Questionnaire about the needs of Greek typography end-users

BACKGROUND:
During the last few decades Greek typography has been strongly influenced by Latin elements. Often, there is substitution of Greek letters by Latin ones. A few designers try to preserve characteristic Greek typographic elements into their work, but overall Greek typography appears to trend towards Latin assimilation.

AIMS:
a) To define what constitutes Greek characteristics in a font,
b) To delineate the needs and problems faced by contemporary Greek typography end-users (graphic designers, book publishers of Greek texts, etc.)

EXPLORATORY QUESTIONS:
a) How do you define or how do you approach the issue of a Greek characteristic in a font?
b) What questions would you have for end-users of your fonts?
c) Why do you design Greek fonts?

P.S. The results of this study, including all information from font-designers and end-users, will be posted in this blog in the end of June.

quadibloc's picture

Well, here is an example from astronomy.

Does the Earth turn on its axis once in 24 hours, while the Moon does not rotate at all?

That is the obvious appearance of things.

But if, in the case of the Earth, you try to explain the Equation of Time, or, in the case of the Moon, you try to explain libration in longitude, you encounter the hidden appearance.

The Earth rotates, with respect to the fixed stars, with a period of 23 hours, 56 minutes, and 4 seconds. Its 365 day journey around the Sun makes the day longer - 24 hours - on average; but since the Earth's orbit is elliptical, and also since the ecliptic and the equator do not coincide, the time of apparent noon as registered on a sundial will move up to 15 minutes back and forth from clock time.

And the Moon's orbit is also elliptical, leading to libration, from combining an orbit and a rotation which both have a period of 24 1/3 days.

Jongseong's picture

There is so much to appreciate and to comment on in this thread, but I'll limit myself to one tiny point.

For instance, η (eta) is not necessary to have a descending part, it may just stand above the baseline (see Centro Serif Regular compared to Centro Serif Italic in my previous post). This is quite acceptable. Greeks write η both with or without descender. You may justify a similar approach with some other letters as well.

In general, I'm of the opinion that up to a point extenders help in reading by aiding in differentiating those relatively fewer letters that have them. A weakness of conventional Cyrillic lowercase in my mind is the paucity of extenders. I'll never get used to the Bulgarian-style yu (ю) with an ascender (are you kidding me?) but I wouldn't mind it if people started actually writing yo (ё) instead of replacing it all the time with ye (е) in Russian, or even if the more prominent old-timey tails of tse (ц) and shcha (щ) came back into fashion.

So I was surprised to see the comment that fewer descenders make reading easier. This got me thinking about the exact role of extenders in helping readability. They of course break up the monotony of letters confined to the x-height; imagine if everything looked like "as no one concerns us even more". But extenders don't have to be too frequent to achieve this role, and past a certain frequency it's likely to become too much.

My subjective opinion is that Greek in its current monotonic orthography has somewhere near the right balance of things happening outside the x-height. I don't think Greek has too many descenders, since many of the letters with descenders like ψ, ξ, ζ, and φ don't occur too frequently.

But here's the thing. I remember posting something about this many years ago, but I'm perfectly happy with η (eta) without a descender. It isn't just about being exposed to so many examples of descender-less η in handwriting and in print, or even that the older handwritten and printed forms of descending η tend to have much shorter descenders than are usual today.

If all the other vowels (α, ε, ι, ο, υ, and ω) are extender-less, it appeals to my symmetry-seeking nature to have a descender-less η as well. Perhaps it's my grounding in the Korean alphabet where all vowels are variations on a set of simple elements that makes me expect some sort of a unity in design that holds the vowel letters together. After all, there's something about the Roman letter y used as a vowel that is not entirely satisfactory for me—it doesn't look enough like a vowel letter due to its descender. I wouldn't be surprised if I'm the only one who feels this way, though.

I do decidedly prefer descending χ, by the way. Confined to x-height, it would be distractingly close to κ.

ilyaz's picture

> the words of Heraclitus are not scientific, they are philosophical.

Frankly speaking, the only scientists I personally know who actively use Greek letters in a continuous flow (for scientific purposes, not when they write letters to their moms) are philosophers. Tell this to them! ;-)

hrant's picture

Brian, you've brought up some interesting things. However to me your stated "symmetry-seeking" (which I actually think we all exhibit) is confined to display typography; when it comes to immersive reading I don't see any positive role for such formalisms. Now, if only the vowels in Greek were confined to the "x-height" I would say that sure, the η (eta) should follow that rule, because there the confinement would be conveying usable information*. But I love the "y"! It would be so nice if Latin used its rich descenders more frequently; the ascenders are more frequent but overall way too plain (only the "f" is pulling its weight).

* Speaking of Hangul, one amazing thing about it is that its four levels of syllable "visual density" are strongly associated with a syllable's grammatical function.

My subjective opinion is that Greek in its current monotonic orthography has somewhere near the right balance of things happening outside the x-height.

Really? But look how underused the ascender space is.

