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

Forty Rules for Pronouncing Written Angosey

Expand Messages
  • Daniel Bowman
    A couple of months ago I discussed my intent to produce a YouTube video showing how to pronounce Angosey. As part of this, I decided to write an automatic
    Message 1 of 16 , Mar 14 8:09 PM
    • 0 Attachment
      A couple of months ago I discussed my intent to produce a YouTube video showing how to pronounce Angosey. As part of this, I decided to write an automatic Angosey orthography to IPA parser. Thus, I had to come up with unambiguous rules that mapped my spelling system to IPA. It turns out it takes forty rules to make this happen. The rules must be applied in order.

      Here's the rule list: http://www.unc.edu/~haksaeng/angosey/angosey_orthography_rules.pdf

      The first set deals with digraphs and trigraphs, followed by vowels and consonants. The final rule simply removes apostrophes.

      And, speaking of apostrophes: if used to distinguish trigraphs from sets of graphs and digraphs (e. g. aya != a'ya), they make this parser completely unambiguous. Unfortunately, I have not been in the habit of using apostrophes in this manner. Well, that starts now!

      The video slides are set up, after a bit more presentation editing I'll be ready to convert them to images and provide the audio. I'll let everyone know when that's up.

      Danny
    • Alex Fink
      ... Ah, that s one of the milder things forty rules could have meant! There are maybe five rules here beyond what would exist in a straightforward
      Message 2 of 16 , Mar 15 3:58 PM
      • 0 Attachment
        On Fri, 14 Mar 2014 23:09:22 -0400, Daniel Bowman <danny.c.bowman@...> wrote:

        >A couple of months ago I discussed my intent to produce a YouTube video showing how to pronounce Angosey. As part of this, I decided to write an automatic Angosey orthography to IPA parser. Thus, I had to come up with unambiguous rules that mapped my spelling system to IPA. It turns out it takes forty rules to make this happen. The rules must be applied in order.
        >
        >Here's the rule list: http://www.unc.edu/~haksaeng/angosey/angosey_orthography_rules.pdf
        >
        >The first set deals with digraphs and trigraphs, followed by vowels and consonants. The final rule simply removes apostrophes.

        Ah, that's one of the milder things "forty rules" could have meant! There are maybe five rules here beyond what would exist in a straightforward one-to-one romanisation with digraphs, at least with a maximalist analysis re diphthongs.

        Of these five, the "aya" rule does not actually appear necessary if the "ay" rule is arranged to precede the "ya" rule, unless there's a subtlety about syllabification this misses. Already I assume there's a subtlety of syllabification distinguishing [Ai.A] from [A.jA]: unless the first one actually has three nuclei [A.i.A] as written, these will sound very similar, perhaps with minor differences in length of transitions or something.

        And then the first four rules are the oddball allophony-type rules you've spoken of before here, which are great in the phonology-as-pure-patterning abstract, but I can't think of phonetic justifications for.

        Oh, and in rule 30, is that really a breathy-voiced [k] -- what does that mean, is it different to a breathy-voiced [g]? And is there really one isolated click written "p"? Is it somehow paralinguistic or confined to exclamations or whatnot?

        Alex
      • Daniel Bowman
        You are correct: the aya rule is redundant. I place that rule in front because it denotes an ambiguity in the current romanization system that I have to
        Message 3 of 16 , Mar 17 6:11 AM
        • 0 Attachment
          You are correct: the "aya" rule is redundant. I place that rule in front
          because it denotes an ambiguity in the current romanization system that I
          have to watch out for when I transform it into IPA. The issue is rare but
          arises due to the following:
          The emotive aspect suffix is "aya," [AiA].
          The partitive suffix is "ya" [jA].

          Now: nouns almost never end in "a" so the spelling "-aya" is seldom
          encountered - but it does exist.
          Since all verbs in the present tense end in "a", the emotive aspect "-aya"
          is quite common.
          This ambiguity is not an issue in Angosey script, where the emotive aspect
          is spelled using the first two glyphs on the first line and the partitive
          suffix is written using the 12th glyph from left on the first line.
          The document I'm referring to is this one:
          http://www.unc.edu/~haksaeng/angosey/angosey_script.pdf (same as what I
          sent you offlist).
          I will write up a nicer one soon.

          The difference between "aya" as emotive aspect and "aya" as partitive
          suffix is syllabification. Ay-a versus a-ya.

          As we've discussed offlist, the "oddball allophones" do require post-hoc
          conhistory justification. I will discuss some possibilities on list
          tonight, if I have a chance.

          Rule 30 is indeed a breathy voiced velar stop pronounced as Korean
          ssang-kyeok. And yes, there is an isolated click, and it is not
          paralinguistic. I passed on to you the trial YouTube slides, but I'll post
          them here as well:
          http://www.unc.edu/~haksaeng/angosey/angosey_alphabet_slides.pdf
          The click is always at the beginning of a syllable. There may be other
          regularities in its usage, I will have to check my lexicon. However, the
          click is one of the rarest phones in the language so there's not much to go
          on.


