UA-J-01: Three Accounts — The Origins of the 2014 Donbas War: the Post-Maidan Southeast, the Crimea Operation, the "Russian Spring", the Girkin Crossing and the Sloviansk Seizure, the Anti-Terrorist Operation and the Summer Escalation, Ilovaisk and the Regular-Forces Intervention, and the Contest Among the Russian-Aggression Account, the Civil-Conflict/Local-Agency Account, and the Western-Provocation/Realist Account (2014–2026)

Status: DRAFTWords: 11,920

1. Key Takeaways

The origins of the 2014 Donbas war are among the most consequential contested questions in contemporary European history, because every later justification narrative was built on an answer to them. Whether the war that began in April 2014 was a Russian state operation, a civil conflict that Russia exploited, or a predictable great-power response to Western expansion determined — in Moscow, Kyiv, and Western capitals — what Minsk-1 and Minsk-2 were understood to be settling, whether the 2015–2021 "frozen conflict" was an internal-Ukrainian or an interstate matter, and how the 24 February 2022 full-scale invasion was narrated by its author ("completing" the defence of Donbas) and by its victim (the continuation of an eight-year-old Russian war). This document applies the corpus's three-account discipline: each account is presented at full evidentiary and logical strength, the scholarly weight is documented honestly, and no winner is declared — while the genuine evidentiary asymmetries among the accounts are stated rather than smoothed over.
The factual spine is itself largely settled; the contest is over causation, weighting, and characterisation. The sequence — Yanukovych's 21–22 February 2014 flight (UA-B-03); the 27 February seizure of the Crimean parliament by unmarked Russian troops and the 16 March "referendum" (UA-C-01); the March–April anti-Maidan / "Russian Spring" mobilisations in Kharkiv, Donetsk, Luhansk, and Odesa; the 6–7 April administrative-building seizures and the DNR proclamation; the 12 April seizure of Sloviansk by the armed group commanded by the Russian FSB veteran Igor Girkin ("Strelkov"); the Ukrainian Anti-Terrorist Operation from 13–15 April; the 2 May Odesa Trade Unions House fire; the 11 May "referendums"; the summer Ukrainian advance; the 17 July MH17 shootdown; the mid-to-late-August Russian regular-forces intervention culminating at Ilovaisk; and the 5 September Minsk Protocol — is documented across Ukrainian, Western, OSCE, and (in fragments and admissions) Russian sources, and is treated operationally in UA-C-02. What divides the three accounts is which links in this chain are load-bearing.
Account 1 — the Russian-aggression account — holds that the war was a Russian state operation from its outbreak: local discontent in the southeast was real but produced no insurgency until Russian cadres, weapons, money, and command arrived. Its evidentiary base is the strongest of the three and has grown continuously since 2014: Girkin's own repeated admissions that his unit's 12 April crossing started the war ("I pulled the trigger of war" [TBD-VERIFY: exact wording of the November 2014 Zavtra interview]); the JIT's forensic attribution of the MH17 Buk to Russia's 53rd Anti-Aircraft Missile Brigade and the Hague District Court's 2022 murder convictions; the captured Russian paratroopers and the Ilovaisk record; satellite imagery, intercepts, and the Bellingcat/Atlantic Council open-source corpus; the Glazyev tapes and Surkov leaks showing Kremlin financing and curatorship; and the ECtHR Grand Chamber's findings that Russia exercised effective control over the separatist-held areas from 11 May 2014 [CORRECTED — Tier-2 block audit, 2026-08-30: the admissibility decision date is 30 November 2022, not the previously TBD-VERIFY-tagged "25 January 2023" — see UA-C-02 source 18, which identifies Application No. 28525/20 (the same Ukraine and the Netherlands v. Russia / MH17 application cited here) as joined with 8019/16 and 43800/14 in the Grand Chamber's 30 November 2022 admissibility decision] and the inter-state merits judgment of 9 July 2025. On this account, the "civil war" framing was not a rival description but a Russian information-warfare objective.
Account 2 — the civil-conflict/local-agency account — holds, in its scholarly form, that the war is inexplicable without genuinely local drivers: the Maidan's victory was experienced in parts of the southeast as an illegitimate seizure of power, and Russia militarised a real domestic cleavage it did not create. Its evidentiary base: Serhiy Kudelia's work on the domestic sources of the Donbas insurgency (state collapse, fear, local elite fragmentation); Yuri Zhukov's economic-geography findings that rebellion correlated municipality-by-municipality with industrial exposure to trade with Russia; survey data showing deep southeastern distrust of the post-Maidan government alongside only minority support for secession [TBD-VERIFY: KIIS April 2014 figures]; and the regional-identity scholarship on the Donbas as a borderland with a distinct Soviet-industrial identity. The account's range must be stated: it runs from rigorous scholarship — whose authors mostly insist that local agency does not negate Russian causation — to its propaganda appropriation as the Kremlin's "civil war in Ukraine" narrative, and the scholarly and the propagandistic versions must not be conflated.
Account 3 — the Western-provocation/realist account — holds that the West's post-Cold War expansion, culminating in the 2008 Bucharest declaration (UA-A-04) and the 2013–14 EU/Russia contest over Ukraine, triggered a predictable great-power response, making the war avoidable through neutrality arrangements. Its canonical statements are John Mearsheimer's "Why the Ukraine Crisis Is the West's Fault" (Foreign Affairs, September/October 2014) and Richard Sakwa's Frontline Ukraine (2015), drawing on the offensive-realism literature, the Kennan-to-Burns lineage of NATO-enlargement warnings, and the negotiated-alternative counterfactuals (the collapse of the 21 February 2014 agreement; the federalisation and neutrality proposals of spring 2014). The account is presented here at full strength together with its standing critiques: the denial-of-Ukrainian-agency problem; the empirical challenges (Russian coercion predating enlargement decisions; NATO membership being effectively off the table in 2013, when the proximate trigger was an EU trade agreement); and the predictive record, including the account's own insistence pre-2022 that Russia would not attempt to conquer Ukraine.
The evidence accumulated since 2014 has rebalanced the historiography asymmetrically, and the document tracks this honestly. The JIT and Bellingcat forensic phase (2015–2019), the Surkov leaks (2016) [TBD-VERIFY: provenance and authentication status], the Glazyev tapes, the Hague MH17 judgment (November 2022), the ECtHR inter-state findings, and above all the 24 February 2022 invasion — which demonstrated that maximal Russian war aims existed independent of any Donbas dynamics — have strengthened Account 1's core claims, narrowed Account 2's scholarly version to a subordinate (though still analytically necessary) role, and damaged Account 3's predictive authority while leaving its counterfactual core unfalsifiable. By 2026 the scholarly weight lies clearly with a Russian-aggression reading that incorporates Account 2's local-agency findings as mechanism rather than cause; Account 3 survives principally in IR-theoretical and policy-polemical registers rather than in the empirical historiography of 2014 itself.
Each account performs identifiable political work, and the work must be separated from the evidence. In Moscow, the civil-war and Western-provocation framings were load-bearing state narratives: they justified denying that Russia was a party to the conflict it directed, structured the Minsk agreements around an internal-Ukrainian fiction, and supplied the 2022 invasion's "protecting Donbas" casus belli. In Kyiv, the aggression account underwrote the legal architecture of "temporary occupation," the refusal of direct talks with the DNR/LNR, and the post-2022 claim that the war began in 2014, not 2022. In Western capitals, the choice among accounts mapped onto policy: sanctions logic and (from 2022) arms-supply logic followed Account 1; pre-2022 "both sides" de-escalation diplomacy and post-2022 negotiation advocacy drew on Accounts 2 and 3. Naming this work is not relativism; it is a precondition for reading any of the literatures.
The hardest analytical problem the three accounts pose is the relationship between necessary and sufficient conditions, and the document keeps that distinction in view throughout. Even the strongest version of Account 1 concedes that Russian operatives worked on real material: regional alienation, a collapsed local party machine, and a frightened, Russian-speaking industrial society. Even the most careful version of Account 2 concedes that no anti-Maidan mobilisation produced sustained armed insurgency where Russian cadres and weapons did not arrive — Kharkiv, Odesa, Dnipropetrovsk, and the rest of the southeast subsided, while only the axis reached by Girkin's group and the cross-border pipeline burned. Even Account 3's structural claim, if accepted, explains at most why Moscow had motive — not why a state with a grievance was entitled to, or bound to, partition a neighbour. The corpus's position is that these are different kinds of claims operating at different levels of causation, that they are routinely confused in public debate, and that disentangling them is precisely what the three-account discipline is for.
The test of this document is the J-block standard: a reader formed by each tradition should recognise their account as fairly stated, and a reader from none should come away knowing where the evidence actually points. The evidentiary asymmetry here is real and is stated: Account 1 rests on a forensic, judicial, and archival record without parallel in the other two; Account 2's scholarly core is solid but limited in what it claims; Account 3 is a structural-interpretive framework whose empirical anchors in the 2014 events themselves are thin. Declaring no winner does not mean pretending the accounts are evidentially equal; it means presenting each at its strongest, attributing every judgment, and letting the asymmetry emerge from the record rather than from editorial assertion.

