Is that what all the hullabaloo of your Great Readability Scandal was all about — Luckiesh & Moss?

Well, why didn’t you *say* so?

Somewhere around here I have a transcript of the original study L&M did for Mergenthaler Linotype Co., ca. 1938. There was a brief dossier included in C.H. Griffith’s papers at the U of KY, Lexington, and I have a copy in my CHG-WAD research/collection.

It includes some comments and correspondence between he and Dwiggins (to whom CHG sent the study for review). As I recall, both men were a bit underwhelmed by the report and the conclusions.

In fact, I seem to recall that their discussion of the matter echoed many of the same points and perspectives that y’all have gone ’round and ’round about here, ad nauseum — except, 70+ years ago.

Plus ça change, plus c’est la même chose. At some point, I’ll have to dig out that file and find some juicy quotes for you.

Kent, I don't know if you are kidding, but Luckiesh and Moss are a big deal scientifically, though not as far as lessons to typographers, because they were only a start. But it was promising start, unlike Tinker, who successfully killed any further reference to Luckiesh, and whose own work I think put a dead hand on further research by reducing all issues to reading speed.

Many factors are likely involved in good visual experience for the reader, and they may at times be in tension with one another. Luckiesh studied three factors: visibility—a threshold contrast measure, reading speed, and fatigue as measured by blink rate.

If you actually read some of Luckiesh and Moss, I think you will see how interesting they are. Frankly it is the first work on readability or legibility that I actually found interesting. Most of the results before and since are equivocal or uninteresting.

Here is one of his more interesting results, in which Luckiesh and Moss compare four weights of the slab serif Memphis:

As you see, reading rate between the medium and bold is pretty much the same. Tinker concludes from similar results that it doesn't matter whether you use medium or bold—instead of thinking that his measure of reading speed might be inadequate. This is the kind of dismissal of typographic craft that has led to the lack of esteem for readability research, and I think that this negative view by people with typographic expertise has been largely justified.

Luckiesh, though, finds that readability of the bold, as measured by blink rate, declines dramatically compared to medium, though it is still above the light weight. The conclusion is that there is some kind of ideal weight range for "normal" text type for extended reading, something which Adrian Frutiger has argued also.

Now this recommendation follows current practice, so it does not result in any change. But it shows that Luckiesh was on to something. He also finds that Bold is more visible than Medium, which would explain why bold is often preferred for titles. It attracts the eye by being the most visible thing on the page even though it is not the most comfortable to read in extended text.

Here you have different factors that reflect the reality of the kind of trade-offs you face in designing a page, and I haven't seen that in any other scientist.

I would be very interested to find out the interaction with Luckiesh and Linotype. They were at least happy enough for Harry L. Gage, vice-president of Linotype, to write the preface to Reading as a Visual Task. Did Linotype provide any funding? Did they ask questions for Luckiesh and Moss to investigate? I know they provided typeset samples. Luckiesh also claimed that his work influenced Newspaper design, so I'd be interested if there is any correspondence on that.

One of the other of Luckiesh's more interesting results was a relatively short line, a bit over one lc alphabet, was more readable than a two alphabet length. This was only for 10/12 Linotype Textype, so he said that a lot more research needed to be done. Still it raises interesting questions.

I want to emphasize that in my view this is not an example of plus ça change. On the contrary, Luckiesh made an advance. No one ever discussed readability specifically as a matter of fatigue before Luckiesh, and afterward it entered the lexicon of typographers in this sense of ease of reading—low effort and low fatigue. That is only one indication of what in fact did change, and did not stay the same. The going round and round is because this important advance was buried by academic politics and misguided ideas about methodology. Luckiesh is in my view a key to stop going round and round, not more of the same.

…the notion of readability as ease of reading was introduced by Matthew Luckiesh…

Wasn't it first introduced by educational psychologists measuring reading ability as a purely textual phenomenon, i.e. correlating length of words and sentences in text to reading ability, indexed in terms of age? By so doing they could recommend the appropriate books for different aged children by analyzing sentence and word length, irrespective of typography.

This meaning of readability is still widely understood today.

Typographers and graphic designers understand readability to mean something else again, i.e. a well laid-out document that is not only easy to read at the level of "immersive text", but also easy to understand and engaging in terms of typographic hierarchy. A magazine page with many different elements, including pictures and diagrams, is thus readable.

That meaning is supported by this:
http://www.readability.com/

However, Luckiesh's eye-blink tests (or whatever) would probably not contradict what this app is doing.

Nick, in 1939 Luckiesh and Moss published their paper “The Visibility and Readability of Printed Matter.” Journal of Applied Psychology Volume 23, Number 6. It is possible that this is the report Griffith and Dwiggins reviewed.

Here is how Luckiesh and Moss start their section on readability: “Theoretical Considerations. The term "readability" is used herein to express the integral effect of physical factors which influence ease of reading. As such, it includes a consideration of all factors which influence visibility as well as certain other factors whose influence is not appraisable by so-called visibility measurements. Among the latter are typographical factors such as length of line, leading, margins and others. It will be noted, however, that the concept of readability is not extended to include a consideration of the comprehensibility of the reading matter. In general, a distinction is made between visibility and readability since the criteria for appraising these characteristics do not necessarily involve and measure all possible factors. For example, the visibility of a given line of print is not appreciably altered by the proximity of adjacent lines of print in practical typography while it is a matter of common experience that the ease of reading is markedly influenced by the space (leading) between successive lines. Nevertheless, in many cases visibility and readability are closely related. It will be noted that we do not use the term "legibility" which has been widely used in connection with reading matter, generally without clear definition and perhaps never with precise definition. Having applied the term, visibility, rather generally to other visual objects and tasks it was logical to apply it to reading. We then adopted the term, readability, as applying to ease of reading, which includes effects of reading upon the reader.”

First of all let me thank Peter Enneson, as I should have in my initial post, for his great help in working on the history of this debate. Once I ran across it, I enlisted his aid on the history, and he has been a fabulous "Sherlock Holmes" on the history and bibliography, as well as valuable sounding board on the issues. I was delighted by the warm reception of my talk at TypeCon, and it owed a lot to Peter.

Nick, there are (at least) two different pretty clear concepts of readability. One was used by Rudolf Flesch in his long ago best seller The Art of Readable Writing. That is the first meaning you refer to, and it may be that Flesch solidified that meaning, which refers to rhetoric, rather than visual appearance.

The meaning in terms of visual qualities of the text that promote ease of reading—low effort or fatigue—was introduced by Luckiesh.

Graphic designers so far as I can see did not use the term in this specific sense until after Luckiesh introduced it in his research. I haven't seen it used in the specific sense of ease of reading (as distinguished from speed and comprehension) before WWII, i.e. before Luckiesh's work.

Graphic designers thus seem to have followed Luckiesh's lead. Works that explicitly define readability in Luckiesh's sense, in addition to my Uncle Ben Lieberman's Types of Typefaces are Tracy's Letters of Credit, Felici's Complete Manual of Typography, and Mitchell & Wightman's Book Typography. Both my Uncle and Tracy had direct connection to Linotype.

The distinction between legibility and readability is a more complicated story. "Illegible" in ordinary English has a clear and specific meaning of the letters and words not being visually identifiable, so that the would-be reader can't decode the meaning of what is written. But the positive term "legible" is not so precise. "Readable" is a broader term that includes layout, rhetoric, etc. So while in ordinary English there are some differences, when it comes specifically to visual qualities of text, there is no clear distinction in meaning between the two terms.

Tinker and others in the US used "legibility" as a technical term to mean reading speed exclusively. Tinker at some stage of his career also used "readability" interchangeably with "legibility". However he refused to recognize the reality of and argued against the reality of any measure of reading comfort other than reading speed. And later in his career he used only "legibility".

Tinker's practice of only recognizing "legibility" and identifying it with reading speed has been generally followed in the scientific community, where "readability" was not recognized as distinctive, once Tinker had done his attack job, particularly in his widely referenced 1963 book Legibility of Print.

