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

Re: The w and hw tengwar

Expand Messages
  • Måns Björkman
    I just thought I d share my thoughts on this interesting discussion. ... People have always used fonts in ways that those who know better object to. Just think
    Message 1 of 14 , Dec 8, 2004
      I just thought I'd share my thoughts on this interesting discussion.


      --- In elfscript@yahoogroups.com, "j_mach_wust" <machhezan@g...> wrote
      (in reply to Johan Winge [>>]):

      > Many tengwar fonts distinguish e.g. three-dot a-tehtar from
      > (unattested) circumflex/caron a-tehtar, and you're also going to adopt
      > the mistaken óre corrected to a rómen. That correction allows an
      > accurate representation of some details in Tolkien's tengwar samples,
      > along with the distinction of several "lambe nuquerna" signs. However,
      > I think the inherent danger of such allographic options is that people
      > are going to distinguish them. I think I've seen modes where the three
      > dot tehtar were opposed to the circumflex/caron. I would say: this is
      > very un-Tolkienian; they would say: it's the fonts!

      People have always used fonts in ways that those who know better
      object to. Just think of the widespread use of foot and inch signs ('
      and ") as quotation marks! And abuse of Tengwar fonts was known long
      before Dan Smith created Tengwar Quenya. I don't think we should
      remove options for those knowledgeable to use them, just to prevent
      the less knowledgeable from making mistakes. With the keyboard layout
      used in most Tengwar fonts, they have ample opportunities to do that
      anyway.


      > > In the future, things will be different, I hope, imagine, and plan,
      > > but for the moment we are not there yet.
      >
      > That's how it is. At least there's already a tengwar font with italic
      > and bold versions!

      Indeed. And a good-looking Tengwar font, too!


      > [...]
      > However, I have a different opinion about the relation between DTS 17
      > and 18: I think that the forms of DTS 17 are more elemental, closer to
      > the underlying ideal, whereas the forms of DTS 18 are less elemental
      > because they're much more conditioned by the pen movements. Therefore,
      > I'm not at all surprised that the "nuquernar" tengwar of DTS 17 are
      > more similar to lambe and rómen. And I don't want to dismiss the
      > samples of DTS 17 because I think they're very important to my
      > argumentation.

      While I do agree that the "decorated verse-hand" of DTS 18 makes it a
      questionable source for the basic shapes of the letters, I don't think
      DTS 17 is more reliable. The vast majority of Tengwar samples
      published to date contains letter shapes consistent with the
      "book-hand", more or less. Why then assume that the "pointed" style
      should display the most elemental forms of the letters, rather than a
      curved one?

      In the case of the /w/-/W/ letters, I see many issues with regarding
      these as turned <lambe> and <rómen>. Firstly, the only character that
      is actually in shape an upside-down version of another character is
      the *<rómen nuquerna> of DTS 17. In DTS 18, the letter is clearly
      distinct from <rómen>. Even if the letter-shapes of DTS 18 are more
      "conditioned by the pen movements", it is (as Johan Winge notes) no
      more difficult to write an turned <rómen> than an ordinary one.
      Secondly, neither turned letter has a close phonological relationship
      with its normal form. If <rómen> was actually *used* for /w/ in these
      texts, I would find it easier to believe that its turned form was used
      for /W/.

      With regard to the Anglo-saxon /W/ sign: I agree with Johan Winge that
      this is probably a <rómen> with raised telco. Regardless wether or not
      the extended series is applicable to the additional tengwar, raised
      stems do represent fricatives in this mode. I based this character on
      <rómen> in Tengwar Parmaite, but alas! I missed the fact that the bow
      is connected to the lower part of the stem.

      Regards,

      Måns
    • Måns Björkman
      I just thought I d share my thoughts on this interesting discussion. ... People have always used fonts in ways that those who know better object to. Just think
      Message 2 of 14 , Dec 8, 2004
        I just thought I'd share my thoughts on this interesting discussion.


        --- In elfscript@yahoogroups.com, "j_mach_wust" <machhezan@g...> wrote
        (in reply to Johan Winge [>>]):

        > Many tengwar fonts distinguish e.g. three-dot a-tehtar from
        > (unattested) circumflex/caron a-tehtar, and you're also going to adopt
        > the mistaken óre corrected to a rómen. That correction allows an
        > accurate representation of some details in Tolkien's tengwar samples,
        > along with the distinction of several "lambe nuquerna" signs. However,
        > I think the inherent danger of such allographic options is that people
        > are going to distinguish them. I think I've seen modes where the three
        > dot tehtar were opposed to the circumflex/caron. I would say: this is
        > very un-Tolkienian; they would say: it's the fonts!

        People have always used fonts in ways that those who know better
        object to. Just think of the widespread use of foot and inch signs ('
        and ") as quotation marks! And abuse of Tengwar fonts was known long
        before Dan Smith created Tengwar Quenya. I don't think we should
        remove options for those knowledgeable to use them, just to prevent
        the less knowledgeable from making mistakes. With the keyboard layout
        used in most Tengwar fonts, they have ample opportunities to do that
        anyway.


        > > In the future, things will be different, I hope, imagine, and plan,
        > > but for the moment we are not there yet.
        >
        > That's how it is. At least there's already a tengwar font with italic
        > and bold versions!

        Indeed. And a good-looking Tengwar font, too!


        > [...]
        > However, I have a different opinion about the relation between DTS 17
        > and 18: I think that the forms of DTS 17 are more elemental, closer to
        > the underlying ideal, whereas the forms of DTS 18 are less elemental
        > because they're much more conditioned by the pen movements. Therefore,
        > I'm not at all surprised that the "nuquernar" tengwar of DTS 17 are
        > more similar to lambe and rómen. And I don't want to dismiss the
        > samples of DTS 17 because I think they're very important to my
        > argumentation.

        While I do agree that the "decorated verse-hand" of DTS 18 makes it a
        questionable source for the basic shapes of the letters, I don't think
        DTS 17 is more reliable. The vast majority of Tengwar samples
        published to date contains letter shapes consistent with the
        "book-hand", more or less. Why then assume that the "pointed" style
        should display the most elemental forms of the letters, rather than a
        curved one?

        In the case of the /w/-/W/ letters, I see many issues with regarding
        these as turned <lambe> and <rómen>. Firstly, the only character that
        is actually in shape an upside-down version of another character is
        the *<rómen nuquerna> of DTS 17. In DTS 18, the letter is clearly
        distinct from <rómen>. Even if the letter-shapes of DTS 18 are more
        "conditioned by the pen movements", it is (as Johan Winge notes) no
        more difficult to write an turned <rómen> than an ordinary one.
        Secondly, neither turned letter has a close phonological relationship
        with its normal form. If <rómen> was actually *used* for /w/ in these
        texts, I would find it easier to believe that its turned form was used
        for /W/.

        With regard to the Anglo-saxon /W/ sign: I agree with Johan Winge that
        this is probably a <rómen> with raised telco. Regardless wether or not
        the extended series is applicable to the additional tengwar, raised
        stems do represent fricatives in this mode. I based this character on
        <rómen> in Tengwar Parmaite, but alas! I missed the fact that the bow
        is connected to the lower part of the stem.

        Regards,

        Måns
      • j_mach_wust
        ... Not the most elemental, for sure, but closer to the underlying ideal, that is, more elemental (perhaps I ve misused this word). Or perhaps I shouldn t call
        Message 3 of 14 , Dec 9, 2004
          I wrote:

          > > However, I have a different opinion about the relation between DTS
          > > 17 and 18: I think that the forms of DTS 17 are more elemental,
          > > closer to the underlying ideal, whereas the forms of DTS 18 are
          > > less elemental because they're much more conditioned by the pen
          > > movements. Therefore, I'm not at all surprised that the
          > > "nuquernar" tengwar of DTS 17 are more similar to lambe and rómen.
          > > And I don't want to dismiss the samples of DTS 17 because I think
          > > they're very important to my argumentation.

          Måns Björkman replied:

          > While I do agree that the "decorated verse-hand" of DTS 18 makes it
          > a questionable source for the basic shapes of the letters, I don't
          > think DTS 17 is more reliable. The vast majority of Tengwar samples
          > published to date contains letter shapes consistent with the
          > "book-hand", more or less. Why then assume that the "pointed" style
          > should display the most elemental forms of the letters, rather than
          > a curved one?

          Not the most elemental, for sure, but closer to the underlying ideal,
          that is, more elemental (perhaps I've misused this word). Or perhaps I
          shouldn't call it an ideal at all, but rather the very core of the
          letters shape, the required form that distinguishes it from other
          letters. The "book-hand" is very close to them, the "decorated
          verse-hand" of DTS 18 is less close to them, and the angular forms of
          DTS 17 are again closer to them.

          I think this explains better why the rómen-like form of DTS 17 than
          your explanation in message #73 (cited by Johan Winge in #4453)
          because your explanation assumes a reanalysis of the form, that is, an
          additional hypothetical step, whereas I assume (in message #4456) that
          the forms are basically identical.

          > In the case of the /w/-/W/ letters, I see many issues with regarding
          > these as turned <lambe> and <rómen>. Firstly, the only character
          > that is actually in shape an upside-down version of another
          > character is the *<rómen nuquerna> of DTS 17. In DTS 18, the letter
          > is clearly distinct from <rómen>. Even if the letter-shapes of DTS
          > 18 are more "conditioned by the pen movements", it is (as Johan
          > Winge notes) no more difficult to write an turned <rómen> than an
          > ordinary one.

