From: Asmus Freytag (asmusf@ix.netcom.com)
Date: Tue Nov 06 2007 - 00:22:43 CST
On 11/5/2007 6:53 PM, John Hudson wrote:
> Michael Everson wrote:
>
>> Actually those rules work really well and reliably in Fraktur.
>
> But ligatures in fraktur writing perform a different function from
> ligatures in the Italian humanist bookhand that is the basis of
> roman/antiqua type. The forms of ligatures in fraktur are also
> different -- unsurprisingly, since differently created strokes can be
> expected to interact and connect in different ways -- and notably the
> form of the fraktur f does not have the same collision potential as
> the roman f.
>
> This is all fine, and I don't see any problem at all with different
> forms of scripts having different rules for ligation. Indeed, that
> makes eminent sense to me because it recognises that ligation is an
> aspect of writing style *not of grammar*. What seems silly to me is
> devising rules for ligation based on a particular style of writing and
> then insisting that those rules also apply for all other styles of
> writing of a given language.
Well, that's what those "silly" Germans have done. If you are used to
having the presence of certain morpheme boundaries interact with
ligation in Fraktur (where a lot more ligation went on, anyway) then
it's quite understandable that you don't want to give up more than you
have to when you eventually make the big switch to Roman/Antiqua.
Because of the way compound words work, they do require that the reader
mentally break (most of) them apart when reading and every subtle bit of
help from typography is welcome in that task. In addition, there are
some ambiguities that can be introduced in the process, where moving the
intra-word splits changes the meaning. Again, having some help in the
typography makes text more readable.
Note that applying the aesthetic sensibilities of anglo-Saxon
typographers, who call for applying ligatures indiscriminately, does
*not* result in the most readable German text (to native readers).
There's a lesson there somewhere. Other typographical devices are
equally non-portable within the same script. Note that, again because of
the Fraktur (and typewriter) tradition, e m p h a s i s created by
increased letter-spacing is still something that's found in German
printed texts (getting less common, I think, but I keep finding
instances). To anyone trained in that convention, the not uncommon
practice of allowing very loose letter spacing in narrow columns which
can be found readily on this side of the Atlantic, creates a sort of
RANdom shOUTiNG effect.
The typographical ideals and what deviations of it are permissible, are
simply different across languages -- even if these languages ostensibly
use the same script. Sometimes, the differences are in your face (French
spacing around punctuation) sometimes they are subtle (angle of accents
French/Polish, ligation in German/Dutch etc.). But they are real and
they exist.
Calling what the other guys prefer "silly" is, well, silly.
A./
> I don't think it makes any sense to say that there is or should be a
> single set of rules for ligation in German that is independent of
> script style. This is not to say that individual languages do not give
> rise to particular orthographic phenomena that require specific rules
> *within writing styles*. The presence of triple-f sequences in some
> German compound words is an example, as is the presence of
> semantically differentiated dotted and undotted i in Turkish. The
> latter has prompted some people to formulate the rule that fi
> ligatures should not be used in Turkish, because it is common in most
> types for the top of the f to swallow the dot of the i in this
> ligature. But the rule is mis-formulated, because if the form of a
> ligature in a particular style of writing or typeface design does not
> swallow the dot then there is no reason why one should not have an fi
> ligature in Turkish. Similarly, I don't think it makes sense to insist
> on general rules for German ligation independent of writing style or
> typeface.
>
> Of course, what this all implies is that handling of rules for German
> ligation should happen at a level somewhere well above text encoding
> -- just as Turkish ligation is --. where it can take into account the
> particular embodiment of the text, i.e. a font.
>
> John Hudson
>
This archive was generated by hypermail 2.1.5 : Tue Nov 06 2007 - 01:26:10 CST