One exchange (a single “speech”) in a real or imaginary conversation between two or more entities, for example, between a an interviewer and the person being interviewed, between a nurse (or doctor) and a patient, between a person and a computer, etc. Each time a new speaker takes over, a new <speech> starts, which names the speaker (<speaker>) and then contains one or more paragraphs (<p>) that hold what speaker said.


In this Suite, a speech is not part of any particular larger element structure; a speech is just one identified fragment of the whole conversation.

A <speech> is modeled to name the speaker, followed by a minimum of one full paragraph to contain the speech’s text, even if what is spoken is only a few words, for example:


In the circumstance in which many voices are heard as one, for example, “All the Kings Men” or “Tom and Jerry”, the combination is considered to be a single speaker.


content-type Type of Content
id Identifier
specific-use Specific Use
xml:lang Language

Related Elements

A <speech> is a container element that names the person, object, or group speaking (<speaker>), followed by one complete utterance, modeled as one or more paragraphs.

Content Model

<!ELEMENT  speech       %speech-model;                               >

Expanded Content Model

(speaker, (p)+)


The following, in order:

This element may be contained in:

<app>, <bio>, <body>, <boxed-text>, <fig>, <license-p>, <named-content>, <p>, <sec>, <styled-content>, <table-wrap>


<p>The participants understood the purpose of their peer 
response groups to be finding mistakes or problems in each 
other&rsquo;s essays. ...Clara, one of the Chinese-speakers, 
explains why she no longer believes the initial positive comments:
<p>I think Aeenoy start this way. I think she always do 
this way, like say some good thing first. And then I know 
the bad thing is coming.</p>
<p>So, why doe she do that?</p>
<p>I think it gives somebody self-esteem ...</p></speech>