[Paper Review] Expressiveness via Intensionality and Concurrency

Thomas Given-Wilson|arXiv (Cornell University)|Apr 2, 2014

Logic, programming, and type systems26 references4 citations

TL;DR

This paper formalizes the relative expressiveness of four computational models—lambda calculus, SF-calculus, pi-calculus, and concurrent pattern calculus—using a computation square where expressiveness increases with intensionality and concurrency. It establishes irreversible encodings showing that concurrent pattern calculus is more expressive than pi-calculus, and SF-calculus than lambda calculus, via homomorphisms and parallel encodings.

ABSTRACT

Computation can be considered by taking into account two dimensions: extensional versus intensional, and sequential versus concurrent. Traditionally sequential extensional computation can be captured by the lambda-calculus. However, recent work shows that there are more expressive intensional calculi such as SF-calculus. Traditionally process calculi capture computation by encoding the lambda-calculus, such as in the pi-calculus. Following this increased expressiveness via intensionality, other recent work has shown that concurrent pattern calculus is more expressive than pi-calculus. This paper formalises the relative expressiveness of all four of these calculi by placing them on a square whose edges are irreversible encodings. This square is representative of a more general result: that expressiveness increases with both intensionality and concurrency.

Motivation & Objective

• To formally analyze and compare the expressive power of four foundational computation models: lambda calculus, SF-calculus, pi-calculus, and concurrent pattern calculus.
• To address the gap in understanding how intensionality and concurrency jointly increase computational expressiveness beyond traditional sequential extensional models.
• To establish a systematic framework using valid encodings, homomorphisms, and parallel encodings to compare these calculi.
• To prove that concurrent pattern calculus is strictly more expressive than pi-calculus, and SF-calculus than lambda-calculus, via separation results.
• To generalize the findings into a computational square model that captures the interplay between extensionality/intensionality and sequential/concurrent computation.

Proposed method

• Adapts and formalizes the concept of valid encodings from prior work to define homomorphisms and parallel encodings for comparing heterogeneous calculi.
• Uses homomorphisms to encode sequential calculi (lambda and SF-calculus) into concurrent ones (pi-calculus and CPC), preserving structural operations like application and parallel composition.
• Employs parallel encodings to map sequential application into concurrent parallel composition with auxiliary machinery, enabling encoding of sequential systems into concurrent ones.
• Applies known results from the literature to prove the validity of encodings, particularly for lambda-calculus into pi-calculus and SF-calculus into CPC.
• Establishes separation results via contradiction: for example, showing that no valid encoding of CPC into pi-calculus exists due to differing behavioral properties like self-reduction and success under replication.
• Constructs a formal computation square with irreversible arrows, demonstrating that expressiveness increases along both dimensions: intensionality and concurrency.

Experimental results

Research questions

• RQ1 How can the relative expressiveness of four key computation models—lambda calculus, SF-calculus, pi-calculus, and concurrent pattern calculus—be formally compared?
• RQ2 What role do intensionality and concurrency play in increasing computational expressiveness, and can this be captured systematically?
• RQ3 Can irreversible encodings be defined between these calculi such that each step in the computation square represents a genuine increase in expressive power?
• RQ4 Why is concurrent pattern calculus strictly more expressive than pi-calculus, and what structural properties prevent such an encoding?
• RQ5 To what extent can the framework of homomorphisms and parallel encodings be generalized to other combinatory or process calculi?

Key findings

• The paper establishes a valid homomorphic encoding from the lambda calculus into the pi-calculus, confirming that pi-calculus can simulate sequential extensional computation.
• A valid homomorphic encoding from SF-calculus into concurrent pattern calculus (CPC) is proven, showing that CPC can simulate intensional sequential computation.
• The encoding from pi-calculus into CPC is both valid and a homomorphism, demonstrating that CPC can simulate concurrent extensional computation.
• A separation result is proven: there is no valid encoding of CPC into pi-calculus, due to a behavioral mismatch in success conditions under replication.
• No homomorphism exists from CPC into pi-calculus, confirming that CPC is strictly more expressive than pi-calculus in the concurrent setting.
• The computation square model is validated, with all arrows representing irreversible increases in expressiveness, and the framework is shown to be generalizable to other calculi.