← All method guides

Method foundations

How to Extract a Reusable Sentence Frame from an Example

Finding a useful sentence is easy. Turning it into a good learning object is harder. If you keep every word, you have a phrase to memorise. If you replace every interesting word with a blank, you have a skeleton that may generate nonsense. Metkagram treats Frame extraction as a constrained editorial step: preserve what carries the relationship, open only the parts that can really vary, then test the result against new examples.

First identify what made the sentence worth saving

Ask what the speaker is doing. Are they correcting a premise, adding a condition, softening certainty, contrasting two causes or buying time? That communicative Move tells you which relationship matters. A sentence can contain several grammatical facts, but they do not all need to become the target of the same Frame.

This protects the extraction from becoming a full parse disguised as a pattern. The learner needs a useful decision, not every property of the sentence at once.

Keep fixed material that carries the construction

Open slots only where natural substitutions are possible. Fixed connective language, required prepositions, clause order or verb forms may be the very thing the learner needs to keep. Replacing them with abstract labels can remove the signal you were trying to teach.

Then make two or three variations. If the Frame survives changes of topic and vocabulary while preserving its meaning and constraints, the abstraction is probably useful. If every variation requires repair, narrow it.

Store the example and the abstraction together

A Frame without examples can become cryptic. An example without a Frame can remain a one-off phrase. Keeping both lets the learner move between concrete language and reusable structure. It also gives reviewers evidence for why the Frame was defined in that particular way.

Metkagram therefore treats extraction as reversible: you should always be able to return from the abstract Frame to natural sentences that demonstrate it.