Lower case ‘a’ from Adobe Caslon Pro, superpos...
Lower case ‘a’ from Adobe Caslon Pro, superposed onto some guides. (Photo credit: Wikipedia)

I was lucky about 25 years ago — I got a job working with someone who was fascinated with typography and who knew it well. I’d dabbled a bit in it, as my college newspaper migrated from phototypesetting to computer typesetting, but it was only in these early career days with an available mentor that I learned about serifs, kerning, counters, leading, ascenders, and descenders, as well as the palpable but subtle differences between superficially similar typefaces (Univers vs. Helvetica vs. Futura, for instance). Being an autodidact, I immersed myself in fonts for many months, buying books, reading magazines, using tools, visiting designers, and so forth. Within a couple of years, I’d gained enough knowledge so that when I needed a job in a pinch, I was able to parley a full-time job setting type for academic textbooks.

Cracking the code on type has helped me see a different part of the psychology of communication, and one that’s often overlooked — the aesthetic aspect and its almost undetectable power. A recent two-part series by Errol Morris in the New York Times sought to explore fonts and their effect on credulity — their ability to generate “truthiness” themselves.

Morris did this in a clever way, writing an initial column ostensibly testing whether readers were optimists or pessimists based on how they responded to statements called out separately from the main text. With the assistance of a computer programmer, Morris was able to show each viewer these statements in one of six randomly generated fonts — Georgia, Comic Sans, Baskerville, Computer Modern, Trebuchet, or Helvetica. Then, through analysis of the more than 45,000 quiz results, he and David Dunning, a psychology professor at Cornell, were able to draw some conclusions.

In short, Baskerville proved to be what they call “the king of fonts,” making the true statements more likely to be accepted and less likely to be dismissed. The p-value was 0.0068.

This may be a surprising result to Morris and Times readers, but it makes complete sense to typographers. In fact, my typography mentor used to mention how Baskerville was so good for conveying information because it “disappeared” to the reader more easily than other fonts with more exotic strokes, x-heights, or designs, allowing almost unconscious absorption of the underlying information. This comports well with Morris’ little experiment — after all, there’s nothing more likely to aid credulity than unconscious absorption of information.

Interestingly, the runner-up to Baskerville was Computer Modern, a font designed for the American Mathematical Society and its TeX typesetting program. It’s widely used in scholarly publishing because of its compatibility with TeX. A a modern typeface, it has perfectly vertical axes and strong contrast between thicks and thins. By comparison, Baskerville is a transitional font, bridging between old style and modern, with some counter offsets and less contrast between stroke weights.

This raises to me a broader question about how type contributes to the authority and credulity of scientific information. Every journal redesign I’ve done has included a desire to change some aspects of the typography, but we could never drift far from an authoritative font. Those with typographic training and experience knew why, but the feeling was universal — there was something “wrong” if the font drifted too far from traditional letterforms and into something conveying more art than information.

The power of type is nicely illustrated in this section from the film Helvetica. In this clip, Michael Beirut (who I was lucky enough to work closely with on one journal redesign) describes how Helvetica itself injected authority and power into corporate communications, launching the current era of corporateness:


Is this typographic power part of why the PDF is still so important to our readers and users?

Look at the typography of the HTML pages of most sites, and you see major deficiencies. It’s often sans serif, with long lines and poor leading, and set in sizes and weights that have no rhyme or reason — or it’s Times Roman with compromised screen contrasts and strange kerning. It’s no wonder readers want the PDF — there, the journal truly manifests itself, in typography, page layout, and sensible allocations of information in an understandable aesthetic.

While Morris believes he tested typefaces, he was actually testing a number of factors that stream from the faces he chose and their setting — kerning, leading, line set, type size, and so forth. He conducted his test on backlit monitors. Would the outcome have changed if paper had been the substrate?

The aesthetics of authority are delicate. Push too far into the adventurous, and you can lose it. Stay too long in an outdated mode of representing authority, and you look irrelevant or stodgy.

As this little experiment shows, even the letterforms you choose could have an impact on how much authority your words carry.

(Hat tip to PMD for the Helvetica clip and to SW’s Twitter feed for the pointer.)

Enhanced by Zemanta
Kent Anderson

Kent Anderson

Kent Anderson is the CEO of RedLink and RedLink Network, a past-President of SSP, and the founder of the Scholarly Kitchen. He has worked as Publisher at AAAS/Science, CEO/Publisher of JBJS, Inc., a publishing executive at the Massachusetts Medical Society, Publishing Director of the New England Journal of Medicine, and Director of Medical Journals at the American Academy of Pediatrics. Opinions on social media or blogs are his own.


39 Thoughts on "The Typography of Authority — Do Fonts Affect How People Accept Information?"

“Th Elements of Typographic Style” is a good starting point. “Just My Type” is a more recent book. I have more, and will add to this once I get to my bookshelves.

Bringhurst’s work is both comprehensive and a pleasurable read. Truly, the Bible of Typography and the text I preach from. Great article! I’ll be sharing this with my type students. As Bringhurst puts it, the purpose of the skilled typographer is to achieve “creative non-interference” by choosing type that embodies “statuesque transparency” – love the comparative type analysis that validates these all-important concepts. Great article!

