Nick's got the right idea. Be sure to play Beethoven during the testing. ;-)

Wow, what is that woodcut?

It's by Reynolds Stone.

More here:
http://www.typophile.com/node/31775

Screw it. Who needs to read anyhow.

(aek chew alley, I agree with you that aesthetics are an important component of legibility. They're also what gets left out of any legibility testing I'm aware of. It is absolutely crucial to any understanding of legibility, readin' ritin' and comprehension in general. Aesthetics is the part of design that gets dismissed immediately by those folks who want to figure out how reading works. After all, designers just want to make things look attractive and beautiful, right?, And this gets in the way of real scientific understanding. Right?. But really - what is, or, are aesthetics, other than the way we humans tend to like to have things organized so that we can comprehend them. A chart, or a dense passage of text that can be described as "Pleasing to the eye" is probably better, i.e. more readable and comprehensible than one that could be described as "A dog's breakfast".

But, how do you test for the aesthetic component in any credible way without spending hundreds of thousands of dollars? In all seriousness, more like Millions of dollars. It's not a small problem. I do know how far a hundred grand will get you and that's really not much to write home about. Personally, I think that good designer or artists understand aesthetics in their bones. It's a non verbal kind of intelligence and unless the results of any such research are applied to further developing those skills for future designers, I really don't see the point. Scientifically design type is like scientifically through basket balls... You may know all the physics, but it doesn't mean you can through a ball.

-=®=-

Russell -

Interesting comments you make there, especially those about the cost of the aesthetics testing. Do you think it's possible though to look at these criteria on a case by case basis and then build almost a kind of process for creating reading materials that maximise the reader experience whilst retaining the highest possible level of legibility. Maybe it's more from a user testing perspective rather than a definitive 'these are aesthetics' perspective?

Aesthetics is the part of design that gets dismissed immediately by those folks who want to figure out how reading works

This is part of my frustrations also, and the idea is to look to see if you can keep a partial focus on the aesthetics, so when it comes to layout and design of the material as opposed to just the 'science' of the text, you can ensure that the legibility isn't compromised. It's all born out of a frustration with complying with standards and legibility, and how you can make them work together.

Sorry if I'm rambling, I've currently got the lurgy and am not thinking quite straight!

uh, oops. what I meant to type at the end was:

"Scientifically designing type is like scientifically throwing basket balls... Knowing all the physics doesn’t mean you can throw a ball. "

-=®=-

In the article Measuring the reading experience in Typo Magazine Issue 22, I wrote about my experiences in attempting to measure aesthetic aspects of typography. It’s known that people perform better on cognitive tasks after being induced into a good mood by receiving a gift or seeing something pleasing. I examined if it was possible to do this with typography and found improvements both on the candle task (a cognitive task) and a reduction in the amount of activity in the corrugator (frowning) muscle after reading with better typography.

It doesn’t cost millions of dollars to do this. In fact I did a live demonstration of this effect at ATypI in Helsinki with the remote associates task (a different cognitive task). After listening to me drone on for a while, I asked the audience to find the word in common with swiss, cottage, and cake (answer = cheese). Perhaps half of the audience found the correct answer. Then after listening to the highly entertaining Herr Spiekermann for a minute, I asked the audience to find a different remote associate, which about three quarters of the audience found. It didn’t cost a penny to run.

Cheers, Kevin

...after reading with better typography...

What were the two type specs?

Nick's image post likens to An Andalusian Dog.

ChrisL

What were the two type specs?

There are five studies with 3 different pairs of document styles discussed in the article. One pair is of good and poor page layout including good and poor image placement. Another pair either included or omitted OpenType layout features such as ligatures and small caps. The third pair had either reasonable h&j, or had severe problems with rivers. There are pictures of the document styles in the Typo article.

Kevin,
I have not seen studies showing benefits of layout since Ulm Hochscule fur Gestaltung journals of the late 1950s [written by Gui Bonsieppe]. Do you know of any more research in that area being done? There are loads of stuff on legibility but little on typographic layout.

ChrisL

There is a lot of work on typographic layout by Tinker & Patterson, Luckiesh & Moss, and others from the 20s to the 60s, but quite a bit less since then. I’m not familiar with everything that’s been done recently, but Mary Dyson at University of Reading has focused much work on typographic layout including line length.

There is also the unnecessarily maligned research by Song & Schwarz showing that your perception of the content itself is impacted by the typeface used. This is another interesting way to study the experience of reading a document.

http://typophile.com/node/46304

Thanks, Kevin!

