How does the serif on a sans-serif i increase legibility/readability?

blank's picture

I am trying to nail down why it is believed that the serif on the lowercase i and j in fonts like FF Info increases legibility and readability. Is it that the serif makes the shape differ more from I and l when viewing conditions are poor and the dot joins with the rest of the letter?

ebensorkin's picture

You will hear many things on this topic if it gets going but let me start with the idea that a distinctive shape is helpful in reading the serif on some lc i & j forms may help enhance distinction. Another thing it may be doing is helping to ensure that even with rougher treatment in spacing that some needed negative space is maintained. The bar may also draw your eye to or in some way highlight the dot - the real marker of difference for i & j. But as far as I know there is no science on this. Just observation & supposition.

Thomas Phinney's picture

The main thing is telling the "aye" apart from an "el" under poor reading conditions....


Nick Shinn's picture

it is believed

By whom?
IMO it's BS, just hype for a new style.

When ITC etc. came out with their big x-height, tight-fit faces in the 1970s, it was said those features improved legibility, but now we look at that stuff and roll our eyes at how nasty it reads.

Whatever the trend in text type is, people will say it's for readability's sake.

dezcom's picture

An eye for an i and an elle for an l. The logic is in the cap I, lowercase l, and number 1 needing differentiation. I think their was some research in the signage arena to support this but in text type, I don't think there is a study to see.


Gary Long's picture

If type designers start adding serifs to sans serif fonts to make them more legible, or for whatever reason, at what point do you say this is not really a sans serif font anymore? It's cheating! Sans serif means "no serifs"! Of course there's nothing wrong with the practice---there are many typefaces that combine elements of serif and sans serif design to good effect. Interesting that I've never heard of a serif font where serif's were removed to enhance legibility.

dezcom's picture

I don't see why the purist approach of categorization must be the only way. Type designers design typefaces. The categories and the rules for defining them are just there for convenience. It is not black and white nor need it be so. Shades of grey are welcomed at least by me. Others are welcome to think differently as it suits them.

Florian Hardwig's picture

It’s cheating! Sans serif means “no serifs”!

That’s why they’re called some-serifs.

dezcom's picture

Or perhaps they are just being serifticious :-)


ebensorkin's picture

This is probably obvious but on the off chance that it is not - the origin of those 'serifs'* is typewriter fonts where they clearly do serve a purpose. I can easily imagine a situation where somebody want to make a new sans & is interested in the techy feel of the mono and 'hey presto!' the semi serif is born. Also, the fact that these 'serifs' clearly help in a mono may have led the designer to extrapolate that they helped in their design too. And depending on the design - they might have.

The interesting question it seems to me is who started it? Was it by any chance... Erik S?

*I don't know that I would call them that actually because they end up being part of the deep structure to my mind...

blank's picture

The interesting question it seems to me is who started it? Was it by any chance... Erik S?

He may have kicked it off in Officina Sans, but I’m hardly an expert on such matters. I guess that I should just ask him, but I hate bothering really busy people too often.

Nick Shinn's picture

Here's part of the research I did for Preface.
I could also have included Bell Gothic (Griffith, 1937), for its serifed I.

Rob O. Font's picture

Thanks again Nick. It doesn't exactly look like the Ascent of Man(s), (place pun here) ;)
Eben is on the correct path, I think: the monospace, brought the machine readable, which brought the techy feel. The "el" (and i, imho), notice, is two things, the rounded base on the left, and the serif on the right. Meta, e.g. barely shows the latter as opposed to OCR-B, but both have a rounded stroke ending on the left? So, some of this, I think also comes from what we're used to from...italic. I also note, that Orator, with it's monospaced some serif and practically no descent must'a been a reading chore I'm glad I never had.


ebensorkin's picture

Nice graphic Nick.

The other thing that interesting is the foot. You might say that thats the thing that's 'missing' in Nick's graphic. A few Monospaces also have them like Osaka:

And the Venetian Calligraphy / Roman upright antecedent:

Bert Vanderveen's picture

Print these letters on a piece op paper: p d q b.
First in a sans. Look at the paper, turn it around, try to recognize the letters.
Now do the same with a serif.
I rest my case.

. . .
Bert Vanderveen BNO

eriks's picture

He may have kicked it off in Officina Sans

I most certainly did not, but I may have re-started the trend. ITC Officina is very much my version of Letter Gothic, which has the serif-i. The monospaced faces had to make the narrow letters wide and the wider ones narrow. So f, t and r had to be extended, i and j got a top serif and the l sometimes both top serif and bottom curl to the right. Same goes for cap I.

With proportional type you still get the benefit of more white space around narrow letters, which supposedly makes serif faces more pleasant (if not easier) to read. In other words: putting serifs on i, j, l or r makes a sans face less sans. The fact that this also avoids common confusion between l, i, I and 1 is really only important for forms, lists, directional signs and other information where one has to pick out single letters rather than read word shapes.

blank's picture

Thanks, Erik. Next time I’ll just email you to begin with ;)

ebensorkin's picture

Thanks Erik!

k.l.'s picture

In other words: putting serifs on i, j, l or r makes a sans face less sans.

I agree.  ;-)

With proportional type you still get the benefit of more white space around narrow letters, which supposedly makes serif faces more pleasant (if not easier) to read.

I cannot see the benefit of adding serifs plus space to some glyphs only. It does not solve anything, but may cause uneven spacing. Why does an i j l or r need more leftside space than an h or m or n or u? If a sans is spaced too tight, all letters need more space. In my desire to know how it feels to design an l-with-a-tail, I tried and considered it a pain: Either you add only little space to the right -- then you need positive kerning to make sure it does not bump into following letters & punctuation marks. Or you avoid these collisions by increasing the l's right space, which results in holes in the text image, with some letter combinations -- then you need negative kerning to compensate for this. No extra trouble without the tail.
[Read as a remark about design technique, aesthetics or personal preferences being irrelevant here.]

Nick Shinn's picture

TheMix and Rotis have semi-serifed lower cases.
Interestingly, TheMix caps are sans, while Rotis are semi.

Ditto Karsten, I don't think that a few serifs added to a sans font will, per se, improve it--there are so many other factors.

ebensorkin's picture

- there are so many other factors

I agree - Serifs per se added to the r i & j etc do not generically* help. Speculating a little, apart from monos; It seems like with a really featureless Sans ( by which I mean an overly geometric one ) I can imagine the block serifs might help. Designs that are 'breaking' rules and are therefore harder to recognize might be candidates as well,** and faces where the design is modular too maybe. What do you think?

* as in 'every time'

** Talking about the Mix vs the Sans.. It seems like the Mix is pretty darn masterful really whereas I think Rotis isn't so much. Rotis may need the extra help. This is just shooting from the hip of course.

Syndicate content Syndicate content