|Title:||Marriage and Names Among Immigrants to Minnesota|
|Publication Info:||Ann Arbor, MI: MPublishing, University of Michigan Library
This work is protected by copyright and may be linked to without seeking permission. Permission must be received for subsequent distribution in print or electronically. Please contact firstname.lastname@example.org for more information.
Marriage and Names Among Immigrants to Minnesota
vol. 1, no. 2, November 1998
Marriage and Names Among Immigrants to Minnesota
This article presents empirical evidence pertaining to the debate about idea of "the melting pot" vs Handlin's thesis about "the uprooted", my aim being to measure to what extent Nordic culture was preserved among the Nordic immigrants to Minnesota. The Minnesota census samples (The Ipums) are used to analyse to what extent immigrants to that state intermarried, even if they had different ethnic or national background. Also, onomastic research shows on the one hand that first names do undergo fashion cycles, on the other that racial and ethnic groups choose from more stable and limited name pools than the national one. Both marriage patterns and names can, therefore, be used as cultural indicators. In order to paint a more longitudinal picture, the behaviour of both first and second generation immigrants is studied.
A basic name study technique is to count the number of names with a specific Nordic form or origin. However, to get a more detailed picture of name patterns in the different ethnic groups, grouping individuals who hold the "same" name is necessary. This is done by listing the rank and number of sampled immigrants reporting each name, while comparing with corresponding figures for their children born in the US, and with Americans of native stock. In order to measure more exactly to what degree the various ethnic name pools differ from that of the native-born Americans, we can use the index of dissimilarity. I have also attempted a multivariate analysis of factors potentially influencing the change in name patterns between the immigrants and the next generation.
Swedish immigrants Gertrud Ryd and Oscar Aronson were married in Norwood, New Jersey in 1928. He originally came to stay with friends in Chicago in 1923, the same year she emigrated to join her aunt and uncle in New Jersey. In 1924 Oscar travelled to New Jersey and moved into her relatives' house. During the next four years they managed to build a house of their own in Norwood. Gertrud and Oscar had agreed to marry already before they left Sweden. She did not care to wait for him to send the promised pre-paid ticket, on the other hand she did not dare follow him all the way to Chicago. Eventually, he consented to leaving his well-paid job in Chicago in favour of joining her and her family.
This article investigates the selection of spouses among immigrants as an indication of how quickly and to what extent they were integrated in American society. For nearly a century the theory of the melting pot has been debated in American historiography. Opponents of this theory, most notably Oscar Handlin, argue that the immigrants were uprooted, that they were holding on to but still losing their European identities. In letters, memoirs and literature each camp find cases in abundance which they use to prove their differing viewpoints. The fact that one quarter of the Nordic and more than half of the Italian immigrants returned to their fatherlands, in itself shows the mixed feelings we expect to find among migrant people. This author believes that quantitative information is a crucial basis for testing the melting pot theory. When starting this research, I hypothesized that both views would find some numerical support. On balance, however, the high rates of return migration, the laments over homesickness in immigrant literature, and the blossoming in the US of immigrant organizations and foreign language newspapers, initially led me to think the quantitative evidence would be in favour of Handlin's hypothesis.
No other source material can match the US censuses in their representativity and coverage of immigrants to the country from 1850 onwards. In the censuses two variables stand out as indicators of the integration process. One indicator is the tendency among immigrants to choose foreign as opposed to US names for their children. To what degree did they stick with first names from their vernacular, or did they prefer the Anglo-American name pool? The other indicator is the degree of intermarriage between partners from different national or ethnic groups. High intermarriage rates (i e well above 50%) should indicate a functioning melting pot, where people rapidly integrate into the American society. This is a two-way cause-and-effect process. Firstly, researchers see high intermarriage rates as a sign of contacts and assimilation between different ethnic groups. Secondly, mixed marriages may produce multi-ethnic offspring, paving the way for further assimilation over the next generation. For instance, one theory lists intermarriage as one out of seven stages in the assimilation process (Gordon 1964). This is disputed by researchers who think degree of assimilation more depends on the ethnic components in the upbringing of the children (Hutter 1990: 147).
Low intermarriage rates can be interpreted as a sign that immigrants primarily socialize with their co-ethnics so that integration levels will remain low for generations to come. As a case in point, intermarriage rates for blacks and whites in the US still stay below one percent. Low rates can also support the in-between hypothesis of the divided heart (cf Skårdal, below) according to which immigrants are divided in their allegiances to the old and the new world respectively. Alternatively, high intermarriage rates would support the newer "salad bowl" version of the melting pot theory, according to which people keep their distinctive stamp even after different ethnic groups have been integrated
Immigrants came from many strands and differ with respect to both physical and mental characteristics. Here I want to focus on the effect of cultural differences between immigrant groups, such as language, religion and customs, rather than differences in physical appearance. Accordingly, I limited my scope to immigrants from the main contributors among North European nations. Germany, and Ireland contributed the highest emigration totals. Relative to population size, however, Norway came second only to Ireland, with nearly one million emigrants. More populous Sweden sent 1.2 million of its inhabitants, Finland less than 400000, surpassing Denmark in emigration totals after the turn of this century.
The bulk of Nordic immigrants settled in the Upper Mid-West, and I chose to focus on Minnesota, since this was the primary destination for Swedes and Norwegians and had an even higher influx of Germans. It was only after it became a state (in 1858) and the civil war had been brought to an end, that the number of settlers peaked. Increased settlement was a consequence of the Homestead act, whereby any grown man could get land for a farm at a nominal price. In line with their history as early emigrants to the US in general, the Germans were the first group to claim land in the state, concentrating their settlements in the easternmost parts towards Wisconsin. The Norwegians came second to Minnesota both in order and numbers, and settled mostly to the south and the west, extending their farming into Iowa and the Dakotas. The bulk of Swedes came third, concentrating settlements in-between the Germans and the Norwegians, particularly to the north of the Twin Cities. The late-coming Finns are primarily found up north by Lake Superior and on the Iron Range. Most immigrants lived on farms, but each group soon formed significant urban elements (Ostergren 1988: 14). When the agricultural frontier was reaching its limits in 1890, the census reported about 750 000 Germans, 250 000 Norwegians and some 180 000 Swedes. There was only one Dane for every six Norwegians. In 1920 Finns outnumbered the Danes by 50%, but the former still numbered less than one third of the Norwegians.
