Shusha pogrom (1905) §

Background

Shusha was not simply another provincial town. By 1905 it was the symbolic capital of highland Karabakh, with an Armenian quarter, a Muslim Azerbaijani quarter, churches, mosques, schools, printing houses and a dense elite culture (Swietochowski, 1985). Its geography amplified danger: neighbourhoods were close enough for sudden attack but socially separate enough for rumour to travel faster than verification. The 1897 Russian Imperial census had recorded the town's population at roughly 25,000, with Armenians a small majority (~56%) and Muslim Azerbaijanis a substantial minority (~44%) (Caucasus Viceroyalty statistical office, 1906). The two quarters were physically contiguous along the central ridge of the town but socially distinct, with shared marketplaces, separate religious institutions, and a thin shared elite that had already been eroding through the 1890s.

The February 1905 Baku massacres had already spread fear across the South Caucasus (Swietochowski, 1985). In Karabakh, Armenian villagers and Muslim Azerbaijani communities watched events in Baku as evidence that imperial order could fail. Armed detachments associated with the ARF became more visible on the Armenian side, while Muslim notable networks and local armed men organised defence in the Azerbaijani quarters (Swietochowski, 1985). de Waal treats Shusha as one of the moments when Karabakh's older pattern of coexistence was replaced by armed communal geography (Thomas de Waal, 2003). sourced opinion

The structural backdrop was an Imperial state that had visibly failed to manage inter-communal tension. The 1903 Tsarist attempt to seize Armenian Church property under the policies of Caucasus Governor-General Prince Grigory Golitsyn (the church-property confiscation) had broken the modus vivendi between the autocracy and the Armenian community; the February 1905 Baku massacres had exposed the inability of the Imperial police to prevent inter-ethnic violence in the Russian South Caucasus's largest city. In May 1905 the Imperial centre appointed Illarion Vorontsov-Dashkov as Caucasus Viceroy with a specific brief to manage what Russian official documents called the "Armenian-Tatar war" (Illarion Ivanovich Vorontsov-Dashkov, 1905). By the time of the August Shusha violence, Vorontsov-Dashkov had been in post fewer than four months and his authority over local Imperial structures was still being established.

The 1905 Karabakh violence sat inside a regional cascade of inter-ethnic disturbances across the Russian South Caucasus that ran from February 1905 through the spring of 1906: Baku (February-August 1905, ~2,000 dead), Nakhichevan (May 1905), Yerevan (1905), Ganja (multiple incidents 1905-1906), and Tiflis (October 1905). Shusha was mid-scale within this cycle: smaller in toll than Baku but larger than the urban incidents in Yerevan or Ganja, and locally foundational in ways the smaller incidents were not.

The event

The August fighting in Shusha unfolded over roughly ten days, 16 to 25 August 1905, with later spasms into late August and early September. Contemporary Russian Imperial police reports, later Soviet-period historiography, and the 1906 Caucasian Calendar administrative summary converge on the following compound chronology, while disagreeing on attribution at every step:

• 16 August: Shooting in or near the Asar bazaar at the boundary between the Armenian and Muslim quarters. Armenian and Azerbaijani accounts dispute whether the first shots came from inside the Armenian quarter, from outside Muslim armed groups, or from a chance encounter that escalated.
• 17–18 August: Defence of the Megrutsian quarter (the Armenian intellectual centre, surrounding the Megrutsian printing press) by armed ARF-affiliated fedayeen. Burning of Armenian homes and businesses on the Armenian-quarter periphery; counter-attacks by Armenian armed elements into the Muslim quarter.
• 19–20 August: Russian Imperial troops arrived in greater numbers — the local garrison had been minimal — and partially restored order. Vorontsov-Dashkov dispatched additional regular forces from Yelisavetpol (Ganja) and reinforced the Imperial police presence (Illarion Ivanovich Vorontsov-Dashkov, 1905).
• 23–25 August: Secondary clashes; further burning of structures. Sustained inter-quarter sniper fire from positions on the ridge.
• Late August / early September: Sporadic attacks into early September; Imperial authorities declared partial martial law and arrested armed leaders on both sides for show, though prosecutions did not consistently follow.

