betaFeedback

WorksPhysics and Astronomy

Time-Travel Paradox in a Bohmian Extension of Deutsch's Approach: Mandatory Tunneling as the "Banana Peel Mechanism"

Abstract

We consider a self-prevention paradox in which a particle travels through a wormhole into its own past, encounters its younger version, and blocks its movement. In a specially constructed one-dimensional model, this collision rules out a classically self-consistent history. Deutsch's quantum approach can resolve such contradictions by imposing a self-consistency condition on the density matrix. The physical interpretation of this approach, however, remains debated. In particular, Dunlap argues that Deutsch's model presupposes the existence of interacting parallel worlds. Here we extend Deutsch's approach within de Broglie-Bohm theory by supplementing the self-consistency condition on the quantum state with a self-consistency requirement for the particle's ontic configuration. This allows us to consider the self-consistent evolution of an individual particle. In a two-channel, narrow-band Gaussian approximation, we find a self-consistent solution in which quantum tunneling is mandatory, despite its arbitrarily small probability in an ordinary scattering process. This realizes the "banana peel mechanism": to avoid a contradiction during a journey into the past, an event that would ordinarily be extremely unlikely occurs nonetheless.

Introduction

Some solutions of general relativity admit closed timelike curves, although whether they can be physically realized remains an open question [5, 6, 7, 8]. In particular, traversable wormholes have been discussed as a possible basis for time machines, whereas the chronology protection conjecture holds that the laws of physics prevent such curves from forming. Without assuming that time machines actually exist, we treat a time loop as a hypothetical physical system and investigate the conditions for its quantum self-consistency.

Consider a particle that travels through a wormhole into its own past and prevents its own motion. If the journey into the past occurs, the younger particle must encounter the older one; if the collision prevents the journey, the older particle should never have appeared. This paradox shows that locally allowed classical motions need not combine into a globally self-consistent history.

Classical models do not necessarily lead to this contradiction: Echeverria, Klinkhammer, and Thorne demonstrated self-consistent trajectories for a billiard ball interacting with its own past version [9]. Classical self-consistency, self-interaction, and the causal structure of wormhole spacetimes have also been studied more generally [10, 11, 12, 13, 14, 15]. By contrast, we deliberately choose a one-dimensional geometry with an obstacle that is impenetrable in the classical limit. The particles cannot simply “pass one another”: only specific channels of motion are available in the classical problem.

Deutsch proposed imposing a self-consistency condition on the density matrix of a system traveling along a closed timelike curve and showed that finite-dimensional systems admit a fixed density matrix [1]. For infinite-dimensional quantum systems, the existence of a fixed density matrix is guaranteed under various additional sufficient conditions [16]. This framework resolves classical logical contradictions at the level of mixed quantum states, but its interpretation remains controversial. Deutsch’s many-worlds reading is historically related to H. Everett’s relative-state formulation and DeWitt’s subsequent development of the many-worlds interpretation [17, 18]. Dunlap [2], in particular, argues that a physical reading of Deutsch’s construction involves additional ontological assumptions about a special kind of interacting parallel worlds.

Other quantum treatments of time loops include path-integral constructions [19] and postselected models [20]. For a comparison of circuit-based approaches, see Allen [21]. A. Everett [22] explored self-consistency in a path-integral model and found that preserving unitarity can make otherwise improbable events necessary – the so-called “banana peel mechanism” [23]. Although such approaches address quantum consistency, they do not by themselves specify a single actual particle trajectory through the loop. This leaves open the question of how an individual particle history can satisfy the return condition.

We propose an extension of Deutsch’s approach within de Broglie–Bohm theory, in which, in addition to the quantum state, there is an actual particle configuration guided by the corresponding wave field [3, 4, 24, 25, 26, 27, 28]. We impose a self-consistency condition not only on the density matrix but also on the particle’s actual ontic configuration – that is, on its real trajectory in configuration space. The central question is whether a self-consistent trajectory for an individual particle can exist in the presence of a time loop.

For the self-prevention problem considered here, we answer in the affirmative: an individual particle can have a self-consistent trajectory in the presence of a time loop. The main result has two parts: an explicit stationary density matrix in a two-channel Gaussian model, and an analytically specified self-consistent ontic configuration lying in the low-probability tunneling branch.

The article is organized as follows. § 2 introduces Deutsch’s density-matrix condition, density-matrix guidance in Bohmian mechanics, and the additional ontic closure requirement. § 3 specifies the two-channel return geometry, explains the classical self-prevention paradox and its possible quantum resolution, and states the Hamiltonian and approximations used in the calculation. § 4 establishes the leading-order matching of the returned packet states and constructs the stationary density matrix of the reduced two-channel model. § 5 derives the ontic fixed point for equal and unequal packet widths and examines the convergence of iterative methods based on the return map. § 6 discusses the scope of the results, their interpretation, and possible broader implications.

General Framework: Quantum and Ontic Self-Consistency

We first state the two self-consistency conditions independently of the particular collision model. The first is imposed on the quantum state, as in Deutsch’s approach; the second is an additional requirement on the actual configuration in a Bohmian description. Both refer to the same prescribed time-loop geometry.

Deutsch's Density-Matrix Condition

In Deutsch’s general formulation [1], a chronology-respecting system interacts with a system traversing a closed timelike curve (CTC). For the self-interaction studied here, we denote the two input subsystems by YY (younger) and OO (older), respectively. The symbols Y,OY,O are labels for the two interacting quantum subsystems; Deutsch’s fixed-point construction does not in general require them to be two stages of the same particle. Let ρY\rho_Y be the prescribed input state and ρO\rho_O the unknown state returning from the future.

The collision is described by a unitary operator V\mathcal{V} acting on the two-particle system. In our geometry, the outgoing younger particle enters the return route and reappears as the incoming older particle. We therefore trace out the outgoing older subsystem and apply the complete return channel G\mathcal{G}, which includes routing through the wormhole mouths:

ρO=G(Tr⁡O[V(ρY⊗ρO)V†]).(1)\rho_O=\mathcal{G}\left(\operatorname{Tr}_O\left[\mathcal{V}(\rho_Y\otimes\rho_O)\mathcal{V}^{\dagger}\right]\right). \tag*{(1)}

Here Tr⁡O\operatorname{Tr}_O traces over the outgoing older subsystem, which does not enter the return route; G\mathcal{G} identifies the outgoing younger state with the incoming older state. This differs in notation from the usual Deutsch formula, where the outgoing CTC subsystem is retained: an output exchange of subsystem labels gives the same construction. The return channel may in particular route spatially distinct output packets to different CTC exit regions.

Writing the resulting completely positive, trace-preserving map as DρY\mathcal{D}_{\rho_Y}, we obtain the quantum consistency requirement

ρO=DρY(ρO).(2)\rho_O=\mathcal{D}_{\rho_Y}(\rho_O). \tag*{(2)}

In finite-dimensional Hilbert spaces a fixed density matrix exists for such a map [1]. In infinite dimension, additional sufficient conditions are required [16]. This equation concerns the quantum state; by itself, it does not specify one actual particle trajectory that returns to its own initial configuration.

