{"id":"65bf808d-8c01-44a4-b5f6-69a8b6b63dbc","arxiv_id":"hep-th/9504090","paper_version":1,"verdict":"CONDITIONAL","confidence":"HIGH","novelty_score":8.0,"correctness_risk":"low","formal_verification":"none","parameter_count":1,"one_line_summary":"Conifold singularities of Calabi-Yau moduli space are resolved in type II string theory by extremal Ramond-Ramond black holes that become massless at the singularity, whose one-loop contribution reproduces the logarithmic metric divergence.","lead":"Strominger argues that the points in a Calabi-Yau string compactification's moduli space where a cycle shrinks to zero size, and where the low-energy description breaks down, are repaired by a Ramond-Ramond black hole that becomes massless there. The result ties black hole physics to moduli-space geometry and explains how a quantum effect can survive at vanishing string coupling.","discovery_kind":"unification","skeptic_critique":{"model":"selfhosted/deepseek-v4.1-flash","headline":"The load-bearing concern is the §4.2 degeneracy postulate: the paper's only decisive support for 'no multiply-charged single-particle states' is the consistency of the resolution it enables, and the cited counterexamples ([11], [27,28]) are acknowledged but not rebutted.","rationale":"The reader's weakest_assumption is correctly located: §4.2's degeneracy postulate is the single point on which the resolution rests. I agree with the verdict and would not change it. Two refinements. First, the paper is more candid than the classifier record suggests: the record reports 'limitations: none found,' but §4.2 says 'the best we can do at present is give some plausibility arguments' and 'We do not know how to derive our assumption from first principles,' footnote 2 concedes multiple-degeneration conifolds are untreated, and the appendix calls its superpotential remarks 'merely suggestive.' These are acknowledged limitations, not new findings. Second, the reader's framing slightly undersells the paper's own resources: the geometric monodromy (2.3) is a theorem for a simple conifold and fixes the one-loop coefficient to 1, so the integer constraint Σ q_i^2 = 1 forces the light spectrum to be one charge-±1 multiplet. The 'assumption' is therefore a derivation in the simple case, and the residual concern is the simple-conifold restriction (footnote 2) and the unreconciled tension with [11] and [27,28], which the paper cites but does not rebut. The coefficient match (4.1) vs (2.4) is a consistency condition, not independent evidence, exactly as the reader says. With hindsight the postulate is correct (the conifold carries one hypermultiplet; later GV invariants), but that is outside the 1995 argument. Net: the concern is real, acknowledged, and does not move the verdict from CONDITIONAL.","tokens_in":11985,"tokens_out":15573,"duration_ms":294369,"concrete_test":"Compute the moduli space of two wrapped D3-branes on the shrinking S^3 at Z1 = 0 (the paper's own suggested check in §4.2) and determine whether the coincident n1 = 2 configuration is a single bound state or a threshold two-particle state. In parallel, verify that the one-loop monodromy from a general light spectrum, F1 → F1 + (Σ q_i^2) Z1, matches the geometric monodromy (2.3) only for a single |q| = 1 multiplet. If the n1 = 2 sector contains a bound state, the coefficient in (4.1) becomes 1 + 4 = 5 and the resolution fails; if it is a threshold state, the §4.2 assumption is confirmed and the concern reduces to footnote 2's multiple-degeneration caveat.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim is conditional on the §4.2 postulate that exactly one hypermultiplet (charge ±1) becomes light at the conifold and that n1 = 2, 3, ... configurations are multi-particle. The paper's stated support is analogy ('t Hooft-Polyakov monopoles; Dabholkar-Harvey string states) plus the sentence 'consistency of the low energy theory seems to require that there are no multiply-charged single-particle states.' That last step is circular as written: the coefficient match between (4.1) and (2.4) is a consequence of assuming only q = ±1 states contribute to the one-loop beta function, so it cannot independently confirm the assumption. The paper concedes this ('We do not know how to derive our assumption from first principles (except from the requirement of a consistent low-energy theory)') and cites [11] (Witten's infinite tower of multiply-charged states in the IIA/11d correspondence) and [27,28] (Dabholkar-Harvey charge-two 'bound state at threshold') as pointing the other way, without rebutting either. What the paper does not exploit is that the Picard-Lefschetz monodromy (2.3), F1 → F1 + Z1, is a geometric theorem for a simple conifold and independently fixes the one-loop coefficient to 1; since a charge-q hypermultiplet contributes q^2, the integer constraint Σ q_i^2 = 1 forces the light spectrum to be exactly one charge-±1 multiplet. So the 'assumption' is actually a derivation for simple degenerations, and the residual concern is narrower: (i) footnote 2 concedes that multiple-degeneration conifolds are untreated, and the abstract's 'generically' is not quantified against that restriction; (ii) the tension with [11] is left open; (iii) the IIA case inherits the argument only via mirror symmetry (§3.2), which is asserted rather than checked at the level of the monodromy.","agreement_with_reader":"agree"},"referee_report":{"model":"selfhosted/deepseek-v4.1-flash","summary":"The paper proposes that conifold singularities in the vector-multiplet moduli space of type II Calabi-Yau compactifications are resolved by massless Ramond-Ramond black holes. The argument has three parts. First, wrapped extremal branes (threebranes in IIB, twobranes in IIA) have a BPS mass M = g5 e^{K/2}|m_I F^I - n_I Z^I| (Eq. 3.7), so the state with n1 = 1 becomes massless at the conifold Z1 = 0 (Eq. 3.8). Second, including this state as a light charged hypermultiplet gives a smooth Wilsonian action whose one-loop beta function reproduces the special-geometry monodromy F1 ~ const + (1/2 pi i) Z1 ln Z1 (Eq. 4.1 versus Eq. 2.4). Third, the effect is argued to be gs-independent because N=2 forbids neutral vector-hypermultiplet couplings after the e^{-2 phi} rescaling of the metric (Section 4.1). The resolution requires exactly one light hypermultiplet, which the paper states as an assumption supported by analogy with 't Hooft-Polyakov monopoles and Dabholkar-Harvey string states (Section 4.2). An appendix argues that a generic N=1 superpotential can have a minimum at a conifold.","tokens_in":12341,"tokens_out":6442,"duration_ms":114714,"significance":"If the picture holds, this is a landmark result: it identifies a generic classical inconsistency of Calabi-Yau string compactifications and shows how a nonperturbative, gs-independent effect cures it. The paper's strengths are real and should be named. The coefficient match between the one-loop beta function and the Picard-Lefschetz monodromy is a parameter-free quantitative check, not a fit. The prediction that a specific BPS state becomes massless at a specific geometric locus is falsifiable within the moduli-space geometry. The gs-scaling argument is explicit and tied to the structure of RR couplings, and the analogy with Seiberg-Witten N=2 Yang-Mills is drawn carefully rather than asserted. The paper also anticipates, correctly, that extremal black holes must be treated as elementary quanta in the quantum theory. The main caveat is that the central claim is conditional on the Section 4.2 degeneracy postulate, which the authors state openly in the abstract ('given a plausible assumption') and in Section 4.2 ('We do not know how to derive our assumption from first principles'). That acknowledgement should be weighed in favor of the manuscript, not against it.","major_comments":[{"comment":"The coefficient match between the one-loop result (4.1) and the monodromy (2.4) is presented as support for the single-hypermultiplet assumption, but as written the match follows from that assumption: the beta-function coefficient is fixed by the assumed light spectrum. The paper can remove the appearance of circularity by noting that the Picard-Lefschetz monodromy (2.3), F1 -> F1 + Z1, is a geometric theorem for a simple conifold and independently fixes the one-loop coefficient to 1. Since a charge-q hypermultiplet contributes q^2, the integer constraint sum_i q_i^2 = 1 forces the light spectrum to be exactly one charge-(+/-1) multiplet for a simple degeneration. This converts the Section 4.2 assumption into a derivation in the case actually treated and strengthens the paper's central claim; it should be stated explicitly.","section":"Section 4.2, Eqs. (2.4) and (4.1)","authors_acknowledge":null,"authors_acknowledge_p":null},{"comment":"The cited counterexamples are acknowledged but not rebutted. Reference [11] (Witten's IIA/11d correspondence) is said to 'seem to require an infinite tower of distinct multiply-charged states', and references [27,28] (Dabholkar-Harvey) are said to contain a charge-two 'bound state at threshold'. The paper should state why these do not apply here: for instance, that the Dabholkar-Harvey charge-two object is a multi-particle state rather than a distinct elementary quantum, and that the IIA/11d tower concerns a different regime or a different