The premise of detecting a spectator's lie was introduced by U.F. Grant in his “Tell the Truth Telephone” from The Sphinx, Vol. 26 No. 7, Sep. 1927, p. 233, in which six spectators are given envelopes with six or seven cards in each. Each person thinks of a card. The performer, seemingly through the aid of radio earphones, can detect when the spectators are lying or telling the truth about their mentally chosen card. See The Fine Print, No. 4, Feb. 1997, p. 92, for a reprint of the article.
Vincent Dalban later posed a different lie detector plot to Theodore Annemann in The Jinx, No. 4, Jan. 1935, n.p. Dalban posed the idea of turning his back to a spectator while she deals from a shuffled deck, naming all the cards aloud as she comes to them. At any point she can lie and name a different card than the one she's looking at, but the magician can spot the lie. Although Dalban didn't have a method for his effect, Annemann published several versions in following issues of The Jinx. The first few solutions relied on full deck stacks or slugs of pre-arranged cards; see Charles Nyquist and Stuart Robson's individual solutions in The Jinx, No. 5, Feb. 1935, n.p. It was Henry Christ who brought forth the now-common solution of having the spectator lie on her selection rather than any random card. This allowed a fully-shuffled deck to be in play, using only a single key card for the method. His handling appeared in The Jinx Summer Extra, 1935, p. 39.
Bringing the lie detector premise to the spelling genre was courtesy of Martin Gardner. His “Gardner's Card Speller” appeared in Joe Berg's Here’s New Magic, 1937, p. 3. The question-based spelling plot incorporated in Gardner's trick — where you first ask and spell the color, then the suit, etc. — stems from Herbert Milton's “The Spelling Card” in The Magic Circular, Vol. 18. No. 186, May 1922, p. 333. There was no lying involved in Milton's trick; all the answers had to be truthful.