Re: use of apostrophes

Date: Tue Jun 04 2002 - 15:52:57 EDT

On 06/04/2002 02:39:15 PM "James Kass" wrote:

>How did the practice of using the apostrophe to represent the glottal stop
>originate? Was it because the writing system originally called for the
>U+02C0 and the apostrophe was visually similar and readily accessible?

The glottal stop has been written in phonetic transcriptions using the
shape shown at U+0294. Pullum and Ladusaw speculate that the origin of that
shape may have been motivated by the Greek smooth breathing mark. That
seems plausible to me. That could, in turn, account for the use of the
shapes reflected by U+02BE and U+02BF for transcribing Arabic hamza and

I'm guessing that the step from the 0294 and 02BE shapes to the 02BC shape
was a combination of (i) similarity in shape, and (ii) convenient access
(the quotation mark / apostrophe being commonly available). When linguists
started working with typewriters many many decades ago, the straight
apostrophe would have been a convenient and obvious choice for representing
glottal stops, and that in turn would have led to the use of ASCII 0x27.
There is one well known exception from the typewriter era, though: it
became common practice among linguists specialising in Mayan languages to
use the digit 7 -- again, a combination of similarity in shape with 0294
and convenience.

>Maybe the answer should be a glyph
>variant. Continue to recommend U+02BC, but provide fonts with
>easily distinguishable glyphs for the characters in question.
>Keyboards could be labelled accordingly.

I'm uncomfortable with this suggestion: if people see a glyph that looks
like (e.g.) 02C0 and then look for it in the charts, they'll conclude that
02C0 was used and start using it. That won't be helpful if the first set of
data was using U+02BC.

The language(s) in question are undergoing some orthography reform as it
is, so they're open to a change in shape. The only requirement put to me
was that they needed one shape per function / character representation /
keystroke sequence. So, any of U+02BE, U+02C0, U+02C8 or perhaps something
else should all work.

- Peter

Peter Constable

Non-Roman Script Initiative, SIL International
7500 W. Camp Wisdom Rd., Dallas, TX 75236, USA
Tel: +1 972 708 7485
E-mail: <>

This archive was generated by hypermail 2.1.2 : Tue Jun 04 2002 - 14:26:02 EDT