Is spacetime quantum? Dami´an Pital´ua-Garc´ıa1, ∗ 1Centre for Quantum Information and Foundations, DAMTP, Centre for Mathematical Sciences, University of Cambridge, Wilberforce Road, Cambridge, CB3 0WA, U.K. (Dated: September 8, 2021) Although the standard viewpoint in theoretical physics is that the unification of quantum theory and general relativity requires the quantization of gravity and spacetime, there is not consensus about whether spacetime must fundamentally have any quantum features. Since the violation of Bell inequalities is one of the most striking features of quantum theory, we assume that the satisfaction of Bell inequalities by spacetime degrees of freedom in a spacelike separated experiment is a sensible condition for spacetime to be called “classical”. Here we show a theorem stating that spacetime degrees of freedom violate Bell inequalities in a background Minkowski spacetime if a few properties of general relativity and quantum theory have a broad range of validity, and if the quantum state reduction upon measurement is a real physical process that is completed superluminally when acting on distant quantum particles in a quantum entangled state. In contrast to the Eppley-Hannah argument for the necessity of quantizing the gravitational field, we discuss the validity of our assumptions, our thought experiment does not require to manipulate or detect gravitational waves, and our theorem does not rely on the conservation of momentum or on the uncertainty principle. I. INTRODUCTION could be fundamentally classical (e.g., [26, 27, 30–35]). For the previous reasons, there is not consensus at present Quantum theory (QT) and general relativity (GR) are of whether the gravitational field, and hence spacetime, our most fundamental theories of physics. They are both must have any quantum properties. Ultimately, this supported by enoumous experimental confirmation for question must be answered experimentally. But experi- over a century. However, QT and GR cannot provide a mental evidence for quantum features of gravity remains complete description of physics. On the one hand, QT missing. predicts that the configuration of matter and energy can Different criteria could be considered to evaluate be in quantum superpositions. On the other hand, Ein- whether spacetime is fundamentally classical. For exam- stein’s equations of GR treat matter, energy and space- ple, natural conditions to impose on spacetime degrees of time as purely classical, which cannot be in quantum su- freedom to consider spacetime as classical could be: that perpositions. The prediction of QT that massive systems they cannot be in quantum superpositions, or that they can be in quantum superpositions has been experimen- cannot be entangled with other systems, or that they tally confirmed in microscopic scales (e.g., [1–6]). Thus, cannot violate any Bell inequalities. it follows that GR needs to be modified, at least, in these Bose et al. [36], and Marletto and Vedral [37] have re- scales. Moreover, it is conceivable that QT requires mod- cently proposed a beautiful experiment that they claim ification in sufficiently macroscopic scales too. Impor- could be implemented in the near future and that could tantly, it has been suggested that the predictions of QT demonstrate that the gravitational field can mediate en- could reach a limit of validity at sufficiently macroscopic tanglement, implying that the gravitational field is non- scales due to effects of gravity (e.g., [7–13]). classical [36, 37]. Assuming that the gravitational field is This motivates the question, is spacetime fundamen- a property of spacetime, as GR has taught us, this would arXiv:2109.02608v2 [quant-ph] 7 Sep 2021 tally classical, or does it have some quantum properties? imply that spacetime is non-classical too. It could be argued that because QT has been experi- Quantum entanglement is a paramount phenomenon mentally confirmed in microscopic scales, spacetime must that does not arise in classical physics. It has various fun- have some quantum properties in these scales too. In damental applications in quantum information science, fact, several theories of quantum gravity have been in- for example, in quantum teleportation [38], superdense vestigated (e.g., string theory [14], loop quantum gravity coding [39], quantum cryptography [40] and quantum [15], and others [16–19]). Furthermore, there are argu- computation [41]. However, the entangled correlations ments claiming that the gravitational field must be quan- between a pair of quantum systems arising in timelike tized (e.g., [20–23]). However, these arguments have been separated experiments can in principle be modelled with refuted (e.g., [24–27]). Moreover, quantizing the grav- purely classical intuitions via local hidden variable the- itational field has several important problems [28, 29]. ories (LHVTs) [42]. For example, two parties communi- Thus, it has been proposed that the gravitational