--

What people are "used to" is only part of the puzzle of readability. Saying that an η (eta) without a descender is just as good for reading as one with a descender* ignores -what I believe to be- a basic premise of reading: that deviation from a rectangle (again consider the low readability of all-caps) conveys information. Saying that Latin is easier to read than Greek because it has fewer descenders is exactly backwards. And trying to make Greek approach Latin by -for example- neutering its descenders is Latinization. Colonialist? Sadly cultural jingoism is alive and well, and sugar-coating our terminology makes things worse, not better.

* In fact -as John once made me realize- the descender-less η is a marked flaw in one of my most favorite Greek fonts, Apollonia.

hhp

ilyaz's picture

> Saying that Latin is easier to read than Greek because it has fewer descenders is exactly backwards.

There is/was a quote I’m trying to locate (Eco? Pinker?) for many years: the premise was that even if one replaces each letter in an English sentence by its bounding rectangle, the resulting image is still kinda readable (as in ▯◽▯ ▯◽◽ ◽◽◽◽ ▯▯◽◽ — only with proper distinction of ascender /descender parts). The theory is/was that the eye when it follows the sentence can recognize such “coarse” shape quite early (when finishing the disambiguation of preceding words, so the following words are still on the low-resolution part of retina), and then this preliminary info may be used when these words come “into focus”.

But then I discovered a text (blog?) of the head (?) of Microsoft type design team. The premise was that when he came to M$, the team was under the influence of the theory I hinted to above; since at the moment he knew that the theory is not very much trustworthy (in view of recent improvement of experimental data), he spent some time to collect information on experiments which more or less debunked the theory above. IIRC, the machine in the experiments would track Rapid Eye Motion, and would blur all the letter further than some distance away from the center of vision (in real time!). IIRC, the speed/reliability of reading in these experiments would not suffer even if the “distance” is 3–4 letters. (Sorry, do not have an exact reference right now.)

So the theory that “macro-shape” of words may be helpful for reading might be not that much substantiated by experiments…

Tatiana Marza's picture

Brian, welcome on board! :)

I will begin to have serious problems in writing my thesis:) I have to close it in few days and still, new, interesting material is coming up. Your theory of vowels having no extenders is an intricate one, worth thinking of :) Do you have the book of French Laurent Pflughaupt, 'Letter by letter'? Is quite a good systemized book on evolution of letters. He explains that "in the crossover to Greek [from Phoenicians, the HETA] was integrated as the HETA consonant, transcibing the [h] sound. It also became the vowel ETA, with a softer pronunciation corresponding to a [long e] in Ionion Greek. This vowel/consonant ambivalence is still evident in French nineteenth-century dictionaries..." (p.75 for ENG version).
This explains, maybe, the descender part of 'η' :)

And, BTW, in letter frequency scale, the 'η' belongs to 14th place. I am not sure if its descender has a negative impact on reading. In any case, to support the theory of descenders making reading more difficult, one should engage in a series of experiments. If any of you has the idea who would be willing to help us, I am in!
Greek population should be asked to read long texts with all ascenders and descenders included, than to switch to a font with less descenders and ascenders. And maybe several switches should occur. But on which criteria one should choose the proper font for both cases? They can not differ in style...
Besides, the fatigue factor should be included, then the emotion condition. Is a scientific project. And for sure the age factor has its significance too (the eldest people are used to descenders, like Chris' mother was used to, and younger are exposed to more Latinized fonts).

_______

Brian, why do you find current Greek orthography monotonic? Because of the abolition of polytonic system?
And when did I miss such evolution of Russian letters? ( writing yo (ё) instead of (е), and the loss of tails of (ц) and (щ)).

Yulia's picture

Hrant, I'm glad you explained theoretically why optimal “x-height” in Greek is larger than in Latin—to what I came through experiment:)
Tatiana, what indeed happens about Russian letters е and ё is that officially both letters “ё” and “е” exist, but in fact words which we pronounce with sound “yo” (for example “ёлка”, “ёжик”) often (but not always) are written with е (like “елка”, “ежик”). You can see it in books, magazines, internet and other media. Though in official documents this situation causes problems: when people with “ё” in their surname have it written with “ё” in some documents (for example passport) and with “е” in other (for example some agreements), they have to change the documents to unify the writing (for some documents may be illegal because surnames with е and ё are considered different).
And about the tails of ц and щ: these letters didn't lose their tails, but the tails are shorter today, than for example in the 18th century. Some fonts have these letters with long tails as stylistic swashed variants, but rarely for default setting (mainly those which are revivals or are influenced by some historical styles). See “Academy” typefamily for example (it is based on “Sorbonne” typeface and Russian typefaces of the mid-18th century).

Tatiana Marza's picture

Yulia,
thank you for your contribution:) But it sounds unbelievable to me that today, in Russia, the 'ё ' is not used anymore. I just opened a website of a Russian newspaper and was trying to find this letter. From time to time I am reading the russian news there. It was astonishing that I didn't realize it. I suppose because I know how to read words with letter and sound 'ё ', and I just missed its disappearance. It happened on a subconscious level. Same case with Greek, with fonts that were latinized. It seems that I had no problem till I was introduced into this chapter. But, especially this site, http://www.protagon.gr/?i=protagon.el.oikonomia&id=24021, where 'γ' is replaced by 'y', literally irritates my brain while reading. It's out of space! I find no reason that it should have an oblique descender, it could be upright.