          2014-03-15 18:58 GMT-04:00 Alex Fink <000024@...>:

          > On Fri, 14 Mar 2014 23:09:22 -0400, Daniel Bowman <
          > danny.c.bowman@...> wrote:
          >
          > >A couple of months ago I discussed my intent to produce a YouTube video
          > showing how to pronounce Angosey. As part of this, I decided to write an
          > automatic Angosey orthography to IPA parser. Thus, I had to come up with
          > unambiguous rules that mapped my spelling system to IPA. It turns out it
          > takes forty rules to make this happen. The rules must be applied in order.
          > >
          > >Here's the rule list:
          > http://www.unc.edu/~haksaeng/angosey/angosey_orthography_rules.pdf
          > >
          > >The first set deals with digraphs and trigraphs, followed by vowels and
          > consonants. The final rule simply removes apostrophes.
          >
          > Ah, that's one of the milder things "forty rules" could have meant! There
          > are maybe five rules here beyond what would exist in a straightforward
          > one-to-one romanisation with digraphs, at least with a maximalist analysis
          > re diphthongs.
          >
          > Of these five, the "aya" rule does not actually appear necessary if the
          > "ay" rule is arranged to precede the "ya" rule, unless there's a subtlety
          > about syllabification this misses. Already I assume there's a subtlety of
          > syllabification distinguishing [Ai.A] from [A.jA]: unless the first one
          > actually has three nuclei [A.i.A] as written, these will sound very
          > similar, perhaps with minor differences in length of transitions or
          > something.
          >
          > And then the first four rules are the oddball allophony-type rules you've
          > spoken of before here, which are great in the phonology-as-pure-patterning
          > abstract, but I can't think of phonetic justifications for.
          >
          > Oh, and in rule 30, is that really a breathy-voiced [k] -- what does that
          > mean, is it different to a breathy-voiced [g]? And is there really one
          > isolated click written "p"? Is it somehow paralinguistic or confined to
          > exclamations or whatnot?
          >
          > Alex
          >
        • Garth Wallace
          ... AIUI, this is a universal about clicks in general: they are never in syllable codas. I have no idea why; it feels easier for me to pronounce a click at the
          Message 4 of 16 , Mar 17 9:21 AM
          • 0 Attachment
            On Mon, Mar 17, 2014 at 6:11 AM, Daniel Bowman <danny.c.bowman@...> wrote:
            > http://www.unc.edu/~haksaeng/angosey/angosey_alphabet_slides.pdf
            > The click is always at the beginning of a syllable.

            AIUI, this is a universal about clicks in general: they are never in
            syllable codas. I have no idea why; it feels easier for me to
            pronounce a click at the end of a word than before a vowel.
          • Daniel Bowman
            And I was ignorant of that rule when I came up with the Angosey click phonology! So it emerged more or less spontaneously. Thus I have to disagree with you -
            Message 5 of 16 , Mar 17 9:36 AM
            • 0 Attachment
              And I was ignorant of that rule when I came up with the Angosey click
              phonology! So it emerged more or less spontaneously. Thus I have to
              disagree with you - it is easier for me to pronounce a click (at least,
              this one) before a vowel than after one.


              2014-03-17 12:21 GMT-04:00 Garth Wallace <gwalla@...>:

              > On Mon, Mar 17, 2014 at 6:11 AM, Daniel Bowman <danny.c.bowman@...>
              > wrote:
              > > http://www.unc.edu/~haksaeng/angosey/angosey_alphabet_slides.pdf
              > > The click is always at the beginning of a syllable.
              >
              > AIUI, this is a universal about clicks in general: they are never in
              > syllable codas. I have no idea why; it feels easier for me to
              > pronounce a click at the end of a word than before a vowel.
              >
            • Alex Fink
              ... Ah, I was hoping you d post that! There is a nice reconstruction problem here, if you don t mind my stating it. (Correct me if I m wrong anywhere.) The
              Message 6 of 16 , Mar 17 11:49 AM
              • 0 Attachment
                On Mon, 17 Mar 2014 09:11:16 -0400, Daniel Bowman <danny.c.bowman@...> wrote:

                >This ambiguity is not an issue in Angosey script, where the emotive aspect
                >is spelled using the first two glyphs on the first line and the partitive
                >suffix is written using the 12th glyph from left on the first line.
                >The document I'm referring to is this one:
                >http://www.unc.edu/~haksaeng/angosey/angosey_script.pdf (same as what I
                >sent you offlist).
                >I will write up a nicer one soon.

                Ah, I was hoping you'd post that! There is a nice reconstruction problem here, if you don't mind my stating it. (Correct me if I'm wrong anywhere.)

                The first two blocks of four lines of characters, after the title, are the letters of two different though largely isomorphic Angosey scripts. The bottom two lines of each block are the _ceremonial_ form, used in calligraphic and decorative contexts; the top two lines are the day-to-day _runic_ form. The first and third lines are the basic forms of the characters, while the second and fourth are used to mark word boundaries.

                Many of the characters in the two scripts look similar enough that one wants to see them as conworld-internally deriving from a common ancestor. How might one reconstruct this ancestor?

                Unfortunately, as far as I know, graphology has not had its Junggrammatiker to formulate the principles of the regularities of orthographic change, and without that I'm at a loss as to whether any reconstruction I might posit is realistic or not. Does anyone here know anything about what pathways orthographic changes actually tends to follow, articulated any more precisely than the conlanging-101 advice "just draw your character a hundred times and the changes will suggest themselves"? As one important case, does anyone know to what extent changes affect whole natural classes of glyphs at once -- or even how to make sense out of that question?

                >Rule 30 is indeed a breathy voiced velar stop pronounced as Korean
                >ssang-kyeok.

                Ah, OK. As I understand Korean phonetics that's not quite breathy, then, though it's close: it's _faucalised_.
                https://en.wikipedia.org/wiki/Faucalized_voice
                Rather than the space between the glottal folds being larger (which is breathiness), the glottis is lowered so the space above the glottis itself is larger. Indeed apparently the Korean ssang- series is actually a little bit stiff, secondarily, i.e. the glottal folds are drawn closer together.