ChrisL

re the cost... Granted, Kevin - All I know about the cost of research has to do with a single project, that was too way expensive and not very productive of useful information. Still, I can extrapolate from that, and say that if one got, say 10% of a useful result at one price, then a wild guess at the cost of 100% a useful result will be at least 10 time as much. :o)

I think that perhaps what your experiment at ATypI showed was that what most of us already intuit from personal and professional experience is valid and common to all of us and it's measurable. But where does one go from there? To prove that rocket prolusion actually works, all you need is a balloon, but building on that knowledge enough to get to the moon is not going to come cheap.

Not that anyone really needs to go to the moon. :o)

-=®=-

Kevin - Thanks some really insightful comments there, I shall have to look up your article, hopefully we have that journal at the uni where I work, as for the other authors you mentioned, are there any particular articles by them that you know of, or is a general search the best way forward?

As far as I am aware with Mary Dyson's research the focus is more on screen factors, but does cover a wide and very interesting range of subjects. I think it might be interesting to see how her results and findings translate to print?

"...the unnecessarily maligned research by Song & Schwarz..."
It would be fascinating to hear when Kevlar could condone necessary malignment in his field, if not here. A two-font, one-document, 20 or so 20-something subjects, dead-end studied to within one thin human hair thickness of anecdotal, and he loves it!

"...showing that your perception of the content itself is impacted by the typeface used..."
NO KIDDING!? Song & Schwarz doesn't do that though, preferring to let the results become muddied by badly conceived type selection and by badly implemented type composition.

"Arial is the best font for reading" (says Song). 'Nuff said? Maybe not. How does finding a remote associate via presentation, have anything whatsoever to do with reading?

There now, I've strained my left corrugator muscle.

Cheers!

There are ways to criticize every study, but sometimes it's more useful to talk about the positive aspects of projects.

The question being discussed is how to measure aesthetics. The typographic conditions in the Song & Schwarz may not be terribly useful, but along with the studies I wrote about in Typo, these are possible methodologies for measuring the reading experience that go beyond letter legibility and reading speed measurements. With these methodologies you could use your own typographic expertise to compare conditions that you find interesting. It's a choice between improving on the existing research or criticizing from the peanut gallery.

What can one say about scientists when they study an area of human endeavour and discover and "prove" what has been common knowledge for hundreds of years, to the point of being a banality? By way of apology for their abject tunnel vision, one might say that their methodology is correct, and that they should be encouraged to improve. Raise your game, indeed!

Scientists do science, which is a methodology that tends to limit the scope of individual experiments so as to ensure a solid basis for furture work, rather than overreaching itself. This narrow experimental focus may indeed look like 'tunnel vision', and it often guaranteed to make any individual experiment appear either insignificant or banal to an outside observer, especially one who may have extensive practical or anecdotal experience of the thing being studied. But the point of any individual experiment is not to answer a grand question, but to see whether the result contradicts or supports the results of other experiments. And occasionally the results will contradict the beliefs of the non-scientific practitioner, and that's when things get interesting for the rest of us.

This narrow experimental focus may indeed look like ’tunnel vision’,

I wasn't conflating a particular instance of tunnel vision (Song & Schwartz) with the scientific method in general.

However, I do think that Blake's Newton is relevant.
Designers have tried to reduce type design to mathematical principle for centuries, but it never works.

There is a quote of Paul Renner (he of all people ;-), but I don't know the correct translation in english: The believe in counting and measuring seduces to mistakes in all kind of arts.

For me science measures only what it sees in nature. Behavioral science only measures what it thinks it sees in human behavior. Humans often think and do things that are not visible or measurable. The products developed by humans can be analyzed by science after creation but not until then. Measuring a product after creation does not necessarily yield a path for science to create a better product of a similar kind. Patterned simplification can lead to measurable malfunction.

ChrisL

There are some really interesting things being said here. [as a designer] I can understand the frustration of trying to reduce design & typography to a set formula, it does in some ways take the creativity out of design, but could you also use it for the reverse effect?

I'm quite interested as it would appear that some of the views here are against using scientific research at all in the design process. Considering this, do you feel that the science could fit into the process, even if it is just to create a process? Does anyone feel that the research would be useful in qualifying design choices to clients, after all, I know how annoying it is when a clients only qualifying statement is "because I like it".

Do you think you could do a study into the emotional reaction to a typeface, a simple "what feeling does it conjure for you?" to see if there is a correlation between opinions to help inform typographic choices?

