"statisquo": british use of statistics in the iraqi kurdish question (1919–1932)

7/31/2019 "Statisquo": British Use of Statistics in the Iraqi Kurdish Question (19191932)

1/67

Fuat Dundar

Statisquo

British Use of Statistics in

the Iraqi Kurdish Question

(19191932)

Brandeis UniversityCrown Center for Middle East StudiesCrown Paper 7 July 2012


2/67

Crown PapersEditorNaghmeh Sohrabi

Consulting EditorRobert L. Cohen

Production ManagerBenjamin Rostoker

Editorial BoardAbbas Milani

Stanford UniversityMarcus NolandPeterson Institute for International EconomicsWilliam B. QuandtUniversity of VirginiaPhilip RobinsOxford UniversityYezid SayighKings College London

Dror ZeeviBen Gurion University

About the Crown Paper SeriesThe Crown Papers are double-blind peer-reviewed monographs coveringa wide range of scholarship on the Middle East, including works of

history, economics, politics, and anthropology. The views expressed inthese papers are those of the author exclusively, and do not reflect theofficial positions or policies of the Crown Center for Middle East Studiesor Brandeis University.

Copyright 2012 Crown Center for Middle East Studies, BrandeisUniversity. All rights reserved.


3/67


4/67

Figure 1. Mosul and Its Districts in the 1930s

Source: This map is drawn based on the ethnographic map, which was most likely prepared in 1931. See Records

of Iraq, V.7, pp. 59697.


5/67

1

Introduction

-200 K v j v v

e Kurds belieethat determining the exact size of the Kurdish population would haerepercussions for issues such as dening the exact borders of the Kurdistanegional Goernment (KG) ascertaining the proper portion to beallotted to them from the national budget and maybe een establishing theproper quota of Kurds in a future Baghdad goernment to ensure that theywould not once again be underrepresented in a future Parliament

To substantiate their political claims, the urds hae attempted to carryout a plebiscite and census, two of the three steps designated by theTransitional Administratie Law TAL, Article 58, in March 4 andthe Iraqi ermanent Constitution Article 14. ased on the results of therequested new plebiscite and census, the urds want to determine the nalstatus of the disputed regions, including hanikin, Sinar, and irkuk, andannex them to the R. Although arious organizations and neighborhood

goernmentsand, especially, the Arabs and Turkomans liing in thesedisputed territoriesoppose the urdish call for a plebiscite and ethniccensus, the urds still insist on one, causing this to become acasus belliinIraq today.3


6/67

2

While awaiting a plebiscite and census, which hae been postponed seeraltimes, the urds hae used historical data to calculate the exact size andpercentage of the urdish population in order to support their nationalistclaims. Aside from demanding a greater proportion of the national

budget, the R, which has claimed that urds are underrepresentedas ciil serants, has also called for equitable representation in the ciilserices, especially in the oilrich city of irkuk.4 Until the nal statusof the disputed territories, especially that of irkuk, is determined by acensus or plebiscite, the urds, like other Iraqi communities, will continueto claim that they are statistically dominant in their own regions. Andto substantiate their claims, all sides will continue to rely on their ownpreferred data. 5

Since statistical data occupy a ery important place in the politicaldiscourse of urdish as well as other Iraqi politicians, and the ethnic censusthreatens Iraqi unity and potentially inites ethnic conict, it is necessaryto focus on the historical data, which were produced during the creation ofIraq.

Problems and Aims

This aper considers the statistical data produced during the ritish periodin Iraq 19193 and analyzes its role in the political context of its time.It examines how political actors, especially the ritish Empire, used thesedata as a scientic tool. It gies attention to the importance attachedto numbers during the creation of Iraq and, most importantly, during

the creation of the urdish autonomous districts. And it demonstratesthat statistics and the census are the most important battleelds in Iraqscontemporary politicsthe result of the ritish manner of exploitingstatistical data during the ritish mandate period in Iraq.

The aper will examine the sets of ritish, Turkish, Iraqi, and League ofNations data, collected between 1919 and 193, that represent the mostimportant statistical data in Iraqi history, other than the mother tongue

data from the 1957 census. In particular, I will discuss how the populationdata on urds collected by the ritish Empire were used to protect thepolitical and military interests of the ritish as well as to maintain thestatus quo. In this way, statistics, ordinarily considered to be a scientic and


7/67

3

obectie tool, become a subectie tool in the serice of political disputes.Moreoer, statistics, which are supposed to be stable and static in character,in fact, become unstable and changeable. The aper will demonstrate thatpopulation statistics are inseparable from their political context, and from

the political aims of whoeer produced them. The ritish/Iraqi statisticaldata were produced to sole a particular political dispute namely, todetermine the political boundaries of Mosul and the fate of its people.When the political context changed, the statistical data were likewise, andsimultaneously, transformed.

The census/statistics literature generally well demonstrates arbitrary changesof ethnic categorization by the political power conducting the census and

producing the statistics. This aper will explore as well equally arbitrarychanges in the political meaning assigned to these ethnic categorizations,and to their numerical size. Obiously, behind these arbitrary changes,both numerical and classicatory, is a common goal, which is to producehegemony and maintain the status quo. The maor aim of this aperis to demonstrate how the ritish used statistical data to support anundemocratic military occupation in Iraq.

While seeral international, bilateral, and domestic negotiations on theIraqi and urdish issues were taking place, statistical data were used notonly to demonstrate ethnic and political facts but also for purposes ofstatistical reasoning.6 The aper shows how the urdish actors7 weresystematically excluded, especially by the ritish/Iraqi powers, during theentire process of preparing and interpreting population statistics. Thisexclusion, I think, would become one of the main historical causes for

urdish nationalisms reactionary and emotional character in the post1958period in Iraqand, ultimately, for the urds insistence on conducting anethnic census in Iraq.

Statistics and numbers became part of the language of political debate withrespect to Iraq beginning with the peace treaties of 1919 and continuingwith the Language Law of 193. This aper will demonstrate the change

in this language and in the statistical data through four chronologicalperiods, which correspond to the thematic and practical eolution ofthe use of statistics and statistical reasoning between 1919 and 193.8The rst chapter of the aper examines urdish nationalism and the


8/67

4

ritish plebiscites in Mosul during the postWorld War I period, whenthe principle of selfdetermination dominated international politics.The urdish problem emerged for the rst time in the internationalarena during the peace conferences. It was considered to be essentially a

population problem: The urds were recognized as a numerous populationthat is, to constitute a maority in a gien region, but were seen asunciilized and incapable of selfgoernment. The main purpose of thischapter is to examine the 1919 and 191 ritish plebiscites in Iraq, and toshow how the ritish authorities interpreted the results to suit their ownpurposes.

The second chapter of the aper explores how statistics became the key

diplomatic issue for both Turkey and reat ritain during the LausanneConference. Following the 191 ritish plebiscites, the Turks reected theannexation of Mosul to Iraq and instead called for a true plebiscite basedon the principle of selfdetermination. This chapter focuses on ritainand Turkeys use of statistical data relating to Mosuls population in theabsence of a true plebiscite to determine the wishes of Mosuls populationbetween 19 and 194. It argues that measuring identity was actually a

way of measuring loyaltyspecically, the political attitude of Mosulspopulation in Iraq toward Turkey and the ritish. Consequently, althoughboth the ritish and the Turks recognized that the urds constituted amaority in Mosul, they interpreted the magnitude and proportion ofthe urdish population dierently, so as to adance their distinctandcontradictorypolitical interests.The third chapter examines the decision of the Frontier Commission of

the League of Nations, which was created to nd a scientic solutionto the dispute between the ritish and Turkish powers oer Mosul. TheLeague of Nations examined the Mosul question in three stages. First, itestablished a commission to inestigate the facts of the disputed area.Second, it appointed a Council committee to attempt mediation. Finally,when that didnt succeed, it xed a proisional frontier line slightly southof the northern boundary of Mosul, dening the military status quo. This

chapter focuses on the commissions inquiry and examines its interpretationof ethnographic and populationstatistics data, in the process pointing uphow the political character of the commission and the personal backgroundof the commissioners aected the character of the inquiry and thecommissions ensuing decisions.


9/67

5

The fourth and nal chapter examines the Languages Law that was createdto nd a solution to the urdish claim in Iraq. The chapter shows howthe Iraqiritish ocers used another set of population data regarding theurds to determine in which districts urdish would be designated the

ocial languageand how, at the League of Nations, the ritish appliedscientic reasoning to their statistical data in order to dismiss urdishcomplaints of underrepresentation in the ciil sericesand, moregenerally, to presere the status quo in Iraq.There is substantial releant literature on the question of Mosul. Whileresearchers recognize the underlying important interrelationship betweendemography and the urdish problem, none of them, with the sole

exception of Shields,9 has analyzed the statistical data and explicated itscentral role. This aper seeks to do so. It utilizes primarily ritish andalso Turkish archial sources that focus on the statistical data producedduring the creation of Iraq, as well as documents that deal with the genesisof the urdish problem. ut because Turkish Ministry of Foreign Aairsdocuments are not aailable to researchers, Turkish ocial documentsare accorded a minor place. The examined documents are mainly from

Ottoman 1914, ritish 1919, 191, 193, 1931, 1947, and ritishsupported Iraqi state sources 194, 193, 1931. These dataalloriginally tabulated for political purposesnot only played an importanthistorical role, but Iraqi communities to this today use them to supporttheir political claims.

