Of computers, myths and modelling : the social construction of diversity, knowledge, information, and communication technologies in Dutch horticulture and agricultural extension

<em></em><em>Introductory chapters</em>In this book two lines of argumentation are developed in relation to what agro-informaticians often describe as 'the limited adoption of Management Supporting Systems in primary agricultural production'. Although -conceptually speaking- the distinction is somewhat problematic, I will speak of a 'theoretical' and a 'practical' line. Figure 1.1 in chapter I shows how these lines of argumentation are interwoven throughout the book.Chapter I is a general introduction to the nature and scope of this book, and provides also some guidance to readers from different audiences. Moreover, and anticipating my later theoretical argument, it makes clear that I distance myself from both realist positions at the ontological level, and from positivism at the epistemological level. Instead, I adopt a constructivist stance, which posits that our understanding of the world is inherently socially constructed. Naturally, this holds for my own understanding of the world as well. Hence, preceding more detailed accounts in relation to specific case-studies, chapter 1 also touches on some broader social dimensions of this research that clarify in the context of which 'negotiation processes' this study was shaped, and which interests, projects, feelings, etc. of the researcher were of importance.In chapter 2, 1 attempt to 'set the scene' by problematizing currently proposed solutions to the limited adoption of Management Supporting Systems (from now on MSS) by farmers and horticulturists. Drawing upon recent theoretical and empirical insights, I conclude that current problem definitions and solutions rest on inadequate <em>unilinear models</em> of, on the one hand, farm development and, on the other, knowledge generation, exchange and utilization.In my practical line of argumentation, my elaborations lead me to identify five practical contributions that extension science and rural development sociology may provide to practitioners in the field of agro-informatics. Hence, I commit myself to providing such contributions in relation to: (1) the generation of relevant classifications of farmers and horticulturists; (2) the development of criteria for the design of MSS that facilitate integration of scientific and other types of knowledge; (3) the assessment of potential contributions of extension workers to the use and development of MSS; (4) the provision of inductive methodologies for identifying relevant information needs; and (5) an appraisal of the types of user-research and user-influence that can be suitably incorporated into methods for MSS-development.In relation to the theoretical line of argumentation, I conclude that a theoretical framework for understanding the use and development of MSS will have to meet two important criteria. First, it needs to allow us to understand interactions in which MSS play a role in the (historical) context of a complex social setting in which a <em>variety</em> social actors are <em>actively</em> engaged. Second, it should help us to conceptualize the <em>social</em> dimensions of knowledge, information, communication and rationality, Furthermore, I propose that management supporting systems or information technologies are best conceptualized as computer-based <em>communication technologies (CT).</em><em></em><em>Part I: Theoretical explorations</em>Chapters 3, 4, 5 and 6 constitute Part I of the book. In search of a theoretical framework that meets the above formulated criteria, various disciplines and fields of study are explored in chapters 3, 4 and 5.In chapter 3, it is argued that the dominant conceptualizations; in computer science fail to meet the two criteria. The same holds for the field of information systems research. Despite important differences, neither the first, second or third wave approaches of management and organization theory, nor hard, soft, critical or autopoietic systems thinkers, provide fully satisfactory conceptualizations. In one way or another, most approaches appear to include elements of determinism, and fail to conceptualize actors as active and historicallysituated agents. Similarly, even if many approaches transcend overtly simplistic 'mechanical' conceptualizations of information, they tend to emphasize the <em>subjective</em> rather than the <em>social</em> dimensions of knowledge and information. That is, due to a focus on the individual, many approaches fail to capture the political, normative and ideological dimensions of knowledge and information. Nevertheless, critical and autopoietic systems thinkers especially, provide some inspiring theoretical concepts and ideas that must somehow be incorporated into a conceptualization of CT-use and development (these include ideas concerning validity claims and the social nature of rationality, the historical and recursive nature of structure, and the concepts of thrownness, blindness and discontinuity).In chapter 4, 1 discuss several frequently used approaches in communication science and extension science. In essence, my elaborations lead me to draw similar conclusions with respect to the two criteria formulated as those arrived at in chapter 3. Even if extension scientists are increasingly aware that extension processes need to be studied in a <em>'multi- actor'</em> context, this assessment has apparently not yet resulted in the development of conceptualizations that are in line with it. That is, the social dimensions of knowledge, information, communication and rationality are insufficiently explicated, and extension scientists often remain to have a rather passive conceptualization of human action. In relation to this, I argue that although extension scientists provide an interesting framework for <em>describing</em> different types of <em>anticipation problems</em> that occur in relation to CT-use and development, the 'diagnostic value' of such descriptions is limited as long as an understanding