Jongseong's picture

Another point I didn't mention is that η can carry an accent as in ή, so it is going to be using the extender spaces in some cases regardless of whether it has a descender.

Hrant, I don't think the ascender space is too underused in Greek. It's way better than Cyrillic at any rate, and it's nice that polysyllabic words will always have a tonos to make use of that space. I do think Latin could use its descender space a bit more, but even that I don't find too problematic. Maybe I'm easy to please, but unless extenders are too few and far in between (as in Cyrillic) for me the gains in usefulness start diminishing with increased frequency. I think it is an overreaction to see fiddling with the descender of a single Greek letter as an instance of jingoism. There is of course the danger that non-descending η will be designed as if it were the Latin lowercase n due to their superficial similarity, and that would be an instance of true Latinization—but I think we would all agree that this is clearly something to be avoided.

I'm pretty sure by now that it is a personal idiosyncrasy of mine to be bothered by vowel letters not confined to the x-height. Maybe it's just my subconscience making up reasons after the fact. I think it must surely be entirely coincidental that in the development of lowercase letters, the vowel letters a, e, o, u, α, ε, ι, ο, υ, and ω became confined to the x-height, just as it must be coincidental that the even numerals 6 and 8 ascend beyond the x-height while odd numerals 3, 5, 7, and 9 descend in old-style figures (and yes, I wouldn't mind seeing 2 and 4 ascend just to preserve that symmetry :) ).

Tatiana, I don't have Pflughaupt's book but I'll try to check it out. That's an interesting note about the vowel/consonant ambivalence of η. And of course, the Latin h comes ultimately from the western versions of the Greek alphabet that did keep the consonant heta. Incidentally, there is also vowel/consonant ambivalence with the Latin y, since it is often used for a consonantal sound as in 'yes' though in some languages like Finnish it is used exclusively for vowel sounds except in borrowed words.

Designing readability experiments is always difficult due to the impossibility of controlling variables. For example, Centro was posted as an example of a design with non-descending η. The problem I have with that design is that η now feels too close to π due to the serif in the left foot of the latter. The effect is subtle, but that introduces the tiniest bit of hesitation in reading and recognizing the letters. Not a problem in display typography, but for immersive reading I imagine there is an effect. One reason I have no problem with non-descending η is that there wouldn't be a danger of it being confused with another letter, but designing η and π to be identical in the lower half does bring that risk. Here I think the fault is more in the design of π rather than η—I can't quite get on board with the serif on π.

Hrant: * Speaking of Hangul, one amazing thing about it is that its four levels of syllable "visual density" are strongly associated with a syllable's grammatical function.

I'm not sure what you're referring to. It is true that CVCC or VCC syllables (with orthographic consonant cluster at the end; C standing for consonant letters and V standing for vowel letters) are most often found in nouns (e.g. 흙 hŭlg 'soil'), verb roots (e.g. 읽다 ilg-da 'to read'), or adjective roots (붉다 bulg-da 'to be red'). But they can also be formed by root+suffix as in 알- al- root of 'to know' + -ㅁ -m nominalizing suffix = 앎 alm 'knowing'. Other than that, CV, CVC, V, and VC syllables occur in all parts of speech, so they are not distinctive in conveying grammatical function. We can't really say anything beyond that the visual density of hangul conveys the underlying phonemic density of each syllable, e.g. 그 is visually much sparser than 끊 ggŭnh as the latter packs more letters into the syllable.

Tatiana Marza's picture

Brian,
I think you are another rare case Hrant was referring to earlier:) How many languages do you know?:)

I don't think the ascender space is too underused in Greek.
I made a small experiment with fonts which retain the asc/desc and which don't. And maybe you have a point regarding the ascender space used in Greek text by tones.
I covered the x-height of these 2 fonts (1st-Parachute's Regal Text Pro, with fewer asc/desc and 2nd is GFS Elpis, with all traditional forms of Greek characters).
The result is interesting. The first thing which came on my mind is "optical noise". I do not agree with this term, because it automatically implies something negative - noise.


Tatiana Marza's picture

And something else.
Panos wrote that There is a total of 10 Greek letters with descenders compared with a total of 5 for Latin .
I wanted to represent this in my paper by an image and for some reason I encounter 11 descenders. Where is my mistake? Is 'ς' not considered a letter with descendant part? Then one could say the same about the descendant part of 'ζ ', and 'ξ'.

I see 9 ascender letters in Latin: b,d,f,h,i and j (I assume they are using ascending space), k, l, t.
And 5 descender letters: g,j,p,q,y.

For Greek, 8 ascenders: β,δ,ζ,θ,λ,ξ,φ and ψ (depending on font, φ can have an ascending part or not, same stands for ψ). So how many ascenders are there?
Descenders, 11: β,γ,ζ,η,μ,ξ,ρ,ς,φ,χ,ψ.
Of course, some of the letters belong to both categories, asc and desc.

