Interesting. I'm glad to hear more exploration of the link between notan and readability.
>When the whites “talk to each other”, boumas are stronger.
Are you assuming that the blacks already talk to each other in chirographic/painterly/object-based approaches? Or are the black parts essentially unimportant?
I would try to get the concepts of black and white out of the discussion, except when talking about concrete things. What I mean is, a character is both black and white, so the black and white are together forming what we see, not the black on the white, or the white on the black, or the red on the off-white-yellowish parchment, or whatever. And because black and white are two ends on this hypothetical line, two extremes, they are actually in a circle, being the same end, but the other side, of a circle, if that makes sense, in that black and white are the same. Ask me more about this circular logic thing some day…
This is just a tangent, perhaps, but the notan discussion might not benefit from specifically discussing things that are the same but identified as opposites. This goes for all concepts that exhibit such characteristics.
— Rob Mientjes
robmientjes.nl / rbmntjs.nl
I was just going to say a similar thing, Rob. One thing that is both scary and beautiful is that with every node we nudge in an outline, we necessarily change both the black and the white; so even from a very practical design-process angle it doesn't really make any sense to think of them as separate things. The outline is not a barrier of Something against the Void – maybe more like a moving membrane between two equal fluids.
Thinking more about this unity of black&white, here's another thought (and this might have something to do with your question, Ben): The re-thinking of the whites as "equal partners" or the other side of the yin-yang or whatever you want to call it, rather than "the rest of the background, where there is no black", conversely also effects more "active" blacks! It makes both sides more active, by making them more intertwined; by making them interact. Look at Legato: Once the whites start to talk, there can be no picket-fence, silent blacks – right?
Hrant: Because letters have to be less themselves for boumas to be more themselves. When the whites “talk to each other”, boumas are stronger.
Maybe you should preface such statements with ‘I believe that...’ or even ‘It is my theory that...’, rather than presenting such ideas as if they are established fact. They are not. Indeed, other theories of how we read directly contradict these ideas (and not just the parallel letterwise recognition theory).
It seems to me that you are confusing categories, because letters are things perceived whereas boumas are perceptual constructs. There is no such thing as a bouma outside of perception in the same way that there are things that we call letters. You can't sit down and design a bouma. A bouma is a perception of multiple letters or parts of multiple letters as a recognised pattern above the individual letter recognition level. It makes absolutely no sense to me to suggest that perception of a pattern of letters or letter parts is somehow enhanced by making the constituent letters ‘less themselves’. That's like saying that face recognition would be easier or more efficient if the individual features -- eyes, nose, mouth etc. -- were somehow less themselves, when it is precisely the individuality of these features and their arrangement on the skull that establishes a recognisable pattern enabling identification.
"It makes absolutely no sense to me to suggest that perception of a pattern of letters or letter parts is somehow enhanced by making the constituent letters ‘less themselves’."
I read "making letters less themselves" as making them less "pure" with regard to their own nature/shape/direction/intent, in favor of adding/stressing uniting, "binding" features that wouldn't be needed for letterwise recognition, but help bouma formation. Like serifs. No?
>every node we nudge in an outline, we necessarily change both the black and the white;
Agreed. In practical terms we are of course adjusting both at once. But as Hrant suggested above, our 'intent' informs where we position the nodes. And if we are focusing on the 'foreground' (or 'marking the black') when drawing we aren't usually paying so much attention to the boundary itself.
I don't know how this all affects readability, and to what extent we need to be conscious of the letterforms when we read. But I love the way this discussion is turning into an exploration of nondualism!
>It makes absolutely no sense to me to suggest that perception of a pattern of letters or letter parts is somehow enhanced by making the constituent letters ‘less themselves’.
In Peter Enneson's theory, normal spacing is sub-optimal for individual letter recognition, because it causes 'crowding'. However, it is optimal for recognition of sub-letter features across the word, as these features hit just of or on a frequency grid. The letters are not "less themselves", but normally tight spacing makes them harder to identify as individuals.
