As with any other TEI text, the actual contents of drama texts are encoded in body, possibly preceded by front matter in front and succeeded by back matter in back. Inside body, the actual text often has specific structures: acts and scenes. Like parts and chapters in prose works, these can be encoded using the generic TEI element div (division). In order to identify the nature of the text divisions, they can be characterised in a type attribute. For text divisions in drama texts, the values act and scene can be used, although you can use your own typology if so desired. For our example text, the major structural divisions can be encoded as follows:

Of course, this markup is far from complete, and even invalid TEI, as the TEI Guidelines stipulate that div text divisions cannot contain unstructured text. We’ll fill in the blanks in the next sections of this tutorial.

One of the most distinctive organisational principles in drama texts is the grouping of text lines in utterances, spoken by the respective characters. Each line of speech can be encoded in a sp (speech) element. When present in the text source, the indications of the speaking character can be encoded in a speaker element. The actual text of a speech line should be further encoded as a paragraph (p) or verse line (l, possibly grouped in lg). If the encoder is undecided about the prose or verse qualities of the text lines, ab (anonymous block) may also be used.A dramatic speech line (sp) cannot contain unstructured text: it should be encoded further as paragraph (p), verse line (l, possibly grouped in lg), or genre-unspecified anonymous block (ab).

For example, the last three speech lines in the example can be encoded as follows:

Notice how in this example, the colons after the name labels have been left out of the transcription. This is an editorial choice that can be motivated by the fact that the colons themselves are purely typographical separators between the labels and the text, that can systematically be reproduced when this transcription is rendered. Likewise, the blank line between both groups of speeches has been omitted as a purely typographic feature. Typographical line breaks can of course be encoded where they occur with empty lb elements.

The sp element provides a more formal mechanism for indicating the speaking character, by means of a who attribute. Its value points to a definition of this character elsewhere in the transcription. The value of this definition’s xml:id attribute is used in the who attribute, preceded with a hash character (#), in order to indicate it as the identifier part of a formal URI reference. For example:

Notice how this example assumes that both characters are identified elsewhere in the transcription, with comp and user as respective values for their xml:id attribute. A common place for such definitions in drama texts is in a cast list, introducing all characters of the play. For the encoding of cast lists, see in this tutorial module. When text is spoken by more than one character, the identification codes for all of them can be listed in the who attribute, separated by white space.

Of course, the encoder is free to choose which identification mechanism will be used for the speakers in the transcription. When the markers are present in the text source, their bare transcription in speaker (see ) may suffice. It may make equal sense to use both systems (transcription of the actual speaker labels in the source text in speaker, and a formal identification in the who attribute), as in . If the source text has no explicit markers of the speaking characters, or if these markers are present but not deemed relevant to the transcription (and could be reconstructed systematically when rendering the transcription), only the formal identification in the who may be used:It is important to point out the different nature of both identification mechanisms for speakers: speaker can only contain a literal transcription of a speaker label in the source text; who can only contain an abstract identification code provided by the encoder.

The encoding of all speeches in our example text would still leave some text structures uncovered: all fragments enclosed in parentheses or square brackets. These text fragments are not meant to be spoken in the performance, but rather convey directions on aspects of the performance. Such stage directions can be transcribed with the stage element. Stage directions can address aspects of the acting, technical circumstances, locations, music, and so on. In order to distinguish between these different types of stage directions, the type attribute can be used on the stage element. The encoder is free to develop an own typology for the kinds of stage directions.

Stage directions can occur in the transcription wherever they appear in the source text. In our example, the stage directions (between parentheses or square brackets) can be encoded as follows:

The TEI Guidelines provide a specialised element for indicating movements of characters: the empty move element. It may be used either to complement implicit movement descriptions in stage directions, or to formally document movement of characters when it is absent from the stage directions. Being an empty element, all information is conveyed in specific attributes. The type attribute can be used to categorise the movement (for example: entrance and exit for stage entrance and exit, respectively; or onStage for movement on stage). The moving character(s) can be indicated in the who attribute, as a white space separated list of formal codes with which they are identified elsewhere in the transcription (see ). The where attribute may be used to specify the direction of the movement. Any system of location information may be used (for example, C for centre, L for stage left, R for stage right), as long as it is identified in the header of the transcription.

In our example, the encoding of the entrance and exit of the user character could be formally enriched with the move element:

It is equally possible to distinguish technical stage directions from general stage directions, using specialised elements. Directions regarding sounds can be encoded using the sound element. The type element can indicate what kind of sound is meant, while the discrete attribute can specify whether the sound overlaps with the surrounding speeches (false), or not (true). For example:

Descriptions of the contents of a screen (usual in transcriptions of screen plays) can be encoded in a view element. In our example, this could apply to the description of both screens on the stage:

Specific camera directions can be encoded using the camera element, whose type attribute can specify the direction:

Directions about text to be displayed on screen (usual in transcriptions of screen plays) can be encoded in a caption element. This could be applied to the directions on both screens on the stage:

Other technical stage directions, not meant for the actors can be encoded with the tech element, with a type attribute for classifying the stage direction. The type attribute can be used for a typology of such technical directions: light (a lighting instruction), sound (a sound instruction), prop (a direction for props), or block (a blocking cue, i.e., related to position and movement on stage).