Jongseong's picture

Tatiana, that is a great little experiment. Note to self: do the same with various languages that use the Latin alphabet like Hawaiian, Finnish, and Vietnamese.

How many languages do you know?:)
Knowing a language is not the same thing as knowing about a language. :) I know the Greek and Cyrillic alphabets and more or less know how to read them out loud in Modern Greek and a number of languages that use the Cyrillic alphabet (though it's difficult to know which syllables are stressed in Russian), but I don't speak any of those languages.

So take my views as that of an outsider, not those of people who live and breathe the Greek language and alphabet, who should be the ultimate arbiters of Greek typography. My native language and alphabet are Korean, so I think that may colour even my approach to Latin typography.

hrant's picture

Ilya, your ingénue post made me smile, because Typophile has in fact seen so much discussion about exactly those things that you could publish a three-volume book! Although it would be hard to read... Kevin Larson (the MS researcher you speak of) himself has mixed it up here in the past.

My view remains that Larson's research has only skirted immersive reading, not fully delved into it. If you're interested in this discussion a really good place to start is issue #13 of Typo magazine, where Larson, Peter Enneson (also a Typophile member) and myself have each presented our views.

In any case, saying that descenders don't help is one thing, saying that they hurt is quite another, and to me pretty ludicrous. The only way that the descender of the η (eta) for example could harm readability is if it were the only descending letter, and without that linespacing could be reduced. Big maybe.

There is of course the danger that non-descending η will be designed as if it were the Latin lowercase n due to their superficial similarity, and that would be an instance of true Latinization—but I think we would all agree that this is clearly something to be avoided.

Sadly I don't think we all agree on that... For one thing people still make such Greek fonts.

About "true" Latinization: To me it remains an issue of intent. For example if you make the η without a descender because you really believe the descender harms readability, then it's not Latinization; but if you do it because you want Greek to align better with Latin text (as opposed to sitting too low overall) then that certainly is Latinization.

Also, don't discount the potential confusion between η and the Latin "n".

But I certainly agree that Cyrillic is in much worse shape.

it must be coincidental that the even numerals 6 and 8 ascend beyond the x-height while odd numerals 3, 5, 7, and 9 descend in old-style figures

Who knows?
To me the important thing to realize is that it's a bad scheme. I prefer the old French style of making the "3" and "5" ascend. And yes, the "2" should ideally ascend (something Gill has actually done). Nothing to do with my personal "symmetry" preferences though!

hhp

J Weltin's picture

Quite interesting, Brian!
Thinking a bit further it wouldn’t to the Latin alphabet any harm having ›n‹ and ›m‹ with a descender (if Greek eta would stick to have none). Wouldn’t that be a nice crossing of cultures? Europe could need that right now …

quadibloc's picture

Since the even 4 descends, I am inclined somewhat in the direction of coincidence. Typewriters with conventional faces often have lining figures that extend slightly above and below the lines following oldstyle figures, which suggests a basis in the shape of the character. The 5 indeed suggests the 2 should ascend, although one could argue that it should descend based on the 6 and the 9.

But rather than that being why it does neither, I would tend to think that the behavior of old style numerals is due to their being slightly closer to the Hindu and Arabic precursors of our numerals. When Arabic numerals were first adopted, various forms were used, and they looked quite different from what we are now used to.

hrant's picture

Jürgen: I've long thought that the middle stem of the "m" should descend.

hhp

eliason's picture

Thinking a bit further it wouldn’t to the Latin alphabet any harm having ›n‹ and ›m‹ with a descender

...aŋ iŋterestiŋg idea...

Jongseong's picture

Sadly I don't think we all agree on that... For one thing people still make such Greek fonts.

I guess I was thinking about the audience of this thread and specifically the people who give enough thought to Greek type design to treat it with the respect it deserves. But you're right, there are some embarrassing instances of uncritical Latinization out there, with the Latin n glyph doubling as η.

For all my arguments in support of non-descending η, I would use the descending η shape by default in my own designs unless I'm specifically targeting Greek-only use. An important reason that I would include Greek in my typefaces is that Greek letters are often used as symbols for technical purposes, for example in particle physics. To make η recognizable for such purposes, the descender is a must. It's somewhat ironic perhaps that for me the best argument for a descending η is to meet the expectations of a Latin-centric audience.

J Weltin's picture

You’re right, Craig :-) How frequent is eng? Besides, the eng’s descender is turning inwards (but is this a must?)
Sorry for getting off-topic.
@ Hrant: why in the middle (m)? The Phoenician Mem had its descender on the right.

And yes: for mathematical setting you’d need eta with descender to recognize it as such. Because ›n‹ means something different.

hrant's picture

Brian, your last sentence made me smile.

the eng’s descender is turning inwards

Which is ideal, because it avoids looking like a "p" or "q"* (or rather, boumas it's in avoid looking like they have a "p" or "q" instead).

* Which BTW I strongly prefer with a rightward curl.

why in the middle (m)?

Because that would pull it away from "rn".

hhp

ilyaz's picture

> Besides, the eng’s descender is turning inwards (but is this a must?)

