Editing someone else

An editor who rewrites a draft into their own voice has produced a worse document and a discouraged writer, which is a poor return on the effort.

The prior question

Before touching anything.

What is this document for, and who reads it?

What stage is it at — is this a structural read or a final polish?

What does the writer want from you, which should be asked rather than assumed?

And what is the deadline, since the depth of editing possible depends on it.

Editing a first draft for commas is wasted work, and editing a final draft for structure is too late.

The levels

Which should be done in order.

Structural: is the argument right, is the order right, is anything missing or redundant?

Sectional: does each section do one thing, and is its point clear?

Line: clarity, length, unnecessary words.

Copy: grammar, consistency, house style.

Proof: typos, numbers, names.

Doing them out of order means polishing text that will be cut, which wastes the editor's time and the writer's goodwill.

Changing versus suggesting

The core judgement.

Change: errors of fact, grammar, consistency, and anything that is simply wrong.

Suggest: anything about structure, emphasis, or argument, which the writer knows more about than you do.

Query rather than change where you are uncertain what was meant, since guessing produces errors that look like the writer's.

And leave alone: choices that are different from yours but not worse, which is the largest category and the one most editors get wrong.

Preserving voice

The discipline.

Before changing a sentence, ask whether it is unclear or merely not how you would write it.

The second is not a reason to change it.

Which means resisting the reflex to smooth every sentence into a house rhythm, since what results is prose that is technically clean and lifeless.

And it means a light hand is generally better than a heavy one, particularly with a writer whose voice is working.

Giving the feedback

How it is delivered.

Start with what is working, specifically, since generic praise is discounted and specific praise tells the writer what to keep.

Prioritise: three things that matter, rather than forty comments of equal weight.

Explain the reason, since a change explained is learned and a change imposed is not.

Ask questions where the problem is the writer's to solve: "what is this section doing?"

And separate the significant from the optional explicitly, so the writer knows where to spend their time.

Comments that work

The wording.

"I lost the thread here" is more useful than "unclear", since it describes the reader's experience rather than issuing a verdict.

"I expected X after this" identifies a structural gap.

"This is the strongest paragraph — could it come earlier?" is a structural suggestion in a form that is easy to accept.

And avoiding "awkward", "flows badly" and similar, which give the writer nothing to act on.

The volume problem

A common failure.

A draft returned with two hundred comments is not actionable and is demoralising regardless of the quality of the comments.

Which means: fix the small things silently, comment on the patterns rather than every instance, and limit the substantive comments to what can be addressed.

"You use the passive throughout — here are three examples" beats forty individual marks.

Being edited

The other side.

Read all the comments before responding to any, since the first ones sting most.

Leave it a day if the reaction is strong.

Distinguish between changes you disagree with and changes you dislike.

Accept the copy-editing without argument, generally.

Defend the substance where you have a reason, which good editors welcome.

And notice the patterns across several edits, which is where your actual habits become visible.

The relationship

Which determines whether any of it works.

Editing is intrusive by nature, and a writer who feels attacked stops sharing drafts.

Which means the tone of the comments matters as much as their content, and that a difficult structural point is better made in conversation than in a margin.

And that an editor who is trusted can be far more direct than one who is not, which is a reason to build the trust before it is needed.