          I don't think you can prove this. I do think it would be possible to
          prove it based on an analysis of the shapes in DTS 18. You'd have to
          show that the same shapes that a true reversed rómen would require are
          found in DTS 18. I don't think they are found.

          On the other hand, it's quite easy to show that there are other
          letters in DTS 18 that have a peculiar form "conditioned by the pen
          movements", e.g. lambe.

          > Secondly, neither turned letter has a close
          > phonological relationship with its normal form. If <rómen> was
          > actually *used* for /w/ in these texts, I would find it easier to
          > believe that its turned form was used for /W/.

          You're saying that the "Bombadil-mode" w and wh tengwar are not
          reversed forms of lambe and rómen but may accidently (because of
          reanalysis) be replaced by reversed lambe and rómen. For unknown
          reasons, Tolkien stopped using them.

          I'm saying that the "Bombadil-mode" w and wh tengwar are reversed
          forms of lambe and rómen but may accidently (conditioned by pen
          movement) look dissimilar. The reason why Tolkien stopped using them
          was because he saw that the sounds they represented weren't related to
          the sounds represented by lambe and rómen. He continued experimenting
          with more consistent representations of w and wh (rómen and "extended
          rómen") and finally settled for vala and hwesta sindarinwa, the latter
          of which externally originated in the "Bombadil-mode" wh-tengwa (as
          suggested by Johan Winge in message #4453), but internally was
          reanalyzed as a variant of hwesta.

          I think my hypothesis has a number of advantages over yours:

          (1) It explains why the forms were abandoned.

          (2) It explains why there was much variations in the sign for w and wh.

          (3) It explains the similarity between the "Bombadil mode" wh-tengwa
          and hwesta sindarinwa.

          (4) It does not imply a reanalysis ot the "Bombadil mode" w and wh
          tengwar as reversed lambe and rómen for the explanation of DTS 17.

          So I believe my hypothesis is more explanatory and less implicatory
          than yours.

          ---------------------------
          j. 'mach' wust
          http://machhezan.tripod.com
          ---------------------------
        • Måns Björkman
          ... I think I understand what you are trying to describe. I don t know if there is any established term for it, though radical form seems appropriate (unless
          Message 4 of 14 , Dec 10, 2004
            --- In elfscript@yahoogroups.com, "j_mach_wust" <machhezan@g...> wrote:
            >
            > [...]
            >
            > Måns Björkman replied:
            >
            > > While I do agree that the "decorated verse-hand" of DTS 18 makes it
            > > a questionable source for the basic shapes of the letters, I don't
            > > think DTS 17 is more reliable. The vast majority of Tengwar samples
            > > published to date contains letter shapes consistent with the
            > > "book-hand", more or less. Why then assume that the "pointed" style
            > > should display the most elemental forms of the letters, rather than
            > > a curved one?
            >
            > Not the most elemental, for sure, but closer to the underlying ideal,
            > that is, more elemental (perhaps I've misused this word). Or perhaps I
            > shouldn't call it an ideal at all, but rather the very core of the
            > letters shape, the required form that distinguishes it from other
            > letters. The "book-hand" is very close to them, the "decorated
            > verse-hand" of DTS 18 is less close to them, and the angular forms of
            > DTS 17 are again closer to them.

            I think I understand what you are trying to describe. I don't know if
            there is any established term for it, though "radical form" seems
            appropriate (unless too strongly pointing backwards in time). Unlike
            you, I would rate the decorated verse-hand higher than the pointed
            style for closeness to that form. The differences between the
            book-hand style and the verse-hand are not very big, in my opinion.


            > I think this explains better why the rómen-like form of DTS 17 than
            > your explanation in message #73 (cited by Johan Winge in #4453)
            > because your explanation assumes a reanalysis of the form, that is, an
            > additional hypothetical step, whereas I assume (in message #4456) that
            > the forms are basically identical.

            But if I have understood your theory correctly, it requires an extra
            hypothetical step for the /w/ tengwa of DTS 17, *as well as* both the
            /w/ and the /W/ tengwar of DTS 18, namely, that these have in shape
            been removed from their hypothetical origins as reversed <lambe> and
            <rómen>. Also, in retrospect, I don't think an intentional approach
            toward the shape of a reversed <rómen> is really required for my
            theory; the pointed-style /W/ tengwa may have gotten this shape for
            many reasons (desire to distinguish it more clearly from <hyarmen>, to
            rely less heavily on a thin /-slanting line, or merely artistic license).


            > > In the case of the /w/-/W/ letters, I see many issues with regarding
            > > these as turned <lambe> and <rómen>. Firstly, the only character
            > > that is actually in shape an upside-down version of another
            > > character is the *<rómen nuquerna> of DTS 17. In DTS 18, the letter
            > > is clearly distinct from <rómen>. Even if the letter-shapes of DTS
            > > 18 are more "conditioned by the pen movements", it is (as Johan
            > > Winge notes) no more difficult to write an turned <rómen> than an
            > > ordinary one.
            >
            > I don't think you can prove this. I do think it would be possible to
            > prove it based on an analysis of the shapes in DTS 18. You'd have to
            > show that the same shapes that a true reversed rómen would require are
            > found in DTS 18. I don't think they are found.

            You are quite right that I cannot prove that a turned <rómen> is as
            easy to write as a normal one; and certainly not that it was as easy
            for Tolkien in the 1930's. I can only speak from my own experiences
            during a decade as a calligrapher and graphic designer.


            > On the other hand, it's quite easy to show that there are other
            > letters in DTS 18 that have a peculiar form "conditioned by the pen
            > movements", e.g. lambe.

            The only difference between the DTS 18 <lambe> and the book-hand form
            of any significance that I can see is that the verse-hand form is
            often written with one continous stroke. Since a large number of
            letters lack even this peculiarity (including <rómen>), I don't regard
            it as very conclusive.


            > > Secondly, neither turned letter has a close
            > > phonological relationship with its normal form. If <rómen> was
            > > actually *used* for /w/ in these texts, I would find it easier to
            > > believe that its turned form was used for /W/.
            >
            > You're saying that the "Bombadil-mode" w and wh tengwar are not
            > reversed forms of lambe and rómen but may accidently (because of
            > reanalysis) be replaced by reversed lambe and rómen. For unknown
            > reasons, Tolkien stopped using them.

            As I see it, the "accident" occurs only once, in DTS 17, where the /W/
            letter appears as an allograph reminiscent of an upside-down <rómen>.
            In all other occurrences, the /W/ letter deviates from that form.
            Nowhere else does it have the inward-closed bow of <rómen>. Further, I
            don't think Tolkien stopped using it: it reoccurs in DTS 37.

            The /w/ tengwa is often quite similar to a turned <lambe>, but the
            only thing speaking for a closer relationship is that similarity (and
            analogy with the supposed "turned <rómen>"). If it is related, why is
            there any deviations at all -- even in DTS 17, where the turned
            <rómen> occurs? Why does it keep the diminished height of <óre>
            instead of swooping to the height of <súle>, as the turned <rómen> does?

            I think it is good practice to assume that the most frequent is normal
            and regular, whereas isolated deviations represent irregularities.
            Given all the evidence, I do not see that the /w/ and /W/ letters are
            reversed <lambe> and <rómen>.

            New material published recently has made me consider a quite different
            origin for the /w/ letter. In "Addenda and Corrigenda to the
            _Etymologies_ - Part Two" in _Vinyar Tengwar_ #46, a tengwa named
            _waiya_ > _vaiya_ is given: it looks like an <úre> with a short tail,
            or slightly like a Q with the tail being straight, horizontal and
            starting at the bottom middle of the circle. The writing is rough, but
            it appears as if Tolkien started the circle at the bottom, drew
            counterclockwise and continued a short distance to the right as he
            reached the bottom again.

            It seems plausible to me that this letter could be the origin of the
            "2"-shaped tengwa found in DTS 16, 17 and 18. It could also suggest an
            alternative origin for <hwesta sindarinwa> (also occurring in this
            source under the name _hwinde_) as being a form of the same tengwa
            supplemented by a raised stem to mark it as a fricative. And this,
            unfortunately, does not fit my "cursive -> straightened" theory.
            Instead, I would assume that the straightened <hwesta sindarinwa> gave
            rise to the cursive form, and that they were used as allographs (a
            development with many parallels in Roman writing, and a few in Tengwar
            writing).


            > I'm saying that the "Bombadil-mode" w and wh tengwar are reversed
            > forms of lambe and rómen but may accidently (conditioned by pen
            > movement) look dissimilar. The reason why Tolkien stopped using them
            > was because he saw that the sounds they represented weren't related to
            > the sounds represented by lambe and rómen.

            Why would Tolkien have started using reversed <lambe> and <rómen> for
            /w/ and /W/ in the first place?


            > He continued experimenting
            > with more consistent representations of w and wh (rómen and "extended
            > rómen") and finally settled for vala and hwesta sindarinwa, the latter
            > of which externally originated in the "Bombadil-mode" wh-tengwa (as
            > suggested by Johan Winge in message #4453), but internally was
            > reanalyzed as a variant of hwesta.