The 1947 booklet of Linotype 'Researches in Readability', which touts their collaboration with Luckiesh, gives his definition of readability, similar to what Peter quotes above, and then defines "legibility" as "quickness of perception", which they say is important in "heading or display types." They also identify legibility with Luckiesh's "visibility", which unfortunately muddles the matter.

By the way, Kent, the 1947 booklet of Linotype is testimony to the fact that someone at Linotype was still a big fan of Luckiesh, (Gage?) though it may not have been Chauncy Griffith, or Dwiggins.

Luckiesh himself defined Visibility, as measured by the Luckiesh-Moss Visibility Meter, which they invented, and Readability, as measured primarily by blink rate. Luckiesh said that the meaning of "legibility" wasn't clear, though often identified with reading speed, and he didn't want to use the term at all.

Lieberman didn't use Linotype's definition, but rather defined "legibility" as "ease of distinguishing between letters". Tracy follows Lieberman on "legibility", but Felici is closer to what Linotype said. Mitchell & Wightman (2005) is back to Lieberman's definition.

Gordon Legge, an outstanding reading scientist today and Tinker's current successor at Minnesota criticizes Tinker for a "behaviorist" identification of readability and reading speed.

To me the important issues are NOT terminological. Reading speed is an important consideration in understanding reading, as is reading fatigue. What terms you use is not so important as being clear. Much of the debate on Typophile I think has been muddled by our not having the element of clarity that Luckiesh brought to these issues.

Now we may distinguish how quickly words can be read in flashing on the screen from the contrast threshold (visibility) as well as reading speed, fatigue and other qualities. The challenge is to understand what is going on when we read, and not to decide the "essential" meaning of one term or the other.

On a personal note, I was gratified to learn at TypeCon that Ronald Arnhold, designer of Legacy, said he still values my Uncle's book highly, and Ed Benguiat said he knew Lieberman very well, and in fact lost out to him in buying the Kelmscott press of William Morris, which was also at one time owned by Goudy.

I've always been impressed by your Uncle's book—one of the few to give full credit to 19th century type designers, avoiding the modernist myth.

Bill — My plus ça change comment was directed at the kind of polarities — the touting, on one hand, and the pooh-poohing, on the other — that I read in many of the exchanges on the subject here and which are reflected in CHG/WAD’s lukewarm review of L&M’s initial study for Linotype 70+ years ago. Not the results of the study, per se. I have no particular comment on the nature or validity of Luckiesh’s work, in general or vis-a-vis his contemporaries and successors.

I have no doubt that some at Mergenthaler found the partnership with Luckiesh and Moss to be worthwhile (note the reference to “executives” in what I quote from CHG below — of which Gage was essentially one). And I’m not implying that it wasn’t valuable. It may likely have been a great advancement in this area of research.

I’m just reporting, historically, that CHG and WAD (the men in the trenches) were underwhelmed by the initial study.

I don’t have time or inclination to present the totality of this dossier right now. I’ll be happy to mail photocopies to you and Peter.

However, for the sake of context in this conversation, here is [most of] the summary that CHG provides in his intro to the dossier:

Early in 1938 executives of the Linotype Company became interested in the work of Dr. Matthew Luckiesh, director of the Nela Park (Cleveland) Laboratory of the General Electric Company, in the field of ophthalmic research, illumination, and in general the science of seeing. [. . . ]

Dr. Luckiesh and his associate, Frank K. Moss, were commissioned to make a series of studies of the anatomy of letters, particularly in relation to what constitutes the optimum of legibility in their configuration, degree of boldness, and thus contributing to the ease and speed of reading. It was also expected that in their researches they might be able to come up with something tangible in the way [of] proven formulae which would bear on the Dwiggins theory of a numerical process for “letter fitting” as an aid for the designer of types.

There is presented here the Report of Luckiesh and Moss to the Company on the result of nearly two years of research on the subject assigned to them.

As a whole, the result was disappointing. This was due, perhaps, to their unfamiliarity with printing, the customs of printers, the esthetic considerations which play an important role in the practice of typography, and lack of understanding of the basic principles of type designing which contribute to the ease and pleasure of reading. Their Conclusions (page 11) were weighted with theories so abstract and impractical in their broad application to the subject under consideration as to become of little value in our work, and the research project was discontinued.

That last paragraph will undoubtedly have a familiar ring ;-)

I think what CHG and WAD reacted to the most was the over-reaching nature of the conclusions that L&M presented as the result of this one avenue of research.

BTW, the study was entitled “Boldness as a Factor in Type-Design and Typography” and studied the results for four weights of Memphis. (The graph you show above was likely derived from this same study, although the typescript I have provides only tables, no graphs.)

In their introduction, L&M state “The specific objective of the present investigation is the determination of the optimum degree of boldness for the Memphis family of 10-point types.”

But then, by the end, from L&M’s Summary and Conclusions:

3. The design of a type of optimum readability may now be guided with reliability and exactness by measurements of visibility within a range which is now fairly definitively known.

The nature of this conclusion should also sound familiar. An exaggerated paraphrase: “We have identified the best performance from a narrow range of options, and now we can tell type designers how they can create the definitive typeface for optimum readability.”

In his endnote to the report in this dossier, CHG comments:

NOTE — The whole of this report merely confirms, in technical jargon, what has been known and practised by type specialists for the past three centuries. A simple “discovery” of the obvious. No action was taken on the recommendations.

Again, the tone should sound familiar.

Following this is the correspondence from WAD. But that will have to wait for another morning.

Kent, haven't you heard—we are just on the threshold of discovering, in the very near future, for the first time, How Things Really Work.
Now where is that damn Higgs boson?

**

The cost of each Apollo flight, in 2011 dollars, was almost $10 billion.

Kent, I find the Griffith response surprising.

Here is what Harry L Gage had to say in his introduction to Reading as a Visual Task in 1942:

“[…] the research partnership of Luckiesh and Moss has made substantial progress toward the ideal alliance of scientist and technician in the graphic arts in the study of reading material. [my emphasis] It has been highly interesting and instructive to serve as their lay collaborator in a series of researches in readability which have developed some of the material for this book.

The industrialist’s impatience for mass action and quick results has been restrained by the demonstration that researches in this field must be thorough and deliberate; that first things must come first; that the awesome number of factors which influence readability require years of extended study before conclusive judgments and ratings of material can be attained.

Several years of work, growing out of three decades of researches in seeing, have been summarized by Dr. Luckiesh thus: “First we have had to develop devices, criteria and techniques which made it possible to study reading material and readability by means of measurements.” This book amplifies and verifies this statement. It substantiates the fundamental soundness of the methods which Luckiesh and Moss have adapted. It should become a basis for the further measurement of the many variants of types, printing materials, graphic techniques and seeing conditions which make the visual task of reading so complex and relatively unexplored.”

Also, the “[w]e have identified the best performance from a narrow range of options, and now we can tell type designers how they can create the definitive typeface for optimum readability” statement that you quote seems uncharacteristic of Luckiesh and Moss. A piece with the identical title “Boldness as a Factor in Type-Design and Typography” was published in the Journal of Applied Psychology in 1940 (Volume 24, Number 2, pages 170–183. It doesn’t contain that statement.

Thanks Kent, it will be interesting to learn the whole story here. As Peter says, in Reading as a Visual Task Luckiesh and Moss are a lot more modest in their claims.

It looks like there was some chest thumping on both sides, as well as some misunderstanding. It may well be that the early report only relied on "visibility" tests, using the Luckiesh-Moss Visibility Meter. Later, the more revealing results came from the blink rate tests. It may be that Luckiesh was chastened by the critical feedback and lifted his game also.

I don't think Luckiesh ever looked into letter spacing, which is what Dwiggins was interested in. Evidently, Gage was not interested in that and didn't ask Luckiesh. Corporate miscommunication. Of course that couln't happen today :)

I don't think the collaboration actually ended in 1938, as Gage refers to several years of work. Further, in the 1947 pamphlet of Linotype, 'Researches in Readability,' again praising the collaboration, it explicitly says that Luckiesh's work did inform their recommendations to newspapers on which types to use. Discussing the revisions of Ionic No. 5, Linotype writes "All of these design developments grew out of years of experience and observation. Research in this field had not arrived to offer further assistance. But current research studies have indicated a preferred selection among several such faces, each designed to meet just these [commercial printing] conditions."