.02. Countries of origin
Denmark, Norway and Sweden make up Scandinavia, which together with Finland and Iceland comprise the Nordic countries. Even if Norway/Sweden on the one hand and Denmark/Iceland on the other were joined in unions and Finland belonged to Russia until 1918, emigrants from these countries originated from five different nations and ethnic groups. Since the Finnish language belongs to a separate language family and the Finns have a special in-migration history and culture, they are more ethnically distinct than the other four Nordic nationalities. The ethnic variety of Northern Scandinavia is, however, not considered fully in this paper. Up north, the Sami formed a distinct ethnic group across national borders, and the Finns constituted significant minorities in Norway and Sweden. The number of Sami emigrants was probably too small to represent a problem - the Sami language is not listed in the US census results. The Finns could cause bigger errors, because they ethnically belonged to Finland, even if they rightly stated their place of birth as Norway or Sweden. The number of Finnish speaking persons from these countries is, however, insignificant in the census samples for Minnesota. This is probably correct, since we know that many Finns emigrating via Norwegian harbours were in fact born in Finland. As could be expected, persons from the Swedish minority in south-western Finland turned out to be the biggest ethnic group born outside their country. Nearly 8% of Minnesotans born in Finland reported Swedish as their mother tongue in 1920. In all probability most of them were ethnically Swedish, and have been grouped accordingly. The Germans should ideally be divided into Protestants and Catholics, but this is not feasible since US censuses never report religion. I have instead chosen to follow the language criterion consistently, including Austrians and Germans in the same group. With the exception of the Swedish Finlanders, therefore, nationality and ethnicity are interchangeable concepts in this study.
Ethnic belonging can be measured objectively with information on birth place, or subjectively by asking individuals to state their cognitive ethnicity. I shall come back to the last method, which was used by the 1980 and 1990 census takers. First, I build on the information found in the earlier censuses about birth places for each person and their parents.
.03. Marriage within and between the immigrant groups
Intermarriage is defined as "marriage of persons deriving from those different in-groups other than the family which are culturally conceived as relevant to the choice of a spouse" (Hunter 1990: 143). In my context then, the two spouses had different ethnic or national backgrounds: their parents were of different ethnic stock, or the bride and groom were born in different countries - and this was not compensated by the fact that at least one of their parents had the same national background as that of the partner. If there is no common nationality in their answers to questions about birth place, the two spouses have out-married. Marriages that report no common ethnic background are called exogamous. Intra-marriage, on the other hand, means that husband and wife have a common national background, either because they themselves were born in the same country, or because they share common nationality through their parents.
Why do some people intermarry, while others choose spouses from within their own ethnic group? On the general level, three reasons for intermarriage have been singled out (Riley 1996: 76). Among in-migrants, especially male ones, availability in the form of skewed sex-ratios is often a major background factor. For instance, most Spanish conquistadores in the new world were men, so they simply had to form alliances with Indian women if they wanted to marry. More often among women, the ambition to climb socially precipitates intermarriage - for instance as a kind of trade-off between beauty and social position. A case in point is that pretty girls from working class districts in Oslo, Norway could be invited to upper class ballroom parties provided they disguised their social origin and place of residence by coming and leaving before the other guests (Kjeldstadli 1995: 164). Physical attraction or love may be viewed as a precipitating factor in such cases, or as the third independent factor behind mixed marriages. Giving examples of the latter should be unnecessary.
The reasons why people choose not to intermarry can be even more complex, and vary extensively with time and place. Most obvious are laws and other regulations directly prohibiting intermarriages between members of different national, ethnic or racial groups. South-African apartheid is the most recent extreme example, but similar antimiscegenation legislation has been in effect in many places. Several US states to the south and west outlawed intermarriage not only between black and white citizens, but also between white people and native Indians or individuals of Asian stock. Such laws are only the formalized versions of racial prejudice among the general public, which is demonstrated when repealing the laws has little effect on the proportion of mixed marriages. Thus, public or private group pressure is in itself a forceful barrier against intermarriage. Such peer pressure may be based on racial prejudice, but also on experience with the difficulties inherent in mixed marital unions. Chances are that conflicts more often lead to divorce when partners have different cultural backgrounds, including different religious convictions, language problems causing misunderstandings etc. This has been cited as the main reason why the US west of the Mississippi tops the world's divorce statistics (Riley 1996: 73).
.04. Source material
As mentioned above, not even the immigration lists can match the censuses with respect to representative coverage of US immigrants. While the quality of both series of sources may be discussed, only the census allows for comparison with the native born population and mirrors family composition at ten year intervals after immigration took place. Therefore, the value of the census in a study of marriage patterns and naming practices can hardly be overstated.
More specifically, I use samples from the national censuses taken in the years 1900, 1910, 1920, 1940, 1950, 1960, and 1980. These records have been computerized by several agencies, and made comparable through standardization by the IPUMS and Census Projects at the Department of History, University of Minnesota (cf Ruggles et al 1995). The 1900, 1910, and 1920 censuses are particularly suited for a study of ethnic behaviour, since they report both the individual's own birthplace, and that of his or her mother and father. While state is given for the native born, foreigners' birthplaces are usually specified at the country level. Regrettably, the equally informative 1930 census has not yet been computerized, while the 1940 and 1950 ones omitted parents' birthplaces except for "sample line persons". The 1960 census reports birthplaces for parents of native-born children, but if both parents were immigrants, the mother's country of origin was unfortunately omitted.
The 1910, 1920, and 1960 censuses state each person's mother tongue. In addition, the earliest two report whether the person speaks English, and also the mother tongue of their parents. (Consistency checks of language against information on birthplace, reveal no significant problems.) Further variables of interest in a study of immigrants, are the number of years spent in the US reported by the 1910 and 1920 censuses. Duration of current marital status was asked in 1900 and 1910. Age at first marriage is given in 1960 for all persons above 14, while number of marriages can be found in 1910 and 1960. I shall come back to the 1980 census, which substituted questions on ancestry for the questions on parents' birthplaces.
