Developing the Horizons of the Mind: Relational and Contextual Reasoning and the Resolution of Cognitive Conflict 9780521817950, 0521817951

Citation preview

This page intentionally left blank

Developing the Horizons of the Mind

Developing the Horizons of the Mind is the first book on Relational and Contextual Reasoning (RCR), a new theory of the human mind which powerfully addresses key areas of human conflict such as the ideological conflict between nations, the conflict in close relationships, and the conflict between science and religion. K. Helmut Reich provides a clear and accessible introduction to the new RCR way of thinking that encourages people to adopt an inclusive rather than an oppositional approach to conflict and problem-solving. Part one outlines the key aspects of RCR theory and supporting empirical data, and Part two offers examples of its application in the modern world. RCR provides a stimulating and challenging tool to several disciplines, including philosophy, psychology, religious studies and education, and this book will be a valuable resource for cognitive scientists, psychotherapists, theologians, educators and all those involved in conflict resolution. .   has had successful careers as a physicist and as a psychologist, winning the William James Award of the APA for contributions to the psychology of religion in 1997.

Developing the Horizons of the Mind Relational and Contextual Reasoning and the Resolution of Cognitive Conflict K. Helmut Reich

          The Pitt Building, Trumpington Street, Cambridge, United Kingdom    The Edinburgh Building, Cambridge CB2 2RU, UK 40 West 20th Street, New York, NY 10011-4211, USA 477 Williamstown Road, Port Melbourne, VIC 3207, Australia Ruiz de Alarcón 13, 28014 Madrid, Spain Dock House, The Waterfront, Cape Town 8001, South Africa http://www.cambridge.org © K. Helmut Reich 2004 First published in printed format 2002 ISBN 0-511-04275-2 eBook (netLibrary) ISBN 0-521-81795-1 hardback

To my grandchildren Michael, Nicole, Nicolas, Natacha May they grow up in a society in which relationships and contexts become more and more recognised and valued

Contents

List of figures List of tables Acknowledgements Introduction

page x xi xiii 1

Part I The Theory of Relational and Contextual Reasoning (RCR) and its Empirical Study 1. Introduction Caveats The nature of relational and contextual reasoning Previous work on relational thinking Summar y of the introduction

2. Development of RCR Anthropology adopted Theories of cognitive development Cognitive development and RCR Unreflected, object-reflecting, and means-reflecting thought Intra-inter-trans – the ‘logic’ of RCR development Input to the present study from earlier work Summar y of RCR development

3. Metaphysical Assumptions and Theor y of RCR Assumptions adopted from the philosophy of knowledge Theor y of RCR Summar y of the metaphysical and theoretical grounding of RCR

4. Empirical Studies of RCR Over view Methodological commonalities Pilot study 1: RCR level descriptions and RCR effectiveness as pragmatic reasoning schema Pilot study 2: Additional RCR inter view problem Pilot study 3: RCR and Piagetian operations

11 11 12 22 23

25 25 26 27 29 32 33 34

35 35 41 46

47 47 49 50 55 59

vii

viii

Contents Pilot study 4: RCR, Piagetian operations, cognitively complex thinking, and evolved logics Discussion of pilot studies 3 and 4 Summar y of empirical studies and outlook

5. Other Thought Forms and Matching Them to the Problem at Hand Other thought forms relevant to RCR Matching the form of thought to the structure of the problem Summary of other thought forms and matching them to the problem at hand

63 72 74

75 75 91 97

Part II Applications of RCR Over view

6. Methodology Method for applying RCR Demonstration of a par ticular search

7. Religion Religion and the nature of human beings Understanding religious doctrines Co-ordination of religious and scientific world views RCR and religious development Conclusions

8. The Archaeology of RCR Fathers of the Council of Chalcedon Vincent van Gogh William James Rainer Maria Rilke Rober t Musil Niels Bohr Conclusions

9. Psychology Psychology as a discipline The case of individual development Psychophysiological phenomena Which music for which purpose? Conclusions

