Modular Construction of Fixed Point Combinators and Clocked Boehm Trees

Fixed point combinators (and their generalization: looping combinators) are classic notions belonging to the heart of lambda-calculus and logic. We start with an exploration of the structure of fixed point combinators (fpc’s), vastly generalizing the well-known fact that if Y is an fpc, Y(SI) is again an fpc, generating the Boehm sequence of fpc’s. Using the infinitary lambda-calculus we devise infinitely many other generation schemes for fpc’s. In this way we find schemes and building blocks to construct new fpc’s in a modular way. Having created a plethora of new fixed point combinators, the task is to prove that they are indeed new. That is, we have to prove their beta-inconvertibility. Known techniques via Boehm Trees do not apply, because all fpc’s have the same Boehm Tree (BT). Therefore, we employ clocked BT's', with annotations that convey information of the tempo in which the data in the BT are produced. BT's are thus enriched with an intrinsic clock behaviour, leading to a refined discrimination method for lambda-terms. The corresponding equality is strictly intermediate between beta-convertibility and BT-equality, the equality in the classical models of lambda-calculus. An analogous approach pertains to Levy-Longo Berarducci trees. Finally, we increase the discrimination power by a precision of the clock notion that we call atomic clock'.

💡 Research Summary

The paper investigates fixed‑point combinators (fpc’s) and their generalisation, looping combinators, from the perspective of structural generation and discrimination. It begins by recalling the classic observation that if Y is an fpc then Y · (S I) is again an fpc, which yields the well‑known Böhm sequence of fpc’s. Using the infinitary λ‑calculus, the authors devise a whole family of generation schemes that go far beyond this single pattern. They introduce modular building blocks (such as S, I, Ω and various parameterised operators) and show how to combine them to obtain infinitely many new fpc’s. The construction is systematic: a “parameterised transformation” Y · M, nested S‑I compositions, and infinite repetitions of certain operators all give rise to new combinators, each with a clear syntactic description.

Having produced a large catalogue of fpc’s, the central problem becomes proving that the newly generated combinators are not β‑convertible to the already known ones. Traditional discrimination tools based on Böhm Trees (BT) fail here because every fpc shares the same infinite BT – a tree consisting of an endless chain of λ‑abstractions ending in a self‑application. To overcome this limitation the authors introduce clocked Böhm Trees. A clocked BT annotates each node with a timestamp indicating after how many β‑reduction steps that part of the tree becomes observable. These timestamps capture the “tempo” of the reduction process, thereby enriching the tree with dynamic information that is invisible in the ordinary BT.

The paper shows that two λ‑terms whose ordinary BTs coincide may have different clocked BTs, and that a difference in timestamps is sufficient to conclude that the terms are not β‑equivalent. By applying this method to the many fpc’s generated earlier, the authors establish β‑inequivalence for a vast collection of previously unknown combinators.

To increase the discriminating power even further, the authors define atomic clocks. While a plain clock records only the number of reduction steps, an atomic clock records the exact shape of each reduction step (e.g., whether it is an η‑reduction, a δ‑reduction, or a specific variable substitution). This fine‑grained information allows the distinction of terms that have identical step counts but differ in the internal reduction pattern. Atomic clocks therefore refine the clocked BT technique and enable the separation of fpc’s that were indistinguishable even with ordinary clocked trees.

The authors also extend the clock concept to Levy‑Longo‑Berarducci (LLB) trees, showing that the same enrichment works for this alternative observational model. Consequently, the paper establishes a new equivalence relation that sits strictly between β‑convertibility (the strongest) and ordinary BT‑equality (the weakest). Two terms that are equal under clocked BTs are guaranteed to be β‑equivalent, but equality of ordinary BTs does not imply equality of clocked BTs.

In the concluding sections the paper summarises four main contributions: (1) a modular, infinitary‑calculus‑based framework for systematically constructing infinitely many new fixed‑point combinators; (2) the introduction of clocked Böhm Trees as a tool for refined behavioural discrimination; (3) the development of atomic clocks that further increase discriminating precision; and (4) the demonstration that these notions apply equally to LLB trees, thereby providing a unified approach to term discrimination across several classical λ‑calculus models.

Overall, the work significantly advances the theory of fixed‑point combinators by providing both a rich supply of new combinators and a powerful, intermediate notion of equality that bridges the gap between syntactic β‑conversion and the coarse observational equivalence given by ordinary Böhm (or LLB) trees. This contributes both to the foundational understanding of recursion in λ‑calculus and to practical techniques for reasoning about program equivalence in functional languages.