Peter also agrees with Hrant--following Noordzij--that the whites in counters are important "features" that the brain reads.
Bill, I think spacing is a separate question. As you say, more tightly spaced letters are not ‘less themselves’: they are letters in a different spatial relationship. The frequency grid theory is one of the things I had in mind when I wrote that other theories of how we read contradict Hrant's assertion re. letters needing to be ‘less themselves’. Perhaps what Hrant actually meant is closer to what you are describing, i.e. letters in a spatial relationship which is not optimal for individual letter recognition but is better for multi-letter or feature pattern recognition. But the phrase ‘less themselves’ suggests to me something else, i.e. somehow undermining legibility at the letter design level in favour of readability, which makes no sense to me. I am perfectly willing to accept that the way in which we perceive letters during immersive reading is different from how we perceive letters in isolation, and that units of recognition in immersive reading may be other than individual letters. But I don't think it follows at all that this either requires or benefits from letters being other than what they are. Again, I think Hrant is confusing categories by treating perceptions of things (boumas) as if they were things perceived (letters, whether individually or as feature patterns on a spatial grid). A bouma is a perception of letters, so it doesn't make any sense to say that boumas benefit from letters being less than letters, and more than it makes sense to say that face recognition benefits from a nose being less than a nose. Indeed, it is difficult even to conceive what such ‘less than-ness’ might look like.
Peter also agrees with Hrant—following Noordzij—that the whites in counters are important “features” that the brain reads.
Well, as I've said before, when we see a bird against the sky, it is the shape of the bird not the shape of the sky that makes it recognisable. The figure-ground relationship is certainly important -- as was interestingly discussed on the previous thread regarding black-white edge effects in Helvetica --, but that is not the same thing as saying that the ground is figure, which seems the implication of talking about counters as features.
John, I have to admit that despite agreeing with much of what you have written just now (I have yet to find a any evidence for bouma based reading.) the bird against the sky analogy just doesn't wash with me.
I don't think the features used in recognizing a letter are "black" or "white". Features are recognized because of the presence of adequate ground and can't don't exist in apart from that ground. It may be that some features required for recognition are in a sense "white" like the opening in the bottom of the lc "e". Also a word space is a feature of a text an quite an important one. Moreover it isn't it the case that we simply tally up features and identify a letter. We look for those features to exist in an recognized order or structure which involves both figure and ground at once. To use the face analogy - neither the eyes nor the mouth can be in the middle. And this again means that it isn't the feature but it's presence in a "correct" context that allows letter recognition to occur. I don't think that asking "what is the figure and what is not" is the right question. I think asking "what are the features required for a glyph to be recognized, and how do they need to be composed to result in recognition?" is.
Maybe I have misunderstood what you are saying. Have I?
>A bouma is a perception of letters
No, the concept of a gestalt perception of words is that we recognize a word-wide pattern of letter features, ordered in space, without first identifying letters.
You seem to get the concept in your comment, but then this seems to indicate that you don't understand the concept of 'bouma'. Whether it is true is another matter, which I regard as an open question.
>bird against the sky
This example misses that identifying a letter in the middle of a word is significantly different from identifying a single shape against an even background. With all the crowding and with regular spatial frequency, something different is going on. In particular some kind of coding of sub-letter features is going on. Those on both sides of the divide--word recognition by single letter vs by word gestalt--agree that recognition of sub-letter features is involved. Whether for example the white counters in the o's in "foot" are a feature that is recognized is to me an open question. What makes you so confident that counters are not used as a basis for either letter or word recognition?
What makes you so confident that counters are not used as a basis for either letter or word recognition?
Well, truth be told, I'm not that confident. If you remember, my stance originated in response to claims about ‘designing the white’ and ‘the white being as important as the black’, which I think tend to be poses rather than describing actual design processes. My point was more about design than about reading: with very few notable exceptions, type designers design the black and the white happens as a byproduct. What we do with that byproduct during reading is, as you say, an open question. I believe variation in stroke weight and counter size -- from very light to very black types -- implies something in this regard, although I'm not sure what.
