Kabadayi, M. Erdem; Erünal, Efe
With the UrbanOccupationsOETR, a European Research Council-funded research project hosted at Koç University 2016-2022, we wanted to highlight the importance of rural economic dynamics to explain differences in long-term regional economic development in the late Ottoman Empire. We provide an Excel dataset on the crop-specific agricultural mix and land area of an Ottoman region, Bursa, in the 1840s. This dataset is the result of a new geosampling methodology we devised, representing a key development in the agricultural and overall economic history of Southeast Europe and the Middle East.
The 1840s serve as a good period to choose for base years mainly due to three main factors to sample economic data on a regional scale. First, due to Tanzimat reforms (planned and only partially accomplished transformation of the Ottoman central administration in the mid-nineteenth century), the 1840s marked a watershed of bureaucratical information gathering. Especially, the temettuat registers were created as a by-product to realize a drastic change in tax collection. With at least in its first iteration, the unsuccessful abolishment of tax-farming by the Tanzimat decree in 1839, the Ottoman central administration aimed to transform the existing indirect and communal taxation with direct and individual modalities. To accomplish this goal, the administration had to survey the tax base, which was in disguise due to centuries-long tax farming practices. The temettuat registers were conducted in the core regions of the empire with the main exception of the imperial capital, Istanbul. Second, the 1840s correspond to the last period before the beginning of drastic territorial losses, primarily in Southeast Europe, which triggered in size and frequency unprecedented waves of emigration and immigration between the core territories of the empire both in Southeast Europe as well as in Anatolia, which continued until the official demise or the implosion of the empire. Third and lastly, the 1840s serves as a very suitable point to assess the dynamics of pre-industrial and ancienne regime agricultural dynamics due to the lack of modern means of mechanization, irrigation, and fertilization combined with extremely rudimentary transport facilities.
The temettuat surveys are invaluable resources for they provide agricultural asset- / crop-type specific agricultural mix information with cultivation area per household. However, extracting their detailed information requires a team and years. To overcome this, we developed a sampling strategy that selected five locations per subdistrict using the Analytical Hierarchy Process (AHP), considering factors of agricultural suitability (85% weight), connectivity to historical roads (within a 500-meter to the closest road or to the Danube, 15% weight, justified by its impact on suitability), and subdistrict population size (chosen villages must represent at least 5% of the subdistrict’s total population).
Our geosampling methodology of the 1840s tax registers (temettuat) is based on contemporary Ottoman population registers. With this geosampling method, we aim to estimate the regional (district (sancak) and subdistrict (kaza)) level total area of cultivation and shares of the agricultural mix for key products. We are using two mid-nineteenth-century datasets: Ottoman tax (TMT) (temettuat) surveys for agricultural asset / crop type and cultivation area and the population (nüfus) (NFS) registers for population-based sampling. Connectivity is based on a detailed and provenly accurate 1940s German military map of Turkey, Deutsche Heereskarte (DHK). The agricultural suitability raster is an amalgamation of the Land Capability Classification (LCC) encapsulating the variables of soil quality and quantity and the Digital Elevation Model (DEM) based on Shuttle Radar Topography Mission with 30-meter-resolution and comprising elevation, slope, and ruggedness data.
In the end, a geosampling initiative was undertaken across six regions in Southeast Europe and Anatolia, namely Ankara, Bursa, Plovdiv, Ruse, Manisa, and Edirne, covering a total of 277 locations with 66,201 households. Our project team entered the economic data from those records into a Microsoft Access database. We employed a specially crafted data entry template to systematically organize the tax survey data into multiple categories.
After geosampling locations, our objective extends to deriving estimates for the total cultivated area within each subdistrict and the district of the regions. To achieve this goal, it is imperative that the data undergoes coding the cultivation areas into a standardized and comparable land-use scheme. We adopted the Corine Land Cover (CLC) nomenclature from the European Union’s Earth Observation Programme (Copernicus), established in 1985 and regularly updated. Our study followed the revised guidelines issued by the European Environment Agency on 10.05.2019. Despite its primary design for contemporary land cover analysis, CLC nomenclature proved well-suited for accurately representing the agricultural tax data and the historical context of the 1845 Ottoman tax surveys.
In our analysis, we coded micro-level cultivated land entries associated with individual households, using CLC’s highest detail level. Successfully, every cultivated land entry was coded into the third level of detail in CLC, encompassing sub-categories such as 2.1 – “Arable land”, 2.2 – “Permanent crops”, 2.3 – “Pastures”, and 2.4 – “Heterogeneous agricultural areas”—all falling under the overarching category of 2 – Agricultural areas. Additionally, we coded entries related to 3.1 – “Forest” and 3.2 – “Shrub and/or herbaceous vegetation associations”, falling under the primary category of 3 – “Forest and seminatural areas.”
Finally, cultivation area expressed in Ottoman measurement units like dönüm (1/9,2 of a hectare) are converted into hectares to ensure consistency and ease of spatiotemporal comparison.
We provide the geosample data of the Bursa region, positioned in Western Anatolia, renowned for its historical and economic significance, large and cosmopolitan population, and diverse geophysical characteristics. This data covers all the geosampled households, the individuals residing in them, and their CLC-coded agricultural assets / crops with quantity / cultivation area.
