Knowing when research is delaying the draft

Name the missing decision

Before opening another source, write what you need to learn and what the answer will let you write. 'Research railways' is too broad to finish. 'Find whether this character could reach the town before evening in the chosen year' creates a bounded task. Once the travel possibility is clear, you can draft the scene and record any remaining uncertainty.

Separate facts that control the plot from details that can wait. A journey time may determine whether two characters can meet. The exact colour of a carriage interior may be replaceable later. Both can be interesting, but they do not deserve equal priority when the draft is stalled.

For nonfiction, identify the claim that needs support rather than collecting material around a general subject. Record what evidence would challenge your current interpretation as well as what might confirm it. A focused research question should narrow the task without predetermining the answer.

Leave a visible boundary

Set a stopping condition before you begin. You might stop after locating an authoritative timetable, comparing two conflicting accounts or establishing that no reliable answer is readily available. A time limit can help with exploratory searching, but it does not make an unresolved factual claim safe to publish. It tells you when to return to drafting with a clearly marked gap.

Use a distinctive placeholder such as '[CHECK: journey time, June 1932]' in the manuscript. Keep a separate list if there are many gaps, and make sure the list can be reconciled with the draft before publication. A placeholder should preserve uncertainty, not quietly turn a guess into a fact.

At the end of a research session, write a few sentences of the passage the research was meant to support. If you repeatedly cannot do so, inspect the obstacle. You may need a structural decision, a clearer argument or permission from yourself to write an imperfect first version rather than another source.

Research remains part of writing when it changes what you can say. When it only enlarges the collection of interesting material, pause and ask which page that material belongs on. An answer can justify the next search; the absence of an answer is a useful reason to return to the draft.