notion of state. As written, the reader is left with the impression that the load-bearing assumption is in tension with published results, which weakens the paper more than the assumption itself does.","section":"Section 4.2","authors_acknowledge":null,"authors_acknowledge_p":null},{"comment":"The BPS mass formula (3.8) carries an explicit factor g5, yet Section 4.1 argues that the black hole physics is gs-independent. The two statements are consistent only if the g5 in (3.8) is a string-frame artifact absorbed by the e^{-2 phi} rescaling to the Einstein frame of (4.3), but the paper does not say this. It should state explicitly that the physical mass and the loop effects are gs-independent in the rescaled variables, so that the g5 in (3.8) and the gs-independence claimed in Section 4.1 do not appear to contradict each other. This is load-bearing because the gs-independence of the resolution is one of the paper's three headline conclusions.","section":"Section 3.1, Eq. (3.8) and Section 4.1","authors_acknowledge":null,"authors_acknowledge_p":null}],"minor_comments":[{"comment":"There are numerous typographical errors that should be corrected: the title reads 'STRIN G THEOR Y'; 'rescaping' should be 'rescaling'; 'Accordnig' should be 'According'; 'intergrating' should be 'integrating'; 'Quantiztion' should be 'Quantization'; 'Bogolmonyi' should be 'Bogomol'nyi'; 'mondromy' should be 'monodromy'; 'willl' should be 'will'; 'D abohlkar' should be 'Dabholkar'; 'a a sigma model' should be 'a sigma model'; 'the a rea' should be 'the area'; and 'proportional the a rea' should be 'proportional to the area'.","section":"Title and throughout","authors_acknowledge":null,"authors_acknowledge_p":null},{"comment":"The range of the index is typeset as 'I, J = 1, .. 12 b3', which appears to be a corruption of 'I, J = 1, ..., (1/2) b3'. This should be corrected so the dimension of the basis is unambiguous.","section":"Eq. (2.1)","authors_acknowledge":null,"authors_acknowledge_p":null},{"comment":"The BPS mass relation is written as 'Mass = Q x (3 - volume)', which is garbled. The intended statement (mass proportional to the brane volume times the charge) should be written out explicitly, since the heuristic identification of mass with wrapped-brane area is used immediately below to argue that the mass vanishes at the conifold.","section":"Eq. (3.5)","authors_acknowledge":null,"authors_acknowledge_p":null},{"comment":"References [3] and [4] are given identical journal citations, 'Nucl. Phys. B42 (1990) 246'. One of these is presumably wrong, and 'B42' is not a plausible Nucl. Phys. B volume number. Reference [5] (Aspinwall, Greene, Morrison) is missing a journal name and gives only 'B416 (1994) 414'. These should be checked against the published versions.","section":"References [3] and [4]","authors_acknowledge":null,"authors_acknowledge_p":null},{"comment":"The statement that a Calabi-Yau space 'has b3 topologically non-trivial three-surfaces' is loose; b3 is the third Betti number, and the number of independent three-cycles is b3 (or b3/2 in a symplectic basis). A precise phrasing would help readers who are not special-geometry specialists.","section":"Section 2","authors_acknowledge":null,"authors_acknowledge_p":null},{"comment":"The IIA case is dismissed with the remark that 'the analysis of the black hole solutions reduces to that considered above' by mirror symmetry. Since the paper itself emphasizes that the IIA physical picture is different (instanton-corrected areas), a sentence spelling out what is and is not carried over by mirror symmetry would be useful.","section":"Section 3.2","authors_acknowledge":null,"authors_acknowledge_p":null},{"comment":"The appendix concludes that V ~ V0 + w1^2 / ln(Z1 Zbar1) 'has a sharp minimum at Z1 = 0'. Since 1/ln(Z1 Zbar1) tends to zero as Z1 tends to zero, the correction vanishes at the conifold and the potential approaches V0; the 'sharpness' comes from the diverging gradient, not from a diverging value. A clarifying sentence would prevent misreading. The appendix's own caveat that perturbative corrections could alter the geometry is appropriate and should be retained.","section":"Appendix A","authors_acknowledge":null,"authors_acknowledge_p":null},{"comment":"The paper does acknowledge its central limitation: the abstract says 'given a plausible assumption on the degeneracies of black hole states', and Section 4.2 says 'We do not know how to derive our assumption from first principles'. This is the right way to present a conditional result, and it should be preserved in revision rather than softened.","section":"Abstract and Section 