Please bear in mind that I'm in no way suggesting that scientific or subjective market research should take over from intuitive decisions on the part of the designer, after all, it is what the training is for, but how do you guys feel about this?

What Chris said.
Market research and product development can use scientific principles to test how well a design might work.
But how a creative product tests (which may be compared to polling) and how it performs when launched are two different things.
There's no such thing as a sure thing, and sequels are rarely as good as the original.
There's no substitute for 10,000 hours of experience.

Science can havr its place and its value. We must, however, guard against results being presented as scientific being placed on a pedestal and viewed as absolute truth because the name "Science" has been attached to it. Science once told us that Pluto was a Planet; now it says, not so. I don't pretend to know which conclusion is correct but it is clear that scientists can error as much as the rest of us. All scientific results need to be scrutinized and questioned. All research methods and procedures need to be examined and questioned without being held in awe. To me, the biggest issue is that the conclusions that are drawn from data are made by humans. The data may be correct but the conclusion may be wrong. If data tells me that regressions in reading are greater when a reader comes across a word in italic, that may indicate that italic is harder to read or it may also tell us that the reader is pausing to contemplate why that word is in italics or just triggers a memory like "I remember reading that book in college." Both may be true but how do we know? There is judgement in analyzing data just as there is in designing a glyph visually. Either can be misjudged. The difference is that we just say the designer made a human error while we are not so quick to be as critical of anything residing in the realm of science.

ChrisL

Science once told us that Plato was a Planet; now it says, not so.

Wasn't it Philosophy that told us that, not Science? ;-)

Whoops, that was supposed to be Pluto! I just corrected it but now my post moved down to the bottom :-/

Sorry!

ChrisL

Chris, the kind of test you are talking about is too all embracing to be realistic/useful as a scientific test. I think you can do useful science in support of typography but ideally it should be done in consultation with type designers so that it is not as naive as it has often been historically. But similarly, all this gnashing of teeth about scientists coming in to stamp all over our nice typographic grass is a bit silly as well and frankly just as naive.

Eben - I'm completely aware that the second half of the tests are very subjective and don't really constitute a scientific test, but the idea is that if you were to couple these with scientific tests for reading speed, comprehension, retention, etc. then you could use them to an advantage. I'm trying to find if there is a benefit in using subjective aesthetics testing to benefit the scientific part of the research, the 'case study' part being "can it benefit the reader experience"

I understand the need to include type designers, or at least designers who are aware of the essentials of type design. I am a graphic designer with a strong interest in typography, I am working on designing my own typefaces, and also work a lot with graffiti and handwriting (one of my interests is the crossover that these have with typography, where the lines blur, etc.) so I'm not too naive, but I don't profess to know all about the area.

I don't think this 'teeth gnashing' is all naive and silly, as it tells a lot about how the designers feel about the process. You get to the point where you can find out exactly why they feel like this and then you can find a way either around, or to turn this to the advantage and use it to aid the process. Surely that's of benefit, analysing problems to find solutions?

>I’m trying to find if there is a benefit in using subjective aesthetics testing to benefit the scientific part of the research, the ’case study’ part being “can it benefit the reader experience”<

This is excellent. But it is completely wrong. However, the fix is simple: simply invert it to 'the case study part being "do [good aesthetics] detract from the reader experience?"'

It stands to reason that good aesthetics are noticeable as such, and the time spent noticing them will be time wasted on admiring aesthetics when all time spent reading needs to be spent . . . reading.

The problem with type designers today is that they are obsessed with harmonious aesthetics to the detriment of readability. Graphic design is not, in many respects, a good foundation for readable type design. Because what the graphic designer finds beautiful, harmonious and attention-getting is a thrust that does not relate at all to what the reader finds easy to read.

I see graphic design as being primarily about a kind of macro experience and type design as being primarily a kind of micro experience. (That's a very bad way of expressing it but the best I can do at this moment -- enough maybe to get the germ of the idea across?) This was brought home to me when I had a glance at Paul Hunt's Reading project. The typeface seems to have tentacles that want to reach into the very fibres of the paper. Here, I thought, one could see a design axis, a design will, a design theory, a design practice, a praxis, that was relevant to reading.

It was brought home to me in a different way when I saw a large dual column edition of Shaw's essays set magnificently in metal Fournier. The pages all but read themselves.

The moral of that experience was one I have felt was paramount since I first tried to design type: nobody, nobody, nobody, is ever going to get reading until they start designing single-size fonts. A single master used at multiple sizes is always a fatally maimed 'monstrico'. Nor is there the slightest justification for single-master/multiple size fonts other than money, money, money.