2. The Factual Spine: February–September 2014

This section establishes the sequence on which all three accounts operate. The operational and institutional detail is treated at anchor depth in UA-C-01 (Crimea) and UA-C-02 (the Donbas war's onset and the DNR/LNR trajectory); what follows is the spine required to read the accounts, with the contested links flagged as they appear.

2.1 The Euromaidan endgame and the post-Maidan transition (21–23 February 2014)

The Euromaidan / Revolution of Dignity (UA-B-03) ended in a compressed 48-hour sequence: the 18–20 February killings of approximately 100 protesters (the Heavenly Hundred, UA-B-04); the 21 February EU-mediated Agreement on Settlement of the Political Crisis signed by Yanukovych and the three opposition leaders, with the Russian envoy Vladimir Lukin present but not signing; Yanukovych's overnight departure from Kyiv on 21–22 February; and the 22 February Verkhovna Rada vote (328 in favour) removing him from the presidency without following the Article 111 impeachment procedure — the procedural irregularity on which the "coup" characterisation has rested ever since. On 23 February the Rada voted to repeal the 2012 Kivalov–Kolesnichenko language law, which had given Russian regional-language status; Acting President Oleksandr Turchynov declined to sign the repeal and it never took legal effect [TBD-VERIFY: Turchynov announced on 3 March 2014 that he would not sign pending a replacement law — confirm date], but the vote itself was symbolically toxic in the southeast and became a fixture of Russian state-media coverage within hours. Every account treats these three days as load-bearing; they disagree about what they bear. For Account 1, the transition was a flawed but genuine parliamentary response to a president's flight, instantly weaponised by Moscow. For Account 2, it was experienced in much of the southeast — whatever its legal character — as the victory of one half of the country over the other. For Account 3, it was the collapse of a negotiated settlement that Western governments had witnessed and then declined to enforce.

2.2 The Crimea operation (27 February – 18 March 2014)

On 27 February 2014, unmarked, professionally equipped troops seized the Crimean parliament and Council of Ministers buildings in Simferopol; under armed presence, the parliament installed Sergei Aksyonov — leader of a party that had won 4.02 per cent in the 2010 Crimean elections [TBD-VERIFY: Russian Unity's 2010 vote share] — as prime minister and called a referendum. Russian forces (the "little green men," denied at the time by Moscow) secured the peninsula; the 16 March "referendum" claimed 96.77 per cent for joining Russia on 83.1 per cent turnout, figures rejected by Kyiv, the EU, the US, and the UN General Assembly (Resolution 68/262, 27 March 2014, adopted 100–11–58); the accession treaty was signed in Moscow on 18 March. Putin denied at his 4 March press conference that the unmarked troops were Russian; on 17 April 2014, in his annual Direct Line broadcast, he acknowledged that Russian servicemen had operated behind the Crimean "self-defence forces" [TBD-VERIFY: the 17 April 2014 Direct Line admission and its exact formulation; Putin elaborated the admission in later statements and the March 2015 Crimea: The Way Home documentary]. The Crimea operation matters to the Donbas question in two ways that all accounts register: it demonstrated a Russian decision, taken at the top, to answer the Maidan with military seizure of Ukrainian territory; and its success at near-zero cost shaped the calculus — Moscow's, the southeast's pro-Russian activists', and Kyiv's — for everything that followed. In the same 17 April broadcast Putin used the term Novorossiya ("New Russia") for the arc from Kharkiv to Odesa, the historical-imperial framing under which the spring's project travelled [TBD-VERIFY: the Direct Line transcript's Novorossiya passage].

2.3 The "Russian Spring": the southeast, March–April 2014

From 23 February onward, anti-Maidan crowds gathered in Kharkiv, Donetsk, Luhansk, Odesa, Dnipropetrovsk, Zaporizhzhia, Mykolaiv, and Kherson — around Lenin statues, war memorials, and regional administration buildings. The mobilisations mixed genuine local protest against the post-Maidan transition with organised pro-Russian networks (the Donetsk Republic movement around Pavel Gubarev, Oplot, Russkoye Yedinstvo) and documented bussed-in participants from Russian border regions (the "tourist protesters" tracked by licence plates and social media from late February). The sequence ran: 1 March, the Kharkiv and Donetsk regional administration buildings stormed, Russian flags raised (in Kharkiv the flag was planted by a Moscow resident [TBD-VERIFY: the widely reported identification of the Kharkiv flag-raiser as a Russian citizen, "Mika Ronkainen"]); 1–6 March, Gubarev proclaimed "people's governor" in Donetsk and then arrested by the SBU; 6 April, coordinated seizures of the Donetsk regional administration and the Luhansk SBU headquarters (with its weapons arsenal), and a brief Kharkiv administration seizure reversed by Interior Ministry forces on 7–8 April; 7 April, proclamation of the Donetsk People's Republic inside the occupied building; 27 April, the Luhansk People's Republic. The differential outcomes are the spine's most analytically consequential pattern: in Kharkiv, Odesa, and Dnipropetrovsk — where Ukrainian authorities acted and no armed cadre arrived — the mobilisations subsided; in Donetsk and Luhansk they were converted into armed statelets. The 2 May 2014 Odesa Trade Unions House fire — the day's street fighting between pro-unity and pro-Russian crowds ending with the pro-Russian group besieged in the burning building — killed 42 people inside the building, with 6 killed in the earlier street clashes, approximately 48 in total [TBD-VERIFY: the OHCHR and 2 May Group counts; figures of 42, 46, and 48 circulate depending on inclusion criteria], and roughly 250 injured. It was the deadliest single event of the spring outside the Donbas, terminated the anti-Maidan trajectory in Odesa, and became the central atrocity referent of the Russian "protect the Russian-speakers" narrative; the stalled Ukrainian investigation has remained a genuine stain on Kyiv's accountability record, documented critically by the OHCHR and the Council of Europe's International Advisory Panel [TBD-VERIFY: the IAP's 4 November 2015 report findings on investigation deficiencies].

2.4 The 12 April Sloviansk seizure: the war's conventional start-point

On 12 April 2014, an armed group of approximately 52 men commanded by Igor Vsevolodovich Girkin ("Strelkov") — a Russian citizen, retired FSB colonel [TBD-VERIFY: Girkin's final rank and reserve status], veteran of Transnistria, Bosnia, and Chechnya, and a participant in the Crimea operation — seized the police headquarters, SBU office, and city administration of Sloviansk in northern Donetsk oblast. The unit had assembled in Crimea and crossed into mainland Ukraine [TBD-VERIFY: the crossing route — Girkin's own accounts describe assembly in Crimea and movement across the Russian border region into Donetsk oblast]. Its coordinated multi-site seizures, modern weapons, and tactical professionalism distinguished it categorically from the crowd actions of March; within days it had taken Kramatorsk and made Sloviansk the military centre of gravity of the insurgency. The date is the conventional start-point of the Donbas war in Ukrainian law and most scholarship, and Girkin's own retrospective statements — that without his unit's crossing "everything would have fizzled out, as in Kharkiv and Odesa," and that "it was I who pulled the trigger of war" [TBD-VERIFY: the November 2014 Zavtra interview's exact wording, widely quoted as "Я нажал на спусковой крючок войны"] — are the single most cited piece of testimony in the entire origins debate. On 13 April the National Security and Defence Council announced, and on 15 April formally launched, the Anti-Terrorist Operation (ATO) under SBU command — the institutional framing (counter-terrorism rather than war or martial law) that preserved the 25 May presidential election and that Ukraine maintained until the 2018 Joint Forces Operation reframing (UA-C-02).

2.5 The 11 May "referendums" and the consolidation of the DNR/LNR

On 11 May 2014, the DNR and LNR authorities staged "referendums" on "self-rule" (samostoyatelnost, an ambiguous term between autonomy and independence) — without electoral rolls, procedural safeguards, or legal basis, in territory only partially under their control. The self-announced results (DNR: 89.07 per cent in favour on 74.87 per cent turnout; LNR: 96.2 per cent on 81 per cent [TBD-VERIFY: self-reported figures vary across DNR/LNR statements]) were rejected by Kyiv, the EU, the US, and the OSCE; notably, Putin had publicly asked on 7 May that the referendums be postponed — a request the organisers ignored or were scripted to ignore, a datum each account reads differently. The 25 May presidential election returned Petro Poroshenko with 54.7 per cent in the first round; voting was impossible in Crimea and in most separatist-held Donbas precincts. On 26 May the first battle of Donetsk airport — involving the Vostok battalion, with fighters from Russia among the dead returned across the border — marked the transition to open warfare in the oblast capital.