4.2","authors_acknowledge":null,"authors_acknowledge_p":null}],"recommendation":"minor_revision","confidential_remarks":"This is a landmark paper whose central physical picture has held up and whose Seiberg-Witten analogy was prescient. The revision I am requesting is about argument structure, not about the result: the Section 4.2 assumption can be recast as a derivation using the monodromy constraint, and the g5 versus gs-independence point in Section 4.1 needs one clarifying sentence. Neither is a reason to delay publication beyond a normal revision cycle. The citation pattern is appropriate for the field and the paper's priority claims are properly attributed to the Seiberg-Witten and Harvey-Strominger work it builds on."},"author_rebuttal":null,"desk_editor":{"model":"selfhosted/deepseek-v4.1-flash","letter":"Short version: this is the paper that established massless Ramond-Ramond black holes as the resolution of conifold singularities, and the central argument holds up better than the abstract's hedge about a 'plausible assumption' suggests. Worth your time, with one presentational caveat and a few scope gaps worth knowing about before you cite it.","headline":"The core argument is sounder than its own abstract admits — the §4.2 'assumption' is really a derivation from the Picard-Lefschetz monodromy, and the paper undersells it.","tokens_in":13108,"tokens_out":1079,"would_cite":true,"duration_ms":19443,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"pith_extraction":{"msc":[],"pacs":[],"model":"selfhosted/deepseek-v4.1-flash","headline":"The conifold singularities that make Calabi-Yau string compactifications ill-defined at low energies may be cured by Ramond-Ramond black holes that become massless exactly where the geometry degenerates.","keywords":["conifold singularities","Calabi-Yau compactification","type II string theory","extremal black holes","BPS states","special geometry","Ramond-Ramond fields","moduli space"],"falsifier":"A demonstration that a charge-two wrapped-brane state at the conifold is a genuine single-particle BPS state, which would add a second contribution to the one-loop beta function and spoil the match between the coefficient in $F_1(Z_1) = F_1(0) + (1/2i)Z_1$ up to a factor and the classical monodromy, would undercut the paper's main quantitative check.","tokens_in":13221,"feed_emoji":"🕳️","tokens_out":2830,"duration_ms":43665,"temperature":0.7,"pith_summary":"Compactify a type II string on a Calabi-Yau space and the space of shapes (the moduli space) typically contains conifold points, where a two- or three-cycle shrinks to zero size, the moduli-space metric diverges, and the low-energy $\\sigma$ model breaks down. Strominger argues that this breakdown is not a failure of string theory but a failure of the effective field theory that kept only the moduli: a brane wrapped on the shrinking cycle is an extremal Ramond-Ramond black hole whose mass is proportional to the cycle's area, so it goes massless at the conifold and must be kept in the Wilsonian action. Keeping it makes the action smooth there, and integrating it back out reproduces the logarithmic singularity of special geometry, $F_1(Z_1) = F_1(0) + (1/2i)Z_1$ up to a factor, matching the classical monodromy. The reason such a quantum effect can repair a classical inconsistency is that Ramond-Ramond couplings do not carry a uniform factor of $e^{-2}$ of the dilaton, and $N=2$ supersymmetry forbids neutral couplings between vector multiplets and hypermultiplets, so the light black hole loops carry no powers of the string coupling $g_s$. If the picture holds, string theory is telling us that extremal black holes must be treated as elementary quanta in virtual loops.","feed_headline":"Massless black holes may patch a string theory breakdown","feed_subtitle":"Conifold points make the low-energy theory singular; wrapped Ramond-Ramond branes going massless there may repair it.","key_machinery":"The load-bearing object is the wrapped extremal brane: a type IIB threebrane (or, by mirror symmetry, a type IIA twobrane) wrapped on a cycle of the Calabi-Yau, which appears in four dimensions as a charged black hole whose BPS mass is $M = g_5 e^{K/2}|Z_1|$ and which therefore becomes massless at the conifold. The second piece of machinery is special geometry: the periods $Z_I$ and $F_I$ of the holomorphic three-form, the monodromy $F_1 = F_1 + Z_1$ around the conifold, and the resulting $F_1(Z_1) = F_1(0) + (1/2i)Z_1$ behaviour up to a factor, which fixes the metric through $G_{I} = _I _J K$ with $K = - (i F^I _I - i Z^I F_I)$. The two are joined by a one-loop computation in the light-hypermultiplet effective theory: the $\\beta$ function of