                >And yes, there is an isolated click, and it is not
                >paralinguistic. I passed on to you the trial YouTube slides, but I'll post
                >them here as well:
                >http://www.unc.edu/~haksaeng/angosey/angosey_alphabet_slides.pdf
                >The click is always at the beginning of a syllable. There may be other
                >regularities in its usage, I will have to check my lexicon. However, the
                >click is one of the rarest phones in the language so there's not much to go
                >on.

                As I responded offlist, one of my little crusades in phonology is that in natlangs that have them, clicks come in full systems with varying accompaniment i.e. phonation and manner, not just isolatedly. So given that Angosey has voiceless and voiced and sometimes faucalised stops, plus nasals, I would expect it to have alongside the tenuis /!\/ all or most of a voiced and a faucalised and a nasal version, and not inconceivably even another one or two.

                On the other hand, it struck me that your example has a /t!\/ cluster (where /t/ is an acceptably imprecise symbol for your [t`]). As one of the posited origins of clicks is from clusters with dorsals, and weak clicks have been recorded as phonetic realisations of such clusters in some natlangs. if you were wanting to get rid of the clicks (as you suggested offlist you might) there might be a way to reanalyse this as a /tk/ cluster where the clicking is subphonemic. You said there were also initial clicks, at least, but maybe this is underlying initial /tk/ -- is that phonotactically not outlandish?

                Alex
              • Daniel Bowman
                ... Correct. Sorry my responses are delayed - it s a busy time of the semester but I want to keep progress on this, however slow. I ve put clearer versions
                Message 7 of 16 , Mar 19 5:42 AM
                • 0 Attachment
                  2014-03-17 14:49 GMT-04:00 Alex Fink <000024@...>:

                  > On Mon, 17 Mar 2014 09:11:16 -0400, Daniel Bowman <
                  > danny.c.bowman@...> wrote:
                  >
                  > >This ambiguity is not an issue in Angosey script, where the emotive aspect
                  > >is spelled using the first two glyphs on the first line and the partitive
                  > >suffix is written using the 12th glyph from left on the first line.
                  > >The document I'm referring to is this one:
                  > >http://www.unc.edu/~haksaeng/angosey/angosey_script.pdf (same as what I
                  > >sent you offlist).
                  > >I will write up a nicer one soon.
                  >
                  > Ah, I was hoping you'd post that! There is a nice reconstruction problem
                  > here, if you don't mind my stating it. (Correct me if I'm wrong anywhere.)
                  >
                  > The first two blocks of four lines of characters, after the title, are the
                  > letters of two different though largely isomorphic Angosey scripts. The
                  > bottom two lines of each block are the _ceremonial_ form, used in
                  > calligraphic and decorative contexts; the top two lines are the day-to-day
                  > _runic_ form. The first and third lines are the basic forms of the
                  > characters, while the second and fourth are used to mark word boundaries.
                  >

                  Correct. Sorry my responses are delayed - it's a busy time of the semester
                  but I want to keep progress on this, however slow.

                  I've put clearer versions online:
                  http://www.unc.edu/~haksaeng/angosey/angosey_ceremonial_script.pdf
                  http://www.unc.edu/~haksaeng/angosey/angosey_runic_script.pdf


                  > Many of the characters in the two scripts look similar enough that one
                  > wants to see them as conworld-internally deriving from a common ancestor.
                  > How might one reconstruct this ancestor?
                  >

                  One possibility is that literacy was restricted to a very small class of
                  people: Temple scribes and royal clerks. A writing system was legislated
                  similar to how hangeul arose in the Choseon dynasty - one version for
                  calligraphy with brushes (ceremonial) and another for large volumes of
                  writing (runic). We can see relatively clear relationships between the two
                  in some cases (for example, "OU"), but divergence in others ("AY"), which
                  suggests some inconsistency - one thought I had was that written records
                  were (accidentally or intentionally) destroyed at least once. This would
                  require reconstruction of the writing system by the two different classes
                  (priestly and clerk), and imperfect recollection could lead to dissimilar
                  characters. However, Angosey spelling is invariant almost without
                  exception, which is hard to explain using the above scenario.


                  > Unfortunately, as far as I know, graphology has not had its
                  > Junggrammatiker to formulate the principles of the regularities of
                  > orthographic change, and without that I'm at a loss as to whether any
                  > reconstruction I might posit is realistic or not. Does anyone here know
                  > anything about what pathways orthographic changes actually tends to follow,
                  > articulated any more precisely than the conlanging-101 advice "just draw
                  > your character a hundred times and the changes will suggest themselves"?
                  > As one important case, does anyone know to what extent changes affect
                  > whole natural classes of glyphs at once -- or even how to make sense out of
                  > that question?
                  >
                  > >Rule 30 is indeed a breathy voiced velar stop pronounced as Korean
                  > >ssang-kyeok.
                  >
                  > Ah, OK. As I understand Korean phonetics that's not quite breathy, then,
                  > though it's close: it's _faucalised_.
                  > https://en.wikipedia.org/wiki/Faucalized_voice
                  > Rather than the space between the glottal folds being larger (which is
                  > breathiness), the glottis is lowered so the space above the glottis itself
                  > is larger. Indeed apparently the Korean ssang- series is actually a little
                  > bit stiff, secondarily, i.e. the glottal folds are drawn closer together.
                  >

                  You're correct, it should be faucalized. The wikipedia article indicates
                  that there's a couple of ways to represent it via IPA, I think I will keep
                  the way I have it now and put in a note when I write up the phonology.

