{"id":"95466ffd-c27b-4913-9e84-9db5bd25ef54","arxiv_id":"hep-ph/0012142","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"Spontaneous symmetry breaking driven by a sudden tachyonic (spinodal) instability completes within one oscillation of the field, and this tachyonic preheating is generic in hybrid inflation.","lead":"Using 3D lattice simulations, this paper argues that when a scalar field suddenly acquires a negative mass squared, symmetry breaking completes within a single oscillation rather than many. The same mechanism, named tachyonic preheating, is claimed to be generic in hybrid inflation models.","discovery_kind":"unclear","skeptic_critique":{"model":"selfhosted/deepseek-v4.1-flash","headline":"The cubic-model result — the prototype for F-term hybrid inflation — is seeded by rare high-amplitude peaks whose abundance depends on the UV tail of the initial fluctuation spectrum; the letter does not show this peak dynamics is lattice-cutoff independent.","rationale":"The reader's weakest_assumption correctly locates the load-bearing soft spot: the classical-statistical lattice evidence and its cutoff/renormalization control, with technical details deferred to [12] (in preparation, so the deferral is a legitimate concern under the citation rule). I agree with that diagnosis and sharpen it to the most specific place it bites. The quadratic-model result is well supported: the analytic estimate (Eqs. 2–3) is structurally parameter-free, the lattice snapshots in Fig. 1 show the distribution splitting into two peaks, and the mechanism (gradient energy ~ m⁴/λ comparable to V(0), so the field cannot return to φ=0) is a sound argument independent of the simulation. The cubic model is the weak link because the instability is peak-seeded rather than mass-seeded, and the paper's own text says the dynamics is dominated by rare peaks whose abundance is exponentially sensitive to the Gaussian tail. That tail is a UV/lattice-cutoff quantity, and the letter reports no cutoff study. Since the cubic potential is the stated prototype for F-term hybrid inflation, this propagates to the 'typically' claim in the abstract. This is a real but not fatal concern: the qualitative picture (fast, non-perturbative, few oscillations rather than dozens) would survive a moderate cutoff dependence, and the contrast with parametric resonance is preserved. Hence CONDITIONAL stands rather than REJECT. The paper also has independent support worth crediting: the analytic growth-time estimate, the explicit statement that the homogeneous stage 'either does not exist at all or ends after a single oscillation,' and the honest acknowledgment of the cubic-model renormalization complication. The concrete test above would settle whether the peak-seeding concern actually lands; if the concentration time and bubble statistics are stable across lattice spacings, the concern dissolves and the letter's claim is as robust as its qualitative framing suggests.","tokens_in":11285,"tokens_out":6548,"duration_ms":131245,"concrete_test":"Rerun the cubic model (4) with λ = 10⁻² at two lattice spacings (e.g. 128³ and 256³ at fixed physical volume, or vary the UV cutoff of the initial spectrum by a factor of two) and compare (i) the time at which P(φ,t) concentrates near the minimum and (ii) the number and size distribution of bubbles/peaks. If the concentration time or bubble count shifts by more than roughly 20–30%, the single-oscillation claim for the cubic/F-term case is cutoff-dependent and the 'typically' language should be qualified.","verdict_should_be":"UNCHANGED","load_bearing_attack":"In the quadratic model (1) the tachyonic mass is fixed at -m², so the growth time t* ~ (1/2m) ln(π²/λ) (Eq. 3) is robust and only logarithmically cutoff-sensitive. The cubic model (4) is different: m²(φ) = -2λvφ + 3λφ² vanishes at φ=0, so there is no fixed tachyonic mass and the instability must be seeded by the initial fluctuation spectrum. Eq. (5) gives δφ_rms ~ λv/(4π²), but the paper itself states (Sec. II.B) that fluctuations 'a few times greater than (5)' grow faster and 'may therefore have greater impact,' and that 'low probability fluctuations with δφ ≫ δφ_rms correspond to peaks of the initial Gaussian distribution.' The number density and amplitude of those peaks are controlled by the far tail of the initial Gaussian, which is set by the UV completion / lattice cutoff. The bubble dynamics of Figs. 4–5 is therefore potentially cutoff-sensitive, and no lattice-spacing or initial-spectrum variation is reported. Because the cubic potential is presented in Sec. III as the prototype for F-term hybrid inflation (g² = 2λ), the 'typically tachyonic, single oscillation' claim for F-term models inherits this sensitivity. The renormalization subtraction is acknowledged (Sec. II.B), but the cutoff dependence of the peak seeding is not addressed. This is the sharpest form