Marriage patterns among immigrants in the US have been dealt with by several scholars, some of them referring to Minnesota in particular. Lowry Nelson (1943) studied the rural Wright County west of Minneapolis. He concentrated on the ten most numerous foreign born groups, making up 91.7% of all people of foreign stock. Among these North Europeans, Nelson found more than two-thirds of all marriages to be endogamous - they married within their own national group. The most numerous groups in Wright County were the Germans, the Swedes and the Finns. Particularly the first and the last of these groups intra-married to a large extent (roughly from 65% to 85%). Only the French, the Irish and the Norwegians showed more inter- than intra-marriage. Nelson explains this with their small numbers, a factor which, however, had little bearing on the very endogamous Polish minority. To isolate the effect of residential propinquity, Nelson singled out communities where one of the groups outnumbered any other nationality. The tendency towards endogamy proved significantly greater in such "majority" communities, lending weight to the hypothesis about the effect of residential propinquity.
Attempting to compensate for the lack of information on religion in US censuses, Nelson split the county into homogeneous areas according to the religious affiliation of nearby schools and churches. This method showed that only 10% of the partners originated in areas with different religions. The result can be questioned because the numbers are affected both by the propinquity factor and by an ecological fallacy: people do not necessarily fancy the majority confession. In addition Nelson admits his data are incidental, since they were collected by asking schoolchildren. But would children disclose any special or dissonant circumstances concerning their parents' marriage or nationality? All the same, his results are sufficiently clear to question the process of the melting pot in the Upper Mid West between the World Wars. With intra-marriage rates exceeding 50%, the possibility seems real that biological as well as cultural assimilation could be postponed indefinitely.
Another study of marriage patterns in Minnesota focuses on a rural part of Kandiyohi County, 85 miles west of the Twin Cities. The historical geographer Rice (1978) limited his scope to the time before 1905, since the state census from that year was the last he could study in manuscript form. Whereas Nelson's study was cross-sectional, Rice's is longitudinal, following groups of individuals through their lives. He achieves this goal by supplementing the census rolls with church registers, family biographies and various membership inventories. The Kandiyohi locality was mostly settled by immigrants from three districts in Sweden, in addition to smaller contingents from Norway and the Netherlands. It is Rice's hypothesis that the biggest and most homogeneous population groups will be the least transient: they will experience both little out-migration and few marriages involving different nationalities.
Emigrants from the same area in Sweden would predominantly found their new settlement together. This may explain why Rice finds support for his hypothesis in the Swedish sub-populations of Kandiyohi. By and large, the male population in two of these communities stayed near home well into their thirties. The men in the third community the Swedes transplanted to Minnesota, however, followed a more transient pattern, as did the women in general. Marriage patterns were consistently conservative, meaning that marriages with non-Swedes were rare, and with non-Scandinavians practically non-existent. About 70% married a partner from their own, local community. Intra-marriage rates for Lutheran church members plummeted, emphasising the church's role both as match-maker and impediment to out-marriages.
No previous study of urban intermarriages in Minnesota has been brought to my attention. Scandinavians made up significant parts of this population element, especially in the Twin Cities, as they also did in Chicago. For Norwegians in the latter city, Lovoll (1988) reports high intra-marriage rates based on the censuses. In 1870, nearly forty years after the first Norwegians settled there, 90% were still intra-married, a figure decreasing slightly to 77% in 1910. For second generation Norwegians, however, the figure was now down to 46%. Lovoll emphasises a strong Lutheran identity, a social life revolving around the church and residential propinquity as the main factors causing the high intra-marriage rates.
McCaa's (1993) findings for New York provide further background for my own results. His study is also relevant because, like me, he uses standardized US censuses on the individual level, in his case those from 1900, 1910, 1960, and 1980. In line with Rice and Lovoll, McCaa finds high levels of endogamy in 1910, although he already sees signs of increased inter-marriage - presumably due to skewed sex ratios. By 1960 high endogamy rates were the exception, mostly found among the more recent immigrant groups from Puerto Rico, Greece and Italy. Be aware, however, that changing definitions of ethnicity in the censuses may contribute to such results. In 1960 people born in the US with one native parent will be defined as natives, whereas the earlier censuses required both parents to be native born. The 1980 census is an in-between case, since now the respondents decided whether to state a mixed or homogeneous background. This may explain the somewhat surprisingly higher rates for intra-marriage in the last census. McCaa stresses the gender gap explanation behind intermarriage, since skewed sex ratios did not lead to significantly higher celibacy rates.
Skårdal (1974) in her dissertation on Nordic immigrant literature, deals surprisingly briefly with "Marriage outside the ethnic group". By stressing a general lack of parental control among immigrants, she relays an impression of widespread out-marriage. It seems literary texts often portray intermarriage as a source of conflicts across generations, although some melting-pot-like instances are cited as well. Widespread exogamy fits nicely into Skårdal's theory of a divided heart - young lovers torn between their ancestral background and new American environment. When marriage conflicts are not treated more extensively in her dissertation, the reason could simply be that this theme occurs infrequently in the literary genre - which never aimed to be representative of the immigrants in a statistical sense.
.06. Marriages in Minnesota
After the main influx of Nordic immigrants, Minnesota, according to the 1910 census, had a population of nearly two million. The first age pyramid in figure 1 depicts main population characteristics, the shaded areas showing a substantial portion of immigrants. In-migrating men swelled the age-groups below 35, causing them to compete numerically with the youngest children, and skewing the sex ratios. Even so, immigrants already formed a minority in all age groups below age 40. Mean age at marriage can be computed from census material with a special technique. In 1910 the mean marriage age was 23.8 years for women and 28.7 years for men. This difference in age at marriage compensated somewhat for the surplus of men among the immigrants. A majority of women above 25 and of men above 30 were married, and in the highest age groups only small proportions stayed single. Skewed sex ratios and mean age at marriage will be discussed at greater length below.
Figure 1: Age pyramids for Minnesota in 1910 and 1960
From the 1910 census sample I extracted the 1347 co-resident, married couples. Tabulating them according to birthplace corroborates the high degree of intra-marriage for the whole state. 85% of the married Swedish women and 74% of the married Swedish men had chosen partners born in Sweden. For Norwegians the percentages were 74 and 64 and for the Germans 70 and 59. Since the Germans came first and the Swedes last, the ethnic intra-marriage rates correspond nicely with the timing of their Minnesota and indeed US pioneer settlements more generally. Rates of intra-marriage were consistently higher for women than for men, chiefly because of the opposite trend among the native born, whose women out-married more. This is in line with contemporary sex roles prescribing household work for women, whereas men continually left home for work and leisure. People born in the "old world" may have stuck more closely to these roles, reinforcing their effect. Skewed sex ratios could also bring about male immigrant intermarriage, since around 55% of the immigrants were men, while the ratio was around 50% for the native born.