British and Ottoman Statistics and the Categorization of

Identities

An important part of the literature on nationalism has focused on theinterrelationship between nationalist ideology and the use of censuses andstatistics. Most academics agree that the adent of modernity has radicallyaltered the conditions of identity formation, and that censuses and statisticshae played an important role in this process. The impact of censuses and

statistics in statebuilding and identity formation has been consideredcrucial, especially in the eld of colonial studies. To this end, the role ofstatistics in establishing and maintaining colonial power has been thesubect of academic studies, most of which hae concentrated on the ritishexperience in India.1


10/67

6

The ritish Indian literature on censuses and statistics in the coloniesunderlines how the ritish brought the intellectual baggage of eighteenthcentury Europe, notably its interest in political arithmetic and statistics,and applied it as an instrument of goernance in their colonies.11 As

Leitan describes, this political arithmetic had been associated with theritish goernment until the mideighteenth century.1 Eighteenthcenturyritain and Europe generally enthusiastically employed statistics, mainlyowing to their pure scientic character and their capacity to be used tocontrol society. The rise of population sciences in Europe in this periodbrought together ideas about population, the deelopment of quantitatietechniques, and the use of demographic data by the state. European states,including ritain, embraced the idea that a growing population was proof

positie of the prospects of a gien state. Subsequently, the state, utilizingstatisticians, deeloped more sophisticated quantitatie techniques toimproe their knowledge of population size and trends. As part of itsduties, the state assumed the responsibility to measure and quantify thesocial bodypart of a larger reimagining and reshaping of goernment inthe new era of liberal and democratic politics.13

It would thus hae been tempting for ritish bureaucrats to imagine thatsound numerical data would make it easier to embark on proects of socialcontrol or reform in the colonies. According to Arun Appadurai, howeer,there were important dierences between the census conducted in theritish metropolis and that in its colonies. The rationale for the ritishcensus at home was oerwhelmingly territorial and occupational, ratherthan ethnic or racial. Meanwhile, in India, the encounter with a highlydierentiated, religiously other set of groups must hae been built on the

metropolitan concern with occupation, class, and religion, all of which werea prominent part of the ritish census. In the colonies, the census tookon a dierent role because the entirepopulation was seen as dierent inproblematic ways.14

oth the quantitatie and qualitatie aspects of the census as a colonialundertaking hae been discussed in the academic literature; but according

to many scholars, the qualifcation of identities by the ritish colonialstatisticians had a more important impact than their quantication.As Eileen Janes Yeo argues, the ritish census in India asked for castealiation despite the diculties of standardizing a classication across


11/67

7

India, and ranked the castes in order of social precedence. Nationalistscomplained that this actually intensied the rialry between castes andconstituted a clear attempt to diide and rule.15 ernard Cohn suggeststhat the ritish colonial census played an important part in South Asian

identity formationa hypothesis strengthened by the fact that it was aninstitution that eery adult male had to encounter during his lifetime.16As enedict Anderson points out, Europeans constructed the categoriesused on their census forms from their own frames of reference and theirown experiencesand in colonial societies, these categories often didnot consider or reect those used by the colonized themseles.17 ut incontrast to Yeo and others, Anderson argues that the real innoation of thecensustakers was not the construction of ethnicracial classifcations, but

rather . . . their systematic quantifcation.18

In comparison with the ritishIndian case, one can ask: Was Iraq aspecial case with regard to the role of ritish colonial statistics? I wouldargue, no, but yes: No in the classifcatory sense, but yes in the numericalsense. No, because Iraq was not an unknown territory for the ritish:They possessed considerable knowledge about the Iraqi population before

their occupation and based their classication of the Iraqi people mainlyon an existing Ottoman taxonomy. Yes, because contrary to the situationin India, the quantication of ritish statistics concerning Iraq was muchmore important than their classication.

The Ottoman censuses, held four times in 1831, 1844, 188193, and1967, were the most important tools of the empires modernizingand centralizing eorts. Throughout this period, population data became

the main instrument of goernmental policy. Howeer, a census suchas those conducted in nineteenthcentury Europe was neer conductedin the Ottoman Empire. None of the Ottoman censuses were conductedduring a short period of time, nor were they conducted in all areas of thecountry. Some censuses took more than ten years to complete; and thepopulations of some areas were neer counted. In addition, census takersneer counted all members of households. Instead, in most cases, the

leaderswhich usually included mullahs, priests, and muhtarsof thesmallest administratie units, were summoned, and the local populationwas determined on the basis of the information they proided In the lastdays of the Empire, representaties of neighborhoods and illages were


12/67

8

also included. The 188193 census, during which women were countedfor the rst time, using the same surey method used throughout theEmpire, can be considered the rst modern Ottoman census. During thiscensus, identity papers were distributed to subects as required by the 1881

establishment of the opulation Registry Administration.19

Aside from the obious military and tax purposes, territorial reasonsmotiated the Ottoman censuses. The censuses were conducted duringa period when the Ottoman Empire was losing territory as a result ofits wars with the reat owers especially with the Russian Empire andalso because of nationalist uprisings. Diplomatic negotiations followingthe wars prooked discussions about the ethnic and religious statistical

composition of the disputed territories. The national moementsespecially those of the nonMuslim communities, such as the Serbs, reeks,and Armenianssubmitted population data supporting their claims. Noneof these moements found the Ottoman censuses credible, insisting insteadthat they did not represent their communitys true size or proportion. Andthere was immense pressure from the reat owers calling for equitableparticipation by preiously excluded nonMuslim communities in the

Ottoman state apparatus.

This territorial factor was eident in the deelopment of the Ottomancensus. The rst Ottoman census, in 1831, classied the Ottomanpopulation according to religious aliation and only dierentiatedMuslims from nonMuslims Reaya. In contrast, by the last Ottomancensus 1967, the number of Christian categories had increased tothirteen, whereas Muslims were neer classied along sectarian i.e.,

Shiite, Sunni or ethnic urd, Arab lines. This increasingly popularclassication was not, as some nationalists took it to be, a diideandruleOttoman power policy, but rather reected the reat owers politicalstruggles regarding control oer each group of Christians within theOttoman Empire. Eery new Christian denomination meant ocialrecognition by the Ottoman power of that community as amilletreligiouscommunity.1 The four Ottoman censuses reected a process of growing

domination by the European statistical methodology and mentality.

As we will see in Chapter Two, while the ethnic and religious taxonomiesof the ritish, Ottoman, and, later Turkish census statistics were identical,


13/67

9

the numerical aspects of the statistical tabulations presented during theMosul disputes diered greatly. In turn, an analysis of the ritish statisticsrequires an inestigation of the earlier Ottoman classication system.While examining that system, one can ask, how did the Turkish state

distinguish between urds and Turks, as the Ottoman goernment did notconsider ethnicity a separate category for Muslims? One might speculatethat either the Ottoman Empire secretly registered Muslims ethnicity orthe Turkish state may hae inented this data. I hae argued in my preiousworks that the Ottoman goernment registered the Muslim populationaccording to its own ethnic and sectarian identities: urds, Shiites, andso on.3 Though this data was registered, howeer, it was neer published.Rather, in the printed Ottoman ocial data, Muslims were classied as a

single group. The rst Ottoman/Turkish statistical data, which classiedMuslims as urds, Turks, or Arabs, were included in the Turkish statisticaltable presented to the Lausanne Conference.4


14/67

10

British Plebiscites in Iraq (191921)

Following World War I, and especially between 1919 and 191, the

Wilsonian principle of selfdetermination played a central role on theinternational political scene.5 It was not until 19, howeerwhenthe Turkish Republic, the new power that emerged from the ruins of theOttoman Empire, demanded selfdetermination for Mosulthat reatritain presented itself as a defender of this principle.6 Until that time,Londons position had been that the frontiers of the future . . . Stateshould, as far as possible, be racial [in todays terminology, ethnic] ratherthan economic or geographical.7

While this political climate sparked a great interest in statistics amongthe Allied powers, it triggered an een stronger interest among nationalistgroups, who wanted to utilize statistical data to claim political rights fortheir historical lands and delineate state boundaries.8 Reecting this globalpolitical context, the ritish Empire conducted two plebiscites in the threeOttoman proinces it occupied, in order to create a single new country,

Iraq. The statistical results of these plebiscites were neer reealed, howeer,nor were their conclusions pertaining to ethnic distribution, making theirresults questionable, both then and now.

