Siegfried Gruber


Regional variation in marriage patterns in Albania at the beginning of the 20th century




Session C12: ANALYSIS OF MARRIAGE




Social Science History Association Annual Meeting

St. Louis, October 24-27, 2002



Projects "The 1918 Albanian Population Census: Data Entry and Basic Analyses" and "Migration in Albania"
email


Introduction

Albania lies according to John Hajnal within the region of the "Eastern European marriage pattern" with a low age at marriage and a virtual lack of celibacy since it is situated east of the line running roughly from Trieste to St. Petersburg (Hajnal 1965:101). He did not deal with Albania as a unit in his study since it was no independent state at the beginning of the 20th century and statistical data was not readily available.
Several scientific publications mention a the low age at marriage in the tribal regions of Northern Albania. Karl Kaser points to a "relatively high age at marriage for women" between 16 and 20 years and between 20 and 23 years for men in the middle of the 19th century recorded by Johann Georg von Hahn (Hahn 1854:I,180;Kaser 1992:254). In the beginning of the 20th century Mary Edith Durham recorded ages at marriage of 13/14 years for girls and of 15 years for boys (Durham 1928: 192). "In northern Albania girls are married as soon as they come to sexual maturity and begin bearing children as soon as they are biologically able. There is no time of peace." (Coon 1950:27). For Southern Albania Johann Georg von Hahn recorded ages at marriage of 12 years for girls and of 15 years for boys in the middle of the 19th century (Hahn 1854:I,143). Karl Kaser explains the low age at marriage with the urgent need for children in the milieu of patriarchy. It was an eminent concern to secure the (male) offspring in view of high infant mortality, high rates of blood feud, an urgent need for manpower and the continuation of the patriarchal line. An early age at marriage in order to "utilize" the female fertility corresponds with this system (Kaser 1995:153). A low age at marriage also favors the formation of joint family households in the absence of neolocality.
Philip E. Mosely distinguished three principal belts of zadruga society. The first one consists of the tribal society of pre-1912 Montenegro and of Northern Albania. The second consists of Bosnia, Herzegovina, western Croatia, northern and central Macedonia and central Albania. The third belt consists of Croatia, Slavonia, pre-1912 Serbia, western and central Bulgaria, southern Macedonia and southern Albania (Mosely 1976:60f.). Therefore Albania extends over all of these three belts with Northern Albania lying in the core zone.
Peter Laslett developed a set of tendencies in domestic group organization in traditional Europe with four sets: West, West/central or middle, Mediterranean, and East. These sets are based on criteria for occasion and method of domestic group formation, procreational and demographic criteria, criteria of kin composition of groups, and criteria of organization of work and welfare (Laslett 1983:526f.). The Balkans are not clearly included in the eastern or Mediterranean set (Laslett 1983: 529). Table 1 brings the criteria for these two sets of tendencies concerning demography and kin composition of groups.

Table 1: Sets of tendencies in domestic group organization in traditional Europe

criterion
 Mediterranean
 East
Age at marriage, female
 Low
 Low
Age at marriage, male
 High
 Low
Proportions marrying
 High
 High
Age gap between spouses at first marriage
 Wide
 Narrow
Proportion of wives older than husbands
 Low
 High
Proportion of widows remarrying
 Very low
 Very low
Proportion of resident kin
 High
 High
Proportion of multigenerational households
 High
 Very high
Proportion of households headed by never-married women
 Low
 High
Proportion of solitaries
 Low
 Absent
Proportion of no-family households
 Low
 Absent
Proportion of simple-family households
 Low
 Low
Proportion of extended-family households
 Low
 Low
Proportion of multiple-family households
 High
 Very high
Proportion of complex-family households (c7 + c8)
 High
 Very high
Proportion of frérèches
 High
 Very high
Proportion of stem-family households
 Low
 Low
Proportion of joint-family households
 Very high
 Very high

In a recently published joint article we investigated household structures in Albania in the early 20th century. This article is based on the data of 12 villages or parts of cities of the Albanian census of 1918. Mean age at marriage for women was for most of the settlements at about 20 years or below and only in urban environments higher. Mean age at marriage for men was ranging from below 20 years to more than 30 years (Gruber and Pichler 2002:360). Female age at marriage does fit into the "Eastern European Marriage Pattern" while male age at marriage does only partly fit into it. This may be the effect of the small size of the sample of the census, which could differ from the average and therefore this paper is intended to look at marriage patterns in Albania at the beginning of the 20th century with greater detail. In addition the questions whether Albania fits into the "Eastern European Marriage Pattern" or not, and which of the set of tendencies developed by Peter Laslett is closer to the Albanian situation, should be addressed.
In a paper about the number of children in Albania I have found much more regional variation than variation between religious groups (Gruber 2001:10). Therefore the regional aspect of variation in marriage patterns will be a major part of this paper. Averages of whole countries can obscure marked differences within these countries and lead to false conclusions. It is very difficult to deal with regional variation and patterns without having them mapped out. Therefore a set of thematic maps will make up the second part of this paper besides the first written part. These maps are not only intended to illustrate results of research, they are also intended to be a means of research. They enable the researcher to see geographical patterns, which are much more difficult to detect with other research tools. The usage of computers facilitates this kind of research by speeding up the production of such maps considerably.