of the reader's weakest_assumption.","agreement_with_reader":"agree"},"referee_report":{"model":"selfhosted/deepseek-v4.1-flash","summary":"The manuscript revisits spontaneous symmetry breaking as a nonperturbative, dynamical process and introduces the notion of 'tachyonic preheating': when a negative mass-squared appears suddenly, modes with k < m grow exponentially, the field distribution rolls to the symmetry-broken minimum, and the transition completes within a single oscillation rather than through many perturbative oscillations. The authors support this with 3D lattice simulations (128^3 and 256^3) of a quadratic model (Eq. 1), a cubic model (Eq. 4), and the hybrid-inflation potential (Eq. 6), supplemented by analytic estimates of the growth time (Eq. 3) and the rms seed amplitude (Eq. 5). They then argue that preheating in hybrid inflation, including SUSY-motivated F-term and D-term models, is generically tachyonic, so the homogeneous oscillation stage ends after one oscillation, and they note possible relevance to disoriented chiral condensate formation.","tokens_in":11655,"tokens_out":4572,"duration_ms":92149,"significance":"If the central claim holds, it displaces the parametric-resonance picture of hybrid reheating with a fast spinodal mechanism and promotes domain structure and gradient energy to leading-order features of symmetry breaking rather than perturbative corrections. The paper's strengths are real and should be named: the analytic estimates in Eqs. (2), (3) and (5) are internally consistent and give a clear, physically transparent criterion for single-oscillation completion; the lattice treatment goes beyond Hartree/1-N methods in a regime where those methods are known to fail; the visual presentation (probability distributions plus the four submitted animations) makes the dynamics checkable by eye; and the extension to F-term and D-term hybrid models yields a concrete, in-principle falsifiable statement about the duration of the homogeneous oscillation stage. The identification of the cubic potential (Eq. 4) as the effective potential along the F-term trajectory is a genuinely useful structural observation.","major_comments":[{"comment":"The cubic model is the prototype for the F-term hybrid-inflation claim in Sec. III, and its dynamics is seeded differently from the quadratic case: because m^2(phi) = -2*lambda*v*phi + 3*lambda*phi^2 vanishes at phi = 0, there is no fixed tachyonic mass, and the instability must grow out of the initial fluctuation spectrum. The manuscript itself states that fluctuations 'a few times greater than (5)' grow faster and 'may therefore have greater impact,' and that 'low probability fluctuations with delta-phi >> delta-phi_rms correspond to peaks of the initial Gaussian distribution.' The abundance and amplitude of those peaks are controlled by the far UV tail of the initial spectrum, i.e. by the lattice cutoff and the choice of initial vacuum distribution, yet no lattice-spacing variation, box-size variation, or initial-spectrum variation is reported for this model. The renormalization subtraction (the induced lambda*v*phi*<phi^2> term) is acknowledged but its implementation is not shown. Because Sec. III identifies the cubic potential with the F-term trajectory (g^2 = 2*lambda), the 'typically tachyonic, single oscillation' conclusion for F-term models inherits this sensitivity. A short convergence check would settle the point: repeat the cubic-model runs at two lattice spacings (or two UV cutoffs of the initial spectrum) and report whether the time to reach <phi^2> ~ v^2 and the fraction of volume in bubbles are stable. If they are, the concern is closed and the F-term claim is on firm ground.","section":"Sec. II.B, Eq. (5) and Figs. 4-5","authors_acknowledge":null,"authors_acknowledge_p":null},{"comment":"The simulations switch on the tachyonic mass at t = 0 with vacuum initial conditions, i.e. an instantaneous quench. In hybrid inflation the transition is not instantaneous: phi rolls through phi_c = M/g with finite velocity, and the effective tachyonic mass m_sigma^2 = g^2*phi^2 - lambda*v^2 grows only as phi moves below phi_c. The manuscript acknowledges in the penultimate paragraph of Sec. III that 'there may exist some particular versions of hybrid inflation in which tachyonic preheating is inefficient, e.g. because of fast motion of the field phi near the bifurcation point,' but gives no quantitative criterion separating the fast-quench regime from the slow-roll-through regime. A criterion would be valuable and is within reach: for instance, the ratio of the rolling velocity |phi-dot|/phi_c to the growth rate m_sigma at the point where the instability sets in, or equivalently the number of e-folds of tachyonic growth available before the trajectory reaches the flat part near the minimum. Even an estimate of the parameter range over which the