field cating subluminally can straightforwardly simulate any quantum entangled correlations. However, as follows from Bell’s theorem [42], the vio- lation of Bell inequalities in experiments performed on a ∗ [email protected] pair of physical systems at spacelike separation cannot be 2 described by LHVTs, defying all classical intuitions. QT U. Since a quantum measurement comprises an interac- predicts the violation of Bell inequalities by some entan- tion between the measurement apparatus and the mea- gled states: every pure entangled quantum state violates sured system, we would expect that QSR and U should a Bell inequality [43, 44], but not every mixed entangled have the same fundamental description. quantum state does [45]. Different approaches towards the quantum measure- The Clauser-Horne-Shimony-Holt (CHSH) inequality ment problem have been proposed. An important view- [46] is one of the simplest and most popular Bell inequal- point is that QSR is merely a Bayesian updating of a ities. It involves Alice randomly choosing and applying probability distribution on acquiring new information one of two measurements labelled by α 0, 1 on a sys- (e.g., [65–69]). Nevertheless, the Pusey-Barrett-Rudolph tem A at spacelike separation from Bob∈{ randomly} choos- theorem [70] (see also Refs. [71–74]) tells us that, given ing and applying one of two measurements labelled by some arguably sensible assumptions, the quantum state β 0, 1 . Alice’s and Bob’s respective outcomes a and is a real physical state, suggesting that QSR is a real b have∈{ two} posible values a,b 0, 1 . The CHSH in- physical process. ∈ { } equality is given by The “many worlds” interpretations of QT [75, 76] claim that only U takes place fundamentally, and that ICHSH 2, (1) the different possible outcomes in a quantum measure- | |≤ ment coexist in reality. However, these interpretations where ICHSH E(0, 0)+ E(0, 1) E(1, 0)+ E(1, 1), and cannot effectively explain the observed probabilities in ≡ − where E(α, β) = P (a = b α, β) P (a = b α, β) denotes agreement with the Born rule in quantum measurements, | − 6 | the expectation value when Alice and Bob choose the and they cannot be verified or refuted in experiment ei- measurements labelled by α and β, for all a,b,α,β ther [77]. ∈ 0, 1 . The CHSH inequality is maximally violated by a In our view, promising approaches to solve the quan- { } pair of qubits in a maximally entangled state, for example tum measurement problem are collapse models, propos- in the singlet state ing that QSR is a real physical process whose explana- tion requires to extend the Schr¨odinger equation (e.g., 1 Ψ− = 0 1 1 0 , (2) [12, 78–88]). Crucially, we note that there are collapse | i √2 | i| i − | i| i models suggesting that gravity mediates QSR (e.g., [7– 12]). achieving the quantum Tsirelson bound [47]: Assuming that QSR is a real physical process, Eppley and Hannah claimed in an influential paper [21] that the ICHSH 2√2. (3) gravitational field must be quantized. Their argument is | |≤ that 1) if a gravitational wave of arbitrarily small mo- Before 2015, the violation of Bell inequalities had been mentum can be used to reduce the quantum state of a confirmed in various experiments (e.g. [48–57]), but sub- quantum particle to measure its position then either mo- ject to at least one of three important loopholes: the mentum is not conserved or the uncertainty principle is locality loophole [42], the freedom-of-choice loophole [42] violated; and 2) if the gravitational wave does not reduce and the detection loophole [58]. Three outstanding ex- the quantum state of the particle then a superluminal sig- periments [59–61] published in 2015 demonstrated the vi- nal can be transmitted by having the wave interact with olation of Bell inequalities closing these loopholes simul- a particle that is entangled with another distant parti- taneously. However, despite being addressed in an ex- cle. This argument has been refuted on different grounds periment [62], the collapse-locality loophole [63] remains (e.g., [24–26, 32, 89]). Ref. [24] shows that the device open [64]. Thus, the results in previous Bell experiments proposed by Eppley and Hannah to measure the posi- could in principle be described by LHVTs. tion of a particle with a gravitational wave cannot be In this paper we assume that the loophole-free satisfac- built even in principle. Refs. [25, 26] assert that even tion of Bell inequalities by spacetime degrees of freedom if QSR is a real physical process, it does not need to al- in spacelike separated experiments is a necessary condi- low superluminal signalling. Refs. [32, 89] argue that tion for spacetime to be sensibly called “classical”. As we strict conservation of momentum does not need to hold will show, under some assumptions, spacetime degrees of fundamentally. freedom can violate the CHSH inequality, implying that Broadly speaking, this paper presents a theorem stat- spacetime has quantum features.