>Do ligatures improve readability?

Kevin Larson, may correct me if I'm wrong, but I think that any improvement in reading performance between texts where the only difference is ligatures vs no ligatures would be far too small to measure using current techniques.

I know the team studied performance where all the typographic stops had been pulled out (ligs, small caps, osf etc.,) vs no typographic niceties and found the difference not measurable even then.

Cheers, Si

In my opinion ligs do improve a conscious or unconscious text access in reading performances, therefore an aesthetic pleasure.

A most interesting and challenging discussion of traditional rules and concepts: Will Harris, Ligature Schmigature. Best I have read. But I bet it's already been mentioned in the previous threads on the subject.

Simon, it sounds like I should contact Kevin Larson for the details. Correct? So in your view would you say that ligatures can be justified on a purly stylitic or historical basis? Or put another way they are basically there to provide color or to ground a typeface in it's historical context?

Alessandro, I have to admit I am leaning in your direction. I favor it. But what basis do you have for saying that?

I think a good f-ligature is one that is invisible when read quickly, and when looked at slowly doesn't look awkward.

The problem of a no ligatures in faces with a rightward extended f is that the clash results in an awkward-looking juxtuposition. The ligature is basically for aesthetics.

The aesthetic dimension is important, particularly for headings and display type, but I wouldn't expect much of an impact on reading.

I stop short when I see an un-ligatured fi. So yes, I think they improve legibility.

>Simon, it sounds like I should contact Kevin Larson for the details. Correct? So in your view would you say that ligatures can be justified on a purly stylitic or historical basis?

No not really, I think that ligatures improve readability but I think it's impossible to prove they do.

I think Kevin would say that when ligatures are used along with small caps, old style figues, correct spacing etc., that you can measure improved performance.

"An arrangement of lines and curves and angles may be beautiful in itself, but unless it suggests a form it is valueless. And the more clearly and definitely it suggests that form, the more we value it." - Talbot Baines Reid, Ars Typographica, Winter, 1920.

In classic serifed faces, the fi and fl combinations create indefinite forms in which the letters are almost-merged. Ligatures solve this problem by fully merging them.

That's a rationalization for something we understand not with mere intellect or gut feeling, but with typographic sensibility.

Eben, basis ?
I like the art of typography----is that an answer ?
For example, I have been lucky enough to touch this book :
http://mitpress.mit.edu/e-books/HP/hyp020.htm
And here is a photo of mine from Tallone's atelier, a book in the process, isn't it beautiful ?
http://www.as8.it/type/formeTallone.jpg

I guess the change in colour when a non-ligatured f and i collide is a good enough reason to suggest they assist reading, just as rivers or double word spaces hinder it.
Tim

> I stop short when I see an un-ligatured fi. So yes, I think they improve legibility.

As would be the case with most of us here. But that is because we are trained to look for them, just as typewriter quotes bother us, and slow down our reading. But to most people, these typographic niceties are not recognized, and would not cause them to "stop short."

I see it as a case, like so many other things, where good typography can give good color to a page, and that can improve reading. Will Harris makes some good points in his article (linked above). I will insist that good color does improve readibility, at least until someone shows me a study that refutes it.

Bad color ... the worst case being the use of bold words sprinkled within text ... means that the eye is attracted to the (bold) words, breaking the normal flow of the text. This has to decrease readability as the eye is jumping about unnecessarily.

Lining figures, fake small caps, full cap acronyms, and other bumps in the road to good color will all do the same thing, although not as badly as bold text.

The other difficulty is that some sets of ligs are done better than others. It might be hard to make a blanket statement about ligatures without measuring the difference between the good ones and the not-so-good ones.
The main job of the lig is to avoid the crash and the hole. To me, it makes sense that a crash and a hole can cause a brief moment to sort out slowing down the reading a tiny amount. Context solves most of it but I find when I am looking at text in a foreign language, the absence of ligs seems most jarring--I have no context to help me there.
The other issue is with typography that is readable enough as words and language but whose meaning becomes enhanced by typographic effects including ligatures. This is probably not the case for emersive reading but might hold true for the things we might dwell over reading a bit longer--covers, ads, menus, art books, display material...

