How to find your thesis after you've done the reading

The assignment says the paper needs a thesis. The handout says a thesis is "a debatable claim that the paper supports." You have read everything, you have a stack of cards, and you have written four thesis sentences, each of which is either something nobody would argue with (the canal was important to the region) or something you cannot actually support (the canal caused the 1830 revolt).

Here is the thing the handout does not say: the thesis is the last claim you write, not the first. It is the sentence that the claims you already have add up to. You do not think it up; you find it in the cards, and there is a procedure for looking.

This guide is the procedure. It assumes you have cards — claims with evidence under them, as in guides 1 and 2. If you are working on paper, you need a table.

Why the thesis written first is a guess

When you write the thesis before the reading, you are predicting what the sources will show. Then the reading either confirms the prediction, in which case you read for confirmation and miss what did not fit; or it does not, in which case you have a thesis and a pile of evidence against it and a due date.

When you write it after the reading but before sorting the cards, it is still a guess, just a better-informed one. You are summarizing a feeling about sixty cards. Feelings about sixty cards are unreliable in a specific way: they are dominated by the last ten you read.

A thesis found from the sorted row of claims is not a guess. It is a description of what the row shows. That is why it holds up.

Step 1: Lay out the row and read only the claims

Put the claim cards in a row — just the claims, one sentence each, evidence beneath but out of the way. Read the titles aloud, left to right.

You are listening for what they have in common. Not the topic; you know the topic. What is the point they share? If every claim is true, what does that make true?

Write candidate answers on cards as they come. Do not judge them yet. You may get four.

Step 2: Test each candidate against the row

For each candidate thesis, go down the row and ask of every claim: does this claim help make the thesis true?

Three answers are possible, and each one is information:

• Yes — the claim is a reason for the thesis. Good.
• No, it is irrelevant — the claim is true and has nothing to do with this thesis. Set the claim to one side for this candidate.
• No, it cuts against it — the claim is evidence the thesis is wrong.

A candidate that gets yes from most claims and cuts against from none is a thesis. A candidate that gets irrelevant from half the row is a thesis for a different paper — a narrower one, or a different one. A candidate that gets cuts against from any claim you believe is not your thesis; it is something you wanted to be true.

Most students find, doing this, that their favorite candidate is the one with the most irrelevant*s, because it is the grandest, and the candidate that survives is smaller and more specific than they hoped. **That is the right result.* A thesis that a paper of this length can actually support is always smaller than the topic.

Step 3: Check that it is a claim and not a topic

The surviving candidate has to pass the test from guide 1: a full sentence with a verb, that someone could disagree with.

• The economics of the canal — a topic. Nobody can disagree with a noun phrase.
• The canal had economic effects — a sentence, but nobody would argue.
• The canal's cost overrun, known to its backers before the vote, is why the county's merchants opposed the railway a decade later — a claim. Someone could say no. Your row of claims is what you would say back.

If the survivor is a topic, look at what the row shows about the topic and write that.

Step 4: Check that it is not one of the claims

Sometimes the thesis you find is already on a card, as a claim, with evidence under it. That is fine, and common. Move that card to the top; it was the thesis all along, and the other claims are its reasons.

But check that it really covers the row. A claim promoted to thesis will often leave two or three other claims stranded — true, supported, and irrelevant to it. Either the thesis needs to be one step more general (what do this claim and those three have in common?) or the stranded claims are a second paper. Set them aside as such. Nothing is wasted; it is next term's paper, and it already has its evidence.

Step 5: Say what it is not

Write one more card: what the thesis does not claim. This paper does not argue the canal caused the revolt; it argues the overrun changed who the merchants trusted.

This card is for you, and it earns its keep twice. When you draft, it stops you drifting into the bigger claim you could not support. And it usually becomes a sentence in the introduction, because telling the reader what you are not arguing is how you keep them from arguing with a paper you did not write.

What if nothing survives?

Sometimes every candidate gets cuts against from something, or the row splits into two groups that share nothing. Three possibilities, in order of likelihood:

1. Two papers. The row has two clusters. Pick the one with more evidence, set the other aside, and find the thesis of the cluster. This is the usual answer.
2. A claim is wrong. A claim you believed cuts against everything else. Look at its evidence. If it is thin, it may not be true, and cutting it may let the rest of the row agree. If it is strong, it may be the most interesting thing you have, and the thesis is the one that explains it — why, despite X, Y — which is a better paper than the one you planned.
3. More reading. The row is too short to add up to anything. This is rarer than it feels. Before reading more, do step 1 again with the counterpoint cards included; often the thesis is in what the objections have in common.

Where it goes

The thesis card sits at the top of the row. It is written into the introduction, which you write last (guide 1, step 7), and restated — not repeated — in the conclusion, where you can now say it with the row behind it.

If you are keeping the cards after the paper (you should), the thesis card is the one to keep on top. Next term, when you pick up the stack, it is the card that tells you what this stack was for.

Doing this in software

Software that works from cards will not find your thesis, and you should be suspicious of any that claims to. What it can do is make step 2 faster: show you the claims alone without their evidence, let you tag each one for, against or irrelevant to a candidate, and show a claim that no other card supports — the row's weak point — before you build a thesis on it. The finding is yours. The row is just easier to see.