            I agree this scenario seems probable in general, though I can't find
            any evidence that the connection between sound /w/ and letter <rómen>
            preceded DTS 17 -- rather the reverse. In DTS 22, written in 1937,
            <rómen> is used for /r/, and it is so used in _The Etymologies_, and
            DTS 17 which of course is the source of "turned <rómen>". The earliest
            occurrence of <rómen> for /w/ that I can think of is DTS 10,
            "possibly" written 1948 but very likely postdating DTS 17.


            > I think my hypothesis has a number of advantages over yours:
            >
            > (1) It explains why the forms were abandoned.

            And I don't think they were. (Possibly <waiya>, if I may use that
            name, but not the cursive <hwesta sindarinwa>.)


            > (2) It explains why there was much variations in the sign for w and wh.

            The variation is not that big, in my opinion. It is usually just a
            matter of swoop. :-)


            > (3) It explains the similarity between the "Bombadil mode" wh-tengwa
            > and hwesta sindarinwa.

            Since my own theory doesn't conflict with yours on this point, I agree.


            > (4) It does not imply a reanalysis ot the "Bombadil mode" w and wh
            > tengwar as reversed lambe and rómen for the explanation of DTS 17.

            I do not reanalyse the /w/ tengwa of DTS 17, which I think is clearly
            distinct from a reversed <lambe>. And I suggest a reanalysis of /W/ as
            reversed <rómen> merely as one possibility out of many.


            > So I believe my hypothesis is more explanatory and less implicatory
            > than yours.

            I believe my hypothesis takes better into account all the forms of all
            the available sources, rather than relying heavily on a few unusual forms.


            Yours,

            Måns
          • j_mach_wust
            ... The differences between the angular style and the book-hand aren t very big either, except for the angles, of course, which I consider to be of secondary
            Message 5 of 14 , Dec 10, 2004
              Måns Björkman wrote:

              > I would rate the decorated verse-hand
              > higher than the pointed style for closeness to that form. The
              > differences between the book-hand style and the verse-hand are
              > not very big, in my opinion.

              The differences between the angular style and the book-hand aren't
              very big either, except for the angles, of course, which I consider to
              be of secondary importance for the "radical form".

              Some more samples of DTS 17 being closer to the book-hand: The tengwa
              hyarmen of the book-hand and of DTS 17 have both a straight longer
              stroke, whereas it's curved in DTS 18; the tengwa esse looks more
              composed in the book-hand and in DTS 17 than in DTS 18; the down part
              of the tengwa rómen has a very marked lambe-like curl in DTS 18
              whereas it doesn't in the book-hand and in DTS 17.

              > > I think this explains better why the rómen-like form of DTS 17
              > > than your explanation in message #73 (cited by Johan Winge in
              > > #4453) because your explanation assumes a reanalysis of the
              > > form, that is, an additional hypothetical step, whereas I
              > > assume (in message #4456) that the forms are basically
              > > identical.
              >
              > But if I have understood your theory correctly, it requires an
              > extra hypothetical step for the /w/ tengwa of DTS 17,

              You're referring to the "diminished height", I guess. Lambe doesn't
              need to be as tall as tinco, and the reversed lambe is also higher
              than the númen-hight, especially in the first instances in DTS 17.
              Towards the end of the sample, the lambes grow taller and the "lambe
              nuquerna" smaller. What's your explanation of the different heights?

              You may also be referring to the diagonal stroke being attached to the
              end of the horizontal one rather than to the middle. A possible
              explanation is that in the "reversed lambe", it's easier to achieve
              the same left-right extension of the "lúva" and the horizontal stroke.
              I have to admit this is an extra hypothetical step.

              > *as well
              > as* both the /w/ and the /W/ tengwar of DTS 18, namely, that
              > these have in shape been removed from their hypothetical origins
              > as reversed <lambe> and <rómen>.

              I see these as a natural transformation due to the swifter pen
              movements. If you insist, you could call this a hypothesis: The shapes
              are conditioned by the general style of the specimen.

              > Also, in retrospect, I don't
              > think an intentional approach toward the shape of a reversed
              > <rómen> is really required for my theory; the pointed-style /W/
              > tengwa may have gotten this shape for many reasons (desire to
              > distinguish it more clearly from <hyarmen>, to rely less heavily
              > on a thin /-slanting line, or

              These arguments exclusively apply to DTS 17; why don't they apply in
              DTS 18 as well? I'd like to have arguments that don't depend on DTS
              numbers or on:

              > merely artistic license).

              ...

              > > > Even if the letter-shapes of DTS 18 are more "conditioned by
              > > > the pen movements", it is (as Johan Winge notes) no more
              > > > difficult to write an turned <rómen> than an ordinary one.
              > >
              > > I don't think you can prove this. I do think it would be
              > > possible to prove it based on an analysis of the shapes in DTS
              > > 18. You'd have to show that the same shapes that a true
              > > reversed rómen would require are found in DTS 18. I don't
              > > think they are found.
              >
              > You are quite right that I cannot prove that a turned <rómen> is
              > as easy to write as a normal one; and certainly not that it was
              > as easy for Tolkien in the 1930's. I can only speak from my own
              > experiences during a decade as a calligrapher and graphic
              > designer.

              I said that it would be possible to prove your hypothesis if it were
              true, that is, not a real prove of course, but finding very strong
              arguments from an analysis of the tengwar shapes in DTS 18. Do the
              shapes required by a reversed rómen occur?

              > > On the other hand, it's quite easy to show that there are
              > > other letters in DTS 18 that have a peculiar form "conditioned
              > > by the pen movements", e.g. lambe.
              >
              > The only difference between the DTS 18 <lambe> and the book-hand
              > form of any significance that I can see is that the verse-hand
              > form is often written with one continous stroke. Since a large
              > number of letters lack even this peculiarity (including
              > <rómen>), I don't regard it as very conclusive.

              (The peculiarity of being written with one continuous stroke?) For
              itself, the manner how lambe is written is not conclusive, but it
              shows a trend if we consider it together with the other differences
              from both book-hand and DTS 17 (see above).

              > > You're saying that the "Bombadil-mode" w and wh tengwar are
              > > not reversed forms of lambe and rómen but may accidently
              > > (because of reanalysis) be replaced by reversed lambe and
              > > rómen. For unknown reasons, Tolkien stopped using them.
              >
              > As I see it, the "accident" occurs only once, in DTS 17, where
              > the /W/ letter appears as an allograph reminiscent of an
              > upside-down <rómen>. In all other occurrences, the /W/ letter
              > deviates from that form. Nowhere else does it have the
              > inward-closed bow of <rómen>.

              That's true. It's the weak spot of my hypothesis that it takes the far
              less numerous sample as the "radical form". But you haven't given me
              any arguments that would explain the reversed-rómen form.

              > The /w/ tengwa is often quite similar to a turned <lambe>, but
              > the only thing speaking for a closer relationship is that
              > similarity (and analogy with the supposed "turned <rómen>"). If
              > it is related, why is there any deviations at all -- even in DTS
              > 17, where the turned <rómen> occurs?

              You're implying that "reversed" letters should be perfectly symmetric
              to their unreversed counterparts. I'm happy I don't have to prove that
              esse nuquerna is a reversed esse -- it might be difficult against such
              a fierce opponent as you! :)

              Of course, the similarity is the most important argument for a closer
              relationship of these letters. But it's not the only thing that speaks
              for a closer relationship: Another hint is the fact these letters were
              abandoned, that is, "lambe nuquerna" was totally abandoned and "rómen
              nuquerna" was reanalyzed and changed in shape. I don't know of other
              abandoned tengwar at all.

              > New material published recently has made me consider a quite
              > different origin for the /w/ letter. In "Addenda and Corrigenda
              > to the _Etymologies_ - Part Two" in _Vinyar Tengwar_ #46, a
              > tengwa named _waiya_ > _vaiya_ is given: it looks like an <úre>
              > with a short tail, or slightly like a Q with the tail being
              > straight, horizontal and starting at the bottom middle of the
              > circle. The writing is rough, but it appears as if Tolkien
              > started the circle at the bottom, drew counterclockwise and
              > continued a short distance to the right as he reached the bottom
              > again.
              >
              > It seems plausible to me that this letter could be the origin of
              > the "2"-shaped tengwa found in DTS 16, 17 and 18.

              Interesting hypothesis. However, the "Bombadil-mode" w-tengwa is
              always open. And why should this tengwa be made more lambe-like?

              > It could also
              > suggest an alternative origin for <hwesta sindarinwa> (also
              > occurring in this source under the name _hwinde_) as being a
              > form of the same tengwa supplemented by a raised stem to mark it
              > as a fricative. And this, unfortunately, does not fit my
              > "cursive -> straightened" theory. Instead, I would assume that
              > the straightened <hwesta sindarinwa> gave rise to the cursive
              > form, and that they were used as allographs (a development with
              > many parallels in Roman writing, and a few in Tengwar writing).

              What development is parallel to a "reversed rómen" shape originating
              in a hwinde/hwesta sindarinwa shape? I don't know any at all.

              If I'm not wrong about the dates, the "Bombadil mode" samples might
              predate the Etymologies. The latter were written in the later 1930s,
              and the former may have been written already in the early 1930s. So
              this sample would also fit in my explanation that hwesta
              sindarinwa/hwinde was a reanalysis of a rómen nuquerna, and it would
              show another "experimental" w-tengwa.