Density-Matrix Guidance in Bohmian Mechanics

In de Broglie–Bohm theory the quantum state is accompanied by an actual configuration Q=(Q1,…,QN)\mathbf{Q}=(Q_1,\ldots,Q_N) [3, 4]. A density-matrix formulation [29] allows the configuration to be guided directly by a configuration-space density matrix W(q,q′)W(q,q'), where q,q′q,q' are configuration arguments. For spinless particles, away from nodes of W(Q,Q)W(\mathbf{Q},\mathbf{Q}), the guiding equation reads

Q˙k=ℏmkIm⁡[∇qkW(q,q′)W(q,q′)]q=q′=Q.(3)\dot{Q}_k=\frac{\hbar}{m_k}\operatorname{Im}\left[\frac{\nabla_{q_k}W(q,q')}{W(q,q')}\right]_{q=q'=\mathbf{Q}}. \tag*{(3)}

Here mkm_k is the mass of particle kk; for internal spin indices one uses the appropriate trace. The matrix WW describes the relevant full configuration, rather than being identified automatically with the reduced ρO\rho_O. When spatial branches do not overlap at the actual configuration, their contributions to the local velocity field vanish, although branch interference during scattering can remain important.

Additional Ontic Consistency

Let XYout\mathbf{X}^{\mathrm{out}}_Y be the actual outgoing configuration of the younger subsystem entering a return route after interaction. Let XOin\mathbf{X}^{\mathrm{in}}_O denote the incoming older configuration at the earlier matching surface. The geometric map C\mathcal{C} identifies these configurations, including spatial reorientation and transport between the wormhole mouths. Our additional ontic self-consistency condition, also referred to as trajectory closure, is

XOin=C(XYout).(4)X_O^{\mathrm{in}}=\mathcal{C}\left(X_Y^{\mathrm{out}}\right). \tag*{(4)}

The outgoing configuration is determined by Bohmian motion under the quantum state satisfying Eq. (2). Thus the mathematical problem is to find a joint solution of the density-matrix condition (2) and the ontic condition (4). A stationary density matrix need not automatically imply a trajectory fixed point.

In the concrete model below, the CTC returns one and the same particle to its own past. The general ontic requirement will reduce to equality of two normalized internal coordinates, zO=zYoutz_O = z_Y^{\mathrm{out}}, on the self-consistent tunneling branch. We will solve the quantum and ontic conditions separately and then show their compatibility within the stated Gaussian approximations.

One-Dimensional Self-Prevention Model

A Particle Traveling into Its Own Past

The particle initially moves from left to right through a region where it may encounter its own older version. After the encounter, it follows one of two routes, to the right or to the left; both lead to the wormhole entrance. The wormhole returns it to a time before its initial passage through the interaction region. The two routes connect to different exit regions in the past, so the encounter determines where the older version appears rather than whether the journey into the past occurs. We prescribe these routes as boundary conditions of the effective model without constructing a specific spacetime metric. The one-dimensional restriction applies to the local collision: the particles cannot pass one another on the side.

Two versions of the particle are therefore present near the interaction region at the same time. We call the segment of its worldline before entering the wormhole the younger version YY, and the segment after emerging from it in the past the older version OO. In the calculation, these are two distinguishable degrees of freedom in an effective two-particle model. They have equal mass mm; their distinguishability can be associated with the particle’s internal clock.

An alternative formulation would let the younger particle enter the wormhole only if it continues to the right: reflection by its older version would then prevent the very journey that brought the older version into its path. Although this expresses self-prevention more directly, it requires quantum states with and without the older version and transition rules between them. We use the two-route formulation above.

To describe the collision quantitatively, we introduce a repulsive two-particle interaction U(xY−xO)U(x_Y-x_O) that depends on the separation between the two simultaneously present versions of the particle. In the classical limit, we choose a potential that forms an impenetrable barrier to their relative motion. Thus, an older version that appears in the younger version’s path does not merely “forbid” its motion by assumption: it physically reflects it. For equal masses, this corresponds to an exchange of velocities in a one-dimensional elastic collision. In the quantum problem, tunneling through the same barrier is possible.

Two Spatial Return Channels

If the younger version continues moving to the right after the encounter, it enters the wormhole along the right-hand route and emerges in the past to the right of itself. We denote this returning version by OAO_A, where AA denotes the right-hand return sector. It moves toward the younger version and becomes an obstacle in its path: the repulsive interaction U(xY−xO)U(x_Y-x_O) creates a barrier that classically reflects the younger particle to the left.

If, instead, the younger version is reflected to the left, it enters the wormhole along the left-hand route and emerges in the past to the left of itself; we denote this version by OBO_B, where BB denotes the left-hand return sector. The exit time, position, and direction of motion are chosen so that, in this case, the returning particle does not lie in the younger version’s path and does not prevent it from moving to the right.

Thus, the outcome of the collision with the particle’s older version determines where that same version emerges in the past:

T⟶OA,R⟶OB.(5)\mathrm{T} \longrightarrow O_A,\qquad\mathrm{R} \longrightarrow O_B. \tag*{(5)}

Here T denotes transmission to the right and R reflection to the left. This rule connects the two physical routes through the wormhole to its exit locations; it does not assume any additional “measurement” or external decision about whether the particle passes through the CTC.

Each route also defines a quantum return map: UT\mathcal{U}_T for the right-going transmitted particle and UR\mathcal{U}_R for the left-going reflected particle. These maps identify the outgoing state of the younger version with the incoming state of the older version at the corresponding CTC exit. They incorporate translation, possible momentum reorientation, and propagation along the respective route. In particular, they can modify the spatial width and quadratic phase of a packet; they need not preserve its shape pointwise. The maps and routing rule are model-defining boundary conditions, not approximations.

Classical Self-Prevention Paradox and Its Possible Quantum Resolutions

The routing rule produces an immediate contradiction in the classically impenetrable regime. Suppose the older version appears to the right (OAO_A). It then collides with the younger version and stops its own earlier motion to the right: the classically impenetrable repulsive barrier reflects the younger version, sending it to the left. But this leftward route leads to the older version appearing on the left, not the right:

OA⟶R⟶OB.(6)O_A \longrightarrow\mathrm{R} \longrightarrow O_B. \tag*{(6)}

Conversely, suppose the older version appears to the left (OBO_B). It does not prevent the younger version from moving to the right; the younger version enters the wormhole along the right-hand route and must appear in the past on the right, not on the left:

OB⟶T⟶OA.(7)O_B \longrightarrow\mathrm{T} \longrightarrow O_A. \tag*{(7)}

Therefore,

OB⟷OA.(8)O_B \longleftrightarrow O_A. \tag*{(8)}

Neither classical alternative reproduces its own initial conditions: each returns the older particle to the opposite sector. Thus the chosen one-dimensional interaction and two-channel geometry admit no classically self-consistent individual history.

There are at least two possible quantum resolutions of this classical self-prevention paradox (see Figure 1). The first possibility involves many-worlds branching:

Figure 1. Schematic wave-packet trajectories through the encounter and the two return channels, including reflection and tunneling (time increases upward). Blue lines represent the younger particle YY, with reflected and transmitted branches YRY_R and YTY_T. The green line represents OBO_B, which emerges on the left and does not interact with YY. In this case, YY passes unchanged; the two labels YY mark its incoming and outgoing portions. Orange lines represent OAO_A, which approaches from the right, and its outgoing branches after reflection or transmission in the relative motion. Dashed blue and orange lines indicate tunneling paths. Dashed grey curves indicate the return routes through the wormhole, connecting the outgoing younger particle to the incoming older particle at AA (right) or BB (left). The circle marks the interaction region.

