Greek stress angle in art nouveau typeface

Thomas Phinney

I've been wrestling with a design problem for the Greek lowercase in this late-19th-century style art nouveau typeface, Cristoforo. The upright has been long done, and the italic is pretty nearly done... but the Greek is bothering me, so I am presumably about to revisit it in both fonts.

Stress is the angle at which the thick vs thin parts of the fonts are emphasized. Traditional Greek lowercase has a different angle of stress than Latin. The average Latin font as having a stress angle of about 30° (oldstyle) to 0° (modern), but old Greek is typically 120° in the lowercase. However, more “latin-style” Greek lowercase with latin stress angles is also common and acceptable, depending on the typeface. (Greek caps don’t get the same stress angle, though. They tend to consistently get the same stress as Latin.) 

This particular typeface is a Victorian Art Nouveau thing with vertical stress in the Latin (and Cyrillic). So the Greek caps also have vertical stress. What should I do for the Greek lowercase?

Latin in the font:

Two different approaches to Greek lowercase, below—currently both are in the font. Top is traditional Greek, bottom is more latinized.

Currently, there is this odd mix. I am thinking that's a bad thing.

Thomas Phinney

Oh, and feel free to make fun of that lowercase delta.

Ray Larabie

The Latin letters are pretty condensed but the Greek isn't. I think you should start by getting the width of the omicron closer to the o. Then make omicron's thins and thick the same as the o. From the Latin letters, I can see that there's emphasis on strokes being either thick or thin and not much in between*. Gradual curves are somewhat minimized. Now look at the left side of your alpha...that gradual thickening doesn't suit the rest of the typeface. Go back to the omicron. If the counter is more of a Contac C capsule shape, it minimizes strokes that aren't thick or thin. This nib logic is hard to follow in the Latin but it's there. It's just a matter of priorities. Thick and thin and reducing the widths in between takes priority over nib logic.

At that point you can start harmonizing the rest of the Greek letters the omicron. For the bottom of the rho and other straight descenders, look to the flag shapes of the E and F... they might look good turned 90 degrees. I'd join the beta bowl the the stem...I think an inner curl is a little too showy, every for such a fancy typeface.

* Except the bottom of the bowl of the g which looks like a mistake to me. Not the tail...I mean compare the bowl to the d.

Andreas Stötzner

I would prefer not to do it at all. Because its not possible, actually.

That Latin typeface has a strong 0° contrast orientation. If you mix that with a traditional Greek sort-of-60° contrast orientation, you’ll end up with a different typeface. You can’t blend oil and water, you simply can’t. To give half the Greeks one direction and the other half the other direction, doesn’t work either, because it destroys the coherence of the Greek minuscule set. On the other hand, to avoid a decision by making two sets, would be not satisfactory to me.

The only possible option I can see is to take the horridness of the Latin as a starting point here. Be so bold and do the Greek ruthlessly with that ‘Latin’ stress pattern and bless the glyphs with all those terrible bits and curls and bobs and try to go as Byzantine as they once used to be. – This would be my playful approach; I doubt you’ll ever get any ‘serious’ solution for this. Maybe, the dirty way is the right one here. But I’m not sure if it is possible, at all.

Kent Lew

You can’t blend oil and water, you simply can’t.

And yet, salad dressing does this every day. It’s technically an emulsion, and it isn’t infinitely stable. But it can be done to some degree.

Shake well before serving. ;-)

Andreas Stötzner

Kent Lew said:

Shake well before serving. ;-)

this is precisely what I meant.

Craig Eliason

Fun but challenging project!
I agree with Ray that the counters need to be narrowed. The "ηικμ" sequence in the second line stands out as much closer to the Latin not only because of the stress but also because of the smaller interior spaces. 
I'd maybe loosen up the head serif on eta/iota/kappa, which obviously matches the Latin but feels out of character with the rest of the Greek. It's tempting to make the parts that seem most "matchable" match, but in effect that can make the more intransigent parts seem even more disagreeable.

Thomas Phinney

It's true that it is something of a wacky thing, to which I just have sentimental attachment because of its usage with some particular books. The Latin was designed in 1892 by Hermann Ihlenburg. I just expanded the character set and added Cyrillic and Greek support. But the Greek lowercase is not “there yet.”

Thoughts about going straight Latinized and making the Greek lowercase narrower make sense to me, I think.

Thomas Phinney

I mean, making the angle of stress and the width of counters better match the Latin lowercase. I'm down with that. The Latin lowercase is just so damn stiff, it feels even more so for Greek. But that may be the right thing for this design. (I am amenable to the idea that even doing Greek is a mistake for this design. Unfortunately, if so, it is a mistake I am committed to!)

Ray Larabie

I'd lean into it and go all the way with the Greek stress.

Andreas Stötzner

Thomas Phinney said:

…it is a mistake I am committed to!

that opens the door for something new and thrilling. Do it!

K Pease

Perhaps a Latin o at 0° deserves a Greek theta and omicron at 90°?

I'd say lean a lot more on flares and make it harder to identify a prevailing angle at all. Look to the numerals for inspiration.

I can spot a few examples of the wrong kind of Latin influence here: beta is not an eszett (close the bottom), and eta is not an eng (its tail would sweep to the right).

Thomas Phinney

So, I also got some comments back from Gerry Leonidas, who is probably more qualified than any of the rest of us in this area. I'm sharing his feedback for anyone who might be interested. I had indeed not seen his recent Medium article, nor the related essay in the book Bi-Scriptual—a book which looks very interesting, for anyone doing multi-script type design (https://www.amazon.com/gp/search?index=books&linkCode=qs&keywords=9783721209822) 

* * * * *

I’m happy for you to post my comments, I just don’t have time myself for discussion forums.

Have you seen <https://medium.com/@gerryleonidas/designing-greek-typefaces-eac0de7767cc>? There’s some relevant information there, especially in the “expanded families” section, which also shows examples of heavy styles. Sérgio Trujillo’s Satira <http://typefacedesign.net/typefaces/year/2015/satira-2/> is a particularly relevant example of how to deal with extreme weight in a modulated Greek. 

Some problems in Cristoforo derive from the behaviour of the (imagined) [writing tool, one assumes—something was left out here] you are using. The tool is lifted for the β, δ, ρ, σ before junctions, whereas written Greek does not require that, and neither does the pattern of the Latin. So, close the counters (in the β the bottom only). In the δ the diagonal is a main stroke, and should be heavier; rotate the stress, and taper the closing stroke of the bowl.

Then, modulate more: for example, the τ correctly has a heavy horizontal stroke, but should have a modulated downstroke, starting thin at the joint, then swelling towards the outstroke.

Have a look at the consistency of the tapering outstrokes (α, ι, μ, τ) for how much you “stretch” the stroke, and the pointiness of the tip; and for ball terminals the angle and sharpness of the joint with the supporting stroke (e.g. δ, ν, υ). Similarly for smoother outstrokes (ζ, κ, ξ). And for in/outstrokes like on the φ and the tip of the middle stroke of the ε: these do not align with any tool behaviour evident in the Latin or other Greek letters (compare the tip of the ρ, the v, w). Overall there needs to be more of a sense of the consistency derived from a tool’s marks, whereas here the stroke thicknesses and in/outstrokes have a variation that suggests an outlines-first approach.

So, I’d go back to writing first, with a high-contrast pen or marker, to see where the modulation would happen. That would probably change the instroke of the λ to a thick stroke, strengthen the middle of ε, give the η an inclined stress, and eliminate the “serif” at the top of the ξ.

Good luck,

Gerry 

and his one later comment: “My pleasure. It’s not an easy style you’re working in…”