Stéphane ROCHE

Associate Professor

CARTA/CNRS - UMR ESO 6590 - UNIVERSITE D’ANGERS

35, Rue de la Barre - 49100 Angers - France

Tel : +33 2.41.36.54.48 - Fax : +33 2.41.36.54.46

I certify that this paper has never been sumbmitted anywhere else for publication.

19/04/99

S. Roche

GIS, tools or social construction? This question amplifies the title of a forum published in the Annals of AAG Nr 2, 1997. Its reality and relevance are no longer challenged at present, when GIT are expanding more and more among local governments (Roche & Humeau 1999a). Yet, very little is known, at this stage, about the building up of GIS and the elements taking part in their social development.

An increasing number of research programs about this topic are undertaken. For example, one of the four research targets encouraged by the Centre for Research in Geomatic of Laval university in Quebec deals with the geomatisation of organizations. This target includes, among other topics, organizational, political and legal aspects of GIS. In France, the new orientations of the research program on GIS of the GDR Cassini open new tracks for reflection on the sociology of GIS. The latest programs of GISDATA focus their investigations on the processes of implementing, expanding and adopting GIT. Some initiatives of the NCGIA are also focussing on the human and social implications of GIS - Initiative 19 (Harris & Wiener 1996), decision and negociation making - Initiative 17 (Densham & al 1995) and the problems of spatial representations - Initiative 21 (Mark & Hengenhofer 1996).

In the French-speaking countries, other authors studied these problems. Several recent studies have explored the implementation processes of GIT in organizations. At the CRG of Laval university, C. Caron (1997) first worked out a descriptive framework of geomatisation projects. He accurately describes the elements interfering in the GIT implanting procedure in organizations. At the EPF of Lausanne, H. Pornon (1997) studied the expanding process of GIS in various organizations and their impacts on power systems. At Angers university, S. Roche (1997) made research on the social construction of GIS in comparative case studies located in French and Quebec towns.

The results of these works show a common element : every GIS is to be treated in its specific social context. Each project is unique and can be understood only in reference to the context in which it is located. "GIS is not simply a tool for processing geographical information. It is a social technology incorporating an entire institutional and intellectual infrastructure that delivers and markets GIS. As such its development cannot be understood apart from the social context in which this occurred" (Sheppard 1995).

However very few researchers clearly refer to the spatial context. The majority of these studies considere the problem as if local and national spatial contexts were not important. As if GIT were in every way similar to information technologies. Besides, studying GIT implementations through the simple relationship with their implied results appear to be incomplete. A GIS is not neutral in the organisation concerned, and various elements of the social context interfere in its development. Sheppard (1995) and Chrisman (1997) insist upon the reciprocal interactions between a GIS and its institutional and cultural environment. It is therefore necessary to analyse the social adoption of GIT in a geographical perspective.

Actually, the relationships between various planners, the local spatial context and GIS tools in a local government are not well known. The aim of this paper is to give evidence of those relationships in the light of four comparative case studies. What are the determinant social, organizational and spatial variables for the social adoption of GIT? As a consequence, our reflection will deal with the socio-spatial embedding of these new technologies and their process of social construction.

After a short account of the methodology adopted, this paper will attempt to show all the importance of spatial representations in the adoption or rejection of GIT by the planning stake-holders (technicians, urban planners, elected members). It will also attempt to give evidence of the determinant character of the close interrelations linking the various perceptions of the stake-holders in the social adoption of GIT.

The results established here specifically proceed from a "science in action" attitude, following the concept developed by B. Latour (1989). This methodology is applied to four comparative case studies : Nantes, an urban French city of 250 000 inhabitants ; Mayenne, a rural small French town of 15 000 inhabitants ; Quebec, a French Canadian city of 220 000 inhabitants and Charny, a small French Canadian town of 15 000 inhabitants. In each case study, the five research variables (fig. 1) were examined and, above all, how they relate to each reference group of stake-holders. In each interview, an analysis of the discourse has been carried out by classifying the stake-holders’ replies in relation to the various variables. The main target was to put into perspective the various discourses and give evidence of the contradictions or similarities between the stake-holders - divergences, convergences, absences or hyper-involvements. The vocabulary was identified in order to define the concepts and carry out a comparative analysis of the definitions proposed. The following analyses attempt to characterise the criteria involved in the participation of the various stake-holders in the social construction of GIS.

Regional planning activities are perceived by the stake-holders interviewed in relation to their mandate, educational background and professional culture. Typically, technicians view planning activities as the building of material infrastructures onto an empty space (construction of houses, new roads, creation of parks; etc.), while urban planners view it more as a theoretical reflection on space, its functions, the relationship between spaces and societies. Local authorities (the elected members) consider it as an essentially political activity when multiple stake-holders interract and negotiate. For example, they emphasize the necessity to regulate the uses of their municipal territory in the heart of economic and political stakes.

