Zompist Bboard Again

I see sasasha’s blog has some more calligraphy… from two years ago, apparently, but somehow I never noticed it in all that time. It looks very nice!

Thank you! I’m really glad you resurrected this thread. The topic is often on my mind. Many a time I have sat down and thought, right, now is the time to make all those tables you want to make of all the ligatures you think might be used in Verdurian script. Then I started learning Georgian, which has many letter-form similarities with Verdurian. I realised that while Georgian calligraphy does involve cursive writing, in general nobody joins the letters up. We do know that in Verdurian, they do ‒ but it’s interesting to note that cursive writing is not a universal tendency.

The historical perspective is also paramount: trends will fluctuate massively over the centuries.

Still, I want to get there one day.

I know not everyone is on the Patreon, but there is quite a bit of interesting new typography stuff up there which I am sure you would enjoy, bradrn.

sasasha wrote: ↑Sun Sep 20, 2026 6:44 pm Then I started learning Georgian

Ooh yes, Georgian! My recent conscript has also been heavily influenced by Nuskhuri in particular.

(Speaking of which, I recently did some more designs for this particular script, but never got around to posting them anywhere. I should do that.)

which has many letter-form similarities with Verdurian.

Does it? I honestly never noticed.

I realised that while Georgian calligraphy does involve cursive writing, in general nobody joins the letters up.

Maybe not ‘in general’, but it’s certainly not hard to find cases where they do. There are many examples here. (Particularly good ones: pages 9, 12, 30.)

it’s interesting to note that cursive writing is not a universal tendency.

My own hypothesis is that it’s strongly related to the writing instrument: narrow nibs and brushes have less friction so they encourage cursive styles (as in modern Latin, mediaeval Greek, the Mkhedruli linked above and the Sinic scripts), while wide nibs have much more friction thus discouraging cursive (as in blackletter Latin, Greek uncials, Georgian Asomtavruli and Nuskhuri, and certain styles of Tibetan). But of course this is just a theory.

(Probably the best counterexamples to the idea are the various cursive Middle Eastern scripts. I don’t really have any explanation for those beyond vague gesturing at ‘areal effects’.)

I know not everyone is on the Patreon, but there is quite a bit of interesting new typography stuff up there which I am sure you would enjoy, bradrn.

I’m sure I would! But alas, I’m not on the Patreon. (For that matter, I haven’t even bought the new Historical Atlas…)

bradrn wrote: ↑Sun Sep 20, 2026 7:53 pm

sasasha wrote: ↑Sun Sep 20, 2026 6:44 pm I know not everyone is on the Patreon, but there is quite a bit of interesting new typography stuff up there which I am sure you would enjoy, bradrn.

I’m sure I would! But alas, I’m not on the Patreon. (For that matter, I haven’t even bought the new Historical Atlas…)

Just as a teaser...

I've created several of these fonts already. Though I don't plan to make them all...

zompist wrote: ↑Mon Sep 21, 2026 1:33 am Just as a teaser...

Very nice! Reminds me of those charts you see of Brahmic scripts.

(I presume some of them are more decorative than practical though — Cont Arin, Čusî, Umbra, Liynnor, etc.)

(drools) Now there's a program just dying to be written.

alice wrote: ↑Mon Sep 21, 2026 2:17 pm (drools) Now there's a program just dying to be written.

?

"Hey, program, here's a letter¹ from a conscript I'm trying to make as good as zompist's. How many interesting random variations and developments of it can you show me so I can get my creative juices flowing?"

¹(aargh, brain fart) What is the correct name for the individual elements which make up a script? Token? Sigil? Grapheme? Graphene?

alice wrote: ↑Mon Sep 21, 2026 2:34 pm ¹(aargh, brain fart) What is the correct name for the individual elements which make up a script? Token? Sigil? Grapheme? Graphene?

I'd say "glyph" or "grapheme". Both could be used for the letters of an alphabet.

To me "grapheme" should be the smallest meaningful element of a script. Thus "é" has two graphemes, <e> and <´>. 如 has two graphemes, 口 and 女. é and 如 are both glyphs.

(Can you still argue about particular glyphs or entire scripts? Of course, this is lingustics!)

I've often wondered if there's anything approaching a theory of glyphs or graphemes; for example, do they fill out an area of conceptual space, like vowels or tones?

alice wrote: ↑Tue Sep 22, 2026 3:05 pm I've often wondered if there's anything approaching a theory of glyphs or graphemes; for example, do they fill out an area of conceptual space, like vowels or tones?

