Text set beside an image to say what it is. A caption decides how the picture is read.
Latin captio (a taking, a seizing; also a catching of the mind), from capere (to take, seize); captio first named the opening clause of a legal document — the part that 'catches' the reader — long before the word attached itself to pictures.
Text set beside an image to say what it is — and in saying what it is, it decides in advance how the image gets read.