The cutting room

Every long piece of writing reaches a day when it has to get shorter, and the day is dreaded for a specific reason: cutting prose feels like destruction. Every paragraph was work. Cutting one means the work is gone, the paragraphs around it no longer join, and the thing that was cut might have been the thing that mattered. So the writer cuts adjectives, and the piece is still two thousand words over, and the deadline arrives.

The fear is correct about prose and wrong about cards. In prose, a cut is a deletion. On cards, a cut is a move — a card taken out of the row and put in a pile — and the card still exists, the pile can be read, and the card can be put back. This guide is about cutting done that way: on the row, not in the draft, with nothing destroyed. Film editors have a name for the place where this happens, and the name is right, because a cutting room is where the film is made, not where it is damaged.

The principle: cut cards, not sentences

A draft that is nine thousand words when it should be six thousand is not a draft with three thousand words of fat evenly distributed through it. It is a draft with too many cards in the row — claims that do not carry the argument, evidence that repeats evidence already given, background that nobody needed, a section that belongs in a different piece. Cutting adjectives cannot find these, because they are not adjectives. Laying out the row can, because on the row each one is a card, and a card that is doing nothing looks like a card that is doing nothing.

So: do not open the draft. Open the row.

Step 1: Ask what each card is for

Go down the row, one card at a time, and ask of each: what would the argument lose if this were gone?

Three answers, and a fourth that is the important one.

• The argument would lose a claim it needs. The card stays.
• The argument would lose evidence a claim needs. The card stays — unless another card under the same claim gives the same evidence, in which case one of them goes.
• The argument would lose nothing, but the reader would be confused. Background, definition, context. The card stays, but it may be shorter than it is, and it may belong immediately before the card that needs it rather than at the front.
• The argument would lose nothing. The card is doing nothing. It is in the row because it was interesting, or because it was work, or because it was there. It goes to the cut pile.

Be honest about the fourth answer. The card you most want to keep for the fourth reason is the one you found in an archive nobody has visited, and it is beautiful, and the argument does not need it. It goes in the pile. It is not gone. It is the first card of the next piece.

Step 2: Find the claim that is a different paper

Sometimes the row is too long not because of stray cards but because it contains two arguments. One claim — often with the most evidence under it, because it was the one you cared about — does not serve the thesis. It serves a thesis of its own. Student guide 4 has the test: does this claim help make the thesis true? If the answer is it is true, and it is interesting, and no, the whole pile — the claim, its evidence, its counterpoint — leaves the row together and becomes its own stack with its own title card. That is not a cut. It is a second paper you have just discovered you had, and it is the most productive thing a cutting session can find.

Step 3: Collapse the evidence

Under each surviving claim, count the evidence cards. A claim with five cards under it is usually a claim with two good ones and three that repeat them. Keep the strongest and the one from the other source; the rest go to the pile with a note — same point as Q14 — because the note is what will let you find them again when a reviewer asks for more.

Two cards from the same source under the same claim is nearly always one card too many.

Step 4: Read the bridges

Now read the row's claim titles aloud in order (student guide 6). The cuts have removed cards, and where they were there are joins that used to be mediated by something and now are not. Some joins hold better than before, because the stray card was in the way. Some do not hold at all, and that is where a bridge sentence is needed in the draft — or where two remaining claims have turned out to be one, and merge.

Step 5: Then, and only then, the draft

Recompile from the shortened row. The sections that correspond to cut cards are gone; the bridges are rewritten at the joins; the introduction is rewritten to promise the shorter order. The cut is done, and it took an afternoon, and nothing was deleted: the pile is beside you, every card in it named and dated with a reason.

Then, if it is still over, cut sentences. There will be fewer to cut, and they will be adjectives.

The pile is part of the work

The cut pile is not waste and should not be treated as waste. Three reasons.

It is the next piece. The claim that was a different paper is in it, with its evidence. The beautiful archival card is in it. Most writers' second article comes out of the first one's cut pile, and most do not know it, because the pile was deleted.

It is the answer to the reviewer. Why did you not consider X? You did. The card is in the pile with the reason: overlaps with claim 3; cut for length. You can say so, and say it with the date.

It is the record. A finished piece has no wrong turns in it. The pile is the wrong turns, and the reasons, and a reader of the process (student guide 7) sees judgment in a pile that a finished piece cannot show.

So the pile has rules of its own: every card in it keeps its number; every card gets a one-line reason for the cut, dated; and the pile is never thrown away. A card can be restored from it — put back in the row — at any time, and the reason it went and the reason it came back are both part of the story.

When the cut is the committee's

This is too long from a committee or an editor is a request with no target attached, and the temptation is to cut evenly, a little from everywhere. Do not. Go back to step 1 with the committee's other comments beside you; the card that is doing nothing is very often the one they also said they did not follow, and cutting it answers two comments at once. Bring the row to the next meeting, with the pile, and say what went and why. A supervisor who can see the cut on the table stops asking for it in the margin.

Cutting to a form

The same piece has several lengths, and each is a cut of the same row: the chapter at sixty pages, the article at six thousand words, the conference paper at three thousand, the talk at twenty minutes with slides. Each is an arrangement — an order of a subset of the cards — and the cutting room is where the arrangements are made: the article's row from the chapter's, the talk's row from the article's. None of them deletes a card. All of them share the pile. The cards were the work, and the forms are what the work is cut to.

Doing this in software

Software that works from cards makes the cut a move and keeps the pile: a card retired from the deck goes to a Cut it can be restored from, with a reason if you give one and the date whether you do or not; it keeps its number, so a reference to it in the record or in a reviewer's question still names it; and the reason appears in the record of the process as a decision, not a deletion. Several arrangements of the same cards — the chapter's, the article's, the talk's — are kept and compared side by side, and the Report counts words by section so the cut can be measured against the limit before the prose is touched. What the software cannot say is which card is doing nothing. It can show a claim with nothing under it and a card in no order; the fourth answer is yours.