OA1⟶R⟶OB2,OB2⟶T⟶OA1,(9)\begin{aligned} O_A^{1} &\longrightarrow\mathrm{R} \longrightarrow O_B^{2},\\ O_B^{2} &\longrightarrow\mathrm{T} \longrightarrow O_A^{1}, \tag*{(9)} \end{aligned}

where the superscripts indicate different branches of the wave density matrix. This is the natural interpretation of the Deutsch model.

The second possibility, explored in this article, is a self-consistent individual history without many-worlds branching. When the older version is on the right, the younger particle may tunnel through the repulsive interaction and take the right-hand route:

OA→tunnelingT→CTCOA.(10)O_A \xrightarrow{\text{tunneling}} \mathrm{T} \xrightarrow{\mathrm{CTC}} O_A. \tag*{(10)}

This is the candidate individual consistent history. It must satisfy the ontic condition of § 2.3, while the full two-channel quantum state must simultaneously satisfy the density-matrix condition of § 2.1.

Hamiltonian and Scattering

With the global routing specified by Eq. (5) and the return maps in § 3.1.1, the local collision is described by the effective two-particle Hamiltonian

H=−ℏ22m(∂xY2+∂xO2)+U(xY−xO),(11)H=-\frac{\hbar^2}{2m}\left(\partial_{x_Y}^2+\partial_{x_O}^2\right)+U(x_Y-x_O), \tag*{(11)}

where UU is a repulsive interaction localized in the relative coordinate and has a finite maximum Umax⁡U_{\max}. Define

Q=xY+xO2,ξ=xY−xO;(12)Q=\frac{x_Y+x_O}{2},\qquad\xi=x_Y-x_O; \tag*{(12)}

the Hamiltonian then becomes

H=−ℏ24m∂Q2−ℏ2m∂ξ2+U(ξ).(13)H=-\frac{\hbar^2}{4m}\partial_Q^2-\frac{\hbar^2}{m}\partial_\xi^2+U(\xi). \tag*{(13)}

The free evolution of the center of mass separates exactly at the Hamiltonian level. Factorization of the two-particle wave function itself in Q,ξQ,\xi generally requires additional properties of the incoming packets; in the symmetric case, it holds for two Gaussians with matching complex widths.

We choose the relative-motion energy to be below the barrier threshold, Erel<Umax⁡E_{\mathrm{rel}}<U_{\max}, so that classical transmission is forbidden: in the classical limit the barrier is impenetrable and reflects the younger version. Let PP be the quantum transmission probability when the older version is on the right, with 0<P<10<P<1; in the regime considered, this process is tunneling.

Approximations

Our aim is to exhibit a self-consistent individual history in an analytically tractable regime. This requires solving both the quantum and ontic consistency conditions, but not a complete treatment of arbitrary scattering states.

Even for simple interactions, such as rectangular or delta barriers, an analytic solution of the stationary scattering problem does not directly solve the full consistency problem. Wave-packet evolution involves momentum integrals, and matching the outgoing and returning states imposes additional integral relations. The return need not preserve Gaussian form or factorization in center-of-mass and relative coordinates.

Direct numerical evolution of the full two-particle density matrix involves four spatial arguments and time, together with the return consistency condition. Reduced numerical descriptions are possible but lie beyond the present study.

We therefore adopt the assumptions below, which permit an explicit analytic construction.

(A1) Narrow-band Gaussian packets. The incoming states of the two versions are Gaussian and initially factorized in Y,OY,O. The relative spectrum is centered at k0k_0 with width σk≪k0\sigma_k\ll k_0. The spectrum of the transmitted branch is

χ~T(k)=t(k)χ~in(k),(14)\widetilde{\chi}_{T}(k)=t(k)\widetilde{\chi}_{\mathrm{in}}(k), \tag*{(14)}

and that of the reflected branch is χ~R(k)=rsc(k)χ~in(k)\widetilde{\chi}_{R}(k)=r_{\mathrm{sc}}(k)\widetilde{\chi}_{\mathrm{in}}(k), where tt and rscr_{\mathrm{sc}} are complex scattering amplitudes. Sufficient conditions for weak distortion are

∣dln⁡∣t∣dk∣∣k0σk≪1,∣d2ln⁡tdk2∣∣k0σk2≪1,(15)\left.\left|\frac{d\ln|t|}{dk}\right| \right|_{k_0}\sigma_k\ll1,\qquad \left.\left|\frac{d^2\ln t}{dk^2}\right| \right|_{k_0}\sigma_k^2\ll1, \tag*{(15)}

with analogous conditions for rsc(k)r_{\mathrm{sc}}(k). The linear spectral phase is incorporated into the spatial shift of the packet. Thus P≃∣t(k0)∣2P \simeq|t(k_0)|^2; more precisely, PP is the average of ∣t(k)∣2|t(k)|^2 over the relative momentum spectrum.

(A2) Factorization of each outgoing branch. For equal masses and nearly constant scattering amplitudes, each outgoing branch approximately factorizes in the single-particle coordinates. In the transmitted channel, the younger particle retains its own packet; in the reflected channel, it acquires the older particle’s packet:

ρY,Tout≃ρY,ρY,Rout≃ρOA.(16)\rho^{\mathrm{out}}_{Y,T} \simeq\rho_Y,\qquad\rho^{\mathrm{out}}_{Y,R} \simeq\rho_{O_A}. \tag*{(16)}

Within the narrow-band approximation, this also holds for different incoming widths: equal widths are not required for factorization of each individual outgoing branch. The full superposition of transmitted and reflected amplitudes is, in general, entangled.

(A3) Branch separation. The two spatial branches OA,OBO_A,O_B are orthogonal to a good approximation, and the outgoing channels likewise separate after scattering. Their contributions to the density matrix are then treated as distinct blocks. In the Bohmian account with a density matrix, branches that have no support at the actual configuration do not affect the local velocity field; interference during the collision, before the outgoing waves separate, nevertheless remains important [29]. The correlation between the younger particle’s outgoing direction and the older particle’s emergence location acts as a record of the outcome (analogous to a which-channel record). The partial trace removes inter-channel coherence in the reduced description; it is not, by itself, an additional physical measurement.

(A4) Fast collision (used only for the ontic trajectory with different widths, § 5.2). During the local interaction, the normalized center-of-mass coordinate changes negligibly. At each approximately fixed center-of-mass coordinate, the conditional transmission probability is taken to be PP, independent of that coordinate, and the normalized transmitted conditional packet is approximated by a Gaussian. These assumptions make a one-dimensional conditional quantile map possible; they do not follow solely from narrow-band scattering.

Self-Consistent Density Matrix

We seek a fixed point of Deutsch’s condition (2) in the reduced, block-diagonal two-channel model. A subtlety is that the right-hand return channel receives contributions from two different processes: unobstructed passage when the older particle is on the left, and tunneling when it is on the right. We must justify identifying their outgoing younger-particle states before combining the contributions.