New data

A recently rediscovered source allows us to compare the ideal with reality, the problem for Southeastern Europe in general and Albania in particular being that except in rare cases, we do not have accurate information from population censuses before the late 19th century. The general scholarly opinion is that in Albania, accurate population censuses were not carried out before the census of 1945 (Lienau and Prinzing 1989:160; Schmidt-Neke and Sjöberg 1993:464).
Michael Schmidt-Neke and Örjan Sjöberg describe the inaccuracies of the earlier population censuses and mention the following problems:

The first population census conducted by the Albanian government was taken in 1923; many of its results on a macro level have been published (Selenica 1928). The next population census was taken in 1930, the original data of which is still available, although not as a compact collection. Parts of it are stored in the Tirana State Archive and parts are dispersed among provincial archives. As mentioned above, the data from this census seems to be rather inaccurate. The census lists include persons who were born later or who moved to a village due to marriage after 1930. In the village of Konaj-Fanë, 11.2 % of the whole population belonged to this group. Another problem is that of under-enumeration. In a study of 10 villages, there was a proportion of 110 men to 100 women. This phenomenon was much more accentuated among Muslims (113.7 men to 100 women) than among Orthodox (107.8 men to 100 women) and Catholic (105.7 men to 100 women) Christians. Most of this imbalance is probably due to the under-registration of girls: in the age group above 19, there was a ratio of 97 men to 100 women (Gruber and Pichler 2002:354). This census is obviously of less value than that organized in 1918 by Graz-based Franz Seiner, an expert on statistics. The quality of this population census has been underestimated in scholarly literature until now.
In January 1916, almost the entire territory of Albania was occupied by the Austrian-Hungarian army with the exception of fringe areas in the south of the country, which were occupied by Bulgarian, French, Italian and Greek troops. This territory can be seen on map 1 . Shkodër, at that time the country's largest city, became the seat of a military administration that left the traditional civil administrative structures unchanged.
The population census was taken on March 1st, 1918. The material was transported to Shkodër and safely stored. It was then processed "with the help of a large number of intelligent young Albanians" (Oberhummer 1921). By the end of September 1918, the data had been double-checked and completions and supplements carried out. However, these activities had to be stopped due to the planned withdrawal of the army in October. The order to destroy the entire census material was neglected with the exception of the district headquarters in Lushnja. Therefore, the material concerning the Berat, Fier, Lushnja and Shkrapar regions (89,142 persons) is missing (Seiner 1992b:5). The aggregate results of the 1916 census are available for the first three of them while for the Shkrapar region there is only an estimation of the population available. Map 2 shows the districts of Albania and the availability of data for them. The surviving material, which covers the major part of the country, is as follows: 435,075 out of the 803,959 (this figure was calculated too high) persons counted in 1923 (54 percent) or 20,096 square kilometers out of the country's total area of 28,748 square kilometers (70 percent). It was very difficult to transport the census material to Vienna. The military administration unit responsible for the delivery agreed to hand out the material to the Austrian Academy of Sciences along with the permission to publish and to work with it. The Academy asked the census director, Franz Seiner, to work out the basic statistics. These tables were published in 1922, supported by funds from the Albanian government (Seiner 1922b). Instructed by the Albanian government, Seiner also separately published the results of the census relating to the tribal areas of northern Albania. On the basis of these results, he prepared the first map on the distribution, size and borders of the tribal territories (Seiner 1922a). One year earlier, the director of the Balkan Commission of the Austrian Academy of Sciences, Eugen Oberhummer, published first preliminary statistical results (Oberhummer 1921). At that time there were also plans to publish the village level data, but the Academy was not able to find adequate funding for the publication. Oberhummer concluded: "If we do not receive outside support, this material indispensable for both scholarly work and any orderly administration in Albania will go to waste." (Oberhummer 1921:63). And this is what happened.
The population census was taken very carefully. It was precisely conducted and prepared over one year in advance. The collection of statistical data began two months after the Austrian-Hungarian army entered the country. In March 1916, a provisional population census was taken, linked to a livestock census and a survey of food supplies. It was badly prepared and full of crucial mistakes. In March 1917, Franz Seiner, a census expert and statistician, was sent to Albania to take over the post of chief census official in the occupied territories. It was his duty to establish a provincial office for statistics and after that to organize a general population census (Seiner 1922b:1f.).
This population census needed huge preparation. First of all, all houses, households, huts and buildings in general had to be counted and the houses provided with numbers, since this had never been done before. A second step was to delineate settlements and fix the borders of villages, as well as to decide how the names of the villages should be recorded in writing. An index of villages, towns and cities with about 1,800 entries was created. An official order for a rough population census was given on May 26th 1917, in order to familiarize the population with the census procedure and to convince people that it would not make sense to under-report the number of children (usually the female ones), since the distribution of necessary food supplies would be based on the data of the population census (Seiner 1922:2). The result was that the census reported 252,794 men and 251,423 women, i.e. 100.5 men to 100 women, or an almost equal distribution (Seiner 1922b:8). This is in contrast to the 1930 census mentioned above, and to the results of all Balkan countries in the 19th century and the beginning of the 20th century, which always reported a large surplus of men.
The process of establishing a naming system for topographical locations was difficult and time consuming, since many locations had names in three or four languages and there may have been significant differences in the pronunciation and spelling of the topographical names by villagers. In addition, many villages did not have a fixed communal name, but just the names of the quarters (mahalle) of the village. Finally, an "Albanian Literary Commission" was established in Shkodër, which worked out guidelines for naming Albanian localities (Seiner 1922b:2f.).
Another problem that had to be solved was how to establish family names and people's first names. Until then, only few Albanian families had a fixed family name. Generally, the mass of the people used only first names, and for distinguishing purposes they added the names of the father and of the patrilineal grandfather. An additional problem was that in the villages the naming pool was relatively small, thus the administration ran into problems. This was why settling on names for the purpose of the population census was made obligatory. The heads of the households were free to choose their names. In only a few cases, the name was selected by the administration; the guidelines of this process were worked out by the scholar of Albanian, Max Lambertz (Seiner 1922:4f.).
The 1st March 1918 was finally chosen as the date for the population census. This date was not chosen arbitrarily since in spring many families left their villages in order to migrate to the mountainous summer pastures for at least half of the year. A "census commissioner" was appointed for every village. This could be a member of the occupation forces or an Albanian military officer; they were supervised by regional and district commissioners. Most of the commissioners were trained by Seiner personally. The registration forms with their 24 columns were in German and Albanian, and the data could therefore be entered in German or Albanian. The village chiefs were obliged to correct false data given by the head of a household (Seiner 1922:2-5).
Evidently everything was done to conduct a precise population census; this is why we can qualify this material as outstanding relative to the conditions of that time and in comparison to the later censuses. Beryl Nicholson investigated the data for the Kreis (district) of Malakastra and stated: "The evidence so far gives reason for optimism that the manuscript census schedules will prove to be a rich source for the study of society in Northern and Central Albania in the early part of this century." (Nicholson 1999:29).