So it looks like to me Luckiesh's work informed their recommendations on which of the Legibility Series was most suited to a newspaper. And I'm guessing this had to do with boldness and the presses. But I will be interested to learn more.

reading...

Bill, I thoroughly enjoyed your talk at TypeCon, and have been reading your article in the American Printing History journal. I think you've done a good bit of history, most particularly in discerning the processes by which Luckiesh and Moss had a lasting impact in the typographic community, while Tinker had a lasting impact in the psychological community.

I think, in light of Luckiesh and Moss, we do need to be very careful how we describe readability. The phrase that your Uncle Ben used is accurate, if one takes into account Luckiesh and Moss, but easily misleading otherwise. I think the word ‘ease’ is unfortunate, because it may suggest a cognitive ease independent of fatigue, rather than relative to fatigue, i.e. an inherent ease by which the same word or short sequence of words might be judged more ‘readable’ in two different typefaces. That would be wrong, I think. Readability is only a factor in extended reading, and as I understand Luckiesh and Moss, it is a measure of effort, not of result, i.e. not of how fast we read or how accurately, but of how easily we do so.

Luckiesh and Moss seem to have been careful to seek ways to isolate effort in their attempt to measure readability through blink rate. Response to their work has not always done so, e.g. CHG's comment, as cited by Kent:

Dr. Luckiesh and his associate, Frank K. Moss, were commissioned to make a series of studies of the anatomy of letters, particularly in relation to what constitutes the optimum of legibility in their configuration, degree of boldness, and thus contributing to the ease and speed of reading.

This association or parallel of ease and speed suggests a misunderstanding of Luckiesh and Moss and a tendency to confuse effort and result. I suspect we all have a tendency to do this, both because reading is experienced holistically and because effort is unconscious and in many cases may be insignificant between different reading circumstances or occasions (affected by many extra-typographic factors, not least of which is the subject matter of the text and the quality of the writing).

Readability is only a factor in extended reading…

That is, in the Luckiesh & Moss terminology.

There is an insurmountable gulf between this and the general understanding of what readability is amongst typographers and graphic designers.

According to the methodology of Luckiesh & Moss (blink rates etc.), it may be possible to determine the readability of a page—involving varying typographic parameters of line length and leading—but what those pressing the issue want (and I would include Bill here, with his interest in sans vs serif), is a measure of the readability of a particular genre of type, or even a typeface, so that they can rationalize their work to themselves, or more likely a client concerned about the reach of an advertisement or the accessibility of a web page.

That was the case, as I understand it, with the ClearType fonts, where the project was greenlighted by Bill Gates, an engineer, on the premise that there would be measurable improvements, and that is what the tests delivered—faster reading (as touted in Now read this).

Nick: There is an insurmountable gulf between this and the general understanding of what readability is amongst typographers and graphic designers.

What is that understanding? Is it really general, i.e. consistent and broadly accepted?

In my experience, the term readability is used very inconsistently by typographers and designers: sometimes as roughly interchangeable with legibility, sometimes associated with ease of reading (with ease seldom being defined), and sometimes in terms of word recognition as something distinct from letter recognition (which is begging the question).

Luckiesh & Moss have a fairly clear definition of readability that is both distinct from legibility and suggests methods of testing that isolate readability as a factor independent of reading and accuracy. Even if this use of the term is rejected in favour of some other terminology, the notion of effort as a measurable factor seems to me important and likely to be overlooked if we don't find a way to draw attention to it. As Bill pointed out in New Orleans, what matters isn't our terminology or our definitions, but what we study. So the questions to ask re. Luckiesh & Moss seem to be ‘Is reading fatigue a phenomenon?’ and ‘Is blink rate a reliable measure of reading fatigue?’ Whether we want to agree to call the measurement of ease vs. effort in reading fatigue a measure of ‘readability’ seems to me secondary to the study.

That was the case, as I understand it, with the ClearType fonts, where the project was greenlighted by Bill Gates, an engineer, on the premise that there would be measurable improvements, and that is what the tests delivered—faster reading

Not quite. The studies cited in Now read this involved ClearType rendering vs. other display techniques, not the CT Font collection specifically. Note also that all these tests were conducted by people who don't believe in readability, who are primarily interested in legibility as measured by speed and accuracy in letter and word recognition. The point of Bill's talk in New Orleans is that there is a potentially measurable* aspect of reading that is independent of speed and accuracy, and hence might contribute to an independent concept of readability as ease of reading. This opens up the possibility of particular typeface settings having higher or lower readability measures independent of legibility measures, e.g. of a setting that scores less well in speed tests but which readers are able to read for longer while experiencing less fatigue.
_____

* I say potentially measurable because one of the questions that must be asked of the work of Luckiesh and Moss must be whether blink rate is a measure of reading fatigue.

John, I agree that Linotype didn't fully understand the distinction that Luckiesh was trying to make. The confusion is also in their 1947 pamphlet, as I quoted above.

For Luckiesh, readability applies to shorter reading times as well, as long as we are above threshold of very short times. Luckiesh did many tests only five minutes in length, but said that ideally they would be an hour.

As to the terms "ease" "comfort" "fatigue" "effort": they all mean something a little different in ordinary English, and are also significantly ambiguous.

The clearest terminology that I've found is what I learned from David DiLaura, an illumination engineer and historian of lighting and illumination. He speaks of the distinction between measuring "performance" and "the psycho-physical costs of performance." For example, he explained to me, you may get just as good a work out of employees with bad chairs, but they will end up with aches and pains at the end of the day. The output is one thing, and the human cost of it is another.

There is a recent book Cognitive Fatigue, which lays out how complex the problem of studying fatigue is, and provides some new models. They discuss such issues as how we compensate for fatigue, and the distinction between time-on-task fatigue from work, and sleepiness.

It is important not to get hung up on terminology. Greater clarity evolves as new theories come forward, and the new technical terms with them. The most important thing, in my view, is that we realize that the reading experience is multi-dimensional, and is influenced by more than one thing. The experience of fatigue or effort is clearly different from a performance measure like reading speed and comprehension. And to me it is pretty clear from Luckiesh that the fatigue dimension is important to the reading experience. It is also important is not to assume that threshold measures, such as from flashing words or very low contrast (visibility) necessarily tell us what is going on supra-threshold.

Nick, if Luckiesh is right that blink rate can, with controls, be a kind of global measure of readability, then I think it is clear that the readability of a text will depend at the very least on spacing factors (line length, leading, letter spacing) as well as the particular type face. Still, we might find useful rules that go across wide categories of text and settings.

I haven't looked at the test of Microsoft's Cleartype, but I do think there's a great scope for testing on screen reading using blink rate.

Nick, I thought quoting the Gage “Foreword” to Reading as a Visual Task might preempt your sense that there is an “insurmountable gulf” between what Luckiesh and Moss are trying to pin down and what typographers and type designers are all about.

Having read Reading as a Visual Task it is clear to me that the intent is the opposite of prescriptive and reductive: “The various factors which determine the visibility and the readability of reading material are always encountered in a complex combination. Therefore, it is essential that each factor be understood and appraised as individually as possible." […] “It is obvious that some factors are major ones and others are minor ones. Furthermore, as none of the factors is ever unaccompanied by others, it is impossible to rank them in a categorical manner. Nevertheless, an attempt is made to separate the various factors into groups and to treat them systematically and consequentially, somewhat in an order of rank."

It seems to me that this concern with “integral effects” is laudable, and perfectly commensurate with the craft knowledge native to type designers and typographers.

I think the Luckiesh / Moss work is seminal on a number of levels. The treatment of visibility and super-threshold visibility in Reading as a Visual Task is a dense but rewarding tour de force that could spark new discussions about getting a handle on threshold and supra-threshold legibility.

