PAPER PRESENTED AT THE XVth CONFERENCE OF THE INTERNATIONAL ASSOCIATION FOR HISTORY AND COMPUTING

University of Tromsø 8th of August 2003

Ole Martin Sørumgård and Arnfinn Kjelland

Putting a main part of Norwegian Local History – the old farm- and genealogical history genre – into the Computer

Content:

Abstract 1

Settlement structure in rural Norway. 2

Period covered and source material used for such bygdebok projects. 3

The method traditionally used when working on a «busetnadssoge»-project 4

Basic functionality of BSS, overview.. 4

Basic structure of BSS, model 5

Implementation of the system and database. 6

Further developments. 6

Appendixes: 6

Literature. 7

 

Abstract

Norway has an old but still flourishing genre of farm- and genea­logical histories. In fact, the genre celebrates its 100th anniversary this year – the first book was published in 1903.[1] The term used in Norwegian is «bygdebok» («-bøker» in pl.). A «bygd» is a small rural community, rarely larger than a parish. A complete bygdebok normally consist of two main parts, often with several volumes each. One part is a general history of the bygd, organized by themes or periods.

The second, upon whom this paper will concentrate, describe the history of all farms and other dwellings in a «bygd», combined with genealogical data of past and present residents. The information is usually presented in a narrative form with dwellings ordered geographically, residents arranged in «families» or other groups and listed below dwellings they belonged to. Further, each individual has cross-references to other places where he or she has stayed, place of birth etc. Because of this it is possible to trace individuals as they moved from place to place, map the pedigree of a selected person and so on. A large number of such books are still published each year in Norway. The area of a project is usually a parish, and the result is usually several large volumes.

Traditionally, manipulation and analysis of the huge amounts of historical data needed as background material are based on more or less manual, highly time consuming methods. In recent years, though, there have been some improvements, especially through faster and cheaper computers available. But each historian has still to rely heavily on his or her own skills and ingenuity in developing the appropriate tools.

It should be noted that historians working in this genre are forced to use somewhat different methods in their analysis, compared to both historical demographers and traditional genealogists. The demographers' approach yields valid results on the statistical level, but not necessarily about individuals in the material that historians and genealogists have to survey. On the other hand, genealogists often map specific families, leaving out the considerable number of individuals with no links to any known family, or no descendants. Clearly, historians in the «farm and genealogy» field have a point of view somewhere between the two others, and must in addition cover the general history of the farms and other dwellings in the researched area.

It should also be noted that there are not two such projects that are identical; they vary considerably as far as the settlement structure, size of population, amount and quality of source material etc. of the investigated areas (parishes) are concerned. Also the willingness of the project organizers (usually the municipality) to fund accurate collection, treatment and analysis of sources and the qualifications of the researcher (often a local historian without formal training in history) can vary to a great deal. The results are books of very variable quality. However, good books in this genre can generate considerable income and contribute to the funding of the project. Such projects – at least 114 are ongoing in 2002 – were funded with a minimum of NOK 21,5 mill. (EURO 2,7 mill.) in Norway this year.

Despite several commercial software packages available, serving both demographers (statistical analysis tools) and genealogists (family record programs), none of them quite fulfils the needs of historians in the field. This is primarily due to their lack of capacity to handle farms and other dwellings, and the fact that they rarely can produce references between persons and places.

The purpose of this project is to develop a useful tool for historians working in this field, a tool to assist researchers in all stages of work, including final presentation of the achieved result as a book manuscript. The title of the system is «Busetnadssoge» («Settlement history»), abbreviated BSS, or BSS-application (under Windows). This application will be able to handle all such projects, regardless of settlement structure, sources etc. in the area investigated; the only limit is population size. BSS is now tested in a bygdebok project of Volda, with a present population of 8.500 (4.556 in 1900, 3.939 in 1801).

Settlement structure in rural Norway

Due to topographical and climatic reasons the area of arable land of Norway is very limited.[2] The agrarian production up to the twentieth century was heavily dependent upon extensive use of natural resources in addition to the arable land, and this fact has constituted a settlement structure of the country rather different from the rest of Europe: In Norway single farms or farms with a few clearly distinguishable holdings are the normal pattern. Villages as in the rest of Europe are not found at all in Norway.

