Re: Ciphers (Was: Berber/Tifinagh)

From: John Hudson (
Date: Mon Nov 10 2003 - 17:53:19 EST

  • Next message: Patrick Andries: "Re: Berber/Tifinagh (was: Swahili & Banthu)"

    At 01:57 PM 11/10/2003, Peter Kirk wrote:

    >Define "cypher", or "cipher", and I will either provide evidence that the
    >Theban script is not one or accept that, on your definition, it is one. In
    >the absence of a definition this discussion is meaningless. Similarly if
    >the definition is simply a whim as you implied, so a personal subjective
    >choice against which there can be no evidence. Was it a whim that Theban
    >and Klingon were rejected?

    There is a lot of philosophical ground between a 'whim' and something that
    is so clearly defined that it engenders no debate. I think a definition of
    cipher that focuses on a deliberate representation of a language with a set
    of signs that is different from that which is the conventional
    representation of the language by the vast majority of its users is
    sufficient. A working definition doesn't need to eliminate all grey areas:
    it is useful enough if it identifies what the grey areas are. We can add to
    this definition that ciphers are *often* intended to be a secret way of
    writing, even if cryptologically weak, or we can suggest that ciphers are
    *usually* used by people who are also familiar with and in other
    circumstance use the same conventional representation of the language as
    everyone else.

    I think it is also important to note the presence in this discussion of
    three different phenomena:

    1. Writing systems that appear to be constructed as ciphers, e.g. Theban.
    2. Writing systems that are borrowed and used as ciphers, e.g. the Masonic
    use of Samaritan.
    3. The use of an unconventional set of glyphs to represent Unicode
    characters, e.g. Philippe's proposed use of Latin letters to display
    Tifinagh characters.

    Michael referred to the latter as a cipher, and so triggered the present
    discussion. I prefer the font developer term 'masquerade' for this
    phenomenon. Although the effect of Philippe's suggestion would be analogous
    to 2, above -- i.e. using the Latin writing system as a cipher for Tifinagh
    --, the intent is really transliteration. Ciphers are appropriately handled
    at the glyph processing level, which is why Theban doesn't need to be
    encoded, but transliteration should occur at the character level.

    John Hudson

    Tiro Typeworks
    Vancouver, BC

    I sometimes think that good readers are as singular,
    and as awesome, as great authors themselves.
                                           - JL Borges

    This archive was generated by hypermail 2.1.5 : Mon Nov 10 2003 - 18:41:31 EST