An endeavor such as this sparks many questions. What did the designers interpret the problems to be? What process did they use to arrive at their solutions? How did their solutions differ? How successful are they as typefaces? How does all this relate to what we know about legibility and dyslexia? Most importantly, did they achieve their goals?
What problems were they trying to solve? Boer says Dyslexie is intended to help dyslexics read with less effort, be less affected by the symptoms associated with dyslexia and make reading more accurate. Sylexiad is positioned by Hillier as an ongoing design investigation to make a typeface that would be preferred by adult dyslexics over alternatives.
The success of these typefaces likely depends on just how you define the goals. If satisfying people with dyslexia is the goal—as Hillier says for Sylexiad—then they may have been quite successful, at least with a lot of their target audience. One Dyslexie customer wrote to Boer, “I showed it to a colleague of mine who happens to have dyslexia, and he literally broke down and started crying because he was able to read your info page aloud.” Hillier’s entire methodology for Sylexiad was focused around studying dyslexic readers' typeface preferences.
But if we look at scientific evidence as to whether using these fonts has (or should have) any measurable effect on dyslexics reading, things are not quite so clear.
WHAT WE KNOW ABOUT DYSLEXIA AND LEGIBILITYThe manual for Boer’s Dyslexie claims that “symptoms of dyslexia are mostly: turning the letters; mirroring the letters; changing the letters.” Unfortunately, although it is a popular belief that these are symptoms of dyslexia, the current scientific understanding of dyslexia says otherwise.
Hillier’s definition (in the thesis for which he developed Sylexiad) matches the current, most common scientific understanding, that dyslexia is “a phonological deficit that influences an individual's ability to decode and encode language.” The phonological part is important. For the most part, scientists believe the difference between people with dyslexia and non-dyslexic readers lies in their confusing similar sounds, not shapes of letters. It seems there is evidence to support this distinction. This theory does not suggest anything about visual processing and mirroring/rotating letters. Although dyslexics may do this, most research suggests they don't do it much more than anybody else at the same reading level.
Figure 1: Christian Boer’s Dyslexie and Dr. Robert Hillier’s Sylexiad displayed at the same nominal point size. (Not all faces of Sylexiad are shown.) Note the erratic/tight spacing across different styles of Sylexiad Serif for combinations such as “bold” “foin” and “bu” and the excess space on the right side of the lowercase letter “l” in Dyslexie.
Mind you, right after the seemingly false statement about symptoms of dyslexia, the Dyslexie manual says, “By emphasizing the shape of each character there will be less confusion between the characters.” Even though mirroring and rotating may not be problems for dyslexics (or anybody else), it is certainly true that letters that are too similar can be easily confused by readers, whether dyslexic or not. The dominant theory of decoding alphabetic text (as opposed to Chinese, for example) today is parallel letter recognition. Although several words worth of letters are seen at once, the letters within each word are recognized and decoded in parallel by the brain (Larson, 2003). Making these letters more easily distinguished from each other is helpful, and experiments have even identified which parts of letters are important (Gosselin, 2011).
One concern is that dyslexia may not be a single thing, but in fact a label for several different problems with reading, some of which are more common than others. Is it possible that there are those for whom letter rotation and mirroring is a key part of their problems, even if they are a minority of dyslexics? Perhaps. But this would not be the first time that science overruled our perceptions of reality (see Copernicus and the sun not revolving around the earth). Apart from the mirroring/rotation issue, I could find no evidence that dyslexics need different kinds of font legibility enhancement than other readers.
Dyslexia aside, what we know about legibility for all readers, dyslexic and otherwise, suggests that when letterforms are designed with closed shapes such as those in Helvetica and Arial, the glyphs are harder to tell apart from each other than in more humanist typefaces such as Lucida Sans/Grande, Myriad, Frutiger and Gill Sans. I believe Dyslexie is enlightened in that it opened up those counters, but misguided in using Helvetica-style shapes as a starting point in the first place instead of a humanist model (or even a geometric one such as Futura).
