Unified Canadian Syllabics

From: chris@languagegeek.com
Date: Mon Feb 09 2004 - 18:28:17 EST

  • Next message: Kenneth Whistler: "Re: Public Review Issue #27"


    I think I posted this to the list last week, but I haven't seen it come up.

    I would like to present to the Unicode community some suggestions for
    missing and mis-named characters related to the UCAS range. To properly
    describe the kinds of characters missing etc., many graphics are required.
    For this reason, I would invite people to see the document at:


    A "words only" description follows here:

    I would like to suggest to the Unicode community the following observations
    relating to the Unified Canadian Aboriginal Syllabics range. My goal (see
    www.languagegeek.com) is to enable all of the North American languages to
    be properly and accurately written on the Internet, and computers in
    general. Here I will focus specifically on the languages which are
    currently using or historically used (and may still be in some communities)

    Some conventions used below. All Unicode character names are in majuscule,
    and “Canadian Syllabics” has been abbreviated to CS. Hexadecimal Unicode
    indices are in parentheses and prefixed with “U+”. All sources cited are
    linked to the languagegeek.com bibliography. A “final” is the Syllabics
    term for a character which represents a consonant only, not a consonant +
    vowel, so CS FINAL GRAVE (U+1420), CS CARRIER H (U+144B) and CS NASKAPI SKW
    (U+150A) would all be examples of “finals”. I use the term “syllabic” to
    refer to a consonant + vowel character. A series is a row of characters on
    a syllabic chart, so in Misnamed Characters Note 1, “tta, tte, tti, tto”
    would be the tt-series.
    Misnamed Characters
    The asterisk ᕯ character (U+156F) appears on the code-page chart as
    **, and is named CS TTH. This is a misreading of the syllabarium chart used
    by the French Missionaries for Chipewyan—probably from the 1904 publication
    Prières Catéchisme et Cantiques en langue Montagnaise ou Chipeweyan. The
    chart in this book has been reprinted in most if not all “scripts of the
    world” type books. Unlike most other syllabics charts, this one does not
    have a column of finals to the right of the consonant-vowel syllabics.
    Instead, it simply has a list of all the finals, which do not correspond
    with the syllabics series on the same row. Thus, the CS WEST-CREE P
    (U+144A) (looks like a prime ') final which appears to the right of the
    “tta” row is not the sound “tt”, but is instead “h”. The blue circled
    asterisk is not “tth”, but is in fact a symbol which indicates a proper
    name, in this case /*adą/ (Adam). A second glitch on the Unicode
    code-page chart is that this character is written with two asterisks “**”,
    when in fact on the chart above, the first asterisk is the character
    itself, and the second is part of the example. I believe this should
    definitively be fixed.

    In the syllbacs chart mentioned above, the final row in the chart is
    labelled “tca, tce…”, (U+1570-73) which corresponds to the modern Roman
    orthography sound /t/ (an aspirated stop). Interpreting “tca” as “tya” is a
    misunderstanding of the French description of what the c represents. The
    Chipewyan Syllabarium page has more info on this. Whether this syllabics
    series is renamed is probably not a high priority.

    In Naskapi, each a-type syllabic character can either be preceded by a
    colon-like character, or have a umlaut-like diacritic. Unicode has labelled
    these as having a long vowel: e.g. (U+1482) CS NASKAPI KWAA. In fact, the
    colon or umlaut does not mark vowel length (Naskapi orthography ignores
    length). Instead, the colon or umlaut simply indicates “wa”. So (U+1482)
    would be better named CS NASKAPI KWA. This is also probably not a high