Data used for this paper

The research project "The 1918 Albanian Population Census: Data Entry and Basic Analyses" aims at making the data of this census available for scholarly research.1 The research project started in August 2000, and we are still adding new data, which means that the results presented in this paper are still preliminary. During the data entry, we divided the settlements into two groups: "normal" settlements and "deviant" settlements. Deviant settlements will be entered completely, while normal settlement will be entered in 5-percent-samples. The following criteria serve as markers for "deviant" settlements:

These criteria often overlap, e.g. size of settlement, occupational structure and cities.

Table 2: Available data according to criteria
Criteria
 Settlements
 Persons
Ethnic minorities
 42
 17,106
Sex ratio
 53
 7,768
Household size
 49
 12,975
Orthodox Christians
 57
 20,206
Cities
 22
 37,312
Occupational structure
 34
 43,322
Size of settlement
 7
 26,076
All “deviant” settlements
 197
 84,009
5-percent-sample
 56
 18,660
All settlements
 253
 102,669

This paper is based on the population that was present at the time of the census and does not include the absent population that was also recorded. Table 3 has the number of persons present at the time of the census according to the highest administrative level, since the paper will deal with regional variation.

Table 3: Available data of persons present by administrative units
Bezirk
 population
 data
 percentage
Kruja
 51,790
 6,823
 13.17
Puka
 32,504
 3,884
 11.95
Shkodra
 84,509
 31,185
 36.90
Tirana North
 110,489
 21,016
 19.02
Zhuri
 100,547
 17,009
 16.92
Tirana South
 31,588
 10,203
 32.30
Berati
 24,409
 5,625
 23.04

435,836
 95,745
 21.97

The different percentages have emerged from several reasons:

Characteristics of Albanian society according to the census of 1918

The Albanian society at that time was almost entirely rural and agricultural. Only two cities, Shkodra (23,000) in the north and Korça (22,000) in the south-east (outside the area occupied by Austro-Hungarian forces) had more than 20,000 inhabitants (Duka 1997:9f.). In the area covered by the census there were only 7 cities with at least 3,000 inhabitants; they accounted for 12 percent of the total population (Seiner 1922b:6). The population was not evenly distributed within Albania, as can be seen in map 3 . The mountainous areas were especially sparsely populated and the districts with a high population density contained a city. Out of more than 1,800 settlements, only 37 (including the cities mentioned above) had at least 20 percent non-agricultural population. Map 4 shows the distribution of the non-agrarian population within Albania in 1918. One can clearly see the concentration of this population at the coast and in the center of Albania where the cities were located. The population was almost entirely Albanian by ethnic affiliation: 95.9 percent (in Upper Albania, the northern and central part of the Austro-Hungarian territory in Albania) according to Seiner (Seiner 1992b:9). Our data shows a percentage of 96.5 percent, but covers also Lower Albania (the southern part of the Austro-Hungarian territory in Albania). The remaining population was mostly of Slavic origin with some Gypsies (Roma). Map 5 shows the distribution of the ethnic minorities within Albania in 1918. The highest percentages were in the eastern districts bordering to nowadays Serbia (Kosovo) and Macedonia, where the Slavic minority lived. There were also Slavic minorities in the cities and in the region around Shkodra, where people of Montenegrin descent lived. The majority of the population was Muslim: 74 percent according to Seiner (Seiner 1922b:10) and 83.5 percent in our data. Map 6 shows the distribution of Muslims in Albania in 1918, where a clear concentration in the eastern parts can be seen. The remaining population was Roman Catholic: 18 percent according to Seiner (Seiner 1922b:10) and 13.5 percent in our data, and Orthodox Christians: 8 percent according to Seiner (Seiner 1922b:10) and 3.0 percent in our data. The small number of Orthodox Christians is due to the fact that most of them lived outside the territory occupied by Austria-Hungary and that a considerable number of the Orthodox villages were in the area of the destroyed data. Map 7 shows the distribution of the Catholics within Albania in 1918 and a concentration of them in the north of the country is obvious. The Orthodox Christians were dominant in the south of the country, as can be seen in map 8 .
The percentage of literate people was almost negligible in rural areas: less than 2 percent of the men and almost no women were able to read and write. In cities a considerable proportion of the male population (about a third) was able to read and write and about one woman in 11 was also able to read and write. The usage of the Latin alphabet was much more common than the usage of the Turkish (Arabic) alphabet, especially among women. Some people were able to use both of them and a small number also recorded using the Greek or the Cyrillic alphabet.

Table 4: Percentage of Literacy in Albania in 1918

 urban men
 rural men
 urban women
 rural women
Latin letters
 23.6
 1.1
 8.7
 0.1
Turkish letters
 9.9
 0.9
 0.3
 0.008
literacy
 28.8
 1.8
 8.8
 0.1

The age structure shows a young population, as can be expected: half of the population was younger than 23 years old. There was also much age-heaping, therefore ages given in this paper should be seen as approximate ages compared to our modern concept of age. The young population means that there was a high percentage of unmarried people: 41.1 percent of the female population and 59.2 percent of the male population. The difference was due to the higher age at marriage for men. The married population comprised 41.0 percent of females and 37.4 percent of males. This astonishing difference was due to the slightly higher number of males in the population and cases of men having more than one wife. Being widowed was much more common for women than for men: 17.9 percent as compared to 3.4 percent. 6.7 percent of the population were absent at the time the census was conducted. This can be attributed to two facts: a tradition of migrant workers in the north-east of the country and the effects of the First World War.
The census showed almost the same number of men (50.1 percent) and women (49.9 percent) in the population, according to Seiner (Seiner 1922b:8). The database has more men (51.2 percent male and 48.8 percent female), which is mostly due to the fact that it includes people absent during counting, and the absent people were mostly men. Map 9 shows the percentage of men within the population of the districts. A male majority was in the north and in the south-west whereas there was a female majority in the central and the southeastern districts. The lowest percentages of males were in the northeast, where there was this already mentioned tradition of migrant workers.
Household size ranged between 4.5 and 8.5 persons for the different districts. The distribution of it can be seen in map 10 . There is no clear pattern throughout Albania, with two exceptions: in most of the northern districts household size was higher than the average (5.5 persons per household) and in the cities household size was below the average.