                  >
                  > >And yes, there is an isolated click, and it is not
                  > >paralinguistic. I passed on to you the trial YouTube slides, but I'll
                  > post
                  > >them here as well:
                  > >http://www.unc.edu/~haksaeng/angosey/angosey_alphabet_slides.pdf
                  > >The click is always at the beginning of a syllable. There may be other
                  > >regularities in its usage, I will have to check my lexicon. However, the
                  > >click is one of the rarest phones in the language so there's not much to
                  > go
                  > >on.
                  >
                  > As I responded offlist, one of my little crusades in phonology is that in
                  > natlangs that have them, clicks come in full systems with varying
                  > accompaniment i.e. phonation and manner, not just isolatedly. So given
                  > that Angosey has voiceless and voiced and sometimes faucalised stops, plus
                  > nasals, I would expect it to have alongside the tenuis /!\/ all or most of
                  > a voiced and a faucalised and a nasal version, and not inconceivably even
                  > another one or two.
                  >
                  > On the other hand, it struck me that your example has a /t!\/ cluster
                  > (where /t/ is an acceptably imprecise symbol for your [t`]). As one of the
                  > posited origins of clicks is from clusters with dorsals, and weak clicks
                  > have been recorded as phonetic realisations of such clusters in some
                  > natlangs. if you were wanting to get rid of the clicks (as you suggested
                  > offlist you might) there might be a way to reanalyse this as a /tk/ cluster
                  > where the clicking is subphonemic. You said there were also initial
                  > clicks, at least, but maybe this is underlying initial /tk/ -- is that
                  > phonotactically not outlandish?
                  >

                  I think that is reasonable. One option is to have several dialects, one of
                  which has /tk/and others that have /t!\. In the /tk/ dialect, "peor" (an
                  example of a click-initial word) would be /tk/eor.


                  >
                  > Alex
                  >
                • Alex Fink
                  ... (Shame no-one else has bitten, though...) ... What is the value of au ? I didn t see it in your forty rules, so is it just [A.V]? ... Yeah, this seems a
                  Message 8 of 16 , Mar 19 8:06 AM
                  • 0 Attachment
                    On Wed, 19 Mar 2014 08:42:48 -0400, Daniel Bowman <danny.c.bowman@...> wrote:

                    >2014-03-17 14:49 GMT-04:00 Alex Fink <000024@...>:
                    >
                    >> On Mon, 17 Mar 2014 09:11:16 -0400, Daniel Bowman <
                    >> danny.c.bowman@...> wrote:
                    >>
                    >> >This ambiguity is not an issue in Angosey script, where the emotive aspect
                    >> >is spelled using the first two glyphs on the first line and the partitive
                    >> >suffix is written using the 12th glyph from left on the first line.
                    >> >The document I'm referring to is this one:
                    >> >http://www.unc.edu/~haksaeng/angosey/angosey_script.pdf (same as what I
                    >> >sent you offlist).
                    >> >I will write up a nicer one soon.
                    >>
                    >> Ah, I was hoping you'd post that! There is a nice reconstruction problem
                    >> here, if you don't mind my stating it.

                    (Shame no-one else has bitten, though...)

                    >Correct. Sorry my responses are delayed - it's a busy time of the semester
                    >but I want to keep progress on this, however slow.
                    >
                    >I've put clearer versions online:
                    >http://www.unc.edu/~haksaeng/angosey/angosey_ceremonial_script.pdf
                    >http://www.unc.edu/~haksaeng/angosey/angosey_runic_script.pdf

                    What is the value of "au"? I didn't see it in your forty rules, so is it just [A.V]?

                    >> Many of the characters in the two scripts look similar enough that one
                    >> wants to see them as conworld-internally deriving from a common ancestor.
                    >> How might one reconstruct this ancestor?
                    >
                    >One possibility is that literacy was restricted to a very small class of
                    >people: Temple scribes and royal clerks. A writing system was legislated
                    >similar to how hangeul arose in the Choseon dynasty - one version for
                    >calligraphy with brushes (ceremonial) and another for large volumes of
                    >writing (runic). We can see relatively clear relationships between the two
                    >in some cases (for example, "OU"), but divergence in others ("AY"), which
                    >suggests some inconsistency - one thought I had was that written records
                    >were (accidentally or intentionally) destroyed at least once. This would
                    >require reconstruction of the writing system by the two different classes
                    >(priestly and clerk), and imperfect recollection could lead to dissimilar
                    >characters. However, Angosey spelling is invariant almost without
                    >exception, which is hard to explain using the above scenario.

                    Yeah, this seems a stretch to me. What it seems you would need for this to work is not the simple destruction of some written records, but the destruction of the entire tradition of writing. If the aftermath of some tumult was merely that the great libraries or archives or whatever had burned down, the literate would still go on writing tomorrow as they had yesterday. On the other hand, if things got bad enough that writing simply ceased to be practiced, as happened in Mycenaean Greece in the Bronze Age collapse ~1200BC, I wouldn't expect the society's lot to turn around quickly enough for imperfect memory of the original writers to be relevant, but rather that these writers would be long dead. E.g. in Greece the new script after a hiatus of some three centuries was entirely different even in principle to the old, the Phoenecian-derived alphabet supplanting the old defective-syllabary-with-scattered-logograms Linear B. (Though to be honest I suppose if there wère gaps of your sort in the historic record they might be overlookable.)

                    I expect there are less violent ways to achieve your ends! Ordinary graphic evolution should be up to the job. I will focus on the medial forms, because the initial forms look to be fairly clearly recently arisen (maybe from fusion of an old interpunct?).