There is an underdeveloped notion in Luckiesh and Moss of the efficiency of the cortical integrational processes within the visual cortex that could link up with work like Denis Pelli’s on the feature detection / combining processes underlying recognition. This could combine with an emerging sense in psychophysics of a connection between blinking and detriments to “feature discrimintion / binding / integration”

> Kent, I find the Griffith response surprising.

I am not surprised that there is a difference in response between “management” and “the factory floor” (which, although Griffith was technically a part of management, was clearly more where his sympathies and viewpoints originated).

I take Gage’s encomium with a grain of salt. Again, that is not to imply that there wasn’t something indeed valuable and worthwhile in the studies. But I suspect, in this case, that the value was much more for Sales & Marketing and Promotion than perhaps for Design & Production.

> It may well be that the early report only relied on "visibility" tests, using the Luckiesh-Moss Visibility Meter.

No, this report very clearly relies upon “eye blinks” and refers to the two metrics of Visibility via the meter and Readability via eye blinks.

> Thanks Kent, it will be interesting to learn the whole story here. As Peter says, in Reading as a Visual Task Luckiesh and Moss are a lot more modest in their claims.

> Having read Reading as a Visual Task it is clear to me that the intent is the opposite of prescriptive and reductive:

I’m reassured to hear that, a decade down the line, their stance had broadened and become more cautious and inclusive. There is not a lot of evidence of that in the initial report. And so we have the response that I quoted. Perhaps we can thank the critique of CHG and WAD with some small part in opening their eyes, so to speak, since what you quote is a reasonable paraphrase of Dwiggins’s primary criticism.

But just for the amusement of the community, here is the outline of the L&M report (from the primary headings):

Boldness as a Factor in Type-Design and Typography

General Considerations
The Measurement of Boldness
The Visibility of Memphis Types
The Readability of Memphis Types
Speed of Reading [which they dismiss as not a meaningful measure]
Summary and Conclusions

So, having exhaustively measured two factors — viz. Visibility and Readability — with respect to boldness of the four-weight family of Memphis, here then are their conclusions, verbatim:

Since our typographical studies have been based invariably upon quantitative appraisals of both visibility and readability, and since each investigation has been prosecuted to the extent necessary to obtain data of a high standard of statistical reliability, it is now possible to present certain definite conclusions pertaining to the utilitarian merit of various type-faces. These conclusions are:

1. A marked enhancement in the readability of the printed page can be obtained by augmenting the boldness of many types which are now being recommended or widely used for body texts.

2. An enhancement in readability is decisively less promising by means of alterations in the configuration of modern type-faces than by utilizing the optimum degree of boldness.

3. The design of a type of optimum readability may now be guided with reliability and exactness by measurements of visibility within a range which is now fairly definitely known.

It is emphasized again that all these conclusions are based upon the intrinsic visibility of the types studied and upon the facility with which they are actually read, and not upon introspective appraisals of the appearance of the printed page.

The design of a superior type -- a "Super Textype" seems to us to be an important future step in our cooperative research program. We are preparing a separate report on the immediate possibilities of further research and of development of an ideal type-face in the light of our current knowledge.

How they came to claim No. 2 is beyond me (and Dwiggins). And No. 3 seems to be quite a bold extrapolation (if you’ll pardon the pun). How they get from the study of one specific family (and size and setting) to “definite conclusions pertaining to the utilitarian merit of various type-faces” is the question that comes under attack.

Again, these are what appear to the very first steps in Luckiesh and Moss’s work in this area, and the very first reactions. I don’t mean to cast aspersions on the totality of their work. Clearly, their perspective evolved considerably by the time you get to the 1947 publication.

I just find this early snapshot interesting, amusing even.

Copies will be in the mail to you tomorrow. Hopefully you will find this unpublished material of value in filling out the historical picture.

Bill: [David DiLaura] speaks of the distinction between measuring "performance" and "the psycho-physical costs of performance."

That is an excellent way of putting it. I was thinking in terms of result and effort, but DiLaura's 'performance' and 'cost of performance' marks the relationship better.

With regard to the influence of Luckiesh and Moss on the development of the legibility/readability distinction in typographic circles:

By coincidence, the TypeCon silent auction items included a copy of the 1951 book Typography and Design, a textbook produced by the United States Government Printing Office for the teaching of intermediate typography skills. There is a chapter devoted to 'Readability', and Luckiesh and Moss' Reading as a Visual Task is cited in the bibliography for this chapter.

> a decade down the line / by the time you get to the 1947 publication<

Actually Reading as a Visual Task was published in 1942.

> here is the outline of the L&M report <

The version published in the Journal of Applied Psychology has the same headings.

> Speed of Reading [which they dismiss as not a meaningful measure] <

The version published in the Journal of Applied Psychology has “relatively insensitive and inadequate one for appraising readability.”

> their conclusions, verbatim <

The version published in the Journal of Applied Psychology has the same conclusions, except that the word ‘definite’ is dropped. The same conclusions appear in Reading as a Visual Task on page 177f.

> How they came to claim No. 2 is beyond me (and Dwiggins). <

Luckiesh and Moss’s partial basis for this is probably the study they did in 1937, published in the Journal of the Franklin Institute [223] entitled “The Visibility of Various Typefaces.” In Reading as a Visual Task the visibility measurements of modern faces are said to “serve as a first approximation of the more illusive [sic] and inclusive characteristic of readability,” and are supplemented by the somewhat more divergent blink rate results, but the “less promising by means of alterations in the configuration” conclusion remains the same where type is not “unduly stylized.”

Kent, this actually doesn't surprise me, and is generally consistent with Luckiesh and Moss, though rather less cautious in wording. Reading as a Visual Task is 1942; the booklet written maybe by Gage? is 1947, and is a tribute to Luckiesh (Moss, the junior partner, died young), with many photos of apparatus.

A few comments on the recommendations. By way of criticism of Luckiesh and Moss, I have argued that they underestimate the importance of reading speed. They don't in fact dismiss reading speed, but do argue that is is less sensitive than measures of fatigue, which is probably correct. They are, though, so proud of their discovery of the visibility and blink rate measures that they don't give proper weight to reading speed as a valuable quality in text. I think this over-doing their case was one cause of Tinker going berserk, and becoming obsessed for the rest of his life with attacking Luckiesh. Tinker had almost exclusively identified reading speed as "the" measure and even definition of legibility.

Indeed, if you look at what they did, rather than their rhetoric, Luckiesh and Moss were better on reading speed than Tinker. They measured reading speed two ways: one by "natural" unhurried reading, and another by flashing lines before people at increasing speed to see at what point they could no longer read them. Peter Enneson has just heard from Gordon Legge, who is in fact Tinker's successor at Minnesota, confirming that that the reading speed test which Tinker used is unreliable, as Peter and I suspected.

If Luckiesh and Moss wrongly downplayed the importance of reading speed, they also misplaced the importance of visibility. If they read the message of their own chart on boldness, above, they would see that visibility and readability can go opposite directions. Again, they recognized this, but because of inventing the Visibility Meter, I think they kept wanting to say that visibility was more closely correlated to readability than it actually was. There is another problem involving visibility, which is that the threshold measures they used cannot be extrapolated as they did, according to current illumination expert Mark Rea, who is so to speak the current "Luckiesh" in the illumination field.

All that being said, the #2, above, does have a good rationale based on their research, though maybe Luckiesh didn't lay out the full case to Linotype as clearly as he later did in the book. They measured a number of type faces, and found that lightness was a problem in readability. And the Memphis test revealed that boldness was a problem too. So this did indicate that there is likely some kind of medium weight that is best for readability of text faces.

This view that there is an ideal medium for text also has been argued more recently by Adrian Frutiger. His recent reworking of Frutiger and Univers involves partly correcting what was done, ironically by Linotype, earlier to the text weight of these faces. The importance of medium weight is also corroborated by the tests on Clearview Highway, where a stroke width to height ratio of 1:5.5 proved to be readable at a greater distance than the past Highway Gothic somewhat bolder weight of 1:5. This is a different kind of test than continuous reading and fatigue, but still it is an interesting agreement.