the gauge field in the same multiplet as $Z_1$ reproduces the coefficient of the logarithm, so the classical singularity is exactly what integrating out a massless charged state produces.","core_discovery":"The paper's central claim is that for type II string theories the conifold singularities of the Calabi-Yau moduli space are resolved by a single light Ramond-Ramond black hole hypermultiplet rather than by new classical geometry. The BPS mass formula $M = g_5 e^{K/2}|m_I F^I - n^I Z_I|$ gives, for the state with $n^1 = 1$ and all other charges zero, $M = g_5 e^{K/2}|Z_1|$, which vanishes at the conifold $Z_1 = 0$ because the wrapped brane's mass is set by the area of the shrinking cycle. The Wilsonian effective action that includes this charged hypermultiplet is smooth near $Z_1 = 0$; the singularity appears only after the state is integrated out, where its one-loop contribution to the gauge coupling reproduces the special-geometry behaviour $F_1(Z_1) = F_1(0) + (1/2i)Z_1$ up to a factor, matching the classical monodromy. Strominger further argues that the effect survives as $g_s$ goes to zero, because the relevant couplings of Ramond-Ramond gauge fields to the dilaton do not come with a uniform $1/g_s^2$ and because $N=2$ supersymmetry forbids neutral vector-hypermultiplet couplings, so the black hole loops are $g_s$-independent.","pith_inferences":["If the resolution is right, the number of light states at a conifold is a consistency condition on the theory rather than an accident, which suggests that conifold points could be used to count BPS states of a Calabi-Yau without doing the hard semiclassical black hole counting directly.","The argument that $N=2$ forbids neutral vector-hypermultiplet couplings does double duty: it explains the $g_s$ independence here and it underwrites a nonrenormalization statement for the moduli metric, so a failure of the resolution would also threaten that nonrenormalization claim.","The charge-two question is the sharpest handle: a single-particle charge-two state at the conifold would shift the one-loop coefficient and break the match with the monodromy, so any independent count of wrapped-brane bound states at threshold is a direct test of the picture.","The paper leaves heterotic conifolds unexplained, which suggests the mechanism is tied to Ramond-Ramond charges specifically; whether an analogous light state exists for heterotic compactifications is the natural next question."],"forward_implications":["Low-energy effective field theories of Calabi-Yau compactifications are incomplete without nonperturbative states, so the classical moduli space is not the right setting for cosmology that rolls through a conifold.","Extremal black holes must be counted as elementary quanta contributing to virtual loops, fixed by their mass and charge rather than by any independent coupling.","String theory admits quantum effects that are not suppressed as $g_s$ tends to zero, because Ramond-Ramond field strengths enter the classical action without a uniform dilaton factor.","Conifold and orbifold singularities are resolved by the same mechanism in spirit: light solitons supply the states the low-energy theory is missing, with spin one-half giving conifolds and spin one giving orbifolds.","For type IIA the resolution is tied to worldsheet instantons, suggesting that in a dual formulation where fundamental strings are solitons the same singularity would look like an ordinary spacetime instanton effect."],"supporting_citations":[{"why":"Supplies the conifold example, the monodromy around it, and the divergence of the moduli-space metric that the paper sets out to resolve.","marker":"[1]"},{"why":"The Seiberg-Witten monopole story that the paper's resolution is modelled on: light charged states, a smooth Wilsonian action, and a one-loop logarithm reproducing the singularity.","marker":"[8]"},{"why":"Provides the extremal p-brane black hole solutions that, when wrapped on Calabi-Yau cycles, become the four-dimensional black holes that go massless.","marker":"[7]"},{"why":"Supplies the Kahler- and $Sp(b_3;)$-invariant BPS mass formula used to show the wrapped brane mass vanishes at the conifold.","marker":"[12]"},{"why":"Source for the special geometry of $N=2$ vector multiplets, including the prohibition on neutral vector-hypermultiplet couplings used in the $g_s$ argument.","marker":"[16]"},{"why":"Source for the period formulation of special geometry and the $Sp(b_3;)$ bundle structure over the moduli space.","marker":"[17]"},{"why":"Related observations on Ramond-Ramond fields and the failure of the usual $g_s$ suppression, which the paper's coupling argument builds