2.6 The summer escalation, MH17, and the August intervention

Through June and July the reconstituted Ukrainian forces, with volunteer battalions, pressed the ATO offensive: Sloviansk and Kramatorsk were retaken on 5 July as Girkin's force withdrew into Donetsk city; by early August Ukrainian forces had constricted the separatist-held territory to the Donetsk and Luhansk agglomerations and were operating along the border. Three developments in this phase carry the heaviest evidentiary weight in the origins debate. First, the supply pipeline: heavy weapons — tanks, MLRS, air-defence systems — appeared in separatist hands from June, documented by NATO imagery and open-source tracking as transiting from Russia. Second, the cross-border fires: from mid-July, artillery and rocket strikes on Ukrainian positions were fired from inside Russian territory, documented by Bellingcat's crater-analysis work [TBD-VERIFY: Bellingcat's 2015–2016 cross-border-attacks reports]. Third, MH17: on 17 July 2014 Malaysia Airlines Flight 17 was destroyed over Hrabove, killing all 298 aboard. The Dutch Safety Board (October 2015) established the cause as a Buk missile; the Joint Investigation Team established (May 2018) that the Buk TELAR belonged to the 53rd Anti-Aircraft Missile Brigade of the Russian Armed Forces, brought from Kursk into Ukraine and back; the District Court of The Hague (17 November 2022) convicted Girkin, Sergey Dubinsky, and Leonid Kharchenko in absentia of 298 murders, finding that the missile and crew came from Russia and that Russia exercised overall control over the DNR from mid-May 2014 [TBD-VERIFY: the judgment's overall-control finding and its date-anchor; the conviction/acquittal pattern across the four defendants — Pulatov acquitted].

In mid-to-late August, with the DNR/LNR facing military collapse, Russian regular forces intervened directly: armoured columns crossed the border from approximately 16 August [TBD-VERIFY: the intervention's start-date range of 14–24 August across sources]; on 24–29 August Ukrainian forces advancing into Ilovaisk were encircled by Russian regular units and destroyed withdrawing along a negotiated corridor that came under fire on 29 August, with Ukrainian official death tolls between approximately 366 (the Rada commission, 2016) and 459 (Ministry of Defence statements) [TBD-VERIFY: reconciliation of the Ilovaisk casualty figures]. Ten paratroopers of Russia's 98th Guards Airborne Division were captured inside Ukraine on or about 25 August [TBD-VERIFY: nine or ten captured; Moscow's "lost their way on patrol" explanation]; Russian soldiers' funerals in Pskov and elsewhere were documented by Russian journalists who were then assaulted or obstructed. On 28 August NATO released satellite imagery of Russian combat units inside Ukraine; Moscow denied everything. The military reversal forced Kyiv to the table: the Minsk Protocol was signed on 5 September 2014 by the Trilateral Contact Group (Ukraine, Russia, OSCE) with DNR/LNR representatives, followed by the 19 September memorandum — the architecture treated at UA-C-04, and the point at which this document's factual spine closes.

2.7 What the spine does and does not settle

The spine settles more than the public debate usually admits: that Russian state personnel seized Crimea before any Donbas violence; that the armed phase began with a cross-border seizure led by a Russian security-services veteran; that Russian heavy weapons, cross-border fires, and ultimately regular formations sustained and saved the insurgency; and that the deadliest single atrocities of the period (MH17; Ilovaisk's corridor; Odesa's fire) each have a documented record. It does not by itself settle the three questions on which the accounts divide: whether the insurgency would have arisen in some form without Russian direction (the counterfactual at the heart of Accounts 1 and 2); how much causal weight the southeast's genuine alienation carries relative to the imported cadre (the weighting question); and whether Russian conduct, however documented, is best explained as unprovoked revisionism or as a response within a longer interactive spiral with the West (the framing question at the heart of Account 3). To those questions the document now turns.

3. Account 1: The Russian-Aggression Account

3.1 Provenance and evidentiary base

The Russian-aggression account is the position of the Ukrainian state in all its legal instruments from the 15 April 2014 Occupied Territories Law onward; of the JIT states and, progressively, of Western governments; and of the dominant stream of the scholarship — Andrew Wilson (Ukraine Crisis: What It Means for the West, 2014), Taras Kuzio (Putin's War Against Ukraine, 2017), Mark Galeotti (Armies of Russia's War in Ukraine, 2019, and his "hybrid war" analyses), Nikolay Mitrokhin ("Infiltration, Instruction, Invasion: Russia's War in the Donbass," 2015), Serhii Plokhy (The Russo-Ukrainian War, 2023), and Paul D'Anieri (Ukraine and Russia: From Civilized Divorce to Uncivil War, 2019, with qualifications discussed under Account 3). Its evidentiary base is of a different kind from the other accounts': not interpretive but forensic, judicial, testimonial, and archival.

Its principal strata, in rough order of accumulation: (i) the contemporaneous operational record — the professionalism of the 12 April seizures; OSCE Special Monitoring Mission reporting; the documented weapons pipeline and cross-border fires; NATO imagery; the captured 98th Division paratroopers and the soldiers' funerals. (ii) The participants' own admissions — above all Girkin, whose November 2014 Zavtra interview and subsequent years of interviews and Telegram commentary state plainly that his unit started the war and that without it the protest wave would have died as it did in Kharkiv and Odesa [TBD-VERIFY: Zavtra exact wording]; but also Alexander Borodai (the Moscow political-technologist who served as the DNR's first "prime minister" and later described himself and the leadership cadre as Russians sent to do a job [TBD-VERIFY: Borodai's 2014–2015 interview statements]), and the parade of Russian citizens — Antyufeyev, Bezler, and others — in the statelets' 2014 leadership. (iii) The forensic open-source corpus — Bellingcat's MH17 and cross-border-attack investigations; the Atlantic Council's Hiding in Plain Sight (2015); the Conflict Intelligence Team's tracking of Russian servicemen. (iv) The intercept and leak record — the SBU-released intercepts of spring 2014; the Glazyev tapes (intercepted late-February/March 2014 calls, released by the Ukrainian Prosecutor General in 2016, in which presidential adviser Sergei Glazyev coordinates financing and crowd-mobilisation for administrative-building seizures in Kharkiv, Odesa, and Donetsk [TBD-VERIFY: the tapes' release date, August 2016, and their authentication status; Russia has not contested their authenticity in court]); and the Surkov leaks (October 2016 –, hacked email caches from the office of presidential aide Vladislav Surkov published by the Ukrainian Cyber Alliance, showing Kremlin curatorship of DNR/LNR personnel, budgets, and media lines [TBD-VERIFY: authentication assessments by the Atlantic Council's DFRLab and independent analysts]). (v) The judicial record — the Hague District Court's 2022 MH17 judgment with its overall-control finding; the ICC Prosecutor's preliminary-examination conclusion that the Donbas conflict was international in character no later than 14 July 2014 [TBD-VERIFY: the OTP's 2016 Report on Preliminary Examination Activities formulation]; and the ECtHR's inter-state findings (Section 6).

3.2 The core claims

First: the war was a Russian state operation from the start, in design even where not in every executor. The account reads the spring of 2014 as one campaign with two theatres. Crimea was the proof of concept: a deniable military seizure, denied and then admitted. The Donbas was the second application — the Novorossiya project — run at lower cost through proxies, with the FSB and GRU providing cadre (Girkin's unit; the "vacationers"), the Glazyev network providing money and political technology, Russian state media providing the "fascist junta" mobilisation frame, and the regular army standing behind the operation as guarantor, deployed openly in August when the proxy layer failed. On this reading the question "did Russia start the war?" is answered by the operation's architecture, not by counting locals among the gunmen: an insurgency whose military founder, first political leader, security chiefs, heavy weapons, operational fires, and strategic rescue all came from across the border is a foreign operation with local participation, not a local rebellion with foreign help.

Second: local discontent was real but causally insufficient — the counterfactual is supplied by the southeast itself. The account does not deny the alienation documented by Account 2's scholarship; it denies that alienation produces armed statelets. Its strongest evidence is comparative and internal to Ukraine: the same post-Maidan shock, the same Russian-speaking industrial sociology, and similar anti-Maidan mobilisations existed in Kharkiv (which Avakov's forces cleared on 7–8 April), in Odesa (where the anti-Maidan movement collapsed after 2 May), and in Dnipropetrovsk (where the regional authorities held the streets) — and in none of them did insurgency follow, because in none of them did the cross-border cadre and pipeline arrive in force. Girkin's own testimony makes exactly this point from the inside. The minority-support survey data (KIIS, April 2014) cut the same way: a region in which separatism commanded a quarter to a third of opinion [TBD-VERIFY: figures] does not secede by referendum; it is seceded by armed men.