McCaa (1993) defines as non-American any US citizen with a foreign born father or mother, because they were often raised in ethnic subcultures. What proportion of intermarriage do we find in the 1910 census according to this limited definition of "American" and broad definition of "foreigner"?
As could be expected, the broad definition increases the degree of endogamous marriages both among the German, Norwegian, and Swedish immigrants, giving the almost consistent result that 4/5 of the couples had married their co-nationals. When the ethnicity of the parents is taken into consideration, most of the differences between the nationalities and the sexes disappear. Seen from a two-generation perspective, the Swedes and the Norwegians are found to marry each other more often than they chose native American partners. The Germans, however, married Americans more frequently. This may at least partly be because they had better access to partners of third generation German stock since the Germans were the earliest settlers. High rates of endogamy in a predominantly native-born US population, can only be explained by intra-marriage among second, third and later generations descending from immigrants. Persons marrying into other groups are on average one or two years younger, indicating a slight tendency that strict endogamy was loosening up. Intra-marriage was, contrary to expectation, as widespread in urban as in rural areas. This may be because of residential propinquity, since the ethnic groups tended to live in separate areas. Probably, abundant access to co-ethnic partners in the cities, counteracted the lack of respect for old social norms that I expected to find there. On the whole, the majority groups who in-married, were not very different from the out-marrying with respect to other variables found in the 1910 census.
Let us now expand our time frame and consider information from the censuses up to 1960, a point in time when most first generation immigrants had passed away. In the 1960 age pyramid of figure 1, immigrants form narrowing bands with decreasing age, and are insignificant in all age groups but the oldest. We also find the sex ratios to be normalized gradually over time, until in 1960 there was a small surplus of women, exactly what we expect to find in a native born population. Since this finding is valid across ethnic or national groups, any effect from skewed sex ratios on intermarriage disappeared over the years. On the other hand, we would expect the rate of intermarriage to have risen, since we deal with national groups that have relatively few distinguishing characteristics and who mostly talked the same language. Birds of a feather tend to socialize.
Table 1 seems to substantiate this hypothesis. In 1920, intra-marriage rates had dropped from around 4/5 to around 3/4, and for the Germans even below 2/3, and the 1910 and 1920 censuses are comparable since the questions about nationality were identical. The drop might be explained by the "All Americanism" drive after the US entered World War I, but since this was less than three years before the census, the tendency towards intermarriage probably started earlier. There is reason to be especially suspicious of the numbers for the Germans, since both first and second generation immigrants could find reason to hide their true ethnic background. As mentioned above, I have not controlled for the ethnicity of third generation immigrants.
Figure 2: Percentage of women from different national groups inter- and intramarrying, according to the 1920 census sample for Minnesota. Based on mother's and father's birthplaces.
Perhaps we can find some of the factors causing the increasing intermarriage rates by comparing endogamous and exogamous couples in the 1920 census (cf figure 2). Just as in 1910, the difference in intermarriage rates between couples who lived on a farm or in a city, was slight. Also, differences between occupation groups seem to be insignificant. Among the younger age groups we find a somewhat stronger tendency to intermarry, but the relationship between intermarriage and command of the English language was stronger. Out of 90 non-English speaking women, only three had out-married. Even if language capabilities may be correlated with other factors such as age and duration of stay in the US, this clear finding indicates the crucial role of language in matchmaking.
The role of religion is difficult to get at, because US censuses never include questions about faith. The small proportion of intermarriages between Germans and the predominantly protestant Nordic immigrants might indicate influence from religion rather than language, at least among second generation English-speaking immigrants. But the predominance of Protestantism in Northern Germany as well as the few intermarriages between Finns and Scandinavians, point in the direction of language rather than confession as the dividing factor.
In 1960, information on parental nationality was less clear than in the 1900 to 1930 censuses, since nationality was mostly given for only one immigrating parent of native born citizens. This late, the number of first and second generation immigrants was lower, so my results may be more spurious. Also, the figures differ somewhat between the national groups. While the proportion of intra-marriages in 1920 was above 2/3, in 1960 it fluctuated around 1/3 - somewhat less for the Norwegians, somewhat more for the rest. By now the majority of citizens were at least two generations removed from their ethnic roots, and the census gives little clue as to the national origin of the immigrating grand- or grand grandparents. Still, it seems natural to interpret the 1960 results as a sign of a continued trend towards intermarriage.
|SMAM - Men||29.8||28.7||27.5||26.1||24.2||23.1|
|SMAM - Women||24.5||23.8||24.7||22.4||22.4||20.1|
.07. Sex ratios
The results in table 2 can be criticised on the grounds that skewed sex ratios are not taken into consideration. After 1900, sex ratios were balanced for the native born, while there was a declining surplus of men among the immigrants. However, differences in marriage age between the sexes can compensate for skewed sex ratios and must be considered concurrently. Immigration caused a surplus of men - 6149 against 5681 women in our 0.5% 1920 census sample. Considering the three year difference in mean age at marriage, we find 2925 men aged over 26 and 2866 women aged over 23. "Removing" people older than 49 from this virtual marriage market creates a balanced sex ratio: 2025 men and 2024 women. Clearly the sex ratios were more skewed in some immigrant groups. For instance, 296 men and 226 women aged over 20 were born in Norway. In order for these immigrants to marry among themselves, the men would have to wait until they were 37, leaving unmarried 70 men aged 20 to 36. The situation was, however, alleviated by the presence of second generation co-ethnics. If we include persons with a Norwegian mother or father, we get a sample of 544 men and 436 women aged over 20. This group could theoretically get along with exclusive inmarriage, if the men waited till they were 27 or 28. Since some of these potential brides were the result of outmarriages among first generation Norwegian immigrants, we face the paradox, that outmarriages in one generation make possible more inmarriages in the next. Table 2 shows that both sex ratios and age at marriage became more balanced over time. By 1920, the skewed sex ratios among Minnesota immigrants were in general balanced by an acceptable difference in marriage age between the sexes.