Casualty figures from the fighting cluster around several hundred dead total, with the per-source breakdown reflecting the standpoint of each account:

Source Armenian dead Muslim dead Method Russian Imperial police reports (1905) ~200 ~100 Police investigation; partial; under Imperial classification Vorontsov-Dashkov correspondence ~200 ~100 Caucasus Viceroyalty reports to St Petersburg Armenian-side period press (ARF Droshak etc.) 300–400 unstated / not lower Partisan; named-victim lists from Armenian community sources Azerbaijani-side later accounts unstated 300–400 (or symmetric) Partisan; emphasises Muslim losses Świętochowski (1985) ~200 ~100 Treats Russian Imperial figures as most authoritative de Waal (2003) range 200–500 (combined) range as part of combined Broad range without attribution split

The material destruction was concentrated in the Armenian quarter. Specific Armenian-quarter institutions damaged or destroyed include the Megrutsian printing press (the Armenian intellectual centre of Karabakh, source of the eponymous quarter name), the Shusha realschule (technical school, with Armenian-leaning student body), and several churches, with the Ghazanchetsots cathedral damaged but not destroyed in 1905 (its full destruction came in 1920) (Caucasus Viceroyalty statistical office, 1906). In the Muslim quarter, named damage records are sparser in the Imperial archive, and Azerbaijani sources emphasise the burning of Muslim homes and shops in what they describe as Armenian counter-attacks deeper into the quarter than the 1906 Caucasian Calendar formally acknowledges contested.

Unlike the later Shusha massacre of 1920, the 1905 violence did not destroy the Armenian presence in the city (Richard G. Hovannisian, 1996). It did, however, weaken the idea of Shusha as a shared urban world. The city that had produced Armenian clerics and printers, Azerbaijani musicians and intellectuals, and bilingual elite interaction became a place whose streets were mapped by danger editorial.

The Russian Imperial state

The 1905 Shusha violence cannot be read without the Imperial state, and most of the later popular Armenian-Azerbaijani memory of the event removes the Imperial dimension entirely editorial. Vorontsov-Dashkov's Caucasus Viceroyalty, reactivated as a separate institution in May 1905 specifically to manage inter-communal violence, was the structural locus of state response. His correspondence with the Imperial centre at St Petersburg produced reports characterising the violence as bilateral with shared organisational responsibility — a framing that has structural consequences for both later historiographies, since both communities have inherited the "Russian neutrality" framing while contesting its empirical accuracy.

The Imperial response also relied on the longstanding strategy of playing communities against one another, which the 1903 Armenian Church property seizure under Prince Golitsyn had made into a structural feature of Russian-Armenian relations and which the Imperial centre had not formally rolled back even after Vorontsov-Dashkov's appointment. Vorontsov-Dashkov himself was relatively moderate and tried to manage rather than instrumentalise the inter-ethnic tension, but his administration was operating inside a longer Imperial pattern in which "Armenian" and "Tatar" (the period name for Azerbaijani Muslims) were instruments of policy as well as objects of it.

Aftermath

The 1905 Shusha violence shaped the military and political reflexes of both sides. For Armenians, Shusha became evidence that Karabakh required armed self-defence and that imperial officials could not be trusted to protect Christian quarters (Richard G. Hovannisian, 1996). The Karabakh ARF cell strengthened in the years immediately following, and Armenian fedayeen networks consolidated their presence across the highland Karabakh districts. For Azerbaijanis, Armenian armed organisation was folded into a narrative of Dashnak aggression and territorial ambition (Audrey L. Altstadt, 1992). Those competing readings would return in 1918–20 during the struggle between the Republic of Armenia and the Azerbaijan Democratic Republic, and again in 1988–92 when control of Shusha became militarily decisive.

The most direct organisational consequence on the Azerbaijani side was the founding of Difai at Ganja in 1906 by Ahmed bey Aghayev. Difai was an explicitly self-defensive organisation, conceived as a Muslim parallel to ARF armed structures, and oriented toward inter-communal mutual aid and armed protection of mixed-population districts. Banned by Russian Imperial authorities by 1909, Difai is significant less for its operational effect than as proto-organisational text of organised Azerbaijani national-defence politics — the form of armed civic organisation that would re-emerge during 1918-20 and again, in transmuted form, in 1988-92 (Ahmed bey Aghayev (Ağaoğlu) and the Difai founding committee, 1906). The Azerbaijani Marxist organisation Hummet (founded 1904, consolidated post-1905) developed in parallel, oriented to working-class organisation rather than self-defence but sharing the same period of Azerbaijani national consolidation. On the Armenian side the ARF's Karabakh field structure consolidated, with continuous fedayeen cells operating in the district through the First World War.

