Semantic inflection is a semantic markup technique that involves the layering extra information about the role that elements serve in markup languages such as HTML.

Although markup elements often have implied semantics (e.g., <table> for tables) many are generic in nature (e.g., <div> for a block of content or <span> for a run of inline text). Semantic inflection is the process of adding more specific information to these elements using attributes like epub:type.

For example, the epub:type attribute could be used to indicate that a section contains a book's frontmatter like this: <section epub:type="frontmatter">. (See the EPUB Structural Semantics Vocabulary for more information.)

Although the ARIA role attribute is used to provide additional context, it plays a more important function in making documents accessible for assistive technologies. It is not a general purpose semantic inflection mechanism, as the roles that can be used, and where they can be used, are restricted. The misapplication of roles can negatively affect the readability of documents.

In markup languages like HTML, semantic markup refers to the use of tagging that identifies the purpose of the content.

Sometimes the element names alone describe the nature of the content (i.e., the purpose is implied by the name, as with table and article). Other times, additional semantics are added to elements using semantic inflection techniques to make the nature explicit or more precise.

The use of semantic markup allows assistive technologies to provide more intelligent navigation of the content (e.g., by making content escapable and skippable).

Skippable refers to the ability for non-visual readers to skip past blocks of content they are not interested in.

A sidebar is an example of a skippable structure. While users might be interested in reading these, they usually want to be able to skip them while reading the primary narrative and return to them later.

The ability to skip structures depends heavily on their being identifiable in the markup (i.e., through the use of semantic markup).

The Synchronized Media Integration Language (SMIL, pronounced smile) is an XML grammar for describing interactive multimedia presentations.

A subset of SMIL is used in both the DAISY and EPUB formats to provide text and audio playback synchonization.

Refer to the SMIL specification for more information.

The spine is the part of the EPUB package document that defines the reading order of the content in the publication.

Its name was chosen because it acts like the spine of a bound publication — holding the publication together in correct order.

Scalable Vector Graphics (SVG) is an XML grammar for describing two-dimensional vector and mixed vector/raster graphics.

SVG images provide a number of accessibility benefits over other formats, beyond just their scalability: text content can be represnted as text, making it available to assistive technologies; image components can be titled and include descriptions of their purposes; and ARIA roles, states and properties can be used to make interactive elements accessible.

Refer to the SVG specification for more information.

SVG content document is a term used in EPUB 3 to refer to any SVG image that conforms to the requirements of that format. EPUB 3 has rules about which version of SVG to use as well as a number of restrictions on the elements that can be used.

Refer to the definition of SVG Content Documents in the EPUB 3 specification for more information.