Types of households

Peter Laslett's set of tendencies of the Eastern and of the Mediterranean type consist of low proportions of nuclear and extended family households and of high proportions of joint family households. The overall proportions for Albania in 1918 were: 25.2 percent of the people were living in nuclear families, 26.4 percent were living in extended families and 45.0 percent were living in joint families. This shows a clear dominance of joint families, although half of the population was living in either nuclear or extended families. The share of nuclear families can be explained by the fact that not all of the men were able to have at least two sons surviving until adulthood, which is the prerequisite for forming a future joint family. A father with only one surviving son could form a joint family together with his son only during the life-time of the father. The continuation of a joint family of co-residing brothers (frérèche) requires that each brother has at least one surviving son. At an age of 30 years only 30 percent of all men lived with at least one child and it was only after an age of 40 years that more than half of the men lived with at least one child. The percentages of fathers with sons rose up to an age of 55 years to 73 percent and reached a peak of 84 percent at an age of 75 years. This means that 16 percent of men failed in their attempt to produce a male heir. In considering that many men died before an age of 75 years this percentage was much higher. At an age of 75 years the peak of the number of co-residing sons (married or unmarried) was reached, too: 2.1 sons (Gruber 2001:8f.). These numbers show that there was simply a demographic constraint against forming joint families. The high number of extended families is due to a high number of widowed mothers living in the household of a married son.
The regional pattern of the distribution of nuclear families in Albania shows a low percentage of people living in nuclear families in northern Albania and much higher percentages in central and eastern Albania. The cities also had higher percentages of people living in nuclear families (33 through 43 percent, map 11 ). Extended families were much more frequent in the central-eastern parts of Albania and in the south-east. The cities also had higher percentages than average. The lowest percentages were to be found in the north-west of the country (map 12 ). The distribution of joint families within Albania confirms the picture of northern Albania as the most patriarchal part of the country: In most of the northern districts more than half of the population was living in joint families. There were also some central Albanian districts with high shares of joint families. The lowest percentages could be found in the larger cities and the south-east of the country. In the cities of Shkodra and Durrës less than 20 percent of the people lived in joint families ( map 13 ).

Proportion married

The proportion married reflects not only marriage patterns but also the age structure of a population, since a high share of children in the population increases the number of unmarried people. 37.4 percent of the male population present at the time of the census were married and 40.7 percent of the female population present were married at the time of the census. The different percentages are due to men being married to more than one wife, due to different overall numbers of men and women and due to higher rates of absence for men. Map 14 shows the different proportion married among the male population by districts. The lowest proportions were in the north-east of the country, where there was a tradition of migrant workers. These absent adult males lowered the number of men and the number of married men within the region. The highest proportions were found in the central part of Albania. Map 15 shows the different proportion married among the female population by districts. There is no clear pattern: the different classes of proportion are spread all over the country. The lowest proportions were in the east of the country, but these districts were not connected. The city of Shkodra also had a low proportion of married women.
Looking at the proportion of married men and women in the age group of 15 to 50 years (Im, the years of reproduction of women and dominant years of reproduction of men) excludes the influence of the number of children and can serve as a measurement for the potential fertility since illegitimate children were virtually absent in this society. 62.2 percent of the men and 72.5 percent of the women were married in this age group. The proportion for females was similar to large parts of Southeastern and Eastern Europe at the beginning of the 20th century, but not the highest, as can be seen from the following table.

Table 5: Proportion married in the female population 15-50 years
Country
 Year
 Proportion
Greece
 1928
 56.6
European Russia
 1926
 62.8
Albania
 1918
 72.5
Romania
 1899
 72.7
Bulgaria
 1905
 73.7
Serbia
 1900
 80.8
(Coale and Watkins 1986:Appendix A).