Agrarian settlement of this type have usually been dated back to about 0 A.D. in Norway. It grew slowly or more rapidly in periods to app. 1300 A.D. Farms occurred in several ways:

·         New, small farms were cleared in earlier uncultivated areas.

·         New, small farms cleared in central areas, but on poorer soil between older farms.

·         Older farms grew during the centuries and were divided into several one-family holdings.

The large number of deserted farms of the late middle Ages was mostly re-settled during the sixteenth and seventeenth centuries. In the eighteenth and nineteenth centuries settlement grew mainly in two ways:

·         Older farms were divided into holdings.

·         New cottar’s places were established either on cultivated land or in the outfields, in order to house a growing group of day workers and their families.

A very significant side of Norway’s rural settlement structure in the farm name system, which developed into the family name system of Norway in the first part of the twentieth century. About 70 % of the family names used in Norway today are farm names.

All farms got their names while settled for the first time, and the main rule is that the farm carries the same name today as it did 700–2000 years ago (of course with significant but traceable changes due to the general change of language from Old Norse to modern Norwegian, heavy Danish influence etc.). Usually it is simple to identify a named farm (a farm with a specific identifiable name) in rural Norway today, either as a single farm or a few holdings of which some originally may have been cottar’s places settled in the eighteenth or nineteenth century.[3]

The names of the holdings under a named farm were also almost as stable, even if changes could occur e.g. after periods when a holding had been deserted. But a more serous problem is that the name of the holding usually wasn’t used in the different government registers –almost always only the name of the main (named) farm.

For the ordinary holdings this isn’t too difficult because one usually knows the land tax of the holding and the name of the owner/peasant through tax lists, censuses etc. But when handling the growing number of cottars’ places through the eighteenth and nineteenth century this might be a growing problem, at least in some parts of Norway where the number of cottar’s places was large.

The basic administrative units of rural Norway up to 1837 were the parish and the tinglag or skipreide (lowest judicial unit). These units often, but not always, covered the same area. However, changing parish and/or tinglag borders rather often creates trouble in the collection of sources for bygdebok projects.

In the twentieth century the settlement structure in most rural areas has changed dramatically; the number of holding based upon agriculture has been reduced severely, and a number of modern houses of suburban types have been built by people employed in other kinds of production and services. Smaller and larger towns have grown up in new areas, mainly linked to modern communication lines.

Period covered and source material used for such bygdebok projects

The history of the farms in this local history genre usually describes the development from the first settlement some time between AD 0 and AD 1300 until today. But for the «genealogical» part of the books, ca AD 1600 is the usual starting point. None might find information about individuals and also a few family relations earlier, but never in any detail. From ca 1600 there are relatively complete tax lists of different types that lists the names of the farmers. From ca 1650 also other sources gradually becomes available:

·         church registers (baptisms, weddings, burials)

·         probate registers

·         a range of land registers

·         censuses

·         judicial journals with information about individuals (defendants and witnesses) and even families (cases on heritage, esp. after the ancient Norwegian odall law which gave the right to inherit the farm to the oldest son)

·         more and more differentiated tax lists (e.g. over cottars that didn’t pay ordinary tax, but have to pay a small amount to the local church)

It is usually possible to use all such sources for a parish project up to the late 1700-es. In the nineteenth century one usually have to leave out certain sources, esp. different tax lists and judicial journals.

From the beginning of the twentieth century sources are less available for other reasons, mainly privacy and the dissolution of the church’ control over family life. Such local history projects usually tries to give the same information up to the present, utilizing open sources like obituaries and other family notices in newspapers and (mainly) by compiling the necessary information between the population living in the area while the project is running.

The method traditionally used when working on a «busetnadssoge»-project

Without going into any detail, the basic method a local historian usually applies in such a project can be summarized to:

·         compiling all relevant sources for his area,

·         copying and sorting the sources that basically covers farm history in paper based folders and

·         starting in one end of the area reconstructing and writing each farm’s history.[4]

As a means of help the author might find it useful to make a scheme of the development of holdings under a named farm, including when the holding changed hands and the name of the owner/user[5] and wife/husband in the different periods, from sources organized by farms, as land and/or probate registers, tax lists and censuses (see Appendix 2).

