Re: Rare extinct latin letters

From: Philippe Verdy (
Date: Thu May 29 2003 - 21:37:41 EDT

  • Next message: Mark Davis: "Re: Not snazzy (was: New Unicode Savvy Logo)"

    Interesting sources here. Another proof that attempts to create a simplified orthograph that "could" be read and written more simply is an old subject for linguists.

    So both sources attempt to create a L molle for the "ille" used in French like the Dutch "ij" digraph. Baf uses a cedilla under the L letter, but Meigret creates a very distinct variant of L which looks like a left-side hook.

    I did not know the "E with hook below" used by Megreit, as a form for a "long/open e", as opposed to the "short/closed E" still used today for final E.

    Megreit does not recode the "ou" digraph, but Baf creates a letter that looks similar to an O vowel with a ogonek.

    Nazalized vowels use a tilde above instead of a n after. The French "ch" digraph is rewritten with the same letter uses for "s" and comming from the Greek final Sigma. There seems to be a confusion in this document between the usage of the cedilla or hook below (depending on the font used in the roman or italic style).

    There are also ligatures for the long form of s (used when it is not final) that I did not see before (the double long s ligature).

    Meigret removes the implicit "u" after q, but it keeps the same letter for the current "u" vowel and "v" consonnant.

    I also see what seems like a strange form variant for a N with tilde (N molle), and looks like a n with a long leg and a rotated i above.

    But when I look more closely, the exact form of each letter is nearly unique and it's difficult to see what would be the prefered intended form (notably because of an approximative positioning of the metal font in the press and irregularities in the manually manufactured fonts, or variations in the quantity of ink and granularity of the paper used).

    So the L molle could be really a L with accute accent or a tilde above. The case of the "N molle" is strange difficult. Phonetically it does not seem to be a modified "eng", but clearly the long leg is present; however its diacritic is ambiguous or does not ressemble to the current form of the tilde (it may just be a form variant of the tilde).

    So the "molle" on L and N may simply be a tilde.

    In Baf, it seems that the A with long trail is a ligature between A and U, used as a distinct letter sorted after "O" because it has a similar sound, only longer. The same type of ligature between E and U is used to create a distinct letter sorted very strangely after Z. So these ligatures can be viewed as a diacritic added to A or E, meaning a more closed and longer sound. (And probably used also for the pair G and U). So the Baf addition is a new diacritic on the right side, which looks like the current Unicode hook.

    Baf also suppresses the Q (probably replaced by K, the same way it replaces the C pronounced K), and uses the cedilla under C as a general way to unify it with the Greek final sigma for the long S sound.

    There's a new ogonek diacritic attached above O, to mean the OU long and closed sound (the combination closely ressemble a Greek gamma). This combination could also be represented by a O with a hook diacritic.

    Both authors seem to segregate the usage of G and J (Baf always uses a hook on the G, and there's no "normal" G). And "superfluous" doubled consonants are removed in their texts.

    From: "Patrick Andries" <>

    This subject seems to come periodically on French typographical lists, so I would like to see what might be the answer of Unicode(unicore) to it.

    What should be done with rare extinct latin letters which usually can't easily be mapped to a single modern letter (i.e. they are not simply glyph variants) ?

    Could letters like l molle ( or long-tailed A (between O and P in Baf's alphabet, letters which I believe cannot be composed from other existing Unicode characters, be considered for Unicode encoding or, since they are no longer used and were rare when introduced, should they be seen as typical candidates for the private use area? Or is the answer again going to be : use a higher level protocol to map to a modern form (of the letter or the modern spelling of the word?) even if you want to preserve and search the original format.

    Incidentally, Baf proposed a letter for OU encoded in Unicode (U+0222 and U+0223).

    This archive was generated by hypermail 2.1.5 : Thu May 29 2003 - 22:15:04 EDT