principles for designing Greek "italics" (aka slanted Greek)

edited January 2020 in Technique and Theory
Historically, Greek typefaces would not include both upright and slanted italic characters like Roman faces do, instead choosing generally either an entirely upright face or an entirely slanted one. Greek typefaces also being traditionally somewhat calligraphic, in every Greek family I have seen with a matching italic/slanted face, the change is much simpler than an italic is for Roman (since the basic calligraphic nature of the letters has not changed). So when designing a Greek font, do you prefer to slant the upright as a basis for an italic, or start again from scratch and, if so, with what principles for matching with an upright Greek/Latin and/or the Latin italics?
Tagged:

Comments

  • It is right that the upright/corsivo ratio, inherited in Latin typography for a long period, had no match in Greek typography for most of the past. In many older Greek typefaces you find calligraphic vigour and (more or less) an upright structure the same time, whereas in Latin type usually calligraphic expression was reserved to the (slanted) Italic sorts, but not applied to the upright Regulars.

    When you want to decide for a new Greek or Latin-Greek design which provides upright and Italic for both scripts, my advice is this: start with reg. Latin. Then do the reg. (upright) Greek and watch Greek models while doing so. Then compare your Greek Regular to the Latin Regular and think about the balance of harmonizing or differentiating them.
    Then turn to the Italics and operate in the same manner.

    A general rule can hardly be given, because it always depends heeavily on the nature of your intended design and the scope of style and usage you want to serve for.

    And try to study materials about this topic as much and thoroughly as possible. Designing a good Greek typeface is hard to master, to design a convincing Greek-Latin match is even more demanding.


  • Chris LozosChris Lozos Posts: 1,458
    Greek italic seems more free to me.  The few Greek italics I have done, I have done from scratch, not with slanting the upright Greek. But that is just me. I tried looking at handwritten letters my mother received from our family in Greece but the hand-written form is too different and I could not make the Latin italic fit with it.
  • John HudsonJohn Hudson Posts: 2,955
    edited January 2020
    A lot of Greek publishing in the 20th Century used what were originally two distinct typefaces in combination: an upright, derived from Didot's 18th Century Greek types, so common as to be referred to by Greeks as aplá, i.e. 'normal', and a slanted derived from 19th Century German types, referred to as Lipsías, i.e. 'from Leipzig'.



    [Image from Atelier Fluxus Virus.]
  • Christian ThalmannChristian Thalmann Posts: 1,938
    edited January 2020
    I mostly just changed the design of θκφ and added some cursive in- and outstrokes. Hope that does it!
  • Rafael CasesRafael Cases Posts: 38
    edited April 2020
    I mostly just changed the design of θκφ and added some cursive in- and outstrokes. Hope that does it!
    What happens if you try using a round beta without a descender in the italic, lunate epsilon for uprights, pomega version of pi for italics and reverse-descender rho in italics?  With that, one creates a greater design distance from the uprights.
  • I seem to remember that lunate /epsilon/ is not recommended... and isn't pomega an archaic and obsolete form?
  • I seem to remember that lunate /epsilon/ is not recommended... and isn't pomega an archaic and obsolete form?
    I'd like to have a survey from the Greeks themselves.
  • John HudsonJohn Hudson Posts: 2,955
    Gerry Leonidas encouraged the ClearType Collection designers to make the rounded beta and the cursive theta the default forms for the italic styles in those families. This works fairly well (presuming one is ignoring some historical French practice to vary the form of beta contextually similar to long s vs s).
  • Nick ShinnNick Shinn Posts: 2,131
    In my Scotch Modern, I made the roman lower case quite stiff and upright, nowhere near as lively as the aplá, but consistent with its the Latin and Cyrillic counterparts. The italic is pretty much the Lipsías style, with its 19th century “pot hook” serifs that so characterize the Latin Scotch Modern italic.

    As the slanted/italic style is generally used for contrast with the upright/roman, it doesn’t seem optimal that they both be “scripty”, and it seemed to me that the traditional aplá/Lipsías relationship is back to front—with the upright/roman being even more scripty than the slanted/italic.