The Allied powers used the selfdetermination principle to shorten thewar, especially after the olsheiks took power.9 In particular, they aimedto utilize the principle to fan the dissatisfaction of minorities within theenemy Habsburg and Ottoman Empires and incite them to rebel.3 Thedeclaration of resident Wilson was thus a turning point in the eolutionof the international political system as well as in the deelopment ofnationalist moements. The principle, which was reconrmed by theAngloFrench Declaration, to encourage and assist the establishment ofindigenous oernments,31 was the key determinant during the eaceConferences of Versailles, Neuilly, Saint ermainenLaye, Trianon, andSers.3 Stateless peoples from all oer the world had high hopes for these

conferences, which appeared to present unprecedented opportunitiesto pursue the goal of selfdetermination. They took their struggle to theinternational stage as their representaties set out for the conferences,whether inited or not, in order to stake their claims in the new world


15/67

11

order. They composed and circulated a ood of declarations, petitions, andmemoranda directed at the world leaders assembled for the conferences inan attempt to shape public opinion throughout the world. Many of thepetitioners drew on Wilsons rhetoric of selfdetermination and the equality

of nations to formulate their demands and ustify their aspirations.33

The fate of the Ottoman minorities was discussed at the Sers Conference.During the conference, almost all the Ottoman minorities, inspired mainlyby the twelfth of Wilsons socalled Fourteen oints, demanded selfdetermination.34 The new world order called for negotiations of nationalboundaries grounded in ciil and scientic discussions based on populationdata proided by the inoled parties. In their memoranda, the nationalist

groups presented statistical tables coering not only their own ethnic groupbut also other Ottoman ethnic groups. As one might anticipate, howeer,each minority submitted data according to which it was always the largestethnic group in a gien area.

In spite of the conferences, howeer, none of the minorities nationalistclaims were realized: Most of the minorities data were exaggerated, and

none of their claims could be implemented owing to the oerlap amongthem. More importantly, with the exception of the United States, thereat owers were focused on securing their own interests, rather thanimplementing the selfdetermination principle. As a result, none of theOttoman minorities were satised with the Sers Treatyone thathad tragic consequences for some minorities, including the Armenians,Assyrians, and reeks.

The spirit of selfdetermination also impacted the urds. Throughout thisperiod, the urds repeatedly petitioned the ritishFrench Commissioners,and during the Sers Conference they presented their wishes for anindependent urdistan. Seeral dierent urdish organizations, in thememoranda they presented at the conference, called for a unied and abigger urdistan, which would include irkuk. Underlying the presence ofa urdish oice at Sers was their unequiocal opposition to the possibility

of their inclusion in a soereign Armenian state. The urds reected theArmenian claim that Armenians were the maority in the contested areas. 35


16/67

12

While presenting their case at Sers, the urds distributed memorandaand maps.36 Although the maps showed the geographical distribution ofthe urdish population, none of them included a detailed statistical tabledocumenting the number of urds and their proportion in relation to

other ethnic groups. Sureyya edirkhan, in the name of the Comit delIndependence Kurde, established in Cairo, presented the only urdishestimate of the urdish population: In a telegram protesting the Armenianclaim, edirkhan reported that there were e million urds in theOttoman Empire. This number was exaggerated, howeer, een if it hadincluded the Iranian urds. edirkhans number was based not on his owninquiry, but rather on the estimate of the prourdish ritish InspectorEdward Noel, known as the urdish T. E. Lawrence.37 Meanwhile,

others, including reat ritain, France, and the Armenians, proided morespecic documentation regarding the size and geographical distribution ofthe urdish population. The Armenian statistics minimized the urdishpopulation numbers and in addition diided the urdish population alongsectarian lines Aleisunni, urdsZaza, nomadssedentary, etc..

From 1918, the ritish prepared seeral statistical tables and maps

highlighting the numbers and geographical distribution of the urds. Theirmain expert was Edward Noel, who was a proponent of a bigger urdistan.He was asked to determine the exact number of the urdish populationwho lied outside of the occupied territories and particularly outside ofMosul; their proportion isis other groups, including the Armenians,Turks, and Arabs; and their geographical distribution.38

The ritish collected their own population data on the urds, since they

wanted to specify the ethnographic borders of the urdish population.Still, there was no unied ritish stance on the matter, as dierent ritishactors held dierent positions on the urds. Noel, using a map that isnow a logomap of urdish nationalists, proposed a reater urdistan,which would include e million urds liing in the Ottoman andersian Empires.39 Arnold J. Toynbee, on the other hand, considered aunited urdistan an impossibility, not for demographic reasons but

rather since it would hae included nonurdish populations of superiorciilizations [Armenians], and who are not capable of running such astate themseles[the urds].4


17/67

13

The Sers Treaty gae preference to proisions regarding the borders ofa Christian Armenia, thereby creating a larger Armenian state that was,ironically, predominantly Muslim and urdish. According to the treaty,the remaining urdish areas lying east of the Euphrates and north of

the frontier of Turkey with Syria and Mesopotamia could then becomean independent urdistan, if they met two stipulations: 1 The maority ofthe population desires independence; and The Council of the Leagueof Nations considers that these peoples are capable of such independence.Furthermore, Article 64 proided that the urds of Mosul roince couldoin this urdish state should they wish to do so.41 The Ankara goernmentled by Mustafa emal, which claimed to reect urdish wishes, reectedthe Sers Treaty, howeer, and so none of these proposals came to pass.

The fate of Mosuls urds was thus left to subsequent discussions betweenthe ritish and the Turks.

On October 3, 1918, one week after the Mudros Armistice, the ritisharmy occupied the Ottoman proince of Mosul. This marked the ritishadministrations rst direct contact with the urds.4 During the rst yearof their occupation, the ritish had an unclear policy regarding the fate of

Mosul and the urds.43

While they wanted to retain Mosul in order bothto safeguard the way to India and to keep its oil resources under theircontrol, the ritish were also under pressure to support Wilsons newinternational system of selfdetermination.

To complicate matters, dierences of opinion existed between the ritishexperts in the Foreign Oce and those in the India Oce. These dieringopinions crystallized at the Cairo conference held on March 13,

191, where the ritish policy in the Middle East was debated. rior tothe conference, Colonel Arnold Wilson, the Acting Ciil Commissionerfor Mesopotamia, had strongly emphasized the strategic importance ofMosul for the ritish Empire, and had recommended attaching Mosulto aghdad.44 During the conference, Hubert Young called for theestablishment of a separate state [urdistan] . . . to function as a strategicbuer against any future emalist threat to Iraq. Opposing this, the

ritish High Commissioner for Mesopotamia, ercy Cox, also called forthe annexation of Mosul to aghdad.45


18/67

14

Each camp defended its policy on the basis of the strategic, economic,and political aspects of the Mosul question. The central issue was theethnographic character of Mosul. The interesting point here is that whilethe dierent sides used the same ritish state data, each ealuated the data

dierently. ercy Cox argued that because of the character of its ethnicdistribution, Mosul should not be separated from aghdad. WinstonChurchill, with the help of Edward Noel, pointed out that there was noserious diculty in drawing the boundary between Southern urdistanand Arab Mesopotamiathat is, between aghdad and asra. Churchill,belieing that the creation of a separate urdish entity was the best way ofdefending against the emalist threat, included all the areas Cox claimedto be nonurdishsuch as irkuk, ifri, and Arbilin his urdish

buer scheme. Cox argued in opposition that an independent urdistanwould leae Arab Mesopotamia with strategically indefensible frontiers.46

The internal ritish discussions ended at the Cairo Conference, where itwas decided that political conditions necessitated that a Sharian ruler beselected to goern Iraq and that the most suitable candidate would be EmirFaisal. ecause of the climate of selfdetermination, howeer, the ritish

goernment fully realized that it could not nominate Feisal, but that hemust be chosen by the people of Mesopotamia.47 Against this backdrop,the two plebiscites organized in Mesopotamia/Iraq during 1919 and 191were used to legitimize ritish imperial dominance in the Middle East.According to the ritish, the plebiscites reealed that the maority of theformer Ottoman proinces of asra, aghdad, and Mosul wished to unifyas one state under the kingship of Faisal. Actually, instead of reectingthe wishes of the Iraqi population, these plebiscites were conducted to

strengthen the ritish status quo in Iraq.

While the ritish conducted two plebiscites in Iraq, they neer respondedto the repeated urdish requests for true plebiscites, especially those fromthe urdish chieftain Shaikh Mahmud arzani, who had led seeraluprisings against ritish authorities.48 The ritish instead used the 1919and 191 plebiscites to ustify their military hegemony and proide

substitutes for the other plebiscites proposed by either urds or Turks;but in reality, the ritish plebiscites were manipulated, and their resultsmisrepresented. For one thing, they were not actual plebiscites at all, butrather plebiscitelike inquiriesand no indiidual oting, either open or


19/67

15

secret, occurred.49 Second, ritish military ocers sering as part of themilitary occupation themseles carried out the plebiscites, enabling them toeasily manipulate the results. Finally, the results of the plebiscites were notpublished in statistical tabular form to show the proportion of otes for and

against and their ethnic distribution.