Today most (or all) authors computerize the most important sources, the church records. The most important censuses (1801, 1865, 1900) have been computerized for the whole country, but some find it necessary to treat also censuses from 1875 and 1891, or older census lists if existing.

After having got a good overview over the named farm by drawing the «holding-scheme» and at least most of the heads of families on the different holdings, the author starts the part of the work that resembles ordinary family reconstitution: a close examination of the church registers and other relevant sources.

It is, however, necessary to stress one particular element in the record linkage process of such projects: before establishing a link the researcher should be positively sure that it is correct. Due to the fact that many of the links in the final product (the published bygdebok) will be revised by genealogists, it is better to avoid an uncertain link than include one that later may prove to be wrong. Too many wrong links in a bygdebok – esp. in the newer parts – might weaken the reputation of the book and harm the sales potential.

The reconstruction of the complete settlement structure of the investigated area (registration/­identification of all dwellings that exist or have existed in the area for some time after AD 1600) is in fact of crucial importance in all such projects whether one works with BSS or traditional methods. This structure (name list) constitutes the main index of a bygdebok of this genre, and the name of the farm usually is the most important identification mark of individuals.

In practice, this reconstruction will not be complete until the end of the project period. During the record linkage process the author will rather often find dwellings that has not being identified before, esp. from the eighteenth and first part of nineteenth century.

Basic functionality of BSS, overview

The BSS system has been designed mainly to help the local historian in handling the sources, register the dwellings, the record linkage operations and manuscript output of such a project.

As a first step, the dwellings found in the source material might be registered in one operation. Alternatively, one can start directly with the computerized source material, which of course must be re-formatted and imported before record linkage can start. Some preparation has to be done prior to searching (normalization of dates, person- and place names). Next, reconstitution of families and persons found in the source material begins, and some brief history might be written about each in the internal editor.

At the same time families and persons are connected to the dwellings where they lived (one or several). This last step can eventually be done at a later stage in the process.

Finally, the resulting material can be exported to a text file and normally further prepared in a traditional word processor or desktop publishing program before it is printed and published.

Basic structure of BSS, model

The overall structure of BSS, or more correct, the underlying database of BSS, can be divided in two abstract levels of information: the lower level of evidential data, consisting of the computerised and imported source material described above. Almost every assertion made in the process of reconstructing families and persons is based on data at the evidential level. Because of this it should remain static throughout the process, and it must be as complete as possible before reconstruction begins.

The second, higher level consists of the researchers conclusions about persons, families and places, as he works through the compiled material and investigates the historical recordings found at the lower level. Appendix 3 presents an overview of the model (leaving out most of the details).

Focus has been set on describing a few basic entities, representing the idea of Persons, Relations and Sites. A Person-entity refers to a real person who can be identified and documented in at least one source (usually several). Persons are related to each other using Relation-entities, defined to represent the relationship between two equal main persons, with or without common children. Site-entities represent dwellings of different types.

In addition, Persons and Relations can be connected to one or more Sites with the use of Group-entities (i.e. a group of persons). By means of this, one is able to describe the life history of persons from birth to death, including what relations they had at a certain time, where they resided in different periods, and migration in and out of the researched area.

At the low level, we find that the model is equipped with Event-entities to represent historical events occurred. Each event usually has statements about individuals, relationships and dwellings involved, which are to be represented by EventObject-entities of an appropriate type. Of course, the model handles all kinds of information present, like source-citations, cross-references, names and dates, age and sex of persons, remarks etc. A special entity has also been developed to take care of eventually additional (structured) data.

Some strictly informative attributes of the entities are shared among several of them. In particular, this is the case for names (Persons and Sites) and dates (Persons, Relations, Groups, Events etc.), and the full-scaled model has its own entities representing this information.

As a researcher investigates evidences found in low-level Event-entities, conclusions he make results in entities of the high level; a Person, Relation, Group or Site. Each EventObject is to be linked to one and only one entity of the high level, thus serving as a documentation of what premises have being used. In return, every conclusion about Persons etc. should in general be based on (linked to) at least one evidence (EventObject) in order to exist. This rule applies at the present only to Persons, otherwise it becomes too strictly.