You can ask: 'where is the money for us to design single-size fonts? We can't afford to spend ten years, rather than one, on a text design that only a few will buy.'

Maybe the answer lies in T.S. Eliot's prescription that the poet needs to have a job separate from poetry.

Certainly, if you look at what is happening today, with the designers who are designing single size type, that is what happens. Two examples: Robert Slimbach's new Garamond (which with four entirely separate optical multiple masters can be permitted, for this argument, to count as single-size) was designed entirely in his spare time over a period of many years. In this he is still following the direction of Sumner Stone who cannot possibly be making enough money from the great Cycles family to compensate him for the endless thousands of hours of development and sheer being that it cost over a period of more than a decade, and continues to cost.

So I guess my larger point is that there is no point in talking about readability until you have thought through some of the obstacles to readability which have been set in motion by the direction that typeface design has taken since, roughly, 1960.

"the kind of test you are talking about is too all embracing to be realistic/useful as a scientific test."...

Eben, of course it is! That is my point and it is not the scientists fault. The reason more tests are not done and that more useful results don't come out is that there are so many variables to control, which are so far not possible to control, as to make both the test too complex or the data too vague. Other issues are that there is very little financial support for this arena of work and the potential of monetary profit is small.

"all this gnashing of teeth about scientists coming in to stamp all over our nice typographic grass is a bit silly as well and frankly just as naive."

Who said that scientists are coming in and taking our turf? Not me. Scientists are not coming in and there is very little work done in the area of typography. My beef is not with the scientists--there are several in my family. My beef is with the rest of us who jump beyond the scientists conclusions and give the studies greater weight than even the scientists who made them. Scientists are not afraid to question each other but we are afraid to question results or to limit the scope of results to the level intended by the scientists. Scientists work incrementally. They take a small piece of the puzzle and analyze what bits they can and build upon the previous work done by others, Other scientists take different chunks either independently or cooperatively. After years, there is a body of work which others look at objectively and may prompt yet another study. This goes on for years. The scientists are not in awe of each other; they just work and read what others have published.

We all need to stop being so full of ourselves and just look at our role pragmatically, even naively. When we pull and push control points about in FontLab, we are experimenting, too. It may not be rigorous scientific method but we control some variable and change others. I have prattled long enough.

ChrisL

Legibility vs. Comfort
Reading Research: What is needed?
and maybe even
Display v. Text

Paul,
Ah, yes, back in the day when we had meaningful discussions here on Typophile :-)
Thanks for finding those threads!

ChrisL

Kevlar:"It’s a choice between improving on the existing research or criticizing from the peanut gallery."

...oh let's not forget the peanut gallery, the Other... who actually measure the humans of specific groups and outfit them with aesthetics to fit. Or, who you callin' peanut gallery, peanut gallery's peanut gallery? ;)

CA: "...it would appear that some of the views here are against using scientific research at all in the design process."

On the contrary, that is how we work: the client feels dark we lighten the type, the clients feel happy we put on the swash. Seriously though, our process of outfitting clients with fonts, a subset of the font selection process for sure, is a scientific progress from an unknown through a growing known, to an optimized legible aesthetic for some kind of foreseen reader experience. We do not do this arbitrarily, but rather progress from small to large glyph repertoires, one to more styles, and from general to specific composition, in search of specific complimentary and contrasting typographic aesthetics, to reach that foreseen reader experience.

I am against most reading research that does not involve reading. I have never been against good scientific research, which I define for one thing, as offering identification of the type(s) used, and the composition(s) and includes in the results, these specifications and examples of the typography as presented to the study subjects. I have also never been against good scientific research on scientific research on design legibility and aesthetics, as Kevlar talks a lot about. But as one can see, sometimes, some forget where the peanut gallery is, and where the use of scientific research is just another live wire in the design business.

Cheers!

Chris: We must, however, guard against results being presented as scientific being placed on a pedestal and viewed as absolute truth because the name “Science” has been attached to it. Science once told us that Pluto was a Planet; now it says, not so. I don’t pretend to know which conclusion is correct but it is clear that scientists can error as much as the rest of us.

The reclassification of Pluto isn't evidence of scientific error: it is evidence of science being done. Scientists are the least likely people to put scientific results on a pedestal, since they know that future science will probably demonstrate the inaccuracies in those results.

where is the money for us to design single-size fonts?

