This is just to notice that the [clone apply] clone apply for recursion is very close to the way the lambda calculus Y combinator reduces via graph rewrites, af

This is just to notice that the [clone apply] clone apply for recursion is very close to the way the lambda calculus Y combinator reduces via graph rewrites, after is translated into a graph [1].The graphical reductions lead to just a pair of two trivalent nodes, one of them a sort of fanout (clone) and another an application [2]. There is no lambda term which can be directly translated into this graph, because of the wiring of the graph. This coincidence is curious. [1] you can check this at https://mbuliga.github.io/quinegraphs/lambda2mol.html#y_comb_id https://mbuliga.github.io/quinegraphs/lambda2mol.html#y_comb... where you can see how the term Y (\x.x) reduces. In the λ>mol textarea delete what is written there and just copy/paste (\g.((\x.(g (x x))) (\x.(g (x x))))) input then click the λ>mol button. Then the start button. Stop whenever you want. [2] the graph akin to a pair clone - apply can be seen at https://mbuliga.github.io/quinegraphs/quinelab.html#howtolab https://mbuliga.github.io/quinegraphs/quinelab.html#howtolab where you copy/paste this graph as written in the mol notation: A input y x^FOE x output y "A" is the apply node and "FOE" is a clone node (say).