Looks like a must. Cutting and pasting from my table of n's:

nNñÑ ńŃʼn♪ ǹǸƞȠ ♪ɴꞑꞐ ñÑᵰᴎ ṅṄṇṆ ņŅṇṆ ňŇɲƝ ŋŊɳⁿ ꞥꞤꝴ♪ ♪ℕꞥꞤ ᶇ♪ȵ♪ ꠸₦꠷꠷ ℕ♪♪♪ νΝ∨⋁ нНњЊ נן׆◌֦ ңҢҥҤ

So, keeping only Latin, and removing diacritics, one gets ʼn ƞȠ ɲƝ ŋŊɳ ꝴ ᶇ ȵ to distinguish. Some of these are obsolete; on the other hand, the keyboard supports only a few paleo-Latin characters yet. (And, on the Cyrillic front, there is also ◌ⷩᵸӈӇ — speaking of the devil, you wanted Cyrillic descenders? ;-)

ilyaz's picture

> Typophile has in fact seen so much discussion about exactly those things that you could publish a three-volume book!

We can discuss things ad infinitum. Then “a really clever guy” comes, and says that he knows a way to measure what we thought was a number of angels on a needlepoint. I have seen my share of unbelievably wise experimentators…

BTW, googling for “microsoft vision descender focus typography” brings http://www.microsoft.com/typography/ctfonts/wordrecognition.aspx as the third reference. I suspect it is the second of the quotes I mentioned yesterday (but my summarization was very flawed). Update: http://typophile.com/node/77618#comment-446202.

Tatiana Marza's picture

Ilya,
did you set this table of letter information?
And do all of you have such kind of information? :) How organized one could be?!!!

Brian, of course is difficult to make readability experiment with latinized texts. But, I think measuring pain http://www.ted.com/talks/dan_ariely_on_our_buggy_moral_code.html, for example, is even more difficult. There are even more variables.
I just wouldn't exclude the possibility of our experiment. Except if, there is an easier way to find an answer to our question.

BTW, could someone enlighten me on the asc/desc numbers? To identify where is my mistake?.... Thank you.

hrant's picture

Ilya, allow me to tell a joke that my dad (a PhD in Soil Mechanics) enjoys:

A scientist wants to measure the effects of amputation on a cockroach. He cuts off one leg and says: "Walk!" The cockroach walks OK, just not very fast. The scientist writes that down. He cuts off two more legs. The cockroach now has three legs. The scientist says: "Walk!" The cockroach struggles, and slowly moves forward. The scientist writes that down, cuts off two more legs, and says: "Walk!" The cockroach struggles immensely and pushes his body forward a few millimeters with his one remaining leg. The scientist writes that down, cuts off the last leg and says: "Walk!" The cockroach does not move. The scientist writes down: "When all the cockroach's legs are amputated, it loses the ability to hear."

Numbers don't lie. Because numbers don't say anything. People looking at numbers say things, and they often see/say what they expect/want. To uncover a Truth, a scientist cannot merely be "clever", he must be wise. And that's a lot to expect.

hhp

Tatiana Marza's picture

:) :) :)

quadibloc's picture

In looking for information on the typeface "Gunaika" (from the name of the images of a very Latinized Greek in Bodoni style) I found this page:

http://missosology.info/forum/viewtopic.php?f=262&t=136286&start=30

and it certainly makes sense that a fashion magazine, trying to appear modern and cosmopolitan (κοσμοπολίτικα!) will eschew typefaces steeped in tradition, but instead seek to be keeping up to whatever is being done in Paris and New York.

However, looking more closely, I see they use n for eta, which is going too far. I could not find more examples of that typeface from that magazine, but I see the newspaper Το Βημα is also a user of a heavily Latinized typeface.

ilyaz's picture

> Ilya, allow me to tell a joke that my dad (a PhD in Soil Mechanics) enjoys

I must confess that decades ago, when I first heard this, I must have enjoyed it a lot too.

For the rest of your post: I do not think it is productive for you to debunk things I did not say. I did not say that smart=wise (often I tease my French friends by being no such distinction in French). I did not say that being a scientist makes one a superior human being. I did not say that one should expect a scientist to be smart or wise.

Still, I have seen many wise people, many of them scientists; some of them experimentators; and some — experimentators in cognitive psychology. And while I consider myself quite inventive in the stuff I do often, I really appreciate how superior the approaches of great experimentators are — compared to what I could think of when I was meditating on the same topics.

Tatiana Marza's picture

John (Q),
here are some recent examples of latinised fonts (not sure which) from journals and newspapers.



quadibloc's picture

I had done some searching, and I found that there were no free Greek fonts that I could find with that level of Latinization... except that one of the fonts on the Greek Font Society site did have the eta without a descender - and that was Artemisia, which I had thought before was a typeface that resisted Latin influence, as it modified the companion Latin to make it more stylistically like the Greek.

In the case of capital upsilon looking like Y, so many of the typewriter faces for Greek available for the Selectric were like this that I was driven to the conclusion that the distinctive upsilon was likely regarded as an affectation of foreigners by the Greeks. I don't know about the eta with a descender - there seems to be evidence either way.