The 1919 plebiscite, which unied the three Ottoman proinces underan Iraqi state, was far from an impartial or truthful one. It was conductedbecause the ritish War Cabinet wanted to announce the result of thisplebiscite to the world as the unbiased pronouncement of the populationof Mesopotamia. In addition, Colonel Arnold Wilson, a ferent defenderof the annexation of Mosul to aghdad, orchestrated the plebiscite5and

it was ritish military ocers who selected the notables and communityrepresentaties to be polled. The rst of the plebiscites three questionswas Do they faour a single Arab state under ritish tutelage stretchingfrom the Northern boundary of the Mosul wilayat [proince] to theersian ulf?51 According to the ritish, in response to this question,those consulted were unanimous in saying that they wished to belong toa State consisting of three vilayets [proinces],5 and the whole country

was agreed that, whateer form of goernment might be set up in Iraq asto which there was a wide diergence of opinion, Mosul should not beseparated from the remainder of Iraq.53

The results of the 1919 plebiscite are highly questionable. As I haeemphasized aboe, the ritish plebiscites were unreliable, owing to themisrepresentation of their results. To hae been reliable, the plebisciteresults would hae required a statistical tabulation showing the distribution

of otes according to their location in one of the three proinces and theethnicity of the respondents. The ritish made use of their statisticaltabulations in the following period 196, when the situation andthe results suited their own interests see succeeding chapters. ut theyrefrained from doing so in the case of these two plebiscites in Iraq in 1919and 191. This aper will reexamine the statistical results of the 1919plebiscite so as to elucidate the statistical realities coneyed by that data.

While conducting the 1919 plebiscite, the ritish collected seentythreepetitions from twentyfour districts for a total of approximately 1,8signatures, mostly from notables, tribal chiefs, religious leaders, and


20/67

16

community representaties.54 Around percent of the signatures werefrom Mosul, corresponding to the proportion of Mosuls population isis the general Iraqi population according to ritish estimations in 1919.The petitions and signatures collected from Mosul were questionable,

howeer, because they did not reect the ethnic and geographicalcomposition of Mosuls population. While the ritish collected twelepetitions signed by 34 notables from Mosul roince, they did notcollect petitions from the entire proince; rather, they collected petitionsonly from three of Mosuls districts: irkuk, ifri, and Mosul CityCentre while excluding Arbil and Sulaymaniyah. urds and Yazidisaccounted for only two of the 57 signatures on these petitions, and bothof them opposed the annexation of Mosul to Iraq. Most of the signatures

belonged to Arabs and Christians who lied in Mosul City Centre and whosupported the annexation of Mosul to aghdad; around 5 signatureswere from people belonging to dierent Christian sects. In sum, Christians,while constituting only eight percent of Mosul according to ritish 191statistics, accounted for seentythree percent of the plebiscite results55whereas a plebiscite intended to ascertain a gien populations desirefor selfdetermination should surely hae reected its correct ethnic and

religious makeup.56

It can be argued that had these petitions reectedthe actual ethnic and religious distribution of Mosul, the results of thatplebiscite would hae been entirely dierentand, indeed, completelyreersed.57

In the 191 plebiscite, which brought about the accession of Emir Faisal,notables and important people from specic localities were gatheredtogether and inited to comment on the plebiscites question: whether they

were in faor of the Emir Faisals candidacy or not. The ritish reportedthat one million58 Iraqis declared their opinion, and [96 percent] otedin faor of the ing.59 ut the ritish hae neer published the detailedresults of this plebiscite. Until now, the otes of the Mosul populationhae remained in obscurity, but according to Turkish experts reports, athird or . . . half of Mosul, including the urdish maority, oted againstFaisal. In truth, as with the preceding plebiscite, the results of the 191

plebiscite had been determined before it een took place. Already at theCairo Conference, the 191 plebiscite was predetermined to choose Faisalas the Arab leader of Iraq.


21/67

17

The international political scene between 1919 and 191 was anextremely ironic period. While on the surface the Wilsonian principle ofselfdetermination dominated, in practice it had little eect, as the statesestablished in the Middle East following World War I were in reality not

based on this principle. Thus, all of the statistical and cartographic actiitiesconducted during the negotiation of peace treaties meant little; realpolitikand ritish imperial interests determined the borders in the Middle East.The ritish also extended preferential treatment to their war allies, likethe Sharian family. The results of the plebiscites in Iraq, the stated aimsof which were to solicit the wishes of Mosuls maority population, were,in reality, far from a reection of the populations true feelingsbut, asreported, they certainly reected the interests of the ritish Empire.


22/67

18

Determining the Fate of Mosul Using Statistics

(192224)

The military success of the Ankara goernment headed by Mustafa emaltriggered another round of discussions during the years 19 to 196reoling around the principle of selfdetermination and also broughtabout a renewed use of ethnic statistics. This new period started whenthe Ankara goernment oided the Sers treaty and forced the holdingof a new peace conference in Lausanne. During the Conference, Ankarainsisted that Mosuls boundaries be determined by a plebiscite. This new

political context had a denitie impact on both the ritish and Turkishstatistical approaches. Although the two sides disagreed oer the fate ofMosul, both were sympathetic to the urds and tried to gain their support,acknowledging in their statistical tabulation that the urds were indeedthe maority in Mosul roince. Turkey reected the results of ritishplebiscites in Iraq and demanded a new plebiscite, while continuing toclaim its rights to Mosul. From 19 until 196, the year that Turkeyrenounced her soereign rights to Mosul, there were seeral bilateral,multilateral, and international discussions held to determine the fate ofMosul, most notably the Lausanne Conference and the discussion at theLeague of Nations. During all of these discussions, the urdish populationissue was the central questionyet urdish representaties were notpresent at any of these meetings.

At the Lausanne Conference, Turkey claimed soereignty oer Mosul,

basing its arguments on military, historical, geographic, economic, political,and ethnographic data.6 Turkey demanded the uncontestable authorityof a plebiscite in the name of urds and Turks. The head of the Turkishcommission in the Lausanne Conference, Foreign Minister smet nn,argued that the Turkish state was the state of both the Turks and the urdsand claimed that Mosuls urds and Turks, who together constituted amaority in the proince, desired to belong to Turkey.nn subsequentlyrequested a new and true plebiscite for Mosul.

The ritish reected this Turkish request, howeer. Aside from claimingthat conducting a plebiscite in Mosul was impractical, the ritish arguedthat the Turks had no right to ask for a plebiscite, since they had [n]eer


23/67

19

encouraged the anquished to demand a plebiscitein any of the territoriesthey had conquered by force of arms.61 In addition, the ritish asserted,the urds hae neer asked for it. oor fellows, they do not know what [itmeans]. The ritish considered a plebiscite to be alid only for a society

with a high stage of education and ciilization. And, furthermore, aplebiscite is needed, they argued, only when we dont know the result.6

The Use of Ethnic Statistics to Determine the Wishes of Mixed

Populations

Once the ritish reected the Turkish demand for a plebiscite in Mosul,ethnic statistics became the most suitable method for determining thewishes of the Mosul population. ased on the assumption that each ethnicgroup would ote in accordance with their identity, both the ritish andthe Turks used statistical tables to counter the others arguments. Whiledoing so, both sides claimed that their statistical data were more accurateand less politicized than the others. In fact, it was the political meaninggien to these ethnic categories, rather than those categories themseles,

that constituted the real focus of discussion and disagreement.

The Turks claimed that their data were based on the 1914 Ottomanregisters. Later, howeer, they claimed that the data was based on the 196and 1916 census.63 Ironically, the statistical tables of the Ottoman ethnicand religious groups, which the Ottoman goernment presented at theSers Conference and which were based on the 1914 Ottoman registers,

lacked data from the proince of Mosul. The lack of data from Mosul wasmainly a result of the Ottoman administrations haing left Mosul withouttaking any state documents especially population registers, since theythought that they would recoer Mosul from the ritish.64 y contrast, theydestroyed or took the population registers of asra and aghdad when theylost those proinces in battle. Since they lacked the corresponding data onMosul, the Turks presented statistical data at the Lausanne Conference thatwere partly based on the preceding data that is, those collected from 1881

to 1893 and were partially inented.