Many of the other modifications in Dyslexie are specifically about protecting against somebody’s brain accidentally rotating or mirroring the letters. This includes the increased boldness on the bottom of letters across the entire typeface. It seems clear that unless the research on dyslexia is all wrong, this won't help most (if any) dyslexics. However, a few changes were also made simply to make letters more distinct, such as adding the tail on the lowercase “l.” The lessons of cognitive psychology suggest these kinds of changes could be helpful to all readers, with or without dyslexia.
Figure 2: Samples in Sylexiad Sans, Sylexiad Serif, Dyslexie, Helvetica, and Lucida Sans/Grande, and Gill Sans, adjusted to the same apparent size. Compare the closed counters of Helvetica and the open shapes in Lucida and Gill. Note also the relatively even proportions of the capital letters in Dyslexie and Helvetica compared to the more varied capital letter widths in Lucida and Gill.
Sylexiad seems to have been created from scratch in form, and is overall less daring but perhaps a little more successful. Hillier concentrated a little more on differentiating individual letters. The “n” and “u” each get a right-side serif, even in the sans version of the typeface. Yet he went with the simpler forms of the letters “a” and “g” instead of the more complex (and hence more easily distinguished) forms. The capital letter “I” and numeral one get horizontal bars to help distinguish them from each other, but the lowercase “l” does not.
PROCESS & SOLUTIONSSylexiad was designed with an ongoing feedback loop from dyslexic readers (what Hillier calls “developmental typeface testing”). This is a great way of developing a typeface, even if not the innovation that Hillier makes it out to be. Prior to Sylexiad, Hillier went through three iterations of previous experimental typefaces called Dine to learn about some specific variables and how dyslexics responded to them. At the end, serif and sans serif versions of Sylexiad were compared against Arial, Sassoon Primary and Times New Roman. The most striking result was not in functionality, but in personal preferences. According to Hillier:
“The majority of non-dyslexic readers preferred: serif style fonts, lowercase forms, large x-heights, medium weights, variable strokes, normal inter-word spacing and familiarity of form. They also preferred Times New Roman as a family. The majority of dyslexic readers, however, preferred: handwritten-style fonts, uppercase (Sylexiad) forms rather than lowercase forms, long ascenders and descenders, light weights, uniform strokes, perpendicular design, generous inter-word spacing and, as with the control, familiarity of form. Unlike the control they preferred Serif Sylexiad as a family.”
Dyslexie got rigorous testing as part of a master’s thesis in which it was compared to Arial for both dyslexics and non-dyslexics (Renske de Leeuw, December 2010, Univ. of Twente). The study is being portrayed in Dyslexie promotions as showing that dyslexics made fewer errors in reading with that typeface. However, that effect did not reach statistical significance. There were various other differences, both positive and negative, that did not reach statistical significance. There were also a few statistically significant effects, both positive and negative, on even narrower and more specific issues. Sadly, the overall evidence does not show that Dyslexie is “better” than Arial (nor vice versa).
REVIEWED AS TYPEFACESSylexiad and Dyslexie are both marginal in meeting my expectations of a professional quality typeface. That isn’t as bad as it sounds, coming from me; I feel that way about a pretty noticeable minority of typefaces sold at retail today, and about 90 percent of free fonts (which is a rant for another time). Given that the designers were graphic designers and not type designers, they arguably didn't do all that badly.
Sylexiad is actually an ambitious 32-font super-family with sans, serif, spaced sans and spaced serif variants, each coming in two four-member families of differing weights (either thin or medium, with matching bolds and italics). The various fonts have a reasonably complete character set, but character coverage is slightly inconsistent across styles. The spaced variants replace the default, overly thin spaces with extra-wide ones. Sylexiad’s general shapes are interesting. The overall impression is like a casual sans and semi-serif extended family. Sylexiad Serif in particular reminds me of several 1970s typefaces with their semi-serifs, such as Seagull and ITC Korinna.