on.","marker":"[11]"},{"why":"The K3 soliton construction whose degeneracies are used as evidence that multiply charged solutions are multi-particle rather than distinct states.","marker":"[14]"},{"why":"Cited as evidence for a picture of conifold behaviour different from the one advocated here, the main competing resolution on the table.","marker":"[2]"}],"fun_headline_variants":["Massless black holes smooth conifold singularities","Light Ramond-Ramond black holes repair the conifold","Conifold cure hinges on assumed black-hole degeneracies","Black-hole loops survive as string coupling vanishes","Type II conifold singularities resolved by massless states"],"cache_read_input_tokens":0,"weakest_assumption_plain":"The argument needs exactly one supermultiplet to become massless at the conifold, and it needs states of charge two and higher to be collections of charge-one particles rather than separate single-particle states; the paper states this as an assumption and offers only plausibility arguments for it.","fun_headline_variants_meta":{"raw":{"variants":["Massless black holes smooth conifold singularities","Light Ramond-Ramond black holes repair the conifold","Conifold cure hinges on assumed black-hole degeneracies","Black-hole loops survive as string coupling vanishes","Type II conifold singularities resolved by massless states"]},"model":"selfhosted/deepseek-v4.1-flash","effort":"low","attempts":1,"cost_usd":0.0,"raw_usage":{"total_tokens":2448,"prompt_tokens":1004,"completion_tokens":1444,"prompt_tokens_details":null,"completion_tokens_details":{"reasoning_tokens":1377}},"tokens_in":1004,"tokens_out":1444,"duration_ms":21975,"image_pages":[],"temperature":1.0,"reasoning_budget":50,"reasoning_tokens":1377,"transport_provider":"selfhosted","cache_read_input_tokens":0,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-10-29T19:32:25.724996+00:00","model_set":{"reader":"selfhosted/deepseek-v4.1-flash"},"falsifier":"A demonstration that a charge-two wrapped-brane state at the conifold is a genuine single-particle BPS state, which would add a second contribution to the one-loop beta function and spoil the match between the coefficient in $F_1(Z_1) = F_1(0) + (1/2i)Z_1$ up to a factor and the classical monodromy, would undercut the paper's main quantitative check.","supporting_citations":[{"cited_title":"A Pair o f Calabi-Yau Manifolds as an Exactly Soluble Superconformal Field Theory","cited_arxiv_id":null,"evidence_quote":"Supplies the conifold example, the monodromy around it, and the divergence of the moduli-space metric that the paper sets out to resolve."},{"cited_title":"Electromagnetic Duality, Mon opole Condensation and Confinement in N=2 Supersymmetric Yang-Mills Theory","cited_arxiv_id":null,"evidence_quote":"The Seiberg-Witten monopole story that the paper's resolution is modelled on: light charged states, a smooth Wilsonian action, and a one-loop logarithm reproducing the singularity."},{"cited_title":"Black Strings and p-branes,","cited_arxiv_id":null,"evidence_quote":"Provides the extremal p-brane black hole solutions that, when wrapped on Calabi-Yau cycles, become the four-dimensional black holes that go massless."},{"cited_title":"Duality Transformations in Supersymmetric Yang-Mills Theories coupled to Supergravity,","cited_arxiv_id":"hep-th/9502072","evidence_quote":"Supplies the Kahler- and $Sp(b_3;)$-invariant BPS mass formula used to show the wrapped brane mass vanishes at the conifold."},{"cited_title":"Lagrangians of N =2 Supergravity-Matter Systems","cited_arxiv_id":null,"evidence_quote":"Source for the special geometry of $N=2$ vector multiplets, including the prohibition on neutral vector-hypermultiplet couplings used in the $g_s$ argument."},{"cited_title":"Special Geometry","cited_arxiv_id":null,"evidence_quote":"Source for the period formulation of special geometry and the $Sp(b_3;)$ bundle structure over the moduli space."},{"cited_title":"String Theory Dynamics In Various Dimensions","cited_arxiv_id":"hep-th/9503124","evidence_quote":"Related observations on Ramond-Ramond fields and the failure of the usual $g_s$ suppression, which the paper's coupling argument builds on."},{"cited_title":"The Heterotic String is a Soliton","cited_arxiv_id":"hep-th/9504047","evidence_quote":"The K3 soliton construction whose degeneracies are used as evidence that multiply charged solutions are multi-particle rather than distinct states."},{"cited_title":"Rolling Among Calab i-Yau Vacua","cited_arxiv_id":null,"evidence_quote":"Cited as evidence for a picture of conifold behaviour different from the one advocated here, the main competing resolution on the table."}],"review_version":1}