                    For instance, there are some correspondences (<a, ay, d, h, hh, k, q, n, r, s> might be instances) in which the ceremonial has a little (or in <a, r, s> not so little) counterclockwise curl at the bottom without a counterpart in the runic. It could be that the historical shapes were more like the runic ones, and these curls were added as like a filter in ceremonialisation, replacing forms that otherwise woulda just had a straight vertical line (or something) there.

                    In the other direction, cases where the runic is simpler than the ceremonial could result from case where the historical shapes were more like the ceremonial ones but a cursive, heavily reduced variant displaced the original in the runic script. Something like this could be behind <i, o, eo, d>, maaaybe <g, m> etc.

                    Or, some developments in one script could have been adopted to differentiate two letters, which was done differently or not at all in the other. This is e.g. what the identity of <eo, h> in runic but not in ceremonial suggests to me. Maybe in the proto-script <eo, h> were identical or fell together, as more or less a simple line, and this is preserved in runic. In early ceremonial, after the curl rule acted, maybe they both looked like current ceremonial <h>. But it was later felt they had to be differentiated in ceremonial, and this was done by making the end of the bottom curl of <eo> longer, bringing it up again through the stem.

                    Proto-<hh> could have been a diacriticked form of <h> (maybe even earlier they weren't distinguished in writing?): maybe it was a shorter horizontal on top of a longer one. Then in runic the two strokes got connected up, while in ceremonial the diacritic stayed separate, ending up on the left, and its presence was enough to prevent the action of the top curling rule. And <k, kh> could have identical in the proto-script, with a more <k>-like form, if the differentiation process went differently in the two scripts: <kh> was distinguished in runic by adding this curly diacritic on the right but in ceremonial by whether the top was curly or straight. Maybe those two ceremonial forms were originally just alloglyphs, like Roman U and V.

                    And so on.

                    Alex
                  • Daniel Bowman
                    What is the value of au ? I didn t see it in your forty rules, so is it just [A.V]? No - that s a mistake, and the consequence of not having one centralized
                    Message 9 of 16 , Mar 19 9:09 AM
                    • 0 Attachment
                      What is the value of "au"? I didn't see it in your forty rules, so is it
                      just [A.V]?

                      No - that's a mistake, and the consequence of not having one centralized
                      resource for me to refer to. This current project will rectify that, at
                      least.
                      AU is a letter in its own right, and needs to be included in the forty
                      rules.

                      >
                      >
                      > Yeah, this seems a stretch to me. What it seems you would need for this
                      > to work is not the simple destruction of some written records, but the
                      > destruction of the entire tradition of writing. If the aftermath of some
                      > tumult was merely that the great libraries or archives or whatever had
                      > burned down, the literate would still go on writing tomorrow as they had
                      > yesterday. On the other hand, if things got bad enough that writing simply
                      > ceased to be practiced, as happened in Mycenaean Greece in the Bronze Age
                      > collapse ~1200BC, I wouldn't expect the society's lot to turn around
                      > quickly enough for imperfect memory of the original writers to be relevant,
                      > but rather that these writers would be long dead. E.g. in Greece the new
                      > script after a hiatus of some three centuries was entirely different even
                      > in principle to the old, the Phoenecian-derived alphabet supplanting the
                      > old defective-syllabary-with-scattered-logograms Linear B. (Though to be
                      > honest I suppose if there wère gaps of your sort in the historic record
                      > they might be overlookable.)
                      >
                      > I expect there are less violent ways to achieve your ends! Ordinary
                      > graphic evolution should be up to the job. I will focus on the medial
                      > forms, because the initial forms look to be fairly clearly recently arisen
                      > (maybe from fusion of an old interpunct?).
                      >

                      Yes, I think medial precedes initial. That is consistent with the "true
                      history" of the language, if I recall correctly I experimented with no
                      interpuncts at first. However, particularly since I was not fluent at all,
                      the writing was hard to decipher. I was lackadaisically studying Arabic
                      writing systems at the time, and that's where I got the idea of an initial
                      form. I found final forms to be overly burdensome and not really necessary
                      for readability.


                      >
                      > For instance, there are some correspondences (<a, ay, d, h, hh, k, q, n,
                      > r, s> might be instances) in which the ceremonial has a little (or in <a,
                      > r, s> not so little) counterclockwise curl at the bottom without a
                      > counterpart in the runic. It could be that the historical shapes were more
                      > like the runic ones, and these curls were added as like a filter in
                      > ceremonialisation, replacing forms that otherwise woulda just had a
                      > straight vertical line (or something) there.
                      >

                      In "real history" developed ceremonial first, and runic as a simplified
                      form. In my conhistory, Angosey was a ritual language first, so I think
                      the "ceremonial" script should be the first alphabetic system to emerge.
                      However, it's a bit ornate and difficult to write, so a linearized,
                      simplified system was developed later.

                      >
                      > In the other direction, cases where the runic is simpler than the
                      > ceremonial could result from case where the historical shapes were more
                      > like the ceremonial ones but a cursive, heavily reduced variant displaced
                      > the original in the runic script. Something like this could be behind <i,
                      > o, eo, d>, maaaybe <g, m> etc.
                      >

                      Yes, I think that is reasonable.


                      >
                      > Or, some developments in one script could have been adopted to
                      > differentiate two letters, which was done differently or not at all in the
                      > other. This is e.g. what the identity of <eo, h> in runic but not in
                      > ceremonial suggests to me. Maybe in the proto-script <eo, h> were
                      > identical or fell together, as more or less a simple line, and this is
                      > preserved in runic. In early ceremonial, after the curl rule acted, maybe
                      > they both looked like current ceremonial <h>. But it was later felt they
                      > had to be differentiated in ceremonial, and this was done by making the end
                      > of the bottom curl of <eo> longer, bringing it up again through the stem.
                      >

                      Yes, and keep in mind "eo" is a late addition to my phonetic inventory in
                      "real history" - I added it some 4 years after locking in the basic
                      phonetics. So it makes sense that originally there was one symbol in the
                      proto script.


