Impossibility of superluminal signalling rules out causal loops in conical spacetimes

This paper resolves a key open question by demonstrating that while operationally detectable causal loops are theoretically possible in (1+1)-dimensional Minkowski spacetime without violating the no-superluminal-signalling principle, such loops are strictly ruled out in higher-dimensional conical spacetimes across classical, quantum, and post-quantum theories, thereby establishing that the relationship between no superluminal signalling and the absence of causal loops is inherently dependent on spacetime geometry.

The Gist

The Big Question: Can You Send a Message to Your Past Self?

Imagine you are in a room with a "time machine" that lets you send a message to your past self. If you could do this, you could create a causal loop: You send a message to your past self, telling them to send the message, which causes them to send it, which causes you to receive it. It's a circle with no beginning.

In physics, there is a golden rule called No Superluminal Signalling (NSS). This is a fancy way of saying: "Nothing can travel faster than light."

For a long time, physicists believed that if you obey the "no faster-than-light" rule, you automatically break any time loops. You can't send a message to the past because that would require breaking the speed limit.

However, a surprise happened.
In a previous study, researchers found a loophole in a very specific, flat, 1-dimensional world (think of a single line). In that narrow world, it is theoretically possible to set up a time loop without breaking the speed limit. It's like a magic trick where the cause and effect are perfectly tuned to hide the fact that a message is traveling "backwards" in time.

The New Discovery: The Shape of the World Matters

This new paper asks: Does this magic trick work in our real world? Our world has 3 dimensions of space (up/down, left/right, forward/back).

The authors say: No, it doesn't work here.

They prove that in any world with more than one dimension of space (like our 3D world), the "No Faster-Than-Light" rule strictly forbids these time loops. If you try to build a causal loop, you will inevitably have to send a signal faster than light, which breaks the rules of relativity.

The Secret Ingredient: "Conical" Geometry

Why does it work in 1D but fail in 3D? The answer lies in the shape of the universe, which the authors call "Conicality."

The Analogy of the Flashlight

Imagine you are holding a flashlight in a dark room. The light beam creates a cone shape.

• In 1D (The Line): Imagine the light is just a line going forward. If you have two people standing on this line, their "future" (where their light beams go) can overlap in a very messy, confusing way. You can arrange them so that their future paths intersect perfectly to hide a time loop. It's like two shadows on a wall that can merge in a way that tricks your eye.
• In 2D or 3D (The Cone): Now, imagine the flashlight beam is a real 3D cone. If you have two people standing apart, their light cones (their futures) intersect, but they have a specific, rigid shape. The authors call this "Conical."

The paper proves that in these "Conical" shapes (like our 3D space), the geometry is too rigid. You cannot arrange the light cones so that they overlap perfectly to hide a time loop. The "shape" of the universe forces the time loop to collapse. If you try to build the loop, the math forces you to break the speed limit.

The "Fine-Tuning" Problem

The paper also explains why the 1D trick worked. It required Fine-Tuning.

Think of it like balancing a house of cards.

• In the 1D world, you can build a house of cards that looks like a time loop. But it is incredibly fragile. If you move one card just a tiny bit (a tiny change in the message or the location), the whole thing collapses, and the time loop disappears.
• The authors show that in 3D space, you can't even build the house of cards in the first place. The geometry of the room (the "Conical" shape) makes it impossible to balance the cards without breaking the rules of physics.

The Main Takeaway

1. In a flat, 1-dimensional world: You can theoretically create a time loop without breaking the speed of light, but only if you set up the universe with extreme, unnatural precision (fine-tuning).
2. In our 3-dimensional world (and any world with 2+ dimensions): The geometry of space is "Conical." This shape acts like a strict bouncer. It ensures that if you try to create a time loop, you must break the speed of light. Therefore, time loops are impossible in our universe if we respect the rule that nothing travels faster than light.

In short: The relationship between "no faster-than-light travel" and "no time travel" depends entirely on the shape of the universe. In our 3D universe, the shape guarantees that time travel is impossible if we obey the speed limit.

Technical Summary

Technical Summary: Impossibility of Superluminal Signalling Rules Out Causal Loops in Conical Spacetimes

Problem Statement
Recent work within the "affects framework" [6, 7] demonstrated that in (1+1)-dimensional Minkowski spacetime, the principle of No Superluminal Signalling (NSS) does not necessarily rule out operationally detectable causal loops (specifically, "affects causal loops" or ACLs). This counter-intuitive result arises because causation and signalling are inequivalent in the presence of fine-tuning, where causal mechanisms are tuned to hide observable signalling despite the existence of causal cycles. However, it remained an open question whether this phenomenon persists in spacetimes with higher spatial dimensions ($d > 1$). Specifically, does NSS rule out all operationally verifiable causal loops in $(d+1)$-Minkowski spacetime for $d > 1$, and what geometric features of spacetime govern this distinction?