Self-Consistency of Gaussian States

Let ρf\rho_f be the normalized outgoing state of the younger particle after unobstructed passage, and let ρt\rho_t be its normalized reduced state after tunneling, with the outgoing older particle traced out. Both are defined at the same asymptotic matching surface, before propagation through the right-hand CTC route. In general, ρf\rho_f and ρt\rho_t need not be identical: their histories involve different correlations with the older particle, and the scattering amplitude may introduce a momentum-dependent phase.

The assumed narrow-band regime makes their leading-order equality explicit. For equal masses, the transmission amplitude depends on the relative wave number k=(pY−pO)/(2ℏ)k=(p_Y-p_O)/(2\hbar), where pY,pOp_Y,p_O are the single-particle momenta. Over the incident spectral width σk\sigma_k, require not merely weak variation of ∣t(k)∣|t(k)| but also small variation of its phase relative to the freely propagated packet:

σk∣dln⁡tdk∣k0≪1,σk2∣d2ln⁡tdk2∣k0≪1.(17)\sigma_k \left|\frac{d\ln t}{dk}\right|_{k_0} \ll1,\qquad\sigma_k^2 \left|\frac{d^2\ln t}{dk^2}\right|_{k_0} \ll1. \tag*{(17)}

The first condition includes the small-displacement requirement σk∣φ′(k0)∣≪1\sigma_k|\varphi'(k_0)|\ll1, where φ(k)=arg⁡t(k)\varphi(k)=\arg t(k); a constant phase is immaterial. We assume analogous weak distortion for reflection, together with approximate factorization of the individual outgoing branches as in Eq. (16). These conditions are fulfilled, for example, in the contact-barrier limit U(ξ)=gδ(ξ)U(\xi)=g\delta(\xi), g>0g>0, where

t(k)=kk+iγ,γ=mg2ℏ2,∣dln⁡tdk∣k0=1−Pk0.(18)t(k)=\frac{k}{k+i\gamma},\qquad\gamma=\frac{mg}{2\hbar^2},\qquad\left|\frac{d\ln t}{dk}\right|_{k_0}=\frac{\sqrt{1-P}}{k_0}. \tag*{(18)}

Thus σk/k0≪1\sigma_k/k_0\ll1 controls the relative variation of the full complex amplitude, including its phase, for this illustrative localized interaction, even if PP is small. For wave packets sufficiently concentrated near k0k_0, the resulting normalized younger-particle states satisfy ∥ρt−ρf∥1=O(σk/k0)\|\rho_t-\rho_f\|_1=O(\sigma_k/k_0) in trace norm (with the return-route and free phases matched). A sufficiently narrow, nonresonant finite barrier can be used instead, provided the same spectral conditions are checked; spatial localization alone does not guarantee them.

In the constant-amplitude limit the transmitted two-particle wave function, after normalization and removal of free propagation, is unchanged to leading order. Consequently, its younger-particle reduced state is independent, to this order, of the detailed incoming state or spatial translation of the older packet. We therefore have

ρf≃ρt≃ρY.(19)\rho_f\simeq\rho_t\simeq\rho_Y. \tag*{(19)}

The right-hand return map UT\mathcal{U}_T acts on either outgoing younger state. The left-hand route applies after reflection; for equal masses, the reflected younger particle approximately inherits the older right-hand packet [Eq. (16)]. The resulting leading-order packet-shape conditions are

ρOA≃UT(ρY),ρOB≃UR(ρOA).(20)\rho_{O_A}\simeq\mathcal{U}_T(\rho_Y),\qquad\rho_{O_B}\simeq\mathcal{U}_R(\rho_{O_A}). \tag*{(20)}

These conditions do not require equal widths: the return maps may broaden the packets and generate quadratic phases, as long as the incoming states remain narrow-band and the branchwise scattering approximation applies. The well-separated spatial sectors OAO_A and OBO_B are distinct from the two nearly coincident contributions ρf\rho_f and ρt\rho_t within the right-hand sector. No spatial separation between ρf\rho_f and ρt\rho_t is assumed.

Stationary Density Matrix

Let pA,pBp_A,p_B be the probabilities of finding the incoming older particle in the right and left spatial sectors, respectively, with $p_A+p_B=1. A younger particle returns on the right either after unobstructed passage from the left-older sector or after tunneling from the right-older sector. These channel transitions are illustrated schematically in Figure 1. Therefore

pA′=pB+PpA,p'_A=p_B+Pp_A,
pB′=(1−P)pA.(21)p'_B=(1-P)p_A. \tag*{(21)}

For an approximately state-independent transmission probability PP, stationarity gives pA∗=1/(2−P)p_A^*=1/(2-P) and pB∗=1−pA∗p_B^*=1-p_A^*.

To verify stationarity of the density operators, we first write the unnormalized right-hand output before identifying ρt\rho_t with ρf\rho_f:

WA′=pBUT(ρf)+pAPUT(ρt),(22)W'_A = p_B\mathcal{U}_T(\rho_f) + p_A P\mathcal{U}_T(\rho_t), \tag*{(22)}

where WA′=pA′ρOAW'_A = p'_A\rho_{O_A}. The two terms correspond to unobstructed passage and tunneling, respectively, with the outgoing older particle traced out in each case. If ρf\rho_f and ρt\rho_t differ, the returned state is a mixture of different packets.

Under the controlled leading-order identification ρf≃ρt≃ρY\rho_f \simeq\rho_t \simeq\rho_Y, both contributions produce the same returned state ρ0:=UT(ρY)\rho_0 := \mathcal{U}_T(\rho_Y). The resulting candidate stationary operator is

ρ~O∗≃12−PρOA⊕1−P2−PρOB.(23)\tilde{\rho}^{*}_{O} \simeq\frac{1}{2-P}\rho_{O_A} \oplus\frac{1-P}{2-P}\rho_{O_B}. \tag*{(23)}

Here ⊕\oplus denotes orthogonal right and left spatial sectors. With WA=pAρOAW_A = p_A\rho_{O_A}, WB=pBρOBW_B = p_B\rho_{O_B}, the block map becomes

WA′≃(Tr⁡WB+PTr⁡WA)ρ0,WB′≃(1−P)UR(WA).(24)\begin{aligned} W'_A &\simeq(\operatorname{Tr} W_B + P\operatorname{Tr} W_A)\rho_0,\\ W'_B &\simeq(1-P)\mathcal{U}_R(W_A). \tag*{(24)} \end{aligned}

Substituting the stationary weights and using Eq. (20) gives WA′≃WAW'_A \simeq W_A and WB′≃WBW'_B \simeq W_B within this leading-order two-channel approximation.

The stationary weights remain valid when ρf\rho_f and ρt\rho_t differ, provided PP is state-independent. Replacing their mixture by a single Gaussian packet requires the complex-amplitude conditions above; otherwise, each spatial sector may contain several packets. The ontic analysis in § 5 uses the leading-order Gaussian state, since trace-norm closeness alone does not control Bohmian trajectories in remote Gaussian tails.

Self-Consistent Bohmian Trajectory

The density-matrix solution of § 4 fixes the statistical CTC state. We next impose the additional ontic requirement (4) on the branch in which the older particle appears on the right and the younger particle tunnels to the right. We first treat equal widths, then allow arbitrary positive width ratios compatible with the Gaussian and fast-collision approximations. No explicit propagation time or spreading law is needed for the latter result.