Also back when Robert Thorne's Bold Face was first introduced in the 19th century, Theodore Low de Vinne (in 1885!) criticized its readability because it didn't have enough white space within. The idea of ideal ratio of stroke to height already goes back to scribal practice, and it is also the general practice not to have bold for extended text.

Thus to me Luckiesh's sense that a medium weight is critically important to readability of text has a lot going for it, even beyond his tests. However, his understanding of the issue was in my view too crude. He did not take into account that the width of characters, letter spacing, and the way thicks and thins are handled, and even leading also affects the visual impression of boldness, and probably also readability. At least that's my hypothesis.

I think Luckiesh's bias for the importance of visibility also led him to put his foot down a little too hard on boldness, contrary to the actual result of his readability tests.

Thus I would say that his claim in #3 to "exactness" on the correct medium was a boast that he couldn't really back up as yet. It may well be that further research and more collaboration would have found out more. Evidently, Griffith and Diggins were unimpressed, and Luckiesh failed to sell Linotype on further research, if I get it right. But it looks like Gage and maybe others were impressed, and pushed Luckiesh, at least up till 1947. Did Gage leave then?

Thanks so much for the materials, which I look forward to!

By the way Peter was also rather skeptical of my great enthusiasm for Luckiesh, but when he actually read Luckiesh, he was as impressed as I was. As you can see, I am definitely critical of Luckiesh, but I also think he made an important breakthrough that has been unwisely and unfairly neglected.

> The version published in the Journal of Applied Psychology has “relatively insensitive and inadequate one for appraising readability.”

Yes, that is the exact wording. I over-generalized my editorial comment.

Sorry about the date mixup.

> I also think he made an important breakthrough that has been unwisely and unfairly neglected.

Very possibly. I leave that to you guys. Again, I am relatively neutral on the science itself.

I was not aware that he was neglected in this field, given that this Linotype connection was known to me, so I apologize for my overly allergic reaction to what I perceived as hyperbole in your TypeCon title.

————————

As promised, for those on the sidelines and in the cheap seats, here is the bulk of Dwiggins’s response to the 1939 L&M report.

The determination of the best weight for a given size is valuable for criticising an existing design (type face). If it can be worked in combination with determinations of optimums of other variables — if it can be driven in team — it will be a valuable help in creating new designs. I think the problem is: how to hitch it up with other “optimums” and make a team — that is what is needed to make the boldness findings valuable in the case of new designs.

[ . . . a brief outline of other variables WAD thinks could be studied . . . ]

All these factors interplay; and the investigator has to keep all the balls in the air at once, as I see it, because each variable influences all the others. I suppose the only way to attack the problem is to take the points one at a time and try them out on the guinea pigs separately, and then put them all together and work a big Act 5 grand finale test. *But*, until they have been tested all working together in a complex, I can’t see that the investigator is ready for his Conclusion No. 3 on page 11. For the same reason I can’t agree with No. 2. Optimum weight alone is not enough to go ahead with.

[restatement of Conclusion No. 2, already quoted above, followed by this footnote:]

(* My God why don’t they talk English? He means that there’s more hope of making types legible by giving them the right heft than by bothering about shapes of the characters.)

They may be right . . . I dunno . . . It certainly puts it up to me. What am I on the payroll for? Why am I fussing about curves, “action”, proportion, balance? . . . Their tests that turned up Memphis as the most legible face certainly prove that “the configurations of the individual characters are relatively unimportant”!

I can understand that the laboratory is trying to rule out all subjective influences, and that it has to get its findings in terms of numerical values — dial reading — muscle-twitches per minute. And maybe that limitation rules out about one-half the facts in the case. It is fair for them to wipe them out, according to the rules of the game. But it isn’t fair to Average D. Reader to force him to read a Super-Textype based on a formula that leaves out a half of the facts.

Nobody knows what goes on in the little grey cells when you read. The process varies from person to person. But it’s a safe bet that a great many people read by means of word-images. BOAT registers on the brain as the idea “boat”, not as a series of letters, b, o, a, t. Somewhere in that notion is the base line of my attack on the legibility problem — an effort to have the letters reach out to their neighbors on each side and join up to make a *word*. And I’ll swear that the configuration of the individual characters has a lot to do with *that* function — family relationships — similarities in anatomical structure — coordinating gestures, so to speak : motions that bind together.

[ . . . ]

I don’t want any of this to make it seem that I am blowing cold on the laboratory end of the game. I’m for getting all the facts *via* eye-blinks that a feller can get together. I have a great sympathy for muscle-twitches laid out on cross-section paper. I’m half scientist you know. (To be sure the artist part cancels out the science part in the estimation of the laboratory crowd, and in commenting on laboratory findings I am sunk before I can run up a flag).

It’s simply that I feel a little shaky about the L. & M. findings because I find the investigators so eager and willing to found a Super-Textype on “boldness” alone : “Now we’ve got a sure basis to work on !” They haven’t. They’ve built one corner of the foundation very nicely . . .

I wish they had stopped earlier in the Report, and just said: “Here’s *one* factor in making a readable type, and we can pin it down for you”. I’d have felt much better.

WAD

Priceless, Kent! Thanks for typing all that out! Dwiggins really did write “in the American voice!” I think it’s very fair, and the force of it not lost on the Luckiesh and Moss of 1942! Perhaps they were made aware of it.

Thanks, Kent! The comments of Dwiggins are completely fair, and if I might say so are in line with I said in my above critique of Luckiesh, namely that his analysis has a core of valuable insight but is too crude. It seems also that Dwiggins was actually quite a bit more positive than Griffith about Luckiesh and Moss.

Dwiggins does credit Luckiesh and Moss with building "one corner of the foundation very nicely," but in effect mistaking one corner for the whole structure. I think Dwiggins's analysis is right on target. But as Peter says, I think Luckiesh was chastened by what he heard from Linotype, as the tone of the book is much more cautious and balanced than what he, elated by his discoveries, wrote in that initial report to Linotype.

Tinker, by contrast, relentlessly attacked that corner, denying that it had any validity. The title of my TypeCon talk, the 'Great Readability Scandal,' had nothing to do with Luckiesh's relationship with Linotype, and everything to do with Tinker's successful suppression in the scientific community of Luckiesh's work. Fatigue researcher John Stern's 1994 review article on the quarrel, and and at least partial vindication of Luckiesh's work after over 40 years, is what motivated my title.

Tinker's approach of reducing the reading experience to one "best" measure I think has had a very damaging effect on subsequent research, and one of my purposes was to rehabilitate the multi-variate approach of Luckiesh. In fact, the approach of Luckiesh allows for the very kind of mult-dimensional research that Dwiggins advocates.

I think science has a lot to benefit from Luckiesh and Moss's work. For typographers, the gain of putting their name and work to the concept of "readability" is I think to add clarity to the concept—that it is a measure of effort or fatigue in reading. And that effort needed to read is an important consideration in typography, in addition to speed and comprehension.

The concept of eye fatigue, derived from blink rate, is misleading.
Fatigue is something we associate with muscles becoming tired, but that cannot be the case with blinking here.
The normal blink rate is 10 times a minute, reduced to 3-4 times when concentrating (focused) on tasks such as reading.
So even if a reader blinks twice as much because the text appears dirty (or whatever), that is still not up to the normal rate of blinking, which one would assume does not cause fatigue.

There is arguably more of a physical toll taken by not blinking as much as normal (dry eyes), but as one is able to read for long periods of time at a reduced blink rate, surely the range of blink rate can vary considerably with no adverse physical effects.

There does appear to be some correlation between hard-to-read text and blink rate, but even the idea that more effort is required is problematic. What Luckiesh & Moss were really talking about is readability as convenience.

If there is eye strain from poor readability, I would expect it to be caused by excess use of the muscles used to focus the lens, but not by regression—because as much effort is required to move forward as to move back.

What if the strain of poor readability is entirely mental?

