Quantum mechanics allows for correlations that defy classical logic, yet it is not the most "non-local" theory mathematically possible. While the no-signaling principle—the requirement that information cannot travel faster than light—prevents instantaneous communication, it does not explain why quantum correlations are capped at specific values like Tsirelson’s bound. A central goal in quantum foundations is to identify the deeper physical principles that define these boundaries.
The Gap Between Causality and Quantum Limits
The no-signaling principle is a surprisingly weak constraint. As Wei Ren notes in Joint measurements and Svetlichny's inequality, while violating certain inequalities would necessitate a breach of causality, there exists a vast "post-quantum" landscape of correlations that respect no-signaling but exceed quantum limits.
To narrow this gap, researchers have proposed more restrictive principles. One leading candidate is Information Causality, which posits that communication cannot grant a receiver more information than the number of physical bits actually transmitted. Wei Ren demonstrates in Bound on genuine multipartite correlations from the principle of information causality that this principle successfully rules out many extreme non-local theories. However, the authors also observe that Information Causality may not be sufficient on its own to uniquely define quantum mechanics, as some super-quantum correlations still appear to satisfy it.
Uncertainty as a Correlation Governor
Another perspective suggests that the limits of non-locality are not external rules, but rather consequences of internal quantum constraints. Stephanie Wehner argues in Relaxed uncertainty relations and information processing that Tsirelson’s bound is a direct result of quantum uncertainty relations. By mathematically "relaxing" these local uncertainty constraints, one can construct hypothetical theories that allow for superstrong correlations. This implies that the difficulty of learning a local state is fundamentally linked to how strongly that state can be correlated with a distant partner.
Similarly, the nature of the measurements themselves imposes limits. Selby et al. show in Correlations constrained by composite measurements that the existence of "parity reading measurements" can recover Tsirelson's bound in generalized probabilistic theories. This suggests that quantum theory might be optimal in its ability to produce correlations given the specific types of observables available in nature.
Topology and Network Constraints
In complex systems, the physical arrangement of participants—the network topology—creates its own set of bounds. Unlike standard Bell tests, quantum networks with independent sources result in non-convex sets of correlations. Renou et al. derive nonlinear bounds in Limits on Correlations in Networks for Quantum and No-Signaling Resources, showing that the causal structure of a network restricts achievable data even for theories more general than quantum mechanics.
These nonlinearities are increasingly important for characterizing non-classical behavior. For instance, Huang and Vontobel explore generalizations of the CHSH inequality using Pearson correlation coefficients, revealing subtle non-classical behaviors that linear inequalities miss. To handle these complexities, Klep et al. introduced "state polynomials" as a mathematical framework to optimize and bound the violations of these nonlinear inequalities.
Bypassing Bounds with Quantum Processes
While quantum states are strictly bounded, the rules change when considering quantum processes. A famous theoretical construct, the Popescu-Rohrlich (PR) box, achieves the absolute algebraic maximum of non-locality (a CHSH value of 4). While no physical state can do this, Plávala and Ziman demonstrate in Popescu-Rohrlich box implementation in general probabilistic theory of processes that PR boxes can be implemented using quantum channels.
However, this "super-quantum" performance comes with a catch. Jenčová and Plávala prove that any qubit-based channel reaching this maximum violation must be entanglement-breaking. This highlights a recurring theme in the field: nature seems to trade one resource for another, ensuring that even when we "break" one bound, we remain constrained by others, such as the inability to transmit information faster than light.
Key Foundational Work
Bound on genuine multipartite correlations from the principle of information causality (2011) — contributed 4 concept elements to this direction.
Relaxed uncertainty relations and information processing (2008) — contributed 3 concept elements to this direction.
Maximal Violation of the Collins-Gisin-Linden-Massar-Popescu Inequality for Infinite Dimensional States (2008) — contributed 2 concept elements to this direction.