ChrisL

Alessandro : I do see beauty there but plenty of beautiful things are not functional. Ideally the two should be married & if you listen to some beauty should precede or fuction should as they prefer. But my question is how did you come to decide this. That is what I mean by basis. You can still count me as sympathetic just skeptical of either point of view having been made in a some kind of solid way.

"The other difficulty is that some sets of ligs are done better than..." Chris that's a very good point and while it clouds the issue somewhat - that's just too frickin' bad. That's reality!

Actually the only really strong arguement I have heard so far against ligatures is that they can make spacing look odd when the basic letter spacing is screwed with. This is Will Harris' aregument. See the link above. But his point of view is overly practical in my view. He is advocating safer type. Not better type if you see what I mean. He is seeking the robust over the perfect and somewhat brittle. I see his point even if I am not completely taken with it.

Reading over everybody's posts the final thing that I am reminded of ( again one of Will's arguements or obsevations ) is that it is possible to simply design letters differently to avoid the problems that would have occured. An f can be made more narrow so collision is avoided.

http://www.will-harris.com/ligatures.htm

So I suppose I aught to reframe the question ( without expecting a solid answer per se ) is it better to redesign the forms or to use ligatures for immersive reading. I bet they wash out as more or less equal. And other issues such as 'is it a revival?' etc. must trump any kind of absolute decision you could make.

Of course now I am think about it I am starting to see other redesigned elements in typefaces which are not ligature related but are related to robustness in spacing such as the seemingly shrinking arm on the 'r' glyph. Anybody care to mention other examples?

Also, does anybody have any specific arguements' against Will's various points re: ligatures?

It sounds like a lot of people are saying the same thing: they're good if they're good.

I think ligatures should fix a problem, and they should disappear. I've seen lots of nice Linotype faces (and others) which are designed to not need ligatures. I think it's especially difficult for young type designers to come around to the idea that ligatures aren't always better -- because they are so fun to make and look at.

As mentioned above, if the ligature itself doesn't match the overall letterspacing, it's trouble.

Once again, I can cite some changes between the old Type 1 version of Utopia and the newer OpenType version to show how two different approaches to ligatures compare. (I always thought the old ligatures were a little obtrusive, despite my liking the design of them. I think especially the first f in the ffi "reaches" a little too much.)

I suspect that Utopia's popularity as a news face might have something to do with this. Given that most news text varies tracking quite a bit to assist justification on narrow columns, if the font has a dotless fi ligature, you will get a variety of ligating and non-ligating f_i combinations in text, with an overall appearance of "quantum dotting", which is probably not a good idea. Therefore the dotted ligature is to be preferred -- at least this was my reasoning for doing it that way in Worldwide.

Eben, and also Christopher for his “they’re good if they’re good”, I leave you your definition of beauty because I like things that are not funtional especially in the semiosphere. What are commonly referred to as ligatures are strictly aesthetic forms, most exist to avoid awkward collisions between characters (or text kerned glyph systems as tradition has it that ligatures should not be used in display type). They have no phonetic or syntactical roles (some editors dislike ‘&’)----they are strictly typographic constructs.

Regarding the point whether it is possible to “simply design letters differently to avoid the problems that would have occured” I would say that italics should be set only with the default designed spacing, as for the Roman, I expect old and young type designers to come around to the idea that there are new possiblities in forms but lettering is a precise art and strictly subject to tradition -- you can’t make letters any shape you like, you can’t make houses any shapes you like, unless you live all by yourself on a desert island.

Eben, I am glad I can count you as sympathetic and skeptical, I have my knowledge, my know-how and my physical being as basis, I'd like to quote Buckminster Fuller for you, “When I’m working on a problem, I never think about beauty. I think only how to solve the problem. But when I have finished, if the solution is not beautiful, I know it is wrong.”

> I stop short when I see an un-ligatured fi. So yes, I think they improve legibility.