.08. The use of ancestry
Following the end of mass immigration, place of birth gradually lost value as an objective ethnic marker. The most obvious alternative is to ask each individual questions about their ethnic identity. Fortunately, this is precisely the method that the Census Bureau started practising. From 1980 the Bureau discontinued the century-long practice of asking for parents' birthplaces, introducing instead the more open-ended question "What is this person's ancestry". In 1980 one, two or three nationalities or ethnic origins could be indicated, but when coding three answers, these were restricted to a list of seventeen choices. Reporting religions, such as "Jewish" was explicitly prohibited, and respondents were told to avoid general answers like "American". Several source critical points have been made about the census' new question on ancestry. Obviously, people show varying levels of consciousness about their ethnic origin. Studies indicate that educated people, people living with their parents, and living in the northern states, give more detailed and more informed answers (Farley 1991).
Lieberson and Waters (1988) analysed ethnicity in the US from the 1980 census. One of their main questions is about inmarriage rates, inmarriage being defined widely as a marriage where both partners report at least some common ethnic background. Thus, a couple has in-married even if they reported "Norwegian" as their second ethnic criterion. In the preceding paragraphs, I started with immigrants from different nationalities and looked for signs of intermarriage. Below, I start with a general picture of homogeneous marriages, while searching for the continued role of ethnic background on the regular US marriage market. The results in table 3 are, therefore, restricted to American born couples in their first marriage. The rightmost column in table 3 contains inmarriage ratios based on all women in the 1980 census for Minnesota reporting a European ancestry. If partners were chosen randomly, the result for all groups would have been 1.0. For instance, the figure 35.83 for Norwegians, means that for women reporting this ancestry, the ratio of Norwegian to non-Norwegian husbands is nearly 36 times greater than the ratio of Norwegian to non-Norwegian husbands among other married women in Minnesota. While Norwegians and Swedes had about the same level of inmarriage, we find a higher level among East Europeans. The Germans, on the other hand, had the lowest rate. With such a big group, many in-marriages would occur just by chance.
Table 3 also specifies the ratios by decade of marriage. All ethnic groups experienced less inmarriage over time. This is especially true for the East-Europeans, including a high, but unknown proportion of Jews. In general, the drop was less dramatic for the ethnic groups that pioneered immigration: The Anglos followed by the Germans, the Irish and the Scandinavians. However, the Norwegian and Swedish pioneers did not act in line with this hypothesis. Even if they came earlier, Norwegians tended to in-marry more often. Maybe Norwegians are more watchful of their ethnic identity than comparable groups? Further interpretation of these numbers is not warranted. The reliability and validity of the ancestry variable are open to discussion, and we have no information about marriages dissolved prior to 1980. Inter-ethnic marriages may be more exposed to divorce, potentially inflating the number of intermarriages in the census.
.09. Minnesota first names
Onomastic research shows on the one hand that first names do undergo fashion cycles, on the other that racial and ethnic groups choose from more stable and limited name pools than the national one. Among historic examples of explicit first name traditions, a case in point are the customs in pre-twentieth century Norway. Here few farmers dared ignore the rules laying down how each child was to inherit their grandparents' names. For instance, a farmer named Hans Olsen, had inherited his patronymic from his father Ole and his first name from his grandfather Hans. Hans Olsen's first son would normally be named Ole Hansen, so that every second owner of the farm had the same name. Such name traditions underlined the oldest son's rights to inherit the farm - the so called allodium. This tie to the farm was strengthened when the farm name evolved into the third part of the name during the nineteenth century, Ole Hansen Berg typically living on a farm called "Berg".
The immigrants met with a combination of different naming traditions and weaker succession rights in America. In addition, severed ties to the old farm and less contact with ancestors meant that traditional naming rules could be dispensed with. A fixed family name substituted the patronymic that changed with each generation. Instead, US naming traditions recommended a middle name of the same kind as the first name. Even more important were the differences between the Scandinavian and American name pools. Since English was the lingua franca between the immigrant groups and in communication with the native born, strange, foreign names could be a real nuisance. Also, the immigrants came under formal pressure to revise their names. Not only were they advised by their American neighbours and other contacts to do so, but some immigrant children starting school were literally re-baptized by their teachers. 'Aasmund' could be called 'Osmund', 'Gunnar' was renamed 'Gilbert' to keep his initial, etc. Even more drastic changes are heard of (Haugen 1953: 207).
Therefore, neither the individual immigrant nor his co-ethnics as a group controlled the naming part of the assimilation process, independently. Even so, they were under no obligation to change their first names. When the census taker came around, the family was free to report any first names with which they identified. Census takers from outside a group might have difficulties understanding and spelling their strange names. Indeed, a number of low-frequency names in the census should probably be blamed on spelling errors made by the enumerators or the data entry personnel. But perhaps we can play down this problem in early twentieth century Minnesota, where most grown-ups were literate, and a North European background was quite ordinary.
My data set consists of the 2667 different first names in the 50% sample from the 1920 census for Minnesota. Nearly a hundred of these names are marked as partly illegible, and only a few of these could be used in my analysis. Also, many records included initials which were thrown out during data extraction, since we have no clue as to what names the abbreviations represent. 4085 sampled persons, representing slightly more than 1/3 of all Minnesotans, had such signs of a middle name, and many names of foreign origin may remain hidden from our analysis here. On the other hand we can argue that any findings based on such secondary names could hardly be given more weight than what we find out about the primary names reported in extenso to the census taker. The most frequent first names were John (380) and Mary (311), while there were 1867 names with a frequency of one, counting all spelling variants as different names. Slightly less than half of the names belonged to men, and considering the skewed sex ratio, this indicates a somewhat bigger female name pool, which is a normal phenomenon.
When analysing the distribution of names there are two basically different approaches, involving distinct classification methods. Onomatologists and other researchers within the humanities, classify the names according to linguistic criteria. For instance 'Gunnar' is a Nordic name and 'Heinrich' a German name because they either originated in a certain area or have a specific linguistic form. Social scientists instead classify names according to criteria related to the name holders (Watkins et al 1994: 177). Then, for instance, Erik is a Swedish name when reported by a Swede, while Erica is American, whenever connected with an American with native born parents. When classified in this way, it is perfectly possible that the same name can be both Scandinavian, German, and English; cf an international name such as 'Nina'.