I've designed several, some as commissions, some as retail.
For instance, a text font for the Globe and Mail newspaper, and a revival of a 19th century book face.

Of course, people will use them at sizes other than for what they were originally created.
However, that is their prerogative.
The idea that a typeface will only work at a given size is mistaken, because it ignores the contribution of the typographer, and the effect that different layouts, media, and reading environments have on the effect of type and process of reading, not to mention the variability of reading skills and style, and content.

This is not to say that the idea of size-specific fonts is wrong, just that text fonts work within ranges, and the specification is as much the typographer's as the foundry's--and online, the reader's. Isn't this appropriately post-modern for the 21st century, multiple authorship?

John,
I quite agree! Scientists are NOT the problem, WE are. Perhaps you missed my other comment a few posts below the one you quoted: "Scientists are not afraid to question each other but we are afraid to question results or to limit the scope of results to the level intended by the scientists"
An example of what I mean is a recent quote taken out of context. There was a study done regarding screen fonts on web sites. The only fonts tested were bundled fonts in certain screen read situations. The quote taken out of context said something to the effect "therefore, Arial is the best font for reading" leaving out all the conditions of the study and making the reader think that Arial is the best font for reading anything anywhere anytime. The researcher did not say or intend this but never the less some non-scientist (I can't recall but I think he was an editor of some sort) jumped all over it and took it as god's gospel because it was "science". You can be sure that a true scientist would have been quick to point out his error.

ChrisL

“The reclassification of Pluto isn’t evidence of scientific error: it is evidence of science being done.” [John Hudson]

This obscures the fact that nothing new about Pluto was actually revealed (though it seemed at first that someting new about Plato was being proposed): the reclassification of Pluto had to do with a redefinition of planethood. Taking this into the realm of metaphor though, if Socrates was the sun, might Plato not be admitted to planethood after all?
See: http://www.space.com/scienceastronomy/060824_planet_definition.html

How my comment about reclassification on account of redefinition applies to the issue at hand is that, if legibility or readbility are operationalized in terms of eye movement patterns (saccadic amplitude and the absence of regressios), or ease of visual word form resolution (as quantified by the size of the word superiority effect), rather than simple reading speed, or the simple discrimination affordance of letter shapes under compromised conditions caused by distance or illumination, what counts as a legible or readible font might be expanded or contracted or reclassified.

Improving the reader’s experience can of course involve all of these things, but often includes much more, including introducing elements that enhance expressive specificity but have measurable — but non-fatal — perceptual processing costs. Some of the ‘much more’ might only be accessible to anecdotal report.

"including introducing elements that enhance expressive specificity but have measurable — but non-fatal — perceptual processing costs. "

Well said, Peter!

When the elements of a fonts design or positioning are purposely strayed from best practices of reading in order to enhance comprehension of meaning, your conditions are met.
An example might be the word "Help" in one instance set in white in Clearview on a plain green background; in the second instance the same word "Help" hand scribbled in red (as if traced in blood). In the second case, the legibility factors are negatively impacted but the communication is enhanced without another word being added.

ChrisL

I'll give you a good example of how typeface design impacts readability. Penguin puts out an excellent set of historical atlases in 9" x 7" softcover format, text by Colin McEvedy. The text was originally printed in Times Roman at about 8, perfectly legible and a pleasure to read. The layout encouraged comprehension. Later, they reprinted these books, same size, but with the typeface changed to one (I forget which right now and I'm too lazy to track it down) with a relatively higher x-height and at slightly larger size, reducing the white space. The design is less attractive and the text less legible, never mind that the point size is bigger. I wanted to sue.

...a good example of how typeface design impacts readability.

Well no it's not, because you go on to contradict yourself:

...the point size is bigger.

How is it possible to know whether it's typeface or type size that's affecting readability?
Or, for that matter, leading.

Pluto! Pluto? ;)

"Scientists are the least likely people to put scientific results on a pedestal, since they know that future science will probably demonstrate the inaccuracies in those results."

Another way of saying it is that 'scientists' who do put results on a pedestal, ready to defend it in all cases, are suspect as is their science.

"How is it possible to know whether it’s typeface or type size that’s affecting readability?"

Exactly, or other composition affecting the results that most studies seem utterly oblivious of. Most don't even bother to publish the compositions used in the studies. These things and more make any amount of maligning of Song & Schwarz, e.g. insufficient.

Cheers!

“How is it possible to know whether it’s typeface or type size that’s affecting readability?”