single-oscillation result survives, or a single lattice run with a nonzero initial phi-dot, would substantially strengthen the 'typically' in the abstract.","section":"Sec. III, Eq. (6)","authors_acknowledge":null,"authors_acknowledge_p":null},{"comment":"The central claim is quantitative ('within a single oscillation'), but the evidence presented in the letter is a set of probability histograms with no quoted lattice spacing, box size, initial-spectrum cutoff, or statistical uncertainty, and the technical details, including the renormalization procedure, are deferred to Ref. [12], which is listed as 'in preparation.' A letter may legitimately defer details, but the specific numbers that would let a reader judge convergence are exactly the ones deferred. At minimum the letter should state the lattice spacing and physical box size in units of m^{-1} (or v), confirm that the 128^3 and 256^3 runs agree on the time at which the distribution settles at |phi| ~ v, and state how the initial fluctuation spectrum is truncated. This is a modest addition and would make the headline number auditable rather than illustrative.","section":"Figs. 1, 3, 5, 6 and Ref. [12]","authors_acknowledge":null,"authors_acknowledge_p":null}],"minor_comments":[{"comment":"The mode-function normalization 1/sqrt(2k) exp(-ikt + i k.x) should be stated as a box or continuum convention, since the overall amplitude of <delta-phi^2> and hence the numerical coefficient inside the logarithm in t_* depends on it. The text notes that t_* depends only logarithmically on the initial distribution, which mitigates this, but the convention should still be explicit.","section":"Sec. II.A, Eq. (2)","authors_acknowledge":null,"authors_acknowledge_p":null},{"comment":"The statement that fluctuations grow 'until the amplitude of delta-phi becomes comparable to 2v/3' would read better with the one-line justification: m_eff^2 = -2*lambda*v*phi + 3*lambda*phi^2 vanishes at phi = 2v/3. As written the reader must reconstruct this.","section":"Sec. II.B","authors_acknowledge":null,"authors_acknowledge_p":null},{"comment":"The time labels (t = 0, 8, 10, 13, 14, 19 in Fig. 1; t = 0, 40, 44, 48, 50, 70 in Fig. 5) are not given units. Please state in each caption that time is in units of m^{-1} (or the appropriate scale for that model), and state the lattice size used for each figure.","section":"Figs. 1, 3, 5, 6 captions","authors_acknowledge":null,"authors_acknowledge_p":null},{"comment":"The caption refers to motion along the ellipse g^2*phi^2 + lambda*sigma^2 = g^2*phi_c^2, but sigma is complex in this model. Please clarify whether sigma in the ellipse denotes |sigma| and whether the plotted distribution is over phi and |sigma| or over phi and one component.","section":"Fig. 6","authors_acknowledge":null,"authors_acknowledge_p":null},{"comment":"Several reference entries need correction: Ref. [14] should read 'Bastero-Gil' rather than 'Bastelo-Gil'; Ref. [20] should read 'Van Proeyen' rather than 'Van Proyen'; and the first author affiliation gives 'Stanford, CA 9 4305' where '94305' is intended.","section":"References","authors_acknowledge":null,"authors_acknowledge_p":null},{"comment":"Typo: 'until its becomes O(v)' should read 'until it becomes O(v)'.","section":"Sec. II.A, paragraph 3","authors_acknowledge":null,"authors_acknowledge_p":null},{"comment":"The four animations (1.gif-4.gif) are cited as part of the submission and hosted at a Stanford URL. Please confirm whether they will be archived with the published version, since a URL alone is not a durable reference.","section":"Submission / supplementary files","authors_acknowledge":null,"authors_acknowledge_p":null}],"recommendation":"major_revision","confidential_remarks":"The qualitative picture is convincing and the quadratic-model result is robust, so my recommendation turns on the F-term/D-term application, which is half of the abstract's claim and rests on the cubic model. The authors already flag the renormalization issue and the possible inefficiency for fast phi motion, so the revision is a matter of showing the controls rather than rethinking the physics. The heavy reliance on Ref. [12] ('in preparation') for the very details a referee would want to check is worth noting to the editor; if that companion paper is not yet accepted, the authors should be encouraged to include the convergence statement in this letter rather than defer it."