Inference To The Best Explanation

Inference to the Best Explanation (IBE) is the logical process of observing a set of facts and concluding that the theory which explains them best is likely true. Unlike strict proofs, this method deals in probabilities and plausibility—just like a detective ruling out impossible suspects to find the culprit. In the driveway scenario, you intuitively used IBE to reject rain (contradicted by the dry street) and accept the car wash hypothesis because it was the simplest explanation that covered all the evidence. Ask this: Does my conclusion rely on the fewest assumptions while explaining the most facts?

Inference to the Best Explanation operates on a specific framework called 'explanatory virtues.' When faced with competing hypotheses, we filter them not just by what is possible, but by what makes the most sense. The primary virtues are usually Simplicity (Parsimony), Coherence (it fits with what we already know), and Scope (it explains the most data points).

Consider a medical diagnosis. A patient presents with a fever and a rash. It could be a rare tropical disease or a common viral infection. Unless the patient recently traveled to the tropics, the doctor infers the viral infection because it is a simpler explanation that coheres with the patient's history.

This applies to technology as well. If your internet stops working, you could hypothesize that the global internet infrastructure collapsed or that your router is unplugged. While both result in no internet, the router explanation is 'better' because it requires fewer massive assumptions.

However, 'best' is subjective. What counts as the best explanation depends on the background information you have. A mechanic might infer a specific engine failure from a noise that a layperson infers is just 'something loose,' simply because the mechanic has a richer set of potential explanations.

Inference to the Best Explanation (IBE) is often used synonymously with 'abductive reasoning,' though scholars sometimes distinguish the two. The term was coined to describe how we accept hypotheses that possess specific explanatory virtues over merely probable ones.

• Gilbert Harman (1965): Argued that all non-deductive reasoning, including induction, is essentially finding the best explanation for evidence, rather than just counting frequencies [1].
• Peter Lipton (2004): Distinguished between the 'likeliest' explanation (highest probability) and the 'loveliest' explanation (provides the most understanding), arguing that seeking the 'lovely' often guides us to the 'likely' [2].
• Douven (2017): Notes that IBE is the primary logic used in scientific realism to justify believing in unobservable entities like electrons because they best explain observable phenomena [3].

A major critique is the 'Bad Lot' objection: if all available hypotheses are poor, the 'best' one is still false. Additionally, humans are prone to 'explanation bias,' where we prefer stories that are emotionally satisfying or narrative-driven over complex, boring, or fragmented truths.