              > > I'm saying that the "Bombadil-mode" w and wh tengwar are
              > > reversed forms of lambe and rómen but may accidently
              > > (conditioned by pen movement) look dissimilar. The reason why
              > > Tolkien stopped using them was because he saw that the sounds
              > > they represented weren't related to the sounds represented by
              > > lambe and rómen.
              >
              > Why would Tolkien have started using reversed <lambe> and
              > <rómen> for /w/ and /W/ in the first place?

              I don't know. He needed tengwar for these two sounds. Could it be that
              this assignment was only made in English? This might explain why he
              was less scrupulous.

              Your hypothesis doesn't explain why he abandoned them if they were
              distinct from rómen and lambe.

              > > He continued experimenting with more consistent
              > > representations of w and wh (rómen and "extended rómen") and
              > > finally settled for vala and hwesta sindarinwa, the latter of
              > > which externally originated in the "Bombadil-mode" wh-tengwa
              > > (as suggested by Johan Winge in message #4453), but internally
              > > was reanalyzed as a variant of hwesta.
              >
              > I agree this scenario seems probable in general, though I can't
              > find any evidence that the connection between sound /w/ and
              > letter <rómen> preceded DTS 17 -- rather the reverse.

              My bad English. (1) "lambe nuquerna" and "rómen nuquerna"; (2)
              comprehension that this doesn't work; (3) experimenting with other
              forms; (4) vala (or other depending on mode) and hwesta sindarinwa.

              > In DTS 22,
              > written in 1937, <rómen> is used for /r/, and it is so used in
              > _The Etymologies_, and DTS 17 which of course is the source of
              > "turned <rómen>". The earliest occurrence of <rómen> for /w/
              > that I can think of is DTS 10, "possibly" written 1948 but very
              > likely postdating DTS 17.

              Of course. DTS 50 and 51 were probably written in late 1946. It's
              interesting that the vala/hwesta sindarinwa solution predates this
              (DTS 25 at the end of 1939). The final forms don't need to be the last
              developed.

              > > I think my hypothesis has a number of advantages over yours:
              > >
              > > (1) It explains why the forms were abandoned.
              >
              > And I don't think they were. (Possibly <waiya>, if I may use
              > that name, but not the cursive <hwesta sindarinwa>.)

              I've missed my own point. Here it is again: My hypothesis explains why
              the "lambe nuquerna" was abandoned.

              > > (2) It explains why there was much variations in the sign for
              > > w and wh.
              >
              > The variation is not that big, in my opinion. It is usually just
              > a matter of swoop. :-)

              It's in comparison to other sounds.

              > > (3) It explains the similarity between the "Bombadil mode"
              > > wh-tengwa and hwesta sindarinwa.
              >
              > Since my own theory doesn't conflict with yours on this point, I
              > agree.

              I've missed my own point again. Here it goes: My hypothesis explains
              the similarities and the differences between the "Bombadil mode"
              wh-tengwa and hwesta sindarinwa.

              ---------------------------
              j. 'mach' wust
              http://machhezan.tripod.com
              ---------------------------
            • Måns Björkman
              ... I agree with you that the angular form of the letters is of secondary importance, but in my opinion, so are the few peculiarities of the decorated
              Message 6 of 14 , Dec 15, 2004
                "j_mach_wust" <machhezan@g...> wrote:
                >
                > Måns Björkman wrote:
                >
                > > I would rate the decorated verse-hand
                > > higher than the pointed style for closeness to that form. The
                > > differences between the book-hand style and the verse-hand are
                > > not very big, in my opinion.
                >
                > The differences between the angular style and the book-hand aren't
                > very big either, except for the angles, of course, which I consider to
                > be of secondary importance for the "radical form".

                I agree with you that the angular form of the letters is of secondary
                importance, but in my opinion, so are the few peculiarities of the
                decorated verse-hand.


                > Some more samples of DTS 17 being closer to the book-hand: The tengwa
                > hyarmen of the book-hand and of DTS 17 have both a straight longer
                > stroke, whereas it's curved in DTS 18; the tengwa esse looks more
                > composed in the book-hand and in DTS 17 than in DTS 18; the down part
                > of the tengwa rómen has a very marked lambe-like curl in DTS 18
                > whereas it doesn't in the book-hand and in DTS 17.

                I agree that the letters in DTS 17 are much more constrained in their
                design than those in DTS 18. This I think is due partly to the
                requirements of adapting the letters to a 'pointed' style, partly to
                space restraints caused by the use of a very broad calligraphy pen.

                But the deviations of DTS 18 are very small, and often not greater
                than in, say, DTS 10 (which I think is a prime example of the
                book-hand). None of the examples you give (<hyarmen> being more
                curved, <esse> being "less composed", <rómen> having a larger lower
                curl) remove any significant traits from the letters.

                I argue that if the /W/ tengwa of DTS 18 is meant to be a <rómen
                nuquerna>, then it has lost the significant trait of the inward bow,
                and thus been distorted more than any other letter in that sample.


                > > > [...]
                > >
                > > But if I have understood your theory correctly, it requires an
                > > extra hypothetical step for the /w/ tengwa of DTS 17,
                >
                > You're referring to the "diminished height", I guess. Lambe doesn't
                > need to be as tall as tinco, and the reversed lambe is also higher
                > than the númen-hight, especially in the first instances in DTS 17.
                > Towards the end of the sample, the lambes grow taller and the "lambe
                > nuquerna" smaller. What's your explanation of the different heights?

                As I see it, the greater <lambe> height and the slighter /w/-tengwa
                height remain consistent, with slight variations. Agreed, the smallest
                <lambe> (in line 2?) is more or less the same height as the tallest
                /w/-tengwa (in line 4?), but generally <lambe> is higher (cp for
                instance the /w/-tengwa and the <lambe> in line 2).


                > You may also be referring to the diagonal stroke being attached to the
                > end of the horizontal one rather than to the middle. A possible
                > explanation is that in the "reversed lambe", it's easier to achieve
                > the same left-right extension of the "lúva" and the horizontal stroke.
                > I have to admit this is an extra hypothetical step.

                Don't you think it might be significant that the diagonal stroke is
                attached to the end of the horizontal in *all* our samples of this
                tengwa, including DTS 16, 17, 18, and 23? If the pointed style is
                indeed closer to the "radical form" than the decorated verse-hand,
                shouldn't we somewhere, in some sample, expect a "genuine" turned <lambe>?


                > > *as well
                > > as* both the /w/ and the /W/ tengwar of DTS 18, namely, that
                > > these have in shape been removed from their hypothetical origins
                > > as reversed <lambe> and <rómen>.
                >
                > I see these as a natural transformation due to the swifter pen
                > movements. If you insist, you could call this a hypothesis: The shapes
                > are conditioned by the general style of the specimen.

                As pointed out above, these would be the only letters so markedly
                distorted in this style.


                > > Also, in retrospect, I don't
                > > think an intentional approach toward the shape of a reversed
                > > <rómen> is really required for my theory; the pointed-style /W/
                > > tengwa may have gotten this shape for many reasons (desire to
                > > distinguish it more clearly from <hyarmen>, to rely less heavily
                > > on a thin /-slanting line, or
                >
                > These arguments exclusively apply to DTS 17; why don't they apply in
                > DTS 18 as well? I'd like to have arguments that don't depend on DTS
                > numbers or on:
                >
                > > merely artistic license).

                Well, suffice to say that the one, singular attested sample of a
                /W/-tengwa with the shape of a turned <rómen> remains a mystery (if
                not a very great one), and that all other samples form a continuum
                leading up to the "cursive <hwesta sindarinwa>" in DTS 37.


                > [...]
                >
                > I said that it would be possible to prove your hypothesis if it were
                > true, that is, not a real prove of course, but finding very strong
                > arguments from an analysis of the tengwar shapes in DTS 18. Do the
                > shapes required by a reversed rómen occur?

                Well, a curve possibly corresponding to the bow of a reversed <rómen>
                may perhaps be found in <silme> or <vilya>, or the "e" tengwa of this
                mode. The rest of a reversed <rómen> (the upward swash and the right
                leg) could of course be written as in the /W/-tengwa.


                > [...]
                >
                > > > You're saying that the "Bombadil-mode" w and wh tengwar are
                > > > not reversed forms of lambe and rómen but may accidently
                > > > (because of reanalysis) be replaced by reversed lambe and
                > > > rómen. For unknown reasons, Tolkien stopped using them.
                > >
                > > As I see it, the "accident" occurs only once, in DTS 17, where
                > > the /W/ letter appears as an allograph reminiscent of an
                > > upside-down <rómen>. In all other occurrences, the /W/ letter
                > > deviates from that form. Nowhere else does it have the
                > > inward-closed bow of <rómen>.
                >
                > That's true. It's the weak spot of my hypothesis that it takes the far
                > less numerous sample as the "radical form". But you haven't given me
                > any arguments that would explain the reversed-rómen form.

                I have, though not strong arguments (as you rightly pointed out). On
                the other hand, I don't think your arguments for the /W/-tengwa in DTS
                18 being a distorted <rómen nuquerna> are very strong either, and in
                my opoinion they are further weakened by the occurrence of a closely
                similar tengwa in other sources.


