Re: Emoji: emoticons vs. literacy

From: Asmus Freytag (
Date: Thu Dec 25 2008 - 17:44:50 CST

On 12/25/2008 2:26 PM, John Hudson wrote:
> Asmus Freytag wrote:
>> Punctuation marks are not the only devices beyond letters that people
>> feel are necessary to employ as distinction between otherwise
>> identical statements. Publishers, editors and authors of books in
>> English, but not necessarily all other European languages,
>> occasionally seem to find it necessary to employ italics to
>> disambiguate between two possible interpretations of a given sequence
>> of words.
> Indeed. But that is typography -- a 'higher level protocol' -- not
> character encoding. The question is not whether emoticons are used or
> may be useful, but whether the kind of information that they seek to
> convey belongs in plain text. I think it is consistent with type size
> selection, font style selection, and other articulatory aspects of
> display above the plain text level.
All of what you describe here as stylistic features has in common that
it affects a span of text. Emoticons, like punctuation, on the other
hand are discrete elements. That's an important distinction when it
comes to deciding the question of how to handle something in the overall
architecture. Except in the rarest of circumstances, something that
affects a span of text does not belong on the character encoding level.
And even when the exceptions are well-motivated, or, in fact, unhappily
prove inescapable, they do break the model at great cost to implementers.

On the contrary, discrete elements, don't break the model, even if
there's a suspicion that they are less "textlike" than other text
elements, and even when they are subject to different typographical
rules than letters or punctuation. When discrete graphical elements come
in (more or less) conventional sets, and have found a (more or less)
conventional use, it may not only be not harmful, but perhaps even
beneficial to treat them like other encoded text elements.

In any case, I don't consider the claim of a special nature of these
emoticons as either compelling or proven in any sense. Quite the
opposite, in fact. I rather view the existence of common plain-text
fallbacks as proof that they are integral part of plain text.

Does that mean I see no problems at all? Not at all. While there is
clearly a conventional sub-set of these things, the very open-ended
nature of their usage needs to be addressed. That's actually true for
all symbols that are not part of a well-governed notational convention
like math or music. I would nevertheless support the encoding of the
conventional sub-set(s) of emoticons, general symbols, etc. while
accepting the need to draw a line. The other problem, not shared with
general symbols, is that of rendering in color, or with effects. For the
core set of emoticons, those aspects thankfully aren't needed, and so
wouldn't need to be considered in encoding.

The third issue, unique to emoticons, is the existence of more or less
canonical ASCII fallbacks. Any argument along the line that *that*
should be the sole representation in plain text needs careful thought.
The advantage of the fallbacks is that they are open-ended, but the
disadvantage is that they constitute their own, non-documented,
mini-markup. One of the advantages of Unicode-based protocols (e.g.
MathML over TeX) was supposed to be the elimination of the need to use
markup for plain text-elements, reserving it for stylistic and other
information around and about the text.


This archive was generated by hypermail 2.1.5 : Fri Jan 02 2009 - 15:33:07 CST