Coding Purpose
For each level there is an introduction which briefly states the scope
and the application/purpose of the level, i.e. what phenomenological aspect
of language or communication is described.
Existing Schemes
This section gives a summary of the discussions in MATE Deliverable
D1.1, reporting what schemes have been looked at and have been taken into
account on the phenomenological level and in terms of markup. There is
a list of schemes looked at a list of phenomena mentioned links (hrefs) to passages in D1.1
Selected Scheme(s)
This section describes what exactly has been chosen as the scheme(s)
to be used for the level. Either one or more existing schemes have been
chosen, parts of different schemes have been selected, or a new scheme
has been developed. In fact, none of the selected schemes existed in the
description format of a coding module or corresponding XML representation
before.
If more than one scheme has been selected - as in Dialogue Acts - the
following structure would appear for each of the schemes described.
Scheme Name
In this sub-section the scheme is described in its overall structure:
Markup Declaration
Here, a hierarchical overview of the phenomena and their embedding structure is given (e.g., phone, tone, break index).Description of Elements
For each phenomenon - which corresponds to an element - within the level (e.g., prosody) the following information is specified.
- Phenomenon name
The name of the phenomenon to be tagged serves as a header of the chapter.
- Description
The phenomenon (element) is described. The selected element name (e.g., <word>) is introduced and described in more or less detail according to the status of the theoretical discussion. Also, the properties of the phenomenon are described for the element.
- Data Source
This section describes what kind of existing information is needed (i.e., what elements can be selected) in order to label phenomena of this type: One may need words for the markup of phrases, one needs utterances or sentences for the markup of dialogue acts etc.
- Segmentation/Selection
What are the criteria important for the segmentation or selection process that that select one or more elements of the existing (base) level of annotation as starting point for the newly to be created element of this level, i.e. what are the selection criteria that make a given sequence of sounds a syllable? Note that in some cases there is a segmentation task. i.e. one segments an existing higher level phenomenon into smaller pieces, or there is a selection task where some elements are chosen to belong to a new kind of phenomenon.
- Assignment
Criteria for selecting attribute values (properties) for the elements are discussed. The attributes and their value ranges are introduced. This information must be provided for each of the attributes a given element can have. Of course, the details given here may vary a lot. This information can also be provided by a kind of decision tree (cf. DAMSL).
- Example
An example is given, starting with a data source example and describing segmented and labelled elements together with their attributes and values.
- Coding Procedure
A recommendation on how to proceed when applying the markup to data. This may also include proposals for software or reliability checks etc. that have proven to be useful.
- Markup Table
In the markup section the element name, the attributes and their values used will be summarized in a table like representation. Please note that there is not a one to one correspondence between the table contents and the underlying XML representation. The table below represents the grid of specification options. In the upper row the element name is specified, in all rows following the element name, for each possible attribute of the given element, the attribute names and the values are listed. The value specification can be represented as the value type (a), as a value set (b) or - in the case of href attributes - as the element class pointed to (c).
<element_name> attribute name value range (a) value type start [FLOAT] age [INTEGER] name [ASCII] (b) value set colour red, green, blue (c) href target href <other_element>
Integrated Example
This section gives an example of the integrated use of all elements, attributes and values described within this coding scheme.
- Coding Procedure
In the individual descriptions of elements and attributes there was a description for recommendations of the coding procedure. In this section general coding procedure guidelines that apply to the whole coding scheme and all of the phenomena covered are provided.
The following chapters describe the level-wise annotation recommendations.