Re: IPA Null Consonant

From: Karl Pentzlin (karl-pentzlin@acssoft.de)
Date: Thu May 29 2003 - 18:42:19 EDT

  • Next message: Kenneth Whistler: "Re: “book end” or <enclosing characters> in most languages?"

    Am Donnerstag, 29. Mai 2003 um 22:35 schrieb Kenneth Whistler:

    KW> Kent:

    >> Others gave references where it in most cases did NOT look at all like the
    >> empty set symbol.

    KW> Gustav Leunbach (1973), Morphological Analysis as a Step in
    KW> Automated Syntactic Analysis of a
    KW> Text.http://acl.ldc.upenn.edu/C/C73/C73-2022.pdf

    KW> uses an empty set symbol to denote a morphological zero.
    KW> (see p. 272). [Typographically, this could arguably
    KW> have been taken from a type tray for a Norwegian ø
    KW> character, rather than from a mathematical symbol font,
    KW> but this is *clearly* not a slashed zero.] And this is
    KW> a document type set the old fashioned way, with actual type,
    KW> in 1973. See:

    KW> http://acl.ldc.upenn.edu/C/C73/C73-2000.pdf

    KW> bearing the publication logo of Firenze.

    KW> A. S. Liberman (1973), Towards a Phonological Algorithm.

    KW> http://acl.ldc.upenn.edu/C/C73/C73-1015.pdf

    KW> uses an empty set symbol to denote a phonological zero.
    KW> (See pp. 196-197 for numerous examples.) These are
    KW> clear examples, and show that this is used symbolically,
    KW> to indicate a "something which is not there". Look at
    KW> the type style. These are included in *italic* word
    KW> citations, but the null set symbol (used to denote the
    KW> phonological zero), is *not* set in italic.

    KW> Harri Jäppinen and Matti Ylilammi (1986), Associative Model
    KW> of Morphological Analysis: An Empirical Inquiry.

    KW> http://acl.ldc.upenn.edu/J/J86/J86-4001.pdf

    KW> Displays a distinctive usages, with an italic epsilon to
    KW> denote a morphological zero. (Not the same as the set theory
    KW> use of epsilon to denote set membership.)

    KW> You can dig further in these archives of old editions of
    KW> Computational Linguistics and other journals from the 1970's
    KW> to find other instances illustrating the use of the empty
    KW> set symbol in linguistics to denote a phonological or
    KW> morphological zero.

    >> From what I've heard on this thread, a slashed zero glyph appears common
    >> in this situation in linguistics.

    KW> See examples cited above.

    >> A slashed zero is completely
    >> unrelated to the empty set symbol.

    KW> This is nonsense. You have found the correct citations
    KW> on the web regarding André Weil's claim to have introduced
    KW> the empty set symbol, as part of the Bourbaki group. And
    KW> for Weil, the source of the symbol may well be Norwegian ø.
    KW> (What the Weil citation doesn't specify is why he chose
    KW> a symbol vaguely reminiscent of a zero, while not actually
    KW> being a zero, to represent the empty set.) And what I pointed
    KW> out earlier is that, in *linguistic* usage, the slashed zero
    KW> glyph is clearly an acceptable glyphic variant of the
    KW> empty set symbol. So to claim it is "completely unrelated"
    KW> is to manifestly ignore actual practice.

    >> The empty set symbol and slashed zero remain unrelated.

    KW> Another bald assertion contradicted by Pullum (1996), who
    KW> *does* relate them, in linguistic usage. Nobody is claiming
    KW> that in *mathematical* usage they are connected, or would
    KW> be acceptable alternative glyphs in a treatise on set theory.

    >> [The EMPTY SET symbol] does not appear to have wandered
    >> into linguistics in any way (except by occasional typographic mistake,
    >> and that does not count), even though there is use of a similar-looking
    >> symbol.

    KW> What you are missing here is that the use of the empty set
    KW> symbol in linguistics is associated with structuralist
    KW> linguistics, which was in intellectual development roughly
    KW> contemporaneously with the Bourbaki group. And structuralist
    KW> morphology, in particular, was influenced by formal set
    KW> theory, and many morphologists borrowed the kind of formalisms
    KW> used by logicians and set theoreticians.

    KW> A phonological zero or a morphological zero has nothing to
    KW> do with numeric values, nor is it conceived of as part of
    KW> a word, per se. It is a pattern gap, an absence, a set with
    KW> no elements. And while I can't track you back, from web
    KW> citations to some earliest usage and give you a morphologist
    KW> explicitly talking about his notational conventions, without
    KW> spending more time at it than I can manage today, I can
    KW> assure you that it is perfectly reasonable and expected to
    KW> find clear examples of use of the empty set symbol in this
    KW> linguistic usage.

    What's the reason that this thread becomes that long and emotional?

    There is a need to express some emptiness/missingness concept by a
    symbol resembling somehow a circle or vertical oval form, overlayed
    by a slash.

    At present, Unicode has not a character which fulfills this need
    uniquely and unanimously (as this thread shows).
    If there was a need to include such a character into Unicode, this
    would have happened long before (considering the many linguists here),
    or at least nobody would have objected to the idea as it was expressed
    in this thread.
    Such, there seems to be *no* need for a dedicated symbol. As a
    consequence, other symbols are sufficient and cannot be called
    "wrong" without looking at the context of their actual usage.

    If you want to express the concept "empty phoneme/morpheme/whatever",
    use any symbol which is unambiguous in *your* context.
    Use U+2205, U+2298, U+A01C or whatever. If these characters are
    missing or ugly in your font, use U+00D8, as long as this is
    unambiguous within *your* text. Or create an OpenType font with your
    favourite glyph for U+0030 U+0338 if you have time and resources.

    *Every* symbol which your readers interpret correct *is* correct.

    - Karl



    This archive was generated by hypermail 2.1.5 : Thu May 29 2003 - 19:30:32 EDT