The case against banning italics

7 January 2019 | Ted Page

Why a blanket ban on italics for accessibility reasons is likely to do more harm than good

For accessibility reasons, some organisations impose a blanket ban on the use of italics in online copy. It will be argued here that this is a mistake.

Although blocks of italic text can be difficult to read for many people, banning italics in all contexts can strip text of much of its semantic richness, making it less readable and therefore less accessible. For the time being, the judicious use of italics will yield the greatest overall benefits, but in the longer run, the solution is to develop reading systems that allow full text customisation.

The research to date

To date there has been only limited research into this area. One of the most influential studies has been Rello and Baeza-Yates, Good Fonts For Dyslexia, 2013 (PDF, 683KB). (Note: this is an untagged PDF which may be more or less inaccessible to some readers.) This study measured the impact of font type on people with dyslexia. It tested 48 subjects, each of whom was given 12 texts to read with 12 different fonts, 3 of which were italic fonts. One of the study’s conclusions was that, for people with dyslexia, “italic fonts decreased reading performance”.

Only blocks of text were tested

In the present context, the most important point to note about the study is that the 12 texts used in the test were all 60 words in length. In other words, the study was conducted only on blocks of text. But as will be seen, text that is conventionally rendered in italics often comes in short phrases, or even in single words or numerals.

WCAG/WebAIM on the use of italics

There are many respected and trusted accessibility resources that advise against the use of blocks of italic text. For example, WCAG Understanding Guideline 3.1 includes an advisory technique for “avoiding chunks of italic text”. Similarly WebAIM advises as follows: “Do not use italics or bold on long sections of text”, but at the same time “use various stylistic elements (italics, bold, color, brief animation, or differently-styled content) to highlight important content”.

International standards

It should also be noted that many conventions for the use of italics are enshrined in international standards, including BS 5605 (citing and referencing published material), BS ISO 690 (bibliographic references and citations), and ISO 999 (content, organization and presentation of indexes).

What you lose when you ban italics

As detailed in a 2008 post in this blog (Italics), italics give meaning to content in many important ways. These include:

  • emphasis (as opposed to bold for strong emphasis)
  • works (book or document titles, acts of Parliament, legal cases, film, TV and radio programmes, paintings, compositions, ship and aircraft names…)
  • foreign words or phrases (tête-à-tête, faux), or in biology for Latin binomials (homo sapiens)
  • stylised text such as “see” and “see also” index cross-references, or index locators that refer to illustrations, maps or diagrams

In some cases, whether you italicise a word or not can completely change the meaning of a sentence. For example, the sentence “He was looking around for a suitable pension”, with no italics, means “He was looking around for a suitable financial plan for his old age”. However, if you italicise the word “pension” it completely changes the meaning of the sentence to something like “He was looking around for a suitable Continental-style boarding house”.

No consensus amongst readers with dyslexia

Even amongst people with dyslexia there is no consensus that a blanket ban on italics would be helpful. For example, Martin Pitt, in a (May 2017) GitHub forum discussion stated:

Monotonous walls of text are no good for anybody. Rich text is fantastic … [it adds] legibility to paragraphs of copy by sprinkling different font weights (e.g. bold), italics and colour. … Thus practices, or dogma, that outlaw potential for variability in text are just a really bad idea.

The long-term solution: full text customisation

The ideal long-term solution to the problems that some readers may experience with italics is, of course, user choice, by way of full text customisation. But how realistic a proposition is this?

Olaf Drümmer’s 2012 paper, How feasible is text customisation for PDF? offers an interesting overview with respect to PDFs, especially in the light of the then recently published PDF/UA standard.

However, I would very much take issue with the paper’s contention that colour customisation and text reflow in PDF only work well for very simple documents. On the contrary, these important features are very much controllable by proficient document authors and I would contend that a PDF that is not optimised for both cannot truly call itself an accessible PDF.

For more details on why this is so, please see PDF accessibility and reading order on the topic of reflow. With respect to colour customisation please see PDF background colour bug returns in Acrobat/Reader DC which explains one of the principal problems, and The InDesign workaround: paste into which provides the solution.  (Note: fancy background colours are best avoided in MS Word which is a great word processor, but a not-so-hot page layout programme.)

I will leave to others more knowledgeable in the field to comment on the feasibility of full text customisation in web browsers.


There is a general consensus that large blocks of italic text are best avoided. However, a blanket ban on the use of italics will undoubtedly strip some texts of much of their meaning. Document producers therefore need to be free to use their judgment, and bring their experience and knowledge to bear on the question of the appropriate use of italic text in online documentation.

Until such a time that all the major reading systems can offer good levels of text customisation, common sense, compromise, good judgment and suitable levels of document editing and typographic skills seem the way to go. By contrast, a blanket ban on italics is an over-reaction to a problem which will, in many cases, do more harm than good.