The proportion of married women in the fertile years was highest in the district of Gramshi in the southeast of the country, where 90.2 percent of the women in this age group were married. This was caused by a very low percentage of widowed women. The east of the country had slightly lower proportions as the rest of the country, but there were no large differences (map 16 ). The proportion of married men in these years was highest in the south and the center of Albania, whereas in many of the northern districts the proportion was lower. But there were also some districts in the north with proportions above average. The lowest proportion was among the male inhabitants of the city of Shkodra with 44.8 percent (map 17 ). This is the effect of the high age at marriage in this city.
The proportion married or unmarried is used for calculating the singulate mean age at marriage and we shall have a look at the proportion married at different ages for both sexes to see the different marriage patterns throughout Albania in 1918. At an age of 15 years already 29.0 percent of all females were married. There were even younger girls recorded as being married. The proportion of women married at an age of 15 years was very unevenly distributed within Albania in 1918: There were districts with no married girls at such an age and there were districts with more than 40 percent of married girls at such an age. The highest rates of married girls were in three central Albanian districts and also some northern Albanian districts had rates above average. No married girls were recorded in the districts of Kruja, Hasi, Gramshi and Peqini South. There was no clear concentration of the districts with no or only very few married girls at an age of 15 years, but three of the districts were the west of the country, two were in the north-east and three were in central/south-east Albania (map 18 ). All the cities had low proportions of married girls at this age.
At an age of 18 years about half of the women were already married (52.5 percent). There were districts with still no woman married (Gramshi) and districts with all the women married at this age (Puka). The districts in the south-east were still the ones with the lowest percentages of married women in such an age, while the three districts in the west were now among the ones with percentages above average. All the north-west of the country had proportions of married women above average, there was only one district with a very low percentage of married women. Out of the three central Albanian districts with the highest proportion of married girls at an age of 15 years two of them were still in the leading group. All the cities had proportions married below the surrounding districts (map 19 ).
At an age of 20 years almost all women were already married: 81.1 percent. The really exceptional case was the district of Kiri in the north-west, where still only 13.6 percent of the women were married. Other districts with rates below average were to be found in the north-east and in the south-east, where already at younger ages the rates were lower than the average. The highest percentages were to be found now clearly in the center of Albania and the cities still had lower percentages than the surrounding districts (map 20 ). At an age of 25 years 89.2 percent of the women were married and there were 8 districts with a proportion of more than 95 percent, including the one with the lowest percentage being married at an age of 20 years. In the north-west there was a compact bloc of these districts, but there were also such districts in other parts of the country. The lowest percentages were to be found in the north-east and south-east of the country ( map 21 ). At an age of 30 years the percentage married among women was 87.4 percent, which is lower than the percentage 5 years younger. This is the effect of increasing widowhood among these ages: At an age of 30 years 10.7 percent of the women were already widowed. There is no clear pattern of the distribution of married women at an age of 30 years throughout Albania, but the central part and parts of the north-west, which had the highest rates at younger ages, were now among the districts with the lowest rates at this age. This is obviously the effect of getting widowed earlier when getting married earlier ( map 22 ). Therefore the proportions married in the fertile years were rather similar in almost all of the districts. Shkodra was at this age the only city with a lower proportion married than the surrounding district, all other cities had now higher percentages than the surrounding districts.
The corresponding proportion is that of the still unmarried women at an age of 30 years. The overall rate was 1.9 percent, which means almost universal marriage for women. There were only two districts(Kiri and Elbasani North ), and the city of Shkodra, with rates of more than 5 percent. There may be some trouble with the data for the district of Kiri, since more than 99 percent of the women were recorded as being married at an age of 25 years. There is no regional pattern of the proportion unmarried at an age of 30 years (map 23 ).
The effect of widowhood can be seen at map 24 , where the proportion of widowed women at an age of 40 years is mapped. The average proportion was 30.1 percent. The highest proportions of widowed women were in the southern half of the tribal area and at the coast. In addition the two southernmost districts with available data also had proportions above average. The three districts with the lowest percentages were not located in one area, but dispersed. All of them have been mentioned earlier: Gora with its high percentage of labor migrants, Kiri with its low percentage of married women at young ages and Gramshi with the highest percentage of married women within the age group of fertile women.
At an age of 20 years 26.6 percent of the male population were already married. Map 25 depicts the distribution of proportions married by districts. The lowest percentages were prevailing in the eastern districts, the district of Kiri, and in the cities. The highest proportions married were in the southwestern district of Malakastra and in the coastal district of Durrës. In both of them more than 60 percent of the male population were already married at an age of 20 years. 5 years later, at an age of 25 years, 41.6 percent of the male population were married. The proportions married at an age of 25 years are mapped on map 26 . The lowest percentages were still in the eastern districts, the district of Kiri and in the city of Shkodra. In the other cities the proportion of married men were above average. The highest proportions of married men at such an age were in the western districts, the district of Oroshi in the center of the tribal region and the southeastern district of Gramshi. At an age of 30 years 59.0 percent of the male population were married. The lowest rates of married men were in the eastern half of the central part of Albania and in the northern districts. At this age all the cities had lower rates than the surrounding districts. The highest rates were still in the western districts, the district of Puka neighboring to Oroshi and in the southeastern district of Gramshi. This distribution can be seen on map 27 . At an age of 35 years 72.9 percent of the male population were married and at this age the eastern districts were no longer the ones with proportions married lower than average. The lowest percentages were in the northern districts and in the city of Shkodra. The districts with the highest proportions were not neighboring, but there is some sort of belt from the district of Shkodra in the northwest to the district of Gramshi in the southeast. Map 28 shows this situation.
The corresponding rate of men which were still unmarried at an age of 35 years is 19.6 percent. There were very high proportions of unmarried men in three corners of the country: in the northwest of the country, neighboring to Montenegro, in the northeast of the country, and in the southeast of the country. The districts with the lowest proportions unmarried were located in the center of the country and at the coast. In addition the southeastern district of Gramshi had also low rates of unmarried men at such an age. The cities had differing rates from each other, Shkodra had very high rates of still unmarried men at such an age: only half of the male population was married at an age of 35 years. The other cities had much lower rates. This can be seen on map 29 .
The percentage of widowed men at an age of 60 years was 12.7 percent, which is much lower than the share of 30.1 percent of widowed women at an age of only 40 years. The highest rates of widowed men were in the northern coastal districts and in the northeastern district of Gora. The lowest percentages were in the eastern districts with the exception of the district of Gora. Map 30 shows this distribution within Albania. These proportions were very much influenced by differing rates of remarriage after becoming widowed. The much higher rates of remarriage among men as compared to women led to much lower rates of widowed men. One can assume that the different rates of widowed men were the result of different rates of remarriage in different districts. Unfortunately there is almost no information about remarriage available in the data used for this paper. The much higher age at marriage of men also contributed to much higher rates of widowed women as compared to men, since normally the older husband died earlier than his younger wife. In Serbia in 1910 the proportion of widowed men at an age of 55 to 59 years was 17.7 percent and therefore higher than in Albania. The proportion of widowed women at an age of 40 to 44 years was 11.1 percent (Sundhaussen 1989:121) and therefore much lower than the proportion for Albania. The much narrower age gap between spouses in Serbia prevented such extreme differences in proportions widowed between men and women.