In general, conclusions made should be self-evidential with respect to the facts it is based upon, and there is no need to record every single assertion made. Instead the model is provided with a Validation-entity for representing somewhat weaker conclusions or assertions that may be proved to be wrong. Then it is necessary to record the assertion made only when it differs from «certain beyond reasonable doubt».

Implementation of the system and database

BSS is implemented in object-oriented Pascal, running on Windows-system computers against a local database consisting of Paradox-tables. This solution will in most cases suffice the needs and resources available for potential customers.

Much effort has been laid down in designing a logic and self-explaining interface, with easy-to-use tools at hand when needed. BSS is intended to be flexible enough to support researchers whatever methods they prefer in their work, but there are some guidelines worth paying notice to.

A typical situation for a researcher is to search for instances of an actual person among still unlinked events, known to have resided in a specific dwelling. The selection has to be narrowed down until it contains only desired candidates. Linking together the records found produces a new Person with attributes like names and dates computed to reflect source data.

Next step could be to repeat the process with searching and linking for a possible partner to the first Person generated, thus imposing creation of a Relation between the two, eventually followed by treating common children in the same manner. The Relation (or family) should then ordinarily be connected to a Site representing the dwelling. Then, move on to the next person supposed to have resided in the same dwelling, and repeat the process. Using this approach one is gradually working throughout the material, place-by-place, person-by-person, until the desired result is achieved.

Since searching and linking heavily rely on recognition and comparison of names (in particular person and place names), one have to get rid of the many anomalies found with respect to writing. To overcome this, a tool for normalization of all types of names has been developed as a part of BSS.

Today (early August 2003), BSS is still under development. Some of the vital solutions have not yet left the drawing board. The needed routines for linking together persons and families are working, and the registration of sites and groups. What remains is mainly the connection between families (persons) and places. As mentioned earlier, BSS is already being tested in a full-scaled, actual project. So far the testing is promising. The plan is to implement what remains of the system during the test period, including a tool for producing final reports of the resulting material.

Further developments

We already have some plans about extensions of the BSS-application, in a long-time perspective. An opportunity to analyse evidences and conclusions with more convenient tools than today would be helpful, perhaps designed as a SQL-interpreter. There is also needs for representing geographical data about places (GIS-data), not accounted for today. Finally, some improvements have to be done with today's solution of importing source material from the computerised historical recordings, as well as development of tools for import and export of structured, analysed data – and maybe web-oriented presentation too.

In addition, one can think of numerous tools in order to assist the researcher in various situations. But there is a turning point where the system is growing too complex, that problem has to be avoided.

Appendixes:

1: The history of a holding/dwelling. «Bygdebok for Lesja» Vol. 2 (1992) page 162f.

2: Scheme, development of holdings and owners/users of a Norwegian farm ca AD 1600–1900, example

3: The BSS model – simplified overview

Literature

Alsvik, Ola: Local history in Norway. At the web site of The Norwegian Institute of Local History: http://www.lokalhistorie.no/English/local-his.html.

Gissel, Svend et. al.: Desertion and Land Colonization in the Nordic Countries c. 1300–1600. Almqvist & Wiksell International, Stockholm 1981.

Johansen, Hans Christian: Urban social and demographic reconstitution. The case of eighteenth century Odense. In History and Computing. Vol. 11 1999: 115–128.

Kjelland, Arnfinn: One Tenant, Several Landlords. The Land Tenure System of Norway until ca. 1800. Paper for the pre-conference «Land, Labour and Tenure: The Institutional Arrangements of Conflict and Cooperation in Comparative Perspective». The University of Leicester, England, August 21–24, 1996.



[1] Ocar Albert Johnsen: «Hurum Herred. En historisk-topografisk beskrivelse». 1903. See more details of the genre by Ola Alsvik.

[2] A more detailed description is found in Kjelland 1996.

[3] The system is described in in Gissel, Svend et. al., Chapter 4: Methods, Hypothesis and Study areas. By Eva Österberg.

[4] Despite the large differences in settlement structure between rural Norway and urban Denmark, the method has some resemblances to the one described by Hans Christian Johansen.

[5] In Norway a relatively large group of peasants owned their farms without interruptions from the Middle Ages. The rest of them bought their farms from private or government landlords and became owners in the period ca 1650–1850.