Postcodes ([+ ...])

Status: Reference Last modified: 2026-10-20 (commit f7e32e22)

A postcode attaches a researcher or analysis tag to an utterance as a whole. It appears on the main tier after the terminator, not on an individual word. Adding a tag does not change the utterance’s words or itself perform an analysis operation.

Syntax and Scope

The reference transcript corpus/reference/content/postcodes-and-freecodes.cha includes this example:

*CHI:	hello . [+ bch] [+ foo]

Multiple postcodes are stored in source order. Preserve that order unless a reviewed operation explicitly changes it; an open vocabulary does not establish that every consumer treats ordering as irrelevant.

The payload is text, not a list of CHAT words. That does not mean arbitrary bytes are accepted: the grammar requires content after [+ and excludes a closing bracket or newline from that content. Parser admission also rejects an empty payload after trimming trailing whitespace. A recognized token or a programmatically constructed Postcode does not certify the whole document.

Common Postcodes

The vocabulary is open-ended. The following conventional labels are also offered by the editor’s completion feature; they are suggestions, not a whitelist:

Consult the corpus protocol and the analysis tool’s options before interpreting or changing a tag. Chatter does not automatically exclude an utterance from every downstream analysis because its postcode is exc, nor does an unfamiliar tag such as foo become invalid solely because it is absent from this table.

This guide does not maintain a corpus-frequency census. Historical occurrence counts are not a vocabulary definition, a current prevalence measurement or evidence that an unobserved parser path is unreachable.

What Postcodes Are NOT

These forms have distinct grammar and validation rules. Free text is not a blanket validation exemption, and selecting another annotation to avoid a diagnostic can change the transcript’s meaning. See Replacements for the word-level alternative and Annotations for scoped forms.

A comment that says “exclude this” is not automatically an utterance-exclusion postcode. Conversely, a postcode does not provide a word-level attachment for a consumer that needs to know which word was annotated.

Not Postcodes: Quotation Markers

Quotation terminators and linkers such as +"/., +"/ and +" are not [+ ...] tags. They have separate grammar and model owners. See Utterances → Terminators rather than interpreting every plus-prefixed form as a postcode.

Position in the AST

The utterance’s main.content.postcodes is a TierPostcodes collection of Postcode values. A postcode owns its text and source span; it is not a ContentAnnotation on a word. Word walkers therefore are not the interface for reading or changing the postcode list.

Dependent tiers have their own payload types and grammars. A bracket-shaped string there is not automatically main-tier postcode metadata: a text tier may retain it as text, while another tier’s grammar may reject it. Consult the CHECK assessment for the documented diagnostic boundary; do not add a second text parser to imitate another tool’s error code.

The model serializer writes [+ , the stored payload and ]. The primary parser trims trailing payload whitespace, so a parse/serialize cycle is not a promise of byte-for-byte source preservation. Keep the original when exact authored spelling or formatting matters.

Tooling Rules

• Add or inspect utterance tags through main.content.postcodes, not the word-level annotation list.
• Preserve unknown values and their order during unrelated transformations. A supported constructor or serializer is not authority to reinterpret a tag.
• Validate the complete candidate after an intentional edit, and review the changed tag against the corpus’s coding protocol. Validation cannot decide whether an utterance should be excluded from a particular research analysis.
• Completion replaces the selected postcode with an explicit source range; the user still chooses whether the offered tag is appropriate.

References

• CHAT manual: Postcodes
• Model and serializer
• Primary parser admission
• Grammar: postcode and annotation_content

This page last changed: 2026-10-20 (commit f7e32e22). The whole book last changed: 2026-10-27 (commit 5bfee2f9).