

                  > but _really weird_. Relatively quiet clicks, and
                  > subtle alterations of facial expression, gesture,
                  > and posture are probably better things to explore.

                  Yes, and we could do other things if English is the language we are
                  working with. Maybe we could slightly undo some of the vowel
                  reduction, or even slightly shift some of the vowel positions.
                  Maybe even toss in a hint of another language, like turning voiced
                  stops into fricatives like Spanish does to have some sort of meta
                  purpose. We could even tie these features to the main message. A
                  click on a word could be "opposite of", while a tone change could
                  mean "this word describes the color", and have other markers for
                  metaphoric speech.
                • MorphemeAddict@WMCONNECT.COM
                  In a message dated 12/1/2007 11:14:28 AM Central Standard Time, ... There is a scene between two Bene Gesserit women where they are having a banal conversation
                  Message 8 of 18 , Dec 1, 2007
                  • 0 Attachment
                    In a message dated 12/1/2007 11:14:28 AM Central Standard Time,
                    Joe-Schelin@... writes:


                    > I was going to say that this concept reminds me of the scene in
                    > 'Dune' where there is polite conversation going on, with double meanings to
                    > some of the words, so there is an extra layer of meaning floating through
                    > the conversation.
                    >
                    >


                    There is a scene between two Bene Gesserit women where they are having a
                    banal conversation in speech and the real conversation in something like
                    fingerspelling or very compact sign language. We get to see subtitles for this mode.

                    stevo </HTML>
                  • Douglas Koller
                    From: MorphemeAddict@WMCONNECT.COM ... Wa! Haragei on steroids. Kou
                    Message 9 of 18 , Dec 1, 2007
                    • 0 Attachment
                      From: MorphemeAddict@...

                      > There is a scene between two Bene Gesserit women where they are having a
                      > banal conversation in speech and the real conversation in something like
                      > fingerspelling or very compact sign language. We get to see subtitles for this
                      > mode.

                      Wa! Haragei on steroids.

                      Kou
                    • Sai Emrys
                      ... If it were really pronounced on every word, then yeah. But if it were used sparsely, I d think it wouldn t be so obvious. ... *nod* ... Agreed. It would
                      Message 10 of 18 , Dec 1, 2007
                      • 0 Attachment
                        On Dec 1, 2007 9:21 AM, Jim Henry <jimhenry1973@...> wrote:
                        > English does use tone, though not phonemically as
                        > "tonal" languages do. I suspect using suprasegmental
                        > tone as your extra channel will make English sound
                        > not just
                        >
                        > > * be anywhere from almost undetectable to unobtrusive (on the level of
                        > > an odd mannerism / speech impediment)
                        >
                        > but _really weird_.

                        If it were really pronounced on every word, then yeah.

                        But if it were used sparsely, I'd think it wouldn't be so obvious.

                        > Relatively quiet clicks, and
                        > subtle alterations of facial expression, gesture,
                        > and posture are probably better things to explore.

                        *nod*

                        > > I see (b) as the most interesting variant, though (c) could also be useful...
                        >
                        > I agree. Evidentiality, validationality, distinction of
                        > exclusivity for we/us/our and number for you/your;
                        > perhaps number marking for nouns that are
                        > invariant in English (deer, sheep, fish)...
                        > Things like that could be useful.

                        Agreed.

                        It would also be interesting to add a layer of modifying or even
                        contradicting the base meaning, so e.g. you could say "Meet my
                        friend3, Bob" ~= "Meet my backstabbing colleague who I pander to
                        'cause he's dangerous, Bob". ;)

                        - Sai
                      • li_sasxsek@NUTTER.NET
                        ... You could even strech a short phrase over a long dissertation. ... Yes, and I ve been pondering this a bit and forgot about gangster language. Organized
                        Message 11 of 18 , Dec 1, 2007
                        • 0 Attachment
                          > [mailto:CONLANG@...] On Behalf Of Sai Emrys

                          > > > * be anywhere from almost undetectable to unobtrusive (on
                          > the level of
                          > > > an odd mannerism / speech impediment)
                          > >
                          > > but _really weird_.
                          >
                          > If it were really pronounced on every word, then yeah.
                          >
                          > But if it were used sparsely, I'd think it wouldn't be so obvious.


                          You could even strech a short phrase over a long dissertation.