Third: the "civil war" framing was itself a Russian information objective, and treating it as a neutral description repeats the operation. The account documents that denial-of-involvement was not incidental but doctrinal: the little-green-men denial later admitted; the "they bought their tanks in a shop" line; the "lost paratroopers"; the Minsk architecture itself, which Russia structured as mediator rather than party. On this reading, every Western usage of "Ukrainian civil war," "separatist conflict," or "Kyiv's war against its own east" between 2014 and 2022 was a measurable information-warfare yield, and the scholarly civil-conflict literature — whatever its authors' intentions — supplied citable authority for it. The account's analysts (Wilson, Galeotti, Kuzio) wrote extensively against the "hybrid war" mystique precisely to insist that the hybridity was a deniability wrapper around a conventional state operation.

Fourth: the 2022 invasion retrospectively confirmed the account's reading of intent. If the Donbas war had been fundamentally a civil conflict that Russia merely exploited, the 2022 decision — a full-spectrum invasion aimed at Kyiv, justified by the "protection" of the very statelets Moscow had claimed for eight years not to control, preceded by the 21 February 2022 recognition speech in which Putin denied Ukrainian statehood as such — is inexplicable. If 2014 was the first, constrained application of a standing revisionist programme, 2022 is its continuation at higher risk tolerance. The account treats this as the strongest kind of confirmation available in historical explanation: the actor's own subsequent behaviour and stated reasoning.

3.3 Internal tensions and the account's burdens

Stated fairly, the account carries burdens. Its monolith problem: the documentary record (including the Surkov leaks and Girkin's own furious testimony against Moscow's half-measures) shows a Russian operation that was improvised, factionalised, and repeatedly surprised — Girkin's crossing appears to have been licensed and enabled rather than centrally ordered as part of a worked-out invasion plan [TBD-VERIFY: the state of evidence on prior FSB authorisation of the 12 April crossing], the Novorossiya project was abandoned in the autumn of 2014, and Putin's 7 May call to postpone the referendums sits awkwardly with a fully scripted reading. The strongest scholarly versions (Galeotti, Mitrokhin) therefore describe an adhocratic, opportunistic aggression — multiple Russian state and para-state actors competing to deliver the Kremlin's broadly signalled wish — rather than a single planned campaign, and the account must resist its own polemical temptation to backdate 2022's deliberateness onto 2014. Its local-erasure problem: in Ukrainian public memory the account has sometimes flattened the real 2014 sociology of the Donbas — the genuine local fighters, the genuine anti-Maidan constituency — into an undifferentiated story of invasion, which complicates both reintegration policy and historical accuracy; the scholarship is better than the public version on this point. And its selection problem: the account's forensic corpus is strongest exactly where international institutions had jurisdiction and victims (MH17), and thinner on the war's inner political history on the Ukrainian side — the volunteer battalions' conduct, the Odesa investigation's failure — which the account's critics read as motivated asymmetry of attention, and which this corpus addresses by documenting those records where they belong (UA-C-02; and the planned J-block document on the 2014–2019 casualty and displacement record).

4. Account 2: The Civil-Conflict/Local-Agency Account

4.1 Provenance, and the essential distinction within the account

The second account must be introduced with a distinction that its own leading scholars insist on: there is a scholarly civil-conflict literature, which argues that local drivers were necessary conditions of the insurgency and that the conflict cannot be explained as invasion alone; and there is a propaganda "civil war in Ukraine" narrative, deployed by the Russian state from March 2014 onward, which used superficially similar language to deny Russian belligerency. The two share vocabulary and almost nothing else. The scholars discussed here — Serhiy Kudelia, Yuri Zhukov, Elise Giuliano, Gwendolyn Sasse, Gerard Toal and John O'Loughlin, Anna Matveeva (whose Through Times of Trouble, 2018, sits at the account's most Russia-sympathetic scholarly edge), and the contributors to Jakob Hauter's Civil War? Interstate War? Hybrid War? (2021) — overwhelmingly acknowledge Russian involvement as documented fact; their claim is about causal weighting, not about Russian absence. Kudelia, the account's most prominent voice after his September 2014 PONARS memo "Domestic Sources of the Donbas Insurgency" and the sharp exchange with Andreas Umland and Taras Kuzio that followed, has stated explicitly that emphasising domestic sources does not absolve Russia [TBD-VERIFY: Kudelia's formulations in the 2014 PONARS memo and the 2016 follow-up literature]. The account's propaganda appropriation is treated in Section 7; this section presents the scholarship.

4.2 The evidentiary base

The account's evidence is social-scientific rather than forensic: survey data, electoral and economic geography, and field research. (i) The survey record. The April 2014 KIIS survey of the eight southeastern oblasts (conducted 8–16 April for Dzerkalo Tyzhnia [TBD-VERIFY: fieldwork dates]) found majority distrust of the post-Maidan authorities in Kyiv and substantial sympathy for federalisation — but only minority support for secession or union with Russia even in Donetsk and Luhansk (roughly 27–30 per cent in Donetsk oblast favouring some form of separation, with clear majorities opposed [TBD-VERIFY: the KIIS April 2014 toplines; the companion Pew Research April 2014 finding that approximately 70 per cent of eastern respondents wanted Ukraine to remain united]). The account reads this two-sidedly: the minority-secessionism finding limits what Russia had to work with, but the majority-distrust finding is the account's core datum — a region that did not want to leave Ukraine also did not regard the new Kyiv government as legitimately its own. (ii) The economic geography. Zhukov's municipality-level analysis ("Trading Hard Hats for Combat Helmets," Journal of Comparative Economics, 2016) found that rebellion within the Donbas correlated strongly with localities whose industries — machine-building above all — depended on exports to Russia and stood to lose from the EU Association Agreement's trade reorientation; economic exposure outperformed ethnicity and language as a predictor of rebel control [TBD-VERIFY: the paper's headline findings]. (iii) The political-opportunity analysis. Kudelia's work locates the insurgency's enabling conditions in the February–April collapse of state authority in the region: the disintegration of the Party of Regions machine, the defection or paralysis of the local police and SBU (documented in the effortless April seizures), the fear cascade produced by Maidan-era violence and the language-law vote, and the availability of local strongmen and marginal counter-elites for whom the crisis was opportunity. (iv) The identity scholarship. The longer Donbas literature — Hiroaki Kuromiya's Freedom and Terror in the Donbas (1998), Sasse's borderlands work, the post-2014 Toal/O'Loughlin surveys inside the separatist-held territories — documents a regional identity that was distinctively Soviet-industrial and Russophone without being straightforwardly pro-Russian-statehood: a region accustomed to viewing Kyiv governments instrumentally and to regarding itself as the economic donor and political hostage of western-Ukrainian projects.

4.3 The core claims

First: the Maidan's victory was experienced in much of the southeast as an illegitimate seizure of power, and that experience was locally real, not manufactured. The account assembles the sequence as it appeared from Donetsk rather than from Kyiv: a president for whom the region had voted at near-90-per-cent levels removed by a procedure that did not follow Article 111, within days of an EU-witnessed agreement that would have let him serve out the year; a parliament that made the repeal of the language law one of its first acts; the Berkut — recruited heavily from the southeast — returning home stigmatised; armed Maidan self-defence groups and the rhetoric of lustration filling the airwaves. None of this required Russian television to be alarming, though Russian television amplified all of it. The anti-Maidan crowds of March were, on this account, a genuine mirror of the Maidan itself — a regional population mobilising against a government it regarded as imposed — and the later insurgency drew its local manpower from that real constituency, not from phantoms.

Second: Russia exploited and militarised a real domestic cleavage it did not create. The account accepts the Girkin record, the pipeline, and the August intervention; its claim is about the material those instruments worked on. The cleavage — regional, economic, linguistic, and memory-political — is visible in every Ukrainian electoral map from 1994 to 2012 (UA-B-01), in the federalisation debates that long predate 2014, and in the survey record above. On this reading, the right analogy is neither pure invasion nor pure rebellion but ignition: Russian cadres supplied the spark, command, and accelerant; the combustible material was local and decades in the making. Causation is therefore genuinely joint, and accounts that delete either element fail — the invasion-only account cannot explain why the Donbas burned when Kharkiv did not without quietly invoking local variables (the depth of the machine's collapse, the border's proximity, the region's sociology), and the rebellion-only account cannot explain why nothing burned anywhere until armed men crossed borders.