The Turkish statistical table indicated the proportions between the ariouselements of the [Muslim] population in Mosul. According to the table,


24/67

20

the urds constituted 5 percent plus the Yazidi 3.6 percent of the totalMosul population; Turks accounted for 9 percent, and Arabs 8.6 percentsee Table 1. ased on these numbers, the Turks claimed that the totalnumber of Turks and urds, including Yezidis, constituted more than

fourfths of Mosuls population and was almost nine times larger than theArab population.65

While these data at rst seem to alidate the Turkish claim to Mosul, uponcloser examination it is apparent that the data exaggerated the size of theTurkish population in Mosul. For instance, while the Turks claimed thatthere were 3,96 Turks in Sulaymaniyah, according to the ritishIraqiocials, who conducted a prolonged and a careful search, there were only

two Turks and no Turkomans. Similarly, the Turks claimed that the twodistricts of Shaikan and Ashair Saba were entirely populated by Turks,whereas the ritishIraqi ocials claimed that they did not nd any Turksat all in those districts.66

Table 1: 1914 Turkish Data as Presented to the Lausanne Conference

Source: House of Commons arliamentary apers: Turkey No. 1 193, Lausanne Conference on Near Eastern

Aairs, 19193. Records of proceedings and draft terms of peace, London, His Maestys Stationery Oce,

193, p. 373.

Meanwhile, the ritish also claimed that they had carefully collectedpopulation data from Mosul by isiting approximately a thousand illagesand had counted the houses and consulted the ocial Turkish censusdocuments.67 After comparing all of the ritish data for Iraq from 1917,1919, and 191, howeer, it seems that their data were not as carefully

collected as they claimed. The rst ritish data on the urds were compiledfrom the 1917 estimation for Mesopotamia, at which point the ritisharmy had not yet conquered a single part of Mosul roince see Table .68The ritish prepared two population data sets: The rst, in 1919, addressed

Kurds Turks Arabs YazidisNon-

Muslims Total Nomads

Sulaymaniyah 62,830 32,960 7,210 0 0 103,000

Kirkuk 97,000 79,000 8,000 0 0 184,000

Mosul 104,000 35,000 28,000 18,000 31,000 216,000

Total 263,830 146,960 43,210 18,000 31,000 503,000 170,000


25/67

21

the religious and sectarian character of the population, while the second,in 191, was mainly concerned with the populations ethnic character. Inthe course of those two years, according to ritish calculations, the entirepopulation of Mosul had increased by eighty thousand people. The ritish

explained this seemingly unreasonable increase in population by regardingit as a direct result of the massie return of refugees and soldiers after WorldWar I see Tables 3 and 4.69

According to the ritish gures from 191, the urds were the maorityin Mosul, constituting 57.9 percent of the population; the percentage ofTurks was 8.4, and of Arabs, 3.6 See Table 4.7 Comparing these datawith those of Turks, the biggest dierence between the two concerned the

size and proportion of the Arabs in Mosul see Figure . According to theritish, who supported the annexation of Mosul to Arab aghdad, thenumber of Mosul Arabs was 6,5, amounting to 3.6 percent of Mosulspopulation. In contrast, the Turks, who opposed annexation, claimed thatthe number of Mosul Arabs was only 43,1, or 8.6 percent of Mosulspopulation.

Figure 2: Graphic Representation of 1914 Turkish Data

%

0

40

%

- , ,

, , , ,

, , , ,

, ,

, , , ,

Source: Data compiled by author


26/67

22

Table 2: 1917 British Data for Mesopotamia

Source: Secret Report entitled Turkey in Europe and Asia, dated October , 1917, Records of Iraq, V.1,

pp. 67586.

Table 3: 1919 British Religious Data

Source: Original title of table was opulation of the Vilayet of Mosul by Religions according to an Estimate

made in 191, House of Commons arliamentary apers, Turkey 1 193, pp .36566.

Table 4: 1921 British Racial Data

Source: Original title of table was opulation of the Vilayet of Mosul by Races according to an Estimate made in

191, House of Commons arliamentary apers, Turkey 1 193, pp. 36566.

Mesopo

tamia

Kurds Turks ArabsSyrian

ChristiansJews Persians

380,000 110,000 1,650,000 60,000 60,000 70,000

Armenians Yazidis Circassians Sabians Chabaks Miscellaneous Total

57,000 21,000 8,000 2,000 10,000 10,000 2,438,000

Sunni Shiite Jewish Christian Other Total

Arbil 96,100 0 4,800 4,100 1,000 106,000

Kirkuk 85,000 5,000 1,400 600 0 92,000

Mosul 244,713 17,180 7,635 50,670 30,180 350,378

Sulaymaniyah 153,900 0 1,000 100 0 155,000

Total 579,713 22,180 14,835 55,470 31,180 703,378

Kurds71 Turks Arabs Christians Jews Total

Arbil 77,000 15,000 5,100 4,100 4,800 106,000

Kirkuk 45,000 35,000 10,000 600 1,400 92,000

Mosul 179,820 14,895 170,663 57,425 9,665 432,468

Sulaymaniyah 152,900 1,000 0 100 1,000 155,000

Total 454,720 65,895 185,763 62,225 16,865 785,468


27/67

23

In light of these discrepant data sets, it is apparent that both the ritishand the Turks cynically exploited both the statistical data and the principleof selfdetermination. As has been emphasized aboe, both the ritish andthe Turks proceeded on the assumption that members of each ethnic group

would ote in accordance with their ethnic identityso the statisticaldata resulted from an imaginary plebiscite. y counting people based ontheir ethnicity, both sides argued that they had determined the Mosulpopulations choice, for Iraq or Turkey. The assumption regarding otingbehaior was itself problematic, and what made it more complicated werethe urds. ecause neither the ritish nor the Turks discussed the wishes ofthe Arabs or the Turkomans, the dispute was oer the wishes of the urds.While both delegations defended the selfdetermination principle in theory,

in practice they both simply supported their own selfinterest. The Turks,referring to the Arabs, asked how it was possible that an insignicantelement who consist [of] less than onefth7 and are a maority in thecapital of a proince [but] [make up] a tiny minority in the proince itself,should decide the fate of the whole proince. The ritish also posed thesame question regarding the Turkish population in Mosul, asking whyMosul should [be] gien back to the Turks, who constitute only one

twelfth of the whole [of Mosul].73

Ironically, the February 195 urdish rebellion in Turkey had an impacton these population statistics discussions. Een though the assumptionthat ethnicity directly determined political wishes and oting behaior didnot change, both sides shifted from emphasizing the ethnic character ofthe population to stressing sectarian diisions. The Turks, who were likelyupset by the urdish rebellion in Turkey, introduced a new demographic

consideration to the argument: In response to the Frontier Commissionsreport at the League of Nations in September 195, the Turks claimed thatwhile the Mosul maority was Sunni like the inhabitants in Turkey, themaority of Iraq referring to aghdad and asra was Shiite, as [were]the inhabitants of ersia.74 Accordingly, the Turks insisted that Mosul,because of its sectarian demographic character, should be attached toTurkey. The ritish reected this argument, stating that there was only a

slight maority of Shiites in the whole area of Iraq. aghdad was diidedequally between the two religious communities. In addition, the ritishpointed out that ing Faisal and the maority of the Iraqi goernment wereSunnis, and that there was no political dierence between the Shiites andthe Sunnites.75


28/67

24

The statistical discussions also extended to the origins of the ethnictaxonomy reected in the ritish and Turkish statistical tabulations.Although both sides taxonomies were identical, their denitions werenot. The Turks alleged that the urds were Turks, basing their claim

on historical, political, and racial arguments. In particular, quoting theEncyclopedia Britannica, they claimed that both urds and Turks were ofa common Turanian origin. Reecting this argument, the ritish claimedthat the urds were not of Turkish origin, but were rather Iranian. Whilethe ritish, drien by their political aims, were ery clear regarding urdishidentity, they were not so clear with respect to Turkish identity: Theysimply assumed that the Turkomans were dierent from the Turks. Withregard to Arab taxonomy, they claimed that there were no distinctions

among Arabs, since they all alike claim the independence of Arabs andshare the Arab ideals and the use of Arabic as their mother tongue.76

The central issue underlying these political discussions was the loyaltyof theurdsthat is, the political meaning of the statistical categories that theyhad tabulated. While the ritish backed their claim that the Mosul urdsdid not want to unite with Turkey by pointing out that the urds had been

continuously rebelling against the Ottoman and Turkish states, the Turkishside claimed that urds not only had the same origin as Turks, but alsoshared the same political inclinations.

As emphasized aboe, both the Turks and the ritish considered compilingethnic statistical data a suitable method for determining the wishes ofMosul people. To present this argument clearly as well as to show howstatistics were part of the discourse within the bureaucratic state apparatus,

I quote from ritish internal correspondence that was intended to preparea response to the Turkish goernment. The ritish argued that whereasthey did not know the wishes of the Iraqi people before the 1919 and 191plebiscites, they already knew the wishes of the Mosul people. This ritishknowledge was manifestly a result of their reductionist approach: Allmembers of an ethnic group, they belieed, hae similar political attitudes.They further argued that they already knew the wishes of half of the

population of Mosul, and they did not need to know the other halfsthat is, the urdswishes, because they considered urds ignorantand lacking a coherent and rational opinion. Thus, a plebiscite wasunnecessary.


