So far we have only talked about different types of text at the paragraph level (‘this paragraph is a heading, that one is a subheading’), but one can also talk about different types of text within paragraphs. For example, in written books, italics (a presentational attribute) is often used to indicate a number of different meanings (semantics, content): stress emphasis, a title of a work (e.g. a book), a defining instance of a term, an idiomatic phrase in a foreign language, a name of a ship, a taxonomic designation, a technical term, a word when you discuss the word itself, etc.

When it comes to such in-paragraph types of texts, it is again possible to apply presentational attributes (such as italics) directly. For instance, if you use an English-language version of Microsoft Word, Ctrl+I and Ctrl+B will toggle italics and boldface, respectively, at the caret (or to the selection). Or one can use styles to separate content and presentation.

Of course, the second approach is the preferred one in most cases (in theory). Indeed, if you want keywords in a textbook to be in bold orange text, you should create a style for this (named ‘Keyword’, perhaps?). This way

• you do not need to repeat the same formatting commands every time you want to highlight a keyword,

• all keywords are guaranteed to have the same formatting,

• you can easily change the formatting of every keyword in the text at once,

• and you can use a software program to create a list of all keywords in the text.

In the terminology of Microsoft Word, styles that apply to paragraphs are called ‘paragraph styles’ whereas styles that apply to spans of text within paragraphs are called ‘character styles’. The latter cannot contain any information about paragraph settings (like paragraph spacing or indentation). There is also a third kind of style: linked style. Those contain paragraph settings and may be used as paragraph styles, but may also be used as pure character styles by ignoring the paragraph settings.

In the following example, two definitions are marked up using a character style (named ‘Definition’, perhaps?). Three good-looking versions of presentational attributes connected to this style are given.

In addition, more specific (and visually complex) in-paragraph types of text, such as those related to computer code fragments, benefit greatly from the use of character styles. For example, you could create a character style named ‘Computer Code’ and give it a monospaced font and a greyish background:

Some padding would probably make this look better, but unfortunately Word does not support padding at the character level.

Although the general rule is to use styles and never apply formatting directly to the text, it is not unreasonable to make exceptions for some kinds of very simple character-level presentational attributes.

Indeed, in practice, you often do use direct formatting in the form of italics and boldface at the character level instead of using character styles. There are a few reasons for this:

• It is somewhat awkward to set character styles the way Word is designed (compared to pressing Ctrl+I, say). Although it is possible to associate keyboard shortcuts to styles, the method of character style selection is not ideal for rapid typing. Indeed, in order to set the style to emphasis (say), you have to select the text that is to be emphasised, and then select the style. I believe you cannot simply write the text before the emphasis, ‘enter emphasis mode’, enter the emphasised text, ‘leave emphasis mode’, and continue with normal text.

• Although it might make sense to restyle keywords and defining instances of terms (italic, bold, underline, background colour…?), in most cases the need to restyle in-paragraph text types is fairly small. For example, stress emphasis is almost always rendered in italics in books (and text in general). Also, setting this presentational attribute is very simple to do directly in the first place (press Ctrl+I).

• More often than not, a Word document is printed or converted to PDF before it is distributed, and in these cases semantic annotation is lost anyway.

In conclusion:

• Inside paragraphs, you use character styles to mark up special kinds of meaning, like emphases, defining instances of terms, titles of works, keywords, and pieces of computer code.

• However, for the very simplest cases, like stress emphases that should be rendered in italics, you might simply choose to apply the presentational attributes directly. The reason is that the alternative (using a character style) is somewhat awkward and the gain of it fairly small.