Third: without the local dimension, the war's durability and the statelets' consolidation are inexplicable. The DNR/LNR fielded tens of thousands of locally recruited fighters across 2014–2021 [TBD-VERIFY: force-composition estimates; by 2015–2016 the 1st and 2nd Army Corps were majority-local in rank-and-file under Russian command architecture]; the territories' administration, however Kremlin-curated, ran on local personnel; and the conflict's eight-year persistence — through Minsk, blockade, and grinding attrition — required a population that, whatever its preferences in March 2014, was progressively bound to the de facto authorities by violence experienced (including from Ukrainian fires on separatist-held cities), by economic separation, and by the absence of a credible reintegration offer. Toal and O'Loughlin's surveys inside the territories found, across 2016–2019, substantial genuine support for the de facto authorities and for integration with Russia — support the account reads as partly produced by the war itself, but real [TBD-VERIFY: the ZOiS (Sasse) and Toal/O'Loughlin survey findings and dates]. The policy entailment drawn by the account's scholars was that any settlement would have to address the region's political reincorporation, not merely Russian withdrawal.

Fourth: the local-agency claim is a claim about necessary conditions, and its scholars' own ceiling on it must be quoted against its abusers. Kudelia's position — that domestic sources were necessary but that the insurgency's military form, survival, and escalation were Russian-dependent — is representative of the scholarly account's centre of gravity. Zhukov's economic findings explain variation within the Donbas, not the existence of the war. Giuliano's work on who actually supported the separatists found the support base concentrated, conditional, and motivated more by fear of the new Kyiv government and economic anxiety than by ethnic-Russian irredentism [TBD-VERIFY: Giuliano's Ethnopolitics/Post-Soviet Affairs findings]. The scholarly account, stated at its own chosen altitude, is a theory of the insurgency's social base — not a theory that the war was Ukrainian-on-Ukrainian.

4.4 Internal tensions and the account's burdens

The account's burdens are threefold. The appropriation burden: its vocabulary was conscripted by the belligerent it least resembles. From Moscow's first framing of the conflict as a civil war through the Minsk architecture's internal-conflict fiction to the 2022 "genocide in Donbas" casus belli, the Kremlin cited the existence of a civil-conflict framing — including, selectively, Western scholarship — as authority. The scholars are not responsible for the theft, but the account as a public object cannot be evaluated apart from it, and several of its authors spent the post-2022 period explicitly disentangling their findings from the Russian narrative. The evidentiary-update burden: parts of the account's empirical foundation aged badly as the archival record grew — the Glazyev tapes showed Russian money and direction behind the "spontaneous" March mobilisations earlier and deeper than the 2014-era scholarship assumed; the Surkov leaks showed the statelets' political life as curated product; and the ECtHR's effective-control findings from 11 May 2014 compressed the window in which a genuinely autonomous local insurgency can be said to have existed to a few weeks at most. The scholarly account has partially absorbed these updates (Hauter's 2021 volume, applying source-critical method to the escalation sequence, concluded the war's escalation was overwhelmingly externally driven [TBD-VERIFY: Hauter's conclusions]); its weaker popularisations have not. The moral-asymmetry burden: even where the account's causal claims hold, "joint causation" sentences can perform false equivalence — a cleavage is not a combatant, and no reading of the survey data converts a 30-per-cent secessionist minority into a belligerent party with standing equal to a state that sent armoured columns across an international border. The careful versions of the account say this; the account is burdened by the versions that do not.

5. Account 3: The Western-Provocation/Realist Account

5.1 Provenance and intellectual lineage

The third account is not primarily a historiography of the Donbas; it is a structural theory of the crisis within which the Donbas war occurred, and it must be presented as such. Its canonical statement is John Mearsheimer's "Why the Ukraine Crisis Is the West's Fault: The Liberal Delusions That Provoked Putin" (Foreign Affairs, September/October 2014), built on his offensive-realist framework (The Tragedy of Great Power Politics, 2001): great powers police their neighbourhoods, NATO's eastward expansion (1999, 2004) and the 2008 Bucharest declaration that Ukraine and Georgia "will become members" (UA-A-04) crossed a repeatedly stated Russian red line, and the EU's Eastern Partnership plus Western democracy-promotion constituted, in Moscow's reading, a single package aimed at detaching Ukraine from Russia's orbit. Richard Sakwa's Frontline Ukraine: Crisis in the Borderlands (2015) supplies the account's fullest book-length version, framing the catastrophe as the collision of a "wider Europe" (EU/NATO monism) with a "greater Europe" (a pan-continental order including Russia), compounded by what Sakwa calls the contest between "monist" and "pluralist" conceptions of Ukrainian statehood itself. The lineage behind both is real and distinguished: George Kennan's 1997–98 warnings that NATO expansion would be "the most fateful error of American policy in the entire post-Cold War era"; William Burns's 2008 "Nyet means nyet" cable from Moscow reporting that Ukrainian membership was "the brightest of all redlines" across the entire Russian elite [TBD-VERIFY: the cable's exact phrasing, published via WikiLeaks]; the 2008-era opposition of the German and French governments to the Membership Action Plan; and the "Everyone Loses" interactive-spiral literature (Samuel Charap and Timothy Colton, 2017), which occupies the account's most empirically careful wing. Stephen Cohen, Jack Matlock, and parts of the European left and the American restraint school carried the account into public debate; after 2022, Mearsheimer's lectures became among the most-viewed political-science content in the world, and the account structured a substantial share of global-South commentary on the war.

5.2 The core claims

First: the crisis was structurally produced by post-Cold War Western expansion into a sphere a declining great power had repeatedly declared vital. The account's causal chain runs: NATO enlargement over Russian objection (1999, 2004) → the 2008 Bucharest declaration, adopted over the explicit warnings of the US ambassador in Moscow and of allied governments → the 2009 Eastern Partnership and the EU Association Agreement, read in Moscow as the economic limb of the same project (the DCFTA being incompatible with Russia's Customs Union project) → the Western embrace of the Maidan — visiting Western politicians on the square, the Nuland–Pyatt leaked call of 4 February 2014 discussing the composition of a post-Yanukovych government — read in Moscow as the orchestration of a coup against an elected, Russia-aligned president → the Crimea and Donbas operations as the predictable, indeed predicted, defensive reflex of a power that could not stop the geopolitical loss of Ukraine by other means. On this reading Putin's conduct in 2014 required no theory of imperial ideology: any Russian leadership, the account holds, would have responded comparably, as Burns's cable had reported the entire elite spectrum demanding.

Second: the Maidan's Western backing made Moscow read Kyiv 2014 as a Western coup, and the 21 February agreement's collapse is the account's pivotal counterfactual. The account places unusual weight on the 48 hours treated at §2.1: an agreement brokered and witnessed by three Western foreign ministers — early elections, constitutional reform, a national-unity government — was overtaken within a day by the Rada's procedurally irregular removal vote, and the Western governments that had witnessed it recognised the successor authorities without insisting on its terms. Whatever the domestic dynamics that actually drove those hours (the Maidan's refusal of the deal after the 20 February killings; Yanukovych's flight as abandonment), the account holds that Moscow's reading — a Western-blessed unconstitutional transfer of power — was reasonable on the visible evidence, and that the episode confirmed to the Kremlin that negotiated outcomes with the West over Ukraine would not be honoured. The subsequent counterfactuals follow: a federalised, militarily neutral Ukraine — proposed in various forms by Kissinger (March 2014), by the Russian foreign ministry's federalisation demands, and implicit in the account's own policy advocacy — would, it claims, have averted the war or ended it early, at the price of constraints on Ukrainian alignment that the account regards as the normal tribute geography exacts from buffer states.

Third: the Donbas war specifically was the foreseeable product of contested-alignment dynamics, not the first move of an unlimited revisionist programme. Mearsheimer's 2014 article argued that Putin had no interest in conquering Ukraine and that the annexation of Crimea plus pressure in the east was coercive diplomacy aimed at blocking Ukraine's Western integration — and, on the account's own terms, the 2014–2021 record fit: Russia declined to annex the Donbas, abandoned Novorossiya, and channelled the conflict into Minsk-2's reintegration-with-veto architecture, exactly what a sphere-of-influence (rather than conquest) objective predicts. The account therefore read the 2014 war as tragic, avoidable, and instrumental — and read the West's post-2014 policy (sanctions plus military assistance plus non-implementation of Minsk's political clauses) as prolonging a war that alignment-neutrality could settle.

5.3 The critiques, stated at full strength

The corpus presents this account's critiques as part of presenting the account, because they are by 2026 inseparable from its standing.

The agency critique. The account explains Ukraine almost entirely as an object — a buffer, a prize, a "borderland" — and Ukrainian choices (the Maidan itself, the 2014 electorate's turn, the consistent post-2014 supermajorities for Western integration) as Western inputs or as noise. Critics across the spectrum (Toal, who is himself critical of NATO policy, called this the "thin geopolitics" problem; Ukrainian scholarship nearly unanimously) note that the account's causal chain requires treating the most consequential mass political action in 2014 — millions of Ukrainians' own conduct — as epiphenomenal, and that "the West did it" and "Russia was made to do it" share a structure in which Ukrainians do nothing. The account's reply — that structural theory operates at the level of great-power incentives, not of small-state preferences — concedes the point in the act of answering it.

