Text-level semantics: italics (i), emphasis (em), bold (b) and strong (strong) in EPUB3

Following Colleen Cunningham's tweet yesterday and the discussion that followed, I thought it worth examining the differences between the text-level semantics of italics (i), emphasis (em), bold (b) and strong (strong) in EPUB3.

italics and em (and cite)

Italics are the most common form of emphasis in books (outside of manuals and guides) and are used for a number of reasons. These are the main ones:

  1. to identify a book title <cite> [update: thanks to @indiamos]
  2. to identify a foreign word <i>
  3. in place of quotation marks to identify a keyword <i>
  4. to identify that the word should be stressed when pronounced <em>

Next to each of the scenarios, I have placed the relevant tag that you would use. And it is logical that EPUB breaks italics into <cite>, <i> and <em> because, aside from any desire to vary the styles applied to these elements, text-to-speech interpreters then know where emphasis should be placed and where it should not. And we also have reference points for when books are being referred to.

Note: To write that EPUB breaks italics into two isn't a very exact use of language, because in actual fact text marked <em> and <cite> needn't be displayed in italics at all, it could be red or in a box, and so on, but I'm sure you understanding the meaning here.

bold and strong

Bold text can be used in books in place of italics or to alert the reader to something. The <strong> tag, similar to the <em> tag, enables us to distinguish between these two uses. And a paragraph beginning in the following way would be a perfect example of the latter use (i.e. the alert or warning):

<strong>Warning</strong>: you really need to know this fact otherwise everything you do will be wrong.

Again, text-to-speech interpreters need to know the difference between bold and strong so that the correct emphasis can be added.


[update: quotation added and text amended following comment from @indiamo on twitter]

Where italics and bold are used in headings as the overall style, the EPUB guidelines are to specify this in the CSS. And this is entirely sensible.

The question is: what happens when you have italics, emphasis, bold or strong within an italicised or emboldened title style? The answer is that where italics or bold exist and repeat the style already applied to the whole of the heading as a 'paragraph-level' style, then you should note the following:
Where headings or running heads are in small capitals or italic, foreign words and titles of books etc. - which would normally be in italic - are sometimes put in quotes. This tends to give foreign words a self-conscious look; and it may be better to omit the quotes even though the word is italic in the text. In some books, words that are italicised in the text are set roman within italic headings. (Judith Butcher, Copy-editing, 3rd edn., p. 142) 
As regards <em> and <strong>, the good thing about them is that we can simply add them to heading text as we would do to any text.

Ongoing debate

[update: section added, following discussion with @dvsch and @indiamo]

As with all things, there is room for debate. There are those who would want to semantically identify all elements and never use styles such as italics and bold at all. But I would want to make two points in relation to this:

  1. as we see from the changing meanings applied to elements and how they are rendered between HTML versions, semantic meanings are not fixed - and so using them might mean future revisions are necessary.
  2. in print the semantics for something being italic is all conveyed through grammar, convention and context. So every semantic variant and addition that needs to be tagged is an additional piece of work for the person creating the eBook. And whether or not it is economically sensible to spend money on these invisibles is the main thing to debate within companies.
Yes I am being good when I use cite, but it might mean that I need to set and apply a separate InDesign character style in order to create a workflow that outputs ebooks using the cite tag. All this requires time, money, education, and so on.

Further reading

Text-level semantics (HTML5, W3C)

Bolding and italics (EPUB3, IDPF)

Endorse on Coderwall