},"author_rebuttal":null,"desk_editor":{"model":"selfhosted/deepseek-v4.1-flash","letter":"Short version: this is the letter that named tachyonic preheating, and the headline claim — symmetry breaking in these models finishes in a single oscillation, so the homogeneous stage of hybrid reheating essentially doesn't happen — is well argued and, for a letter, adequately supported. It's a standard citation for good reason and it deserves to be one. What's actually new is the reframing. Symmetry breaking here is a nonperturbative, gradient-energy-dominated process, not a perturbative roll with small corrections. The paper makes the case that Hartree and 1/N treatments miss defect production and rescattering, and that the domain structure is leading order rather than a correction. The analytic estimates are clean and internally consistent: the growth time t* ~ (1/2m) ln(pi^2/lambda) in Eq. (3) is robust and only logarithmically sensitive to the initial spectrum, and the cubic-model seed estimate in Eq. (5) is sensible. The lattice movies aren't decoration; they show the percolation picture directly. The authors are also honest that their own 1998 paper already covered the g^2 >> lambda limit. Soft spots, in order of weight. First, the technical backing — lattice parameters, the renormalization subtraction for the cubic model, error analysis — all lives in ref. [12], listed as in preparation. That's the main reason this reads as a letter announcing a program rather than closing it. Second, the figures are qualitative histograms with no error bars and no lattice-spacing scan. On the stress-test point about the cubic model: it's the sharpest concern, but I'd temper it. The bubble seeding comes from the tail of the coarse-grained IR field, whose variance is set by modes below the tachyonic scale and by box volume, not by the UV lattice spacing; the tail probability is exponentially suppressed in (dphi/dphi_rms)^2 regardless of cutoff. So the qualitative picture should survive. What isn't shown is a lattice-spacing scan, and the renormalization subtraction is exactly where a cutoff dependence would enter. Worth asking for, not a refutation. Also, the F-term conclusion doesn't rest on the cubic toy alone — they ran F-term and D-term lattices directly. Who should read it: anyone working on reheating, hybrid inflation, or defect formation, plus the DCC crowd. Send it to referees; the companion paper is where the real scrutiny belongs.","headline":"The letter that named tachyonic preheating: the single-oscillation claim holds up qualitatively, but the technical backing sits in a companion paper still in preparation.","tokens_in":12231,"tokens_out":2937,"would_cite":true,"duration_ms":62042,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"pith_extraction":{"msc":[],"pacs":["98.80.Cq"],"model":"selfhosted/deepseek-v4.1-flash","headline":"Symmetry breaking may complete within a single oscillation of the field, not through a long slow roll toward the minimum.","keywords":["tachyonic preheating","spinodal instability","spontaneous symmetry breaking","hybrid inflation","lattice field simulation","cosmic strings","parametric resonance","disoriented chiral condensates"],"falsifier":"A lattice simulation of the same models with a different cutoff and an explicitly converged renormalization scheme, or with a gradual rather than instantaneous sign change of the effective mass, that shows the homogeneous field completing several oscillations before backreaction sets in would undercut the central claim; conversely, a hybrid-inflation parameter regime found to oscillate many times before the fluctuations dominate would delimit its generality.","tokens_in":25406,"feed_emoji":"🌌","tokens_out":4284,"duration_ms":83235,"temperature":0.7,"pith_summary":"The paper revisits the textbook story of spontaneous symmetry breaking — a field sitting at an unstable maximum with a negative mass squared, then rolling down to a new minimum — and argues that the process is far faster and more violent than perturbation theory suggests. Using three-dimensional lattice simulations of the classical field, the authors show that the exponentially growing long-wavelength modes reach the size of the symmetry-breaking scale $v$ within a time $t_*$ that depends only logarithmically on the coupling, and that the resulting gradient energy of the domain structure is comparable to the entire initial potential energy $V(0)$. The practical consequence is that the field distribution settles near $|v|$ after a single oscillation, so the usual picture of many slow oscillations about the minimum does not apply. The authors then apply this to hybrid inflation, where inflation ends at a bifurcation point and the triggering field is unstable: they argue that preheating there is typically tachyonic rather than parametric-resonant, so the stage of coherent oscillation of the inflaton fields ends after one oscillation. If that is right, the standard treatment of reheating in hybrid models — resonant particle production over many oscillations — is describing the wrong mechanism.","feed_headline":"Symmetry breaking can finish in one field oscillation","feed_subtitle":"Lattice simulations suggest hybrid-inflation reheating is a sudden spinodal collapse, not a slow resonance.","key_machinery":"The load-bearing mechanism is the spinodal (tachyonic) instability