Teaching Argument Writing with Frankenstories

When we make an argument, we are trying to build a logical ladder from ground truth to some kind of contested claim. This evidence, or these truths, mean that this claim is also true.

But how do we assess good reasoning? What do we mean by logical?

In an ideal world of pure reason, we could use deductive reasoning or formal logic to assess an argument.

This is the kind of reasoning where the structure of the argument itself is what leads to the conclusion. For example:

• All A's are B's.
• X is an A.
• Therefore, X is a B.

You can make A, B, and X anything you like, and as long as the first two statements are true, the conclusion must also be true.

• All bananas are fruit.
• This black squishy thing on my desk is a banana.
• Therefore, this black squishy thing on my desk is a fruit.

Of course, you can argue about the truth of any of those supporting claims (maybe that black squishy thing can't be called a banana anymore), but the argument structure is valid.

If that black squishy thing really is a banana, then you can't argue that it isn't a fruit. That's what valid means.

However, in real-world arguments, formal logic is limited.

In ordinary social arguments, we have less certainty, more noise, time constraints...

For example, how would you assess the structure of this argument?

• You often like books set on boats.
• Piratica by Tanith Lee is a book set on a boat.
• Therefore you should read Piratica by Tanith Lee.

By deductive standards, we'd have to chuck it out. There are all sorts of ways that conclusion could be wrong: 

• Maybe the subset of boat-related books you don't like are ones about pirates.
• Maybe every copy of Piratica got burned to ashes in a freak firestorm.
• Maybe there are just other things to do with your time that are more important to you than one specific book recommendation.

But it's not a bad argument!

It's very reasonable to recommend books to people based on the stuff they like.

We call this kind of reasoning informal or inductive. The goal isn't 100% absolute certainty; rather it's to discover (or persuade an audience of) what's most likely.

Instead of validity, inductive reasoning is judged on how strong or weak the relationship is between the main claim and its supporting reasoning and evidence.

The problem is that this is a lot harder to establish than validity, which is why so many arguments just never seem to end.