In my ICTVC 2004 presentation I said a tabulation of the perceptual processing impact of manipulating type-form microvariables is maddeningly elusive. I tried to suggest we look at specific determinants of ‘perceptual discrimination affordance’ in letters or specific determinants of the effective and automatic formation of ’word-level codes’ with words, like lateral interference or crowding, response-bias, cue value and computation costs as some of the areas that are effected.

It is not enough to change a variable, measure the effect along a single raw dimension like reading speed and then conclude ’readability’ is improved. This becomes blinding. We go around in circles. We stop trusting our cultivated vision and craft wisdom. We kid ourselves into thinking we have understood readability in this way. We've gained no understanding at all of what our type-forming micro-decisions actually do in perceptual-processing terms. We can sense and experience what occurs, but we haven't deepened our vision and craft wisdom with evidential knowledge about what occurs. Yet, intelligent design requires just that.

We need to know, for example how spacing affects the formation of word-level codes. How does too wide or overly tight spacing disrupt this? Why? What are the tolerences?

Or we need to know, does decreasing stroke contrast, or making it variable effect the cue-value of distinctive features in such a way that word-level codes are harder to assemble in the visual cortex.

Many sciences have acheived this level of specificity and particularity. Why not the science of reading?

This reminds me of the old show "What the Fed Chief Said?", before it didn't matter.

"...I said a tabulation of the perceptual processing impact of manipulating type-form microvariables is maddeningly elusive."
Learning how users employ their eyes and brains to read, is hard?

"I tried to suggest we look at specific determinants of ‘perceptual discrimination affordance’ in letters or specific determinants of the effective and automatic formation of ’word-level codes’ with words, like lateral interference or crowding, response-bias, cue value and computation costs as some of the areas that are effected."
If you break the rules of good typography, like over or under track, that's your best route to finding out what's gone wrong?

Cheers!

David, it is easy to take the piss, and no doubt tempting with such thick prose - but actually lateral interference or crowding, response-bias, cue value and computation costs are all specific and distinctive ideas which have emerged from psychological testing. I am about to study this crowding thing and with any luck the others because I am going to run a test to see if some CALT shapes I am going to design as replacements to the regular letters aid in speed and or accuracy when reading. The final shape of the test and the extensiveness of it are TBD. And as I said I need to come up to speed in the primary lit for this kind of testing. But I am hoping to find something nifty! If you would be willing to talk to me about it as it's getting designed that would sure be "pukka" as they say here in Old Blighty. Wadda ya say?

“If you break the rules of good typography, like over or under track [...]”

David, I think it's not so hard to see that over-tracking causes the word to fall apart as an integral object-like unit, and under-tracking leads to crowding, but wouldn't it be nice to know precisely why both these things disrupt reading, or how exactly having an integral object-like unit improves reading?

Explain to me in perceptual psychophysical or perceptual-mechanical — perceptual science — terms why it is important — if it is — that the cluster of letters form an optically integral or cohesive object-like unit. Why might it be important that the word as a visual unit resist easy deomposition into it's constitutent letters? Isn’t this the implicit belief that motivates type-design? Talk to me about the work that the visual cortex performs in reading that makes this desirable or mandatory.

Knowing this might give weight to our fetishistic attention to spacing.

and under-tracking leads to crowding, but wouldn’t it be nice to know precisely why both these things disrupt reading ... or how exactly having an integral object-like unit improves reading?

As far as I see, rather than describing too tight spacing, "crowding" actually results from good spacing -- "crowding" in the way Eben uses it is formation of "gestalt", or what you seem to call an "integral object-like unit" which, with the subject at hand, may safely be called "word". The latter part of the quote I don't understand though. What "how exactly"? Ideal reading conditions means words that are distinguishable and recognizable as such. What riddle is there that needs to be solved?
By the way, I am amused by the term "crowding", especially when confronting it with the older "gestalt". Both acknowledge the same phenomenon but approach it from opposite directions. It looks like the "crowding" is rooted in Anglo-American philosophic tradition which is obsessed with reducing everything to smallest particles as if these had some built-in evidence, but gets trapped in the un-answerable question, how come that we are able to get from these particles to a unity? From sensual impressions to conceptions, ideas, objects? (The "reference" problem.) "Crowding" nicely reflects this, alluding to a bunch of particles lumped together somehow. "Gestalt" instead starts out with concepts, ideas, objects -- the unity which "crowding" acknowledges, wants to explain, but does not really know how. From a "gestalt" perspective, perception is perception of objects, period, and particles are recognized only in as far as they form unity.
Off-topic.