After all, do athletes experience tired eyes, when they are using their eyes to the max to follow the ball?

I think we assume there is eye strain involved in reading, because intense reading causes eyes to dry out, and hold one's eyelids open more than normal—which tires the eyelid muscles (the relaxed position being closed). But blinking more when reading would ameliorate that.

Therefore, text with good readability is fatiguing to the eye, because of the effort required to "hold the pose" of eyelids open, for long periods of time.

On the other hand, text with poor readability may cause fatigue to the muscles which focus the lens, as more careful inspection of the sharp details of letters is involved.

Bill, by identifying readability with effort or fatigue, aren't you in danger of making the same mistake Tinker made by identifying legibility with speed?

I think what needs to be shown is that *a* factor in the elevation of blink rate as a result of time on task is "ease of visual word-form resolution." This has to be thought of roughly in feature detection / feature binding** / “role-unit level statistics” gathering terms. I would think the kind of attention to “coordinating gestures” that Dwiggins aptly describes, makes the detection / binding / statistics-gathering process go more smoothly, that is more efficiently. We have to become acquainted with what happens in the “little grey cells” within the visual cortex, that affects effort and results in fatigue.

You write as if readability is the measure and fatigue and effort are the critical factors. I think construction, contrast, size, boldness and spacing are the critical factors, blink rate is the measure, and visual word-form resolution is what is affected. The effect on visual word-form resolution of manipulating the critical factors is the pivotal concern for questions of typographical readability.

[**I’m unhappy with the feature detection / feature integration “functional ontology” as you know, but that's somewhat beside the point here.]

Nick: The concept of eye fatigue, derived from blink rate, is misleading.

Note that Bill did not say eye fatigue. This is something that I started talking with him about over lunch after his talk, and perhaps a topic to take up again. I wondered what Luckiesh and Moss meant by fatigue, and certainly we should clarify that before making assumptions or drawing conclusions based on what you or eye might think the term means. Something that Bill mentioned is ‘cognitive fatigue’, which obviously implies something other than muscle fatigue. It is also worth bearing in mind that, due to the direct connection of the visual cortex, the eyes are in a sense part of the brain.
_____

Peter: Bill, by identifying readability with effort or fatigue, aren't you in danger of making the same mistake Tinker made by identifying legibility with speed?

Was that a mistake? It seems to me that Tinker's mistake was to ignore effort and focus only on results. Speed (with accuracy) is a great performance test, and if we want to identify that with legibility I don't think that's a bad use of the term. If you're concerned that the process of word recognition involves something other than individual letter recognition, that just implies that the term legibility must apply to units of recognition other than individual letters. That's a terminological shift, but it seems to me a sensible one, and far better than trying to identify legibility with recognising individual letters and readability with recognising something else. Why not say that legibility is associated with recognition (letters, features, relationships), and readability is associated with the effort involved in that recognition. In other words, legibility is performance and can be tested as such (speed and accuracy), and readability is cost and should be testable as such (and hence my first question to Bill after his talk: ‘Is blink rate a reliable test of effort in reading?’).

You write as if readability is the measure and fatigue and effort are the critical factors. I think construction, contrast, size, boldness and spacing are the critical factors, blink rate is the measure, and visual word-form resolution is what is affected. The effect on visual word-form resolution of manipulating the critical factors is the pivotal concern for questions of typographical readability.

I think one of the great things about rediscovering the origins of the legibility/readability distinction in typography is that it frees us from the unending cycle of debate, here and elsewhere, about where to draw the line between them and how they relate to visual word-form resolution. We've been going round and round for years now trying to associate readability with something that happens in word recognition, distinct from the role of legibility. If we shift readability into the domain of effort -- how easy/hard is it to read a given typographic treatment --, we can deal with it both conceptually and empirically as distinct from legibility, and in the process free-up legibility from a too-narrow association with letter recognisability.

John, I want to associate legibility with basic affordance and readability with efficient processing in neural encoding, activation and transmission terms. I think good typography and type design contributes to that. My attraction to Luckiesh is that his visibility / readability disjunction appears to provide a precedent or stimulus for this.

> what Luckiesh and Moss meant by fatigue <

L&M talk about fatigue of the extrinsic eye muscles, pupillary changes, heart rate and nervous muscular tension but they relate rate of involuntary blinking what Ponder and Kennedy summarize as “the mental tension of the subject at the time.” “In all probability the movements constitute a kind of relief mechanism, whereby nervous energy, otherwise utilized, passes into a highly facilitated path.”

Stern and an associate Orchard associate Ponder and Kennedy's term "mental tension" with the more usual term "attention" According to Orchard and Stern Ponder and Kennedy "inferred that persons inhibit blinking while actively engaged in information abstraction whether in response to simple stimuli or during the course of reading."

I'm taking Orchard and Stern to mean visual information abstraction, because how else would it become a gauge of typographic factors.

Peter: "Bill, by identifying readability with effort or fatigue, aren't you in danger of making the same mistake Tinker made by identifying legibility with speed?"

Peter, please don't take what I say as definitional, or we will get into terminological debates again, and these are usually fruitless. We can avoid getting entangled in terminological debates. I follow Popper's idea that theory is primary, and that trying to define terms precisely before developing theories is a trap, a trap that Popper calls "essentialism," deriving from Aristotle's faulty idea science is a matter of defining essences.

My model is Newton's law of motion F=ma, where 'F' is force, 'm' is mass, and 'a' is acceleration. None of these terms is defined in physics, and if you look at Max Jammer's books "Concepts of Force" "Concepts of Space," etc., you will see that nobody understands them fully, much less defines them precisely. And definition is inherently limited because it involves other terms that either are undefined or shift the burden again to other terms used in the definitions.

Newton's law expresses a relationship, and the meaning of the other terms is investigated and refined by testing and other theories relating to different forces, motion, and so on.

What is normal in the progress of science is *demarcating* one term from another. For example, Newton for the first time clearly distinguished mass and weight. Weight is the force of gravity on an object; its resistance to motion under force is its mass. So, for example, an object with the same mass will weigh different amounts on the earth and moon, but respond in motion to the same amount of force in the same way.

In the case of Luckiesh, I think the basic conceptual clarification is the distinction between performance and the costs of performance. I like this language better than Luckiesh's, but the concept is clearly there in the chart he often reproduced in different forms:

The middle darker portion represents the costs of performance, which you can call effort, fatigue or whatever. The idea is that brighter light, better typography, etc., can move you to the right on the chart. The important thing is not the terminology, but the distinction between performance measures and psycho-physical costs, whatever you call them.

I am not stuck on the terms "ease" or its inverse "effort" or "fatigue". The important thing is the demarcation between performance and its costs gives us some additional clarity that we didn't have before, as John has said.

I think it is clear that blink rate does partly indicate cost of reading performance. But "What exactly is blink rate measuring?" is a good question. For we know blink reacts to many things, including auditory tasks, and not just visual ones. Luckiesh and Moss tried hard to keep factors other than typographic ones constant, so they wouldn't confound the result. And they got cleaner results than their rivals, who had less careful controls.

But the question remains what exactly are the causes of increased blink rate with say type that is lighter or bolder than ideal for extended text. For example are they muscle fatigue, fatigue from processing the visual cortex, or from elsewhere in the brain?

These are questions that cannot be decided by definition, but need to be investigated by hypothesis and experiment. My own hypothesis, influenced by recent papers, is that one of the factors is that when there is more "noise" relatively to "signal" in set type that the visual cortex has to use more dopamine and other brain chemicals in processing the image. Then the brain cells take more time to get the fully renewed chemical supply from the blood. As blinks partly suppress visual processing, they give the circulatory system more time to refresh the brain. Incidentally, John Stern felt that duration and amplitude of blink was a more sensitive measure of fatigue, and I am not set on blink rate either. But I do see blink as reflecting brain processes, and not just muscular strain.

So I see increased blinks as they relate to typographic factors—as opposed to brightness, glare, contrast etc.—as primarily caused by the need for more time to refresh brain chemicals.

I don't know how to test this, but it seems to me it ought to be testable.