                > > The /w/ tengwa is often quite similar to a turned <lambe>, but
                > > the only thing speaking for a closer relationship is that
                > > similarity (and analogy with the supposed "turned <rómen>"). If
                > > it is related, why is there any deviations at all -- even in DTS
                > > 17, where the turned <rómen> occurs?
                >
                > You're implying that "reversed" letters should be perfectly symmetric
                > to their unreversed counterparts. I'm happy I don't have to prove that
                > esse nuquerna is a reversed esse -- it might be difficult against such
                > a fierce opponent as you! :)

                I'm only fierce when whining doesn't work ;-)

                I don't require that turned forms of letters are in perfect symmetry
                with their normal form. In fact, even if we did not know it for a
                fact, <esse nuquerna> has many characteristics that strongly imply
                that it is a turned <esse>: it has the same sound value as <esse>, it
                is used as an allograph of that tengwa in several sources, and in all
                sources it is consistently drawn as two bows turned opposite to those
                in <esse>, clearly hinting at a relationship. The supposed <rómen
                nuquerna> lacks all these characteristics.

                I think my point was not made very clearly in my previous post. As I
                see it, the only thing speaking for a relationship between the
                /W/-tengwa and <rómen> is that they are perfectly symmetrical in shape
                in DTS 17. Would you be so willing to believe in a <rómen nuquerna> if
                the /W/-tengwa looked like, say, a mirrored <hyarmen>? The theory of a
                <lambe nuquerna> seems to me to rest on even shakier grounds, since
                the /w/-tengwa lacks even the perfect symmetry with <lambe> (in a
                source of writing supposedly very close to the "radical form").


                > Of course, the similarity is the most important argument for a closer
                > relationship of these letters. But it's not the only thing that speaks
                > for a closer relationship: Another hint is the fact these letters were
                > abandoned, that is, "lambe nuquerna" was totally abandoned and "rómen
                > nuquerna" was reanalyzed and changed in shape. I don't know of other
                > abandoned tengwar at all.

                Well, that might be just a matter of scarce sources. But speaking of
                abandoned tengwar, the initial "schwa" tengwa comes to mind (the one
                looking like a stem with a horisontal bar connecting it to the
                following tengwa). (In "a merry", DTS 16-18.)


                > > New material published recently has made me consider a quite
                > > different origin for the /w/ letter. In "Addenda and Corrigenda
                > > to the _Etymologies_ - Part Two" in _Vinyar Tengwar_ #46, a
                > > tengwa named _waiya_ > _vaiya_ is given: it looks like an <úre>
                > > with a short tail, or slightly like a Q with the tail being
                > > straight, horizontal and starting at the bottom middle of the
                > > circle. The writing is rough, but it appears as if Tolkien
                > > started the circle at the bottom, drew counterclockwise and
                > > continued a short distance to the right as he reached the bottom
                > > again.
                > >
                > > It seems plausible to me that this letter could be the origin of
                > > the "2"-shaped tengwa found in DTS 16, 17 and 18.
                >
                > Interesting hypothesis. However, the "Bombadil-mode" w-tengwa is
                > always open. And why should this tengwa be made more lambe-like?

                A good question, and one that I won't delve into at the moment. It is
                not crucial for the current discussion. :-)


                > What development is parallel to a "reversed rómen" shape originating
                > in a hwinde/hwesta sindarinwa shape? I don't know any at all.

                I can't think of any either, but I don't find that very surprising
                given that we only have two document pages that constitute samples of
                the pointed style (except the "Namárie" headline in DTS 20).


                > If I'm not wrong about the dates, the "Bombadil mode" samples might
                > predate the Etymologies. The latter were written in the later 1930s,
                > and the former may have been written already in the early 1930s. So
                > this sample would also fit in my explanation that hwesta
                > sindarinwa/hwinde was a reanalysis of a rómen nuquerna, and it would
                > show another "experimental" w-tengwa.

                It is probably true that DTS 16-18 predate the Etymologies, and the
                <waiya>-tengwa might be "experimental", as you say. Still, it doesn't
                feel to me like a probable scenario that <hwesta sindarinwa> was
                developed from reanalysis of <rómen nuquerna>. Note that it is the
                "upright", normal <hwesta sindarinwa> of DTS 9 that is found in the
                Etymologies, not the cursive form of DTS 37.

                (By the way, if the Etymologies postdates DTS 16-18, then <waiya>
                would be developed from the /w/-tengwa of the "Bombadil-mode", rather
                than the other way round. The later tengwa is thus "closed", made
                *less* "lambe-like" as you put it. This should be an interesting topic
                of research for a later time.)


                > > > I'm saying that the "Bombadil-mode" w and wh tengwar are
                > > > reversed forms of lambe and rómen but may accidently
                > > > (conditioned by pen movement) look dissimilar. The reason why
                > > > Tolkien stopped using them was because he saw that the sounds
                > > > they represented weren't related to the sounds represented by
                > > > lambe and rómen.
                > >
                > > Why would Tolkien have started using reversed <lambe> and
                > > <rómen> for /w/ and /W/ in the first place?
                >
                > I don't know. He needed tengwar for these two sounds. Could it be that
                > this assignment was only made in English? This might explain why he
                > was less scrupulous.

                Now we are simply guessing; but I have the feeling Tolkien usually had
                his own language foremost in his mind when developing his scripts --
                either that, or phonetic applicability in general. Of the published
                Tengwar and Sarati texts written in the first half of the 20th
                century, the vast majority is in English.


                > Your hypothesis doesn't explain why he abandoned them if they were
                > distinct from rómen and lambe.

                As stated earlier, I don't think he abandoned the /W/-tengwa seen in
                DTS 16-18 -- in my opinion the /W/-tengwa of DTS 37 is the same one.
                As for the /w/-tengwa, I concede that it may very well have been a
                passing idea, but I don't think it must have had anything to do with
                <lambe>.


                > > > [...]
                >
                > My bad English. (1) "lambe nuquerna" and "rómen nuquerna"; (2)
                > comprehension that this doesn't work; (3) experimenting with other
                > forms; (4) vala (or other depending on mode) and hwesta sindarinwa.

                See my comments above.


                > > In DTS 22,
                > > written in 1937, <rómen> is used for /r/, and it is so used in
                > > _The Etymologies_, and DTS 17 which of course is the source of
                > > "turned <rómen>". The earliest occurrence of <rómen> for /w/
                > > that I can think of is DTS 10, "possibly" written 1948 but very
                > > likely postdating DTS 17.
                >
                > Of course. DTS 50 and 51 were probably written in late 1946.

                Right. I didn't think of ol' Lowdham. :)


                > It's
                > interesting that the vala/hwesta sindarinwa solution predates this
                > (DTS 25 at the end of 1939). The final forms don't need to be the last
                > developed.

                I'm may be confused, but I can't find any evidence in DTS 25 (The
                Doors of Durin (Inscription Drafts)) of the <vala>/<hwesta sindarinwa>
                solution. What am I missing?


                > > > I think my hypothesis has a number of advantages over yours:
                > > >
                > > > (1) It explains why the forms were abandoned.
                > >
                > > And I don't think they were. (Possibly <waiya>, if I may use
                > > that name, but not the cursive <hwesta sindarinwa>.)
                >
                > I've missed my own point. Here it is again: My hypothesis explains why
                > the "lambe nuquerna" was abandoned.

                I am sure I have commented on this somewhere already. ;-)



                > > > (2) It explains why there was much variations in the sign for
                > > > w and wh.
                > >
                > > The variation is not that big, in my opinion. It is usually just
                > > a matter of swoop. :-)
                >
                > It's in comparison to other sounds.

                Yes, you are right about that. The /w/ sign of DTS 16-18 does show an
                unusual amount of variation.


                > > > (3) It explains the similarity between the "Bombadil mode"
                > > > wh-tengwa and hwesta sindarinwa.
                > >
                > > Since my own theory doesn't conflict with yours on this point, I
                > > agree.
                >
                > I've missed my own point again. Here it goes: My hypothesis explains
                > the similarities and the differences between the "Bombadil mode"
                > wh-tengwa and hwesta sindarinwa.

                While I agree that <hwesta sindarinwa> may have originated in a
                cursive form, I am not convinced that that cursive form in turn must
                have originated in a <rómen nuquerna>. It is clear to me that the
                /W/-tengwa of the DTS 16-18 period had an allograph (in the pointed
                style) that looked exactly like an upside-down <rómen>. But all other
                allographs of this grapheme are more similar to the /W/-tengwa of DTS
                23 and 37.

                To return to the origins of this thread: in a font that follows the
                'pointed' style, the /W/-tengwa (for lack of a better name) should, in
                my opinion, appear as a turned <rómen>. But if the font follows rather
                the book-hand, the /W/-tengwa should be based on the forms in DTS 37,
                23, and 18.

                Yours,

                Måns
              • j_mach_wust
                I fear we won t come to an agreement on the reversed rómen/lambe hypothesis . My hypothesis that they re reverted rómen/lambe seems not to convince you, and
                Message 7 of 14 , Dec 15, 2004
                  I fear we won't come to an agreement on the "reversed rómen/lambe
                  hypothesis". My hypothesis that they're reverted rómen/lambe seems not
                  to convince you, and your hypothesis that they're independent tengwar
                  won't convince me. I see underlying rómen and lambe, and I find
                  arguments for it; you see independent tengwar, and you find arguments
                  for it.

                  Måns and me wrote alternatly:
                  > > Some more samples of DTS 17 being closer to the book-hand: The
                  > > tengwa hyarmen of the book-hand and of DTS 17 have both a straight
                  > > longer stroke, whereas it's curved in DTS 18; the tengwa esse
                  > > looks more composed in the book-hand and in DTS 17 than in DTS 18;
                  > > the down part of the tengwa rómen has a very marked lambe-like
                  > > curl in DTS 18 whereas it doesn't in the book-hand and in DTS 17.
                  >
                  > I agree that the letters in DTS 17 are much more constrained in
                  > their design than those in DTS 18.

