I rassled my way through this subject quite a few years ago, when I was working on a model of an American frigate from the Revolutionary War period. It's a complex topic.
The basic developmental progression seems to have begun with the lateen yard and its "full," triangular sail. That configuration eventually morphed into the modern gaff-and-boom-rigged spanker. Just how the process worked seems to have varied a bit from place to place, and the various steps in it didn't happen suddenly.
Here's a quote from the relevant volume in the Conway's History of the Ship series, The Line of Battle: The Sailing Warship, 1650-1840. The chapter on rigging was written by Karl Heinz Marquardt. (I have reservations about some of his stuff, but I do have confidence in his writings about the history of rigging in the eighteenth century.) From p. 127:
"Of all the major sails in a ship, the mizzen course underwent the greatest modification. Originally a fore and aft lateen sail (and not altered in most Continental ships for the duration of the period under review), by the early seventeenth century in English and Dutch ships it had acquired a bonnet to its foot, which gave the sail a four-cornered look. Around 1680 the bonnet was integrated into the course and, strictly speaking, the sail was then a settee rather than a lateen sail. After 1730 the large mizzen course's part before the mast was gradually dispensed with and the new fore edge, or luff, was laced to the mizzen mast. The change came first to smaller ships and was progressively extended to all, with the last of the larger ships being altered during the 1780s.
"Smaller craft had carried a short gaff rather than the more unwield mizzen yard from early in the century, but the sail usually had a vertical leech and a clew which belayed inside the taffrail (or a little ouside it on fixed outriggers). The gaff can trace its ancestry to the 1620s, with the first gaff rigged vessels being Dutch boeiers; evolved from the diagonal sprit yard, for a time it was also known as a half-sprit. During the American Revolutionary War, the relatively small mizzen began to be replaced with a larger gaff-and-boom sail 'in the form of a brig's main sail'. The new sail came to be called the spanker, and gradually took over from both the loose-footed gaff and the truncated sail set from the long mizzen yard. Nelson's flagship during the Battle of the Nile in 1798, HMS Vanguard, was the only ship of th eline still carrying a mizzen yard during that action.
"Preceding the spanker, and often confused with it, was the driver, at first an additional fine weather square sail, hoisted to the peak of the mizzen yard. After about 1780 in English merchantmen it became a mizzen's fore and aft extension, but in the Royal Navy the term was applied to a larger temporary fore and aft sail of brig main sail shape hoisted to the gaff instead of a small loose-footed mizzen. An ancilliary boom was essential to extend the clew of the driver, but the sail disappeared when the permanent spanker rendered it redundant. Continental ships seem to have carried the square driver for much longer."
I've seen other sources that disagree with that explanation in a few minor points, but I think Mr. Marquardt has the story straight.
The crossjack, or crojack, yard seems to have been conceived initially as an apparatus for spreading the foot of the mizzen topsail. In the beginning, when the mizzen sail extended forward of the mast, there obviously was no room for a square sail set to the crossjack yard. Eventually it occurred to somebody that a square sail could be set there, and it was, with rare logic, called the crossjack. But I'm not sure exactly when that happened. James Lees, in his authoritative book The masting and Rigging of English Ships of War, 1620-1860, makes no mention of such a sail. The American and British clipper ships (e.g., the Flying Cloud and the Cutty Sark) certainly set crossjacks. But I can't recall having bumped into a contemporary picture of a warship setting one. Such a sail didn't actually accomplish much. If the ship was running before the wind the crossjack would block the wind from the main course, and if she was working to windward the spanker would block half the wind from the crossjack. Contemporary pictures (both paintings and photos) frequently show the crossjack furled when virtually all the other square sails are set.
Bill's observation of the mizzen yard being secured above the crossjack yard in that painting of the Victory is most interesting; I'd never noticed such a thing before. My first inclination was to think the artist had simply made a mistake - and I guess that's possible. But after thinking about it for a minute I don't see why that configuration wouldn't work. The mizzen yard wouldn't need to be raised or lowered under any normal circumstances. (The ritual of swinging it around to the opposite side of the mast every time the ship came about became obsolete when the part of the sail forward of the mast got chopped off). The mechanism for securing the crossjack yard to the mast, according to Mr. Lees, was quite simple: a rope truss arrangement, rather than the bulkier parrels used on the fore and main courses. Such a truss could easily be passed between the mizzen sail and the mast. And whether the crossjack yard or the mizzen yard was on top surely wouldn't make any difference to the handling of the ship.
Or maybe the artist goofed. But the rest of the painting suggests pretty strongly that he knew what he was doing.
Fascinating, if ultimately trivial, stuff.