Other good starting points — “A Tally of Types” (Morrison), “Basic Typography: Design with Letters,” and “The Visible Word” (Spencer).

“The Elements of Typographic Style” by Robert Bringhurst is another classic. And if you really want to get into the stories of various fonts and font families, I’d suggest Alexander Lawson’s “Anatomy of a Typeface.”

On a related subject–book design–I’d suggest Richard Hendel’s “On Book Design” and Adrian Wilson’s “The Design of Books” as good places to start.

And apologies for the redundancy on the Bringhurst recommendation–I so reflexively think of it as “The Bringhurst” that I breezed right over the title! So take this as a second to Kent’s recommendation in the first place.

I have written my bachelor’s thesis on this theme so I have gathered lots of useful articles and books. I would really recommend to start with Simon Garfield’s “Just My Type” – it’s an easy to read and really interesting book on the topic, there is a good chance you will fall in love with typography after reading it. To continue – Eva Brumberger has written a beautiful article called “The persona of typeface and text” with lots of examples and links to the other authors. More complex information can be found in the articles of Theo van Leeuwen and Hartmut Stoeckl

Very nice piece. As a librarian working with e-journals in their many iterations, I’ve added “context is king” to my long-held “content is king” belief. You’ve given an elegant explanation of how the typographical element provides context. In the journal world, so does the structure of a journal issue. An article in clear pdf loses some meaning when presented in isolation from the rest of the journal.

I was a little disappointed that Morris didn’t discuss genre, as that seems to be the obvious direction to go—i.e., that certain typefaces or typographic styles are associated with certain types of documents and that readers’ perceptions are influenced by how well those conventions are followed or transgressed. Here’s an interesting article that discusses this more in depth, although it is more focused on page layout arrangements than typeface choice: http://www.aaai.org/Papers/Symposia/Fall/1999/FS-99-04/FS99-04-002.pdf.

Thanks for drawing attention to the much neglected area of typography. As you have alluded, choosing the font is the easy part. Just as important is line breaking and page breaking algorithms, knowing which dash to use (hyphen, en-dash, em-dash, minus), subtle uses of space, etc. I would like to conjecture that the reason PDFs are regarded as more authoritative than html pages is that the page make-up engines that are used to make a conventional page are more sophisticated than those used to create an html page. It is quite possible to have better line breaking engines for html and I think that will give it more authority.


Thanks for this post and the links. Design is a woefully underappreciated aspect of scholarly publishing. So many researchers see it as a question of aesthetics rather than a question of functionality and fail to understand why it’s so important. You particularly see this attitude expressed in the poor choices so many researchers make in designing their slides when they give talks, so often making their slides incomprehensible and unreadable, thus undermining the credibility of their research (really CERN? Comic Sans?).

As a species, we’ve been doing art a lot longer than we’ve been doing formalized science. Just as there are clear protocols for experimental techniques, there are also clear protocols for visual design. Some things work and some things don’t work well at all (try autoclaving bleach or using a yellow font on a white background).

When a researcher is giving a talk or writing a paper (and designing figures for either), they’re essentially telling the story of their research. Those who understand the techniques of art and can effectively visually present their data (see the works of Edward Tufte as a starting point) have a leg up on those who present things in a thoughtless and often unclear manner (your recent rant about chart junk as a good example).

Given the incredible competition that researchers face and the importance of reputation and credibility in that market, one would think they’d be quicker to take advantage of something so helpful to success.

Beatrice Warde made this very point in her Crystal Goblet essay (1930) —

“Printing demands a humility of mind, for the lack of which many of the fine arts are even now floundering in self-conscious and maudlin experiments. There is nothing simple or dull in achieving the transparent page. Vulgar ostentation is twice as easy as discipline. When you realise that ugly typography never effaces itself, you will be able to capture beauty as the wise men capture happiness by aiming at something else.”


To augment Greg’s great Beatrice Warde’s quote, here’s a favorite of mine from Stanley Morison (designer of the Times New Roman, which is only overused because it is so good) along the same lines (and btw when he refers to “printer” it is not inappropriate to read “typographer”):

The primary aim of printing is not to be an art, but to be the most responsible of our social, industrial, and intellectual mechanisms; it must, like a transport system, be most disciplined, most rational. Nevertheless, if it is allowable to define art, in this connection, as the application of knowledge, reason, and skill to the service of writers and readers, it may not be rash to hope that some of the past, present, and future productions of the printer will, as multiplied productions of reason and skill, be counted worthy to rank as an Aspect of Art.

Also a good example of a statement that is both grandiose and self-effacing at the same time. 😉

An interesting post.

But you may also want to wind the clock back a little further and remember Eric Gill (Gill Sans and Perpetua) who, in his Essay On Typography (http://books.google.co.uk/books/about/An_Essay_on_Typography.html?id=44Yq6UplAbAC) wrote that ‘Letters are things, not pictures of things, A is A and B is B’.

