The Stemma
Reconstruct a family tree from a curated set of undocumented text witnesses, then argue the strongest case against your own tree.
Due: 30 April 2027 · Weight: 30%
You are given a set of undocumented witnesses from one documented source family. Reconstruct the family tree. Then argue against it.
Due 12:00, Friday 30 April 2027 (Canberra time). About 10–12 hours. The curated pack is released when the week 6 workshop ends, at 14:00 on Wednesday 31 March 2027; this brief is available from the first day of the course.
What you submit#
A proposed tree or partial graph, an evidence table mapping each inferred edge to the shared feature that justifies it, a supported partial archetype with gaps explicitly marked, and a 600-word argument for the strongest alternative tree.
How it is marked#
Your reasoning is marked against the evidence in the pack, not against similarity to a hidden answer. “The evidence does not resolve this edge” can earn full credit where the evidence genuinely does not resolve it.
The pack’s true source and production log are released after the assessment closes, as feedback. A stemma is a family tree of copies, built from the mistakes they share; week 6 teaches the method and its limits.
Extensions, late work and declaration of tools: see the course policies.
The spec
The brief poses the problem and leaves room for a response. The spec is the fixed part: what a marker checks. Some lines can be checked mechanically; the rest call for judgement.
- every proposed edge names the specific shared feature that supports it
- portions you cannot resolve are marked as unresolved rather than guessed
- you present the strongest competing tree the same evidence allows
How it is marked
| Criterion | Weight |
|---|---|
| Evidence for relationships | 35% |
| Reconstruction within evidence limits | 20% |
| Competing explanation | 25% |
| Confidence and limitations | 20% |