I agree, but only for display sizes. When reading text settings (less than 13 pt), I think my eye is just too caught up in decoding words to usually notice a colliding f and i or l.

What REALLY stands out, though, is an fi-lig or fl-lig that is too cramped. How many times have we all seen a word with loose tracking applied but a "stands out like a sore thumb" auto-inserted fi-lig? Yeesh. That's the real peeve for me.

As a rule, I say "use 'em at display sizes, but only when the ligs a) avoid a bad collision and b) are close in kerning to the way the letters would appear without a lig.

--Michael.

------------------------------------------------------
// love what you do or do something else. //
Michael Ebert -- graphic designer, jazz saxophonist, horror movie devotee
------------------------------------------------------

... I’d like to quote Buckminster Fuller for you, “When I’m working on a problem, I never think about beauty. I think only how to solve the problem. But when I have finished, if the solution is not beautiful, I know it is wrong.”

I love that quotation. That's modernism right there, letting functionality -- a perfectly solved problem -- be beauty. Also, not bad advice for any designer.

Ah oh, Christopher used the "M" word. Duck! :-)

ChrisL

I for one don't have a problem with ligs (if they're good), and i agree one shouldn't consciously see them.
That goes for f-ligatures, mainly.
The revival of historical ligatures, tho, is a completely different thing. I find ct and st ligatures, as beautiful as they may be, if used in plain body copy very distracting. I think a typographer should have very good reason to use these – because of context, content, etc.–, otherwise leave them on the shelf.

I think the value of ligatured forms has to be measured in terms of neurological computation costs within the visual cortex (not reading speed) and judged in terms of the manipulation of saliency and the strategic assignment of cue value, not aesthetics first and foremost.

The brief moment of sorting out referred to by Chris is because the 'crash and hole' create disproportionate areas of saliency in the total figuaral word-object, upsetting the balance of cue-values of criterial recognitional features.

It is this imbalance that compromises readability. The aesthetics of it is another dimension.

How do you measure in 'neurological computation costs'? When you say 'manipulation of saliency' and 'cue value' I think you might be saying something like : '...manipulation of the ability of the brain to rapidly resolve, or recognize glyphs or ligatures. Am I with you?

How can a type designer measure the balance of cue values? Right now I think most (or perhaps all ) type designers use aesthetic criteria to judge that balance. Even if it might be ideal to do otherwise, is there another option today?

"type designers use aesthetic criteria to judge that balance."

That is only part of it. Granted, it is not a scientific measure used by type desighners but their trained eyes are capable of seeing disruptions in visual pattern that are caused by missfitted neighboring glyphs. They either fix the problem with adjustments to the outlines or with font metrics--fixed or kerning based. If this fails, ligatures may be a solution. I don't think these decisions are based so much on the aaesthetics of the type as they are on what the type designer feels is the function of the type. Anyone can debate if the type designer is successful in fixing problem crashes and holes (people like Kevin may be able to measure the effectiveness in some more concrete way) but I don't think the typical type designer would call his an aesthetic decision.

ChrisL

My basic point was that type designers should think of what they are doing when they manipulate the contrast in and connection, or role-relationship rules of the role architectural components of letters as adjusting cue-values and manipulating saliencies, that is, as making salient what has criterial cue value toward visual wordform resolution. Experienced type designers have an intuitive grasp of this, or so I think.

Chris's 'crash and hole' create areas of proportionally higher visual salience in areas of the word map, and this disproportion exacts a neural computation cost. The word 'measure' was a bit unfortunate. I meant it in the sense of gauge, or 'thought of in terms of'. I'm introducing a set of heuristics that have a perceptual processing signature--a framework for acting, a system of constructs convenient for gauging efforts and channelling action.