The empirical critique. Key links in the chain are weaker than the account's confidence suggests. Russian coercion of Ukraine predates the Bucharest declaration (the 2003 Tuzla island confrontation; the 2006 and 2009 gas cut-offs; the long pre-2008 pressure on the Black Sea Fleet question). NATO membership for Ukraine was moribund by 2013: Ukraine's 2010 non-bloc law (UA-B-01) had taken it off the table, no MAP existed, and public support for membership in pre-2014 Ukraine polled at minority levels — it was Russia's annexation of Crimea that converted Ukrainian opinion [TBD-VERIFY: pre-2014 vs post-2014 NATO-support polling series]. The proximate object of the 2013–14 contest was an EU trade agreement, not a military alliance, and the EU is a poor fit for offensive-realist threat logic. The "coup" reading of February 2014 sits badly with the documented record of the Maidan's internal dynamics (UA-B-03) and of Yanukovych's flight. And Putin's own stated justifications evolved away from the account: the 18 March 2014 Crimea address and the July 2021 essay "On the Historical Unity of Russians and Ukrainians" ground the claim in civilisational-historical entitlement, not NATO — the 2021 essay barely engaging enlargement at all [TBD-VERIFY: characterisation of the essay's NATO content].

The predictive critique. The account's standing was damaged by its own record. Mearsheimer wrote in 2014 that Putin was too smart to attempt to conquer Ukraine and that Russia lacked the capacity to do so [TBD-VERIFY: exact formulations in the 2014 article and 2015 lecture]; the account's adherents broadly dismissed the US intelligence warnings of late 2021 as alarmism; and the 24 February 2022 invasion — aimed at Kyiv, justified by denial of Ukrainian nationhood, launched against a state whose NATO accession was nowhere in prospect — was precisely the unlimited-aims behaviour the account had ruled out. The post-2022 adaptation (the invasion as preventive war against Ukraine's de facto NATO integration) preserved the framework at the cost of unfalsifiability: a theory that predicts restraint, and then accommodates its opposite, is doing interpretation, not prediction.

The asymmetry critique. "Provocation" performs analytical work in one direction only. The account treats Russian red lines as structural facts to be respected and Ukrainian (or Polish, or Baltic) security preferences as destabilising choices to be managed; it counts NATO's openness to Ukraine as a cause of war but does not count Russia's standing demand for a sphere of influence as one. Critics note that the same realist toolkit, applied symmetrically, generates the opposite policy conclusion drawn by the states nearest Russia: that the unprovoked-aggression record (Chechnya, Georgia 2008, Crimea) made hedging toward NATO rational — which is why enlargement was demand-driven from the east, not supply-driven from Washington.

5.4 What survives of the account

Presented at full strength, the account retains three durable contributions that the corpus records without endorsing its causal headline. It correctly identifies that Western policy 1991–2014 never resolved the contradiction between enlargement's open door and the unwillingness to defend the states in the doorway — Ukraine in 2014 was in the worst position the policy mix could produce: encouraged, exposed, and unprotected. It correctly reports that the Russian elite's hostility to Ukrainian Western alignment was broad, early, and consistently communicated — the warning record (Kennan, Burns) is genuine even if the inference from warning to justification is not. And its counterfactuals, though unfalsifiable, force the historiography to specify what exactly is being claimed when the war is called "unprovoked": not that Moscow lacked stated grievances, but that the grievances licensed nothing, and that the decision to convert them into the armed partition of a neighbour was made in Moscow and nowhere else.

6. The Evidence Since 2014: The Historiographical Rebalancing

The origins debate did not stand still after September 2014; it was progressively re-litigated by forensic investigation, leaked archives, and international courts, and then violently reframed by the 2022 invasion. This section tracks what each evidentiary wave did to each account's standing — the rebalancing tracked honestly, including where it cut against the corpus's strongest-evidenced account.

6.1 The forensic phase: JIT, Bellingcat, and the open-source revolution (2014–2019)

MH17 made the Donbas the proving ground of open-source forensic investigation. Bellingcat's photo-and-video geolocation of Buk 3x2's journey from the 53rd Brigade's Kursk base into Ukraine and back (2014–2016), the Atlantic Council's Hiding in Plain Sight (May 2015) on Russian servicemen and equipment in the Donbas, the Conflict Intelligence Team's soldier-by-soldier identifications, and Bellingcat's cross-border artillery analyses together created something the 2014 information environment had lacked: a public, checkable, cumulative evidentiary record independent of any government. The JIT's milestones — the 28 September 2016 launch-site finding (Pervomaiskyi, separatist-held); the 24 May 2018 attribution of the TELAR to the 53rd Anti-Aircraft Missile Brigade of the Russian Armed Forces; the 19 June 2019 naming of Girkin, Dubinsky, Pulatov, and Kharchenko — converted open-source method into prosecutorial product. Following the May 2018 JIT attribution, the Netherlands and Australia formally held the Russian Federation responsible under international law for its role in the downing [TBD-VERIFY: the 25 May 2018 joint state-responsibility announcement]. The effect on the accounts was direct: Account 1's claims about the supply pipeline and Russian personnel moved from intelligence assertion to documented public record; Account 2's scholarly wing absorbed the record (its claims, as noted, never depended on Russian absence); the Russian state's serial counter-narratives — the fabricated radar claims, the doctored imagery, the succession of mutually exclusive alternative culprits documented and debunked across 2014–2018 — became themselves part of the record, studied as information warfare.

6.2 The archival phase: the Glazyev tapes and the Surkov leaks (2016–)

Two leak corpora moved the debate from the battlefield to the Kremlin's paperwork. The Glazyev tapes — intercepted calls from late February and March 2014, released by the Ukrainian Prosecutor General in August 2016 in the context of the treason case against Yanukovych [TBD-VERIFY: release date and proceeding] — record presidential adviser Sergei Glazyev arranging financing for anti-Maidan actions, instructing local actors in Kharkiv, Odesa, and Donetsk to storm administrative buildings and generate appeals for Russian assistance, weeks before the April seizures. If authentic — and they have been treated as authentic across the scholarly literature, with Moscow dismissing rather than rebutting them [TBD-VERIFY: authentication status and Russian responses] — they date organised Russian direction of the "Russian Spring" to its very first days, substantially before the period in which the 2014-era civil-conflict literature located autonomous local mobilisation. The Surkov leaks — email caches from the office of Vladislav Surkov, the presidential aide who ran the Donbas portfolio, hacked and published by the Ukrainian Cyber Alliance from October 2016, assessed as authentic by the Atlantic Council's Digital Forensic Research Lab among others [TBD-VERIFY: DFRLab's authentication finding; the leaks' multiple tranches through 2017] — document Kremlin curatorship of DNR/LNR cabinet lists, budgets, media messaging, and even the statelets' "constitutional" texts. Together with the Hauter volume's source-critical reconstruction of the escalation sequence (2021), the archival phase shifted the scholarly centre of gravity: by the late 2010s the live question in the serious literature was no longer whether the insurgency was Russian-directed but when direction became control, and how much the local substrate mattered within a Russian-run operation — Account 2's strong versions narrowing toward Account 1's qualified versions.

6.3 The judicial phase: The Hague, Strasbourg, and the ICC (2021–2025)

Three judicial tracks converted the record into law. The District Court of The Hague (17 November 2022), convicting Girkin, Dubinsky, and Kharchenko of murdering MH17's 298 occupants, made findings beyond the verdicts: that the Buk and its crew came from the Russian armed forces, and that from mid-May 2014 the Russian Federation exercised overall control over the DPR — a finding made for the purpose of the conflict's legal qualification [TBD-VERIFY: the judgment's overall-control reasoning and date anchor]. The European Court of Human Rights, in the consolidated inter-state cases Ukraine and the Netherlands v. Russia, held at the admissibility stage (Grand Chamber decision, 30 November 2022 [CORRECTED — Tier-2 block audit, 2026-08-30: prior text read "25 January 2023"; corrected to match the internally-decisive resolution per UA-C-02 source 18, which identifies Application No. 28525/20 as joined in the Grand Chamber's 30 November 2022 admissibility decision]) that Russia had effective control over the separatist-held areas of the Donbas from 11 May 2014, by virtue of its military, political, and economic support and the presence of its personnel — and in its merits judgment of 9 July 2025 found systematic violations of the Convention attributable to Russia across the 2014–2022 Donbas conflict and the post-2022 full-scale war, including in respect of MH17 [TBD-VERIFY: the 9 July 2025 Grand Chamber judgment's principal holdings; Russia ceased participating in the proceedings after its March 2022 expulsion from the Council of Europe, and the judgment's enforceability is moot in practice]. The ICC track ran differently: the Office of the Prosecutor's preliminary examination had concluded as early as 2016 that the Ukraine situation included an international armed conflict — direct engagements between Russian and Ukrainian forces indicating an IAC from no later than 14 July 2014, and the shelling from Russian territory supporting the same conclusion [TBD-VERIFY: the OTP's 2016–2020 preliminary-examination formulations] — but the court's eventual arrest warrants (from March 2023) concerned post-2022 conduct. The judicial phase's net effect: the "civil war" characterisation of the post-May-2014 conflict is now untenable as a matter of adjudicated international law, whatever its remaining life in political rhetoric; the February–April 2014 window — the weeks between the Maidan's victory and Girkin's crossing — remains the only period for which an autonomous-local-mobilisation description retains any legal or evidentiary footing, and the Glazyev tapes contest even that.