29/67

25

One half of the population of the Mosul Vilayet consistsof urds. Of the other half, threesixths are Arabs. Thatthese Arabs desire to remain in the Iraq State will scarcelybe disputed. Twosixths are composed of the nonMuslim

minority, iz: the Yezidis, the Jews and the Christians. TheYezidis hae repeatedly expressed their desire to remain inIraq, and the same is the case with the Jews and the ChaldeanChristians. The Nestorian Assyrians, it is true, would preferto be included neither in Turkey nor in Iraq, but [that they]would prefer, gien suitable safeguards, to remain under a Statewhich will enoy, for a time at any rate, a considerable measureof ritish adice and protection, is not open to question. The

remaining onesixth consists of Turkoman. They are contentedand prosperous under the rule of the Iraq State, and thereis eery reason to beliee that they would remain so. Let itbe assumed, howeer, purely for the purpose of the presentargument, that they are unanimously desirous of a return toTurkish rule. Thus the position is that the wishes of onehalfof the population of the Mosul Vilayet are wellknown and

that in their case, the plebiscite is completely unnecessary. Theother half is composed of urds, the great maority of whomare ignorant tribesmen from whom no coherent expression ofopinion could be obtained.77

These bilateral discussions between the two biggest military powers inthe Middle East with regard to Mosul ended without diplomatic success.After Turkey and the ritish Empire failed to reach a diplomatic agreement

within the stipulated nine months, the Mosul question was submitted tothe League of Nations on August 3, 194.


30/67

26

The Frontier Commission (1925): Determining the

Mosul Populations Wishes

In the aftermath of the futile Mosul negotiations, political deelopments,especially those in Turkey, had a signicant impact on the use of statisticaldata. During this period, Turkish nationalism, which in the past had beeninclusie and had emphasized religious identity, became a more exclusieidentity following the abolishment of the Caliphate in 194 and theurdish uprising in 195. After the abolishment of the Caliphate, theurds were forced to reealuate their links with Turkish culture and their

political attitude toward the Turkish Republic. Whereas up to this point,the urds allied with the Turks primarily because of the existence of theCaliphate, once it was abolished, deep tensions were reealed, particularlyfollowing the outbreak of the February 195 urdish uprising in Turkey.This political change had a denitie impact on both the ritish andTurkish statistical approaches to the Mosul question.

After negotiations failed to sole the Mosul question, the League of Nationsset up the Frontier Commission in 195 to determine the sentiments ofthe local population in Mosul. The Commission was authorized, based onits ndings, to recommend a solution that the League would implementafter consulting with the Iraqi, Turkish, and ritish goernments.78 Theappointment of an inquiry commission actually marked a diplomaticsuccess for ritish diplomacy, since ritish ocials had argued thatappointing a commission was a procedure which the ritish oernment

had always thought would proe more eectie for the solution of thedispute than a plebiscite.79

To determine the fate of Mosul, the Commission attempted to adopt abalanced and scientic approach. This scientic approach utilized a seriesof tests, based on geographical, ethnic, historical, economic, and strategicfactors, and also drew on the Commissions interpretation of the Mosulpopulations wishes. The Commission traeled to Mosul in February 195and stayed until March. During this trip, the Commission questionedthe Turkish and Iraqiritish parties. After the conclusion of the trip, theCommission submitted a nal report to the League of Nations, which wasslated to denitiely decide the matter.


31/67

27

From the outset, howeer, the Frontier Commission was not nearlyas scientic as it purported to be. For one thing, the composition ofthe Commission was problematic: more political than scientic. TheCommission was made up of representaties from Sweden, Hungary, and

elgium8

three minor countries who were deemed to rarely hae thecourage or experience to handle such questions with real skill, udgment,and impartiality.81 Moreoer, while none of these countries had anyobious direct economic interests in the disputed area, all were erydependent on their trade with ritain.8

eyond this, the indiidual proles of Frontier Commission membersfurther skewed the Commissions decisions. One of the three Commission

members, Count aul Teleki, was also a scholaran expert in geographyand his academic works were largely known to the international scholarlycommunity. He was on that account a key member of the Commission,during the research period as well as during the writing of the report. Forexample, the Commission based its conclusion that the Yazidis were anentirely distinct people from the urds on the fact that while the urdswere Muslim, the Yazidis were not. This conclusion was a departure from

that endorsed by both the Turks and the ritish during the preiousbilateral negotiationsnamely, that the Yazidis were in fact urds.83 Whilethe reason for this departure is open to speculation, it seems likely that itcame about as a result of aul Telekis cultural and national background.84Telekis countryas was the case in all of Central Europeiewed religionas the most important marker of national identity, and it is likely thatTelekis emphasis on religion at the expense of ethnicity impacted theCommissions decision.85

Other aspects of the Commissions makeup also likely impacted itsdecisions. Although seeral ritish, Turkish, and Iraqi Arab assessorsaccompanied the Commission during its inquiry tour, there were nourdish assessors. In addition, during its factnding mission to Mosul, theCommission did not isit arious rural areas and also did not meet specicsegments of the local population. Although they questioned eight hundred

persons of some education and inuence, most of these indiiduals wereselected from lists submitted by the ritish and Turkish assessors.86 Whilethe statistics proided by these representaties are unknown, it is likely thatthey did not accurately reect Mosuls dierse population.


32/67

28

In its nal report, the Commission found the ritish arguments moreacceptable than the Turkish ones. This conclusion, howeer, was basedon partial and probably skewed data. In addition to not conducting anystatistical inquiries in the eld, the Commission examined only three

statistical sources of data: the 1914 Turkish data, the 191 ritish data,and the 194 Iraqi last census data, which were actually ritishIraqi data.87 ased on this already aailable eidence and without eenseeing the details of the census, the Commission determined that the Iraqilast census was nearer to the truth.88 In its report, the Commissiondenitiely concluded that the greater part of the population of thedisputed territory is undoubtedly urd about eeighths. The urds aretherefore numerically the most important factor. . . .89 On the basis of this

statistical data, the Commission concluded, If the ethnic argument alonehad to be taken into account, the necessary conclusion would be that anindependent urdish State should be created, since the urds constituteeeighths of the population see Table 5.9

Since the ethnic data were not taken into consideration, the other mostimportant fact to be determined scientically was the wishes of the

Mosul population. It seems that the Commissions expert, Teleki, triedto map the geographical distribution of the polled population, whichnumbered around 8. In a sample map that shows the geographicaldistribution of Arbils population, a maority of the population of the citycenter were proturques see Figure 3.91 Considering all of Mosul, theCommission determined that Mosuls people were more in faor of Iraqthan of Turkey, but that this was not due to any feeling of solidarity; itwas instead on account of economic reasons, and also because of the desire

to presere the continuity of the ritish mandate. If not for these factors,the Commission concluded, the Mosul people would [hae] preferred[to] return to Turkey. In sum, according to the Commission, the Mosulpopulations priorities were rst economic and second politicaland onlylastly ethnic.9


33/67

29

Figure 3: Geographical Distribution of the Opinions of Arbils Population:Pro-Arabs, Pro-Turks, Pro-Kurds, Discretes

Source: Istn & bor, Der ungarische eograph l Teleki als Mitglied des Mossulommission, pp. 175.

Following these statements, the Commission noted the impracticality ofholding a plebiscite in Mosulwhich happens to hae been the opinionthat the ritish had held since the Mosul dispute emerged. In supportof this argument, the Commission obsered that aside from the fact that

a plebiscite might create ethnic conict, it would not reect the truewishes of indiiduals, since the existing social organization is medieal orfeudal . . . [indiiduals] follow [the opinions of] their tribal chiefs. . . . 93 Inother words, the Commission argued that the results of a plebiscite wouldbe determined by the communities leaders rather than the communitiesthemseles. Ironically, this claim was not raised in the past, when boththe ritish and the Commission solicited the opinions of representatiesof communities and especially of the Christian representaties. Only once

the communities under discussion were Muslim communities, in which thetribal chiefs also assumed traditional religious roles, was the Commissioncritical of their opinions.