You don't need a nueralcomputationcostometer to practice mindfulness of computation costs / saliency / cue value in action. Though it might help to have empirical verification of what things have cue value. I think things like the overshoot of ascenders and descenders beyond the x-height and baseline; abscence of closure in the 'c', the 'n', the 'u' etc. are a start

I agree with Chris that avoiding 'crashes' with overhanging f's and other letters does benefit readability. I just think the benefit in a text type is tiny compared to many other more important factors. When read quickly at text sizes, I suspect that eg fi crashes would not usually be noticed and so not slow down a reader. In Peter's terms the crashes are not very 'salient' architectural features.

Thus the aesthetic dimension--you don't want the word to look like a mess when you look closely--becomes more the decisive factor.

Bill, we might not notice the crashes, much like we don't notice regressive eye movements. They don't break through to our overt awareness unless we are attuned to them, or scan the setting with the eyes of a practiced typographer. But I would suggest that they are registered in the early layers of the visual cortex as activation spikes at a level disproportionate to the cue value importance of information in the stimulus area where they occur. This makes necessary--or so I think--additional neural computational work in accomodating the activation provoked by the stimulus complex--with its rogue area of over-dense, distracting or clotting visual information--to the neural code for a known or familiar word.

The receptive fields of the visual cortex benefit from--does it's work better with--clean, well-managed saliency maps, with no areas where contour interactions become disruptive.

Bill, to follow up, essentially I'm averse to downplaying the perceptual processing benefit in favour of the aesthetic benefit. In the case of ligatures they are both operative in our engagement with the text. In isolated instances the costs might seem inconsequential. Over time and a large text the costs might be calculable. Especially if the type of problem ligatures try to solve is endemic to the font.

I would suspect that the perceptual processing costs are proportionate to the aesthetic costs.

When you gauge the extent of the gap in the bottom of the seriffed 'n' are you assessing it's functionality in perceptual processing terms or it's membership in a set of formally cohesive, mutually fitable forms? I would venture it is both, and that they do not need to be in tension.

>Over time and a large text the costs might be calculable.

You may be right Peter. I really don't know, of course. I am just guessing that this is not a big factor. Reflecting on this again, let me partly retract what I said. I was reacting to the fact that I and other people don't notice the fi crash in Times Roman in MS Word, where they don't have ligatures. There are just millions of documents like this. However, if fi ligatures were all over the place in a language, it would be important. But I wonder if then it might be better just to have no overhanging f.

On the gap at the bottom of a seriffed n. Funny thing, I was just working on that yesterday. I think it has a lot to do with having the 'n' look balanced and even in color within words. I think here the aesthetic and readability considerations largely coincide. The balance and even color give the neutral background against which only the salient features of the n stand out as different and are easily and instantly registered by the eye and brain, without any distraction. Of course, connecting the bottom of the n completely would also change the architecture, so that apprearance would need to be avoided.

"The balance and even color give the neutral background against which only the salient features of the n stand out as different and are easily and instantly registered by the eye and brain, without any distraction."

Elaborating, I might say: the even colour is the expression of a rhythmically coordinated field in which only the features criterial for visual wordform resolution are properly salient, and registered without distraction or interference. [ 'registered' could be substituted with: assert their proper cue-value ]

The strongly terminated f, shown centre and right, is better disambiguated from adjacent descenders, but at the cost of requiring ligatures.

Traditional serifed typography, which did not use extra leading, was and is dependent on type forms where the f ligatures are absolutely essential. One sees this notably in type associated with the Morison years at Monotype, as in his little book on type practice, which is set in a solidly leaded Bembo, if my memory is correct.

So, there are different kinds of ligatures. Those that are used in text to reapportion areas of saliency in the total figuaral word-object, so that the balance of cue-values for criterial recognitional features is maintained, (or they "look better" than the default glyphs), are aides to readability (fi, at least). Those ligatures that are used in text to embellish the total figuaral word-object, so that the balance of cue-values of criterial recognitional features is disturbed to the reader, should do so for some purpose, or risk doing so for nothing, or less.

To some extent, I think this is true about type beyond ligatures. If you have a chance to "read" today's NY Times OpEd piece on String Theory (that'd be the print version, I dunno if they publish type set on a spool in pdf...), you can see what I mean. ;)

Thanks, Nick, interesting insight.