Similarly, regarding spatial representations, technicians perceive the municipal territory as a material and physical space deemed to comply with the daily needs of its inhabitants. Urban planners consider it more as a space for theoretical analysis and studies. The local authorities view their municipal territory as a place where various stake-holders’ strategies occur as well as a variety of cultural, social, economic events. The manner in which the various stake-holders perceive the role of planning activities therefore proceed from their own spatial representations (fig. 2). However, differences appear according to the municipal contexts explored.

On the one hand, the differenciations noted are not as important in smaller municipal organizations like Mayenne or Charny as in larger structures. Quebec and Nantes are very hierarchised and complex municipalities. There, the activities of the various stake-holders are clearly defined and differentiated from one another. Everyone's mandate focuses on a specific target. The "planning schedule" effect of big towns generate a very fragmented working organization. Thereby, some stake-holders have a fragmented vision of planning activities. Even inside each group, perceptions appear to be different. An urban planner in charge of checking land tenure and administration rules applications does not perceive planning and the municipal territory in the same manner as the one whose work consists of actually thinking about the main orientations of urban development. Local authorities themselves have in charge specific themes bearing some consequences on their conception of planning. On the contrary, in more modest organizations, the stake-holders have to deal with different things. Their activity is less fragmented. The differences in perceptions related to the mandate tend to fade slightly.

On the other hand, the stake-holders' professional mandate clearly has more influence than their educational background on their perceptions. For example, in Nantes, some professionals in charge of the social development of urban areas are trained as surveying engineers. However, their perception of the municipal territory is much more theoretical than the one of some urban planners whose activity amounts to controlling the land tenure process.

Besides, in Quebec towns, a territory is perceived by most stake-holders as a "consumer goods" meant to adapt to the needs of its population and their spatial and temporal evolutions. This Anglo-Saxon cultural heritage is absent from French towns where the territory is perceived as a historico-cultural patrimony destined to be preserved. Here again, this phenomenon is setting bounds to the differences between the various groups' perceptions.

Finally, these differences appear to be strongly linked with the local geographical context. The impoverishment of the central areas of Quebec, the saturation of the municipal territory of Charny, the social and economical destabilisation of the town of Mayenne or else the "disorganization" of the urban areas of Nantes deeply impresses the spatial perceptions of the stake-holders of these towns. These local geographical characteristics also influence their vision of planning activities. Planning is defined by the different stake-holders in reference to local planning problems that they deem of major concern and more precisely to their mandate in the actions undertaken in order to solve those problems.

Whoever the stake-holders and whatever the cases might be, the perception of the usefulness of GIT is very positive, apart from those who consider that the development and large implementation of those tools will constitute a major risk for their statute and function in the municipal organization. The case studies show that the perception of the usefulness of GIT is strongly influenced by the professional mandate. As in the case of spatial representations, these differentiations are more perceptible in Nantes and Quebec than Charny and Mayenne. The stake-holders interviewed mention the usefulness of geomatic tools in reference to their own role in the municipal planning. Technicians consider GIT as spatial data processing and mapping tools. Planners, in a more theoretical way, consider them as means to cross-compare multi-source and multi-topic data. The aim is to extract spatial representations helping in the understanding of urban dynamics. Local authorities perceive GIT more as media of explanation and communication (fig. 2). It is particularly true in Nantes and Quebec where the new policy of urban development is generalizing public consultations of the citizens.

The differentiations mentioned above are progressively fading with the level of direct use of GIT. Two common elements therefore appear in the four cases : (1) the development of GIT entails a larger spreading of maps, even among stake-holders who did not traditionally have access to them ; (2) the time saved in the various steps of data production allow the stake-holders to make trials and find more time to reflect and compare various scenarii.

The main and often the only use of GIT is the production of maps and plans. But this situation reflects two different realities, particularly in towns where the degree of use is the highest and the most generalized :

2) On the other hand, the indirect users (i.e. planners and a few elected members) perceive and attempt to use GIT to produce real models of the territory or the urban area they have to study. To this effect, beyond the physical dimensions, they expect GIT to enable them to compile and compare various informative elements in order to produce true spatial representations closer to territorial realities and their dynamics, networks, population mobilities, etc. But due to the lack of socio-economic data and their difficulty in handling these new tools, the planners remain more reserved about their ready usefulness.

The first notable difference between the French and Quebec municipalities concerns the concepts and the terminology used. The municipal stake-holders of Quebec and even Charny (where the level of diffusion and use of GIT is somehow quite low) own a more extended theoretical knowledge on the usefulness, potentials and limits of geomatic tools than their French homologues. However that quite deep knowledge does not mean that it is always levelled with a degree of practical use of GIT more important than in France. In that aspect, the discrepancy is real.