Age difference between spouses

The mean age difference between spouses was 9.6 years. The lowest age differences were in the districts at the Adriatic coast including the city of Kruja. In the district of Kavaja men were on the average only 5 years older than their wives. On the contrary the highest mean age difference was in the southeastern district of Shinapremtja with 15.1 years. In the city of Shkodra men were on average 14.1 years older than their wives. In the northernmost districts the mean age difference was also more than 11 years. You can see the regional distribution of these age differences in map 31 . These high age differences meant that in about half of the marriages the husband was at least 10 years older than his wife (46.4 percent). In map 32 you can see that the lowest proportions of women married to men at least 10 years older were in the central parts of Albania. The highest proportions were in the two regions with the highest age differences: in the district of Shinapremtja (77.6 percent) and in the city of Shkodra (72.5 percent). In general the proportions in the northern districts were higher than the proportions in the southern districts. The proportion of husbands who were at least 20 years older than their wives was 14.1 percent. The lowest percentages were again in the coastal districts and the highest in the district of Shinapremtja (30.2 percent) and in the city of Shkodra (24.7 percent). In addition three of the northernmost districts had proportions of more than 20 percent. These proportions are mapped in map 33 .

Singulate mean age at marriage

In table 6 you can see the singulate mean age at marriage for men and women by district and city.

Table 6: Singulate mean age at marriage by district and city in Albania 1918
districts
 men
 women
Kruja
 27.9
 18.4
Matja
 22.1
 n.a.
Ohëri
 n.a.
 n.a.
Oroshi
 25.7
 16.6
Puka
 24.6
 16.5
Kiri
 29.5
 19.4
Lezhja
 24.6
 20.7
Malcija e madhe
 28.2
 17.4
Shkodra
 27.2
 19.9
Durrësi
 21.5
 18.2
Elbasani North
 25.3
 20.9
Kavaja
 20.7
 16.0
Peqini North
 27.5
 19.0
Tirana
 24.3
 17.4
Dibra e poshtër
 27.3
 18.8
Gora
 26.2
 20.6
Hasi
 27.9
 19.7
Ljuma
 n.a.
 n.a.
Malcija e Gjakovës
 26.9
 17.3
Elbasani South
 26.4
 21.9
Gramshi
 24.5
 19.4
Peqini South
 26.0
 20.2
Shinapremtja
 29.5
 21.3
Malakastra
 25.3
 17.8
Berati
 n.a.
 n.a.
Fieri
 n.a.
 n.a.
Lushnja
 n.a.
 n.a.
Shkrapari
 n.a.
 n.a.
Kruja, city
 28.5
 20.5
Shkodra, city
 33.5
 22.5
Durrësi, city
 28.2
 19.8
Kavaja, city
 26.5
 19.0
overall
 26.4
 18.8

The SMAM for women by district was ranging from 16.0 to 22.5 years. The lowest SMAMs were to be found in the north of the country and in two central districts. In the four cities the SMAM was always higher than in the surrounding districts and in Shkodra it was the highest for the whole of Albania. The other districts with the highest SMAMs were in the eastern districts of Gora, Shinapremtja, Elbasani North, and Elbasani South and in the coastal district of Lezhja. The female SMAMs are mapped on map 34 . The SMAM for men by district was ranging from 20.7 years to 33.5 years. The regional pattern is not clear: The lowest SMAMs were in two coastal districts and one northern/central district. The cities had higher SMAMs than the surrounding districts. The higher ages at marriage in cities were also prevailing in neighboring Serbia. In the period from 1891 to 1910 the average urban age at marriage for men was about 4 to 5 years higher than the average rural age at marriage. The average urban age at marriage for women was about 2 years higher than the average rural age at marriage (Sundhaussen 1989: 150). The highest SMAMs were in the city of Shkodra (33.5 years), the northwestern district of Kiri (29.5 years) and in the southeastern district of Shinapremtja (29.5 years). Most of the other districts with SMAMs above average were on the borders of the tribal area in Northern Albania. The male SMAMs are mapped on map 35 . There are nine combinations of high, average, and low SMAMs for men and women and all of them were existing for the different districts of Albania. The most frequent combinations were high male SMAM and high female SMAM, average male SMAM and high female SMAM, and low male SMAM and low female SMAM. Combinations of high male SMAM and low female SMAM and of low male SMAM and high female SMAM were only seldom. The connection between female SMAM and the proportion of married women at a certain age can be seen most effectively in map 19 with the proportion of married women at an age of 18 years. A low female SMAM was clearly connected with high rates of married women at an age of 18 years.

Polygamous marriages

The majority of the Albanian population was Muslim, who were allowed to be married to more than one wife at the same time. 6.2 percent of married Muslim men were actually married to more than one wife. In the cities polygamous marriages were less frequent than in the countryside. The highest percentage was in Kavaja with 3.6 percent, while in Shkodra only 0.5 percent of the married men were married to more than one wife. This was the lowest percentage of all the districts and cities of Albania. The lowest rates of polygamy were in the southern districts, in the northwestern district of Malcija e madhe, and in the northeastern district of Gora. The highest rates of polygamy were recorded in the central districts of the tribal area, although the high rate for the district of Puka may be due to the low number of cases.
Of course polygamy was and is not allowed for Christians, but in the data entered until now we have found 13 cases of Christian men married to two wives.