The truncated f was done for Linotype, I believe. Monotype I believe had an 'edge' over Linotype in allowing kerning, so they probably made the most of it for that reason among others you point out. You could get ligatures on the Lino keyboard though it seems they were probably more trouble--perhaps someone knows about this.

Isn't the truncated f common now in sans faces? Why is that, given that there is no problem with kerning and ligatures in digital type?

You could get ligatures on the Lino keyboard

Also two-letter "logotypes" to "kern" character combinations such as "V_a".
Do we also have to debate whether kerning can be scientifically proven to improve readability?

Why is that, given that there is no problem with kerning and ligatures in digital type?

Firstly, it's easier to design without ligatures.
Also, extra leading is no more work than solid leading, and is the norm (note software defaults). So "fingers" are a not a problem.

> To some extent, I think this is true about type beyond ligatures.

Totally.

> Isn’t the truncated f common now in sans faces? Why is that

Tradition usually has a lot to do with that sort of thing. Plus Modernism wreaks its usual havoc, seeing as how it prefers sans to begin with. Also note that sans fonts tend to have larger x-heights, making a [relatively] large beak on the "f" fit fine.

> Do we also have to debate ...

You don't.

BTW, the notion that metal type was usually (or even often) set without leading is incorrect (except in newspapers). And if you've ever set metal type you know why: the lead keeps those little sorts from falling into the bottom line like rotten teeth! The lead physically holds the line together.

But Linotype, different story, since the type was the line itself (besides being limited in other ways compared to "foundry" -coldmetal- type). And Monotype, being mechanically tighter than human hands, I'm guessing could get by without leading as well.

hhp

Also: some designers (a while back) used to shift the tittles on the "i" and "j" to the right, to accomodate a well-endowed "f", in order not to force ligature use on the compositor.

hhp

With regards to metal type, I said "extra" leading.
You're right about newspapers. The Linotype, casting quite literally a line of type, made solid leading very practical. Although it's interesting to note that the Linotype manual of 1923 flaunts copious leading throughout.
For tight leading I was thinking of pieces like this Monotype promotion of 1924:

I've been 'lurking' in this thread and have learned a lot from so many erudite posts. One thing I've been wondering, though, is whether this discussion should first be approached from the other end, so to speak.

Before diving into the complexities of how ligatures and other lettershape features affect word image and, ultimately, readability, shouldn't one look into the expectations of people from different cultures and printing traditions?

To me, Patty's comment says it all:

I stop short when I see an un-ligatured fi. So yes, I think they improve legibility.

My reply would be, "Well, yes, but that's because you're a type nut, like the rest of us."

People who have lived their whole lives in places with little or no typographic tradition (like Brazil) are not at all unlikely to never have laid eyes on a 'standard' ligature. Couldn't it be that ligatures cause more of an imbalance to the word image that those people expect to see than so-called 'crashes'? After all, one can only process ligatures without any conscious effort if one is intimately familiar with them. In other words, if all one has seen their whole lives is 'unsightly' collisions between f's and i's, perhaps that's what they expect to see.

one can only process ligatures without any conscious effort if one is intimately familiar with them

I think this idea is a mistake. It's true that familiarity ( culture) has it's effects. Culture is not a minor aspect. It's a big deal. It is however, just one part of the puzzle. I ( and I suspect most of us here - correct me if I err) think that there is such a thing as better & worse design. I would be willing to bet that a well designed ligature will beat crashing & holes in what ever population you toss it into. And a poorly designed one will not. Assuming you are not stressing the font out with heavy spacing as Will harris ( see above) points out. Also, all the ligatures pairs and triples are different. an fi is not a ct. I would argue that the ct is more significantly an artifact of the chrographic origin of the lower case. It's more vestigal and has less obvious visual utility. Where as the fi ( when it is designed well) has quite alot of visual utility.

Well said Eben.

ChrisL

I (and I suspect most of us here - correct me if I err) think that there is such a thing as better & worse design.

