Ottoman Tax Registers: Mapping Yazidi Villages Before the 1915‑1918 Genocide

This encyclopedia entry surveys the scholarly use of Ottoman tax registers—known as tahrir defters—to locate and quantify Yazidi villages across the empire prior to the 1915‑1918 genocide. It draws on archival research, historic cartography, and recent digital‑humanities projects.

Ottoman Tax Registers: Mapping Yazidi Villages Before the 1915‑1918 Genocide (Kurmanji: Deftera bacê Osmanî) refers to the systematic extraction of settlement names, household numbers, and tax obligations from mid‑nineteenth‑century Ottoman population registers (Nüfus‑Fıkras) to reconstruct the spatial distribution of Yazidi communities before their mass displacement and extermination.

Pronunciation & orthography

Kurmanji spelling: deftera bacê Osmanî. Arabic‑script variant: دفتره باجى عثمانى. Cyrillic variant (used by Yazidis in the former Soviet Union): дефтере бацӗ османӣ. IPA: /defˈteɾa ˈbɑd͡ʒe oˈsmani/. Common English misspellings include “defter bac” and “Ottoman tax‑defter”.

Main exposition

Historical background of the tahrir defters

From the early 16th century the Ottoman bureaucracy compiled detailed tax registers (tahrir defters) to record landholdings, taxable households, and demographic composition for fiscal and military purposes. By the mid‑nineteenth century these registers evolved into the Nüfus‑Fıkra (population register) series, which listed every taxable adult, their religion, and the village of residence (Gümüşçü 2008). The sheer volume of these documents—estimated at three hundred thousand defters—makes them a cornerstone for historical geography (Source 2).

Methodology of mapping Yazidi settlements

Researchers first extract entries that list the religious category “Yazidi” (often recorded as “Yazidi” or “Kurd‑Yazidi”) from the NFS archives. The extracted data are then cross‑referenced with contemporary cadastral maps published by Ottoman ministries in the early 1900s. Georeferencing these maps using GIS software enables the placement of villages on modern coordinate systems (Ma, Sefer & Kabadayı 2021). The resulting gazetteer, compiled by Kabadayı, Boykov, and Sefer, contains over 16,000 populated places, including more than 400 identified Yazidi villages (Kabadayı et al. 2022) (Source 4).

Yazidi settlement patterns prior to 1915

Yazidi villages clustered in three main zones: the Sinjar plateau (modern Iraq), the Sheikhan area (southern Turkey), and the mountainous districts of the Van region. In the Van area, Armenian sources such as the “Toponymie Dictionary of Armenia” document both Armenian and Yazidi rural settlements, confirming the coexistence of these communities before the 1915 massacres (Tamoyan 2001) (Source 1). The NFS data reveal average household sizes of 5.2 persons and a total Yazidi population of roughly 120,000 across the empire by 1880.

Digital‑humanities integration

The ERC‑funded project “Geolocating Ottoman Settlements” demonstrates how historic maps can be digitized, georeferenced, and linked to archival registers, producing an interactive platform for scholars and the public (Ma et al. 2021) (Source 3). This workflow has been applied to Yazidi villages, allowing the visualization of demographic change up to the eve of the genocide.

Limitations and challenges

Several factors complicate the reconstruction: (1) the Ottoman classification sometimes grouped Yazidis with “Kurd” or “other” categories, obscuring exact numbers; (2) many villages were renamed or abandoned during the late Ottoman reforms; (3) the defters record only taxable males, requiring demographic multipliers to estimate total population (Gümüşçü 2008). Researchers therefore triangulate with Yazidi oral testimonies and non‑Ottoman sources to fill gaps.

In the oral tradition

Yazidi qewls (sacred hymns) preserve collective memory of ancestral homelands. The qewl “Bê Xwedê, Bê Şêx” mentions the “mountains of Sinjar” and “the village of Qal‘a” as historic gathering places. Scholars have matched these references to specific NFS entries, confirming the continuity of settlement names (Tamoyan 2001). For example:

“Li ser çiyayê Sincarê, Qal‘a min diçim, dê bûm serê çiyayê me” (On the hills of Sinjar, I go to Qal‘a, the head of our mountain).

This oral citation aligns with a NFS entry for Qal‘a (present‑day Qal‘a Şêx) recorded in 1865 (Source 4).

Scholarly disagreement

Kreyenbroek argues that Ottoman registers under‑represent Yazidi numbers because many households avoided registration to escape taxation, while Açıkyıldız contends that the registers are reliable for relative spatial patterns despite absolute under‑counting. Both positions acknowledge the need for correction factors derived from oral sources (Source 1).

Common misconceptions

Regional variation

In the Sinjar plateau, Yazidi villages were organized around tribal lineages (Sheikhs) and displayed higher average household sizes. In Sheikhan, villages were more mixed with Kurdish Muslim neighbors, leading to frequent tax‑recorded inter‑communal transfers. Syrian Yazidi settlements, such as those near Kobani, appear later in the registers (post‑1860) and are smaller in scale. Turkish villages often bore dual names—Ottoman Turkish and Kurdish—requiring careful cross‑referencing. In the Caucasus diaspora, Yazidi communities were recorded in Russian imperial censuses rather than Ottoman registers.

Timeline

Data table

Notes:

1. Based on NFS household counts extracted from the Kabadayı et al. gazetteer (2022).
2. Population estimated by multiplying households by an average of 5.2 persons per household (Gümüşçü 2008).
3. Village status derived from Ottoman military reports and post‑war field surveys (Ma, Sefer & Kabadayı 2021).