late 14c., "a taking, seizure," from Old French capcion "arrest, capture, imprisonment," or directly from Latin captionem (nominative capito) "a catching, seizing, holding, taking," noun of action from past participle stem of capere "to take" (from PIE root *kap- "to grasp").
Used from mid-17c. especially in law, and there via its appearance at the head of legal document involving seizure, deposition, etc. ("Certificate of caption"), the sense was extended to "the beginning of any document," and thence to "heading of a chapter or section of an article" (1789), and, especially in U.S., "description or title below an illustration" (1919).