Nicholas Bodley wrote:

>On Sun, 09 Oct 2005 21:26:47 -0400, Don Osborn <dzo@...> wrote:
>
>
>
>> but the resulting similarity of everything meant you
>>had to pay very rapt attention to each form in order to read text in
>>it (or so it seemed to me; this was not like Chinese characters which
>>have many unique forms and resulting combinations, but rather used
>>positioning and number of dots and straight lines to make the
>>distinctions, as I recall).
>>
>>

That was what I thought when I followed the link earlier this evening to
that "Norquay Script." It's mostly Tengwar. Now Tengwar is a very
pretty alphabet, and an intelligently-designed one... but as an actual
usable writing system it's pretty poor. Half the letters look like...
well, the other half. On the one hand, you can learn some simple logic
and work out the sounds, on the other hand, who has time to apply logic
when reading at speed? It lacks practicality.

>Within reason, the more easily-distinguished the glyphs are from one
>another, I should think the better the readability, *but* at the same
>time, a neatly-bordered and "smoothly-flowing" run of text helps (I think
>) readability. (I'm probably mixing up readability and learnability to
>some degree.) Surely, Myanmar/Burmese has some very-prominent
>interruptions in its general borders! For some time, it has seemed to me
>that CJK glyphs are created to minimize similarity of appearance; there
>seem to be relatively few that look almost the same, but I really don't
>know.
>
>

Not sure what you mean by "smoothly flowing." Connected letters
probably are not a help to readability. Letters that vary wildly in
features and shape and size might be easy to distinguish, but they would
also be hard to read in any quantity, as the page would be a riot of
shapes. You need to have some consistency and quietness as well. You
can see some of this in the development of square Hebrew script.
Looking at some early inscriptions, even though the underlying shapes of
the letters is more or less the same as it is today, it looks very crude
and wandering: some letters just way out of size, etc. A lot of what
went on in developing the script was regularizing the size and features
of the letters. That's a whole area of study too: how *features* of
letters fell together, so *parts* of letters look like each other. It's
handy for type-design, too: get an i that looks good and you already
have most of the l done, as well as significant pieces of b, d, u, k, p,
and a bunch of others (in most styles). You see this in Hebrew, as the
letters shin, `ayin, tet, nun, zayin, gimel, tsadi formed a
common-feature group in quite a few periods of development (see
Yardeni's Book of Hebrew Script) and still today in Torah scrolls those
seven letters (and final forms)--and no others--are decorated with
little triple crowns on their heads.

>Otoh, much of Thai and Armenian, neither which I read at all, but can
>identify, seems to consist largely of paired vertical strokes. To my
>ignorant eyes, the letters look similar enough (I think) that they are
>somewhat hard to distinguish. I'm reminded of a small illustration of part
>of a Latin ms. which had almost mechanically-perfect, evenly-spaced
>vertical strokes in some words, such as "minimum"; it was shown to explain
>why dots on minuscule i's were needed.
>
>

Ref. B.L. Ullman, "Ancient Writing and its Influence":

mim numinum niuim minimi munium nimium uini muniminum imminui uiui
minimum uolunt

"The very short mimes of the gods of snow do not at all wish that during
their lifetime the very great burden of (distributing) the wine of the
walls to be lightened" That's Ullman's translation; take it up with him
if you don't like his Latin. He does mention in brackets just after
"burden" that "[munium is neuter singular]" if that helps.

Set that in a nice blackletter (as did Ullman) and good luck.

>As well, the size of distinguishing details relative to that of the whole
>glyph seems to matter; details seem rather small in Thai, for instance.I'm
>reminded of some Fraktur faces in which the f and the k seem to be made
>about as similar as one can make them, almost as if it were a matter of
>pride. One comment was that in a certain face and size, the two letters
>differed by only 40 millionths of a square inch!
>
>

In which case why bother? Especially if we shift to consider
handwriting, it's perfectly possible to have a script (or more likely, a
style) in which some letters simply *aren't* distinguishable. There
were periods in Hebrew where some of the similar letters were even more
difficult to distinguish. I myself discovered, poring over a
handwritten work in Samaritan Hebrew (which is vastly different from
square Hebrew) that in the particular handwriting and style I was
reading, there simply was no reliable way to distinguish Het from
Tsadi. Most of the time you can get away with this, and the script can
even continue on without conflating the letters permanently. Such
things started to happen in Arabic, so they added the dots, I think.

>Some time back, it occurred to me that some elaborate glyphs, especially
>South and Southeast Asian, took some time to write! Those people were not
>in a hurry, it seems to me. (The opposite seems to be modern cursive
>Hebrew; it looks really "fast and simple.")
>
>

Nothing beats Arabic for looking "fast." Modern Hebrew cursive has the
"advantage" (by the logic given above) of not being a connected script,
so the letters are still recognizably distinct even to beginners--though
that probably starts being a little less true in really fast writing by
fluent natives. And in the past it was even less so; Yardeni brings
several examples of quite cursive Hebrew styles (most of which I can't
read at all).

Sorry. Guess I'm feeling talkative today.

~mark