Superscripted letters, as they are typically used in modern texts, need to be distinguished from a class of characters which appears in very old texts whose typography attempts to imitate manscript conventions. This class of characters includes composite characters, often consisting of a letter with a small letter or letters attached to it. These may be directly over the main letter, or they may be in the position of a conventional modern superscript. The encoding of these letters is discussed in the entry on “Special characters: Brevigraphs and diacritical marks”.
Conventional superscripts, on the other hand, consist of a small version of an ordinary letter, positioned close to the top of the larger letters it accompanies. We encode this kind of superscripted letter with <hi>, using the sup() keyword of the rend= attribute:
Note that the argument (which goes in between the parentheses after sup) is left blank. It could be used to indicate the exact amount of superscription, but the WWP does not use this feature.
There are certain cases where it is not possible to use the rend= attribute to capture superscription: for instance, when a footnote number is superscripted, since the number is captured as part of the rend= attribute, it cannot itself have a rend= attribute (see the example below). Instead, the superscripted number is captured as an entity reference (&sup-1;, ², etc.).