Open quantum dynamics without Complete Positivity: a criticism
Fabio Benatti, Dariusz Chrusciński ; Saverio Pascazio · 2026
Read Open quantum dynamics without Complete Positivity: a criticism by Fabio Benatti, Dariusz Chrusciński ; Saverio Pascazio (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2605-18639.txt.
Introduction
Fabio Benatti and two co-authors challenge, in a 2026 paper, complete positivity as a requirement for open quantum dynamics. They weigh its physical motives and limits, including plans to limit non-completely-positive maps to matching initial states. Using isotropic states as a test case, they show domain limits tighten fast as system size grows. The study exposes a built-in flaw in the matching-based approach to non-completely-positive dynamics.
Catalog record
| CREATOR | Fabio Benatti, Dariusz Chrusciński ; Saverio Pascazio |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 26660 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2605.18639
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Open quantum dynamics without Complete Positivity: a criticism Fabio Benatti, Dariusz Chrusciński ; Saverio Pascazio Abstract The requirement of complete positivity is very often regarded as a fundamental consistency condition for the description of open quantum dynamics. We critically examine this requirement and discuss both its physical motivations and its limitations. We analyze proposals based on restricting the domain of non-completely positive maps to subsets of compatible initial states. Using isotropic states as a concrete example, we show that such domain restrictions become increasingly severe with growing system dimension, revealing an intrinsic weakness of the compatibility-based approach. 1 Introduction and motivations In Physics, when one formulates an evolution law in terms of a differential equation, one implicitly assumes that the law is valid for any initial condition. This is Laplace’s legacy. A good example is the fundamental equation of mechanics F=ma=m\ddot{x} . Physics students are taught how to solve this equation for every initial condition x(t=0)=x_{0},\quad\dot{x}(t=0)=v_{0} , where x_{0} and v_{0} are the initial position and velocity, respectively. This is called a Cauchy problem and one requires that the solution exists and is unique \forall\;x_{0},v_{0} . When the solution does not exist or is not unique one faces a serious problem. An example is \ddot{x}=6x^{1/3} , with x(t=0)=0,\quad\dot{x}(t=0)=0 , that admits the two solutions x_{1}(t)=0 and x_{2}(t)=t^{3} . One would be tempted to assume that a potential of the type V\sim x^{4/3} does not exist in Nature (one would say that it is “unphysical”), but unfortunately such a potential can be conceived, and this has generated controversy [ 1 ] . There is a mathematical way out of this conundrum: the function x^{1/3} is not Lipschitz near x=0 and hence the uniqueness theorem does not apply; therefore one assumes that physical forces and potential must satisfy the Lipschitz condition and removes ambiguity-generating initial conditions from the outset. This does not solve the physical problem though. A more physical way out would be to argue that an equation like the afore-mentioned one does not describe the behavior of a point particle with infinite accuracy, more so close to the origin, where the effects of microscopic details and friction (and even symmetry breaking) become crucial. If duly taken into account, the microscopic features of the environment would yield a unique solution. Why do physicists not like the above example? Probably because one does not like the idea that the mechanical particle can “decide” its own trajectory, in a world, that of classical mechanics, that is deterministic. This attitude goes very far: a Cauchy formulation, for a physical law describing time evolutions, is taken to be valid even for partial differential equations, say the wave equation. Also in this case, one wants that the solution exists and is unique. In full generality, abiding by such a request leaves two possible scenarios: i) either one should be open to restricting the possible initial conditions to those that allow for the existence and uniqueness of trajectories, or, ii) one should modify the dynamics in order to allow for all possible initial conditions. The aim of this brief introduction is to show