Case 1: A 65-year-old Catholic of the Kurbini tribe living with 2 wives (aged 60 and 30 years) and no children.
Case 2: A 60-year-old Catholic of the Lurja tribe living with 2 wives (aged 60 and 40 years) and one son and one daughter (aged 10 and 13 years). We do not know who their mother is.
Case 3: A 34-year-old Catholic of the Lurja tribe living with 2 wives (aged 31 and 39 years) and one son (5 years). We do not know who his mother is.
Case 4: A 66-year old Orthodox Christian of the Durrësi area living with 2 wives (aged 63 and 38 years) and 5 stepsons.
Case 5: An 64-year-old Orthodox Christian of the district of Dibra e poshtër living with 2 wives (both aged 60) and one married son, one unmarried daughter, and one widowed daughter-in-law.
Case 6: A 47-year-old Orthodox Christian of the Elbasani area living with 2 wives (40 and 25 years old), a 15-year-old son, 2 daughters (10 and 4 years old).
Case 7: A 50-year-old Orthodox Christian of the Elbasani area living with 2 wives (40 and 35 years old) and a married son (16 years old) and three daughters (1 from the first, 2 from the second wife).
Case 8: A 60-year-old Orthodox Christian of the Elbasani area living with 2 wives (40 and 50 years old) and 1 daughter (15 years old), according to order in household the daughter of the first wife.
Case 9: A 50-year-old Orthodox Christian of the Elbasani area living with 2 wives (40 and 30 years old) and a daughter (4 years old) from the first wife and a son (3 years old) from the second wife.
Case 10: A 34-year-old Orthodox Christian of the Elbasani area living with 2 wives (30 and 25 years), no children.
Case 11: A 50-year-old Orthodox Christian of the Elbasani area living with 2 wives (50 and 37 years), 2 married sons from the first wife, 3 sons and 1 daughter from the second wife.
Case 12: A 60-year-old Orthodox Christian of the Elbasani area living with 2 wives (50 and 30 years old), 1 son (16 years old) - obviously the son of the first wife.
Case 13: A 55-year-old Orthodox Christian of the Elbasani area living with 2 wives (50 and 30 years old), a married son from the first wife, 3 sons and 1 daughter from the second wife.

Under the assumption that the second wife was married because the first had not born a son or no children at all, the men were successful in having at least one son from the second wife in nine cases. In three cases the men still had no co-residing son and in another case the man had only stepsons. It is interesting to see, that in four cases the first wife clearly had born a son and the man nevertheless married another wife despite the Christian ban on it. It is also interesting to see that we have found a lot of such cases outside of the tribal area with its strong emphasis on the continuation of the male line. Muslims obviously were more frequently marrying a second wife in the tribal areas than in the other parts of Albania. The regional distribution of polygamous marriages among married Muslim men can be seen on map 36 .

Inter-marriages

This paragraph was intended to deal with religious and ethnic inter-marriages in Albania, but their number is so small that it will be only a short paragraph. Only 0.04 percent of married men were married to a wife of a different religious confession and only 0.05 percent of married men were married to a wife of a different ethnic affiliation. Religious inter-marriages were recorded only in two districts of the north, in the southernmost district, and in the cities of Shkodra and Durrësi. In the majority of the cases the husband was Roman-Catholic and the wife was Muslim. In about 10 percent of the cases the husband was Muslim and the wife was Roman-Catholic. Ethnic inter-marriage was recorded mostly in those districts where there were more ethnic minorities than on average. In most of the cases the wife was Albanian and the husband was a member of an ethnic minority. Albanian men married much less frequently wives of an ethnic minority. The cities of Shkodra and Durrësi had "high" rates of ethnic inter-marriages, too.

Conclusion

The marriage patterns of women in Albania in 1918 clearly fit into the "Eastern European marriage pattern" with a low age at marriage and a very small proportion of women, who never married. The marriage patterns of men in Albania in 1918 were different, since the overall SMAM was 26.4 years and at an age of 35 years 19.6 percent of them were still unmarried. The very low ages at marriage recorded by travelers and ethnographers in the 19th century and in the beginning of the 20th century cannot be confirmed. There were many cases of young girls already being married at an age of 15 years, but they were still a minority. There might have been a rise in age at marriage since the middle of the 19th century, but the low ages at marriage recorded in the 19th century were maybe the lowest, at which marriage actually happened and no averages. These low ages at marriage might have been also some sort of ideal reported to foreign researchers, but an ideal that was only met by a minority of women in reality. The relatively high age at marriage for men is in complete contradiction to the reported low ages at marriage and to the assumption, that a low age at marriage was needed for ensuring the continuation of the patrilineage. This high age at marriage also lowers the proportion of people living in joint family households. On the other hand the wide or sometimes even very wide age gap between husband and wife ensured the male domination within the household very effectively. One factor contributing to the higher age at marriage as compared to other Balkan countries could be the existence of polygamy. Men, who were married to more than one wife, lowered the number of women available for marriage for still unmarried men. In the predominantly Christian societies of most of the other Balkan countries there was no such effect.
Albania in 1918 fits better into the Mediterranean set of tendencies in domestic group organization in traditional Europe according to Peter Laslett than into the East set of tendencies. Male age at marriage was rather high and not low as in the East set and therefore the age gap between spouses at first marriage was wide and not narrow as in the East set. There were only few wives older than their husbands and the proportion of households headed by never-married women were also very low.
Out of the regional analysis the following very general results can be obtained: In the cities the proportion of nuclear families was higher and the proportion of joint families was lower than in rural settings. The age gap between the spouses was higher in urban areas than in rural areas. This was the effect of the considerably higher male age at marriage, which was more than compensating a slightly higher female age at marriage. The northern, tribal districts of Albania were characterized by a larger proportion of joint families and lower proportions of nuclear families. Women were marrying earlier and for men at higher ages the proportion still unmarried was higher. The age gap between the spouses was higher and the proportion of men with more than one wife was higher, too. The percentage of men within the population was higher and the households contained more persons. This allows to confirm the notion of the tribal regions of being more patriarchal than the other parts of the country, but one should always bear in mind the considerable differences within the tribal districts and that for many variables one can find similar proportions in other districts, too.