Equal Gaussian Widths

Relative-Coordinate Quantile

First consider the symmetric case in which the packets of the younger particle and the older particle on the right have matching complex Gaussian width parameters. Let σ>0\sigma> 0 denote their common spatial standard deviation before the collision. Their actual positions are

XY=XˉY+σzY,XO=XˉO+σzO.(25)X_Y = \bar{X}_Y + \sigma z_Y,\qquad X_O = \bar{X}_O + \sigma z_O. \tag*{(25)}

Introduce the normalized variables

zQ=zY+zO2,r=zY−zO2.(26)z_Q = \frac{z_Y + z_O}{\sqrt{2}},\qquad r = \frac{z_Y - z_O}{\sqrt{2}}. \tag*{(26)}

For independent incoming Gaussians, both variables are standard normally distributed. Denote the standard normal cumulative distribution function by

Φ(a)=12[1+erf⁡(a2)].(27)\Phi(a)=\frac{1}{2}\left[1+\operatorname{erf}\left(\frac{a}{\sqrt{2}}\right)\right]. \tag*{(27)}

Its density is ϕ(a)=e−a2/2/2π\phi(a)=e^{-a^2/2}/\sqrt{2\pi}.

In one-dimensional Bohmian relative motion, trajectories do not cross [30]. This ordering underlies the quantile description of Bohmian trajectories [31] and the identification of incoming positions leading to transmission [32]. With the chosen orientation, the transmitted branch corresponds to incoming quantiles qin>1−Pq_{\mathrm{in}}>1-P, where qin=Φ(r)q_{\mathrm{in}}=\Phi(r). Once the branches have separated, one-dimensional ordering fixes the quantile qTq_{\mathrm{T}} of the actual trajectory within the normalized transmitted packet. The additional assumption that this packet remains Gaussian relates that quantile to its normalized outgoing relative coordinate rTr_{\mathrm{T}}:

qT=Φ(rT)=Φ(r)−1+PP.(28)q_{\mathrm{T}}=\Phi(r_{\mathrm{T}})=\frac{\Phi(r)-1+P}{P}. \tag*{(28)}

Here Φ−1\Phi^{-1}, used below, denotes the inverse cumulative distribution function.

Ontic Fixed Point

In the symmetric factorized problem, the normalized center-of-mass coordinate is preserved, so

zYout=zQ+rT2.(29)z_{\mathrm{Y}}^{\mathrm{out}}=\frac{z_{\mathrm{Q}}+r_{\mathrm{T}}}{\sqrt{2}}. \tag*{(29)}

By assumption, the CTC map preserves the normalized internal coordinate of the Gaussian packet, even under free spreading. Therefore, trajectory closure requires

zO=zYout.(30)z_{\mathrm{O}}=z_{\mathrm{Y}}^{\mathrm{out}}. \tag*{(30)}

To visualize this condition, define the return map R(zO)=zYout\mathcal{R}(z_{\mathrm{O}})=z_{\mathrm{Y}}^{\mathrm{out}} at fixed PP and zYz_{\mathrm{Y}}. For equal packet widths, Eqs. (26) and (28) give the explicit map

R(zO)=zY+zO2+12Φ−1[Φ((zY−zO)/2)−1+PP].(31)\mathcal{R}(z_{\mathrm{O}})=\frac{z_{\mathrm{Y}}+z_{\mathrm{O}}}{2}+\frac{1}{\sqrt{2}}\Phi^{-1}\left[\frac{\Phi\left((z_{\mathrm{Y}}-z_{\mathrm{O}})/\sqrt{2}\right)-1+P}{P}\right]. \tag*{(31)}

It is defined on the transmitted branch, zO<zY−2Φ−1(1−P)z_{\mathrm{O}}<z_{\mathrm{Y}}-\sqrt{2}\Phi^{-1}(1-P). The outgoing younger particle becomes the incoming older particle after return through the CTC. An individual history therefore closes at an intersection of R(zO)\mathcal{R}(z_{\mathrm{O}}) with the diagonal R=zO\mathcal{R}=z_{\mathrm{O}}, as illustrated in Figure 2.

Figure 2. Ontic return map (31) in the symmetric Gaussian model with equal packet widths, for P=0.2P = 0.2 and zY=0z_Y = 0. The solid curve gives the returned older coordinate R(zO)=zYout\mathcal{R}(z_O) = z_Y^{\mathrm{out}} as a function of the incoming older coordinate zOz_O. Its intersection with the dashed diagonal R=zO\mathcal{R} = z_O gives the self-consistent history, with zO∗≃−1.368z_O^* \simeq-1.368. Equal axis scales make the diagonal a 45∘45^\circ line. The map is defined on the transmitted branch, zO<zY−2Φ−1(1−P)z_O < z_Y - \sqrt{2}\Phi^{-1}(1-P); the shaded region lies outside this domain.

Equations (26) and (30) imply rT=−rr_T=-r. Substitution into Eq. (28), using Φ(−r)=1−Φ(r)\Phi(-r)=1-\Phi(r), gives

r∗=Φ−1(11+P).(32)r_*=\Phi^{-1}\left(\frac{1}{1+P}\right). \tag*{(32)}

For a given internal coordinate zYz_Y of the younger particle, we obtain

zO∗=zY−2Φ−1(11+P).(33)z_O^*=z_Y-\sqrt{2}\Phi^{-1}\left(\frac{1}{1+P}\right). \tag*{(33)}

The resulting fixed point lies specifically in the tunneling channel:

Φ(r∗)−(1−P)=P21+P>0.(34)\Phi(r_*) - (1-P) = \frac{P^2}{1+P} > 0. \tag*{(34)}

As P→0+P \to0^+, the coordinate r∗r_* moves into the right tail of the normal distribution, while zO∗→−∞z^*_O \to-\infty for fixed zYz_Y. In the strict classical limit P=0P=0, no finite tunneling fixed point exists.

Figure 3 shows the dependence of this ontic fixed point on the tunneling probability for equal packet widths and several fixed values of zYz_Y. The zY=0z_Y=0 curve includes the fixed point illustrated in Figure 2 at P=0.2P=0.2.

Figure 3. Ontic fixed point zO∗z^*_O as a function of the tunneling probability PP, evaluated from Eq. (33) in the symmetric Gaussian model with equal packet widths (λ=1\lambda=1) for zY=−2,−1,0,1,2z_Y=-2,-1,0,1,2. The probability axis is logarithmic. As P→0+P \to0^+, zO∗→−∞z^*_O \to-\infty; no finite fixed point exists at P=0P=0. Open circles mark the limiting values zO∗=zYz^*_O=z_Y as P→1−P \to1^-. The curves differ only by the vertical shift zYz_Y.

Convergence of Direct Iterations of the Return Map

The return map (31) also defines a numerical procedure for finding an ontic fixed point. Here we examine its local convergence in the equal-width Gaussian model, with P∈(0,1)P \in(0,1) and zYz_Y held fixed. The map is smooth on its transmitted domain, and its fixed point lies strictly inside that domain by Eq. (34).