  • Christian ThalmannChristian Thalmann Posts: 1,938
    edited April 2020
    Does the round beta need that strange gap between the two bowls, or can I use the shape of Bulgarian /ve/?
    Oh, and is the tailed /rho/ a cursive form, rather than just an alternative one?
  • Does the round beta need that strange gap between the two bowls, or can I use the shape of Bulgarian /ve/?
    Oh, and is the tailed /rho/ a cursive form, rather than just an alternative one?
    The beta looks better, at least in the point of view of design distance between the uprights and italics.  How about using the two-stroke alpha (resembles a wide single-storey Latin a) in the upright and keep the italic alpha?  Once it's done, have the text tested with a Greek designer.
  • Deactivated UserDeactivated User Posts: 11
    edited April 2020
    I don't want to be intrusive here, and of course my opinion's worth much less than a native Greek reader's, but I think you might be suggesting too many alternate forms for a "normal" Greek typeface like this. I agree with you that the tailed rho and round beta are fine, and I'm ambivalent about the two-stroke alpha, but I wouldn't expect to see lunate epsilon or pomega as the default forms in any font, especially a conventional humanist sans-serif like this. And they have their own unicode points if you're dead-set on including them.
  • I don't want to be intrusive here, and of course my opinion's worth much less than a native Greek reader's, but I think you might be suggesting too many alternate forms for a "normal" Greek typeface like this. I agree with you that the tailed rho and round beta are fine, and I'm ambivalent about the two-stroke alpha, but I wouldn't expect to see lunate epsilon or pomega as the default forms in any font, especially a conventional humanist sans-serif like this. And they have their own unicode points if you're dead-set on including them.
    We're exploring the interactions of the glyphs if Greek uses the upright / italic treatment, and with the understanding of what the Greek type foundries release, the Greeks are looking for an italic variant.

    My parameters to create the italic are the number of strokes in a ductus, and fluidity.  If there are variant forms for a glyph in Greek, these too are to be explored.
    Whichever is more angular, has more strokes required to be written, or whichever form requires lifting of the pen, these become the uprights.  Case in point, the angular sigma was revived due to classical Greek scholarship in Western Europe from the 1400's to the 1800's.  If one from is more fluid, requires less strokes or requires almost no lifting of the pen, these become the italic forms.
    However, the constraining factor for fluidity is that the strokes must not have overarching and extremely flourishing elements like swashes or else, these will affect the overall grey colour.  The Greeks want Greek typefaces to have the same notan and overall grey colour as Latin typefaces, that is, under Tatiana Marza's paper.

    Based on my understanding, the lunate epsilon is a mere typographical variant for Greek two-tiered epsilon, but these glyphs in the scientific world are different mathematical variables.  The pomega is archaic, but it can be revived the way the angular sigma is revived.  Since Greek is rich in glyph variants, they are all to be explored and tested.  But the ultimate opinion on readability still lies with the native Greek readers, the scholars and scientists who also use them.
  • The /a/-like alpha smacks a bit of Latinization to me, and I hate it when I use an alpha in Latin text and it looks like an /a/ (e.g. Minion). 
  • Rafael CasesRafael Cases Posts: 38
    edited May 2020
    The /a/-like alpha smacks a bit of Latinization to me, and I hate it when I use an alpha in Latin text and it looks like an /a/ (e.g. Minion). 
    If you saw Andreas Stötzner's design process of Andron Greek in 'Questionnaire about the needs of Greek typography end-users' in Typophile, you'll understand it:

    https://web.archive.org/web/20170818015143/http://www.typophile.com/node/101331

    I also have an article from Gerry Leonidas in Medium which shows the two-stroke alpha in Candara:

    https://medium.com/@gerryleonidas/designing-greek-typefaces-eac0de7767cc
  • Rafael Cases said:
    If there are variant forms for a glyph in Greek, these too are to be explored.
    Whichever is more angular, has more strokes required to be written, or whichever form requires lifting of the pen, these become the uprights.
    This sounds like an interesting academic exercise, but for a typeface meant for everyday use, I wouldn't want considerations like that to dictate to me which shapes I should use for the upright. Especially since Italics are not as much a thing in Greek as they are in Latin, and therefore the uprights are going to see the lion's share of practical use.
  • Christian Thalmann said:
    This sounds like an interesting academic exercise, but for a typeface meant for everyday use, I wouldn't want considerations like that to dictate to me which shapes I should use for the upright.
    That's why academia and industry, despite their differing directions, need surveys to verify the claim.  Once the numbers are here, we can decide.
  • Christian ThalmannChristian Thalmann Posts: 1,938
    edited May 2020
    Looks like Gerry fully approves of the two-stroke alpha. Interesting!
    BTW, Rafael:
    Once it's done, have the text tested with a Greek designer.
    Like any Greek typeface, mine will be reviewed by an expert upon ingestion into Google Fonts. Looking forward to that!
  • John HudsonJohn Hudson Posts: 2,955
    edited May 2020
    I would distinguish two forms of similar appearance, only one of which I would call 'two stroke'. A lot of the forms that Gerry shows, are based on the first ductus, which is continuous (cursive):



    Both this ductus and what I think of as the 'fish ductus' with the crossed strokes on the right are found in early types and the manuscripts on which they were based.

    To spot the influence of this ductus in a typeface, look at the way the bottom of the bowl flows towards the top of the stem on the right, e.g.:


  • Nick ShinnNick Shinn Posts: 2,131
    edited May 2020
    The ductus could be the same (“one stroke’) in both your written examples, John—the difference being that in the second one, the pen has been lifted off the paper, to touch down again at the top of the right stem. 