But the devil is in the implementation. Sylexiad has many awkwardly lumpy curves in the individual glyphs. The spacing is erratic and tends towards overly tight (see Figure 2). Accent placement is poor, although not awful. The Sylexiad italics are simply the upright faces slanted twelve degrees, without any optical corrections.
For format, one can choose Mac PostScript Type 1 or Mac TrueType implementations of Sylexiad. I was surprised that there were no OTF or TTF fonts available for a typeface that was made in 2007. The families don’t group properly in InDesign CS5, which made them harder to work with.
Figure 3: The original “g” and “n” in Sylexiad Serif Medium Bold (top) and how a type designer might improve their curves (bottom). Straight-to-curve transitions at the top and bottom of the “g are more gradual, but the over-gradual transition on the “n” is firmed up and the right shoulder of the “n” is raised. The round stroke meeting the vertical at the center right of the “g” is more tapered to avoid the “dark spot” that resulted in the original.
Dyslexie is both more and less successful than Sylexiad, and ambitious in different ways. The fonts arrived as TrueType TTFs, which work on both Mac and Windows. There were just two fonts, a regular and a bold, with no italics. In typical non-DTP applications (including web browsers), users will get faux italics. The typeface has no kerning.
As mentioned earlier, the shapes in Dyslexie are essentially those of a modified Helvetica. Unfortunately, adding weight to the bottoms of all the letters, plus making vertical strokes wider at the bottom, just looks horrid. Dyslexie looks like a gross mutation of Helvetica and Comic Sans. There’s no point in redrawing the letters, because the basic idea is just broken, from an aesthetic viewpoint. What’s amazing is that when set in text, it still works perfectly well in terms of readability—so at least it is not achieving the opposite of its goal.
Part of Dyslexie’s solution to make letters easier to recognize was to make things bigger. In fact, almost everything except the numbers was enlarged.
The astute reader may, at this point, realize that when you make everything in a font bigger piecemeal, it is pretty much the same as making the whole font larger for a given point size. This doesn’t really achieve anything other than being able to claim it’s being set at 12 point when it is effectively larger, for purposes of comparing against some other typeface. Hey, let’s just rescale all the world's fonts 25 percent bigger for the “same” point size and improve legibility around the world! Of course, this has problems (witness Figure 4).
ACHIEVEMENTWith my endless complaints about these typefaces, it would be easy to lose sight of an essential point, as I did for a bit while scrutinizing them. Designing a typeface is a vast amount of work, and the people who designed these typefaces did so primarily to try to solve a problem that vexes millions world-wide. I applaud their motivations and their efforts. Both Hillier and Boer were also quite helpful in supplying fonts and material for this review.
Figure 4: Line-to-line collisions with Dyslexie, set with default settings in InDesign CS5.
Nonetheless, I can’t wholeheartedly recommend either Sylexiad or Dyslexie for dyslexics, people creating written materials for dyslexics or anyone else. Of course, dyslexics may react positively to these typefaces, especially if told a special typeface is being used for their benefit. Or they may simply prefer these letterforms even if there is no measurable effect (though as noted in my last article [CA July/August 2011], the effects of better type can be hard to measure).
Yet there is a lesson to be learned from the ways these attempts fail, particularly when contrasted with a spectacularly successful counter-example: Check out Clearview Hwy, the highway signage typeface, and the process used to create it. Typefaces created for challenging special purposes benefit from being designed by experienced type designers, in concert with domain experts and cognitive psychologists.
Would expert input from cognitive psychologists with expertise on reading and dyslexia have simply aborted the projects? Or with the help of a type designer might they have gone in a new or different directions, focused more on traditional legibility concerns, useful to a broad spectrum of users beyond dyslexics? There may be multiple answers. I hope future attempts at a typeface for dyslexics reveal them. ca