Consider the direct iteration

zO(n+1)=R(zO(n)).(35)z_O^{(n+1)}=\mathcal{R}\left(z_O^{(n)}\right). \tag*{(35)}

Differentiating Eq. (31) and using rT=−r∗r_T=-r_* at the fixed point gives its linear error multiplier:

R′(zO∗)=−1−P2P.(36)\mathcal{R}'(z^*_O)=-\frac{1-P}{2P}. \tag*{(36)}

Thus direct iteration is locally convergent for P>1/3P > 1/3, with errors alternating in sign. For 0<P<1/30 < P < 1/3, the fixed point is locally repelling under this iteration. In particular, for the example P=0.2P = 0.2 in Figure 2, the linear error multiplier is −2-2.

At P=1/3P = 1/3, the multiplier is −1-1, so linearization alone is inconclusive. With δn=zO(n)−zO∗\delta_n = z_O^{(n)} - z_O^*, expanding two successive iterations yields

δn+2=δn+2(1−r∗2)δn3+O(δn4),r∗=Φ−1(3/4)≃0.67449.(37)\delta_{n+2} = \delta_n + 2(1-r_*^2)\delta_n^3 + O(\delta_n^4), \qquad r_* = \Phi^{-1}(3/4) \simeq0.67449. \tag*{(37)}

The positive cubic coefficient implies weak local repulsion at this boundary value as well.

Convergence can be restored by mixing the returned coordinate with the previous iterate, also called under-relaxation:

zO(n+1)=(1−ω)zO(n)+ωR(zO(n)),0<ω≤1.(38)z_O^{(n+1)} = (1-\omega)z_O^{(n)} + \omega R\left(z_O^{(n)}\right), \qquad0 < \omega\le1. \tag*{(38)}

For any nonzero ω\omega, this modification preserves the fixed-point equation. Its local error multiplier is

qω=1−ω1+P2P.(39)q_\omega= 1-\omega\frac{1+P}{2P}. \tag*{(39)}

The condition ∣qω∣<1|q_\omega| < 1 gives the sufficient local convergence range

0<ω<4P1+P,0<ω≤1.(40)0 < \omega< \frac{4P}{1+P}, \qquad0 < \omega\le1. \tag*{(40)}

For every P∈(0,1)P \in(0,1), choosing ω=2P/(1+P)\omega= 2P/(1+P) sets qω=0q_\omega= 0 and gives at least quadratic local convergence (ω=1/3\omega= 1/3 for P=0.2P = 0.2).

More generally, for a smooth scalar return map with derivative d=R′(zO∗)d = R'(z_O^*), mixing gives the local criterion ∣1−ω+ωd∣<1|1-\omega+\omega d| < 1. This permits a suitable mixing parameter to stabilize an oscillatory direct iteration when d<0d < 0, although the appropriate parameter depends on the map. The explicit bounds above apply to the equal-width model. All convergence statements are local: a sufficiently close initial iterate remains within a contracting neighborhood in the transmitted domain, whereas distant iterates may leave that domain.

Different Widths: Fast-Collision Approximation

Normalized Variables

Let λ=σO/σY>0\lambda= \sigma_O/\sigma_Y > 0 denote the ratio of the spatial widths of the older right-hand and younger packets at the collision. The return maps may generate such a difference and a nonzero quadratic phase. At the start of the collision,

XY=XˉY+σYzY,XO=XˉO+σOzO.(41)X_Y = \bar{X}_Y + \sigma_Y z_Y, \qquad X_O = \bar{X}_O + \sigma_O z_O. \tag*{(41)}

Let Σ=σY2+σO2\Sigma= \sqrt{\sigma_Y^2+\sigma_O^2} and define the orthogonal normalized coordinates

s=σYzY+σOzOΣ,s = \frac{\sigma_Y z_Y+\sigma_O z_O}{\Sigma},
u=σOzY−σYzOΣ.(42)u = \frac{\sigma_O z_Y-\sigma_Y z_O}{\Sigma}. \tag*{(42)}

Here QQ, ξ\xi are evaluated at the actual Bohmian positions xY=XYx_Y = X_Y, xO=XOx_O = X_O. The quantity s=2(Q−Qˉ)/Σs = 2(Q-\bar{Q})/\Sigma is the normalized center-of-mass coordinate. The variable uu is the normalized relative coordinate at fixed ss; indeed,

ξ−ξˉ=σY2−σO2Σs+2σYσOΣu.(43)\xi-\bar{\xi}=\frac{\sigma_Y^2-\sigma_O^2}{\Sigma}s+\frac{2\sigma_Y\sigma_O}{\Sigma}u. \tag*{(43)}

For factorized incoming Gaussians, ss and uu are independent and standard normally distributed.

Conditional Quantile and Closure Equation

Under the fast-collision and conditional Gaussian assumptions (A4), the normalized transmitted relative coordinate is mapped according to

uT=Φ−1(Φ(u)−1+PP).(44)u_{\mathrm{T}}=\Phi^{-1}\left(\frac{\Phi(u)-1+P}{P}\right). \tag*{(44)}

Within the same approximation, the incoming and outgoing internal coordinates of the younger particle are related by

zYout=s+λuT1+λ2.(45)z_Y^{\mathrm{out}}=\frac{s+\lambda u_{\mathrm{T}}}{\sqrt{1+\lambda^2}}. \tag*{(45)}

Since free Gaussian propagation along the CTC preserves the normalized internal coordinate of an individual particle, ontic self-consistency again requires zO=zYoutz_O=z_Y^{\mathrm{out}}. Substituting the coordinate transformation (42) into this condition and using Eq. (44) gives a scalar closure equation with coefficients

α=λ−1+1λ,β=(λ−1)1+λ2λ.(46)\alpha=\lambda-1+\frac{1}{\lambda},\qquad\beta=\frac{(\lambda-1)\sqrt{1+\lambda^2}}{\lambda}. \tag*{(46)}

The resulting equation is

Φ(u)−PΦ(βzY−αu)=1−P.(47)\Phi(u)-P\Phi(\beta z_Y-\alpha u)=1-P. \tag*{(47)}

Once the solution u∗u_* is found, the older coordinate is given by

zO∗=λzY−1+λ2 u∗.(48)z_O^*=\lambda z_Y-\sqrt{1+\lambda^2}\,u_*. \tag*{(48)}

Existence of a Root in the Gaussian Map

For any λ>0\lambda>0, we have α=λ+λ−1−1≥1\alpha=\lambda+\lambda^{-1}-1\ge1. Define

f(u)=Φ(u)−1+P−PΦ(βzY−αu).(49)f(u)=\Phi(u)-1+P-P\Phi(\beta z_Y-\alpha u). \tag*{(49)}

Differentiation gives

f′(u)=ϕ(u)+Pαϕ(βzY−αu)>0.(50)f'(u)=\phi(u)+P\alpha\phi(\beta z_Y-\alpha u)>0. \tag*{(50)}

As u→−∞u\to-\infty, f(u)→−1f(u)\to-1, whereas as u→+∞u\to+\infty, f(u)→Pf(u)\to P. Therefore, for any P∈(0,1)P\in(0,1), λ>0\lambda>0, and fixed zYz_Y, Eq. (47) has a unique finite root in this approximate map. Equation (47) also implies Φ(u∗)>1−P\Phi(u_*)>1-P, so the root lies in the transmitted channel.