I don't think you err at all; in fact, I probably have a lot more to learn from you than the other way around. I present my case here half hoping that someone will prove me wrong! :-)

I'm all for making choices based on what I believe to be 'better design'. I just don't know that I would claim such decisions to be right or wrong in a more absolute, universal sense -- not that I think you're doing that, mind you. In other words, I'm open to research that could prove my choices to be wrong.

But even if incontrovertible 'hard data' were to prove, say, that for a certain target audience lining figures make for better readability than text figures, I would still use the latter, especially in extended text. In fact, I think a lot of people do this consciously all the time: if it were proven once and for all that serif faces are more readable than sans-serif ones, or that faces with high contrast and vertical stress are more readable than low-contrast ones with an oblique axis (for a given target audience), Bodoni wouldn't all of a sudden become the most popular face among well-informed designers.

To sum up, I believe in better and worse design; I'm just not sure that better design always corresponds to better readability.

What team?

Re: To sum up, I believe in better and worse design; I’m just not sure that better design always corresponds to better legibility.

Your may be right. It isn't clear to me what you mean by better design.

So let's take that apart: I think a design can be said to be good if it does something very specific - very well. But it's really complex.

( BTW- you might want to search Typophile for threads about legibility and readability. They are different too. )

Even just speaking about legibility is complex. Legibility should probably be considered in the context of culture as you pointed out, on what paper, with or accomodating what feeling or style, and at what size, and.... many more things. With type there are many many many somethings that could be done well and some of the specifics are not compatible.

For instance: Type that works wonderfully well looking just at legibility at 72 point we know does not work well at 9. And vice versa. It would be silly to expect one design to do both.

I suspect when you say good design you might mean good style or maybe beauty. That's complex too. Style by definition has to coexist with other attributes. Some attributes of style like weight or serifs can impact legibility positively or not depending on how they are handeled. But you are certainly right to think that sometimes a design can be beautiful at the expense of legibility. Is that your point? If so, keep in mind that Beauty is not just one relevant measure in some cases - it may be the primary one depending on the purpose of the face or the design in which it is used.

This doesn't mean that there are no generalities that can be noticed and which support legibility. They exist, but in a very provisional or contextual way rather than absolute one.

But coming back to a pervious point of yours (if I understand it). I agree that people who design type don't have nearly enough hard data to work with yet. That may change. And to be sure some of what we do is probably neither helping nor hurting legibility. But that's not necessarily the problem that it may sound like; legibility being just one way to measure the success of a typeface.

But even in our relatively impovrished state (from a hard data point of view); there is quite alot that intuition, cultivated sensitivity, and awareness of context can offer in the persuit of legibilty. Pehaps you will be one those folks who will chew on the data! Or create the data. That would be great!

However I don't think that proof is what will emerge from testing or that 'proof' will help anyone, but rather I hope that testing will illumminate the way that characteristics of type can be traded off against each other to differing and hopefully increasingly deliberate effect. No matter what the test results may show - a prescription for Bodoni ( or any other face ) will never be 'the right one'. But I do think testing may eventualy help us to know more precisely what a type design is likely to be able to do. I put it that way because then there is the question of use. That's a whole additional area!

This is way too long. Sorry about that. I hope it is helpful and not a pain.

Cheers!

BTW- you might want to search Typophile for threads about legibility and readability. They are different too.

I am well aware of the difference. Here the distinction didn't seem important, since no contrast is being made between one and the other. In any event, I've edited my posts so that the word 'readability' is now used throughout.

I suspect when you say good design you might mean good style or maybe beauty. That’s complex too.

Hmm... Not really, no. I mean, designers usually strive for style and beauty, yes, but that shouldn't be their ultimate goal. After all, a designer has a problem to solve: finding the best possible way to communicate a certain message or messages.

But what is the best way to communicate a message? Doing it in such a way that the reader will be able to process the verbal message with maximum efficiency? Or finding the most effective solution to convey a certain mood? Or trying to strike a balance between one and the other? That all depends on the designer's goal, of course: sometimes it's mostly about being as clear and matter-of-fact as possible, sometimes it's mostly about creating a certain atmosphere, sometimes both are important.