Re. boumas, I did not deny the role of sub-letter features; indeed, I made explicit reference to parts of letters in my initial post. My point is that parts of letters are, well, parts of letters and I find Hrant's notion of somehow making letters ‘less than themselves’ unconvincing.
"with very few notable exceptions, type designers design the black and the white happens as a byproduct."
I don't believe the above, John, what is your basis for thinking it is just a "few notable exceptions"?
I have no idea what the majority may do nor can I see how you would "know" either.
Personally, as a type designer, I think it is impossible for me to see the black without seeing the white. I am not talking about reading performance, only design.
I think John is right about what we do unless you take extraordinary steps to do otherwise. Which I think we do in a limited sense when we check our spacing/kerning. But by then our willingness to reconsider shapes may be limited. It might be nice to have tools that encouraged/supported a different process. And we may one day. But until that happens...
Also on a practical level I am not sure that optimizing notan and evenness of white space is nearly as important or primary as I once did. I think making sure the feature set is obvious enough, and arranged well enough to support rapid letter recognition is actually the 1st consideration (letter level legibility). Once you have that then it is worthwhile to think about the harmony of the features, spacing in words, between words, lines, and ultimately notan. I should also say that I don't mean to diminish the value of visual evenness, notan or what have you. Instead it I am commenting on the question of priority.
Nevertheless what i am saying does seem to be essentially in contradiction to Hrant's recently stated point regarding letters being 'less than themselves'. I think the letters have got to be extraordinarily themselves in order to support letter level legibility. Not in showy or noticeable sense but in a deep sense. And this deep sense is culturally mediated, perceptual/optical rather than being in any way essential/mystical. The reason for this is that letter recognition is primary.
Maybe I am misreading what Hrant means by this. It is a very squishy way of putting things.
Perhaps I should also explain where I get this idea from. It is a Pelli study: Pelli DG, Tillman KA (2007) Parts, Wholes, and Context in Reading: A Triple Dissociation. PLoS ONE 2(8): e680. doi:10.1371/
Also as Peter Enneson pointed out it is worth noting that Kevin Larson probably doesn't completely buy the whole word aspect of this study. ( Kevin?) But that is a separate issue.
Once again for Eben's benefit, I will repeat, "I am not talking about reading performance, only design."
On the question of "designing the whites", I believe that this is something that Frutiger has been fond of saying. I have shared some of John's skepticism about this. But I am noticing, in working on a sans, that in sans, the whites in the counters generally do seem to be more assertive than in a serif, particularly in how they resonate with other counters.
Am I mistaken in thinking that part of an education in type design is learning to see the white inside of and between letters?
Designing the whites, as you call it, is more apparent in a sans because sans don't so much follow a writing tradition as a drawing or construction tradition. The prevalence of monoline over contrasted sans further forces the issue and the bolder the weight the more the urgency. I find it totally impossible to design a bold sans without an equal partnership from black and white forms--to me, they are interdependent alter egos and inseparable anyway.
Traditional roman faces born of writing and calligraphy actually need the same uniformation but it is less visible due to the long ingrained penstroke tradition.
due to the long ingrained penstroke tradition
and due to the fact that shapes like those on the right read as more familiar shape Gestalts, I'd guess.
Peter: Am I mistaken in thinking that part of an education in type design is learning to see the white inside of and between letters?