10. Education Who controls the educational system? Teaching the investiture contest Stimulating RCR in the classroom Concluding remarks

101

103 103 104

116 116 120 126 129 132

133 133 134 136 137 139 140 143

145 145 149 151 152 156

157 157 158 159 163

Contents

11. Social Issues Overcoming illegal use of drugs Nuclear power Ways to solve old problems and create new work Concluding remarks

12. Conclusions This volume Postscript

ix

165 165 174 181 184

185 185 187

Appendix 1: Inter viewing techniques

191

Appendix 2: Scoring manual for RCR

194

References Index

199 219

Figures

2.1 Evolution of cognition 3.1 ‘Figure–ground’ shift of the number of cubes 3.2 Noncompatibility according to Bedau and Oppenheim (reproduced by kind permission of Kluwer Academic Publishers, Dordrecht, The Netherlands) 4.1 Schematic sketch of the snail task 5.1 Venn diagrams of class sets 5.2 INCR group 5.3 Through becoming, being and nonbeing transform into new being and new nonbeing 7.1 Changes when moving from one to two (reproduced by kind permission of the British Journal of Religious Education) 7.2 Correlation between religious judgement stages and levels of RCR

x

page 30 44

45 60 79 83 86

121 131

Tables

1.1 The four structural levels of the model of thought processes page 18 4.1 Description of RCR levels (reproduced by kind permission of S. Karger A.G., Medical and Scientific Publishers, Basel, Switzerland) 52 4.2 Mean scores of RCR scores of pilot study 1 (reproduced by kind permission of S. Karger A.G., Medical and Scientific Publishers, Basel, Switzerland) 53 4.3 Developmental logic of RCR 54 4.4 Mean scores of RCR scores of pilot studies 3 and 4 58 4.5 Mean scores of RCR scores of pilot study 3 62 4.6 Frequencies of individual scores concerning Piagetian operations and RCR levels (reproduced by kind permission of John Wiley & Sons, Inc., New York [owner of Jossey-Bass, Publishers, San Francisco, the original publisher]; Hogrefe, Verlag fur ¨ Psychologie, Gottingen, ¨ etc.; Ernst-Reinhardt-Verlag, Munich) 63 4.7 Frequencies of individual scores of levels of cognitively complex thought and RCR levels 66 4.8 Frequencies of individual scores concerning levels of (meta-) logical thinking and RCR levels 70 4.9 Minimum stages/levels of other competencies for a given RCR level (reproduced by kind permission of the Ernst-Reinhardt-Verlag, Munich) 72 5.1 Various operations based on formal binary logic 78 5.2 The sixteen binary operations 80 5.3 Level of cognitive complexity from an analysis of diplomatic notes (reproduced by kind permission of Sage Publications, Inc., Thousand Oaks, CA) 85 5.4 Main differences of various forms of thought 89

xi

xii

List of tables

7.1 Frequencies of individual scores of RCR levels and intelligibility judgement of the Chalcedonian Definition (reproduced by kind permission of the British Journal of Religious Education) 7.2 Frequencies of individual scores of RCR levels and intelligibility judgement of the Doctrine of the Holy Trinity (reproduced by kind permission of the British Journal of Religious Education) 7.3 Frequencies of individual RCR levels and levels of co-ordinating biblical and scientific views (reproduced by kind permission of Pabst Scientific Publishers, Lengerich, Germany) 9.1 Classification of guidelines for functional music