But even in our relatively impovrished state (from a hard data point of view); there is quite alot that intuition, cultivated sensitivity, and awareness of context can offer in the persuit of legibilty. Pehaps you will be one those folks who will chew on the data! Or create the data. That would be great!

I don't believe cultivation and intuition will ever be entirely supported by hard data. My point is that even if research proves your cultivation and intuition to be 'wrong', I don't think you should give them up. 'Better design' will never be entirely objective, even if you do make a conscious effort to solve a problem rather than just 'make things pretty'. Besides, you can always try to educate the audience a little.

This doesn’t mean that there are no generalities that can be noticed and which support legibility. They exist, but in a very provisional or contextual way rather than absolute one.

I completely agree. That's why, in my opinion, looking for universally applicable generalizations can be a thankless job. Rather than ask whether ligatures improve legility/readability -- a 'universal' question -- perhaps it would be more to the point to ask whether that is true for, say, French baby boomers that read Le Monde.

But I do think testing may eventualy help us to know more precisely what a type design is likely to be able to do.

If you're talking about general guidelines, I agree. But not even those general guidelines will be universal; you'll always need to try your design on a focus group if you want to determine how well it works in comparison to something else.

In the end, though, no tradition is absolute and universal, but that doesn't mean you should abandon tradition. It's a little like religion: keep doing it, but never hold it to be a universal truth.

By the way, if anyone thinks all I've said here is a load of c*** or that I've disrupted the discussion that was going on without being knowledgeable enough about the subject, feel free to say so.

David Berlow gets the the shift in referential framework I'm trying to encourage.

A Modest Proposal: Let's reserve the terms legibility and readability for base-level experiential judgements about writing and texts (with both language use and typographical formatting being factors in the perception of readability), where they are more at home. Familiarity with the script, facility at visual wordform resolution, adequate discrimination affordances of role architectural particulars and proper saliency of criterial features are (co)operative within these base-level experiential judgements of the typographical component of the readability of texts. Let science isolate, denominate and focus in on them.

Than let's ask more interesting and testable questions like: do serifs, tight spacing and ligatures promote or inhibit visual wordform resolution / promote 'response bias (to the whole) collapse"?

In Erik's context, my hypothesis would be ligatures improve, in the nature of the case, the performance of texts set in types that require them because of the structure of the visual cortex and the mechanics of visual wordform resolution. Another example: to be resistant to response bias collapse, unseriffed text requires tighter control of spacing. So be aware of that when you use it. But don't avoid sans serif when it makes aesthetic and journalistic sense to use it. (There are other things to be aware of vis a vis unseriffed faces when used for extended texts.)

I don’t believe cultivation and intuition will ever be entirely supported by hard data. ( Erik )

Clearly that's true. It's also true that the vast majority of what we find useful in letter design today comes from that source- cultivation and intuition. So we should respect it. That respect needn't suggest a phobia about science or testing of course.

< cite> Perhaps it would be more to the point to ask whether that is true for, say, French baby boomers that read Le Monde. ( Erik )

I am less convinced of the importance of culture as a relvant factor than you are. I think other directly observable ( an measurable) factors in type are better places to look for specificity. That said I think regional testing would be good too. The thing is I don't have any confidence that anything like the grand ideas we have espoused for is likely to be pursued in the near future. I hope I am wrong. But we shall see.

reserve the terms legibility and readability ( Peter)

I don't think anyone can control these terms use except in their own speech and writing.

I would like you to expand on what you mean by:

' facility at visual wordform resolution' and 'adequate discrimination affordances of role architectural particulars ' and 'proper saliency of criterial features' and 'mutatis mutandi' and 'response bias collapse'

for my sake and others.

And by expand I mean I would like for you to give examples to illustrate what you mean. I can guess what you mean but I don't know that that would be too helpful.

Than let’s ask more interesting and testable questions like: do serifs, tight spacing and ligatures promote or inhibit visual wordform resolution ( Peter )

Now that's more like it! I even think we could get even more particular while we are at it if we want,and that that would probably be a good idea.