To support the ritish argument regarding the impossibility of holding aplebiscite, the Commission argued that although the urds were by far thelargest group, the ethnic mix in the region was so complicated that it would

be impossible to draw a border along ethnic lines. ut in determining thatit was too dicult to draw racial boundaries between Arabs, Turks, andurds, the Commission took a problematic approach. In the analyticalsection of its report, the Commission alluded to how the three races


34/67

30

intermingle and how a mixed population is gradually being formed. Yet,despite making this claim, the Commission oered ust one example ofthis intermingling: According to the Commission, the Bayat tribe was 65percent Turkish and 35 percent Arab and included intermarried couples

composed of members of both races.94

Yet, without proiding the actualproportion of mixed marriages, the Commission arried at the conclusionthat so great is the confusion of races in the disputed territories, thatit would be dicult to draw ethnic borders. Instead of examining theintermarriage issue with statistical methods, the Commission drew theirconclusion by generalizing from a single example. As Adolphe Quetelethad pointed out much earlier, this approach has its limits, since social factscannot be generated by arbitrary or exceptional examples.95

The Commission also proposed that the russels line, which was the borderdrawn in the middle of the positions of the ritish and Turkish armiesat the end of World War I, be used to sole the frontier problem. Thissolution actually diided the urds and, to a lesser degree, the Turkomanswho lied in Iraq from those in Turkey. Accordingly, it seems that theCommission found it easier to draw boundaries that would separate the

urds from each other than to create boundaries that would separate urdsfrom Arabs, who were already diided climatically and topographically seeFigures 4 and 5.96

Table 5: 1925 League of Nations Data

Kurds Turks Arabs Christians Jews Yazidis Total

Sulaymaniyah 189,900 0 75 0 1,550 0 191,525

Arbil 170,650 2,780 11,700 3,900 2,750 0 191,780

Kirkuk 47,500 26,100 35,650 2,400 0 0 111,650

Mosul 88,000 9,750 119,500 55,000 7,550 26,200 306,000

Total 496,050 38,630 166,925 61,300 11,850 26,200 800,955

Source: League of Nations, Question o Frontier between Turkey and Iraq : Report Submitted to the Council by the

Commission instituted by the Council Resolution o September 30, 1924enea : League of Nations : C.4. 195,

VII. no. 14, pp. 7677.


35/67

31

Figure 4: Comparison of Turkish and British Population Statistics

In the end, the ritish strategy of employing the League as an adunctof ritish diplomacy97 proed to be successful, as the Commissionsconclusions were clearly proritish. The Commission reected a

plebiscite in Mosul and drew boundaries between the Arabs, the urdsand the Turks.98 In addition, the Commission proposed that the russelsdemarcation line should become the permanent border between Turkeyand Iraq. These borders were practically identical to those Curzon hadproposed much earlier.99

The Commission awarded Mosul to Iraq on the condition that the ritishmandate be extended to twentye years and that certain guarantees

be gien to the large urdish population. On December 16, 195, theCouncil of the League of Nations decided unanimously that Mosul wasa part of Iraq, with a permanent border to be drawn along the russelsline. In addition, the Council instructed the mandatory goernment ofIraqalong with the ritish, as the mandatory powerto guarantee thatocials of the urdish race be appointed for the administration of theircountry. The Council also instructed that the urdish language would

be the ocial language in the Mosul administration and especially in theschools and the courts, and it stated that urds would be allowed to takepart in the administration of Iraq.1

20%

30%

40%

50%

60%

0%

10%

Kurds Turks Arabs Non-Muslims

1914 Turkish Data 1921 British Data


36/67

32

Although the Commission criticized a few ritish arguments, it concludedwith a proritish stance that faored the status quo in the region. TheCommission thought that the ritish successfully understood the wishesof Mosul people, and that it was therefore unnecessary to conduct a

plebiscite. The Councils decisions spelled an end to the scienticdiscussions regarding the plebiscite, and to the resort to statistical data andthe selfdetermination principle in the serice of the military status quo.Subsequently, on June 5, 196, Turkey and ritain agreedto the Councilsrecommendations, and ritain included a pecuniary compensation of7, to Turkey.


37/67

Figure5:

Eth

nographicMapofMosulCompiledbytheFrontierCommission

Source:Infact,th

ismapwasdrawnbyTeleki,who

haddrawnseveralethnographicm

apsforthedisputedterritoriesinEuropee.g.,theethnographicmapofHungarybased

onthedensityofpopulation,accordingtotheCensusof1910.

Foranexaminationof

hisethnographicmaps(including

theMosulone),seeIstvn&Gbor,Derungarische

GeographPlTele

kialsMitglieddesMossul-Kommission,pp.172

5.


38/67

34

Statistics in the Delineation of Kurdish Districts

(1932)

y 196, after Turkey had relinquished its claims to Mosul, the urdishquestion turned into a problem common to Turkey, Iraq, and theritish Empire. The League of Nations 196 promises became thebasis for a rising urdish nationalism, which obliged the ritishbackedIraqi monarchy to grant the urds a certain degree of autonomy. Oerthe course of the following years until 193, the urds demanded aclarication of their rights and threatened ritish interests in Iraq with their

widespread unrest.11 These political changes coincided with changes in thestatistical presence of urds in the ritish and Iraqi tabulations, as both thenumber and proportion of urds were lower in both tabulations than inthe earlier data.

The ritish, who when determining the frontier between Turkey and Iraqclaimed to be ery precise and scientic, lacked that precision and scientic

approach when later determining the borders of the urdish districts andthe size and proportion of the urdish population in Iraq. This lack ofprecision was likely aected by a change in the ritish mentality, which wasreected in their policy change toward the urds. In reecting the Turkishdemand for a plebiscite in Mosul, the ritish had based their politicalarguments on racial and ethnic considerations. Following the creationof Iraq, howeer, the ritish started to examine the urdish populationand deise subsequent ritish policy primarily through the urdish

language. This chapter examines the use of statistical data regarding theurdish population when the ritish and Iraqis determined the urdishdistricts in northern Iraqand analyzes the dierential ritish and Iraqiinterpretation of these data.

When the details of the 193 AngloIraqi treaty became public, theurdish regions became scenes of unrest, out of fear that the aghdad

goernment might be more antiurdish after the exit of ritish troops.On the other hand, urdish notables submitted an aalanche of petitionsto the ermanent Mandates Commission.1 The maority of the petitionshighlighted the underrepresentation of urds in the state apparatus


39/67

35

and the need to more clearly delineate their rights and the borders ofthe urdish districts. For example, the National Central Committee inSulaymaniyah claimed that the maority of administratie and executieocials of urdish districts were Arabs, who press[ed] and intimidate[d]

the urds. As a result of these actions, the Committee argued that oncethe mandatory regime ended, aghdad would make the urds situationworse than [in] the Turkish period. Accordingly, it demanded theformation of a urdish oernment under [the] superision of [the]League of Nations,13 contending that the urds hae the ability to forman independent State that will be composed of a sucient number ofinhabitants . . . more than a million souls . . . [which is] greater than that ofthe Sunni Arabs who constitute . . . [the] carcass of the Iraqi ingdom.14

The petition, which was signed by e urdish deputies, demanded thatthe Iraqi goernment draft a special law and determine the boundaries ofurdish areas.15

These urdish petitions to the League of Nations panicked the ritish,who wanted to leae Iraq as soon as possible. In response, they decided tourgently undertake a fresh and independent inquiry. The ritish feared

that as a result of the incessant urdish petitions, tensions would increasein this extremely important section of the world and culminate in ageneral urdish uprising.16 To preent this from happening, the ritishorganized an inquiry in Mosul, sent a counterpetition to the Leagueof Nations, and proposed to the Iraqi goernment that it promulgate aLanguage Law. After the League of Nations reected the urdish demandson Noember 17, 193, the ritish army conducted a military operationin Sulaymaniyah and forced Sheikh Mahmud arzani out of Iraq.

Meanwhile, the Iraqi goernment enacted the Local Languages Law no.74, May 3, 1931, according to which the urdish language became theocial language of certain districts and the primary or secondary languagein the courts in certain other districts.17 In addition, during May 193,the Iraqi goernment made a formal declaration regarding minorities tothe League of Nations, in which it promised that all racial, religious andlinguistic minorities shall be equal before the law.18 This declaration was

promulgated on July 13, 193.


40/67

36

British Statistical Inquiries and the Language Law

In response to the urdish petitions, the ritish conceied the idea for

the Local Languages Law, which was enacted in 1931. To persuade theLeague of Nations that ritain was the mandatory power responsible forthe Iraqi state, Sir inahan Cornwallis, the adisor to the Iraqi Ministryof the Interior, made a careful examination of urdish complaints.With the help of Iraqi state ocers, Cornwallis calculated the number andproportion of both the urdish population and urdish ocials in theurdish districts.19 The ritish relied on these statistical data in theirinquiries and in their internal correspondence, as well as in their externalcorrespondence with the Iraqi goernment and the League of Nations.The ritish described the urds as an uneasy, suspicious, insatiablepeople, whereas the urds presented themseles as realistic and scientic.Moreoer, the ritish generally described the urdish allegations asnationalist and claimed that they were based on feeling.