123

125

127 156

Acknowledgements

No book is complete without giving thanks to all those who contributed in one way or another to the work described and the writing of the book. (However, the responsibility for the actual content remains entirely mine.) Thanks are particularly in order in this case, as I embarked upon a new field of inquiry, and thus needed more help and interaction with colleagues than usual. Fritz Oser made available the facilities of the School of Education of Fribourg University, introduced me to actual empirical work in developmental psychology, and was a helpful discussion partner all along. Wolfgang Edelstein, Hans Fischer, August Flammer, Dedre Gentner, Jean-Blaise Grize, Rolf Hagedorn, Siegfried HoppeGraff, Karen S. Kitchener, Deirdre Kramer, Rolf Oerter, Philibert Secretan, Thomas Bernhard Seiler, Victor F. Weisskopf, and Phillip K. Wood were each invaluable in the early stages for critical comments and suggestions for the next steps. Carol Rausch Albright, Henry Babel, Michael A. Basseches, Paul-Richard Berger, Mark H. Bickhard, Ronnie Blakeney, Thomas Bornhauser, Anton A. Bucher, Thomas J. Burke, Michael Chandler, Philip Clayton, Veit-Jakobus Dieterich, Lutz Eckensberger, Helmut J. Efinger, Reto Luzius Fetz, Ernst Peter Fischer, Anne Foerst, James W. Fowler, Peter C. H¨agele, Philip Hefner, Heinz S. Herzka, Stefan Huber, John Hull, Bruce Hunsberger, Michael E. Hyland, Christopher B. Kaiser, Gisela Labouvie-Vief, Thierry Magnin, David Moshman, Ehrhard Muhlich, ¨ W. Jim Neidhardt, Erwin Nickel, Karl Ernst Nipkow, Willis F. Overton, Arthur Peacocke, Martin Rothgangel, Robert J. Russell, George Scarlett, Gerhard Schurz, Friedrich Schweitzer, Kevin Sharpe, Jan D. Sinnott, Bernard Spilka, Maria Spychiger, Peter Suedfeld, Kalevi Tamminen, Eberhard Todt, Peter Valentin, Hendrika Vande Kemp, Harald Walach, Christoph Wassermann, Michael Welker, and David M. Wulff over the years stimulated progress by way of their remarks and suggestions. I learned much from more advanced (anonymous) colleagues through discussions especially after my presentations at various conferences in Europe and North America, and I benefited significantly from relevant Internet discussions. Whenever I quote here xiii

xiv

Acknowledgements

(a large part of ) a pertinent posting (in agreement with the author), I indicate the author’s name. However, in some cases, I simply follow up hints, summarise the gist of a discussion, or use some expressions without individual acknowledgements. I am indebted to all participants of those discussions, especially to Ric Barr, David R. Burwasser, Michael Cavanaugh, Thomas L. Gilbert, Ursula Goodenough, Paul Harrison, William Irons, Rex Kerr, Edwin C. Laurenson, Steve Petermann, and V. V. Raman. My colleagues at the School of Education, in particular Wolfgang Althof, Franz Baeriswyl, Traugott Els¨asser, Aloys Niggli, and Roland Reichenbach, helped in numerous ways to make me a social scientist and educator. Anton Bucher (pilot study 1), Birgitta Michel (now Mrs Thenen – pilot studies 2 and 3) interviewed grade-school children in the local Swiss-German dialect (which I do not speak), an essential ingredient of an interview bringing out the authentic views of these children; both of them and Ornella Di Loreto (now Mrs Miller – pilot study 4) also participated in the scoring. Philip K. Wood, Richard Klaghofer, and Bernd Kersten in succession were instrumental for getting the statistics done. Without Anke Schroder ¨ the classroom work would not have been the opportunity for such a fruitful experience as it has been. And clearly, without all those interview respondents and participating pupils as well as students and older adults, there simply would be no book. My wife Ursula discussed many issues with me, put up gracefully with plenty of unavailability of her husband, and – given our differing biographies, outlooks and temperaments – provided a good number of opportunities to apply RCR and thereby refine it. Without Emily Wilkinson it would have been more difficult to get the book off the ground and flying. Peter F. Bucher-Roth, Michael Cavanaugh, Kristen E. Kann, Bernd Kersten, Hans Koppel, ¨ Ueli Simmel, Geoffrey Scobie, Bernard Spilka, Joy Stephens and the anonymous reviewers gave valuable feedback on earlier versions of the present text. Sarah Caro, Sophie Read and Gillian Dadd dealt gracefully and effectively with my typescript at Cambridge University Press. A number of figures and tables could be included due to the kindness of the permission-granters (pp. x–xii) and the respective authors. Finally, the Hochschulrat of the University of Fribourg provided financial support for this research. I express my special thanks to all of the above.