It may be good for the health of Greek type design, though, if Latinization turns out to be a passing fad of the present rather than an enduring trend. Incidentally, Ignacio among the Majuscule typefaces on the GFS site illustrates Greek uncial - however, it's too much like Greek capitals to suggest much of a proper direction for Greek lower case. Except as small capitals like Cyrillic - so it may explain why Cyrillic lower case is the way it is.

Of course, this thread has inspired me to do searches from which I learned things. Thus, I thought New Hellenic marked a genuine Greek direction; neither foreign-imposed archaism (Porson) or slavish copying of the Latin model. But now I see that it was designed by Monotype in Britain, not by anyone in Greece. Its success in Greece still indicates it has some value in showing what is valid to Greek readers, but I'm disappointed.

hrant's picture

Ilya, if you've stopped laughing at that joke, maybe something has gone wrong. I've known it for over 30 years and it still brings a smile to my face. And I remember once telling it to one of the directors of the MIT Media Lab: she laughed so long and hard that I was actually a bit embarrassed for her. What I'm getting is: laughing at scientists who believe that gigabytes of data collected by spending millions of dollars can overcome lack of real insight never gets old.

Tatiana, "great" examples! :-/
That first one has so many tell-tale signs of getting Greek letters by hacking Latin ones that I have to show the same sort of thing for Armenian:


It's a slide I used at ISType last year*, showing what I think it should look like above versus what it actually looks like below.

* http://typophile.com/node/91821

there were no free Greek fonts that I could find with that level of Latinization

I don't know how hard/deep you looked, but actually in general highly Latinized fonts in any script tend to not be put for sale. This is because: they typically contain stolen outlines and it seems a bit much trying to make money from that; they take so little time to make it feels OK to give it away. Maybe you found none because such fonts often stay "internal" to the designer.

BTW, I think Artemisia is actually the same design as the Apollonia (which was also called Appolonia) I mentioned...

if Latinization turns out to be a passing fad

Wishful thinking. :-/
As long as Greeks admire Western Europe and the US (because that's what's on TV) it will remain strong.

hhp

Tatiana Marza's picture

Hrant, I find even more radical substitutions in Armenian font than in Greek:)
The interesting fact is that although less latinized Greek fonts are designed today (compared to '70s - '90s), there is still a high request for them. So, indeed, as long as Greeks admire Western Europe and the US (because that's what's on TV) it will remain strong.

quadibloc's picture

Searching on γραμματοσειρά instead of font helped me finally find some more examples of Greek typefaces with n as eta:

SK Berkeley stood out; two other examples were Magenta's versions of Futura and Antique Olive.

EDIT: This page, which illustrates many free and pay Greek fonts: http://www.polytoniko.org/fonts.php?newlang=en was my source.

ahyangyi's picture

Frankly speaking, the only scientists I personally know who actively use Greek letters in a continuous flow (for scientific purposes, not when they write letters to their moms) are philosophers. Tell this to them! ;-)

You might want to read this. It is one of the most important paper in Distributed Systems.

quadibloc's picture

Every now and again, a scientific monograph will introduce itself, or chapters, with a quotation - and these quotations can be in foreign languages, or even in Greek. And then there's Φωτεινή Μαρκοπούλου-Καλαμαρά, a theoretical physicist who I am sure uses Greek letters in a continuous flow often in daily life - but you already covered that possibility.

Jongseong's picture

To make it clear, I do not view the mere fact that η does not have a descender as an indication of Latinization.

There is a lot of variability in early Greek minuscule writing with respect to the baseline—θ would frequently descend instead of sitting on the baseline as it does today, κ may have an ascender, etc. A non-descending η is common in these early stages; see the images in the table at the Wikipedia page on Greek minuscule. Many early Greek typefaces have η without a descender and others have very short descenders.

So one need not invoke the influence of Latin to explain the non-descending η. To me, Apollonia (GFS Artemisia) is not a Latinized design at all. Notice how Apollonia's η does not have the typical serifs of an upright Latin n.

In contrast, of the three images that Tatiana posted as examples of Latinization, the first and the third are clear examples of substitution of the Latin n for η. Maybe not as painfully obvious as some of the Armenian substitutions in Hrant's example, but these are clearly n's masquerading as η's. The first image is particularly grating as other than κ, η is the only Greek letter with full serifs. The third image at least is consistent in going all out with the serifs.

The second image, assuming we're talking about the sans serif, does not strike me as an unambiguous case of Latinization (though I'm not a fan of the non-descending χ as I mentioned earlier), unless you are willing to level the claim against practically all instances of sans serif designs of simplified geometry.

quadibloc's picture

To me, there is Latinization and Latinization. If the Greeks themselves want to use the same typefaces as Latin script users, and they can adapt the letter shapes to them in a way they find valid, I might be mildly bemused. When all the available Greek faces are based on a pattern set by foreigners, the fact that this pattern is one of conscious archaism rather than imposition of their own script doesn't mitigate matters for me.

So an authentic interpretation of an indigenous model is what I see as the ideal.