Thus I agree with you, Peter, that one part of blink rate, which hopefully we can isolate in testing, does reflect brain processing costs. But this is a matter which is not helped at all by trying to make it a question of definition. The idea of costs of performance is somewhat vague and it should be until we learn more what they actually are. That's a problem for further scientific inquiry, and not a definitional issue.

Bill, I'm not an essentialist. Essentialism is your bug-a-boo and you sometimes like to pin it in me because of my phenomenological sympathies. Like you and many others I'm looking for terminological clarity and stability. I think terminology does matter, but theory is where the rubber hits the road. Having a shared terminological matrix [construct system / functional ontology] adequate to the task of opening up the domain is of great ancillary importance. Terminological agreements are like social contracts and have to feel meaningful if they're to stick. Why is crowding called crowding?

My point was to say that just saying “[readability] is a measure of effort or fatigue in reading” or psychophysical cost is insufficient. It doesn't tell me much, or at least not enough of what we need to know. Effort and fatigue are fuzzy. They just point us in a general direction or tell as approximately where to look. At the very least readability has to be seen as a qualitative marker used in assessing functional conformities between aspects of items in a stimulus set and visual processing architectures of a distinct sort, i.e., the sort we use in visual information processing during continuous reading. Non-conformities my lead to an expense of effort which over time might contribute to fatigue, but I don't know what it means to say “[readability] is a measure of effort or fatigue in reading”

For the luckieshian broadening of the field to be of use today, and its terminological initiative to become cognitive-scientifiaclly stabilized by being securely anchored to actual neuro-physical functioning, we need to focus more directly on the nature and brain-anatomical location of the actual costs — which are the relevant costs when we are trying to assess typographical and type-design specific questions, and what exactly is responsible for the costs, why does the progressively more frequent dopamine infusion become necessry. For example does bad spacing throw out the narrow phase alignment necessary for efficient processing? Or, to use Dwiggin’s example, do dissimilarities in anatomical structure throw off the channel tuning necessary for efficient quantization?

You try to move in this direction by referring to signal to noise ratios.

Just to flesh out, because my edit function stops after 15 or so minutes: [line 4] theory, in the sense of conjectures and refutations (using tests) or confirmations, is where the rubber hits the road...

To be clear, I'm not rejecting the association of readability with ease. I'm just asking, what in perceptual processing terms or neuro-mechanical terms, or perceptual psychophysical terms does this ease consist of. Perhaps there are more kinds of ease in reading that have to do with typography and design than the ones I specified above, for example those involved with moving along the line, and those involved in text-navigation.

I also react to your phraseology when you use “is a measure of” and “refers to”. In Reading as a Visual Task Luckiesh and Moss write: “[t]he term readability is used to express that attribute of reading materials which governs [my emphasis] the relative ease with which different materials may be read by subjects possessing normal vision and exhibiting normal responses.” and in the earlier passage quoted above it is clear that Luckiesh and Moss see this ease as an integral effect, or a summation of the effects of various typographic factors. To give a measure of these effects, that is, to provide some quantitative data, blink rate is used. It was the most sensitive indicator they could find.

My point was that strictly speaking, readability is not the measure, blink rate is; readability doesn't refer to effort, greater effort is what problems with a text's readability results in. I want to preserve the sense of readability as an attribute which governs, or qualitative marker. But I don't want to put to fine a point on it. This is after all just terminological.

>Like you and many others I'm looking for terminological clarity and stability.

Peter, I'm not looking for terminology at all, but rather for a deeper understanding of how we read. Terminology is necessary tool but arguing about it to me is a distraction. I think terminology will be more settled when the theories are better, not the other way around. I see terminology as sometimes a help and sometimes a hindrance, but not the heart of the matter. I see a focus on terminology, rather than try to answer questions about reading, as a distraction.

As I have said, if thinking about terminology helps you get new ideas, fine, because you have good ideas, but the reality is that this not the only way to get new theories, and this approach I think makes your writing much harder to understand.

We don't have to use the terms "readability" and "legibility" at all. The question of how reading speed, comprehension, and fatigue are affected by typographic factors are important questions whatever terminology you use.

To go back to Newton's law of motion, you can also formulate it in different terminology, such as work and energy. But it's the same theory, equivalent in meaning. So terminology is not the heart of the matter.

Luckiesh said that reading comfort or ease is affected by typographic factors, and causes variations in blink rate. That is a theory you can formulate many ways, using different terminology. I agree with you that to advance this theory, we need more understanding of the causality of blinks and the nature of eye and brain processing during reading.

Bill, you seem to think I am trying to get terminology settled first and than build theory. Just the opposite is true. Any terminological refinement I've proposed was generated by immersing myself in the literature, running up against disconnects in the models proposed to illuminate the data, and trying to move forward toward a more sustainable and intellectually satisfying alternative theory.

Anyway, forget about it, as my father used to say.

Where were we? Oh yes, there is a necessity to talk about visibility in the nuanced manner Luckiesh and Moss do, corrected perhaps by Rae; there is the necessity to address reading speed or output; there is a need to delve further into the intricacies of blink rate, ease, effort and fatigue.

On reading rate I have the following.

One of the outcomes of the debates between Luckiesh and Tinker is the recognition that Tinker was concerned mostly with the ‘knee‘. When size or illumination levels are plotted against reading speed the graphs show a steep rise, than a wide plateau, and than a gradual decline. See the lower curves on the two graphs earlier in this thread. The knee is the transition area between the sharp ascent and the flat plateau.

One way of characterizing the Luckiesh / Tinker debate is to say that Tinker was concerned almost exclusively with the knee, while with readability Luckiesh and Moss were concerned with factors affecting what happens on the plateau, and with threshold and supra-threshold visibility factors they were concerned with factors determining what happens below and up to the knee.

Is there a home for the typographical term legibility in this scheme?

Is the knee an inflection point where visibility issues become less decisive and readability issues take over?

And John, one of the determinants in setting the location of the knee and in the fact that rate stays at a constant level through a range of illuminations and sizes is ‘the efficiency of the eye’ and maintaining “clear seeing.” The efficiency of the eye was explored in detail and with tests by Ferree and Rand before Tinker and Paterson or the Luckiesh and Moss team hit their stride. Both teams were aware of the Ferre and Rand work. Luckiesh and Moss regard the efficiency of the eye as a confounding factor in reading speed. They claimed it accounted for the relative insensitivity of rate of reading as a criterion.

The “efficiency of the eye” concept brought to my mind your notion of readerability!

Both teams were aware the efficiency of the eye in maintaing clear seeing under adverse or sub-optimal conditions has a cost. Ferree and Rand even tried to control blink rate. Tinker didn't ignore the cost but tried to accommodate it by suggesting safe zones. He made no attempt to quantify these. Luckiesh and Moss of course tried to quantify the cost.

This is very interesting information.

Not being familiar with this history, I had simply felt that legibility, being much easier to measure than readability (in the sense of comfort in reading), had tended to push readability aside without help.

The fact that there was early research that showed that readability could also be measured objectively, rather than left to the subjective experience of typographers, is something I find very encouraging - because it gives typographers a way of defending themselves against a misguided emphasis on legibility alone.

Obviously, a typeface can be highly legible, but distracting and ugly at the same time.

"Obviously, a typeface can be highly legible, but distracting and ugly at the same time."

Hmmmm, not so obviously to me.

Peter: Perhaps there are more kinds of ease in reading that have to do with typography and design than the ones I specified above, for example those involved with moving along the line, and those involved in text-navigation.

It seems to me that we need to develop what we mean by ‘ease’ (and hence by ‘effort’), while being aware that it might provide difficult to isolate different kinds of ease in a meaningful way (meaningful to type design that is). Note that I don't mean develop what we mean in terminology, but identify the phenomena that constitute ease.

The “efficiency of the eye” concept brought to my mind your notion of readerability!