    There is no knowing what ductus the pen follows once it leaves the paper, merely by looking at the marks made. But it does seem reasonable to assume a smooth trajectory. In fact, the change of direction at the top of your left example would be, as pure path without thickness, a sharp angle, or a very tight hairpin curve, slowing to a virtual standstill, as one does, for the abrupt change of direction, as tantamount to a fresh stroke as lifting pen off paper.

    Also, your written example is not really “one stroke”, as the beginning of both your alphas is extremely unlikely in calligraphy, involving pushing the nib north-westerly against the paper (for right-handed scribes). Although with a ball-point nib, it is possible to do that. So, with a split-nib pen, to write your left alpha would require a ductus of three strokes, not one.

    I believe that both types of alpha—this and the “fish”—would, in cursive Greek with a split-nib pen, start at the top right and proceed in a clockwise direction, easing up in pressure while lightly dragging up past 9 o’clock while simultaneously rotating the nib angle to prepare for going heavy across the top. The really different method is therefore the bottom of the three types you showed, which is too heavy on the left to accommodate any kind of up-stroke there, and is distinguished, by this stress, as the Latin construction, with its steady rhythm of strong downward stems. 
  • Well, I've added a two-stroke /alpha/ to Ysabeau as part of SS01, or the Infant cut, respectively. I rather like it. Up to the user to decide, then.

  • For what it's worth: I far prefer the first of these as a Greek reader, though it's not my native language.
    Well, I've added a two-stroke /alpha/ to Ysabeau as part of SS01, or the Infant cut, respectively. I rather like it. Up to the user to decide, then.



  • Nick ShinnNick Shinn Posts: 2,131
    edited May 2020

    Further to me previous post, which was based on my own calligraphy (mostly broad-nibbed), I consulted The Calligrapher’s Bible, by David Harris, and discovered that the number of strokes depends on both the nib/brush and the style.
    From left to right: Italian Copperplate, English Roundhand Copperplate, Italic. 
    Therefore (and although Harris does not cover Greek styles), from the perspective of “what it looks like”, one-stroke for the Greek “fish” style and two-stroke for the style shared with Latin do seem appropriate.
  • Well, I've added a two-stroke /alpha/ to Ysabeau as part of SS01, or the Infant cut, respectively. I rather like it. Up to the user to decide, then.

    Looks like it's time to test this with the native Greek readers.  The scholars outside Greece might like the more traditional versions of the glyphs, as stated by Uwe Waldmann.  Gerry Leonidas has set this standard; he stated in the Medium article that native Greek readers are major stakeholders, but all scholars and scientists outside Greece using this alphabet are also major stakeholders.
  • Nick ShinnNick Shinn Posts: 2,131
    Let’s not forget the ecology of taste. It’s rather like the single bowl/double bowl thing in Latin /g and /a.

    I wonder what percentage of new Latin types now have both, with one as an OpenType alternate. I’ve done that in quite a few typefaces. The trick is to decide which should be the default.

    It’s not much effort to provide an alternate /g; alternate /a requires a bit more work.

    But in polytonic Greek, alternate /α is a daunting prospect!


  • The trick is to decide which should be the default.
    That's why I always provide spin-off fonts embodying the most common OpenType variants! That single-storey /alpha/ is definitely going in my infant cut along with the single-storey /a/.
    (BTW, having designed that two-stroke /alpha/, I'm wondering whether I shouldn't make a larger-scale infant mode for Ysabeau with tailed /d/l/u/ to go with that a tailed single-storey /a/... currently I only have a tail in infant /l/.)
    But in polytonic Greek, alternate /α is a daunting prospect!
    How so? Once you've set the anchors, isn't it just a matter of copying, mass renaming and mass regenerating?
  • Nick ShinnNick Shinn Posts: 2,131
    It’s still a lot of characters and glyphs, that one has to manufacture and inspect, and one may want to tweak the position of the marks.
  • Does the round beta need that strange gap between the two bowls, or can I use the shape of Bulgarian /ve/?
    Oh, and is the tailed /rho/ a cursive form, rather than just an alternative one?
    I just stumbled on this data on page 8 of this paper by Giannis Haralambous:

    https://www.researchgate.net/publication/2417759_From_Unicode_to_Typography_a_Case_Study_the_Greek_Script

    According to him, the tailed rho is more commonly viewed in the italic versions, but his work is viewed in the context of scholarly Greek typeface design.
  • Rafael CasesRafael Cases Posts: 38
    edited October 2020
    I seem to remember that lunate /epsilon/ is not recommended... and isn't pomega an archaic and obsolete form?
    With a Greek letterer encountered, I remember him discouraging the pomega, reserving it for calligraphy and cursive forms.  With my limited knowledge and some intuition, using / pomega for / pi is probably the equivalent to using the cursive lowercase / r in Latin text typefaces.
Sign In or Register to comment.