I'd say: only conscripts (which includes Han'kul).

Ideally glyphs in a script should maximally contrast, so none are very similar to each other, and also aren't mirror images. They also shouldn't be hard to draw— glyphs in two or more pieces are just asking for trouble as people will connect them. (See: caoshu.)

But glyphs arise from a miscellaneous set of drawings; they change shape or orientation; they are mangled by graphic rearrangements (e.g. the 90° rotation in cuneiform) or by changes in medium; new letters are needed; sound changes spoil the structure; people insist on adding embellishments; people insist on simplifying to the point of confusion. So any structure gets lost.

bradrn wrote: ↑Sun Sep 20, 2026 7:53 pm

sasasha wrote: ↑Sun Sep 20, 2026 6:44 pm Then I started learning Georgian

Ooh yes, Georgian! My recent conscript has also been heavily influenced by Nuskhuri in particular.

(Speaking of which, I recently did some more designs for this particular script, but never got around to posting them anywhere. I should do that.)

Yes, I remember this script appearing on the board; it is beautiful. I’d like to see any developments.

which has many letter-form similarities with Verdurian.

Does it? I honestly never noticed.

Well, ო <o> is essentially Verdurian <š>, and რ <r> is rather like Verd. <ž>. The closest correspondence is ს <s>, which is almost exactly like Verdurian <s>. I also think comparisons can be drawn between პ, ჰ, კ, ც (<p’ h k’ ts>) and Verd. <a>.

Another, vaguer, comparison: both scripts have curves at the beginnings or ends of letters that don't easily lend themselves to joining, because they curve in on themselves away from the start point of a new letter. In Georgian (mkhedruli), ი ო ღ ლ ჩ პ ჰ at least fall into this category (and many more by a slightly wider definition); in Verdurian, at least š ž č d.

Latin script has a lot of minims and more joinable letters in general (in such a way that nuskhuri more closely resembles it, IMO). Though it does have <b p s>.

What’s more interesting, then, is the differences in joining strategies that exist between mkhedruli and Latin. You could join <ი ო> etc. on in a fairly similar way that we join <b p>, but to my understanding this isn't often done in modern handwriting. If they do join they leave off ending the letter before they turn for the inward curve and use a short diagonal line right and down. And letters with left-curving descenders in Georgian like <ქ ე ყ უ კ ვ>, when joined, do not use loops but rather snake right and up in a 180° (then usually right and down in another to start the next letter).

These kinds of observations have led to a lot of head-scratching on my part, dreaming up what kinds of culturally entrenched joining strategies might develop for Verdurian. It’s an area that zomp got right early on, with the handwriting samples we have, since the letters display rather unexpected joining strategies (see <b> in the Lord of the News example, off the top of my head). I am sure that a schema for Verdurian handwriting would be as much about what is quintessentially Verdurian in the strategies to join letters as what is quintessentially Verdurian about the letters themselves.

And this basically holds for any conscript, so it’s useful to consider.

I realised that while Georgian calligraphy does involve cursive writing, in general nobody joins the letters up.

Maybe not ‘in general’, but it’s certainly not hard to find cases where they do. There are many examples here. (Particularly good ones: pages 9, 12, 30.)

This is a great document, I’ve been looking for something like this for a while. Thanks!

I should have specified modern handwriting, really. My understanding is that these kinds of cursive hands basically belong to the past and modern Georgians might struggle to read them. I might be wrong about this, but this is based on trawling the internet for Georgian handwriting samples and finding so, so many examples of modern hands with no observable joins whatsoever; plus anecdotal evidence like reddit comments saying they don’t teach cursive in schools and no one really bothers with it anymore.

There certainly are ornamental/calligraphic Georgian hands that do use more ligatures and multi-letter joins, but they appear to be relegated to the status of curios (and of course historical artefacts) these days.

it’s interesting to note that cursive writing is not a universal tendency.

My own hypothesis is that it’s strongly related to the writing instrument: narrow nibs and brushes have less friction so they encourage cursive styles (as in modern Latin, mediaeval Greek, the Mkhedruli linked above and the Sinic scripts), while wide nibs have much more friction thus discouraging cursive (as in blackletter Latin, Greek uncials, Georgian Asomtavruli and Nuskhuri, and certain styles of Tibetan). But of course this is just a theory.

(Probably the best counterexamples to the idea are the various cursive Middle Eastern scripts. I don’t really have any explanation for those beyond vague gesturing at ‘areal effects’.)