Yes, you are mistaken. :)
Eben mentioned spacing and kerning, which is the context in which one might think the whites and their relationships are most critical. But of course this is typically a secondary process after the letters have been drawn, during which initial process the internal whites have been defined primarily relative to the black not relative to other whites. Usually, there is an early stage in the design when one is testing a small number of trial letters in word or pseudo-word settings, but then pretty much everyone seems to jump straight into ‘filling out the glyph set’ and only returns to word settings in any systematic way at the kerning stage. And what is happening during kerning? One way to think about it is certainly in terms of balancing the whites, but if that were the case -- or all of the case -- then autospacing would be a breeze. It isn't a breeze because letter shape intrudes on the relationship of the whites. So I'm inclined to think of spacing and kerning in terms of grid arrangement, in which the relationship of the whites is one part but again secondary to the relationship of letter features, which need to remain independently legible. This is why I am not wholly convinced by David Kindersley's assertion that spacing should override letter distinction, even if this means allowing letters to overlap, and I think his illustration of this principle uses very carefully selected letterforms to make the argument more convincing than it would otherwise be.
I'm afraid I don't know what you mean by see -- presumably the italics indicate something like ‘really see’ or ‘see in some way other than normal people see’. Sometimes, when designing letters I think to myself ‘That's a nice counter shape’, but only after I've determined that the black is doing its job. Everyone sees the whites, and without seeing the whites we couldn't see the blacks, but that doesn't mean that the whites are featural. It is interesting that the bubble test -- sorry I don't have the paper reference handy; I know we've talked about it in another thread -- led to the identification of letter features necessary for recognition and that all those features were parts of the figure -- the black letter --, not parts of the ground.
Chris, I'm willing to believe that the approach to whites in a low contrast, sans serif design might differ from that in serif text types. There are fewer other factors, and more direct interplay of figure/ground relationships and edge effects. In fifteen years of designing type for a living, I have yet to design a sans serif type from scratch -- I have made Cyrillic and Greek extensions to existing sans types --, so I'm not in a position to confirm this from personal experience.
>learning to see the white inside of and between letters?
Peter, from my experience there is definitely a "learning to see" experience, but it is hard to pin down exactly what one is learning to see. The "learning to see" is definitely a matter of learning how to evaluate letters in the presence of other letters, particularly in words. Basically it is a matter of what design of the letter best makes words "come together" as a gestalt.
The whites may be a big part of this, but the fundamental experience is one of sensing visually rightness or wrongness in the presence of other letters, and analysis is secondary. When I analyze what might be changed to make a letter better I personally try to look at everything: weight of black strokes at different angles, relationship of terminals, letter widths, curves, stroke contrast, and size, harmony and vibrancy of the whites. But when it works it works, and when it doesn't it doesn't. So the analysis is sort of feeling your way to a solution, but claims about what is most important are always a little suspect, because all the factors interact, and nobody really knows.
I do agree with Chris that the relationship of whites is also important in serif faces. I just noticed that its importance is more obvious for sans.
So basically my view is that Frutiger's "designing the whites" is misleading, but still he has a point.
"Peter: Am I mistaken in thinking that part of an education in type design is learning to see the white inside of and between letters?"
I was just about to post my one word answer to your post when John's post showed up :-)
Oh, my one word response was "Yes!"
By see I meant: subject to a kind of optical-grammatically gaugeing vision. My assumption is that this has to be done for the white and black equally, together.
I was reacting to the ‘white as by-product’ idea. I think Hrant and Noordzij are absolutely correct in trying to counter this idea. I thought it no longer existed amoung professional designers.
I thought contours are defined and manipulated; the black is gauged; contours are adjusted; the whites are guaged; contours are readjusted, and so on until a kind of resolution of countervailing pressures is acheived.
John and Chris, is that a Yes, with the inflection Regretably?
My "yes" meant you were correct but the double negative thing messed it up. Yes, you must learn to see and no, not regrettably but thankfully.
Peter, speaking from my still very limited experience as a fresh learner, I've found that I have indeed been learning to "switch" my vision around in the design process, i.e. trying to perceive the white as figure and the black as ground*; this has actually been very illuminating and has, I think, given me better results. Among other things, it has also led to a minor spacing epiphany.
(* Inverting the screen and zooming in on countershapes is a nice crutch to make this work better.)
However, I'm not sure type design education has directly helped me with this… I'm not sure generally how far one can be *taught* to see something – doesn't one inherently have to discover it oneself, actively, because one is looking to see it?