Introduction

The main purpose of this monograph is to present the findings of fifteen years of continuous research on a particular postformal form of thought, namely, ‘Relational and Contextual Reasoning’ (RCR). RCR is particularly helpful when one seeks to co-ordinate two or more competing theories about the same phenomenon or issue. An example of usefully applying RCR would be when one is debating whether to attribute an outstanding athletic or artistic performance to native endowment or to training. RCR will clarify the extent to which the two kinds of explanations are needed, bring out any links between them, and elucidate the respective explanatory potential in the context considered. Secondary aims of the monograph are (a) to stimulate further study of RCR, (b) to demonstrate its potential for solving particular problems better than other forms of thought, and (c) to encourage use of RCR and its broader application. Given these main and secondary aims, arranging the material in a coherent manner was not obvious, apart from (c), to which Part II is devoted. Considerations (b) were finally moved to a later chapter (Chapter 5), to be presented after the main aim and secondary aim (a) are met. The research on relational and contextual reasoning to be reported was originally triggered by the following observation. Whereas many adolescents espouse either a religious or a scientific world view when trying to understand what goes on around and inside them, some manage to ‘combine’ both views in some fashion. The question that intrigued me was, ‘How do they do it?’ The answer I came to after looking at other possibilities was that those adolescents use relational and contextual reasoning, a term that I adopted after other trials for reasons to be discussed below. Before I fully reached that insight, however, I had first to work my way through theories of reasoning already proposed. Until the 1970s, Piagetian formal operations were considered by many researchers to be the high end of individual development of reasoning. The label formal operations indicates that certain formalisms have been developed by an individual which can be used for solving a class of problems 1

2

Developing the Horizons of the Mind

irrespective of their particular content. Such operations involve a number of aspects, such as exploring possibility space, hypothetico-deductive theory building, and checking a solution for its internal and external logical consistency. As considered here, the exclusive use of formal binary logic constitutes the characteristic core of Piagetian formal operations (but see Labouvie-Vief 1980). This is exemplified by the central system of sixteen binary operations, to which I shall come back in detail in chapter 5. In the early 1980s, a category of more highly developed thought, called ‘postformal operations’, became a topic of interest to a small group of psychologists (e.g., C. N. Alexander, P. K. Arlin, Ch. Armon, P. B. Baltes, M. A. Basseches, A. Blasi, J. M. Broughton, M. J. Chandler, M. L. Commons, C. Gilligan, H. Koplowitz, D. Kramer, G. Labouvie-Vief, E. J. Langer, F. A. Richards, J. D. Sinnott). A number of volumes on that subject were published in fairly rapid succession (e.g., Commons, Richards and Armon 1984; Commons, Sinnott, Richards and Armon 1989; Commons, Armon, Kohlberg, Richards, Grotzer and Sinnott 1990; Alexander and Langer 1990). But those publications appear to have dwindled to a trickle (e.g., Sinnott 1998) without having resolved the central issue of the distinguishing characteristics of postformal operations and their relations with Piagetian formal operations. No consensus currently exists regarding those characteristics and relations. I would argue that insufficient attention is paid to logic in this debate. Postformal operations may in principle share much with Piagetian formal operations, with the exception of formal binary logic. In my view, postformal thinking is based on logics different from formal binary logic. For that reason, as will be shown, fully developed RCR, with its specific logic, is postformal. A further characteristic of postformal thought is suggested by the established use of the word ‘post’ (as in ex post factum) implying that fully developed postformal thinking arises after the Piagetian formal operations are mastered. This, however, does not exclude a development of less developed stages of various other thought forms in parallel with Piagetian stages. I bring to this work thirty years of experience of research in physics and engineering, together with seventeen years in social science, principally psychology. With this background, it is perhaps not surprising that philosopher physicist Niels Bohr came to my mind when I came across the adolescents who ‘combined’ religious and scientific world views. Among other issues, Bohr discussed the paradox of the wave-like and particlelike behaviour of light in terms of complementarity – that these contextdependent behaviours do not contradict, much less exclude each other, but instead ‘complete’ each other, and both pictures are needed for a full explanation in non-mathematical terms. I later became aware of William James’s account of complementary phenomena concerning (a) memory