The longer-term contradiction is that both memories are rooted in real fear while each erases the fear of the other side editorial. Armenian accounts often underplay how intimidating ARF armed structures could appear to neighbouring Muslim communities; Azerbaijani accounts often underplay the vulnerability of Armenian civilians in mixed towns where state protection failed. The atlas's Shusha 1905 dispute page surfaces the three principal historiographical readings — Armenian, Azerbaijani, and academic / Russian-Imperial-archival — alongside their respective primary sources, rather than collapsing them into a single narrative.

Khosrov-bek Sultanov — the young Karabakh Muslim notable who would, fifteen years later, become the Azerbaijan Democratic Republic's governor-general of Karabakh and the perpetrator of the 1920 destruction of the Armenian quarter — was active in his community's notable networks during this period. The line from his 1905 milieu to 1920 perpetrator role is one of the structural illustrations of how the early-century inter-communal violence consolidated armed political identities across the next decade and a half (Thomas de Waal, 2003, Ch. 4).

Historiography

The academic literature on the 1905 Shusha violence sits across three principal readings, none of which has fully integrated the others, and the atlas's editorial position is that each reading captures something the other two minimise editorial. The principal works are Świętochowski's Russian Azerbaijan, 1905-1920 (1985), de Waal's Black Garden (2003) historical chapters, Altstadt's The Azerbaijani Turks (1992), and Bournoutian's Karabakh-focused historiography. Each works against a different set of primary sources and arrives at a recognisably distinct frame (Swietochowski, 1995). sourced opinion

Świętochowski (1985) treats the 1905-06 violence in the South Caucasus as a sequence of inter-communal disturbances mediated by Russian Imperial state failure rather than as a sustained inter-ethnic war. His Shusha account is built on Russian Imperial police reports cross-checked against the focused 1905 monograph and treats the Vorontsov-Dashkov correspondence's ~200 Armenian / ~100 Muslim dead enumeration as the most evidentiarily authoritative figure (Swietochowski, 1995, Ch. 2). sourced opinion Świętochowski's structural argument is that the Imperial state's earlier reliance on Russification (under Golitsyn) and on the 1903 church-property seizure created the conditions in which the 1905 revolution-period state weakness translated into communal violence — the Imperial framework, not the underlying inter-ethnic relations, was the proximate cause.

de Waal (2003) places the 1905 violence inside a longer arc of Karabakh-specific inter-communal trauma running from 1905 through 1918-20 and 1988-94, and treats the city of Shusha as a particular site where each cycle's violence produced the structural conditions for the next (Thomas de Waal, 2003, Ch. 4). sourced opinion de Waal's account broadly endorses the Russian Imperial casualty range (200-300 combined dead, with Armenians the larger share) and reads the violence as the moment when Karabakh's older inter-communal coexistence was decisively replaced by an armed-communal urban geography. The 2020 recapture of Shusha by Azerbaijani forces and the subsequent treatment of the city as a state-cultural project closes the arc de Waal had identified as open (Thomas de Waal, 2003). editorial

Altstadt's Azerbaijani Turks (1992) places more emphasis on the Imperial state's deliberate cultivation of communal asymmetry — the policy of arming local Muslim notables while restricting Armenian armed organisation under the church-property crisis — and on Difai as a legitimate self-defence response rather than as a separatist Azerbaijani-national consolidation. Her ~100 Muslim / ~200 Armenian breakdown is consistent with Świętochowski but the framing is closer to the Azerbaijani-national reading (Audrey L. Altstadt, 1992). sourced opinion Bournoutian, writing from the Armenian-academic position, emphasises the pre-1905 demographic continuity of the Armenian community of Karabakh and the operational responsibility of state-aligned Muslim irregulars in the destruction of Armenian quarters and rural settlements — a frame closer to the Armenian-state reading though distinguished from it by the use of Russian-archival primary sources rather than ARF-press secondary material (George A. Bournoutian (trans.), 1994). sourced opinion