combined with a classical-statistical lattice treatment of the field. Because the growing modes acquire occupation numbers $n_k ∼ π^2/λ ≫ 1$, the quantum state is legitimately replaced by an ensemble of classical waves obeying the relativistic wave equation on a $128^3$ or $256^3$ grid, evolved with the LATTICEEASY code. The argument turns on two estimates that make the effect leading-order rather than a correction: the growth time $t_* ∼ (1/2m)ln(π^2/λ)$, which is only logarithmically sensitive to the initial fluctuation spectrum, and the gradient energy of the resulting domain structure, $∼ k^2δφ^2 = O(m^4/λ)$, which matches the initial potential energy and therefore blocks any coherent return to the symmetric state. In the cubic model the same role is played by the field-dependent mass $m^2_{eff}(φ) = -2λvφ + 3λφ^2$, which vanishes at $φ = 0$ and makes the onset of instability depend on rare peaks of the initial distribution, giving the bubble-like growth seen in the simulations.","core_discovery":"The central claim is that spontaneous symmetry breaking driven by a suddenly imposed tachyonic mass is a one-oscillation event. For the quadratic model $V(φ) = -m^2φ^2/2 + λφ^4/4$, the authors estimate that all modes with $k < m$ grow until $t_* ≈ (1/2m)ln(π^2/λ)$, at which point their occupation numbers are $n_k ∼ π^2/λ ≫ 1$ and the field has broken into domains of size only logarithmically larger than $m^{-1}$. Because the domain walls carry gradient energy of order $m^4/λ$ — the same order as the initial false-vacuum energy $V(0) = m^4/4λ$ — the field cannot return coherently to $φ = 0$, and the simulations show the probability distribution $P(φ,t)$ collapsing onto the two minima after the first oscillation. For the cubic model $V = -λvφ^3/3 + λφ^4/4 + λv^4/12$, the growth is more like bubble formation seeded by rare high peaks of the initial Gaussian fluctuations, but the outcome is the same: relaxation to the minimum within one oscillation. Applied to hybrid inflation, $V(φ,σ) = λ(σ^2-v^2)^2/4 + g^2φ^2σ^2/2 + m^2φ^2/2$, the authors report lattice runs for the non-supersymmetric model and for SUSY-motivated F-term and D-term variants showing that the homogeneous fields are destroyed by backreaction of the tachyonic fluctuations after a single oscillation, so preheating in hybrid inflation is generically tachyonic.","pith_inferences":["If the mechanism is as generic as claimed, the relevant dividing line between 'tachyonic' and 'resonant' reheating is the ratio of the quench rate to the tachyonic growth rate $m$; a smooth, slow change of sign of the effective mass should interpolate between the two pictures, which is a testable statement about intermediate regimes.","The logarithmic dependence of $t_*$ on $λ$ suggests the one-oscillation outcome is robust to the details of the initial fluctuation spectrum, but it also implies the outcome is sensitive to the initial dispersion in a way that a different vacuum choice could shift.","Since gradient energy at the level of $V(0)$ is produced, the transition is a candidate source of a stochastic gravitational-wave background from the resulting field inhomogeneities, an observable the paper does not discuss."],"forward_implications":["Preheating in hybrid inflation should be modelled as spinodal decay of the homogeneous fields rather than as broad parametric resonance, which changes the predicted spectrum and occupation numbers of produced particles.","The stage of coherent oscillations of the inflaton fields in hybrid models ends after a single oscillation, so scenarios that rely on many oscillations to build up resonant occupation numbers need re-examination.","Because a large fraction of $V(0)$ goes into gradient energy and domain structure, topological defects such as cosmic strings (for a complex $σ$) are produced as a leading-order outcome of the transition, not as a rare by-product.","Gravitino production in hybrid inflation should be computed from scattering of the classical waves generated by tachyonic preheating, not from decays of a coherently oscillating condensate.","The same one-oscillation picture is proposed for the formation of disoriented chiral condensates in heavy-ion collisions, where a rapidly cooled plasma suddenly finds itself on an unstable side of its effective potential."],"supporting_citations":[{"why":"Supplies the lattice method for simulating classical scalar waves with large occupation numbers, which is the technique the whole argument rests on.","marker":"[10]"},{"why":"The LATTICEEASY code used to run the 3D simulations reported in the figures.","marker":"[11]"},{"why":"Listed as in preparation and cited for the detailed description of the calculations, lattice parameters, and renormalization procedure deferred from this letter.","marker":"[12]"},{"why":"Defines the hybrid inflation potential and bifurcation-point structure