Introduction

3

and (b) the stream of thought (Reich 1998). Along with these leads, my own career(s) continually encouraged me to look at things from differing points of view and then to work towards a coherent ‘story’. Having become aware of the possible existence of relational and contextual reasoning, I interviewed students and some professional physicists on issues with a ‘structure’ similar to that of religious vs. scientific world views. For example, I asked them (1) about whether the change from the Romanesque to the Gothic church architecture had spiritual, or economic causes; (2) whether kidney pain is best relieved by surgery, or by drinking a certain type of herb tea; (3) whether the reported crash of a glider was due to naturally explainable causes, or to ‘fate’ as foretold by the pilot’s horoscope. Along with collecting these data, I also studied various types of logic, the debates on the interpretation of quantum theory in physics, as well as various views on the relationships between science and religion/theology. Slowly I came to postulate hypotheses about RCR (initially called ‘thinking in terms of complementarity’ – Oser and Reich 1987; Reich 1994b), and then worked on clarifying them through empirical work and analyses of the results. After understanding RCR better, I tried to elucidate its ‘composition’. My current view is that it shares ‘components’ with other thought forms, namely, with Piagetian operations, cognitively complex thought, and dialectical as well as analogical thinking. Therefore, I deal here also with these thought forms after having established the distinctness of RCR. Let me return for a moment to the differing views of religion and science. Is one right and the other wrong? Often both are aiming to ‘explain’ the same phenomenon, as for example in the case of the origin of the universe. Using a Latin term, the phenomenon to be explained – here the origin of the universe – is designated as the explanandum. Whoever works on the explanatory task in the examples given (and in structurally similar ones) and employs relational and contextual reasoning, should keep the competing theories distinct. For instance, when a scientific explanation is (still) missing, to introduce divine action as part of a ‘scientific’ explanation is not appropriate. Yet, all (partial) theories should be used fully (in their context). This may be referred to as ‘both-and’ reasoning. When applying (partial) theories, one may find that one or the other theory has more explanatory power under some conditions, and less under others. In other words, one may find that context affects the explanatory efficacy of a partial theory. Fully developed relational and contextual reasoning will elucidate the relations the partial theories have with the explanandum and with each other as well as the details of the context dependence. These relationships involve a trivalent logic: two statements about the same explanandum are

4

Developing the Horizons of the Mind

either compatible (both true concurrently), incompatible (never both true ‘simultaneously’) or noncompatible (not compatible simultaneously, but one is ‘true’ in one context respectively at one point in time, the other in a different context or at a different time). I employ both the terms ‘complementary’ and ‘complementarist’ in this monograph. The distinction between the two terms as I use them here is as follows. Complementary merely indicates that the various parts, aspects, activities, etc. ‘complete’ each other, yet they are inherently independent. By contrast, complementarist refers to aspects, states, activities, events, views, explanations, etc. which are complementary and intertwined, that is inseparable because intrinsically linked (= entanglement as described in quantum physics by Heisenberg’s principle of indeterminacy). Examples of the latter would be native endowment and the efforts to produce fruits of practising an art or skill, or the wave-like and the particle-like nature of light. As these example and others show, as a rule the links are not of the cause–effect type but of other types. A given link may be one of ‘kinship’, of information transfer, of symbiosis, of mutual limitation, and so on. The appeal of studying RCR and using it appropriately goes beyond purely academic intellectual interest. As one looks around, examples abound where either/or thinking (rather than both-and reasoning) has undesirable consequences. I would argue that rigid adherence to either/or thinking has impeded the full realisation of the potential of psychology, education and religion, among other fields, and has hindered better resolutions of societal problems like fighting illicit use of narcotics. In these and other comparable cases, RCR offers a method or a pathway toward more encompassing and fruitful results. In sum, RCR is not needed for solving crossword puzzles nor similar, conceptually simple tasks having just one well-defined solution. Rather, it helps in dealing with highly complex, often controversial problems of the kind just indicated. How does RCR help? RCR helps one to analyse the various aspects of a problem and their ‘internal’ relationships as well as the role of the context and thence to bring out the respective dominant explanation. Doing so contributes to developing the horizons of the mind. This comes about in particular because – where applicable – the use of the ‘structural’ trivalent RCR logic frees one from the limitations of formal binary logic. This may also permit one to resolve cognitive dissonances or even conflicts. Once a teacher grasps the nature of RCR and its developmental logic he or she can stimulate RCR in the classroom step by step. In particular, the teacher can further students’ ability to differentiate and to integrate statements about what is or might be the case, and can help students to become conscious of different types of logic used in establishing and connecting statements.