of <em>why</em> and <em>how</em> such problems emerge is lacking. Thus, I propose that there is a need to enrich both communication science and extension science with sociological conceptualizations of human action, communication, knowledge, information and rationality. Hence, I reject claims made by authors who -on the basis of a sharp distinction between 'knowledge for action' and 'knowledge for understanding'- argue that studies aimed at generating 'knowledge for understanding' are almost inherently of little use to practitioners.Although it appears in chapter 5 that an actor-oriented sociology of rural development provides promising conceptualizations of the social actor, human action, knowledge and ignorance, I argue that there are weaknesses as well. The approach generates a number of important analytical concepts, but it is often unclear how they are to be theoretically connected. Moreover, the conceptualization of social structure leans towards 'actor voluntarism'. Another issue is that actor-oriented sociologists have so far insufficiently reflected on their own role in the production of social change, so that in its present form the approach has little to offer practitioners.In a search for more comprehensive frameworks, I follow suggestions made by some authors in previous chapters, and continue chapter 5 with an evaluation of the prospects of Habermas' theory of communicative action, and Giddens' theory of structuration. Even if Habermas' framework is becoming increasingly popular among extension scientists, I conclude that it fails to meet the two criteria formulated, and thus that it is unsuitable for both improving our understanding of the use and development of CT in agriculture, and more generally- for helping extension scientists to deal with 'multi-actor' intervention contexts. In contrast, a constructivist interpretation of Giddens' theory seems to meet the criteria much better. Giddens proposes that all social interaction has a communicative dimension and that the production of meaning (and therefore the production of knowledge and information) is <em>inherently</em> connected with the operation of power and normative sanctions. Furthermore, Giddens' theory demonstrates how actors are <em>actively</em> involved in (re)producing social structure. Also, it allows me to identify mutual knowledge (as inherently connected with mutual ignorance) as the key modality of structure, and therefore as underlying the existence of structural properties and the operation of power in society. In all, I conclude that Giddens' theory offers a much more promising and/or systematic insight into the interrelations between action, structure, knowledge, communication and rationality than the other sociological approaches discussed.In chapter 6, I attempt to clear the ground for more empirical forms of investigation. Building on the insights arrived at in the theoretical explorations, I formulate a set of interrelated preliminary theoretical propositions with respect to how the use and development of CT should be understood. Most importantly, it is proposed that CT-mediated communications must be looked at as politically and normatively laden negotiation processes, which are inherently connected with the (re)production of structural properties in society. The social 'codes' incorporated into such technologies are both constraining and enabling and can be renegotiated and creatively dealt with so that largely unintended consequences can easily emerge. Moreover, such technologies are best understood as playing a -never fully neutralrole in actors' reflexive monitoring of action, rather than in processes of 'rational decision making'.Building upon the theoretical explorations, I present the overall 'theoretical' problem statement as:To what extent do anticipation problems originate from: (a) the social nature of CT-development processes, and (b) insufficient recognition of the social dimensions of knowledge, information, communication and rationality.At the 'practical' level we must add the question of how an understanding of such social dimensions can help to improve the anticipatory nature of communication technologies. Inspired by the theoretical framework adopted, I thereby translate earlier defined envisaged practical contributions into more specific guiding questions. I also formulate additional questions aimed at gaining an understanding of the role of the social scientific researcher in processes of social change.Following this, I introduce a variety of methodological guidelines that follow from my theoretical approach. Here, it is argued that empirical studies must focus on day-to-day social practices at social interfaces. These practices must be studied in a context-sensitive manner, whereby attention must be given to actors' diverging interpretations, projects, rationalizations, and attempts to create space for manoeuvre. Furthermore, I adopt a casestudy approach.Finally, I argue that the consequences of my (meta-)theoretical framework are mainly methodological, for it proposes that a more down to earth understanding of social life must be rooted in the experiences and life-worlds of the actors that are studied. Thus, I argue that only if the methodological guidelines indeed help to increase both our understanding of the use and development of communication technologies, <em>and</em> our capacity to make practical contributions, can we conclude that (elements of) the theoretical framework, and/or (some of) the preliminary theoretical propositions are plausible and relevant.