                  At a first glance I thought you were admitting that DTS 17 is closer
                  to the "radical form" and that therefore, you'd have to agree that the
                  unambiguous "rómen nuquerna" of DTS 17 absolutly must be considered.
                  At a second glance, I realize that you'll still not agree since you've
                  only conceded that it is "constrained" and not that it would be closer
                  to the "radical form"...


                  > But the deviations of DTS 18 are very small, and often not greater
                  > than in, say, DTS 10 (which I think is a prime example of the
                  > book-hand).

                  I was only thinking of (that is, comparing to) the chart from app. E,
                  DTS 9.

                  > None of the examples you give (<hyarmen> being more
                  > curved, <esse> being "less composed", <rómen> having a larger lower
                  > curl) remove any significant traits from the letters.
                  >
                  > I argue that if the /W/ tengwa of DTS 18 is meant to be a <rómen
                  > nuquerna>, then it has lost the significant trait of the inward bow,
                  > and thus been distorted more than any other letter in that sample.

                  The significancy of the traits is a circular argument: They only
                  become significant because you believe that it's different tengwar.


                  > If the pointed style is
                  > indeed closer to the "radical form" than the decorated verse-hand,
                  > shouldn't we somewhere, in some sample, expect a "genuine" turned
                  > <lambe>?

                  That's right, my hypothesis implies a still more "radical", but
                  unattested form where lambe and "lambe nuquerna" look identical. I'd
                  only expect a "genuine" "reversed lambe" on DTS 9.


                  > > > the pointed-style /W/
                  > > > tengwa may have gotten this shape for many reasons (desire to
                  > > > distinguish it more clearly from <hyarmen>, to rely less heavily
                  > > > on a thin /-slanting line, or
                  > >
                  > > These arguments exclusively apply to DTS 17; why don't they apply
                  > > in DTS 18 as well? I'd like to have arguments that don't depend on
                  > > DTS numbers or on:
                  > >
                  > > > merely artistic license).
                  >
                  > Well, suffice to say that the one, singular attested sample of a
                  > /W/-tengwa with the shape of a turned <rómen> remains a mystery (if
                  > not a very great one), and that all other samples form a continuum
                  > leading up to the "cursive <hwesta sindarinwa>" in DTS 37.

                  In science, you only need a single counterexample for a theory to be
                  rebutted. (For sure, this is not science but philology, that is,
                  interpretation.) I claim that my interpretation produces no mysterious
                  left-overs.


                  > > Do the shapes required by a reversed rómen occur?
                  >
                  > Well, a curve possibly corresponding to the bow of a reversed
                  > <rómen> may perhaps be found in <silme> or <vilya>, or the "e"
                  > tengwa of this mode. The rest of a reversed <rómen> (the upward
                  > swash and the right leg) could of course be written as in the
                  > /W/-tengwa.

                  I've found a sample of a straightened rightward bow, that is, of the
                  same alternation as between the "rómen nuquerna" tengwar of DTS 17 and
                  18: Compare DTS 20 rómen with other rómen, or compare the first three
                  occurences of rómen in the second line of DTS 51 with the fourth one
                  (you'll probably insist that it's unique)!

                  The Lowdham-texts show also another example of a sign with variable
                  heighth: the Anglosaxon &-sign: It's short in DTS 50 and tall in DTS 51.

                  And that's not enough: There are also samples of that look like an
                  inverted _2_: In the second part of DTS 50, the deleted one on line 25
                  (you'll probably argue it isn't a lambe at all) and the very last one
                  of line 27 (you'll probably argue it's really composed like the others).

                  The mere existence of single instances of these variations make it
                  more probable that the same kind of variation is found in the supposed
                  "lambe nuquerna" and "rómen nuquerna".


                  > I don't think your arguments for the /W/-tengwa in
                  > DTS 18 being a distorted <rómen nuquerna> are very strong either,
                  > and in my opoinion they are further weakened by the occurrence of a
                  > closely similar tengwa in other sources.

                  You're speaking of the wh-tengwa in DTS 37, right? This tengwa isn't
                  any more similar to the one of DTS 18 than the one of DTS 17 is.
                  However, the similarities of DTS 17 and 18 are undoubtingly much
                  bigger than the ones between DTS 18 and 37, so I tend to associate the
                  wh-tengwa of DTS 18 with the one of DTS 17. Especially because DTS 37
                  uses the same w-tehta as other later modes, vala, I believe that the
                  wh-tehta of DTS 37 is a cursive form of hwesta sindarinwa (coexisting
                  with the straight form and probably predating it as an intermediate
                  form between from the "rómen nuquerna").


                  > I don't require that turned forms of letters are in perfect symmetry
                  > with their normal form. In fact, even if we did not know it for a
                  > fact, <esse nuquerna> has many characteristics that strongly imply
                  > that it is a turned <esse>: it has the same sound value as <esse>,
                  > it is used as an allograph of that tengwa in several sources, and in
                  > all sources it is consistently drawn as two bows turned opposite to
                  > those in <esse>, clearly hinting at a relationship. The supposed
                  > <rómen nuquerna> lacks all these characteristics.

                  It's obvious that it lacks the first two characteristics of bearing
                  the same sound value and being an allograph. It's an important point
                  in my argumentation that the supposed "rómen nuquerna" and "lambe
                  nuquerna" have different sound values and are not allographs of rómen
                  and lambe. Looking at DTS 21 (the only sample I remember that has both
                  except for the charts), however, I think that the differences in shape
                  between esse and esse nuquerna are as big as the differences in shape
                  between the lambe/rómen etc.: Esse looks like two silme, esse nuquerna
                  like a _3_.

                  And still another difference between an upright tengwa and a reversed
                  one: In all instances of malta in DTS 18, the closing horizontal
                  stroke begins in the telco, whereas in all instances of nwalme, it
                  exeeds it, most of the times very markedly. I believe it's common for
                  reversed forms to deviate from the upright forms. Have you ever seen
                  anybody drawing a schwa or a reversed a (the IPA signs corresponding
                  to [6] or to [Q])? I don't call these "distorsions", but "natural"
                  deviations.


                  > > Of course, the similarity is the most important argument for a
                  > > closer relationship of these letters. But it's not the only thing
                  > > that speaks for a closer relationship: Another hint is the fact
                  > > these letters were abandoned, that is, "lambe nuquerna" was
                  > > totally abandoned and "rómen nuquerna" was reanalyzed and changed
                  > > in shape. I don't know of other abandoned tengwar at all.
                  >
                  > Well, that might be just a matter of scarce sources. But speaking of
                  > abandoned tengwar, the initial "schwa" tengwa comes to mind (the one
                  > looking like a stem with a horisontal bar connecting it to the
                  > following tengwa). (In "a merry", DTS 16-18.)

                  Right, this tengwa wasn't used in DTS 24 any more. There is also much
                  variance in the representation of schwa! But the schwa representation
                  was already complex in the "Bombadil modes", so I'm not surprised that
                  it changed, whereas according to your hypothesis, there were two
                  clearly distinct tengwar for /w/ and /W/.


                  > > What development is parallel to a "reversed rómen" shape
                  > > originating in a hwinde/hwesta sindarinwa shape? I don't know any
                  > > at all.
                  >
                  > I can't think of any either, but I don't find that very surprising
                  > given that we only have two document pages that constitute samples
                  > of the pointed style (except the "Namárie" headline in DTS 20).

                  It doesn't need to be one of the "few" parallels in tengwar, I'd also
                  be interested in one of the "many parallels in Roman writing" you alluded.


                  > It is probably true that DTS 16-18 predate the Etymologies, and the
                  > <waiya>-tengwa might be "experimental", as you say. Still, it
                  > doesn't feel to me like a probable scenario that <hwesta sindarinwa>
                  > was developed from reanalysis of <rómen nuquerna>. Note that it is
                  > the "upright", normal <hwesta sindarinwa> of DTS 9 that is found in
                  > the Etymologies, not the cursive form of DTS 37.

                  If I am to uphold that hwesta sindarinwa is a reanalysis of the
                  supposed "rómen nuquerna", then I have to assume that the
                  "intermediate" form seen in DTS 37 already existed when the
                  etymologies were written (as written above).


                  > (By the way, if the Etymologies postdates DTS 16-18, then <waiya>
                  > would be developed from the /w/-tengwa of the "Bombadil-mode",
                  > rather than the other way round. The later tengwa is thus "closed",
                  > made *less* "lambe-like" as you put it. This should be an
                  > interesting topic of research for a later time.)

                  Indeed! (I hope that one day we will be able to look back and wonder
                  how we could speculate so much on so few published samples! ;-)


                  > I have the feeling Tolkien usually
                  > had his own language foremost in his mind when developing his
                  > scripts -- either that, or phonetic applicability in general.

                  I share the feeling that most of it was based on English!


                  > Of the
                  > published Tengwar and Sarati texts written in the first half of the
                  > 20th century, the vast majority is in English.

                  Published, but alas, not publicly available... when PE 13 was still
                  available, I wasn't that much into Tolkien's scripts yet, and now I
                  regret it so much!