The Bursa region comprised 591 geolocated settlements in 12 subdistricts in 1840. Notably, the city of Bursa, serving as the major urban center and regional capital, was intentionally left out of the sample. Additionally, the subdistrict of Pazarköy, with its 14 settlements, was excluded. Despite being initially part of the Bursa region in population registers, it became attached to the northern neighboring Kocaili district in 1845. Consequently, out of the 576 remaining settlements, we geosampled 55 populated places from 11 subdistricts, covering 3547 households, representing 12% of the total households in the region, totaling 30,518. The variables of the tax surveys of the geosampled locations were read, extracted, and entered into the customized Microsoft Access database.
In the Bursa region, there are a total of 13,344 entries for agricultural assets coded with CLC across all 55 sampled populated areas. Our dataset includes all these entries and covers 3,325 households (out of the total of 3,547 sampled households) that owned these assets. Top of FormBottom of FormThis allows for a comprehensive analysis of the agricultural mix and land area at a detailed level.
The tax survey data was transcribed in Turkish using modern Turkish spelling and punctuation to keep the nuances of the original source. That said, because the original register information is largely presented in a standardized fashion and grouped under detailed variables, the data can easily be translated into other languages and coded into specific coding schemes.
The categories and descriptions of the variables of the geosample dataset for the Bursa region are as follows:
Category | Variable | Description |
GeoCode | “GeoCode” | UniqueID belonging to a specific geosampled location |
Location | “Longitude” & “Latitude” | Geographical coordinates used to specify the precise location of a geosampled location on the Earth’s surface |
Geographic unit of entry | “Region” & “SubDistrict” & “Location” | Geographic unit of entry, including region (district/sancak); subdistrict (kaza); and geosampled location as they appear in the population registers |
Unique key/ID | “HouseID” | Unique and consecutive ID belonging to a specific household, automatically generated by Microsoft Access |
Register specifics | “RegisterNo” | Archival code of the population register whose data is being entered |
“Household” | Number of the household (specified by the registers as Menzil, Persian word for house), as appears in the register | |
Unique key/ID | “HouseID” | Unique ID belonging to a specific household, automatically generated by Microsoft Access, which links the individuals to households and connects the “tblHouse” and “tblIndividual” |
“IndivID” | Unique ID belonging to a specific individual, automatically generated by Microsoft Access | |
Ethno-religious identity | “EthnoReligiousIdentity” | Ethnoreligious identity of the individual as given in the register |
Individual interrelationships | “RelationtoHouseholdHead” | Shows the individual’s relationship or lack thereof to the first recorded male within the household. If he was the “_firstRegisteredMale,” then he is recorded as such. The type of relationship to the “_firstRegisteredMale”, such as son (oğlu), grandson (hafidi/torunu), tenant (kiracı), or slave (gulamı/kölesi), or lack of it (in the case of a new individual moved to the household) are specified. |
“FamilyName” | The individual’s family name, in the rare occasion before the official introduction of family names (not to be confused with title) | |
“Title1” | Title(s) appearing before an individual’s name | |
“NameI” | The individual’s name(s) | |
“Title2” | Title(s) appearing after an individual’s name | |
“Conjunction” | Arabic patronymic ibn, bin, and veled (equivalent to the “-son” suffix in English, meaning the son of “NameII”) and Turkish patronymic oğlu (the opposite, meaning the son of “NameI”, appearing rarely), that links “NameI” and “NameII” | |
“TitleII1″ | Title(s) appearing before the father’s name (if the conjunction is “oğlu”, then the title of the son (the individual himself) | |
“NameII” | The father’s name(s) (if the conjunction is “oğlu”, then the name of the son (the individual himself) | |
“TitleII2” | Title(s) appearing after the father’s name (if the conjunction is “oğlu”, then the title of the son (the individual himself) | |
Occupation | “OccupationI” | Standardized version of the occupation [e.g.,barber] |
“OccupationIStatus” | Status of the occupation [e.g., apprentice]. If an occupation did not have a status, then “statüsüz” (no status) was entered. | |
“OccupationII” & “OccupationIIStatus” | Versions of the descriptors above if the individual is employed in multiple occupations | |
Agricultural Asset / Crop | “AgrID” | UniqueID belonging to a specific agricultural asset / crop belonging to an individual |
“Agriculture” | Type of the agricultural asset / crop | |
“CLCAgricultureCode” | CLC-code of the agricultural asset / crop | |
“CategoryUnit” | Is applicable when the quantity of an agricultural asset or crop is specified using specific terms (“aded”, “res”, “eşcar”, “sak”), and when the area of an agricultural asset or crop is described in vague terms (“bab”, “kıta”) | |
“Unit” | Quantity of the “CategoryUnit” | |
“CategoryArea” | The land area type of the agricultural asset / crop in Ottoman measurement units, like “Dönüm” | |
“Area” | Quantity of the “CategoryArea” | |
“Total area of cultivated land (CategoryArea) in hectares (if applicable)” | Quantity of the “CategoryArea” converted into hectares |
If you would like to use the dataset, please use the credentials specified below:
M. Erdem Kabadayı and Efe Erünal, “UrbanOccupatiponsOETR_1840s_Ottoman_Bursa_District_TMT_geosample_dataset” (Zenodo, July 1, 2024), https://doi.org/10.5281/zenodo.12610779.