Introduction

5

Urged on years ago by colleagues and one editor of a leading professional journal to publish this work as a monograph, I deferred doing so until the conceptual basis of RCR was sufficiently clarified, the empirical data well established, and my experience with applying RCR sufficiently promising, particularly in a classroom setting. I am fairly confident that those criteria have now been met. Given RCR’s potentially wide-ranging use, I have written the present volume, in particular Part II, with an audience in mind comprising not merely experts or students of developmental psychology, psychology of education, and cognitive science, but also of interested persons from other fields for whom RCR might be relevant and helpful in practice. A brief discussion of the style of my presentation may be appropriate at this point. Let me begin by citing two exemplars of what I aspire to do in writing this monograph. The first is Ren´e Descartes whose ‘Discourse on Method’ was written in a style that itself illustrated the discourse’s content: his writing was systematic, formal, analytic. The second model is Søren A. Kierkegaard, the Danish philosopher theologian, who documented his revolt against the formal, petrified Church of his day. His writing on that subject is unsystematic, aphoristic, sometimes even disjointed.1 In the same vein, I have attempted to write this monograph in such a way that it expresses RCR stylistically as well as thematically. While the thematic treatment of RCR is fairly obvious, my stylistic demonstration of RCR may require a further word of explanation. The most significant choice a writer may face, apart from the relative formality of his or her style, is whether to proceed deductively after having presented the main thesis ‘up front’ (risking mental overload of the reader), or inductively, presenting the arguments one by one and the resulting thesis as conclusion (risking losing the reader on the way because it is not clear where one is going). In this work exploring RCR, I alternate between partial deduction and partial induction in an effort to emphasise the both-and importance of the two methods for gaining insight. In other words I attempt to make full use of both methods in a complementarist 1

For persons deeply knowledgeable about the Qur’an (which I am not), the style of the holy book of Islam (in Arabic) is another example of a match between content and style: the style is said to express both the sweetness of city dwellers’ sedentary placidity and the forcefulness of Bedouins’ migrating roughness; the rhythm of the syllables echoes that of both prose and poetry – the pauses, while different from those in either prose or poetry, exhibit a harmonious and rhythmic symmetry; the words chosen are neither trivial nor overly rare – they are the expression of an admirable nobility; the sentences are phrased in such a way that the smallest number of words renders thoughts of extreme richness; intellect and feelings/emotions are brought in ‘together’ in such a way that the narration, arguments, doctrines, laws, and moral principles are both intellectually convincing and emotionally engaging (Schimmel 1991, p. 11, quoting an Egyptian scholar; translation from German by K.H.R. as throughout this monograph).

