The origins of RFT are described in the "personal prologue" to the first book on RFT:
Hayes, S. C., Barnes-Holmes, D., & Roche, B. (2001). Relational Frame Theory: A Post-Skinnerian account of human language and cognition. New York: Plenum Press.
That prologue is attached below.
From about 1981 thru early 1986 Steve Hayes and Aaron Brownstein, a well known basic behavior analyst, ran their labs jointly together at the University of North Carolina, Greensboro. Despite the fact that Hayes was a clinical psychologist and Brownstein was a basic behavior analyst the team was highly cohesive and productive. While Steve was trying to develop an analysis of verbal rules and verbal meaning Aaron suggested he consider the relatively recent and not well known phenomenon of stimulus equivalence. Hayes reports being shocked by the phenomenon and the absence of a process-oriented account of it. Unwilling to leave it as a mere outcome he reports spending almost an entire week pondering the processes that could possibly give rise to equivalence. In a kind of "lightbulb moment" he came to the simple idea that relating is likely an instance of generalized instrumental learning (that is, an operant), much as with generalized imitation.
Hayes excitedly presented his analysis to Brownstein, who thought the idea was plausible and very interesting. The two soon laid out nearly 20 studies that would test the idea. That list has been lost to history, but Hayes claims that when he last saw it in the early 2000's nearly every study had in fact been done and all were supportive.
The first paper that began to lay out the ideas behind RFT (and ACT and Functional Contextualism for that matter) was this one: Hayes, S. C. (1984). Making sense of spirituality. Behaviorism, 12, 99-110. It is attached below. That paper was already being written in 1983 and the lab worked on these ideas for months if not years so RFT would have to have begun somewhere in the 1981 thru 1983 period. Precisely when has been lost to history.
The first full presentation of the model was in an invited address co-authored with Aaron Brownstein, : Hayes, S. C. & Brownstein, A. J. (May 1985). Verbal behavior, equivalence classes, and rules: New definitions, data, and directions. Invited address presented at the meeting of the Association for Behavior Analysis, Columbus, OH.
A jointly authored Hayes and Brownstein paper was drafted before that talk but publication was delayed because Hayes had accepted a job at the University of Nevada, Reno as Director of Clinical Training and was busy planning the move. The paper was still being rewritten for submission for publication when Aaron died unexpectedly while dancing with his wife, on April 12, 1986. It is attached below in two sections because the file to too large to attach it as one.
Brownstein was a better known figure at the time and Hayes felt that it was not right to assume that Aaron would have approved of the details of the analysis. He also feared that others might think we was using Aaron's reputation to get attention to the piece. Thus he continued to revise the paper. It was delivered in revised form in June of that year in this paper: Hayes, S. C. (June 1986). What is a verbal stimulus? Invited address presented to the Summer Institute on Verbal Relations, Behaviorists European Summer Academy, Bad Kreuznach, West Germany. The proceeding of that conferences were submitted for publication sometime in 1987 but due to extreme delays in publication by a skeptical publisher Hayes and his new wife Linda Parrot Hayes eventually started a new publishing company, Context Press, in large part so as to finally bring that volume into print. The proceedings appeared in 1991in the first volume produced by Context Press: Hayes, S. C. (1991). A relational control theory of stimulus equivalence. In L. J. Hayes & P. N. Chase (Eds.), Dialogues on verbal behavior (pp. 19-40). Reno, NV: Context Press. A footnote to that chapter acknowledged it great debt to Aaron and had he lived people would likely think of him as a co-founder of RFT.
Parenthetically, Context Press continued to grow and was eventually purchased by New Harbinger Publications and still exists today.
So where did RFT come from? Hayes summarizes his views in the personal prologue of the first RFT book this way:
Where RFT came from? Putting together these sources of influence —radical functionalism, rigorous experimental orientation, the lead provided by rule-governance and simple equivalence, and the analytic creativity of basic behavior analysts —RFT is a fairly natural extension of the field itself as I was trained to view it. (Hayes, 2001, p. 9).
As a historical side note:
Hayes has long claimed that anyone who really understands generalized imitation understands the RFT idea that relating is an operant. Generalized imitation is an example of the kind of radical functionalism that is inside RFT. To understand generalized imitation, start with Don Baer, one of the founders of applied behavior analysis. Don's mentor was Jack Gewirtz who championed he idea of generalized imitation (Gewirtz, J. L., & Stingle, K. G. (1968). Learning of generalized imitation as the basis for identification. Psychological Review, 75(5), 374–397. https://doi.org/10.1037/h0026378). You can read about his background in a long oral interview done for the Society for Research in Child Development (https://www.srcd.org/sites/default/files/file-attachments/gewirtz_jacob_interview.pdf). Its a bit of a shock to realize that Jack Gewirtz was trained by Beth Wellman at the Iowa Child Welfare Station -- the heroic woman who almost single handedly stood up to the eugenic lies of Louis Terman when the eugenicists were claiming all intelligence is genetic. A good book on that horrible history is Marilyn Brookwood's "The Orphans of Davenport: Eugenics, the Great Depression, and the War over Children's Intelligence". Stated another way, RFT up with a rich functional behavioral tradition.
Linking RFT to that wing of behavioral thinking may also help make sense of some of the resistance to RFT ideas even today ... but that is a story still to be told.