In his report, Cornwallis addressed the number and proportion of urdish

ocials in urdish areas of Mosul. According to Cornwallis, there were34 urdish ocials in these areas, constituting 44 percent of the totalnumber of ocials. Cornwallis noted, howeer, that if the statistics ofirkuk liwa [a subproince] are excluded, the remaining 79 urdishocials accounted for 55 percent of the total. In other words, Cornwalliswas coneying that urdish ocials constituted a maority of the total.While the urds did in fact account for the maority of ocials, a closerexamination reeals that the urds in the urdish areas constituted 7percent of the total population, whereas the Arabs, who were only 8 percentof the total population in those areas, constituted 3 percent of the ocialssee Table 6.11


41/67

37

Table 6 : Ethnic Distribution of the Ofcials Employed in Mosul

Gazetted Non-Gazetted Total

K. T. A. C&J K. T. A. C&J K. T. A. C&J

Sulaymaniyah 19 8 1 2 90 14 11 12 109 22 12 14

Arbil 18 5 4 1 95 23 34 - 113 29 38 1

Kirkuk 16 11 18 2 29 107 40 6 45 118 58 8

Mosul (Kurdish

Qadhas)

11 3 14 4 46 8 43 35 57 11 57 59

Source: See the condential Memo of Lord . C. asseld, Secretary of State for the Colonies, dated April 3,

193, CO 73/157/1, pp. 36.

In an October 4, 193, letter to the League of Nations, the ritish arguedthat the urdish complaint was false. To persuade the League of Nations,statistical data occupied a signicant portion of the ritish report andof ritish political arguments. Using Sulaymaniyah as an example, the

ritish argued that of 157 total ocials there, 19 were urdish and only1 were Arabs. Een as they made this claim, the ritish acknowledgedthat in other urdish districts, the proportion of urdish ocials wasnot so high. Nonetheless, they reasoned that the increase in the numberof ocials of nonurdish race was a recent deelopment, and thata considerable maority of the nonurds spoke urdish and werescarcely distinguishable from the urds in the areas they administer.To the ritish, the increase in the number of nonurdish ocials was

not due to the policy of the Iraqi Mandate State, but was rather a resultof urdish ignorance and of some of the Turkoman goernors policies.In other words, while acknowledging that urds were underrepresentedin the goernment, the ritish reected the urdish claim that thisdeelopment was the fault of the ritish. The ritish stated that to ndsuitable and qualied urds was increasingly dicult, because theiragricultural pursuits and primitie existence are not calculated to t them

for oernment serice.111

In fact, the ritish noted, the urds wereunderrepresented not only in skilled obs, but in unskilled ones as well.


42/67

38

When submitting their claims to the League of Nations, the ritishpresented only one side of the results of their statistical inquiries. Forexample, when the ritish refuted urdish complaints and stated that theproportion of Arab ocials was only 8 percent, they failed to mention

that according to that ery same statistical tabulation, the proportion ofurdish ocials in Sulaymaniyah was only 69 percent, while the urdishpopulation was 99 percent of the total population.

Cornwalliss most dubious data in his statistical tabulations concernedirkuk. According to Cornwallis, the proportion of urds in irkukwas 49.5 percentthat is, not the maority. In contrast to this gure,howeer, the ritish ethnographic map of 1931 indicated that the

urdish proportion in irkuk was 57 percent. To present the urds asa minority, therefore, Cornwallis carried out two simple falsications.First, he increased the number of Jews from ,47 to 8,47. Second, hedecreased the number of urds from 77,68 to 67,73. The number ofJews in irkuk had neer been more than 3, in other ritish data,and according to 1917 ritish data, the number of Jews was ,47. Thus,Cornwallis inated the number of Jews to coer up the fact that the urds

were a maority in irkuk and, in turn, to decrease the extent of urdishocials underrepresentation. Second, ritish statistical data showed thatonly 8 percent of the ciil serants in irkuk were urds, despite the factthat urds constituted 57 percent of the total population; so to coer upthis underrepresentation, the ritish misrepresented their own data. Theritish found it extremely important to coer up the underrepresentationspecically in irkuk, since by 197 they had laid the basis there fora petroleum industry that opened up a huge ob market, and had thus

assumed a signicant leel of responsibility for the area See Table 7.

Table 7: British Data on Kirkuk, 1930 and 1931

Kirkuk Arabs Kurds Turks Christians Jews Other Total % Kurds

British1930 data

26,561 67,703 28,741 1,228 8,472 0 136,705 49.5

British1931 data

26,561 77,608 28,741 1,228 2,472 192 136,802 56.7

Source: For 193 data, see CO 73/157, p. 36; for 1931 data, see Records of Iraq, V.7, pp. 59697.


43/67

39

The absence of Yazidis in the ritish statistical tabulations was alsoproblematic. In the period following 196, the ritish had argued thatmother tongue, not race, had to be taken into consideration. Consequently,the ritish were supposed to include the Yazidis, who were mostly

urdophone, in their statistical data. Inclusion of Yazidis, howeer, wouldhae increased the proportion of urds, so they were excluded.11

In short, the most important problem with regard to the ritish statisticaldata was their inaccuracy, extending to the number and proportion ofurds in the urdish areas of Mosul. The 193 ritish data did notcorrespond to either their preceding or their subsequent data. Accordingto the 191 ritish data inoked during the Mosul negotiations, there

were 454,7 urds including 3, Yazidis, and they constituted 58percent of the total population. The 194 data, which the ritish andIraqis presented to the League of Nations Commission in 195, depictedsimilar statistics, as the urdish population in Mosul including nomadsand Yazidis was found to be 516,94, constituting 65 percent of thepopulation. According to 193 data, howeer, which were prepared inresponse to urdish complaints, the number and proportion of urds were

substantially lower 393,, equal to 55 percent of the population.

While this discrepancy can be explained by technical reasons pertaining tocensus taking and statistics, additional data from a later period shows usthe other reason why they were dierences between the 194 and the1931 data. These statistical data, prepared and published in 1947 by CecilJohn Edmonds, were designated for ritish inner circulation only and werenot intended for diplomatic purposes.113 The data present the number and

proportion of urds in Mosul roince as essentially comparable to thenumbers and proportions gien in the data of 191 and 194. These1947 statistical data were prepared following World War II, when thepolitical situation was unstable and the Soiet role in the Middle East hadincreased.114 According to these data, by 1947 the number of urds withinthe boundaries of the old Mosul roince was 84,4, amounting to 63percent of the population. While it is eident that the urdish population

had increased from 191 to 1947, the proportion of urds is nearlyidentical to that found in the 194 data.115


44/67

40

This ritish manipulation of data was certainly drien by politicalpurposes. If the 194 Iraqiritish data had been used, the ritish wouldnot hae been able to easily disproe the urdish allegationsand thediscrepancy between the proportion of urds in the population and the

number of urdish ocials would hae warranted an increase in the latter.According to the ritishIraqi data of 194, the proportion of urdsin the urdish districts was oer 83 percent, not 74 percent as reported in193. The proportion of urdish ocials in the urdish areas of Mosulthat the ritish presented 44 percent therefore had to be reealuated inlight of the proportion shown in the 194 data 83 percent.

The ritish recommendation to the Iraqi goernment with regard to

remedying the urdish situation proed to be the most importantritish impact on the urdish question. Following his inquiry into theurdish allegations, Cornwallis drew up a list of recommendations thathe persuaded the Iraqi goernment to adopt.116 In accordance with the196 League of Nations directie, Cornwallis proposed a LanguageLaw, by which the urdish language would become the ocial languagein the administration, the school system, and the ustice system in the

predominantly urdish qadhas. This legal proposition was problematic,because it was based on the criteria of language and not race. In fact,though the League of Nations had stipulated that the ocials of urdishrace should be appointed for the administration o their country, theritish replaced the race criterion with the language criterion when theyproposed to the Iraqi goernment that language should be made the test ofemployment. The ritish reasoned that we hae ustied the new policyof making language and not race the test of employment on the grounds

that it is in the best interest of the urds.117 The reality, howeer, wasthe total opposite. For example, the decision to appoint urdishspeakingArabs in place of the urdish oernor of Sulaymaniyah, Tawq Wahbieg, and the urdish police chief caused maor chaos in the urdishdistricts of Mosul.118

Last but not least, the most important eect of ritish statistical

manipulation was on the delineation of urdish districts, in which urdishwas introduced as the ocial language. Deputies of urdish origin hadrequested this delineation before the ritish left, and this delineationclaried urdish rights and districts. Deputies of urdish origin had


45/67

41

requested from the ritish, before they left Iraq, the delineation of urdishrights and the borders of the urdish districts.119

Since 1919, the urdish regions of Mosul had been called by many

dierent names, including Southern urdistan, urdish country, and theurdish district or urdish zone. According to the League of Nationsinstructions, the ritish goernment, as the designated mandatory power,was supposed to institute guarantees regarding local administration to theurdish population and was therefore obligated to demarcate the urdishdistricts.1 The ritish used their population data to delineate thesedistricts. The problem was not in the use of these statistical data, but ratherin the criteria used to dene terms such as maority, predominance,

and considerable proportion that were intended to best determinethe urdish districts. Although the ritish had used the maoritycriteria during the Mosul negotiations, they now started to use the termpredominance. They also proposed that the Iraqi goernment apply theLanguage Law in the predominantly urdish qadhas. Interestingly, theCouncil of Ministers of Iraq promised to apply the law in areas where theurds constitute a maority of the popula

"statisquo": british use of statistics in the iraqi kurdish question (1919–1932)

Documents