                      >
                      > Proto-<hh> could have been a diacriticked form of <h> (maybe even earlier
                      > they weren't distinguished in writing?): maybe it was a shorter horizontal
                      > on top of a longer one. Then in runic the two strokes got connected up,
                      > while in ceremonial the diacritic stayed separate, ending up on the left,
                      > and its presence was enough to prevent the action of the top curling rule.
                      > And <k, kh> could have identical in the proto-script, with a more <k>-like
                      > form, if the differentiation process went differently in the two scripts:
                      > <kh> was distinguished in runic by adding this curly diacritic on the right
                      > but in ceremonial by whether the top was curly or straight. Maybe those
                      > two ceremonial forms were originally just alloglyphs, like Roman U and V.
                      >

                      Proto <<hh>> is indeed a derivation of <h>, and earlier writing will not
                      show any distinguishing symbol because they were not seen as distinct
                      phonemes.

                      >
                      > And so on.
                      >

                      Interesting analysis! Funny to see all these patterns in this writing
                      system.

                      >
                      > Alex
                      >
                    • Alex Fink
                      ... I don t think that necessarily follows. If Angosey was first a ritual language, there will have been another language, let me call it language B, which
                      Message 10 of 16 , Mar 19 10:31 AM
                      • 0 Attachment
                        On Wed, 19 Mar 2014 12:09:06 -0400, Daniel Bowman <danny.c.bowman@...> wrote:

                        >In "real history" developed ceremonial first, and runic as a simplified
                        >form. In my conhistory, Angosey was a ritual language first, so I think
                        >the "ceremonial" script should be the first alphabetic system to emerge.
                        >However, it's a bit ornate and difficult to write, so a linearized,
                        >simplified system was developed later.

                        I don't think that necessarily follows. If Angosey was first a ritual language, there will have been another language, let me call it language B, which the same people used when not engaged in ritual. Then the following scenario seems plausible to me in the abstract: The proto-Angosey alphabet was first used for language B, and in form it could well have been closer to runic than ceremonial. Then the ritual use of Angosey started. Befittingly, to write this new ceremonial language a ceremonial form of the script was devised; independently, language B continued still being written in the plain script. Later, as the usage of Angosey spread, people started applying the plain script to Angosey as well, when used in plain contexts, and this is the origin of the runic script.

                        Alex
                      • And Rosta
                        ... The probable-conlang Damin apart, I can t think of a counterexample to that. But can we be sure this not just (a relic of) an areal feature? In Livagian,
                        Message 11 of 16 , Mar 19 1:05 PM
                        • 0 Attachment
                          Alex Fink, On 17/03/2014 18:49:
                          > As I responded offlist, one of my little crusades in phonology is
                          > that in natlangs that have them, clicks come in full systems with
                          > varying accompaniment i.e. phonation and manner, not just
                          > isolatedly.

                          The probable-conlang Damin apart, I can't think of a counterexample to that. But can we be sure this not just (a relic of) an areal feature?

                          In Livagian, whose phonology is neither patently engelangy nor based on natlang models, there are four clicks, one for each of the four Place series. They are phonetically nasal. Phonologically they behave like obstruent nasals, which means they don't bear tone.

                          --And.
                        • Alex Fink
                          ... Incidentally, on hearing that Angosey started as a ritual language, I thought of a comparison to Damin as well. If early Angosey phonology was had
                          Message 12 of 16 , Mar 19 4:00 PM
                          • 0 Attachment
                            On Wed, 19 Mar 2014 20:05:29 +0000, And Rosta <and.rosta@...> wrote:

                            >Alex Fink, On 17/03/2014 18:49:
                            >> As I responded offlist, one of my little crusades in phonology is
                            >> that in natlangs that have them, clicks come in full systems with
                            >> varying accompaniment i.e. phonation and manner, not just
                            >> isolatedly.
                            >
                            >The probable-conlang Damin apart, I can't think of a counterexample to that.

                            Incidentally, on hearing that Angosey started as a ritual language, I thought of a comparison to Damin as well. If early Angosey phonology was had invented ~ artificial elements in the same way, then its quirks of are maybe not so bothersome. (Modulo evolution.)

                            >But can we be sure this not just (a relic of) an areal feature?

                            Probably not: of course the presence of any clicks at all, beyond paralinguistic phenomena and Damin, is already areal and we have only one data point.

                            But the analogy with pulmonic egressive stops has some weight for me. Is the fact that they often come in several phonation series an areal feature?

                            I usually blame the design of the IPA at this point: the click letters are designed in a way that doesn't draw attention to click accompaniments. If the IPA were more like Kirschenbaum this way and I could write [|\_v] as "[d_!]" and [O\_v~] as "[m_!]" and so on, one imagines click systems in the typical conlang would look different.

                            >In Livagian, whose phonology is neither patently engelangy nor based on natlang models, there are four clicks, one for each of the four Place series. They are phonetically nasal. Phonologically they behave like obstruent nasals, which means they don't bear tone.

                            What sorts of pulmonic egressive plosives and nasals are there in Livagian? And is it worth asking what they all are underlyingly (ISTR there's lots of fusion between the phonemic and surface levels)?