The contested questions in the academic literature: the precise casualty breakdown (200 Armenian / 100 Muslim is the modal Imperial figure; ARF-press counts run 300-400 Armenian; Azerbaijani-side counts run symmetric or higher Muslim); the attribution of first-shots responsibility at the Asar bazaar on 16 August (Armenian and Azerbaijani accounts disagree); the degree to which the Russian Imperial garrison's late intervention represents structural Imperial failure or active Imperial complicity; the role of the ARF fedayeen presence in escalating the violence beyond the initial flashpoint (Swietochowski, 1995). editorial The atlas's full multi-perspective treatment of these disputes, with sourced critique on each viewpoint, is at the Shusha 1905 dispute.

Soviet-era memory

The Soviet seven decades treated the 1905 Shusha violence in characteristic fashion: as a "Tsarist tragedy" produced by Imperial divide-and-rule, with both Armenian and Azerbaijani working classes presented as objects of manipulation rather than agents of violence. This framing solved the political problem of incorporating the events into the SSR national histories without inflaming inter-republic relations, but it also stripped the events of analytical specificity. The Armenian SSR's Hayrenagitatsyun (homeland-studies) textbooks and the Azerbaijani SSR's parallel curricula nonetheless diverged on attribution, with each republic's textbooks emphasising the victimisation of its own community and treating the other side's casualties as either much smaller or as collateral to Imperial provocation. Specific Armenian and Azerbaijani writers in the 1960s and 1970s — including Sero Khanzadyan on the Armenian side and several Azerbaijani historians associated with the post-Aliyev (Heydar) academic apparatus — reopened the question with material drawn from their respective republic archives, but without breaking the underlying compromise framing (Audrey L. Altstadt, 1992). sourced opinion

The post-1988 reopening of the Karabakh conflict also reopened the historiographical struggle over 1905. Armenian publications since 1988 have emphasised the 1905 violence as the founding moment of organised anti-Armenian aggression in highland Karabakh; Azerbaijani publications have emphasised it as the moment when ARF aggression broke an older pattern of inter-communal coexistence. Neither framing engages directly with the Russian Imperial archival material, and neither has been fundamentally revised in the post-1991 academic environment of either country.

Memory and politics

Shusha's 1905 violence matters because it began the conversion of a cultural city into a territorial symbol (Svante E. Cornell, 2001). Later Armenian memory would mourn the loss of Shusha as a centre of Karabakh Armenian life; Azerbaijani memory would elevate Shusha as a cradle of national music and elite culture. Both claims are historically grounded. Their collision is precisely why Shusha has such unusual force in the atlas: it is a shared place remembered as if it could only belong to one community editorial.

The structural legacy of 1905 in Shusha — the mapping of urban space by danger, the consolidation of armed organisations on both sides, and the discrediting of Imperial institutions as guarantors of inter-communal peace — set the conditions for the 1920 destruction of the Armenian quarter under Sultanov's governorship, for the operationalisation of Shusha as the artillery platform that would shell Stepanakert in 1991-92, and for the symbolic weight of the 2020 Azerbaijani recapture of the town under Ilham Aliyev (Thomas de Waal, 2003). editorial

The post-2020 Azerbaijani state project around Shusha — declaring it the "cultural capital" of Azerbaijan in 2021, the reconstruction programme funded through the Heydar Aliyev Foundation, the May 2022 Shusha Declaration with Turkey, and the symbolic placement of state cultural events in the city since 2020 — closes the arc the 1905 violence had opened: the Armenian-cultural Shusha of the 19th century gave way to the segregated Shusha of 1905-20, then to the Armenian-only Soviet-era Stepanakert-shadow Shusha of 1923-92, then to the Armenian-cultural Shusha of 1992-2020, then to today's Azerbaijani-cultural Shusha, with each transition mediated by violent population removal (Broers, 2019). editorial Each community now mourns a Shusha that, at the moment of its loss, was theirs as the only legitimate claimant; the structural finding of the 1905 historiography is that this was true for both communities sequentially across a century (Thomas de Waal, 2003). editorial

The full multi-perspective treatment of the 1905 historiographical disputes is at the Shusha 1905 dispute in this atlas. editorial