A Latinized typeface designed in Greece by Greeks will still testify to what aspects of the character shapes are genuinely relevant to Greek readers - so such a typeface is highly informative to outsiders trying to design Greek typefaces.

But the worst thing is not such a Latinized typeface; the worst thing is a typeface designed by foreigners who are telling the Greeks they should print with this interpretation of Byzantine script. Since that's what Greek has been digging itself out of, Latinization seems to me like small potatoes in its case.

hrant's picture

one need not invoke the influence of Latin to explain the non-descending η.

But Brian, when Panos says "Large number of letters with descenders. There is a total of 10 Greek letters with descenders compared with a total of 5 for Latin. This automatically makes Latin easier to read." what do you think his motivation is in making the η not descend? And is he alone in looking -maybe even fishing- for reasons to make Greek more Latin-looking?

And most of all, beyond leveraging precedent (which BTW can be done in earnest or as a ruse) and beyond Intent there's the matter of what effect it's having on the ground. Should type designers only worry about being paid by corporations to short-change their native culture? To be fair, Greek culture isn't highly endangered (although I do think it's threatened) so the issue of forestalling Latinization isn't as urgent as for Armenians for example. But there is good reason to worry.

I agree with you that Tatiana's second example isn't very Latinized. However:
- What effect is that "x" having on Greek culture?
- What are some reasons to make the χ like that? Are all of them good reasons?

And Apollonia/Artemisia is indeed not at all Latinized (in fact as John S. pointed out its Latin is Hellenized) but that doesn't make its non-descending η a good idea.

this pattern is one of conscious archaism

I agree that's just as bad as Latinization. But it's not worse.

The future of a people can hinge neither on the past, nor foreigners, nor mere materialism.

hhp

quadibloc's picture

@hrant:
I agree that's just as bad as Latinization. But it's not worse.

I will clarify something here, though. Conscious archaism or Latinization are equally bad when carried out by foreigners who don't understand the nature of the script. (Actually, Latinization is worse in that case, as it is likely to be carried out with overt imperialistic intent.)

Latinization carried out by native speakers of Greek who at least know what they're doing is, in my opinion, not only less bad, but even considerably less bad. In that case, what is genuinely important is likely to be preserved, and with any luck, mostly only that which is incidental and superficial will be lost.

Thus, look at "alpha" in the Latinized Greek typefaces we've seen in this thread - it is still a single-story a, not a Latin double-story a, and there are rather un-Latin things going on in the serifs. As long as the baby isn't thrown out with the bathwater, I can't get too upset.

Jongseong's picture

Hrant, I am puzzled by Panos' argument about descenders as well, the premise of which I don't agree with in any case. I know you place great importance on intent and motivation, and it does seem that Panos is pointing to an aspect of Latin to mimic in Greek typeface design. I hope we can hear his take on this issue again.

John (quadibloc), when you say conscious archaism, are you referring to the non-descending η? That this is a recent innovation with historical precedent? Maybe someone can fill us in on Greek handwriting since the early years of printing, because my impression was that a non-descending η continued to be at least an option in handwriting even as Greek typefaces in the West settled on the descending form. A non-descending η makes it easier to connect letters in cursive—you could write την in two strokes if you wanted to.

hrant's picture

native speakers of Greek who at least know what they're doing

1) This might sound strange coming from me, but: I wouldn't draw such a firm line between foreigners abroad and natives living in the country. For one thing, what about somebody (like Gerry Leonidas) who is an expat? Or a Russian who moves to Greece because he loves making Greek type. Often "converts" are more Catholic than the Pope! Any culture belongs to the whole planet to some extent - we're all in this together. And as important as nativity is, it doesn't guarantee really grasping the needs of one's culture. Just because a person is closer to something does not give him the right to abuse it more.

2) When they don't know how reading works, they can't really know what they're doing. A text font isn't just about formal harmony.

hhp

dezcom's picture

My dilemma is with the η looking like a Latin "n". This may be my problem as an English language speaker who also speaks Greek (albeit poorly). Since Latin is my primary script, I am far more likely to have that confusion than a person who only reads Greek script. So the problem may be small for native Greeks but large for non-native speakers of Greek who also read any of the Latin languages.
Since there is absolutely no good beneficial reason to cut off the descender of the η, why would someone want to? I have seen no science that infers that removal of descenders helps readability--if so, everyone in India has a big problem.

hrant's picture

Really, what proportion of Greeks is unfamiliar with the Latin alphabet?

why would someone want to?

To pretend Athens is Paris.

hhp

dezcom's picture

Practically none of the Greeks in the World are unfamiliar with Latin. So all the more reason to not cut the tail off of the η.

My point was to avoid a MAJOR differentiation issue, just leave the η alone. There is no good to come from it.

@Hrant, I once gave a little presentation here which I titled "An American in Piraeus" where I discussed the problems of an American born Greek designing Greek type ;-)

Jongseong's picture

Chris, you're assuming that the default form of η is that with the descender. This is of course true for most Westerners since this is the only form they are familiar with. I don't think it's surprising that Greek typefaces made in the West settled on the form with the descender in the last couple of centuries to differentiate the η from the Latin n.