References

Ansley J. Coale and Susan Cotts Watkins, eds (1986) The Decline of Fertility in Europe: The Revised Proceedings of a Conference on the Princeton European Fertility Project. Princeton, NJ.
Carleton S. Coon (1950) The Mountains of Giants: A Racial and Cultural Study of the North Albanian Mountain Ghegs. Papers of the Peabody Museum of American Archaeology and Ethnology, Harvard University, Vol. XXIII- No. 3. Cambridge, MA.
Valentina Duka (1997) Qytetet e Shqipërisë në vitet 1912-1924. Tirana..
Mary Edith Durham (1928) Some tribal origins, laws and customs of the Balkans. London.
Siegfried Gruber (2001) The "Tradition" of Many Children in Albania... and the Evidence: the Census of 1918. Paper at the SSHA meeting 2001 in Chicago.
Siegfried Gruber and Robert Pichler (2002) Household structures in Albania in the early 20th century. In: The History of the Family, Vol. 7, No. 3, pp. 351-374.
Johann Georg von Hahn (1854) Albanesische Studien. Jena.
John Hajnal (1965) European Marriage Patterns in Perspective. In: D.V. Glass and D.E.C Eversley, ed.: Population in History. London, pp. 101-143.
Karl Kaser (1992) Hirten, Kämpfer, Stammeshelden. Ursprünge und Gegenwart des balkanischen Patriarchats. Wien-Köln-Weimar.
Karl Kaser (1995) Familie und Verwandtschaft auf dem Balkan. Analyse einer untergehenden Kultur. Wien-Köln-Weimar.
Peter Laslett (1983) Family and household as work and kin group: areas of traditional Europe compared. In: Richard Wall, ed.: Family forms in historic Europe. Cambridge etc., pp. 513-563.
Cay Lienau and Günter Prinzing, eds (1989) Albanien. Beiträge zur Geographie und Geschichte. Münster.
Philip E. Mosely (1976) The Distribution of the Zadruga within Southeastern Europe. In: Robert F. Byrnes, ed.: Communal Families in the Balkans: The Zadruga. Notre Dame, London, pp. 58-69.
Beryl Nicholson (1999) The Census of the Austro-Hungarian occupied districts of Albania in spring 1918: A preliminary note on the manuscript census schedules (Centre for Scandinavian Studies Papers 5). Newcastle upon Tyne.
Eugen Oberhummer (1921) Mitteilung über die erste Volkszählung in Albanien. In: Anzeiger der Akademie der Wissenschaften in Wien, philosophisch-historische Klasse 57/1920. Wien, pp. 61-68.
Michael Schmidt-Neke and Örjan Sjöberg (1993) Bevölkerungsstruktur. In: Klaus-Detlev Grothusen, ed: Albanien (Südosteuropa-Handbuch, Vol. VII). Göttingen.
Franz Seiner (1922a) Die Gliederung der albanischen Stämme. Graz.
Franz Seiner (1922b) Ergebnisse der Volkszählung in Albanien in dem von den österr.-ungar. Truppen 1916-1918 besetzten Gebiete (Schriften der Balkankommission, Linguistische Abteilung XIII). Wien-Leipzig.
T. Selenica (1928) Shiperia më 1927. Tirana.
Holm Sundhaussen (1989) Historische Statistik Serbiens 1834-1914 mit europäischen Vergleichsdaten (Südosteuropäische Arbeiten 87). München.

Maps

Map 1: Comparison of the territory of the census of 1918 and today's borders

territory of the census of 1918

Map2: Available census data for the districts of Albania in 1918
available census data for the districts of Albania in 1918

Map 3: Population density
population density

Map 4: Percentage of non-agrarian population
non-agrarian population

Map 5: Percentage of ethnic minorities
ethnic minorities

Map 6: Percentage of Muslims
percentage of Muslims

Map 7: Percentage of Catholics
percentage of Catholics  

Map 8: Percentage of Orthodox Christians
percentage of Orthodox Christians

Map 9: Percentage of Men
percentage of men

Map 10: Household size
household size

Map 11: Percentage of people living in nuclear families
percentage of people living in nuclear families

Map 12: Percentage of people living in extended families
percentage of people living in extended families

Map 13: Percentage of people living in joint families
percentage of people living in joint families

Map 14: Percentage of married men within the male population
percentage of married men within the male population

Map 15: Percentage of married women within the female population
percentage of married women within the female population

Map 16: Percentage of married women within the female population 15-50 years
percentage of married women within the female population 15-50 years

Map 17: Percentage of married men within the male population 15-50 years
percentage of married men within the male population 15-50 years

Map 18: Percentage of married women at age 15
percentage of married women at age 15

Map 19: Percentage of married women at age 18
percentage of married women at age 18

Map 20: Percentage of married women at age 20
percentage of married women at age 20

Map 21: Percentage of married women at age 25
percentage of married women at age 25

Map 22: Percentage of married women at age 30
percentage of married women at age 30

Map 23: Percentage of unmarried women at age 30
percentage of unmarried women at age 30

Map 24: Percentage of widowed women at age 40
percentage of widowed women at age 40

Map 25: Percentage of married men at age 20
percentage of married men at age 20

Map 26: Percentage of married men at age 25
percentage of married men at age 25

Map 27: Percentage of married men at age 30
percentage of married men at age 30

Map 28: Percentage of married men at age 35
percentage of married men at age 35

Map 29: Percentage of unmarried men at age 35
percentage of unmarried men at age 35

Map 30: Percentage of widowed men at age 60
percentage of widowed men at age 60

Map 31: Mean age difference between spouses
mean age difference between spouses

Map 32: Percentage of marriages with an age difference of at least 10 years
percentage of marriages with an age difference of at least 10 years

Map 33: Percentage of marriages with an age difference of at least 20 years
percentage of marriages with an age difference of at least 20 years

Map 34: SMAM for women
SMAM for women

Map 35: SMAM for men
SMAM for men

Map 36: Percentage of polygamous marriages of Muslim men
percentage of polygamous marriages of Muslim men

1 The team of the research project consists of Helmut Eberhart, Karl Kaser, Siegfried Gruber, Gentiana Kera, and Enriketa Papa. The research project is financed by the Austria Science Fund (Fonds zur Förderung der wissenschaftlichen Forschung). I am currently employed by the affiliated project "Migration in Albania" which is financed by the Austrian National Bank (Jubiläumsfonds der Oesterreichischen Nationalbank).