Re: What makes a good script? (was: Re: What is this list...)

Doug_Ewell@... wrote:

>
> Peter T. Daniels <grammatim@...> wrote:
>
> > ... However, your characters are far, far too similar in appearance
> > to each other to constitute a practical script. (That's the other
> > main objection to the winning entry for Shaw Alphabet, too.)
>
> Oddly enough, I had just been wondering about the uniformity of my
> letters and the effect on readability right before I received Peter's
> message.
>
> I remember reading somewhere about studies that showed that Latin-
> script text written in all lowercase is easier to read than all
> uppercase, partly due to the "more distinct letterforms" of lowercase.
> I wonder if that's why I find lowercase Cyrillic text so tiresome to
> read -- most letters are simply small versions of the uppercase
> letters, unlike Latin and Greek.

More than that, even, there's the paucity of descenders and ascenders,
and the monotonous pair of serifs above and below like trolley tracks.
There was lots of experimentation with Cyrillic typography in the early
days of the Revolution, and probably with calligraphy as well, but we
all know what happened to experimental arts there soon after.

> As a non-expert, it hadn't occurred to me that there was such a thing
> as a level of uniformity so excessive that it would render a script
> "impractical" or unsuccessful.
>
> This leads to the bigger question (and if this isn't on topic for
> Qalam, nothing is):
>
> What features, traits, or attributes make a "good" script?
>
> Peter states that too much uniformity among letterforms (in Ogham,
> Shavian, or my Ewellic) can make a script excessively difficult to
> read. It seems to me that several living scripts, especially Thaana
> (Maldivian) and Ethiopic but perhaps also Thai and Myanmar, have this
> problem to a certain extent as well.

I don't think there can be much problem with Ethiopic, since it's barely
changed in 1500 years; I don't know about Burmese, but in Thai, lots of
the letters are virtually unused, or occur in only a handful of words.
(There's been _lots_ of imaginative Thai typography.)

> Is there such a thing as *too little* uniformity among letterforms?
> Do any existing scripts exhibit this problem?

Funny you should ask, that's what my paper in the 1996 Hofstra
conference on Islamic calligraphy is about (it'll be published any year
now, but they have apparently lost my illustrations). Proto-Sinaitic (if
we want to include it in the sequence; cf. Brian Colless, these days)
and earliest Phoenician are pretty irregular, but they get neatened up
because scribes have esthetic impulses -- see how beautifully rhythmic
Late Punic inscriptions are.

> What about number of characters? Logographic and hieroglyphic scripts
> require the reader to learn hundreds or thousands of characters,
> compared with the relative few in alphabets, abjads, and syllabaries.
> It would certainly seem easier to learn 25 to 40 symbols than ten
> thousand. But too few symbols may mean that more of them are required
> to represent a given language adequately, that supplementary
> diacritical marks must be introduced, or that phonemic spelling
> principles are abandoned (as in English), all of which impair reading.
> Also, many alphabets have adopted separate uppercase and lowercase
> forms, which doubles the number of distinct letterforms but apparently
> improves readability.

English has not abandoned "phonemic spelling principles," but it has
grown in a way that has made them quite complex (and I feel -- but
haven't yet demonstrated -- that the amount of "logography" in English
is about on a par with that in Chinese or Japanese: that is, you need to
memorize stuff like bomb/comb/tomb and the dozen-odd <ough> words, and
all that sort of irregularity will add up to a few thousand forms at
most, and of course the irregularity is concentrated in the commonest
and shortest words). For the regularities, see Carney, *Survey of
English Spelling* (Routledge, 1994).

> What about directionality? Most scripts today are LTR. Is there
> anything inherently more difficult about reading RTL? What about top-
> to-bottom (or Ogham bottom-to-top)? Some early scripts were
> boustrophedon (alternating LTR and RTL). This seems easier to write
> but harder to read; is that true?

Some South Arabian inscriptions are written boustrophedon on walls
hundreds of feet long -- definitely easier for the reader, too. Derrick
de Kerkhove made some suggestions about brain lateralization and script
direction that seem pretty implausible and don't seem to have been
followed up on.

> What other issues are there? This is clearly a subject on which the
> experts (particularly Daniels and Bright, but others as well) can
> share some of their expertise with those of us who are interested and
> curious, as our presence on this list should demonstrate.

--
Peter T. Daniels grammatim@...