For λ=1\lambda=1, we have α=1\alpha=1 and β=0\beta=0, so the general fixed-point equation (47) reduces to the symmetric result (32)–(33). Thus, unequal widths do not remove the root within the fast-collision Gaussian map.

Discussion and Conclusions

In this work, we considered the self-prevention paradox in Deutsch’s time-loop model, supplemented with a Bohmian description of an individual particle. In the classical version of the model, no self-consistent trajectory exists: the spatial return channels lead to a contradiction. In the quantum description, a stationary density matrix is found; in the Bohmian description, there is a self-consistent individual history in which the younger particle tunnels through its interaction with the older one. For any nonzero tunneling probability, the resulting history must pass through this channel, even though such an event may be extremely unlikely in ordinary scattering. At zero tunneling probability, this type of individual solution disappears. This realizes the “banana peel mechanism” [23]: the global requirement of self-consistency selects a locally unlikely event as necessary for a contradiction-free history.

The key methodological step is the choice of approximations that make both the quantum and ontic self-consistency conditions analytically tractable. Within this approximation, we explicitly construct a stationary density matrix and a self-consistent individual history, using the quantile description of Bohmian motion for the latter. This provides an analytic demonstration of the tunneling resolution of the paradox.

The result is established for the prescribed one-dimensional, two-channel model under the assumptions of § 3.3. Whether such individual histories persist beyond these approximations remains an open question. A relativistic extension would additionally require a specification of the spacetime structure used to define Bohmian configurations, including the role of spacelike foliations [33].

These results raise a broader question: how do interpretations of quantum mechanics affect physical reality in the presence of closed timelike curves? In ordinary circumstances, without closed timelike curves, standard interpretations of quantum mechanics agree on observable predictions. Deutsch’s time-loop model, however, requires an additional self-consistency condition, making the question of what counts as a physical history consequential. In a many-worlds reading, the solution is described in terms of interacting parallel branches of history. In a Bohmian reading, one can speak of the trajectory of an individual particle; in the model considered here, that trajectory passes through a low-probability tunneling channel. Thus, the presence of closed timelike curves makes interpretation more than a way of talking about the quantum state: it affects the physical account of how the paradox is resolved and what kind of self-consistent history is realized.

This motif can be cautiously extended to a macroscopic time-travel paradox. In discussions of the “grandfather paradox,” the banana peel mechanism means that unlikely accidents may stop the time traveler [23]. If, however, the self-prevention is caused by a reliable, deterministic mechanism, it is harder to see how that mechanism could fail in a classical description. In a many-worlds reading of Deutsch’s approach, the contradiction is resolved through interactions between parallel branches of history. In a Bohmian reading, one can imagine a different scenario: individual components of the mechanism tunnel, causing it to fail. This is only a possible analogy with the result of our model, not a conclusion about the feasibility of closed timelike curves or macroscopic tunneling.

Use of Artificial Intelligence

OpenAI’s ChatGPT was used to support analytical derivations, mathematical calculations, consistency checks, and the preparation and editing of the manuscript. The author reviewed and verified all AI-assisted content, independently reproduced the central calculations, and takes full responsibility for the scientific content and conclusions of this work.

Notation and Abbreviations

CTC

closed timelike curve. Y,OY,O

younger and older interacting subsystems; in the model, the two versions of the particle. OAO_A is the right-hand older sector and OBO_B is the left-hand older sector.

T,RT,R

transmission and reflection outcomes of the younger particle; PP – the tunneling probability when the older particle occupies sector AA.

xY,xOx_Y,x_O

coordinate arguments of the two-particle wave function; XY,XOX_Y,X_O – actual Bohmian positions; XˉY,XˉO\bar{X}_Y,\bar{X}_O – packet centers.

m,ℏ,H,Um,\hbar,H,U

particle mass, reduced Planck constant, Hamiltonian and repulsive interaction potential, respectively.

Q,ξQ,\xi

center-of-mass and relative coordinates; Qˉ,ξˉ\bar{Q},\bar{\xi} – their mean positions; Ψ\Psi – the two-particle wave function.

k,k0,σkk,k_0,\sigma_k

relative wave number, its central value and spectral width; pY,pOp_Y,p_O – single-particle momenta; t,rsct,r_{\mathrm{sc}} – complex transmission and reflection amplitudes; χin\chi_{\mathrm{in}} – incoming spectral amplitude.

ρY,ρO\rho_Y,\rho_O

prescribed younger and self-consistent older density matrices; ρOA,ρOB\rho_{O_A},\rho_{O_B} – normalized older states in the two spatial sectors; DρY\mathcal{D}_{\rho_Y} – quantum map for a cycle.

ρf,ρt,ρ0\rho_f,\rho_t,\rho_0

normalized outgoing younger states after unobstructed passage and tunneling, and the common returned state ρ0=UT(ρY)\rho_0=\mathcal{U}_T(\rho_Y), respectively.

UT,UR\mathcal{U}_T,\mathcal{U}_R

return channels associated with the transmitted and reflected outcomes; pA,pBp_A,p_B – sector weights; WA,WBW_A,W_B – their unnormalized density-matrix blocks.

V,GV,\mathcal{G}

collision unitary and complete quantum return channel; Tr⁡O\operatorname{Tr}_O – partial trace over the outgoing older subsystem.

σY,σO,σ\sigma_Y,\sigma_O,\sigma

positive spatial standard deviations of the younger and older packets and their common value in the symmetric limit; Σ=σY2+σO2\Sigma=\sqrt{\sigma_Y^2+\sigma_O^2} – combined width; λ=σO/σY\lambda=\sigma_O/\sigma_Y – width ratio.

zY,zO,zQ,r,s,uz_Y,z_O,z_Q,r,s,u

normalized internal coordinates, including symmetric center-of-mass and relative coordinates zQ,rz_Q,r, and unequal-width coordinates s,us,u; rT,uTr_T,u_T – transmitted relative coordinates.

qin,qTq_{\mathrm{in}},q_T

incoming relative-coordinate quantile and the quantile within the normalized transmitted packet.

R\mathcal{R} ontic return map R(zO)=zYout\mathcal{R}(z_O)=z_Y^{\mathrm{out}}, at fixed PP, zYz_Y, in the equal-width model.

ω,qω\omega, q_\omega

mixing parameter and local error multiplier of the mixed return-map iteration.

Φ,ϕ,erf⁡\Phi, \phi, \operatorname{erf}

standard normal cumulative distribution, its density and the error function; Φ−1\Phi^{-1} – the inverse normal cumulative distribution.

α,β\alpha, \beta

scalar coefficients in Eq. (46); ff – auxiliary function; superscripts in and out – incoming and outgoing quantities; * – value at the self-consistent solution.

W(q,q′)W(q,q')

full configuration-space density matrix in the guiding equation; Q=(Q1,…,QN)\mathbf{Q}=(Q_1,\ldots,Q_N) – actual configuration; mkm_k – mass of particle kk; ∇qk\nabla_{q_k} – its configuration gradient; C\mathcal{C} – geometric return map; XYout,XOin\mathbf{X}_Y^{\mathrm{out}},\mathbf{X}_O^{\mathrm{in}} – actual outgoing younger and incoming older configurations.