                  > > Your hypothesis doesn't explain why he abandoned them if they were
                  > > distinct from rómen and lambe.
                  >
                  > As stated earlier, I don't think he abandoned the /W/-tengwa seen in
                  > DTS 16-18 -- in my opinion the /W/-tengwa of DTS 37 is the same one.
                  > As for the /w/-tengwa, I concede that it may very well have been a
                  > passing idea, but I don't think it must have had anything to do with
                  > <lambe>.

                  Are you going to tell me you believe that the wh-tengwa of DTS 18 is
                  an allograph of hwesta sindarinwa? Is it the same tengwa? If it is,
                  what happened that it changed so much? If it is not, why was it
                  replaced? (See also above.)


                  > I may be confused, but I can't find any evidence in DTS 25 (The
                  > Doors of Durin (Inscription Drafts)) of the <vala>/<hwesta
                  > sindarinwa> solution. What am I missing?

                  My bad, I thought of DTS 24, but I had the wrong DTS number in mind (I
                  still don't know them all by heart :-).


                  > > My hypothesis explains why the "lambe nuquerna" was abandoned.
                  >
                  > I am sure I have commented on this somewhere already. ;-)

                  Yes you have. You've said: "As for the /w/-tengwa, I concede that it
                  may very well have been a passing idea". From this, I understand that
                  you can't explain it.


                  > > > > (2) It explains why there was much variations in the sign for
                  > > > > w and wh.
                  > > >
                  > > > The variation is not that big, in my opinion. It is usually just
                  > > > a matter of swoop. :-)
                  > >
                  > > It's in comparison to other sounds.
                  >
                  > Yes, you are right about that. The /w/ sign of DTS 16-18 does show
                  > an unusual amount of variation.

                  I've missed the point again (gee, it's difficult to talk about these
                  things without using your fingers!): Looking at all the tengwar
                  samples, the sounds /w/ and /W/ have a bigger number of signs they can
                  be represented with (and not just in terms of allography) than the
                  other sounds. My hypothesis takes this into account.


                  > > My hypothesis
                  > > explains the similarities and the differences between the
                  > > "Bombadil mode" wh-tengwa and hwesta sindarinwa.
                  >
                  > While I agree that <hwesta sindarinwa> may have originated in a
                  > cursive form, I am not convinced that that cursive form in turn must
                  > have originated in a <rómen nuquerna>.

                  The cursive form hwesta sindarinwa originated in would be the one of
                  DTS 37.


                  > To return to the origins of this thread: in a font that follows the
                  > 'pointed' style, the /W/-tengwa (for lack of a better name) should,
                  > in my opinion, appear as a turned <rómen>. But if the font follows
                  > rather the book-hand, the /W/-tengwa should be based on the forms in
                  > DTS 37, 23, and 18.

                  I disagree! I'd suggest to have two different wh-tengwar for each
                  font: In the angular style, a "rómen nuquerna" and a hwesta
                  sindarinwa; in the upright book-hand style a "rómen nuquerna" (similar
                  to the one in DTS 18, but maybe a little bit more like the one in DTS
                  17) and a hwesta sindarinwa; in a cursive book-hand style a "rómen
                  nuquerna" (as in DTS 18) and a hwesta sindarinwa (as in DTS 37).

                  I believe it's important to keep these apart, just as you should keep
                  the w-tengwa of the "Bombadil-modes" apart from later w-tengwar.

                  ---------------------------
                  j. 'mach' wust
                  http://machhezan.tripod.com
                  ---------------------------
                • Måns Björkman
                  I suspect you are right in saying that we won t come to an agreement regarding the w/wh tengwar. (On that much we agree! ;-) Perhaps future publications will
                  Message 8 of 14 , Jan 12, 2005
                    I suspect you are right in saying that we won't come to an agreement
                    regarding the w/wh tengwar. (On that much we agree! ;-) Perhaps future
                    publications will prove one of us right, perhaps not. I think I will
                    leave it at this for now.