    Missing Characters

    According to the Naskapi Lexicon, there is no symbol NASKAPI WOO (U+1416),
    but there is a “wi”. This character look similar to U+140E CS WI, but is
    different—the dot is higher up on the left side. “wi” may need to be added.
    “woo” may be on a different Naskapi chart I have not seen.
    In Blackfoot, a raised “equals sign” is used much as the “CS FINAL MIDDLE
    DOT” (U+1427) is in Cree: to indicate a /w/ between the consonant and vowel
    of the syllabic. A raised = with “CS BLACKFOOT KA” (U+15BD) before it,
    gives the sound /kwa/. This character is vital to writing Blackfoot, should
    be added.
    Carrier Dene
    A few finals are missing from Unicode which are used in Carrier.
    Information for Carrier is from Poser 2000. There is an important graphical
    distinction between the finals used for /s/ and /s(+macron-below)/ (in the
    Roman Orthography version). The former is a small serif “s” written
    mid-line, while the latter is a small sans-serif “s” written mid-line.
    Unicode lists only one version (U+1506) CS ATHAPASCAN S. A second
    character, an upside-down mid-line small “h” is used for load words with
    /f/ or /v/ sounds. These two finals should be added.
    In examples of Carrier, the finals are virtually always mid-line. This is
    purely stylistic, but see Dene note 2 about how final placement is
    phonetically important. This example is from Morice 1894, note that all of
    the finals are mid-line, not top-line as in Unicode.
    Dene (Chipewyan, Slavey, Hare, Beaver)
    In early Chipewyan texts, a-finals — like those used in Eastern dialects of
    Cree — were used instead of the western-style finals which were employed in
    all later Chipewyan texts. For the most part, Unicode includes these finals
    under names such as CS K (U+1483), however, some of the Chipewyan series
    are not to be found in Cree, and are thus absent. The two missing finals
    are a raised small version of CS WEST-CREE LA (U+154D). and a raised /ga/
    syllabic. The g-series here corresponds to CS SAYISI HE, HI, HO, HA
    (U+15C0–15C3). Though no longer used today, for historical purposes, these
    characters may be added.
    In Dene texts, it is vitally important to distinguish between the position
    of the finals, top-line, mid-line, or bottom-line. The same shape character
    may have a completely different value depending on how high up from the
    baseline it is. Take for example the Chipewyan character which resembles CS
    WEST-CREE M (U+14BC). The top-line is a diacritic which indicates that the
    following sound is pronounced "like a fricative" or "with a velar fricative
    release", e.g. /ł/ instead of /l/. The mid-line version has the sound

    The example above comes from the Syllabic tradition instigated by French
    Catholic Missionaries. The English Anglican style of writing Dene is
    different in some important ways. With regards to final placement, where
    the French system distinguishes top- and mid-line finals, the English
    differentiate between top- and bottom-line. For example, if a top-line CS C
    is combined with a following CS WEST-CREE LE (U+1544), it produces /tli/
    (where /tl/ is a single phoneme). The base-line final is /t/ alone. A
    top-line CS SAYISI TH (U+14A2) indicates the following l-series is to be
    pronounced /ł/, but a baseline CS SAYISI TH is pronounced /th/.

    This system holds true for many other finals as well. These examples should
    make it clear that the Dene Syllabics system distinguishes top-line from
    mid-line and baseline finals in very important ways. I have not included
    all of the characters which can change position to make different phonemes
    (but I could if necessary). As for how to render this in Unicode, I see two

    i)Encode three unique characters for each final, i) top-line, ii) mid-line,
    iii) baseline. First, this solution is not terribly encoding-efficient.
    Also, it sets up different encodings for glyph variants of different
    styles. For example, some Cree fonts place the CS FINAL RING (U+1424) at
    the top-line, others at mid-line. This differentiation should be based on
    the font, not on Unicode encoding. An opentype substitution would be a much
    better solution for stylistic differences I think.

    ii)Add three non-spacing “characters” to the UCAS range, which tell the
    software where to put the final: top, mid, or base. This also allows much
    more flexibility for writers who want control over how high up their finals

    English-tradition Dene syllabics also includes an “accent”. This accent can
    be either acute or a tilde (depending on the font). It appears in extremely
    limited contexts (in my data, always above CS THA, for the postposition
    meaning “with”). I don’t believe a series of pre-composed accented
    syllabics is required, however, a non-spacing CS ACCENT would be
    sufficient. Note that using COMBINING ACUTE ACCENT (U+0301) and COMBINING
    TILDE (U+0303) may not be a good solution, as in Dene, there is no real
    orthographical difference between acute and tilde (much like Greek tilde
    and reverse-breve). Beaver also has the accent, but it does not appear
    above the syllabic, instead it shows up after the syllable (looking exactly
    like CS FINAL ACUTE [U+141F], which I believe suffices to represent the
    Beaver acute), and much more frequently. I do not have data as to what the
    Beaver acute represents phonologically.

    Dene also requires the dot accent diacritic, although not to mark long
    vowels. In Chipewyan (English tradition) and South Slavey, a dot on an
    o-vowel syllabic changes the vowel to “u”. While in Beaver, a dot over an
    a-vowel syllabic is pronounced as a “y” off-glide to the vowel. Although
    these characters and others should have unique Unicode encodings to be
    consistent with the rest of the UCAS range, if there is a non-spacing dot
    diacritic, this ought to suffice.

    Several dialects of Ojibway in Northern Ontario have a unique means of
    writing finals. It follows the logic of the Eastern finals (a small
    top-line version of the a-syllabic), but instead it employs the i-version
    of the syllabic. The entire series of i-finals should be added to Unicode.
    Some dialects of Ojibway use the raised “l” and “r” finals (U+14EB and
    U+1551) as distinct characters, while others place these above an n-series
    syllabic as a diacritic. Two non-spacing characters would be required to
    accommodate this.
    The UCAS code-chart incorrectly labels U+141E as a Moose Cree (Y). In fact,
    the Moose Cree y-final is a small ring diacritic located above the syllabic
    character. The Inuktitut characters like CS AAI (U+1402) and CS PAAI
    (U+1430) would be read /iy/ and /piy/ in Moose Cree.