Hmm. I’m not sure I agree, except in broad stokes (ha, ha). Yes, the type and width of the writing implement greatly influence a variety of features within a script. But you provide your own counterexamples. Also, writing with wide nibs seems to actually lead to joins in my own experiments: the high width variance with stroke direction makes it rather natural to flick to the next letter’s start point on the narrow angle of the nib. I find that a very narrow nib can frequently split or warp on an upstroke, too. And finally, from your example on p.30 to now the nibs available to Georgians have, if anything, got narrower and more precise, but their handwriting has got less, rather than more cursive.

sasasha wrote: ↑Tue Sep 22, 2026 5:44 pm Another, vaguer, comparison: both scripts have curves at the beginnings or ends of letters that don't easily lend themselves to joining, because they curve in on themselves away from the start point of a new letter. In Georgian (mkhedruli), ი ო ღ ლ ჩ პ ჰ at least fall into this category (and many more by a slightly wider definition); in Verdurian, at least š ž č d.

Very insightful point!

And letters with left-curving descenders in Georgian like <ქ ე ყ უ კ ვ>, when joined, do not use loops but rather snake right and up in a 180° (then usually right and down in another to start the next letter).

There’s some loops on page 30 of that PDF… though then again that sample also seems borderline illegible, so who knows.

These kinds of observations have led to a lot of head-scratching on my part, dreaming up what kinds of culturally entrenched joining strategies might develop for Verdurian. It’s an area that zomp got right early on, with the handwriting samples we have, since the letters display rather unexpected joining strategies (see <b> in the Lord of the News example, off the top of my head). I am sure that a schema for Verdurian handwriting would be as much about what is quintessentially Verdurian in the strategies to join letters as what is quintessentially Verdurian about the letters themselves.

I agree with all of this, though I never actually put it into words. The letter which particularly struck me in zompist’s samples was final ⟨-n⟩ joined from the bottom rather than the top. In this light I think it’s fair to say that a lot of my early experimentation was around trying to find interesting joining strategies.

I should have specified modern handwriting, really. My understanding is that these kinds of cursive hands basically belong to the past and modern Georgians might struggle to read them. I might be wrong about this, but this is based on trawling the internet for Georgian handwriting samples and finding so, so many examples of modern hands with no observable joins whatsoever; plus anecdotal evidence like reddit comments saying they don’t teach cursive in schools and no one really bothers with it anymore.

For that matter, you could say the same about English. I suspect the influence of computers. (It’s rather a pity in my opinion.)

But yes, I agree that cursive seems much less pervasive in Georgian than in Latin, Cyrillic or Arabic. Which, given that those three scripts have surrounded Georgia for much of its history, is a rather interesting observation.

Also, writing with wide nibs seems to actually lead to joins in my own experiments: the high width variance with stroke direction makes it rather natural to flick to the next letter’s start point on the narrow angle of the nib.

It’s interesting that was your experience. To me that only happens with very specific styles which encourage it, most notably Latin italic script.

Relatedly: it’s interesting to note that blackletter scripts got more broken up and less connected as they evolved. As far as I can tell, the cursive Kurrent script evolved only after the transition to a narrow pen.

I find that a very narrow nib can frequently split or warp on an upstroke, too.

I believe this is why narrow nibs only became really widespread with improvements in writing technology, namely the transition to quills and then metal nibs. Even then, with flexible nibs, upstrokes must be made with minimal pressure (whence that distinctive ‘copperplate’ look). As far as I’m aware, it’s only with modern pencils, ballpoint pens and round-nibbed fountain pens that all strokes become equally easy to write.

(How the mediaeval Greeks and Georgians made their thin nibs is something I’ve been curious about for a while. However they did it, the style seems not to have become widespread beyond those two scripts and Estrangelo Syriac.)

zompist wrote: ↑Tue Sep 22, 2026 3:38 pm

alice wrote: ↑Tue Sep 22, 2026 3:05 pm I've often wondered if there's anything approaching a theory of glyphs or graphemes; for example, do they fill out an area of conceptual space, like vowels or tones?

I'd say: only conscripts (which includes Han'kul).

Ideally glyphs in a script should maximally contrast, so none are very similar to each other, and also aren't mirror images.

I've got the impression that basically all natscripts have a certain stylistic consistency, though. I've never seen a script (unless you count Asian texts incorporating Western letters) where, let's say, some of the glyphs looked like Latin letters, some of them looked like Arabic letters, some looked liked cuneiform, some looked like Chinese glyphs, and some looked like Mayan glyphs.