6.4 The 2022 reframing — and the duty of honesty about hindsight

The 24 February 2022 invasion settled, for most of the field, the question of Russian intent toward Ukraine — and thereby rebalanced the 2014 historiography more forcefully than any document could. Putin's 21 February 2022 recognition address and the invasion's stated aims ("demilitarisation and denazification"; the protection of the Donbas statelets; the denial, in the July 2021 essay and the recognition speech, that Ukraine possessed authentic statehood) demonstrated that maximal revisionist objectives existed and were actionable independent of NATO accession dynamics or Donbas battlefield events. Account 1 absorbed 2022 as confirmation. Account 2's scholarly wing largely re-described its own findings as a theory of the insurgency's social base within a Russian war — several of its authors publishing explicit post-2022 clarifications [TBD-VERIFY: post-2022 statements by Kudelia and others repositioning the 2014 literature]. Account 3 bifurcated: its empirically careful wing (Charap among others) acknowledged the predictive failure while defending the interactive-spiral history; its polemical wing extended the provocation frame to 2022 itself, at increasing distance from the documentary record.

The duty of honesty cuts both ways, and the corpus states it: 2022 is evidence about 2014, but it is not a licence to rewrite 2014 teleologically. The 2014 operation was constrained, deniable, and partially improvised; the decision-making of 2014 should not be retrofitted with 2022's openly annexationist programme, any more than 2014's deniability should be allowed to launder 2022's candour. The historiographical rebalancing the evidence supports is specific: the war of 2014 was, on the accumulated forensic, archival, and judicial record, a Russian state operation conducted through and upon a real Ukrainian regional crisis — an operation whose author spent eight years denying it and then, in justifying its sequel, finally claimed it.

7. The Three Accounts in Dialogue: Convergences, Divergences, and Political Work

7.1 Where the accounts converge

Three convergences are larger than the polemics admit. First, on the February–March substrate. All three accounts agree that the post-Maidan transition produced a genuine legitimacy crisis in the southeast: Account 1 concedes the substrate while denying its sufficiency; Account 2 makes it central; Account 3 treats it as the local expression of the structural contest. No serious version of any account holds that the Donbas of March 2014 was politically quiescent, or that the language-law vote and the Article 111 irregularity were Russian inventions. Second, on the August intervention. No account disputes that Russian regular forces entered Ukraine in August 2014 and decided the campaign; Ilovaisk is in every account's record. Divergence concerns characterisation — culminating act of an aggression (Account 1), escalation rescuing a client insurgency (Account 2's framing), or the enforcement move of a coercive-diplomacy strategy (Account 3) — not occurrence. Third, negatively but importantly, on Minsk's fiction. All three accounts, for different reasons, agree that the Minsk architecture misdescribed the war it regulated: Account 1 because it cast the aggressor as mediator; Account 2 because its "special status" clauses addressed the local dimension only as Russia's instrument; Account 3 because it embedded a political settlement (federalisation-by-another-name) that Kyiv could not implement and the West would not enforce. The shared diagnosis that Minsk codified an untruth is one reason its 2022 collapse surprised none of the three traditions (UA-C-04).

7.2 Where they genuinely divide

The divisions, stripped to essentials, are three. The counterfactual division: would armed conflict have occurred in the Donbas absent Russian cadres, weapons, and command? Account 1 answers no, with the Kharkiv/Odesa/Dnipropetrovsk comparison and Girkin's testimony as evidence; Account 2's careful versions answer "not in this form, probably not at all as war — but the political rupture was real and would have produced a serious crisis regardless"; Account 3 declines the question's premise, holding that Russian action was itself structurally produced and so cannot be counterfactually deleted without also deleting its causes. The weighting division: among the documented causal elements — Kremlin decision, security-services entrepreneurship, local cleavage, Western policy environment — the accounts assign systematically different loads, and no factual discovery can fully arbitrate a weighting dispute; this is the irreducibly interpretive core of the controversy. The normative division: the accounts disagree about what kind of act the war was — international aggression (with its legal entailments of state responsibility, reparation, and the inapplicability of "both sides" framing); a militarised civil rupture (with its entailments for reintegration and amnesty politics); or a system effect (with its entailment that responsibility distributes across the system's architects). These are not merely emphases; they generate incompatible policy and justice conclusions, which is why the origins question remained live through every phase of the war it began.

7.3 The political work each account performs

In Moscow, the civil-war and provocation framings were operational assets before they were interpretations. The civil-war framing underwrote eight years of non-belligerency fiction — sustaining deniability at Minsk, in sanctions litigation, and before domestic audiences for whom the Donbas was charity, not war. The provocation framing supplied the external-audience track: NATO expansion as the deep cause, deployed from the 2007 Munich speech through the December 2021 treaty ultimatums to the 2022 war messaging, effective precisely because it incorporated genuine Western voices (Kennan's warning, Mearsheimer's lectures) as authorities. The two framings were logically in tension — a civil war needs no provocation theory; a provoked response is not someone else's civil war — and were run simultaneously anyway, a tension characteristic of information operations, for which internal consistency matters less than audience segmentation.

In Kyiv, the aggression account was constitutive of wartime statehood. It grounded the legal architecture (occupied-territories law; the 2018 reframing of the ATO as defence against armed aggression; the refusal to negotiate with the DNR/LNR as principals); it organised memory (the war dated from 2014, making 2022 an escalation, not an outbreak — a dating with material consequences for veterans' status, casualty accounting, and the international narrative); and it disciplined politics (framings that conceded a civil-conflict element were policed, sometimes to the detriment of honest engagement with the Donbas's actual 2014 sociology and with the eventual reintegration problem — a cost Ukrainian scholars and officials have themselves periodically acknowledged [TBD-VERIFY: instances — the debates around the 2021 transitional-justice draft legislation]).

In Western capitals, the account chosen tracked, and partly determined, policy. The 2014–2016 sanctions architecture was built on Account 1's factual core (the EU's sectoral sanctions were tied to Minsk implementation, i.e., to Russian conduct); but pre-2022 Western diplomacy — the Normandy format's both-sides choreography, the Franco-German pressure on Kyiv toward the Steinmeier formula — operated on a working blend of Accounts 2 and 3, treating the war as a conflict to be balanced rather than an aggression to be reversed. After 2022, Account 1 became the declared basis of Western policy; Account 3 migrated to the policy-critique opposition, the restraint school, and significant parts of global-South opinion, where it remains the dominant frame for the war as a whole. The accounts, in short, have never been merely academic: each has been, at some point between 2014 and 2026, the operating assumption of a government.

7.4 The corpus position

The corpus's adjudication discipline for J-block documents is to declare no winner while refusing false symmetry, and the origins question shows why both halves matter. The three accounts are not equally evidenced: Account 1 rests on a forensic-judicial-archival record that the other two do not possess and have largely been forced to accommodate; Account 2's scholarly core survives as a necessary theory of mechanism within it; Account 3 survives as a structural interpretation whose empirical anchors in 2014 are thin and whose predictive record is poor. But they are also not answering the same question: one is a finding about agency and operation; one about social conditions; one about systemic context. A reader who holds all three in view — Russian operation, real cleavage, real structural antagonism — is closer to the historiography's 2026 state than a reader who has merely picked the best-evidenced one and discarded the rest. The dishonest versions of each account are the ones that use their own level of analysis to deny the others' findings: the aggression account that erases the Donbas's 2014 sociology; the civil-war account that erases the Glazyev tapes and the 53rd Brigade; the realist account that erases the Ukrainians. This document is built so that none of those erasures survives contact with its record.