It has been argued that the latter method is more objective, since the national or ethnic origin of many name forms can be disputed (ibid). However, the grouping of individuals with the same name also entails subjective judgements. Some name forms necessarily must be put into the same group even if their spelling in the census is not identical. Then we must choose between several levels of standardization. The most moderate alternative is to only eliminate what we believe to be spelling errors. This is called the graphemic level. Usually also names with different spellings but with the same pronunciation ('Nils' and 'Niels') will be normalized, on what has been called the phonemic level. For some purposes we may want to standardize to the lexical level, disregarding sound variants of the same name, for instance different vowels. In that case also 'Nels' belongs with the two written variants above, even if it was pronounced differently. An even more radical approach is to bundle all names that have the same root together in normalization on the etymological level. By this criterion, such distant graphemic and sound variants as 'Peggy' and 'Margaret' will be grouped together. The fuzzy standardization of names, together with disputable definitions of ethnicity in the census, make it as defendable to categorize by name form as to categorize by origin of the individual.
Names must, naturally, be classified in accordance with the research questions. My aim is to measure to what extent Nordic culture was preserved among the Nordic immigrants to Minnesota. One way of doing this is to count the number of names with a specific Nordic form or origin. The twentieth century Finnish name pool is distinct, but up to about 1900 they often used Scandinavian names. The latter have a pre-Christian, old Norse origin, preserved in the Icelandic saga literature. Many of these names such as 'Idar', 'Ragnar' etc., are distinct from the name pools of other countries. Since the occurrence of such names crested around the turn of the century, they are frequent in the lists of emigrants. I coded these Finnish and Scandinavian marker names as two specific name groups. My third group is the common American name, containing names brought over by English-speaking immigrants. These can be viewed as the background noise from which I aim to distinguish the marker names belonging to other cultures. Among these a subgroup of biblical names is singled out, since many of these were used across cultures. With the church as a primary meeting place and the location for baptisms, I would expect such names to be frequent among all Christians. A fourth group of first names are German forms such as 'Heinrich', while names such as 'Ali' or 'Francois' go into an international category. Since all coding of some names can be disputed, I have arranged for the interested reader to easily inspect the list of names and codes. The absolute and relative number of names in each group are listed in table 4.
|Number of Names||Number of Persons|
Table 4 is based on a representative sample of inhabitants, not names, in 1920 Minnesota. Thus, the rightmost panels probably give representative results for the population of name holders. It is, however, as likely that the leftmost panels give a skewed picture of the distribution in the total name pool. Outside the groups of American and Biblical names, many names were infrequent ones, often used by one person only. Thus, Nordic and German marker names would probably increase their relative share of the total name pool if we could work from the whole census. Assuming higher frequencies, it is more likely that the lion's share of American and biblical names are included in the sample. The representative numbers based on persons show that only ten percent of the state's population were enumerated with Nordic or German marker names in 1920. A significantly greater group had biblical names, while more than two thirds were noted with names from the general American pool. Naturally, the marker names were used by the immigrants themselves, but how did they name their children?
Figure 3: First names by parents' birthplaces in 1920 census sample for Minnesota
Figure 3 presents the first names of the persons in the 1920 sample according to the birthplaces of their parents. Attempting to maximize the effect of foreign birthplaces, I classify as American only individuals with both parents born in the US. Also, the latest arriving immigrant group is given priority, so that the children of a Scandinavian-Finnish couple will be classified as Finnish. Figure 3 shows that the Scandinavian names were mostly used by Scandinavians, while Finnish, German and American names were more common in other national groups. The numbers for the Finns are too small to carry much weight, and the narrow definitions of American and German mixed marriages may explain the results for these names. American names dominated in all national groups: American names were nearly five times more common than Scandinavian ones among Scandinavians. From about 1870 a diffusion of old Norse names can be seen in the Norwegian censuses, starting in the cities and spreading to the country-side. This change in naming traditions may have come too late to influence the emigrants to Minnesota, many of whom left from the Scandinavian country-side to take up homesteads in the couple of decades after the Civil War.
Among the Finns and the Germans, the use of names from the vernacular is even less widespread. Finnish names must have sounded particularly unfamiliar in the Anglo-American ear, and they had just been reintroduced in the homeland. High-running anti-German sentiments during World War I, together with their early immigration is the probable background for the latter group. Reconstructing the figure with a wide definition of "American", requiring only one native-born parent, raises the number of American couples significantly, while the proportions using Nordic or German names remain stable. Thus, the predominance of American names cannot be explained by dominant native-born partners in mixed marriages.
There are significant differences between foreign born immigrants and their children. While 17% of the Scandinavian immigrants bore Scandinavian names, the proportion is half that among US born persons with a Scandinavian parent. Old-fashioned biblical names were also more common among the immigrants than their children. As many as two-thirds of the children had American names, while biblical and Scandinavian names account for around ten percent each. My data does not corroborate the hypothesis that children were given new American names when they started school - I found the lowest proportion of vernacular names among the youngest age groups. In line with the early German immigration, the difference between first generation immigrants and their offspring is much less. The tendency to baptize a higher proportion of their off-spring with biblical names, can be explained by the greater attention given to holy persons within Catholicism.
What we observe here is a tendency towards ethnicization rather than assimilation (Morawska 1994: 325). Ethnicization is the social construction of a new ethnic identity, a process whereby a national or ethnic group after migrating develops a uniquely reformed composition of cultural characteristics, rather than adopting wholesale the culture of the majority. In a name context this means that the immigrants did not simply choose names that were popular among the native-born. They rather reformed the name pool by concentrating on a specific sub-sample of American names, by changing the name forms and introducing names from the old vernacular. Such signs of group-specific naming cannot be detected with the method of classifying names that I used above.
To get a more detailed picture of name patterns in the different ethnic groups, counting individuals holding the "same" name is necessary. As discussed above, this entails some form of name normalization. I have chosen to standardize to the phonemic level, eliminating spelling errors and decorative spelling of names that are pronounced in the same way. This will eliminate some of the bias introduced by different enumerators' decisions on how to spell names. Further standardization to the root level cannot be recommended since this would block the possibility of finding specific name forms that differentiate between ethnic groups. Thus the two names 'Christina' and 'Christine' will be coded as distinct names in the data base, while forms such as 'Christena' or 'Christana' will be put together with 'Christina'. The first is standardized according to the rule about similar pronunciation, the second is considered a spelling error. The rather arbitrary basis for the latter decision is a frequency of one in the census sample, and the failure to find the name in other contexts.