Than let’s ask more interesting and testable questions like: do serifs, tight spacing and ligatures promote or inhibit visual wordform resolution / promote ‘response bias (to the whole) collapse”?

Sounds good to me. I would actually be interested in collaborating on a research project like this. What methodology(ies) do you suggest should be used to investigate these questions?

Eric, I'm long on hypotheses but short on methodologies. One approach might be to see how the magnitude of the Word Superiority Effect varies with spacing, ligatures in fonts that need them, and style of stroke termination. Another might be to measure the responsiveness of the Visual Word Form Area under conditions of 'priming' with these variations. The best thing would be to develop a criterial test for what I refer to as 'letter-wise slot processing' verses the more direct and supercessionary pattern-based recognition style of visual word-form resolution that I claim emerges when reading becomes fluent.

Eben, 1) I misused and miss-spelled the latin phrase mutatis mutundis, thinking it meant something like 'in the nature of the case.' 2) Expanding on 'facility at visual wordform resolution,’ ‘adequate discrimination affordances of role architectural particulars,’ ‘proper saliency of criterial features’ and ‘response bias collapse’:

'facility at visual wordform resolution,’
I use this instead of 'word recognition' because the term 'word recognition' doesn't isolate and reference clearly enough the foundational perceptual processing moment' of the sense-following mechanisms in reading. It can mislead us into thinking that the interesting and important part of what we are dealing with in reading is mainly a cognitive letter-assembly mechanism.

In my scheme, facility at visual wordform resolution is a learned perceptual skill which develops of its own accord, without conscious effort because of the optical-grammatical structure of writing and, once solidified, proceeds automatically and under the radar of consciousness. It becomes the 'response bias' of the visual system when it is confronted with separated written or printed text. It begins to supercede the letter-wise routine we must master to become skilled readers. The visual binding forces of serifs and rhythmic spacing make breaking the system's learned bias toward a kind of wholistic processing an effort-intensive process neuro-mechanically. These visual binding forces broker or catalyze the perceptual learning that must occur. Bad spacing and the abscence of serifs may make breaking the wholistic respones bias less difficult. They may threaten it, or they may actually privilege a more letter-wise neuro-mechanical routine. When the more letter-wise neuro-mechanical routine is induced or privileged by the typographic realization of the text, I speak of 'wholistic response bias collapse".

‘adequate discrimination affordances of role architectural particulars,’
Visual wordform resolution requires that the visual system be able to 'make out'--and the visual cortex register and integrate--the shapes of visual components of the word at least up to the level of letter parts and counters. My word for letter parts is 'role architectural particulars'; my word for things like counters is 'role architecturally evoke form'. I use them because we need a set of generic terms for things like bowls and stems and shoulders and counters. Because the stems and counters 'afford' or lend themselves to being 'made out' by the visual sytem, I use 'discrimination affordance'. When the lighting is too dim or the contrast with the ground too low, or the type too small for role architectural components to be registered or connected with effectively or made out, the 'discrimination affordances' or 'discrimination affordance quotients' of the font's component letters are inadequate for the condition of use. The discrimination affordances of some fonts are greater at threshlds of size or illumiation or degradation than others.

‘proper saliency of criterial features’
Even if the discrimination affordances are optimal for a given condition of use, the type may not perform well, because certain parts of the visual wordform that are perhaps marginal to perceptually resolving it's identity are given too much visual weight by idiosyncracies in the shaping of the letters. The forms have to be balanced against eachother so that all the parts that are collectively important to resolving the perceptual identity of the stimulus have equal visual weight. In these contexts I speak of 'proper saliency' and 'cue value'. And it is important that salient holes and crashes not draw neural processing attention away from areas that are criterial for visual resolution of the identity of the stimulus. This mucks up processing from the retinal receptors up through the visual cortex. Salient here means salient for the visual system at a receptor and neurological level--not necessarily salient for conscious vision)

‘response bias collapse’
[see above]

Hope this helps!