8. Conclusion

The question of how the Donbas war began is the hinge on which the entire 2014–2026 Russo-Ukrainian record turns, and it has been contested with a persistence that ordinary historical questions do not attract, because every answer carried operational consequences: for Minsk's design, for sanctions law, for the 2022 invasion's justification, and for the terms on which any eventual settlement will be narrated. This document has applied the three-account discipline to it: the factual spine of February–September 2014 established with its operational detail delegated to UA-C-01 and UA-C-02; each account presented at full strength with its provenance, evidence, claims, internal tensions, and burdens; the post-2014 evidentiary waves — forensic, archival, judicial, and the brutal natural experiment of 2022 — tracked for what they did to each account's standing.

The honest summary of that tracking is asymmetric, and the asymmetry is the finding. The Russian-aggression account has been progressively confirmed in its core claims by every class of evidence that has emerged since 2014 — by the participants' admissions, by the open-source forensic corpus, by the leaked Kremlin paperwork, by two international courts, and by the aggressor's own 2022 conduct and stated reasoning — while requiring qualification in its texture: the operation was adhocratic and opportunistic rather than scripted, and it worked through a real Ukrainian regional crisis whose reality the account's public versions have sometimes been reluctant to grant. The civil-conflict account has been narrowed but not nullified: its propaganda form is dead as description, killed by the record; its scholarly form survives as what it always claimed to be — a theory of the insurgency's social base and of the war's local mechanism, indispensable for explaining variation, recruitment, and durability, and explicitly subordinate on the question of cause. The Western-provocation account has lost the empirical argument about 2014 and the predictive argument about 2022 while retaining its grip on a structural register that evidence cannot reach: as long as the counterfactual of a neutral Ukraine cannot be run, the account will persist, and its genuine contributions — the unresolved contradiction in Western policy, the documented warning record — will remain part of any complete history.

Declaring no winner, under this discipline, therefore means something precise rather than evasive. It means the corpus does not collapse three different kinds of question — who acted, on what material, in what system — into one ranking. It does not mean the evidence is balanced, and this document has said plainly where it points: the war that began at Sloviansk on 12 April 2014 was started, sustained, rescued, and ultimately claimed by the Russian state. The standard this document sets for itself is the J-block test in its hardest form: that a reader formed in Moscow's tradition would find the provocation and civil-war cases stated better here than their own propaganda states them; that a reader formed in Kyiv's tradition would find nothing conceded that the record does not concede; and that a reader formed in neither would close the file knowing both what is contested and what, by 2026, honestly no longer is.

Primary Sources Consulted

Wilson, Andrew, Ukraine Crisis: What It Means for the West (Yale University Press, 2014).
Sakwa, Richard, Frontline Ukraine: Crisis in the Borderlands (I.B. Tauris, 2015).
Mearsheimer, John J., "Why the Ukraine Crisis Is the West's Fault: The Liberal Delusions That Provoked Putin," Foreign Affairs 93:5 (September/October 2014).
Kudelia, Serhiy, "Domestic Sources of the Donbas Insurgency," PONARS Eurasia Policy Memo No. 351 (September 2014); with the responding exchange by Andreas Umland and Taras Kuzio (2014–2016).
Zhukov, Yuri M., "Trading Hard Hats for Combat Helmets: The Economics of Rebellion in Eastern Ukraine," Journal of Comparative Economics 44:1 (2016).
Giuliano, Elise, "Who Supported Separatism in Donbas? Ethnicity and Popular Opinion at the Start of the Ukraine Crisis," Post-Soviet Affairs 34:2–3 (2018).
Hauter, Jakob (ed.), Civil War? Interstate War? Hybrid War? Dimensions and Interpretations of the Donbas Conflict in 2014–2020 (ibidem/Columbia University Press, 2021).
Mitrokhin, Nikolay, "Infiltration, Instruction, Invasion: Russia's War in the Donbass," Journal of Soviet and Post-Soviet Politics and Society 1:1 (2015).
Galeotti, Mark, Armies of Russia's War in Ukraine (Osprey, 2019); and "Hybrid War or Gibridnaya Voina? Getting Russia's Non-Linear Military Challenge Right" (2016).
Kuzio, Taras, Putin's War Against Ukraine: Revolution, Nationalism, and Crime (2017).
D'Anieri, Paul, Ukraine and Russia: From Civilized Divorce to Uncivil War (Cambridge University Press, 2019; 2nd ed. 2023).
Plokhy, Serhii, The Russo-Ukrainian War (Allen Lane, 2023).
Charap, Samuel, and Timothy J. Colton, Everyone Loses: The Ukraine Crisis and the Ruinous Contest for Post-Soviet Eurasia (IISS/Routledge, 2017).
Matveeva, Anna, Through Times of Trouble: Conflict in Southeastern Ukraine Explained from Within (Lexington, 2018).
Toal, Gerard, Near Abroad: Putin, the West, and the Contest over Ukraine and the Caucasus (Oxford University Press, 2017); with O'Loughlin, John, and Toal, survey research inside the DNR/LNR (2016–2019).
Sasse, Gwendolyn, ZOiS survey research on the Donbas (2016–2019); and Russia's War Against Ukraine (Polity, 2023).
Kuromiya, Hiroaki, Freedom and Terror in the Donbas: A Ukrainian-Russian Borderland, 1870s–1990s (Cambridge University Press, 1998).
Dutch Safety Board, Crash of Malaysia Airlines Flight MH17 (Final Report, 13 October 2015); Joint Investigation Team findings (28 September 2016; 24 May 2018; 19 June 2019); District Court of The Hague, MH17 judgment (17 November 2022); JIT Closing Statement (2023).
European Court of Human Rights, Ukraine and the Netherlands v. Russia (apps. 8019/16, 43800/14, 28525/20 and 11055/22): Grand Chamber admissibility decision (30 November 2022 [CORRECTED — Tier-2 block audit, 2026-08-30: prior text read "25 January 2023"; corrected to match the Tier-2-audited date]) and merits judgment (9 July 2025) [TBD-VERIFY: citation details].
International Criminal Court, Office of the Prosecutor, Report on Preliminary Examination Activities (2016–2020), Ukraine sections.
Czuperski, Maksymilian, et al., Hiding in Plain Sight: Putin's War in Ukraine (Atlantic Council, May 2015); Bellingcat MH17 and cross-border-attack reports (2014–2018); Conflict Intelligence Team publications.
The "Glazyev tapes" (intercepts released by the Prosecutor General of Ukraine, 2016) and the "Surkov leaks" (Ukrainian Cyber Alliance, 2016–2017), with Digital Forensic Research Lab authentication analyses [TBD-VERIFY: provenance documentation].
Girkin (Strelkov), Igor, interview with Zavtra (November 2014) and subsequent interviews and statements (2014–2023).
Kyiv International Institute of Sociology (KIIS), survey of the southeastern oblasts (April 2014, for Dzerkalo Tyzhnia); Pew Research Center, Ukraine surveys (April–May 2014).
OHCHR, Reports on the Human Rights Situation in Ukraine (2014–2022), including casualty accounting and the 2 May 2014 Odesa documentation; Council of Europe International Advisory Panel, Odesa report (2015).
Putin, Vladimir, Address on Crimea (18 March 2014); Direct Line (17 April 2014); "On the Historical Unity of Russians and Ukrainians" (12 July 2021); Address of 21 February 2022. Kremlin transcripts.
Burns, William J., cable "Nyet Means Nyet: Russia's NATO Enlargement Redlines" (Embassy Moscow, February 2008, via WikiLeaks); and The Back Channel (2019).

UA-A-01: Orange Revolution (November 2004 – January 2005)
UA-A-04: 2008 NATO Bucharest Summit and the Membership Action Plan
UA-B-01: Yanukovych Presidency (2010–2014)
UA-B-03: Euromaidan / Revolution of Dignity (November 2013 – February 2014)
UA-B-04: The Heavenly Hundred — February 2014 Maidan Killings
UA-C-01: 2014 Crimea Annexation (Ukrainian Perspective)
UA-C-02: Donbas War Onset and the LPR/DPR (April 2014 – 2022)
UA-C-04: Minsk-1 and Minsk-2 Agreements
UA-D-04: Wartime Mobilisation, Conscription Law, and Manpower Architecture (2022–2026)
UA-D-05: 2025–2026 Ceasefire Negotiations, Minerals Deal, and Peace Architecture
UA-F-01: Ukraine–NATO Trajectory (2008–2024)
UA-K-01: 1994 Budapest Memorandum Nuclear Disarmament Decision
UA-N-01: Ukraine in International Perceptions (1991–2026)
RU-K-01: The February 2022 Full-Scale Invasion Decision
RU-M-01: The Regime's Ideas — Sovereign Democracy, Russian World, and the Civilisational Turn
UA-B-05: back-reference added by symmetry sweep
UA-E-01: back-reference added by symmetry sweep
UA-E-03: back-reference added by symmetry sweep
UA-D-03: The 2021 Russian Force Buildup and the Pre-War Diplomatic Sequence (March 2021 – 24 February 2022) — back-reference added by symmetry sweep