                    Yours,
                    Måns



                    --- In elfscript@yahoogroups.com, "j_mach_wust" <machhezan@g...> wrote:
                    >
                    > I fear we won't come to an agreement on the "reversed rómen/lambe
                    > hypothesis". My hypothesis that they're reverted rómen/lambe seems not
                    > to convince you, and your hypothesis that they're independent tengwar
                    > won't convince me. I see underlying rómen and lambe, and I find
                    > arguments for it; you see independent tengwar, and you find arguments
                    > for it.
                    >
                    > Måns and me wrote alternatly:
                    > > > Some more samples of DTS 17 being closer to the book-hand: The
                    > > > tengwa hyarmen of the book-hand and of DTS 17 have both a straight
                    > > > longer stroke, whereas it's curved in DTS 18; the tengwa esse
                    > > > looks more composed in the book-hand and in DTS 17 than in DTS 18;
                    > > > the down part of the tengwa rómen has a very marked lambe-like
                    > > > curl in DTS 18 whereas it doesn't in the book-hand and in DTS 17.
                    > >
                    > > I agree that the letters in DTS 17 are much more constrained in
                    > > their design than those in DTS 18.
                    >
                    > At a first glance I thought you were admitting that DTS 17 is closer
                    > to the "radical form" and that therefore, you'd have to agree that the
                    > unambiguous "rómen nuquerna" of DTS 17 absolutly must be considered.
                    > At a second glance, I realize that you'll still not agree since you've
                    > only conceded that it is "constrained" and not that it would be closer
                    > to the "radical form"...
                    >
                    >
                    > > But the deviations of DTS 18 are very small, and often not greater
                    > > than in, say, DTS 10 (which I think is a prime example of the
                    > > book-hand).
                    >
                    > I was only thinking of (that is, comparing to) the chart from app. E,
                    > DTS 9.
                    >
                    > > None of the examples you give (<hyarmen> being more
                    > > curved, <esse> being "less composed", <rómen> having a larger lower
                    > > curl) remove any significant traits from the letters.
                    > >
                    > > I argue that if the /W/ tengwa of DTS 18 is meant to be a <rómen
                    > > nuquerna>, then it has lost the significant trait of the inward bow,
                    > > and thus been distorted more than any other letter in that sample.
                    >
                    > The significancy of the traits is a circular argument: They only
                    > become significant because you believe that it's different tengwar.
                    >
                    >
                    > > If the pointed style is
                    > > indeed closer to the "radical form" than the decorated verse-hand,
                    > > shouldn't we somewhere, in some sample, expect a "genuine" turned
                    > > <lambe>?
                    >
                    > That's right, my hypothesis implies a still more "radical", but
                    > unattested form where lambe and "lambe nuquerna" look identical. I'd
                    > only expect a "genuine" "reversed lambe" on DTS 9.
                    >
                    >
                    > > > > the pointed-style /W/
                    > > > > tengwa may have gotten this shape for many reasons (desire to
                    > > > > distinguish it more clearly from <hyarmen>, to rely less heavily
                    > > > > on a thin /-slanting line, or
                    > > >
                    > > > These arguments exclusively apply to DTS 17; why don't they apply
                    > > > in DTS 18 as well? I'd like to have arguments that don't depend on
                    > > > DTS numbers or on:
                    > > >
                    > > > > merely artistic license).
                    > >
                    > > Well, suffice to say that the one, singular attested sample of a
                    > > /W/-tengwa with the shape of a turned <rómen> remains a mystery (if
                    > > not a very great one), and that all other samples form a continuum
                    > > leading up to the "cursive <hwesta sindarinwa>" in DTS 37.
                    >
                    > In science, you only need a single counterexample for a theory to be
                    > rebutted. (For sure, this is not science but philology, that is,
                    > interpretation.) I claim that my interpretation produces no mysterious
                    > left-overs.
                    >
                    >
                    > > > Do the shapes required by a reversed rómen occur?
                    > >
                    > > Well, a curve possibly corresponding to the bow of a reversed
                    > > <rómen> may perhaps be found in <silme> or <vilya>, or the "e"
                    > > tengwa of this mode. The rest of a reversed <rómen> (the upward
                    > > swash and the right leg) could of course be written as in the
                    > > /W/-tengwa.
                    >
                    > I've found a sample of a straightened rightward bow, that is, of the
                    > same alternation as between the "rómen nuquerna" tengwar of DTS 17 and
                    > 18: Compare DTS 20 rómen with other rómen, or compare the first three
                    > occurences of rómen in the second line of DTS 51 with the fourth one
                    > (you'll probably insist that it's unique)!
                    >
                    > The Lowdham-texts show also another example of a sign with variable
                    > heighth: the Anglosaxon &-sign: It's short in DTS 50 and tall in DTS 51.
                    >
                    > And that's not enough: There are also samples of that look like an
                    > inverted _2_: In the second part of DTS 50, the deleted one on line 25
                    > (you'll probably argue it isn't a lambe at all) and the very last one
                    > of line 27 (you'll probably argue it's really composed like the others).
                    >
                    > The mere existence of single instances of these variations make it
                    > more probable that the same kind of variation is found in the supposed
                    > "lambe nuquerna" and "rómen nuquerna".
                    >
                    >
                    > > I don't think your arguments for the /W/-tengwa in
                    > > DTS 18 being a distorted <rómen nuquerna> are very strong either,
                    > > and in my opoinion they are further weakened by the occurrence of a
                    > > closely similar tengwa in other sources.
                    >
                    > You're speaking of the wh-tengwa in DTS 37, right? This tengwa isn't
                    > any more similar to the one of DTS 18 than the one of DTS 17 is.
                    > However, the similarities of DTS 17 and 18 are undoubtingly much
                    > bigger than the ones between DTS 18 and 37, so I tend to associate the
                    > wh-tengwa of DTS 18 with the one of DTS 17. Especially because DTS 37
                    > uses the same w-tehta as other later modes, vala, I believe that the
                    > wh-tehta of DTS 37 is a cursive form of hwesta sindarinwa (coexisting
                    > with the straight form and probably predating it as an intermediate
                    > form between from the "rómen nuquerna").
                    >
                    >
                    > > I don't require that turned forms of letters are in perfect symmetry
                    > > with their normal form. In fact, even if we did not know it for a
                    > > fact, <esse nuquerna> has many characteristics that strongly imply
                    > > that it is a turned <esse>: it has the same sound value as <esse>,
                    > > it is used as an allograph of that tengwa in several sources, and in
                    > > all sources it is consistently drawn as two bows turned opposite to
                    > > those in <esse>, clearly hinting at a relationship. The supposed
                    > > <rómen nuquerna> lacks all these characteristics.
                    >
                    > It's obvious that it lacks the first two characteristics of bearing
                    > the same sound value and being an allograph. It's an important point
                    > in my argumentation that the supposed "rómen nuquerna" and "lambe
                    > nuquerna" have different sound values and are not allographs of rómen
                    > and lambe. Looking at DTS 21 (the only sample I remember that has both
                    > except for the charts), however, I think that the differences in shape
                    > between esse and esse nuquerna are as big as the differences in shape
                    > between the lambe/rómen etc.: Esse looks like two silme, esse nuquerna
                    > like a _3_.
                    >
                    > And still another difference between an upright tengwa and a reversed
                    > one: In all instances of malta in DTS 18, the closing horizontal
                    > stroke begins in the telco, whereas in all instances of nwalme, it
                    > exeeds it, most of the times very markedly. I believe it's common for
                    > reversed forms to deviate from the upright forms. Have you ever seen
                    > anybody drawing a schwa or a reversed a (the IPA signs corresponding
                    > to [6] or to [Q])? I don't call these "distorsions", but "natural"
                    > deviations.
                    >
                    >
                    > > > Of course, the similarity is the most important argument for a
                    > > > closer relationship of these letters. But it's not the only thing
                    > > > that speaks for a closer relationship: Another hint is the fact
                    > > > these letters were abandoned, that is, "lambe nuquerna" was
                    > > > totally abandoned and "rómen nuquerna" was reanalyzed and changed
                    > > > in shape. I don't know of other abandoned tengwar at all.
                    > >
                    > > Well, that might be just a matter of scarce sources. But speaking of
                    > > abandoned tengwar, the initial "schwa" tengwa comes to mind (the one
                    > > looking like a stem with a horisontal bar connecting it to the
                    > > following tengwa). (In "a merry", DTS 16-18.)
                    >
                    > Right, this tengwa wasn't used in DTS 24 any more. There is also much
                    > variance in the representation of schwa! But the schwa representation
                    > was already complex in the "Bombadil modes", so I'm not surprised that
                    > it changed, whereas according to your hypothesis, there were two
                    > clearly distinct tengwar for /w/ and /W/.
                    >
                    >
                    > > > What development is parallel to a "reversed rómen" shape
                    > > > originating in a hwinde/hwesta sindarinwa shape? I don't know any
                    > > > at all.
                    > >
                    > > I can't think of any either, but I don't find that very surprising
                    > > given that we only have two document pages that constitute samples
                    > > of the pointed style (except the "Namárie" headline in DTS 20).
                    >
                    > It doesn't need to be one of the "few" parallels in tengwar, I'd also
                    > be interested in one of the "many parallels in Roman writing" you
                    alluded.
                    >
                    >
                    > > It is probably true that DTS 16-18 predate the Etymologies, and the
                    > > <waiya>-tengwa might be "experimental", as you say. Still, it
                    > > doesn't feel to me like a probable scenario that <hwesta sindarinwa>
                    > > was developed from reanalysis of <rómen nuquerna>. Note that it is
                    > > the "upright", normal <hwesta sindarinwa> of DTS 9 that is found in
                    > > the Etymologies, not the cursive form of DTS 37.
                    >
                    > If I am to uphold that hwesta sindarinwa is a reanalysis of the
                    > supposed "rómen nuquerna", then I have to assume that the
                    > "intermediate" form seen in DTS 37 already existed when the
                    > etymologies were written (as written above).
                    >
                    >
                    > > (By the way, if the Etymologies postdates DTS 16-18, then <waiya>
                    > > would be developed from the /w/-tengwa of the "Bombadil-mode",
                    > > rather than the other way round. The later tengwa is thus "closed",
                    > > made *less* "lambe-like" as you put it. This should be an
                    > > interesting topic of research for a later time.)
                    >
                    > Indeed! (I hope that one day we will be able to look back and wonder
                    > how we could speculate so much on so few published samples! ;-)
                    >
                    >
                    > > I have the feeling Tolkien usually
                    > > had his own language foremost in his mind when developing his
                    > > scripts -- either that, or phonetic applicability in general.
                    >
                    > I share the feeling that most of it was based on English!
                    >
                    >
                    > > Of the
                    > > published Tengwar and Sarati texts written in the first half of the
                    > > 20th century, the vast majority is in English.
                    >
                    > Published, but alas, not publicly available... when PE 13 was still
                    > available, I wasn't that much into Tolkien's scripts yet, and now I
                    > regret it so much!
                    >
                    >
                    > > > Your hypothesis doesn't explain why he abandoned them if they were
                    > > > distinct from rómen and lambe.
                    > >
                    > > As stated earlier, I don't think he abandoned the /W/-tengwa seen in
                    > > DTS 16-18 -- in my opinion the /W/-tengwa of DTS 37 is the same one.
                    > > As for the /w/-tengwa, I concede that it may very well have been a
                    > > passing idea, but I don't think it must have had anything to do with
                    > > <lambe>.
                    >
                    > Are you going to tell me you believe that the wh-tengwa of DTS 18 is
                    > an allograph of hwesta sindarinwa? Is it the same tengwa? If it is,
                    > what happened that it changed so much? If it is not, why was it
                    > replaced? (See also above.)
                    >
                    >
                    > > I may be confused, but I can't find any evidence in DTS 25 (The
                    > > Doors of Durin (Inscription Drafts)) of the <vala>/<hwesta
                    > > sindarinwa> solution. What am I missing?
                    >
                    > My bad, I thought of DTS 24, but I had the wrong DTS number in mind (I
                    > still don't know them all by heart :-).
                    >
                    >
                    > > > My hypothesis explains why the "lambe nuquerna" was abandoned.
                    > >
                    > > I am sure I have commented on this somewhere already. ;-)
                    >
                    > Yes you have. You've said: "As for the /w/-tengwa, I concede that it
                    > may very well have been a passing idea". From this, I understand that
                    > you can't explain it.
                    >
                    >
                    > > > > > (2) It explains why there was much variations in the sign for
                    > > > > > w and wh.
                    > > > >
                    > > > > The variation is not that big, in my opinion. It is usually just
                    > > > > a matter of swoop. :-)
                    > > >
                    > > > It's in comparison to other sounds.
                    > >
                    > > Yes, you are right about that. The /w/ sign of DTS 16-18 does show
                    > > an unusual amount of variation.
                    >
                    > I've missed the point again (gee, it's difficult to talk about these
                    > things without using your fingers!): Looking at all the tengwar
                    > samples, the sounds /w/ and /W/ have a bigger number of signs they can
                    > be represented with (and not just in terms of allography) than the
                    > other sounds. My hypothesis takes this into account.
                    >
                    >
                    > > > My hypothesis
                    > > > explains the similarities and the differences between the
                    > > > "Bombadil mode" wh-tengwa and hwesta sindarinwa.
                    > >
                    > > While I agree that <hwesta sindarinwa> may have originated in a
                    > > cursive form, I am not convinced that that cursive form in turn must
                    > > have originated in a <rómen nuquerna>.
                    >
                    > The cursive form hwesta sindarinwa originated in would be the one of
                    > DTS 37.
                    >
                    >
                    > > To return to the origins of this thread: in a font that follows the
                    > > 'pointed' style, the /W/-tengwa (for lack of a better name) should,
                    > > in my opinion, appear as a turned <rómen>. But if the font follows
                    > > rather the book-hand, the /W/-tengwa should be based on the forms in
                    > > DTS 37, 23, and 18.
                    >
                    > I disagree! I'd suggest to have two different wh-tengwar for each
                    > font: In the angular style, a "rómen nuquerna" and a hwesta
                    > sindarinwa; in the upright book-hand style a "rómen nuquerna" (similar
                    > to the one in DTS 18, but maybe a little bit more like the one in DTS
                    > 17) and a hwesta sindarinwa; in a cursive book-hand style a "rómen
                    > nuquerna" (as in DTS 18) and a hwesta sindarinwa (as in DTS 37).
                    >
                    > I believe it's important to keep these apart, just as you should keep
                    > the w-tengwa of the "Bombadil-modes" apart from later w-tengwar.
                    >
                    > ---------------------------
                    > j. 'mach' wust
                    > http://machhezan.tripod.com
                    > ---------------------------
                  Your message has been successfully submitted and would be delivered to recipients shortly.