Raphael wrote: ↑Wed Sep 23, 2026 1:37 am

zompist wrote: ↑Tue Sep 22, 2026 3:38 pm

alice wrote: ↑Tue Sep 22, 2026 3:05 pm I've often wondered if there's anything approaching a theory of glyphs or graphemes; for example, do they fill out an area of conceptual space, like vowels or tones?

I'd say: only conscripts (which includes Han'kul).

Ideally glyphs in a script should maximally contrast, so none are very similar to each other, and also aren't mirror images.

I've got the impression that basically all natscripts have a certain stylistic consistency, though. I've never seen a script (unless you count Asian texts incorporating Western letters) where, let's say, some of the glyphs looked like Latin letters, some of them looked like Arabic letters, some looked liked cuneiform, some looked like Chinese glyphs, and some looked like Mayan glyphs.

I’d say there are definitely typological consistencies with scripts (and hands).

In other related thoughts, aren’t many other natscripts essentially conscripts? Georgian, Armenian, and Cherokee (and I think more) purport to have named inventors.

Raphael wrote: ↑Wed Sep 23, 2026 1:37 am

zompist wrote: ↑Tue Sep 22, 2026 3:38 pm Ideally glyphs in a script should maximally contrast, so none are very similar to each other, and also aren't mirror images.

I've got the impression that basically all natscripts have a certain stylistic consistency, though. I've never seen a script (unless you count Asian texts incorporating Western letters) where, let's say, some of the glyphs looked like Latin letters, some of them looked like Arabic letters, some looked liked cuneiform, some looked like Chinese glyphs, and some looked like Mayan glyphs.

Mostly true, but think about why. Such a script would require changing every few seconds from a round quill, to a wide nib, to a stylus, to a brush, to a chisel. Scripts are designed for one tool only, and even move away from inconvenient variation in its use. E.g. over the centuries the angles used in cuneiform narrowed to a small angle, so you didn't have to rotate the stylus much.

Even so, I'd say there are exceptions. E.g. look at some of the examples of hanja mixed with han'kul. Japanese is written with three visually distinct writing systems.

For that matter, modern Latin alphabets are examples too! The Carolingian script might well have looked jarring when it was new, mixing ancient forms of the letters (the "capital letters") with the simplified forms that had developed since ("lower case").

zompist wrote: ↑Wed Sep 23, 2026 3:29 am Even so, I'd say there are exceptions. E.g. look at some of the examples of hanja mixed with han'kul.

I’m going to say that that image is a case of incompetent font matching. Note that the smaller hangul blend in nicely with the hanja, while the bigger hangul don’t — it’s because the hangul font has much thicker strokes than the hanja font.

Though, oddly enough, this doesn’t invalidate it entirely, because the mismatch goes back to the very beginning (where if anything it was more jarring): https://commons.wikimedia.org/wiki/File ... _pages.jpg

For that matter, modern Latin alphabets are examples too! The Carolingian script might well have looked jarring when it was new, mixing ancient forms of the letters (the "capital letters") with the simplified forms that had developed since ("lower case").

Wasn’t the Carolignian script unicase? Roman script was used only for drop capitals, which were rubricated or illuminated anyway — the whole point was for them to stick out.

To my understanding, script mixing in the modern sense took place only with the humanist scripts. And, interestingly, the humanists didn’t even realise the cases weren’t contemporaneous — they took the Carolignian miniscules to be of Roman age. That said, they still thought of the capitals as ‘sticking out’: e.g. https://commons.wikimedia.org/wiki/File ... voglio.jpg.

bradrn wrote: ↑Wed Sep 23, 2026 6:17 am Wasn’t the Carolignian script unicase? Roman script was used only for drop capitals, which were rubricated or illuminated anyway — the whole point was for them to stick out.

Not so far as I can see. You can see examples here. Majuscules are not only used for drop capitals, but for initial letters, as well as for entire initial lines. And even a few individual words are capitalized, e.g. Gloria and Maria in the gospel text pictured on that page.

zompist wrote: ↑Wed Sep 23, 2026 3:09 pm

bradrn wrote: ↑Wed Sep 23, 2026 6:17 am Wasn’t the Carolignian script unicase? Roman script was used only for drop capitals, which were rubricated or illuminated anyway — the whole point was for them to stick out.

Not so far as I can see. You can see examples here. Majuscules are not only used for drop capitals, but for initial letters, as well as for entire initial lines. And even a few individual words are capitalized, e.g. Gloria and Maria in the gospel text pictured on that page.

Interesting, thanks!