Suppose you open a scene and find a note you wrote the previous evening: “This is awful.” The note has no location, no example and no proposed repair. It tells you how you felt while reading. It does not tell you whether the problem is an unclear speaker, a missing decision or a paragraph that repeats what the dialogue has already shown.
Here, the inner critic is a familiar writing metaphor for the evaluative commentary that can accompany the work. The useful craft question is how to turn that commentary into observations you can act on. You do not have to believe every negative judgement, and you do not have to replace it with a positive judgement before revising.
Begin by identifying the passage the verdict refers to. Read enough surrounding text to understand its purpose, then ask what effect you intended. If the scene should show two flatmates negotiating an unpaid repair bill, can the reader tell who is being asked to pay and why the request is difficult? Those questions are narrower than whether you are good at dialogue.
Separate an observation from an interpretation of your ability. “Both speakers sound alike in these six lines” points to a possible revision. “I cannot write people” makes a claim much larger than the evidence. You can work on the six lines without deciding anything final about yourself as a writer.
Make the observation specific enough to test. Perhaps both flatmates explain the same background facts in complete formal sentences, although each wants a different outcome. One wants payment today; the other wants time to check whether the repair was authorised. Giving their replies those different purposes may produce clearer voices than adding a favourite expression to every line.
A useful editorial note might therefore read: “The exchange explains the bill twice. Let one person ask for payment and the other question who agreed to the work.” That note identifies both a problem and an experiment. It leaves open whether the experiment will succeed, which is appropriate for a draft.
Preserve what already works. If a character keeps flattening the bill on the table while avoiding the amount, that action may carry useful tension. Note it alongside the problem. A review that records only defects can encourage you to remove the very detail that gives the scene its particular life.
You can also decide that a judgement is not yet useful. “Too quiet” might reflect a mismatch between the scene’s purpose and your mood while reading it. Ask whether the quiet scene changes a relationship or a decision. If it does, its scale may be appropriate. If it does not, identify the missing change instead of adding an argument solely to increase volume.
The result of this pass should be a short list of page-level questions. You might ask whether the bill is understandable, whether the speakers pursue different aims and whether the ending changes what happens next. You can answer those questions through reading and revision. A broad verdict about your talent offers no comparable stopping point.
Drafting often involves discovering material whose purpose is not fully clear yet. Evaluation asks whether that material achieves an intended effect. The tasks can overlap, but repeatedly switching between them may leave you polishing an opening before you have discovered the scene’s decision.
Try setting a limited drafting task with a clear endpoint. For the invented flatmate scene, write the conversation until someone proposes a way to settle the bill. Allow the early version to contain ordinary phrasing or a missing descriptive detail. The immediate purpose is to discover the exchange, not to certify every sentence as final.
Here is a rough version:
“The repair cost eighty,” Noor said. “We must pay it because the sink was broken.”
“I know the sink was broken,” Eli said. “But I do not think the repair was necessary at that price.”
Noor felt angry because Eli did not want to pay. Eli felt worried about money.
The scene is understandable at a basic level, but the prose explains the emotional result instead of developing the disagreement. Noor’s objective and Eli’s concern can be made more concrete. Keep the invented amount as part of the example; the scene makes no claim about an actual repair price.
A revision might read:
Noor slid the bill between Eli’s mug and the breadboard.
“Can you send your half today?”
Eli read the total. “Who agreed to the extra part?”
“The old one was cracked.”
“That wasn’t what I asked.”
Noor took her phone from her pocket, then put it beside the bill. “All right. We’ll call and ask what was authorised.”
The revision gives each reply a purpose. Noor wants a transfer. Eli wants an explanation of the additional work. The phone provides a possible next action. Their disagreement has not vanished, but it has moved from competing declarations to a step they can take.
Review the result against the specific note. Does the exchange avoid repeating the sink’s condition? Can the reader distinguish the two immediate aims? Does the ending change their options? If so, the revision has addressed the targeted problem. You may still want to adjust rhythm or add context elsewhere, but that is another question.
Do not assume that showing an action means all explanatory narration must disappear. A sentence about an earlier agreement may be necessary for the reader to understand Eli’s objection. The choice depends on what the scene has already established. An editorial rule that forbids every direct explanation can make the passage less clear rather than more subtle.
You can keep separate notes while drafting. If you notice that the amount has changed from an earlier chapter, mark the continuity question and continue to the scene’s endpoint if the discrepancy does not prevent it. Later, check the amount deliberately. This prevents a small uncertainty from repeatedly interrupting the larger task while still ensuring it gets resolved.
At the end of the session, identify the next job in plain terms. “Check whether Eli knew about the replacement part before this scene” is a useful return note. It directs the next review towards evidence in the manuscript instead of requiring you to reproduce the mood in which you wrote the original criticism.
A bounded review session can focus on one kind of problem. You might read for speaker clarity, then for the scene’s final decision. Trying to judge plot, grammar, voice, pacing and every emotional implication at once can make it difficult to understand why a passage feels wrong. Separate questions help you connect an impression to a repair.
Use examples from the page in your notes. Quote the line where you lose track of the speaker or identify the paragraph that explains a point already clear in dialogue. This gives the later revision a location. It also lets you check whether a change actually addresses the concern instead of merely producing different wording.
Keep a record of successful repairs. You need not turn it into a catalogue of achievements. A brief note that “giving Eli a question about authorisation clarified the scene” preserves a method you can try elsewhere. The record concerns the work, rather than requiring a new global judgement that you are now excellent at dialogue.
Feedback can provide another source of evidence if you ask a focused question. An unfamiliar reader might tell you who they think should pay, what each flatmate wants or why the conversation ends with a phone call. Their answers reveal what the text communicates. They need not agree with either character for the scene to be understandable.
Distinguish confusion from preference. A reader who cannot tell whose phone is on the table may have identified an attribution problem. A reader who dislikes Noor’s impatience may be responding as intended. Ask what in the passage produced the reaction before treating every negative response as a defect to remove.
A separate invented example can show why this matters. Suppose a character declines an invitation politely while hoping the host will ask again. One reader finds the character evasive. That response may fit the scene’s purpose. Revising the character into complete openness to avoid disapproval could erase the conflict you wanted to write. The question is whether the behaviour is intelligible and useful to the story.
Set an endpoint for the review. If your task was to clarify who speaks, inspect the revised passage and move on when the attribution is clear. You can schedule a later pass for rhythm. Without an endpoint, each repaired sentence can become the occasion for finding another vaguely unsatisfactory one, and the scene never returns to the draft as a whole.
Preserve versions when experimenting with substantial changes. A saved earlier copy makes it possible to compare the original intention with the new result. Sometimes a technically smoother revision removes a hesitation or awkwardness that belonged to the character. Recovering it is an editorial choice, not a failure to improve.
For your next session, replace one broad negative note with a question that refers to a specific passage and intended effect. Make one revision aimed at that question, then reread the result in context. Record what changed and what still needs attention. Leave the page with a next action you can name, rather than another verdict you cannot use.