                          > It would also be interesting to add a layer of modifying or even
                          > contradicting the base meaning, so e.g. you could say "Meet my
                          > friend3, Bob" ~= "Meet my backstabbing colleague who I pander to
                          > 'cause he's dangerous, Bob". ;)

                          Yes, and I've been pondering this a bit and forgot about gangster
                          language. Organized crime organizations often have ways of
                          speakings about things without actually speaking about them because
                          they know there's always the possibility of someone listening in on
                          their conversations. Without a direct statement, they avoid
                          incriminating themselves. Even the most innocent sounding
                          conversation could actuallly mean something like "whack that
                          backstabbing ...".

                          Sonny: How's Paulie?
                          Clemenza: Oh, Paulie... won't see him no more.
                        • John Vertical
                          This, or more specifically your idea A (conlang over a limited channel) is something I ve pondered too. The specific context I have been considering is a
                          Message 12 of 18 , Dec 1, 2007
                          • 0 Attachment
                            This, or more specifically your idea "A" (conlang over a limited channel) is
                            something I've pondered too. The specific context I have been considering is
                            a different one however: a religious ritual language based on repeating a
                            mantra.

                            I had not considered body language (and indeed, I would like this to remain
                            writable), and randomly inserted sighs, clicks, etc don't sound enticing to
                            me either, but tones would be OK. I had however also planned to use some
                            manoeuvres you did not (explicitely) mention:
                            * Phoneme splits. Say the host language is English - one could then split
                            eg. /tS/ into 18 different consonants based on alveolar vs retroflex vs
                            alveolo-palatal; tenuis vs aspirated vs ejectiv; and plain vs labialized.
                            * Syntax changes. The fact that the "carrier" here is a single mantra and
                            not the complete host language givs quite a bit of leeway that might not be
                            otherwise available.
                            * Synonym variation, probably including arcaic and/or dialectal allomorphy.
                            In order to be compatible with phoneme splits, this would probably have to
                            specify wide semantic categories, while the latter would then further
                            specify them down...
                            I suppose technically, these mostly have more to do with your idea "C",
                            actually, since the resulting code would not be adaptable for an arbitrary
                            carrier phrase.

                            Alas, before I get the host language to an usable state (and I might need
                            the conreligion too...) this isn't going to progress much.

                            John Vertical
                          • Sai Emrys
                            ... Check out David Peterson s Sign Language IPA . ... Indeed you could! ... Depends; what sort of language & mantra? Is this one of those quote-unquote free
                            Message 13 of 18 , Dec 3, 2007
                            • 0 Attachment
                              On Dec 1, 2007 2:06 PM, John Vertical <johnvertical@...> wrote:
                              > I had not considered body language (and indeed, I would like this to remain
                              > writable),

                              Check out David Peterson's "Sign Language IPA".

                              > * Phoneme splits. Say the host language is English - one could then split
                              > eg. /tS/ into 18 different consonants based on alveolar vs retroflex vs
                              > alveolo-palatal; tenuis vs aspirated vs ejectiv; and plain vs labialized.

                              Indeed you could!

                              > * Syntax changes. The fact that the "carrier" here is a single mantra and
                              > not the complete host language givs quite a bit of leeway that might not be
                              > otherwise available.

                              Depends; what sort of language & mantra? Is this one of those
                              quote-unquote "free variation of word order" deals?

                              > * Synonym variation, probably including arcaic and/or dialectal allomorphy.
                              > In order to be compatible with phoneme splits, this would probably have to
                              > specify wide semantic categories, while the latter would then further
                              > specify them down...

                              Again possible. However, this one would be somewhat more difficult to
                              keep language-agnostically meta; you would need some kind of
                              high-level theory about how one does these splits and which means
                              what, and I don't know of anything that could handle it.

                              (I would however be quite interested if you come up with one!)

                              > Alas, before I get the host language to an usable state (and I might need
                              > the conreligion too...) this isn't going to progress much.

                              *laugh* Endemic problem, eh?

                              But at least a couple of your proposed ideas could probably be taken
                              chunkwise. One of the nice features of my proposal is that it's highly
                              modular; you could implement one part of it mostly independent from
                              having to do the rest, and have it still be usable.

                              Your phoneme split point seems to me the most viable for such
                              treatment, and English is ripe for exploitation in this manner, given
                              how many things we don't mark at all. It would be somewhat hit-or-miss
                              whether a given word happened to have a phoneme that could be
                              segmented appropriately, but that could just lead to strategic word
                              choice by the participants... (e.g. picking something with 'r' in it
                              so as to be able to leverage a retroflex vs trilled vs etc etc etc
                              distinction)

                              - Sai
                            Your message has been successfully submitted and would be delivered to recipients shortly.