Learn
How to evaluate an argument before it moves a decision
Evaluating an argument means checking whether its reasons actually support its conclusion, separately from whether you like the conclusion. Most weak decisions rest on an argument that sounded complete and was missing a piece nobody thought to look for.
The parts of an argument
In 1958 the philosopher Stephen Toulmin set out the parts of a practical argument. The claim is the conclusion being urged. The grounds are the facts offered for it. The warrant is the link explaining why those facts support that claim, and it is very often left unstated.
A complete argument also carries a qualifier, saying how strongly the claim holds, and a rebuttal, saying when it would not.
Most business memos state the claim and the grounds and leave the warrant for the reader to supply.
Four questions to ask
Are the grounds true, and how do we know? Do the grounds bear on this claim, or on something nearby? Would the warrant survive being said aloud? Under what conditions would the claim fail?
Asking the third question is often enough. A warrant that sounds reasonable when implied can sound absurd once stated.
Applying it to a memo
Break the memo into its separate claims, one sentence each. For every claim, write the grounds offered and the warrant that would connect them. Mark the grounds as checked or unchecked, and the warrant as stated or assumed. The claims with unchecked grounds and assumed warrants are where the memo is weakest, and there are usually fewer of them than its length suggests.
The common mistake
Judging the argument by its conclusion. People rate arguments as stronger when they agree with where they end, a pattern psychologists call belief bias. A useful discipline is to put an argument for a conclusion you dislike through exactly the same four questions, and to notice whether your standard has drifted.
How the room weighs claims
In a Decision Room, every argument is held as separate claims rather than paragraphs. Each option's Champion makes its strongest case, the Challenger attacks the weakest warrant behind it, and the Inspector marks every claim supported, partly supported, unverified, or contradicted. The Judge decides what survived and writes the Verdict on that basis, so the Human Decision is made from tested claims rather than from the most fluent memo.
The survey that liked the mockup
A memo argues that customers will pay $99 a month for a new tier, because 68 percent of 250 surveyed customers rated the mockup very appealing. The claim is the $99 price. The grounds are the survey. The unstated warrant is that finding something appealing means paying $99 for it. Said aloud, the warrant fails, since the survey never mentioned price. The 170 enthusiastic respondents are evidence of interest, not of willingness to pay. A better test would show the price to 50 customers and count how many sign up.
Questions
- What is a warrant in an argument?
- The reasoning that connects the evidence to the conclusion. It explains why the facts offered should lead to the claim, and it is the part most often left unstated.
- How is argument evaluation different from fact-checking?
- Fact-checking asks whether the grounds are true. Argument evaluation also asks whether true grounds actually support the conclusion, which is where many confident arguments fail.
Bring Argumont a Decision worth challenging.
Start with a paid 60 day pilot on your own real Decisions.