The second difference in the practical use of GIT is linked with the size of the municipal organizations. In Mayenne and Charny, only a few technicians use GIT in a direct and autonomous manner. The professional specificities mentioned above are not perceptible, which is not the case of Nantes and Quebec where the use of GIT is much more widespread and the non technicians in particular enjoying a high degree of autonomy. GIT are perceived as tools offering extensive help in the understanding of the functioning of a town and the territorial problems. According to the stake-holders interviewed, GIT enable to vary the scales of observation, build various hypotheses of development and test various planning scenarii.

These statements show that the large majority of the stake-holders have a positive perception of GIT. Nevertheless a few differentiations remain which are linked to professional cultures and specific themes studied. Generally, the perception of GIT is very good and the stake-holders interviewed are either already users or potential users. Those who are regular and active users are satisfied, the others wish to increase their level of autonomy.

The geographical information is used in three major fields :

• for consultation purposes in order to draw a portray of a territory.
• being the basis of various planning scenarii, it therefore constitutes a support to decision-making (even if no simulation tool, expert system, multicriteria analysis, etc. is available). This second field of use, mainly mentioned by planners and local authorities, is complementary to the first one.
• the digital geographical information constitutes a help to debate and account for planning projects to third persons. This use is only mentioned by local authorities, planners at a strategic level or those who have to communicate the projects to the population.

The majority of the stake-holders interviewed consider that the quality of the digital geographical information (its aestheticism, the use of colours, the ways to represent, the possibility to choose different scales, etc.) is definitely better. This information is of easier reach, more supple, more adaptable to different uses and more readily available. It enables to overcome the traditional segmented topical categories and offers more reliable multi-scale and multi-topic representations. According to the stake-holders interviewed, the information is more faithful and objective with the GIT. This greater credit brought to geographical information is closely linked to the symbolism and added value currently granted to new geographical information technologies.

On the whole, the stake-holders consider that the digital geographical information is highly representative of the territorial reality it is supposed to describe eventhough many of them regret that the use of these new tools gives rise to the standardization of mapping production. Some stake-holders even confess that they no longer feel the need to go to work-fields to get a good idea of the context of the study. The digital geographical information is therefore quite well perceived. They consider it as good quality and faithfully restituting the municipal territory. They grant it with an almost total confidence. However, this statement deserves to be slightly nuanced.

The stake-holders interviewed in Quebec and Charny are less unanimous about the quality and reliability of this information than the French are. This difference might be related to the level of theoretical reflection, higher in Quebec than in France and the history of the development of geomatic in that country (Roche 1997).

Other differentiations appear, less linked to the place (France or Quebec) than to the size of the municipal organizations. At Nantes, but above all at Quebec, opinions are more moderate. The planners are of the opinion that if the information is exactly reflecting the physical and material dimensions of the space, it is however more limited as regards the third and fourth dimensions (time) such as mobilities, mobilities, urban space developments, networks : those dimensions remain essential for a good understanding of the life in urban areas. In the same way, urban planners confess that the quality of the information is not uniform for all the areas of a municipal territory. They regret spatial and topical lack of balance (often reflecting the political priorities of a municipal team, for example the unsatisfactory representation of the northern urban areas of Quebec in favour of the central areas).

Besides, planners regret as well that the information is not always updated due to the complexity of the GIS co-ordination and maintenance structure. By searching universality and attempting to meet all the expectations, the information ends in not corresponding to any specific need. Such is the case for the corporative maps of the town of Quebec. Those critics are precisely more scathing in Quebec where the GIT are more advanced than in Nantes.

In Nantes and Quebec, other modes of use of the geographical information are now appearing. Many planners perceive the spreading of the digital geographical information as a means to expand wider the spatial approach among professionals who were not yet using it. Some planners are of the opinion that it is a new and more representative way to approach and understand territorial phenomenons previously more difficult to understand via non spatially localised statistical data.

It appears that the perception of the role of planning activities (2) is linked mainly with the stake-holders’ mandate in the municipal organization (6) and their spatial representations (1). The above results also show to what degree the perception of the usefulness of GIT (3) depends on the stake-holders’ basic professional training. It is mostly conditioned by the function exercised (for instance two professionals of identical educational background have different perceptions depending on whether they work for urban planning or for land tenure and administration). We found out that variable 3 also depends on the perception of the role of planning activites in which stake-holders are involved. These analyses also reveal much the stake-holders’ sensitivity to the symbolism of new technologies is an important element in their perception of the quality of digital geographical information (4).

Two other variables also influence the stake-holders’ perception of GIT. The degree of theoretical knowledge of GIT (7) is determining. In Nantes, but mostly in Quebec, where GIS are more developed, the highly trained stake-holders are undoubtedly the most critical and lucid about the inputs, of these new technologies. The perception of the potential implications of the diffusion of GIT (8) on the stake-holders’ position in the organization is also conditioning the perception of these new tools and consequently these stake-holders’ degree of use and appropriation. The technicians of Nantes observatory perceive in the diffusion of GIT from the Direction of GI a risk for their autonomy, their mode of functioning and their position of information producers within the general direction for planning. In fact, they are relatively hostile to the tools offered and their mode of adoption is one of rejection (a form of resistance to change).

