When a process works normally, there may be little reason to discuss its condition. Failure creates distinctions. A surface can be chipped, crazed, blistered, warped, scored, pitted, contaminated, delaminated, burned, glazed, gouged, buckled, torn, cracked, or simply out of tolerance. Each description points toward a different cause, consequence, or repair decision.
Software has similarly elaborate language for failure: timeout, deadlock, race condition, regression, memory leak, stale cache, dependency conflict, corrupted state, dropped packet, authentication failure, and countless others. Medicine, logistics, machining, agriculture, printing, construction, hospitality, and aviation all develop rich failure vocabularies because problems demand classification.
That creates a useful learning strategy. If you want to understand a profession quickly, ask how its workers describe things going wrong. The answers reveal which properties they monitor and which variations change their response. Failure language turns abstract expertise into observable categories.