---

6.1 A Practical Model Of Presentation ¶

It helps to separate four questions that are often mixed together:

Layer	Question	Example
Meaning	What happened or what is this?	A level-two heading was reached during navigation.
Presentation	What should convey that meaning?	Speak the heading in a named voice and play a section cue first.
Resources	Which voice and sound realize that request?	Resolve the logical voice through the active palette and the cue through the active sound pack.
Output	What can this machine produce?	Route speech to an installed engine and apply supported spatial behavior.

An Emacs mode or Emacsvox integration supplies facts about an object or event. Facts can describe a role such as a heading, an event such as completion, a state such as unread, or an attribute such as heading level. Context also says whether you are navigating, reading continuously, or receiving a state change. This allows the same object to receive concise feedback during continuous reading and more orientation when you move to it deliberately.

Presentation rules match those facts and context. A rule can choose a voice, add speech or a cue, insert a pause or tone, request spatial placement, or suppress something that is unhelpful. Several matching rules can compose; a stronger, narrower rule need change only the relevant part rather than repeat the complete presentation.

The user-visible configuration layers are:

1. A fixed internal baseline that preserves established Emacsvox behavior.
2. Automatic integration behavior supplied by the active mode or package.
3. Enabled Presentation Options that add optional behavior in a stable weakest-to-strongest order.
4. Personal, session, and current-buffer overrides. These are the strongest layers and are intended for precise changes.

A presentation profile is not another rule layer. It records a complete set of selected options, sound pack, voice palette, and spatial choices so you can switch those choices together.