                            Alex
                          • And Rosta
                            ... Depends on the number of series. Four would be areal; three, probably; two, not. I m sure you can answer your own question better than I can, tho. Of
                            Message 13 of 16 , Mar 19 5:51 PM
                            • 0 Attachment
                              Alex Fink, On 19/03/2014 23:00:
                              > On Wed, 19 Mar 2014 20:05:29 +0000, And Rosta <and.rosta@...> wrote:
                              >
                              >> Alex Fink, On 17/03/2014 18:49:
                              >>> As I responded offlist, one of my little crusades in phonology is
                              >>> that in natlangs that have them, clicks come in full systems with
                              >>> varying accompaniment i.e. phonation and manner, not just
                              >>> isolatedly.
                              >> But can we be sure this not just (a relic of) an areal feature?
                              >
                              > Probably not: of course the presence of any clicks at all, beyond
                              > paralinguistic phenomena and Damin, is already areal and we have
                              > only one data point. But the analogy with pulmonic egressive stops
                              > has some weight for me. Is the fact that they often come in several
                              > phonation series an areal feature?

                              Depends on the number of series. Four would be areal; three, probably; two, not. I'm sure you can answer your own question better than I can, tho. Of course distinctions described as based on phonation aren't always really based on phonation, and phonation distinctions interact with prosodic distinctions like gemination, but the point you intend to make is simply that pulmonic egressive stops tend to come in multiple series (more so than, say, fricatives).

                              > I usually blame the design of the IPA at this point: the click
                              > letters are designed in a way that doesn't draw attention to click
                              > accompaniments. If the IPA were more like Kirschenbaum this way and I
                              > could write [|\_v] as "[d_!]" and [O\_v~] as "[m_!]" and so on, one
                              > imagines click systems in the typical conlang would look different.

                              You may be right, but another reason for small click inventories is the difficulty of learning to utter big ones. Livagian's clicks are (extrafictionally) nasal because I, like most people who don't speak click languages, find it much easier to integrate nasal clicks into a multisegmental speech stream.

                              >> In Livagian, whose phonology is neither patently engelangy nor
                              >> based on natlang models, there are four clicks, one for each of the
                              >> four Place series. They are phonetically nasal. Phonologically they
                              >> behave like obstruent nasals, which means they don't bear tone.
                              >
                              > What sorts of pulmonic egressive plosives and nasals are there in
                              > Livagian?

                              Primary allophones of /p t k q/ are [b d g ?].
                              Primary allophones of /m n N l / are [m n N l].
                              Click counterparts are bilabial, dental, lateral and retroflex (with palatal an allophone of retroflex).

                              There are also series of
                              * voiceless fricatives (including [T], I'm pleased to say)
                              * voiced fricatives (including [D], ditto)
                              * tonic /m n N l/
                              * plus /r/, plus 5 semivowels, plus 10 vowels

                              So there's just one series of pulmonic egressive plosives, tho there are two fricative series distinguished by phonation.

                              > And is it worth asking what they all are underlyingly (ISTR
                              > there's lots of fusion between the phonemic and surface levels)?

                              /qp, qt, qk/ give preglottalized
                              /pq, tq, kq/ give ejective
                              /pf, tT, kh/ give voiceless aspirated
                              /qf, qT, qh/ give voiceless affricates
                              /pb, td, kg/ give voiced affricates
                              /pqf, tqT, kqh/ give ejective affricates

                              iirc, /pg/ and /ph/ give voiced and voiceless bilabial trills and /pqh/ a velaric egressive bilabial trill.

                              The phonology has been stable for years. Perhaps when time permits I will essay in one of them online pdfs a public description of it. (It is very simple.)

                              --And.
                            • Daniel Bowman
                              When I wrote up the Angosey phonology, I was a little unclear on how click consonants worked. It looks like p is indeed an alveolar click with a retroflex
                              Message 14 of 16 , Mar 19 6:37 PM
                              • 0 Attachment
                                When I wrote up the Angosey phonology, I was a little unclear on how click
                                consonants worked. It looks like "p" is indeed an alveolar click with a
                                retroflex stop. This is based on me clicking away at my laptop whilst my
                                wife attempts to watch American Idol.



                                > Incidentally, on hearing that Angosey started as a ritual language, I
                                > thought of a comparison to Damin as well. If early Angosey phonology was
                                > had invented ~ artificial elements in the same way, then its quirks of are
                                > maybe not so bothersome. (Modulo evolution.)
                                >

                                > >But can we be sure this not just (a relic of) an areal feature?
                                >

                                It certainly could be. Angosey is placed in a region with high linguistic
                                diversity. It could be that surrounding branches of the proto-Angosey
                                language tree have a richer click inventory. This never occurred to me,
                                but I can entertain it. I assume it is reasonable to have a language that
                                starts with several click accompaniments and later loses all except /t/?

                                >
                                >
                                > I usually blame the design of the IPA at this point: the click letters are
                                > designed in a way that doesn't draw attention to click accompaniments. I


                                Yes, this confused me, and that's why my first transcription lacked them.
                              • Jörg Rhiemeier
                                Hallo conlangers! ... I think clicks as a whole *are* an areal feature of sub-Equatorial Africa, even if the area was shattered by the spread of Bantu
                                Message 15 of 16 , Mar 20 5:53 AM
                                • 0 Attachment
                                  Hallo conlangers!