Gill believed that purity in letterforms, and the fact they have no referents, conveyed the ultimate authority, that of God. Maybe ultimately misguided, but a great polemic on the vitality of type…

And don’t forget Gill’s mentor Edward Johnston (http://www.ejf.org.uk/)

Gill had a point — there is an archetype in every letterform. However, as we all know, you’re going to look out of place if you wear a swimsuit to church.

There’s one constant with typographers — they love to talk about type. There are so many books about typography, it’s amazing there are letters left for the rest of us.

I was very surprised that Computer Modern fared so well. I have to look at far too much of it, and I think it is a horrifically ugly and badly proportioned font that looks like it was, well, designed by a computer scientist.
A lot of newer fonts–such as most of the iOS library–are also ugly, so perhaps the preference for Computer Modern can be attributed to deteriorating aesthetic standards caused by the widespread use of electronic media, in the same way MP3 is now regarded as an acceptable format for audio.

Nice, but raises a question: “Does a broad audience have greater trust for content set in Baskerville because most of what we’ve *read* in Baskerville is serious stuff?” And vice versa for Comic Sans.

The question “What prompts a printer / designer to choose Baskerville rather than Comic Sans or other for publishing serious works?” remains. But the printer’s choice may condition trust for the rest of us based on the seriousness of what we’ve previously read in Baskerville rather than deep perceptual cues.

Kent, I’m not sure I follow how randomizing which typefaces were displayed speaks to whether a typeface has preexisting semiotic baggage that influences how the reader will read it.

I don’t think that this one occurrence of Comic Sans by CERN largely affected their credibility since they already had so much cultural capital within the scientific community and this is pretty small. However, the fact that they were largely ridiculed for it does point to a common and widespread perception that serious science is not supposed to happen in Comic Sans. And I think that’s the point Greg is getting at.

If the same message in a different typeface generates significantly different credulity scores, then you might assume that the typeface affected credulity.

Yes, I think Greg, you, and I agree on this. The question Greg was asking is why the typefaces garners different reactions. Morris seems to think that it is something intrinsic in the design of the different typefaces, but Greg is suggesting that it might be that readers are used to seeing credible material presented in Baskerville and so when presented with something else in Baskerville, are then more inclined to feel that it, too, is credible.

I think typefaces are a blend of the two — they take advantage of certain intrinsic aspects of vision and reading (serifs aid the establishment of a baseline and encourage more fluid reading ala cursive; counters provide visual contrast that aids recognition) and culture (we learn to associate certain fonts with high-class venues and others with cheaper or childlike venues).

An interesting inverse experiment in this regard is to look at typefaces not from your culture. Kanji, for instance, can look cheap, computerized, or elegant even though you don’t understand it, have no appreciation for stroke weight or order, etc. In these cases, we infer certain things from the aesthetic, but not being fluent, we can’t take advantage of the way fonts help the eye.

Kent, I note that in your article where you mention the typefaces you do so in their specific typeface [format]. But Baskerville is in a ore modern sans serif Trebuchet like typeface. The only Baskerville font i can find is more akin to Times New Roman.
Is there a Baskerville font version that looks like the type you’ve used in your article?

I wasn’t very clear – or I don’t see how shuffling fonts in this experiment text helps. My question is: Assume that over my lifetime I’ve read a lot of scientific journals, newspapers (many in Baskerville, Times, etc), and comic books, ads, and less serious stuff (many in Comic Sans or similar typefaces).

To what extend does my lifetime experience reading “serious” stuff set in Baskerville condition my expectation that a sample text block typeset in Baskerville is more credible than a typeface that’s not frequently chosen for serious stuff.

Or, how much of the (fairly small) effect is due to conditioning based on what I’ve read in my lifetime, how much is due to differences in perception based on design artifacts of the different typefaces?

As you said: “Every journal redesign I’ve done has included a desire to change some aspects of the typography, but we could never drift far from an authoritative font.”

A much simpler example of how type and message blend struck me — the use of ALL CAPS. In legal agreements, it’s used supposedly for emphasis but actually has the effect of obscuring information. But in email, it’s universally (at least for English-speaking people) equated with shouting. I once worked with someone who would sign her emails with her name in all caps, and it always looked like she was shouting her name. So, even within typefaces, there are ways to modulate the voicing, from the SHOUT of all caps to the whisper of e. e. cummings.


You may find the article that I mentioned above interesting. The author discusses page layout arrangement more than typeface choice, but he talks about how readers come to expect certain graphic design features to be associated with certain sorts of written objects. He does a very brief meta-analysis of the layout of the conference proceedings in which the article appears: “when I opened the AAAI template it immediately became apparent from its layout that it was of the genre ‘academic paper.’” (http://www.aaai.org/Papers/Symposia/Fall/1999/FS-99-04/FS99-04-002.pdf)

I agree that typeface could certainly be one of the markers of “serious” written material.

Thanks Joel That’s what I was trying to ask – not very clearly.

This type preference bit might start early. In my first year of reading I came on a book set in a sans serif typeface. I remember thinking, “there’s something wrong with this book.” I didn’t trust it. You’ve revived a memory from 50+ years ago, so thanks for that, wherever this argument lands.

Comments are closed.