The most popular names among Scandinavian and German immigrants are listed in the appendix. Panels a and b list the top 25 names of Scandinavian men and women respectively, while the corresponding lists for Germans in panels c and d are limited to 20 due to their lower number in the 1920 census sample. The first and second columns show the rank and number of sampled immigrants reporting each name, while columns four and five give corresponding figures for their children born in the US. Finally, columns six and seven list the rank and the number of persons given the same names among children born in America of native stock. In this way the panels differentiate between how popular the names were among first generation immigrants, the second generation and finally third and later generations. To what degree did the names of the immigrants differ from those of their children and from other native born Americans?
Again the changes are not dramatic among the Germans. In line with American traditions, 'Mary' comes up as the most popular girls' name instead of 'Anna'. The decreasing popularity of a name like 'Augusta', is no surprise considering its connotations. Especially among boys, some names remain popular among second, but not third generation immigrants. The same tendency towards a limited name pool conservatism can be found among the Scandinavians, with names such as 'Olga'. Really striking, however, is the increased frequency of boys with the names 'Carl' and 'Oscar', ranking as number one and four, respectively. The finding holds true for second-generation Norwegians and Swedes alike, the two names even beating 'Ole' which was the most popular name among the Norwegian immigrants. 'Oscar' and 'Carl' were the names of the Swedish kings who also ruled Norway until 1905. Thus, the naming practices in the early twentieth century were influenced more by symbolic names from the time of embarkation, than by the medieval kings' names which won contemporary popularity in newly independent Norway. The Swedish queens' names: Sofie, Viktoria and Louise had no similar influence, maybe because they were less noted in the immigrant community. Another explanation seems more likely, however. Other name studies show boys' names to be more serious and traditional, while girls get names more in line with phonetic and semantic trends of fashion (Lieberson et al 1992). Also, the increasing use of 'Anna' may have developed due to German influence. In addition, Scandinavian immigrants introduced "new" names like 'George' (rank 5) and 'Mabel' (rank 3) from the American name pool.
.10. What degree of difference?
In order to measure to what degree the various ethnic name pools differ from that of the native-born Americans, we can use the index of dissimilarity. This index sums the differences between the proportions of people having each different name in each ethnic group as compared with the native born. The index of dissimilarity is 0 when the groups compared have identical name patterns, and 1 when their first name customs are totally different. In a previous study Watkins and London (1994) found rates higher than 0.7 for Jewish males, while female rates were a little lower. The same sex difference was found among Italians, although their rates were about 0.05 to 0.1 points lower, indicating a name pool a bit more like the general American one.
|Native born||0,355||0,342||Native born||0,623||0,652|
|Native born||0,336||0,350||Native born||0,719||0,597|
My own findings for Scandinavians and Germans are on a different level, cf table 5. Scandinavian immigrants have dissimilarity indexes between 0.4 and 0.5, while their second generation and all Germans have rates between 0.3 and 0.4. This is in line with our previous findings; the initial immigrant groups had name pools more similar to the Americans. It comes as no great surprise, therefore, that the second generation immigrants have lower rates than the first generation. As we saw with the Scandinavian royal names, men tend to have more special names, giving them higher indexes than women in the same ethnic group and generation. There is one exception to this, German women had an equal or a higher index. This might be explained by their high proportion of biblical names, indicating that religion was slowing down the acculturation process.
To sum up: The name pools used by the immigrants were neither brought unchanged from Europe nor adopted uncritically from the American surroundings. The more widespread use of American names for the immigrants' offspring points in the direction of assimilation. However, a closer inspection of the name frequency lists rather tells the tale of an ethnicization or acculturation process, involving the composition of name pools specific to each ethnic group.
Theories from social psychology focus on chains of migration linking the birthplaces and destinations of migrants. They depended upon each other for information about favourable places to go, to pay for the journey, and to provide housing and work when they arrived. Such migration chains have been called "the uncle effect", since they were often created by family connections. Gertrud's reluctance to gamble with her future by leaving her aunt and uncle and joining Oscar in Chicago is understandable. The fact that he was a Swede may have been less decisive than a relationship on which trust could be built. When some immigrants intermarried, they risked not only their foothold with the national group, but more importantly with valued social connections. Such breaking up was easier for later generations firmly rooted in the American society. Even so, a sociological survey of Norwegian descendants in the Upper Mid-West show social relations between co-nationals, the so-called gossip-circles centred on their Lutheran churches, to be vital even into the 1950's (Munch 1954).
What about the theories of the melting pot and the uprooted? Our 1920 figures for intermarriage show that the melting pot was slowly beginning to function in Minnesota, but we have no proof of its effectiveness until 1960. This is too late to support the melting pot theory to be valid for the immigrants. Of course they can themselves have felt uprooted, even if their descendants were later on amalgamated. But maybe the whole debate about the melting pot versus the uprooted is beside the point as far as the immigrants are concerned? In my opinion, the social position of the immigrants is well taken by the book title A Community Transplanted about neighbours from Swedish Dalarna who built new contiguous settlements in the Upper Mid-West (Ostergren 1988). The essence is that most immigrants came in groups, joining relatives and friends from their old local societies. They were uprooted only in a limited sense, since they brought their roots along and reconfirmed old social ties through neighbourhoods, gossip-circles, and intra-marriage.
Because of skewed sex ratios, some men had to find brides among the native born women, but this involved a clear minority of immigrants. These connections were more important in a qualitative than in a quantitative sense, forming bridge-heads into the American society for friends and relatives of the intermarrying immigrant. Complimentary to the "uncle-effect" in emigration, this "aunt-effect" in assimilation made it easier for later generations to intermarry. For the bulk of immigrants and their children, however, the melting pot was primarily an economic phenomenon. While cultivating old co-ethnic connections on the one hand, they benefited economically from their integration into the American society on the other, receiving an Anglo-American education, earning higher wages in the cities and producing larger crops on their new farms. They climbed the status ladder From Peasants to Farmers as another book title goes (Gjerde 1985). This partial integration into the American society certainly influenced the name pool, by nearly eliminating specific Scandinavian names already in the second generation. By concentrating on a specific subset from the American name pool, however, immigrants were still able to assert some ethnic identity - just like attendance to churches with a distinct national or denominational background was in line with American religious traditions.
