KBS, GIS and documenting indigenous knowledge

Rhodora M. Gonzalez

As a result of the growing recognition of the role of indigenous knowledge in sustainable development, we are now faced with the huge task of documenting and disseminating that knowledge, in the same systematic way that Western knowledge is generated, documented and disseminated. Rapid modernization underscores the urgency of the task at hand. At the same time, the present widespread use of computers offers a range of tools which facilitate certain aspects of the work. In some cases, this means formalizing and incorporating non-quantitative or descriptive local knowledge into a knowledge base. In other cases, it involves transforming existing databases, in order to approximate the world model of the local inhabitants.

More and more researchers, policy makers, managers and development professionals are now using computerized information systems and decision-support systems in the management of natural resources and the environment. However, these systems operate on models which are influenced by the perspective from which the world is interpreted within a certain discipline. When the ultimate goal is to place indigenous knowledge alongside mainstream development strategies, it becomes necessary to develop ways of integrating indigenous knowledge into such computerized systems. In managing their resources and the environment, local people make use of common-sense knowledge, also known as rules of thumb or heuristics. It is this expertise that must be gathered and modelled in such a way that it can be compared with Western models. This comparison is so important because many Western-trained environmentalists and resource managers end up advising on and sometimes even managing the resources of other countries and cultures. This will help them to understand the 'folk-constructed' world within which decisions are made and behavioural patterns formed (Warren, 1975; Klee, 1980).

KBS and GIS
Efforts along this line are apt to employ tools developed with the aid of advances in computer science research into artificial intelligence, specifically those pertaining to the building of knowledge-based systems (KBS). These are computer programs which contain both declarative knowledge (facts about objects, events, and situations) and procedural knowledge (information about courses of action), with a view to emulating the reasoning processes of human experts in a particular domain or area of expertise (McGraw and Harbison-Briggs, 1989). The challenge in building an indigenous knowledge base lies in understanding and reasoning with the aid of largely abstract, qualitative observations of the local environment. These include heuristics (from the Greek word heureka or Eureka, meaning: 'I have discovered'), rules that are typically less than precise and are sometimes called rules of thumb. Among indigenous peoples and local communities, these rules are passed on from one generation to the next and are gradually refined into a system for understanding the world around them: in effect, a world view (Pawluk et al, 1992). As in KBS, these rules of thumb can be represented as follows:

IF <premise> THEN <conclusion and/or action>.

Schoenhoff's 'barefoot expert' (1993) and Furbee's 'folk expert system' (1989) have demonstrated the usefulness of KBS technology in documenting and disseminating indigenous knowledge, and making possible a comparison with corresponding Western knowledge.

Another computerized tool, known as geographic information systems (GIS), has gained widespread use in natural resource management, because of its ability to perform a rapid analysis of spatial data from multiple sources, to display scenarios and to facilitate data exchange. A GIS is designed for the collection, storage, and analysis of objects and phenomena, whereby geographic location is an important characteristic (Aronoff, 1989). Information on the natural resources of an area, which is usually presented in the form of maps with accompanying tables, can now be efficiently processed using GIS-assisted procedures. Inputs for GIS include digitized maps, aerial photographs, satellite images, field survey reports and census data. However, these data sources are influenced not only by the limitations imposed by the instruments and the methods used, but also by the world view of surveyors, cartographers and others. Outputs are also presented in the form of maps which reflect delineations made by their creators. They determine which terrain features or objects are relevant, and these then appear as geometric descriptions such as shape, size and position. Thematic attributes consistent with the classification systems they use are also evaluated. These world models, which describe the natural resources of a region, then become the basis for development plans and policy formulations. Using indigenous knowledge during the survey itself ensures that the most important factors determining resource value and management practice will be captured, which will help to identify the types of development most likely to be sustained by the local population (Tabor and Hutchinson, 1994).

In many cases, development professionals have to rely on the available database and resource maps of an area. Combining GIS techniques with KBS technology can facilitate the incorporation of indigenous knowledge, which will provide a better view of the world model of local resource managers. The key lies in establishing indigenous criteria for the definition and description of local resources, and then finding a way of translating these into a similar object in the Western model; this is known as context transformation (Molenaar, 1993). It will be clear that in a database, object descriptions have their own particular context, which will depend on the aims--and world view--of the user. GIS capabilities for the storage and integration of multiple databases are eminently suited for documentation, as well as for comparing indigenous knowledge with corresponding Western knowledge. This process is illustrated in the following example, which makes use of a digitized soil map (1:20,000) and database of Neguev (Costa Rica). The town was evaluated using the farmers' criteria for describing their soils and defining potential uses.

Soil map and database of Neguev
The farmers of Neguev associate observable characteristics of their soils with the performance of their crops (Van Uffelen, 1990). Colour is an important criterium in distinguishing soil types, as indicated by the local names:
Tierra Negra: (black soils)
Tierra Bermeja: (brown soils)
Tierra Colourada: (red soils)
Tierra Muy Roja: (very red soils)
Suamposa: (swampy or poorly drained soils)

The brown, red and very red soils are an elaboration of the indications for the red soils. Other criteria used by farmers are location or altitude (high or low places), humidity (wet or dry), texture (sandy or clayey) and structure (hard or soft). Their criteria are based on colour criteria according to which high places display red soils which are dry and clayey, while the capa-dura phenomenon (hard and dry layer) is also common. Black soils, which are more fertile, soft and humid, are generally found in lower areas and near river banks. Red soils are further distinguished by their location within the terrain. This explains such local names as tierra bermeja alta, tierra bermeja mediana and tierra bermeja baja. The farmers have also evaluated their lands according to the performance of their most important crops:

As in any attempt to build knowledge-based systems, success will depend on the ability to elicit knowledge from the knowledge source (whether this is a human expert or a document). Experts know how to do this, and base their decisions on the specific circumstances, although they may not be able to articulate the reasons behind those decisions (Olson and Rueter, 1987). This is why a KBS cannot replace an expert: it is incapable of encompassing the same amount of knowledge. A KBS is designed to assist in solving a problem that requires a degree of expertise, and to make that knowledge easily available to others. This crucial stage was carried out using a combination of techniques (interviews, questionnaires, observations and triad tests) to elicit the knowledge of the farmers. Soil samples were identified by both farmers and soil scientists, in order to establish where the farmers' criteria corresponded to those of the scientists. The colour criteria were translated into colour codes used by soil scientists (the Munsell soil colour chart**1), while the various locations were translated into slope classes (Table 1).