Methodology and Framework
The authors utilize the affects framework, a theory-independent formalism for causal modelling that operates on directed graphs and probability distributions without assuming specific underlying physical mechanisms (classical, quantum, or post-quantum).

1. Causal Models and Signalling: The framework defines causal models via directed graphs with observed and unobserved nodes. Signalling is formalized through "affects relations" ($X \models Y$), where intervening on variables $X$ changes the distribution of $Y$. Higher-order affects relations ($X \models Y | \text{do}(Z)$) capture signalling conditioned on interventions.
2. Irreducibility: To prevent trivial or redundant causal inferences, the framework employs irreducibility conditions (Irred1 and Irred3). Irred1 ensures that every element in the signalling set $X$ has a causal path to some element in $Y$; Irred3 ensures every element in the conditioning set $Z$ has a causal path to $Y$.
3. Spacetime Embedding: Causal models are embedded into spacetime, modeled as a partially ordered set (poset) $T$ representing light cone structures. NSS is formalized as a compatibility condition: for an irreducible affects relation $X \models Y | \text{do}(Z)$, the joint future of $Y$ and $Z$ must be contained within the joint future of $X$ ($\bar{F}_s(Y) \cap \bar{F}_s(Z) \subseteq \bar{F}_s(X)$).
4. Conicality: The authors leverage a recently identified order-theoretic property of spacetime called "conicality" [11, 12]. A spacetime is conical if the joint future region of a finite set of points uniquely determines the location of all points in that set contributing non-trivially to the intersection. This property holds for $(d+1)$-Minkowski spacetime when $d > 1$ but fails for $d=1$.

Key Contributions and Results
The paper provides a definitive proof that in any conical spacetime, NSS rules out all operationally detectable causal loops (ACLs). The proof proceeds in two main steps:

1. Reduction to 0th-Order Relations: The authors prove (Lemma A.7) that any set of higher-order affects relations satisfying compatibility conditions in a conical spacetime can be reduced to an equivalent set of 0th-order affects relations ($X \models Y$) without losing causal inference power or violating NSS constraints. This simplifies the problem by eliminating the complexity of higher-order conditioning.
2. Proof of Impossibility in Conical Spacetimes: For 0th-order relations, the authors demonstrate that a non-degenerate, compatible embedding of an ACL leads to a contradiction in conical spacetimes.
   • In a causal loop, the compatibility condition implies a chain of subset relations among the joint futures of the involved variables (e.g., $\bar{F}(A) \supseteq \bar{F}(Y) \supseteq \bar{F}_s(AB)$).
   • Crucially, in conical spacetimes, the compatibility condition combined with the conicality property enforces strict subset relations (e.g., $\bar{F}(A) \supsetneq \bar{F}(Y)$).
   • Tracing the loop leads to a contradiction where a set is strictly contained within itself (e.g., $\bar{F}_s(AB) \supsetneq \bar{F}_s(AB)$), which is impossible.
   • Consequently, any compatible embedding of an ACL in a conical spacetime must be degenerate (i.e., all variables in the loop occupy the same spacetime location), rendering the loop operationally undetectable.

The paper further generalizes this result by defining "conical embeddings," showing that even in non-conical spacetimes, a causal loop compatible with NSS requires a specific form of fine-tuning in the embedding itself, where the joint futures of distinct event sets align precisely.

Significance and Claims
The authors establish that the relationship between the relativistic principle of NSS and the absence of causal loops is inherently dependent on the geometry of spacetime.

• Dimensional Distinction: The paper resolves the open question regarding $d > 1$, proving that unlike the $(1+1)$-dimensional case, NSS strictly forbids operationally verifiable causal loops in higher-dimensional Minkowski spacetime and any conical spacetime.
• Fine-Tuning Requirement: The results highlight that for causal loops to exist without superluminal signalling, two forms of fine-tuning are required: one within the causal model (to hide signalling) and one within the spacetime embedding (to satisfy the geometric constraints of non-conicality).
• Theoretical Scope: These findings hold for classical, quantum, and post-quantum theories, as the affects framework is theory-independent.
• Future Directions: The authors note that while their inference rules are "complete" for unknown system cardinalities, future work could explore whether stronger rules exist for fixed systems (e.g., binary variables) or whether the results extend to cyclic models satisfying alternative separation properties like $\sigma$- or $p$-separation. They also suggest the framework's utility in studying the emergence of acyclic causal order and time direction from general operational structures without presupposing a background spacetime.