Data Taxonomy for Pain Language

A compact annotation taxonomy capturing pain qualities, metaphor domains, entailments, and relational structure.

Pain descriptions are linguistically rich but structurally inconsistent. Patients describe sensations as burning, stabbing, heavy, or like being attacked. Without a controlled taxonomy, these expressions remain difficult to compare, annotate, and operationalise.

Key idea

A compact taxonomy improves annotation agreement, reduces QA friction, and enables downstream modelling.

What Is a Taxonomy?

In annotation workflows, a taxonomy functions as a constrained menu of categories. Rather than inferring meaning ad hoc, annotators select from defined options, preserving interpretive consistency.

Reduced ambiguity during annotation
Improved inter-annotator agreement
Structured outputs suitable for modelling

Core Category Structure

1 · Pain Qualities

Descriptors of sensation: burning, stabbing, throbbing, dull.

2 · Body Location

Spatial grounding: pelvis, lower back, abdomen, knee.

3 · Intensity Markers

Severity indicators: mild, severe, unbearable.

4 · Figurative Domains

Metaphorical mappings: knife, hammer, fire, beast, weight.

5 · Temporal Markers

Duration and recurrence: constant, comes and goes, worse at night.

Metaphor Entailments

Beyond surface tagging, metaphors carry structured entailments — conceptual implications that shape how pain is experienced and interpreted.

Violence / Attack
→ External agent
→ Victim positioning
→ Aggression / relentlessness

Weight / Pressure
→ Constancy
→ Immobilisation
→ Oppression

Heat / Fire
→ Escalation
→ Spread
→ Consumption

Entailments provide a second modelling layer, enabling higher-level pain profiles within rule-based systems.

Relational Encoding

Pain expressions rarely occur in isolation. Categories interact.

Intensity modifies quality: “sharp pain”
Location grounds metaphor: “burning knife in my knee”
Temporal markers shift severity: “worse at night”

Encoding relations maintains fidelity to natural patient language and strengthens modelling robustness.

Why Compact Design Matters

Oversized schemas reduce annotation efficiency and increase disagreement. A lean taxonomy:

Shortens annotator training time
Simplifies quality assurance
Supports scalable NLP integration

Conclusion

A compact taxonomy transforms narrative pain descriptions into structured, reproducible data. It functions as infrastructure: enabling agreement at annotation level and interpretability at application level.

When aligned with rule-based modelling, such taxonomies provide a transparent alternative to opaque language processing systems.