Table 1: Summary of IK criteria for red soils articulated with scientific soil units
IF
THEN
COLORand/orSLOPE CLASS SOILTYPE
5YR 4/4andCMuy roja alta
5YR 4/4andEorDE Muy roja mediana
5YR 4/4andDorCD Muy roja baja
10YR 3/3andB Colorada alta
10YR 3/3andDorE Colorada mediana
10YR 3/3andCorCD Colorada baja
10YR 4/3andAorB Bermeja alta
10YR 4/3andF Mermeja mediana
10YR 4/3andE Bermeja baja

Legenda slope classes

A:
Level of nearly level
B:
Gently sloping or very gently sloping
Undulating or gently undulating
C:
Sloping or strongly sloping
Rolling or gently rolling
D:
Moderatly steep (single)
Hilly (complex)
E:
Steep
F:
Very steep
Formulated as IF/THEN rules, the farmers' criteria guided the use of GIS capabilities in processing both descriptive and locational soil data. An ancillary programme was created to automate the procedure. Individual soil units on the map were easily reclassified to incorporate the farmers' knowledge of their soils. The resulting map was further evaluated by means of their crop performance criteria, which were translated to fit the transformed soil database. For example:
IF <the soil is very red> IF <colour = 5YR 4/4>
AND <is located on high ground> AND <slope = C OR E>
THEN <it is good for pineapples> THEN <crop group =3>.

GIS functions create maps out of the stored information. These can then be displayed, to show spatial distribution (see figure 1) and to allow comparison with the original scientists' soil map. In this way additional similarities and differences between the farmers' classification systems and those of the scientists are obtained.

Conclusion
The combination of different types of information--whereby the scientists' wording and the farmers' practical experience are integrated--provides a more complete description of the soils. This is helpful in defining appropriate development interventions capable of enhancing sustainable agricultural production. The information provided by local resource managers results in policies which are more sensitive to their conditions. When their information was incorporated into computerized decision-support systems and the resulting maps were displayed, the farmers were still able to recognize their inputs, which gave them confidence in their ability to solve problems. The ability of GIS to create scenarios out of stored information has yet to be developed, but this promises still more possibilities for the management of local resources and the environment. Most important of all, farmers and scientists are learning from each other's expertise, which leads to the validation and enrichment of both scientific and indigenous methods--a useful contribution in the continuing search for sustainable development alternatives. Furthermore, the availability of this expertise in computerized form makes it easier to disseminate and replicate in similar situations. Although much more than commitment is necessary for the mainstreaming of indigenous knowledge, development workers dedicated to finding new ways to document and disseminate indigenous knowledge could harness the available technology, such as a GIS-KBS link-up, and benefit from the advantages it offers.

Rhodora M. Gonzalez
Wageningen Agricultural University
Department of Land Surveying and Remote Sensing
Hesselink van Suchtelenweg 6
6703 CT Wageningen
The Netherlands
E-mail: gonzalez@itc.nl

or

Philippine Rural Reconstruction Movement
940, Quezon Avenue
Quezon City 1103
The Philippines

References
Aronoff, S. (1989) Geographic information systems: A management perspective. Ottawa: WDL Publications.

Furbee, L. (1989) 'A folk expert system: Soils classification in the Colca Valley, Peru', Anthropological Quarterly 62(2):83-102.

Klee, G. (1980) World systems of traditional resource management. London: Winston & Sons.

McGraw, K. and K. Harbison-Biggs (1989) Knowledge acquisition: Principles and guidelines. New Jersey: Prentice-Hall.

Molenaar, M. (1993) 'Object hierarchies and uncertainties in GIS or why is standardization so difficult?', GIS 4(6):22-28.

Olson, J. and Rueter, H. (1987) 'Extracting expertise from experts', Expert systems 4(3):152-168.

Pawluk, J. et al (1992) 'The role of indigenous knowledge in agricultural development', Journal of soil and water conservation 4(47):298-302.

Schoenhoff, D. (1993) The barefoot expert: The interface of computerized knowledge systems and indigenous knowledge systems. Connecticut: Greenwood Press.

Tabor, J. and C. Hutchinson, (1994) 'Using indigenous knowledge, remote sensing and GIS for sustainable development', Indigenous Knowledge and Development Monitor 2(1):2-6.

Van Uffelen, J. (1990) Conocemientos endogenos y cientificos para la determinacion del capacidad de uso del suelos de Neguev. Costa Rica: AZP.

Warren, D. (1975) Indigenous knowledge systems for activating local decision making groups in rural development. Iowa: Iowa State University.

Endnotes
**1 A system of colour notation devised by Albert H. Munsell that identifies colours in terms of three attributes: hue, value and chroma. This system was further developed by scientists and has gained wide acceptance.


Back to: top of the page | Contents IK Monitor 3(1) | IK Homepage
Suggestions to: ikdm@nuffic.nl
(c) copyright Nuffic-CIRAN and contributors 1995.