the similarity of such a dilemma with the request of complete positivity (CP) in open quantum dynamics, specifically in the case of quantum dynamical semigroups. There, restricting the dynamics to those enacted by the generators of the so-called Gorini-Kossakowski-Sudarshan-Lindblad (GKLS) form [ 2 , 3 , 4 ] goes along the lines of the second scenario ii) depicted above, namely paying the price of selecting as physically consistent only specific dynamical maps in order to use them to describe the time-evolution of any initial quantum state. Note that CP plays a crucial role in this context, because the system undergoing the dynamics could be entangled to parts of its environment, and a positive (non-CP) dynamics would not guarantee the positivity of the density matrix. See the following Section for additional considerations on this issue. On the other hand, a quantum counterpart of the classical restriction of the initial conditions , scenario i) above, was proposed by a number of authors [ 5 , 6 , 7 , 8 , 9 ] and by the advocates of the so-called ”slippage of initial conditions” [ 10 , 11 , 12 ] . Though not equivalent, all these proposals share a common spirit in that they restrict the initial states to those which are compatible with the dynamics, be it only positive (namely, not completely positive) or not even positive. Here, states compatible with a given dynamics are those states that remain states, that is, density matrices with positive spectrum adding to 1 for all times. We shall focus upon the first scenario above; to start with, we briefly survey the issue of complete positivity as it comes to the fore when dealing with open quantum systems. 2 Notation and additional considerations The evolution of a closed (isolated) quantum system is described by the so-called von Neumann equation for the density matrix \rho_{t} , describing the statistics of its state at time t , starting from an initial condition \rho at time t=0 : \dot{\rho}_{t}\;=\;-i[H,\rho_{t}]\quad\longleftrightarrow\quad\rho_{t}\;=\;U_{t}\rho_{0}U^{\dagger}_{t}\,;\quad U_{t}=\exp{(-i\,H\,t)}\ . (1) If the quantum system is “open”, namely not isolated, and immersed in an environment with which it interacts, in accordance with the request of Complete Positivity (see below), the one-parameter group of unitary maps \mathcal{U}_{t}:\rho\mapsto\mathcal{U}_{t}[\rho]=\rho_{t} (2) is replaced by a one-parameter family of maps \rho_{t}\;=\;\Lambda_{t}[\rho]\;=\;\sum_{\alpha}K_{\alpha}(t)\,\rho\,K_{\alpha}^{\dagger}(t),\quad\sum_{\alpha}K_{\alpha}^{\dagger}(t)K_{\alpha}(t)=\mathds{I}\ . (3) These maps are said to be in Kraus-Stinespring form [ 13 , 14 ] . Notice that if the summation consists of a single addendum K(t)=U_{t}=e^{-iHt} , ( 3 ) reduces to ( 1 ). Notice also that ( 3 ) is not a differential equation, but rather a “snapshot” of the quantum state at a particular time t . The above specific form guarantees not only that \Lambda_{t} is positive, but also that it is Completely Positive. Namely, a map of the form \Lambda_{t} as in ( 3 ) not only preserves the positivity of all states of an open system S in the course of time, as it should do in order to allow for the association of its eigenvalues with probabilities. What is more to it is that, when tensorized with the identical map {\rm id}_{n} , \Lambda_{t}\otimes{\rm id}_{n} also preserves the positivity of all states of the compound system consisting of S and an arbitrary, dynamically inert and dynamically decoupled, auxiliary n -level system S_{n} . Notice that, though not evolving and not interacting with S (this is expressed by the factorization of the dynamical map of S+S_{n} into the given one of S and the trivial one of S_{n} ), yet this auxiliary system can be statistically correlated to S . Indeed, the reason behind the request of Complete Positivity, namely behind the fact that \Lambda_{t} must be of the Kraus-Stinespring form ( 3 ), is that, otherwise, there certainly exists [ 20 , 19 ] an entangled state \rho^{ent}_{S+S_{n}} of the compound system S+S_{n} some of whose eigenvalues becomes negative at some time, precluding the interpretation of \Lambda_{t}\otimes{\rm id}_{n}\Big[\rho^{ent}_{S+S_{n}}\Big] as a physically consistent state. Two aspects are worth stressing: the first one is that that not all entangled states of S+S_{n} are incompatible with the compound