                                  On Thu, 20 Mar 2014 00:51:12 +0000 And Rosta writes:

                                  > Alex Fink, On 19/03/2014 23:00:
                                  > > On Wed, 19 Mar 2014 20:05:29 +0000, And Rosta <[log in to unmask]> wrote:
                                  > >
                                  > >> Alex Fink, On 17/03/2014 18:49:
                                  > >>> As I responded offlist, one of my little crusades in phonology is
                                  > >>> that in natlangs that have them, clicks come in full systems with
                                  > >>> varying accompaniment i.e. phonation and manner, not just
                                  > >>> isolatedly.
                                  > >> But can we be sure this not just (a relic of) an areal feature?
                                  > >
                                  > > Probably not: of course the presence of any clicks at all, beyond
                                  > > paralinguistic phenomena and Damin, is already areal and we have
                                  > > only one data point. But the analogy with pulmonic egressive stops
                                  > > has some weight for me. Is the fact that they often come in several
                                  > > phonation series an areal feature?
                                  >
                                  > Depends on the number of series. Four would be areal; three, probably;
                                  > two, not. I'm sure you can answer your own question better than I can,
                                  > tho.

                                  I think clicks as a whole *are* an areal feature of sub-Equatorial
                                  Africa, even if the area was shattered by the spread of Bantu
                                  languages (most of which don't have clicks). We know nothing
                                  about the languages spoken, for instance, in the Congo, before the
                                  Bantu family took over, but I consider it likely that many of them
                                  did have clicks.

                                  Alas, we have no data on how clicks evolve from other speech sounds.
                                  The "Khoisan" languages (actually three families and two isolates
                                  of which we do not know whether they are related to each other)
                                  do not have clickless cognates; and even in Dahalo and those Bantu
                                  languages which have clicks, the etymologies of the words which
                                  include clicks are apparently difficult, and nobody so far managed
                                  to trace back the clicks to Proto-Cushitic or Proto-Bantu phonemes,
                                  respectively. At least, that seems to be the state of research,
                                  judging from what I have seen, but I am no expert on these languages.

                                  One hypothesis is that clicks are related to the coarticulated
                                  consonants of the /kp/ type that are so common in western Africa,
                                  but apparently, nobody knows for sure.

                                  --
                                  ... brought to you by the Weeping Elf
                                  http://www.joerg-rhiemeier.de/Conlang/index.html
                                  "Bêsel asa Éam, a Éam atha cvanthal a cvanth atha Éamal." - SiM 1:1
                                • Paul Roser
                                  ... While I haven t read very much from Khoisan specialists regarding the potential genesis of clicks, I had read a few scattered articles years back regarding
                                  Message 16 of 16 , Mar 20 9:11 PM
                                  • 0 Attachment
                                    On Thu, 20 Mar 2014 13:53:35 +0100, Jörg Rhiemeier <joerg_rhiemeier@...> wrote:

                                    >Hallo conlangers!
                                    >
                                    >On Thu, 20 Mar 2014 00:51:12 +0000 And Rosta writes:
                                    >
                                    >> Alex Fink, On 19/03/2014 23:00:
                                    >> > On Wed, 19 Mar 2014 20:05:29 +0000, And Rosta <[log in to unmask]> wrote:
                                    >> >
                                    >> >> Alex Fink, On 17/03/2014 18:49:
                                    >> >>> As I responded offlist, one of my little crusades in phonology is
                                    >> >>> that in natlangs that have them, clicks come in full systems with
                                    >> >>> varying accompaniment i.e. phonation and manner, not just
                                    >> >>> isolatedly.
                                    >> >> But can we be sure this not just (a relic of) an areal feature?
                                    >> >
                                    >> > Probably not: of course the presence of any clicks at all, beyond
                                    >> > paralinguistic phenomena and Damin, is already areal and we have
                                    >> > only one data point. But the analogy with pulmonic egressive stops
                                    >> > has some weight for me. Is the fact that they often come in several
                                    >> > phonation series an areal feature?
                                    >>
                                    >> Depends on the number of series. Four would be areal; three, probably;
                                    >> two, not. I'm sure you can answer your own question better than I can,
                                    >> tho.
                                    >
                                    >I think clicks as a whole *are* an areal feature of sub-Equatorial
                                    >Africa, even if the area was shattered by the spread of Bantu
                                    >languages (most of which don't have clicks). We know nothing
                                    >about the languages spoken, for instance, in the Congo, before the
                                    >Bantu family took over, but I consider it likely that many of them
                                    >did have clicks.
                                    >
                                    >Alas, we have no data on how clicks evolve from other speech sounds.
                                    >The "Khoisan" languages (actually three families and two isolates
                                    >of which we do not know whether they are related to each other)
                                    >do not have clickless cognates; and even in Dahalo and those Bantu
                                    >languages which have clicks, the etymologies of the words which
                                    >include clicks are apparently difficult, and nobody so far managed
                                    >to trace back the clicks to Proto-Cushitic or Proto-Bantu phonemes,
                                    >respectively. At least, that seems to be the state of research,
                                    >judging from what I have seen, but I am no expert on these languages.
                                    >
                                    >One hypothesis is that clicks are related to the coarticulated
                                    >consonants of the /kp/ type that are so common in western Africa,
                                    >but apparently, nobody knows for sure.

                                    While I haven't read very much from Khoisan specialists regarding the potential genesis of clicks, I had read a few scattered articles years back regarding something like glide hardening in South African Bantu languages where, for example, /sw/ and /tw/ would 'harden' to [sgw] and [tkw], and since a number of Khoisan languages (notably Taa, N|uu, and Zhu|'oansi, if memory serves) have pharyngealized vowels, I could see something like 'pharyngeal hardening' leading to the genesis of an initial series or three of clicks, with more series being added through cluster reduction or other mechanisms.

                                    Scungric has two series of clicks, voiced and voiceless nasalized and are probably the 'fortis' counterparts of plain voiced and voiceless nasals, though I haven't thought through all of the details regarding how fortition relates to the non-nasal consonants.

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