It's just that I haven't seen any evidence yet that the Greeks themselves ever stopped writing the η without a descender. If they've been using this form all along, you can't accuse them of cutting off a descender that wasn't there in the first place.

But I haven't seen much about Greek handwriting in Greece during the Ottoman period, and today's Greek handwriting seems to derive from copying printed forms (hence the ahistorical fish-shaped alpha) rather than historical handwritten forms, much like a lot of today's Latin handwriting.

John Hudson's picture

It's a frustration that Greek books on palaeography, of which there are many excellent ones, all seem to stop around AD 1600. The best are the Repertorium ger griechischen Kopisten 800–1600, each issue of which consists of three volumes devoted to manuscripts housed in a different country: one descriptive volume (in German), one volume of photographic plates, and most usefully a volume illustrating the characteristic form of letters occurring in each manuscript. A quick flip through the latter volume in the French issue reveals both descending and non-descending η in use. Of particular interest, I think, is that the manuscript non-descending η is often very difficult to distinguish from the cursive κ, such that in many of the manuscripts that feature the non-descending η the scribe uses a different form of κ with a Latin K construction. Where both the non-descending η abd cursive κ are used, it is very difficult to distinguish them other than by textual context. This may provide some insight as to why the descending η became the common typographic form: to contrast with the form of κ that was adopted in the cursive types as the norm for that letter.

ilyaz's picture

> You might want to read this. It is one of the most important paper in Distributed Systems.

Never read anything Lamport’s outside of LaTeX. However, μαδζ∂ωριτῐσετ does not look too Greeky for me. Did Lesley know Greek? Or is it that my non-knowledge of Greek shows? (That ∂ is kinda suspicious; I though that nowadays Unicode contains most of the paleo-Greek, and I can’t place it…)

quadibloc's picture

By conscious archaism, I was thinking of Porson and of the earlier faces more closely modeled on Byzantine script, not faces without a descender on the eta.

@hrant:
This might sound strange coming from me, but: I wouldn't draw such a firm line between foreigners abroad and natives living in the country. For one thing, what about somebody (like Gerry Leonidas) who is an expat? Or a Russian who moves to Greece because he loves making Greek type. Often "converts" are more Catholic than the Pope! Any culture belongs to the whole planet to some extent - we're all in this together. And as important as nativity is, it doesn't guarantee really grasping the needs of one's culture. Just because a person is closer to something does not give him the right to abuse it more.

Well, I do admit that knowledge of the Greek language isn't coextensive and coterminous with Greek residence or nationality. But I didn't think it necessary to discuss corner cases to make it clear that being an expert typographer is not enough to give one knowledge of the gestalt of a script not one's own.

One can make good guesses with insight and study, though; a failed attempt by an outsider can give inspiration to someone with the right background to achieve something that otherwise would not exist.

One shouldn't abuse one's own script, but one's compatriots would presumably reject abuse for what it is.

As for pretending Athens is Paris... ah, others would wish to pretend New York, or Montreal, or Berlin, or London, or even Beijing or Pyongyang is Paris! Eventually, though, Greek type designers may develop a rich palette that has no need to rely on foreign models. But I can't demand they wait for that because I am charmed by their quaint indigenous characteristics.

And the trouble is that the Greek typefaces we're used to were made by foreigners - so I am without a real reference for what Greek should look like at the present time.

dezcom's picture

:...But I haven't seen much about Greek handwriting in Greece during the Ottoman period, and today's Greek handwriting seems to derive from copying printed forms ..."

Jongseong,
Somewhere I have a stack of hand written letters from my family in Greece to us here in the USA. They are all pre 1960s but as I recall, none of them had an η without the descending tail. When I find them, I will scan and post some samples.

"This may provide some insight as to why the descending η became the common typographic form"

John, all the more reason then to keep the tail. This goes to show that older is not always better ;-)

Jongseong's picture

Somewhere I have a stack of hand written letters from my family in Greece to us here in the USA. They are all pre 1960s but as I recall, none of them had an η without the descending tail. When I find them, I will scan and post some samples.

That would be great, Chris!

In terms of handwriting, the confusion between non-descending η and κ is an excellent reason for preferring the η with a descender.

To keep playing the devil's advocate, strategies for disambiguating similar letters in handwriting are not necessarily translated to typography.

In Cyrillic handwriting, strokes may be added above or below т and ш respectively to tell them apart because they can be identical when scribbled, but since in print these letters are easier to distinguish without additional strokes, these are not used, much as in print the strokes that some people add to the numeral 7 or the letter z in handwriting are not used. Likewise, there is no danger of confusing the printed forms of non-descending η and κ, so there is no necessity to add the descender in print.

However, if η with a descender was indeed established as the exclusive handwritten form due to the above reason, then that may have contributed to the same form being adopted exclusively in print in later centuries. If this indeed became the only form used in handwriting or print, that would mean that from then on, η canonically had a descender.

The problem I have with this narrative is that a lot of today's Greek handwriting uses non-descending η. This may be an innovation, a possibility that I admit didn't occur to me at first. But since a non-descending η was already in common use centuries ago, I can't be sure.

Syndicate content Syndicate content