From: Kent Karlsson (kentk@md.chalmers.se)
Date: Thu May 29 2003 - 08:19:01 EDT
Jim Allan wrote:
> Kent Karlsson posted:
>
> > And I (still!) very strongly disagree. The empty set symbol stands
> > for the empty set (also written {}). But there is no set here, let alone
> > an empty one. Possibly an empty string (of phonetic symbols?).
> > Written as '' or "" in your favourite programming language, and
> > conventionally written as a lowercase epsilon (ε) in math contexts.
> > (That does not make the empty string equal to a string consisting
> > of the letter ε, of course!)
>
> No one claims that empty set symbol used by logicians for empty set is
> used by linguists with exactly the same meaning.
True; but I claim they are not using the empty set symbol at all...
> The glyph ! is use by mathematicians to indicate "factorial", by
Yes. Not just the glyph, but the exclamation mark character;
there is no math exclamation mark character defined...
> phoneticists to indicate an aleveolar or postalveolar click, and by
Those have separately encoded *letters* in Unicode. The one looking like an exclamation mark is:
01C3;LATIN LETTER RETROFLEX CLICK;Lo;0;L;;;;;N;LATIN LETTER EXCLAMATION MARK;;;;
(The other click letters apparently should have different-looking glyphs.)
In addition, when the integral sign is used in IPA, it's got its own letter character:
0283;LATIN SMALL LETTER ESH;Ll;0;L;;;;;N;;;01A9;;01A9
> programmers in c and some other programming languages to
> indicate "not".
Just because they wanted to use an ASCII/EBCDIC character that
would be present in *all* extensions of ASCII or EBCDIC. Otherwise the
logical not sign (U+00AC, ¬) would have been a better choice.
> Such overloading of symbols between disciplines (and even within
> disciplines) is common.
Granted (but I don't see any evidence of that for the empty set symbol).
> > But capital "slashed o" (U+00D8) is not mentioned... And that letter
> > would be entirely appropriate for this purpose **in the contexts** where
> > it would stand for a "null consonant" (or empty string) in linguistics.
>
> It is not clear to me why the empty set symbol, which at least as the
> idea of emptiness associated with it, should be more inappropriate for
> use in linguistics for null character(s) than capital O-slash (Ø) which
The empty set symbol is a math symbol, not expected to ever occur (properly)
in a word-like context. Capital O with stroke, however, is a letter, and can easily
and without any problems occur in a word-like context.
> is a consonant in a real language and as such as no suggestion of
...a vowel...
> emptiness about it, especially not to linguists who recognize its
> lowercase form as part of IPA.
IPA and other phonetic writings are AFAIK always lowercase; so the
uppercase form can be used in another meaning in those contexts. Indeed,
even open/closed variants of the same letter are used in different (though
non-null) meanings in IPA.
> Also, linguists might be dealing with Norwegian and may wish to use
> actual Norwegian spelling in their explanations.
Then you are in trouble, if you/they want to use so similar (or indeed same)
symbol for two different things. In such cases one would choose to use a
completely different symbol (letter) for "empty", like Greek Capital Omega,
or something even more distinctive.
> > But capital "slashed o" (U+00D8) is not mentioned... And that letter
> > would be entirely appropriate for this purpose **in the contexts** where
> > it would stand for a "null consonant" (or empty string) in linguistics.
>
> Almost *any* character not otherwise used could be
> appropriate **in the contexts**.
>
> > It 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.
>
> Can this supposition be documented?
Others gave references where it in most cases did NOT look at all like the
empty set symbol.
> I thought the opposite, that the slashed zero form that sometimes
The empty set symbol has nothing to do with a slashed zero, never has.
> appears in linguistics was a variant mathematical null set symbol, that
> the evolution was the opposite to what you suggest.
Sorry for picking on every statement you make, but there is no such thing
as a "null set" or a "null set symbol" (null and empty aren't the same).
> > I think it would be less problematic to use the letter Ø for the empty
> > set (in a math context), than to use the EMPTY SET symbol (Ø) for any
> > linguistic entity in a word-like linguistic context.