On a theoretical level, the findings about conservative marriage patterns and changing name practices can be reconciled through the employment of Tönnies' classic distinction between Gemeinschaft and Gesellschaft or community and society. The former concept denotes the informal social relations that we find in families, neighbourhoods etc, while the latter is the more formal type of social groupings found in the workplace and in associations of different kinds. There is no clear-cut distinction between the two types of relationships, and one of Tönnies' points is how one form can evolve into the other, for instance when members of a political party become friends, or when co-ethnics form cultural associations. During the first couple of generations, the recreation of social relations fostered in the old country dominated the primary social Gemeinschaft, the scene where most marital alliances were formed. In order to function more smoothly in the American Gesellschaft, it was a tolerable sacrifice to adjust some cultural expressions like the name pool, especially when put under pressure by the Anglo-American majority. The split between these social worlds makes a thesis about the divided heart arguable in a new sense for the immigrants, while later generations gradually blended social ties from Gemeinschaft and Gesellschaft in the salad bowl.
.12. Name lists
Alba, Richard D (1990) Ethnic Identity: The Transformation of White America. New Haven.
Alhaug, Gulbrand (1991) "Personnamngransking som middel til å oppspore slektsrøter i Noreg" (Name research as a method for tracing Norwegian Ancestors), in Helleland, Botolv (ed) Norwegian Language in America. Oslo. 8 [pp173-191]
Farley, Reynolds (1991) "The New Census Question about Ancestry: What Did It Tell Us?". Demography 28(3): 411-429.
Gjerde, Jon (1985) From Peasants to Farmers. The Migration from Balestrand, Norway, to the Upper Middle West. Interdisciplinary Perspectives on Modern History. New York.
Gordon, Milton M (1964) Assimilation in American Life. The Role of Race, Religion and National Origins. New York.
Hajnal, John (1953) "Age at Marriage and Proportions Marrying". Population Studies 7: 111-132
Handlin, Oscar (1951) The Uprooted. The Epic Story of the Great Migrations that made the American people. Boston.
Haugen, Einar (1953) The Norwegian Language in America. A Study in Bilingual Behaviour. Vol I, The Bilingual Community. Philadelphia.
Hutter, Mark (1990) "Introduction". Journal of Comparative Family Studies 21(2): 143-150. [Special issue on intermarriage]
Holmquist, Jude Drenning (ed) (1981) They Chose Minnesota. A Survey of the State's Ethnic Groups. St Paul: Minnesota Historical Society.
Kjeldstadli, Knut og Jan Eivind Myhre (1995) Oslo - spenningenes by. (Oslo - a city of tensions.) Oslohistorie. Oslo.
Lieberson, Stanley & Waters, Mary C (1988) From many Strands. Ethnic and Racial Groups in Contemporary America. New York.
Lieberson, Stanley & Eleanor O Bell (1992) "Children's First Names: An Empirical Study of Social Taste". American Journal of Sociology 98(3): 511-554.
Lovoll, Odd (1988) A Century of Urban Life. The Norwegians in Chicago before 1930. Northfield.
McCaa, Robert (1993) "Ethnic Intermarriage and Gender in New York City". Journal of Interdisciplinary History XXIV(2): 207-231
Mitchell, B R (1992) International Historical Statistics. Europe 1750-1988. Third edition.
Morawska, E (1992) "America's immigrants in the census monograph. Where can we who do it differently go from here?". In S.C Watkins (ed) After Ellis Island: Newcomers and Natives in the 1910 Census. New York. [pp 319-349]
Munch, Peter A (1954) "Segregation and Assimilation of Norwegian Settlements in Wisconsin". Norwegian-American Studies and Records XVIII, NAHA, Northfield, MN. [pp 102-140]
Nelson, Lowry (1943) "Intermarriage among nationality groups in a rural area of Minnesota". American Journal of Sociology, 485): 585-592.
Ostergren, Robert (1988) A Community Transplanted. The Trans-Atlantic Experience of a Swedish Immigrant Settlement in the Upper Middle West, 1835-1915. Madison.
Pryce, W T R (ed) (1994) From Family History to Community History. Cambridge University Press.
Rice, John G (1978) "Marriage behavior and the persistence of Swedish communities in rural Minnesota". In Hasselmo, Nils (ed) Perspectives on Swedish Immigration. Proceedings of the International Conference on the Swedish Heritage in the Upper Mid-West, April 1976. Duluth/Chicago. [pp 136-150]
Riley, Glenda (1996) Building and Breaking Families in the American West. Albuquerque.
Ruggles, Steven and Russel R Menard (1995) "The Minnesota Historical Census Projects". Historical Methods [pp 6 -10].
Ruggles, Steven (1995) "Sample Designs and Sampling Errors". Historical Methods [pp40 - 46].
Ruggles et al (1995) Integrated Public Use Microdata Series. IPUMS-95 Version 1.0. Vol 1: User Guide. Department of History, University of Minnesota. Newer version available via the Internet at http://www.hist.umn.edu/~ipums/.
Skårdal, Dorothy Burton (1974) The Divided Heart. Scandinavian Immigrant Experience through Literary Sources. Oslo.
Sowell, Thomas (1996) Migrations and Cultures. A World View. New York.
Steinmetz, Suzanne K (1990) Marriage and Family Realities - historical and contemporary. New York.
Watkins, Susan Cotts & Andrew S London (1994) "Personal Names and Cultural Change: A Study of the Naming Patterns of Italians and Jews in the United States in 1910". Social Science History 18(2): 169-209
.14. Tables in Appendix
|Native born Foreign born||X||Scandinavian parents||Native parents|
|Native born Foreign born||X||German parents||Native parents|
|Native born Foreign born||X||German parents||Native parents|
Top men's and women's names among Scandinavian and German immigrants compared with those of their children and Americans of native born parents. Source: 1/200 sample from 1920 census for Minnesota.
Gunnar Thorvaldsen, email@example.com