<em>Part II: Empirical investigations</em>Chapters 7, 8, 9 and 10 constitute Part II of the book. First, I elaborate in chapter 7 on the 'structural setting' in which the case-studies presented later must be understood. I critically discuss: (a) the institutional efforts to stimulate the use and development of CT in primary agricultural production; (b) current evaluations of the present state of the art in this respect; and (c) dominant methods for the development of CT in agriculture. Furthermore, I express my theoretical reservations with respect to both the frequently made distinction between 'structured', 'semi-structured' and 'ill -structured' problems, and the classifications of different types of CT that are based on this differentiation. As an alternative, I propose a classification of CT which is based on two theoretically-inspired distinctions, namely: (a) the separation between what I call the 'external' and 'internal' CT-design, and (b) a differentiation between various types of (normatively and/or politically laden) models that are more or less implicitly incorporated. In the closing section of chapter 7, I make an attempt to theorize the broader (macro) context by drawing upon Knorr-Cetina's notion of <em>summary representations.</em> First, I identify a 'web' of interrelated summary representations (in the form of classifications) that agro-informaticians frequently draw upon. Then, I make plausible: (a) that these classifications have historical, political, normative and ideological connotations; (b) that they are drawn upon in particular social practices; and (c) that they consequently underlie a number of important structural properties of the domain of agroinformatics. Finally, I illustrate that the above-mentioned classifications are inherently connected with actively and intentionally constructed areas of ignorance, and hence, that selfreferential processes occur in the domain of agro-informatics.In chapter 8, I present a case-study on the use of CT that provides dairy farmers and extension workers with a large number of parameters and norms relating to the performance of particular farms. In this study, I operationalize diversity among farmers in terms of styles of farming. By means of both quantitative and qualitative analysis, I show that farmers belonging to different styles tend to differ with respect to: (a) the parameters and norms that they focus on; (b) the goals that they formulate vis-à-vis the norms; and (c) the types of comparisons that they make with the help of the CT. Also, it appears that there are plausible connections between the specific ways of dealing with the CT, the different knowledge networks that farmers are part of, and the specific strategic notions that underlie the different farming styles. Moreover, I elaborate briefly on the way extension workers use the CT, and on some contradictions in their evaluation of farms and farmers belonging to different styles.In the concluding section of chapter 8, 1 reflect on theoretical issues and on the three practical contributions that I wanted to deliver, and which guided the selection of the case. Most importantly, I arrive at the following conclusions. First, I conclude that extension workers may facilitate farmers' learning on the basis of CT, but that in this process they bring in a <em>different</em> (rather than <em>different levels of</em> ) expertise and analytical capacity. Second, I argue that increasing standardization and formalization of agricultural knowledge into more and more complex models underlying CT, poses serious risks. In relation to this, I conclude that the prospects of bringing about a much needed integration of knowledge from different epistemic communities are rather limited in the internal design of a CT, and that such integration is more realistically achieved by means of an adequate external design. Third, I assess that the type of CT under consideration can be particularly useful in providing an <em>agenda for discussion,</em> that the quality of such discussions can be especially high in farmers' study clubs, and that -in the light of diversity- the provision of normative parameters in such discussions is confusing, and therefore of limited value. Fourth, I argue that existing methods for identifying 'homogeneous' target-categories rest on inadequate theoretical models, and fail to grasp the social dimensions of making such classifications. Although the classification into styles of farming proves more insightful than current classifications, there are grounds for arguing that classifications starting from different dimensions could have increased my capacity to make practical contributions in relation to the use and development of CT. Finally, I analyze why carrying out this case-study did not contribute much to social change, and develop conditions that must be met in the next case-study in order to be more effective in this respect.Chapter 9 provides a qualitative analysis of enterprise registration and comparison practices among cucumber growers. The diversity observed in relation to these practices, leads me to develop several classifications of cucumber growers. Further qualitative analysis shows that each practice-based classification of diversity helps to both reveal the social dimensions of enterprise registration and comparison practices, and generate concrete design-criteria that any CT aimed at supporting these practices should have to meet. In relation to these criteria, I evaluate the prospects of one postal package and two CT-based packages that were actually designed to support enterprise registration and comparison activities. I conclude that the three packages are characterized by meaningful differences, and that the relative success of the packages can indeed be plausibly explained with reference to the design-criteria formulated.In the concluding section of chapter 9, 1 reflect on theoretical issues and on the four practical contributions that I wanted to deliver, and which guided the selection of the case. Most importantly, I arrive at the following conclusions. First, I conclude that the making of <em>several</em> classifications of diversity on the basis of knowledge-related practices has helped to generate a sharper insight into different rationalities, knowledge networks, communication patterns and CT-use, than appeared possible with the help of the single classification into styles of farming presented in chapter 8. Moreover, the exploration of several classifications provided more insight into how diversity is socially constructed. Second, on the basis of a comparison of the CT discussed in