This is an old revision of the document!
Invented by Max Holden. See his “The New Knife and Selected Cards” in Bagshawe's The Magical Monthly, Vol. 2, No. 10, July 1925, pp. 199-200.
A more rudimentary handling of the principle was published in 1927 by Walter B. Gibson in his Two Dozen Effective Practical Card Tricks, p. 31, under the title “The Wrapped Pack”. Gibson has a knife inserted into a wrapped deck to stab between two chosen cards, which are on the top and bottom. As the deck is separated at the point of insertion, the halves are confused, so that the top and bottom cards are mistaken for those above and below the knife blade. This idea is also the harbinger for Garcia's similar two-card stab and Bill Simon's “Business Card Prophesy”. Another early description of the cross-the-cut force appears in Victor Farelli's 1931 MS., Controlled Coincidence, where Farelli sites the Holden trick.
Some doubt of the Holden attribution has been raised by Wesley James, who points out this passage from Annemann's 202 Methods of Forcing: “An old timer but very practical and deceiving force is to have the card on top and the deck on table. Ask to have it cut anywhere and when this is done, performer carelessly puts the lower half on upper half but crosswise. It is left this way for a few minutes while trick is continued until time for selection to be revealed when top half is taken off and top card of lower half (?) turned over. Really deceptive.” Why didn't Holden, who published the booklet, correct this statement? However, until evidence of the force predating Holden's description is found, he is still the only known candidate.
Bolstering this indirectly is “A New Force” in Wrinkles by S. Wilson Bailey and Harold A. Osborne (1910), p. 38. The Bailey-Osborn procedure is this: Force card is on the bottom; spectator cuts off half the pack, picks up lower portion and deals it onto the cut-off stock (thus bringing the force card to the top), then cuts the deck again; magician asks for a half-pack to be designated, and equivoques the upper stock, the top card of which is then used. This is an admittedly labored procedure, and it could be argued there would be no reason to suggest it if the cross-the-cut force had been known at the time.