>
> But it *is* being so used and has been used for quite some time. The
Well, no...
> word "problematic" is puzzling. What problems does this usage cause?
Math formulas often contain letters; indeed it is extremely common, not to
say ubiquitous. But math symbols, on the other hand, never occur in words.
> From the web page http://www.cl.cam.ac.uk/~mgk25/ucs/secs.html by
> Markus Kuhn on the empty set symbol:
>
> > # Used in technical drawings and on product descriptions. Note that
> > # DIAMETER SIGN is an exact circle while EMPTY SET is often a digit zero,
> > # both with a stroke.
>
> Markus Kuhn distinguishes the diameter symbol from the empty set symbol
They have different origin (even though a unification would have been possible, years ago).
> but considers slashed zero as just a variant of the empty set sign,
I cannot imagine why. They have no historic or typographic relationship
other than looking slightly similar (if your vision is blurred).
> presumably from the same kind of glyph variation I have also seen in
> practise and which Ken Whistler commented on.
From what I've heard on this thread, a slashed zero glyph appears common
in this situation in linguistics. A slashed zero is completely unrelated to the
empty set symbol. The latter is, however, closely related (as a matter of original
design) to the capital o with stroke (and indeed capital o with stroke is used to
denote the empty set, nothing wrong with that).
> The web page http://www.brl.org/formats/rule18.html provides official
> Braille translations for IPA type symbols, including a braille symbol to
> be used for either the slashed zero or round slashed circle
> glyphs with the notation:
>
> > slashed zero, null or empty set
The empty set symbol is still completely unrelated to a slashed zero (even if
Markus and somebody else managed to confuse them; a mistake easily made).
> Personally I prefer the slashed zero for null character(s) in
> linguistic contexts.
Fine. So use that. (Which is entirely different from the empty set symbol.)
> I don't know if it also occurs in mathematical contexts as a null set
> symbol.
Only in bad typesetting, if at all.
> Whether in linguistics the slashed zero should be considered a glyph
> variant of the mathematical empty set sign or whether the
I think not.
> slashed zerio
> is a symbol unto itself (distinct from both the empty set sign and
> normal zero) is something for practising linguists to argue over or
> agree on.
>
> That the slashed zero glyph (used for null character(s) in linguistic
> texts) is to be distinguished from normal zero in linguistic texts is
> easy enough to demonstrate.
>
> Are there also linguistic texts that distinguish slashed zero
> from the
> mathematical empty set sign, giving different meanings to each?
>
> If so, then someone who wishes for Unicode to include slashed
> zero as an
> independant character should make a formal proposal to Unicode with
> sources to back up the difference in use.
"Slashed zero" is already representable as <DIGIT ZERO, COMBINING LONG
SOLIDUS OVERLAY>. Would that representation somehow be inappropriate?
(Such as having digits in words...)
> Even if linguists in general feel that the empty set form which often
> appears for (null character(s)) is a kludge for the proper
> slashed zero
> empty character symbol, a reasonable proposal could be
> presented, backed
> by one or more linguistic organizations.
>
> At least slashed zero might be made available as a variant of
> the round
> empty set symbol through a variation selector ... if it is
> *asked for*.
The empty set symbol and slashed zero remain unrelated.
> But that is for those who use such notation regularly to decide.
>
> But I doubt you will find any linguist who would consider the Norwegian
> capital slashed O as anything other than a kludge replacement for
> either the standard round empty set symbol or the slashed zero symbol.
Again, (sorry for the repeat, but it seems necessary), the empty set symbol
as the capital o with stroke are historically closely related, and either can
be used to denote the empty set (and so can {}). The empty set symbol
is unrelated to a slashed zero (and both are unrelated to the diameter sign).
I have yet to see anyone quote a linguistic texts that *explicitly* says that
they use the empty set symbol for this "empty" linguistic entity.
> Jim Allan
/kent k
This archive was generated by hypermail 2.1.5 : Thu May 29 2003 - 09:01:04 EDT