Finally, the mode of adoption of GIT (5) by stake-holders depend on their perception of the usefulness of GIT and the quality of geographical information. It also depends on their level of knowledge (which may encourage them to extend their use or the contrary) and their perception of the implications of the GIT development in their departments. As a consequence, the perceptions of the functions of the territory and of the role of planning therefore participate in the perception of the usefulness of GIT and of geographical information and actually condition the social adoption of these technologies.

The identification of the relationships between the various perceptions of the municipal stake-holders enabled us to better understand the effective role played by GIT in the municipal planning activities (Roche et al. 1996). We could therefore note four modes of social adoption:

1. GIT are the modes of expression of a new technical power. They assist in reintroducing the technical dimension of planning projects in the debates. They somehow complexify geographical data processing. In that way, they offer to those who control those technologies (the technicians but also some planners who use them directly) an important position in the planning reflection process.
2. On the other hand, these technologies are perceived by the planners as agents which facilitate the communication and the negotiation between stake-holders. They allow each one - whatever his job - to recover, through the spatial representation displayed, his own perception of the municipal territory and facilitate the dialogue.
3. GIT also integrate the logics of the various stake-holders (each one according to his duties) as the instruments enabling a better influence of the decision-making process. It is particularly the case for the local authorities but also for some planners at a strategic level who have in charge to show, explain and defend the municipal projects to the population.
4. More generally, these tools allow each one to defend his vewpoints and aims. The GIT, due to the symbolism referred to them are perceived as tools bringing to geographical information further objectivity and reliability than traditional techniques. They are therefore perceived as ensuring more accuracy, less possible contest.

The results obtained in this study give evidence of very close relationships between the various stake-holders’ perceptions (of the territory, of their role in land planning activities, of the usefulness of GIT, etc.) and their degree of implication in the use of GI Tools. This study also reveals that the links between the five research variables generate modes of GIT social adoption depending on the kind of stake-holders. In a finer way, within these groups, differences appear along with the new variables discovered in the course of the study : the professional culture, the level of theoretical knowledge on GIS and the perception of the organizational and social implications of GIT diffusion.

At last, this analysis shows all the interest of a comparative geographical approach. At this stage, in fact, the prominent role of the variables 1, 7 and 8 - which reflect the spatial but also organizational context - on the relationships "perception - adoption - use" lay the foundations of further reflection on the social construction of GIS. The more a municipal organization becomes complex and hierarchised, the more the jobs and tasks contain very accurate professional functions and the more the modes of social adoption of GIT are differentiated according to the groups of stake-holders. The larger will a town be, the greater the diffusion of these tools will be perceived as power stakes, the greater the risks of competition for their appropriation will be.

Even further, some cultures (particularly the Anglo-Saxon one) generate spatial perceptions closer to the representations issued from GIT (geometrical and static). Therefore the social appropriation of these tools benefit more to the stake-holders concerned. This type of spatial approach is the one developed in the United States - the layer-cake view of the world (Goodchild 1995, Chrisman 1988) and has already become the standard model of GIS software the most used worldwide. Nevertheless the territory is modelled by GIS, despite the limits particular to these tools, according to the space perceived by the actors. The structuring of data into fragmented corporative cartographic layers reflects a mode of functioning of the municipal departments of Quebec deeply rooted in the culture of the town long before the diffusion of GIT. Somehow, the nature of the geographical data in a database, their structuring, the choices of formal representations, etc., reflect the political priorities of a municipal government regarding its territorial planning.

A new approach of the study of GIS appears from these results : GIS are genuine social constructions (Chrisman 1995), they reflect spatial practices which are deeply rooted in their development context (geographical, organizational cultural, etc.). GIS only are what those who use them make out of them. They are the result of the differentiated social adoption of the technologies associated to them and of the geographical information. GIS follow the same rules as computerized systems: "... A successful computerization is a delinquent process, colonised by the social stake-holders who are constantly cheating with it and there lies precisely the token of its success as this computerization process is not rejected, buried in the cardboard boxes of central services, cemeteries or museums. It will have been assimilated" (Pave 1989).

Undoubtedly, this paradigmatical change should offer us elements of reflection to undertake steps better adapted to the always delicate assessment of the potential users' expectations about these new technologies. The aim is to concentrate on more accurate analyses on the relationships between the needs of local stake-holders (whether it be real or perceived) faced everyday to more and more complex problems of territorial management, and the services offered by geomatic tools. Somehow, the "upside-down" approach of the reflections generally led on this topic.