Traditional philosophers identify two kinds of reasoning:
Inductive reasoning from specifics to a general conclusion about all of them: Many samples of ocean water are salty, so ocean water must be salty (but there could be exceptions).
Deductive reasoning from a general warrant and reason to a specific claim: Ocean water is always salty; this water is from the ocean, so it must be salty.
But the more common kind of reasoning is abductive—reasoning that begins with a hypothesis that might explain the data in question. We then test that hypothesis using whatever reasoning seems appropriate. Abductive reasoning is problem-driven, a kind of reasoning that begins with a hypothesis that is the tentative solution to a problem.
Each kind of reasoning is vulnerable to cognitive biases.
When you think inductively, you risk basing a conclusion on too few instances. You avoid that risk by gathering more evidence than you think you need and by learning something about statistical sampling and analysis.
When you think deductively, you risk formulaic thinking, applying a rote warrant to every situation.
When you think abductively, you risk fixating on the first hypothesis that springs to mind. Guard against that by holding your earliest hypotheses delicately, by imagining more than one, and by deliberately seeking out evidence that disconfirms your favorite one.