    To be consistent, Unicode would have to encode a large number of
    pre-composed glyphs of a Moose Cree syllabic with the y-ring on top—not to
    mention those in combination with CS FINAL MIDDLE DOT (U+1427). Instead,
    perhaps a y-ring non-spacing diacritic would be useful. But U+141E is not a
    Moose Cree y. Some speakers prefer to place the small ring not above the
    syllabic character, but to the right like other finals. This small ring is
    also absent from Unicode, and should be included. This small ring should
    not be confused with the larger CS FINAL RING (U+1424), which in the
    example above appears just to the left of the tail of the arrow.
    The Woods Cree dialect (labelled by Unicode as TH-Cree, U+15A7–U+15AE) does
    not use an Eastern-style, raised a-final, as shown in U+15AE. To my
    knowledge, U+15AE would not be used by anyone. The final I have seen used
    is missing from Unicode. It resembles a raised not-equals sign. This
    character is required to write this dialect of Cree.
    I am curious to know why the n-series of Cree syllabics (U+14C0–14D2) is
    missing half of the w-dot characters, namely: nwi, nwii, nwo, nwoo (both
    eastern and western versions).
    For Y-Cree, please see note 2 under Questions below.

    All languages:
    The correct form of the hyphen in Syllabics is a shortened equals sign.
    Should this hyphen get its own Unicode encoding? A regular equals sign
    doesn’t necessarily look right, and will not wrap properly. A single dash
    hyphen is unacceptable because it would conflict with CS FINAL SHORT
    There are several questions I have about certain characters in the UCAS

    I mentioned above that there are several diacritics which perhaps should be
    non-spacing diacritics. Is it wise to use the standard Roman orthography
    non-spacing range? Or should different scripts have their own accentuation
    (like Greek does).

    What is the source for U+141D CS Y-CREE W. In Y-Cree dialects, the final
    “w” is U+1424 CS FINAL RING. In some dialects of Y-Cree, the y-final is a
    smaller top-line small dot instead of the CS WEST-CREE Y (U+1540). This
    small top line dot combines with the CS FINAL MIDDLE DOT (U+1427) to end up
    looking like a colon. Is U+141D supposed to represent this character?
    Would anybody know the sources used for the characters referenced as
    “Sayisi”? I am completely unaware of certain characters, such as U+14BE,
    U+14BF, U+1541. In general, the Sayisi characters match the
    English-tradition Dene syllabics (as one would expect), but I have never
    come across the three characters above.

    Why are CS FWAA (U+155A–B), CS THWAA (U+1567–8), and CS RWAA (U+154E–154F)
    part of UCAS? What about “fwe, fwi, fwii, fwo, fwoo, fwa”, “thwe…”, and
    “rwe…” (both eastern and western versions)? Why do the long ā glyphs
    merit inclusion? I guess that someone was reading directly off a syllabics
    chart, not realising that the 6 glyphs above were just examples of entire
    series with all the vowels. It would have been more useful to include
    mid-dot syllabic composed characters for CS TH-CREE THE… (U+15A7–D), these
    th-cree th syllabics being in common use, where CS F…, CS TH…, and CS R…,
    are for load words only in Cree. Furthermore, perhaps a combing “mid-dot”
    would be useful for those syllabics which were not encoded as composed
    characters: e.g. “fwe”. Using CS FINAL MIDDLE DOT (U+1427) plus the
    syllabic character (in this case, CS FE [U+1553]) would not space properly.

    In some Dene systems, super script F, V, r, and l are used as finals to
    indicate these sounds from European languages. Carrier Dene also uses a
    regular serif roman “r” for loan words. Should these be encoded in UCAS?
    Are they still technically Roman glyphs?

    That’s about all I can think of at the moment, there may be a few other
    issues I have temporarily forgotten. I would appreciate comments and
    suggestions as to how some or all of these ideas can be integrated into the
    Unicode Standard. All of the missing characters above can be found in the
    Aboriginal Serif font, which can be downloaded from this site. It is a free
    font (although the hinting on the syllabics is a bit off). Other Syllabics
    fonts will be available on this site in the future. My apologies if any of
    these issues have previously been discussed on this or other fora.

    Chris Harvey

    mail2web - Check your email from the web at
    http://mail2web.com/ .

    This archive was generated by hypermail 2.1.5 : Mon Feb 09 2004 - 19:16:28 EST