to which the tachyonic preheating analysis is applied.","marker":"[4]"},{"why":"Earlier study of hybrid reheating that noted the single-oscillation behaviour in the $g^2 ≫ λ$ limit, which this paper generalises.","marker":"[5]"},{"why":"Source of the SUSY F-term inflationary potential whose flat-direction effective potential matches the cubic model analysed here.","marker":"[13]"},{"why":"Prior analysis of preheating in the F-term model, providing the parametric-resonance picture that this paper argues is superseded by the tachyonic channel.","marker":"[14]"},{"why":"Cited for the contrasting quartic case $V = V(0) - λφ^4/4$, where the symmetric phase decays by tunnelling and bubble formation, used to place the cubic model between the quadratic and quartic extremes.","marker":"[16]"}],"fun_headline_variants":["Symmetry breaking completes in one oscillation on the lattice","Tachyonic preheating ends hybrid inflation in one oscillation","Spinodal instability breaks symmetry in a single oscillation","Domain walls block a coherent return to the false vacuum","Hybrid inflation reheating is tachyonic, not resonant"],"cache_read_input_tokens":0,"weakest_assumption_plain":"The result rests on treating the quantum field as an ensemble of classical waves on a lattice with a sharp switch-on of the negative mass squared at $t = 0$, and on the one-oscillation outcome being independent of the lattice cutoff and of the renormalization subtraction — a procedure the paper states is needed for the cubic model but defers to a longer paper still in preparation.","fun_headline_variants_meta":{"raw":{"variants":["Symmetry breaking completes in one oscillation on the lattice","Tachyonic preheating ends hybrid inflation in one oscillation","Spinodal instability breaks symmetry in a single oscillation","Domain walls block a coherent return to the false vacuum","Hybrid inflation reheating is tachyonic, not resonant"]},"model":"selfhosted/deepseek-v4.1-flash","effort":"low","attempts":1,"cost_usd":0.0,"raw_usage":{"total_tokens":1606,"prompt_tokens":1034,"completion_tokens":572,"prompt_tokens_details":null,"completion_tokens_details":{"reasoning_tokens":507}},"tokens_in":1034,"tokens_out":572,"duration_ms":10111,"image_pages":[],"temperature":1.0,"reasoning_budget":50,"reasoning_tokens":507,"transport_provider":"selfhosted","cache_read_input_tokens":0,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-10-29T17:59:42.777612+00:00","model_set":{"reader":"selfhosted/deepseek-v4.1-flash"},"falsifier":"A lattice simulation of the same models with a different cutoff and an explicitly converged renormalization scheme, or with a gradual rather than instantaneous sign change of the effective mass, that shows the homogeneous field completing several oscillations before backreaction sets in would undercut the central claim; conversely, a hybrid-inflation parameter regime found to oscillate many times before the fluctuations dominate would delimit its generality.","supporting_citations":[{"cited_title":"Phys. Rev. Lett. 77, 219 (1996); Phys. Rev. Lett. 79, 1607 (1997); Phys. Rev. D56, 653","cited_arxiv_id":null,"evidence_quote":"Supplies the lattice method for simulating classical scalar waves with large occupation numbers, which is the technique the whole argument rests on."},{"cited_title":"LATTICEEASY: A Program for Lattice Simulations of Scalar Fields in an Expanding Universe","cited_arxiv_id":"hep-ph/0011159","evidence_quote":"The LATTICEEASY code used to run the 3D simulations reported in the figures."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Listed as in preparation and cited for the detailed description of the calculations, lattice parameters, and renormalization procedure deferred from this letter."},{"cited_title":"Phys. Lett. B259, 38 (1991); Phys. Rev. D49, 748","cited_arxiv_id":null,"evidence_quote":"Defines the hybrid inflation potential and bifurcation-point structure to which the tachyonic preheating analysis is applied."},{"cited_title":"Garc ´ ıa-Bellido and A","cited_arxiv_id":null,"evidence_quote":"Earlier study of hybrid reheating that noted the single-oscillation behaviour in the $g^2 ≫ λ$ limit, which this paper generalises."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Source of the SUSY F-term inflationary potential whose flat-direction effective potential matches the cubic model analysed here."},{"cited_title":"Phys. Rev. D60, 103517","cited_arxiv_id":null,"evidence_quote":"Prior analysis of preheating in the F-term model, providing the parametric-resonance picture that this paper argues is superseded by the tachyonic channel."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Cited for the contrasting quartic case $V = V(0) - λφ^4/4$, where the symmetric phase decays by tunnelling and bubble formation, used to place the cubic model between the quadratic and quartic extremes."}],"review_version":1}