chapters 8 and 9, 1 conclude that as the complexity of a CT increases, the more both the anticipation of diversity and the integration of knowledge from different epistemic communities depend on an adequate external design. Third, it emerges that CT themselves can -in time and space- play a role in facilitating the integration of knowledge from different epistemic communities. In order to achieve this, it is important to design CT in such a way that they can serve as an agenda for discussion, and facilitate (joint) processes of learning. Moreover, the case-study illustrates that such learning processes may constitute a process of 'structural change'. Fourth, the conclusion drawn in chapter 8 with respect to the role of extension workers is reinforced. Fifth, the study indicates that information needs emerge in a continuous and often routine-like process of learning. In relation to this, information needs are not only diverse, but also socially constructed, and often subject to rapid change. Therefore, my earlier plea to use inductive rather than deductive methods for the identification of such needs is reinforced. Moreover, I argue that it is more useful to identify knowledge and information-related <em>practices</em> and <em>types</em> of information requirements rather than specific information needs. Finally, 1 argue that in this case-study the activities of the researcher did indeed lead to a large number and variety of practical conclusions, and an impact of these on the actual course of CT-development events. 1 suggest that this phenomenon can be attributed to: (a) the fact that growers and researcher enrolled each other, whereby the latter became involved in an existing local project; (b) the political commitment of the researcher; (c) the fact that the practical problem statement was formulated by the growers; (d) the research methodology adopted; and (e) the institutional arrangements under which the study took place.In chapter 10, I try to reconstruct the development histories of two CT-based packages that were designed to support enterprise registration and comparison activities. By doing so, I hope to gain insight both into why and how anticipation problems emerge, and into how software developers, growers, social scientists and extension workers may contribute to the development of adequate CT. On the basis of the reconstructions arrived at, I draw the following conclusions. First, CT-development processes emerge as complex arenas of negotiation and enrolment, in which cooperation and conflict need to be examined in the (historical) context of diverging and changing interests, resource bases, normative convictions, theoretical beliefs, spatial characteristics, etc. That is, CT-development processes are inherently social in nature, and are likely to constitute interface situations. Within and through these processes actors attempt to create longer term outcomes or structural properties, whereby they are often confronted with unintended consequences of their own and other actors' previous actions, and compelled to adapt initial goals, change routes, create new coalitions, etc. Second, CT-development processes can at the same time be fruitfully considered as being inherently social processes of learning. Third, in relation to the prospects of soft systems methodologies for enhancing such learning processes, the case- study indicates that efforts to create a negotiated 'consensus' may have counterproductive consequences when boundaries are chosen in such a way that actors with too widely diverging goals, interests, convictions, etc., are included. This seems to hold especially during the early stages of innovation processes. Fourth, even if -in practice- there appears to be only a gradual difference between project-oriented CT-development methods and prototyping, the use of planned approaches towards CT-development may at times obstruct rather than stimulate the development of an appropriate innovation in that the procedures adopted prevent the rapid inclusion of learning experiences. Finally, I argue that, if CT-development processes are learning processes, it makes sense to <em>organize</em> them as such as well. Thus, I set out to develop a 'learning-oriented' method for CT-development which incorporates methods that originate from actor-oriented sociology, extension science and prototyping approaches. The method that I propose is not so much aimed at realizing particular predefined <em>goals</em> by means of formal planning, but rather to test the feasibility and desirability of particular <em>ideas,</em> and adapt these where necessary. To this end, the method is designed as an open-ended procedure in which interveners, social scientists and various categories of prospective beneficiaries have different learning responsibilities.<em>Part III: Discussion and conclusions</em>In the concluding chapter, the theoretical line of argumentation is rounded off with, on the one hand, a reflection of the overall 'theoretical' problem statement, and on the other, some more specific conclusions in relation to the various fields of study that were covered in Part I of the book. In essence, I conclude that it has become plausible that a lack of correspondence between, on the one hand, the various models that are (more or less implicitly) incorporated in CT-design, and on the other, the models that are actually drawn upon in the context in which such CT are supposed to be used, indeed originates from both the social nature of CT-development processes, and insufficient recognition of the social dimensions of knowledge, information, rationality and communication among those that develop CT. The practical line of argumentation is brought to an end with the formulation of recommendations for practitioners in relation to the five practical contributions that I envisaged to make by means of my study.Eventually, I conclude that a constructivist actor-oriented 'communication paradigm' constitutes an attractive perspective for both understanding and improving CT-development interventions in agriculture and horticulture.