Indeed. And now, thanks to Bill's research on L&M, we're faced not only with how well people perform the reading task, but also how well they compensate for factors that make the task more difficult. This really does suggest to me that if type designers and typographers are to positively affect readability, it is in the area of reduction of the need for such compensations, rather than in the area of word form resolution per se. Unless the text is functionally illegible, we will resolve the word forms, because that is what we do when we are reading, and we may even do so with relatively consistent speed and accuracy, but what it costs us to do so will vary depending on both text-intrinsic (type size, type weight, line length, leading) and -extrinsic (lighting, motion, noise) factors.

In my last post above I wrote: “one of the determinants in setting the location of the knee and in the fact that rate stays at a constant level through a range of illuminations and sizes is ‘the efficiency of the eye’ and maintaining “clear seeing.””

and maintaining “clear seeing” should have been in maintaining “clear seeing.”

I also asked: “[i]s the knee an inflection point where visibility issues become less decisive and readability issues take over?”

Bill criticized Luckiesh and Moss for not doing enough with reading speed and trying too hard to correlate visibility and readability. Had they recognized that Tinker and Paterson’s measure of choice might help them with this by revealing an inflection point… [well, who knows where their conversations might have gone]

[John] "This really does suggest to me that if type designers and typographers are to positively affect readability, it is in the area of reduction of the need for such compensations[…].”

Yes!

Yes, John's "readerability" is a key factor here. We have great ability to compensate for fatigue, and the questions are:

1.) On the scientific side, how can we detect and measure when we are compensating for fatigue?

2.) As type and graphic designers, how do we minimize the unnecessary typographic causes of reading effort and fatigue?

One of the complexities here is made clear by the Cognitive Fatigue book: when we are enjoying a difficult task, we don't get nearly as tired doing it. When we are bored or otherwise don't want to do it the fatigue factor goes way up. One of the theorists on fatigue, David Hockey of Sheffield U., has a model in which cognitive fatigue is basically a competition for attention.

Luckiesh was careful to try to keep the reading material uniformly interesting, so that factor would not interfere with the results.

I agree with Luckiesh that global features of the typeface, design features that go throughout the face, rather than individual characters, are the main thing in readability. However, regarding "boldness" as single feature is too simple. The curves that Dwiggins writes about and weight distribution are two sides of the same coin. And character shape and width and letter spacing all contribute to overall typographic color. And in the real world aesthetics matters a lot to fatigue because of the interest factor.

Peter, I don't understand what Luckiesh was writing about vis a vis Tinker's interest in the "knee". The reading speed tests are all beyond threshold conditions. Could you quote the passage?

Bill, the part about the role visibility (including level of supra-threshold visibilty) plays above and below the knee is my addition. There has to be a basic level of supra-threshold visibility to get reading-speed up to the knee.

Luckiesh's criticism was that using illumination values that merely take reading speed to the knee is the wrong metric for determining ideal lighting values; it ignores the additional benefits to ease [pe: efficient processing] that can accrue with additional lighting. It is the wrong metric because reading speed fails to discriminate performance characteristics, such as ease, above the knee, i.e., on the plateau.

Bill, Luckiesh’s passage about the “knee” is in “A Reply to Dr. Miles A. Tinker” Illuminating Engineering 1948 [V 43, page 888]: “Tinker also concerns himself with the “knees” of curves, which may be partially justified when one is interested in the compromises involved in recommended footcandles rather than in ideal footcandle-levels.”

Just to provide context to Peter's latest posts, Tinker's quarrel with Luckiesh started before the issues over testing optimal typography. It started back in the 1930's with a dispute about best levels of illumination. Tinker wanted to measure only performance, and visual acuity doesn't change rapidly once above threshold levels of illumination. Luckiesh said yes, but people complain of fatigue and eye strain at low levels of lighting. He wanted better measures above threshold, measures of fatigue to indicate comfortable illumination levels. He did several measures, eventual finding blink rate. This measure was also able to distinguish typographic factors more sensitively than the performance measure of reading speed. Tinker attack Luckiesh on both fronts.

Tinker carried on both disputes, eventually bringing the illumination issues to the Illuminating Engineering Society, of which Luckiesh had once been president. Tinker lost that battle in the short term, but Luckiesh's methodology was also lost in this field. However, the higher illumination levels that Luckiesh had argued for—and still higher ones—are still recommended in the standards issued by the Illuminating Engineering Society.

So for Tinker's purposes -- measuring reading speed --, illumination could be isolated from results and ignored if a) above threshold and b) consistent across testing. And that seems to make sense insofar as he didn't care about effort/fatigue, which he was also isolating and ignoring.

I suppose, though, that there is some sort of threshold relationship also between effort and speed, i.e. a point at which the compensations of effort fail and reading speed is affected. This implies that factors affecting effort, such as illumination level, can't be fully isolated in speed testing unless above an effort threshold as well as a legibility threshold.

The mid thirties debates were about visual acuity and light levels, not reading. If I remember rightly, reading speed is affected by light levels above threshold, just not a lot. In fact, Tinker didn't control for lighting level, at least prior to his 1940 book. While writing it he evidently became aware of Luckiesh's work on reading. He claims in the introduction he will address Luckiesh, and then doesn't. He also brags that he doesn't control for illumination—they were in bright rooms with daylight, darker with light bulbs etc. He says that it's "realistic". So much for scientific controls.

I think the main thing fouling up Tinker's results is his weird speed of reading test, though there are other more minor confounds, including illumination level and distance from the eye to the page.

In a private letter to Tinker—Sutherland quotes it in her PhD thesis on Tinker—Luckiesh dismisses prior work on reading because it didn't control for illumination. He doesn't say "this means you, buster" but he may as well have. Luckiesh also said he didn't like public disputation, but rather to let the work speak for itself.

Evidently, it didn't speak for itself, because this turned out to work to Tinker's advantage. Luckiesh didn't write on the topic after his retirement in 1949, and Tinker, who was younger went on attack Luckiesh, with the Legibility of Print (1963) being the work most people looked to to summarize previous work.

Thanks, Bill. And here I was thinking that Tinker was simply too narrow in his focus, whereas his failure to even control for illumination suggests he was doing sloppy science.

[Bill] “The mid thirties debates were about visual acuity and light levels, not reading.”

Bill, I wonder if you meant “not typography” in the above.

Tinker’s first (to my knowledge) critical comment on Luckiesh’s work is in 1934 in “Illumination and the Hygiene of Reading” [Journal of Educational Psychology / 1934 / Volume 25 / 665-680]. The first sustained effort to question the value of Luckiesh and Moss's work is in 1935 in “Cautions Concerning Illumination Intensities Used For Reading” [American Journal of Optometry / 1935 / Volume 12 / Number 2 / 43-51].

Luckiesh and Moss’s The Science of Seeing was published in 1937. It contains a chapter on “Reading as a Task,” which includes a foray into questions of typography. In Tinker’s review of this book he extends the criticism of the Luckiesh and Moss work to their material on typography. I suspect that it is the discussion of typography in The Science of Seeing that lead Linotype to link up Luckiesh and Moss, probably before the Tinker review came out in 1938. I haven’t been able to locate a personal copy of the book or access a library copy, but I am able to deduce some of what it contains from google books snippet view.

The Tinker review includes this statement: “The failure to consider many important contributions from other laboratories weakens the book considerably, especially in the discussions of legibility of print and factors in hygienic illumination.”

In two articles in 1945 / 1946 The reliability and validity of blink rate as a criteria becomes front and centre. This is after Luckiesh had challenged Tinker in a letter to prove that speed of reading is a criterion of ease of reading.

From the Italian Wikipedia (primary source of human knowledge): ;-)

«Italian language, unlike English, does not have two different words to describe the legibility of calligraphy or of a typeface (legibility) from the reading fluency as a function of linguistic structure (readability). The same text may be legible but not readable».

It’s very likely that – if this is the main concept expressed by “readability” you find it expressed in italian with «comprensibilità», which is to say «intelligible» or «understandable».

The fact that the notion of readability as ease of reading, which is a specific meaning, was introduced by an englishman may explain the reason we do not have it in italian. I still don’t understand why it causes so much unrest here at Typophile… :P