References

  1. [1]Deutsch, D. Quantum mechanics near closed timelike lines. Physical Review D 44, 3197–3217 (1991). DOI.pubmed.ncbi.nlm.nih.gov/10013776
  2. [2]Dunlap, L. The metaphysics of D-CTCs: On the underlying assumptions of Deutsch’s quantum solution to the paradoxes of time travel. Studies in History and Philosophy of Modern Physics 56, 39–47 (2016). DOI.arxiv.org/abs/1510.02742
  3. [3]de Broglie, L. La nouvelle dynamique des quanta. In: Électrons et photons: rapports et discussions du cinquième Conseil de physique Solvay, 105–132 (1928).
  4. [4]Bohm, D. A Suggested Interpretation of the Quantum Theory in Terms of “Hidden” Variables. I. Physical Review 85, 166–179 (1952). DOI.DOI
  5. [5]Gödel, K. An Example of a New Type of Cosmological Solutions of Einstein’s Field Equations of Gravitation. Reviews of Modern Physics 21, 447–450 (1949). DOI.
  6. [6]Luminet, J.-P. Closed Timelike Curves, Singularities and Causality: A Survey from Gödel to Chronological Protection. Universe 7, 12 (2021). DOI.DOI
  7. [7]Hawking, S. W. Chronology protection conjecture. Physical Review D 46, 603–611 (1992). DOI.
  8. [8]Morris, M. S., Thorne, K. S., Yurtsever, U. Wormholes, Time Machines, and the Weak Energy Condition. Physical Review Letters 61, 1446–1449 (1988). DOI.pubmed.ncbi.nlm.nih.gov/10038800
  9. [9]Echeverria, F., Klinkhammer, G., Thorne, K. S. Billiard balls in wormhole spacetimes with closed timelike curves: Classical theory. Physical Review D 44, 1077–1099 (1991). DOI.pubmed.ncbi.nlm.nih.gov/10013968
  10. [10]Friedman, J. L., Morris, M. S., Novikov, I. D., Echeverria, F., Klinkhammer, G., Thorne, K. S., Yurtsever, U. Cauchy Problem in Spacetimes with Closed Timelike Curves. Physical Review D 42, 1915–1930 (1990). DOI.pubmed.ncbi.nlm.nih.gov/10013039
  11. [11]Novikov, I. D. Time Machine and Self-Consistent Evolution in Problems with Self-Interaction. Physical Review D 45, 1989–1994 (1992). DOI.pubmed.ncbi.nlm.nih.gov/10014571
  12. [12]Friedman, J. L., Morris, M. S. Existence and Uniqueness Theorems for Massless Fields on a Class of Spacetimes with Closed Timelike Curves. Communications in Mathematical Physics 186, 495–529 (1997). DOI.DOI
  13. [13]Arefeva, I. Ya., Ishiwatari, T., Volo\-vich, I. V. Cauchy Problem on Non-globally Hyperbolic Spacetimes. Theoretical and Mathematical Physics 157, 1646–1654 (2008). DOI.DOI
  14. [14]Fewster, C. J., Higuchi, A., Wells, C. G. Classical and Quantum Initial Value Problems for Models of Chronology Violation. Physical Review D 54, 3806–3825 (1996). DOI.
  15. [15]Clément, G., Gal’tsov, D., Guenouche, M. NUT Wormholes. Physical Review D 93, 024048 (2016). DOI; arXiv:1509.07854.DOI
  16. [16]Tumulka, R., Weixler, J. Fixed Points of Completely Positive Trace-Preserving Maps in Infinite Dimension. arXiv:2411.14800 (2024). DOI.arxiv.org/abs/2411.14800
  17. [17]Everett, H., III. “Relative State” Formulation of Quantum Mechanics. Reviews of Modern Physics 29, 454–462 (1957). DOI.
  18. [18]DeWitt, B. S. Quantum Mechanics and Reality. Physics Today 23(9), 30–35 (1970). DOI.
  19. [19]Politzer, H. D. Path Integrals, Density Matrices, and Information Flow with Closed Timelike Curves. Physical Review D 49, 3981–3989 (1994). DOI.pubmed.ncbi.nlm.nih.gov/10017401
  20. [20]Lloyd, S., Maccone, L., Garcia-Patron, R., Giovannetti, V., Shikhano, Y. Quantum Mechanics of Time Travel through Post-Selected Teleportation. Physical Review D 84, 025007 (2011). DOI.DOI
  21. [21]Allen, J.-M. A. Treating Time Travel Quantum Mechanically. Physical Review A 90, 042107 (2014). DOI.DOI
  22. [22]Everett, A. Time travel paradoxes, path integrals, and the many worlds interpretation of quantum mechanics. Physical Review D 69, 124023 (2004). DOI.DOI
  23. [23]Smith, N. J. J. Bananas Enough for Time Travel? The British Journal for the Philosophy of Science 48, 363–389 (1997). DOI.DOI
  24. [24]Drezet, A., Amblard, A. Can de Broglie–Bohm Mechanics Be Considered Complete? Entropy 27, 399 (2025). DOI.DOI
  25. [25]Struyve, W. Pilot-Wave Theory and Quantum Fields. Reports on Progress in Physics 73, 106001 (2010). DOI.
  26. [26]Dürr, D., Goldstein, S., Zanghì, N. Quantum Equilibrium and the Role of Operators as Observables in Quantum Theory. Journal of Statistical Physics 116, 959–1055 (2004). DOI.
  27. [27]Beck, C., Lazarovici, D. The POVM Theorem in Bohmian Mechanics. Entropy 27, 391 (2025). DOI.DOI
  28. [28]Aleshin, A. M., Nikitin, V. V., Pronin, P. I. Beable-Guided Measurement Theory. arXiv: 2404.09934 (2024).arxiv.org/abs/2404.09934
  29. [29]Dürr, D., Goldstein, S., Tumulka, R., Zanghì, N. On the Role of Density Matrices in Bohmian Mechanics. Foundations of Physics 35, 449–467 (2005). DOI.
  30. [30]Oriols, X., Martín, F., Suñé, J. Implications of the Noncrossing Property of Bohm Trajectories in One-Dimensional Tunneling Configurations. Physical Review A 54, 2594–2604 (1996). DOI.
  31. [31]Brandt, S., Dahmen, H. D., Gjonaj, E., Stroh, T. Quantile Motion and Tunneling. Physics Letters A 249, 265–270 (1998). DOI.
  32. [32]Sanz, A. S., Miret-Artés, S. Setting Up Tunneling Conditions by Means of Bohmian Mechanics. Journal of Physics A: Mathematical and Theoretical 44, 485301 (2011). DOI.DOI
  33. [33]Dürr, D., Goldstein, S., Norsen, T., Struyve, W., Zanghì, N. Can Bohmian Mechanics Be Made Relativistic? Proceedings of the Royal Society A 470, 20130699 (2014). DOI.DOI

Paper details

vlnc:W_ygv6bqdy
Version
v1 · latest
Published
Fresh — added to the Hub within the last 30 days
Added
License
CC BY 4.0
Usage

views · 11

downloads · 0

A view counts once per browser per day, a download each time the PDF is requested.

  • Physics and Astronomy

OpenAI’s ChatGPT

View statement
Looking for related papers…
Contents