Right. Spacing is a good example of something difficult to get right until you actively learn to see. I'm not able to see it as directly as the black but I can certainly sense and fine tune its rhythm. (Though I don't pretend to be terribly accomplished of course!)
[I've notified an admin about closing the open <em> tag in John's 10 Nov 2.39pm post]
Edit: Looks like it's been fixed now. (Thanks, Jared.)
Hrant, would you suppose that both the white and the black have to have the gestalt quality of well-formedness? Just the white? Just the black? Neither, just the combination?
John, you wrote: It is interesting that the bubble test [...] led to the identification of letter features necessary for recognition and that all those features were parts of the figure — the black letter —, not parts of the ground.
I'm not sure that this is what the following classification image using the bubbles technique on the word ‘javel’ suggests:
Interestingly, but maybe counter intuitively, it's from and paper by Blais, Fiset, Arguin, Jolicouer and Gosselin entitled Skilled Readers Process words Letter by Letter in Nearly Optimal Sequence
Not that I've read all the cool contributions to
this thread (shame on me, yes) but I wanted to
add something related that I just discovered:http://eames.houseind.com/fonts/cover_numerals.html
Hrant, that's cool indeed – reminds me a bit of this:http://www.typophile.com/node/67628
So the big question remains @hrant: Are you going to create a font out of it?
And actually the Eames' numerals are astonishingly beautiful. Very very nice indeed.
Not from actual drawings, but from the *idea*, hopefully one day.
Detail from Tankard's "Footnote #17"* about Fenland**, page 2:
@hrant:Quickie trivia question: what (mainstream) writing system has glyphs
based (partly) on the shapes of the mouth/tongue/lips when saying the
sound in question?
Korean hangul, of course, although the non-mainstream system of "Visible Speech" is likely known to more people in North America as having this attribute.
I should have noticed this before:http://typemedia2011.com/kunihiko
To quote what used to be a segment on Letterman: Is this anything?
 It’s worth considering what the right-most hairline on the main stem of the cap R does to the notanic consistency of the word.
Wow, the examples in the two last posts are really nice ...inspirational. I kinda have been on that path before...
...but not really.
What Okano did with Quintet is masterful. Every character is encompassed with just one line. It acquires a very unique rhythm.
The article ILT dedicated to Quintet can also be of interest: http://ilovetypography.com/2012/01/13/interview-font-designer-kunihiko-okano/
All one has to do is go back centuries to look at the construction of Versals. Nothing new. And this form of calligraphic drawing has been going on for eons. Once again nothing new!
Peter: Nice. And what about Raffia Initials? What I wonder though is whether those are a different animal, since they don't seem to be trying to make surfaces; to me it looks like laying down multiple pieces to create the impression of a letter. But maybe they're all on the same continuum.
BTW concerning Quintet, I wonder if Okano realizes how fundamentally different it is to the rest of KABK's output. Looking at the images in the ILT interview I would say that going from that double-pencil "B"* to the shapes in the actual font is a qualitatively huge leap. It's like walking versus flying.
Michael, can you provide examples? Of course it's not impossible that drawing the outlines (as opposed to the bodies) of letters chirographically is old news, but lacking any examples I would have to think you're wrong.
Being the first claim to the invention of napkin-drawing on a surface, I do not think the insignificance of this should be overestimated.
Dates back to the 1600's Hrant. Because of their shape they had to be "drawn" (outline) with a fine cut quill and filled in. The shapes we unachievable with a wide pen, although they were a nod to the Trajan Roman incorporating the nuances of the Uncial.
They were used to "start a verse."
There is a contemporary adaptation in the Chesapeake Corporation logo on my site, typerror.com, if you flip through the lettering section. They are quite popular today amongst practitioners.
Michael, unless you can provide an actual image few people can decide if you're right.
Of the examples I've seen the oldest is Excoffon's (I think). But frankly to me this isn't about who was first; I'd just like to grasp the implications.