6

Developing the Horizons of the Mind

or linked manner, to illustrate how RCR furthers understanding through iterated changes of the viewpoint. Finally, a caveat: this monograph is not necessarily a fascinating read; in fact, some readers may need extra motivation to keep reading. One of the difficulties with RCR is its ‘invisibility’, which is comparable to that of scaffolding: when the building or the renovation is done, almost no trace is left of the scaffolding. Similarly, once RCR has done its work and a solution to the given problem has been found, there is mostly no trace left of how the solution came to be found. To use another rubric, one may compare RCR with the number zero (without claiming for RCR the importance of the zero). Peter Bernstein (1998, pp. 32–3) wrote: The concept of zero was difficult to grasp for people who had used counting only to keep track of the number of animals killed or the numbers of days passed or the number of units travelled. Zero had nothing to do with what counting was in that sense. As the twentieth-century English philosopher Alfred North Whitehead put it, ‘The point about zero is that we do not need to use it in the operation of daily life. No one goes out to buy zero fish. It is in a way the most civilised of all the cardinals, and its use is forced on us by the needs of cultivated modes of thought.’

Similarly, millions live their lives without having heard of relational and contextual reasoning, and without ever using it. My claim is that, were they to use RCR, they would better their chances for improving personal relationships, tackling complex social problems such as getting people to follow good health habits, and dealing more effectively with social and political situations in strife-torn areas such as Northern Ireland, the Balkans, the Middle East, and elsewhere. There is another possible reason for finding this volume off-putting. Many of us egotistically think that anything we say is both complete and consistent (a violation of Godel’s ¨ theorem, by the way), and therefore incorruptible, unchangeable, and not to be questioned. And now comes an author who potentially challenges that view. How does he have the nerve to do that? While understanding such a reaction, I still hope that for serious thinkers, researchers and scholars, the considerations presented here together with the empirical data and their interpretation should open minds to the parameters of organised thought. Some readers may find that I argue like someone for whom everything becomes a nail because I have a hammer in my hands: indeed, I do use relational and contextual reasoning in many different situations, under widely varying circumstances, and in differing modes, for instance to obtain a result of psychological research, to formulate a hypothesis or a desideratum, or to enable a retroduction.

Introduction

7

Although my understanding has reached a level which makes communication of the results reasonable, I do not claim that this volume constitutes the final word on the issues discussed. Rather, I present something to think about, to be explored jointly, this in the hope that others will also contribute to the progress of RCR and its applications. The organisation of the volume is as follows. Chapter 1 presents fully developed RCR in a basic way so that the sequel becomes understandable. It includes a structural analysis of RCR in terms of elementary operations, conjunctive operations, composite operations, and complete forms of thought. Chapter 2 discusses background knowledge needed for understanding the development of RCR: the general ontogenetic development from the child’s searching to understand the world to the adolescent’s argumentative description to the mature adult’s balanced views which imply an awareness of the power but also of the limitations of the human mind. Piaget’s concept of intra-inter-trans, the logic of RCR development, is introduced at that point. Chapter 3 deals with the philosophy of knowledge adopted, and the theoretical underpinnings of RCR. Chapter 4 reports the basic empirical data. Chapter 5 discusses the other thought forms of concern (Piagetian logico-mathematical thinking, cognitively complex thought, dialectical as well as analogical thinking), and expounds on the need to match the type of thought that one uses in analysing and solving a problem to the structure of the problem itself in order to obtain best results. The reason to choose just these thought forms is twofold: on the one hand, as already mentioned, those forms share ‘components’ with RCR. On the other hand, due to their difference in characteristics and ‘performance’, they underline the fact that the choice of an appropriate thought form matters. For both reasons they need to be known in their own right, not just as contrasts to RCR. Part II, that is Chapters 6 through 11, discusses applications of RCR. My conclusions are presented in Chapter 12; the visions of Reginald Victor Jones and of Daniel Goeudevert complete that last chapter. Appendix 1 and Appendix 2 deal with technicalities of RCR interviews and their scoring, respectively. Part II can be read without first reading Part I, but understanding Part II fully might be easier after reading Part I.

Part I

The Theory of Relational and Contextual Reasoning (RCR) and its Empirical Study

1

Introduction

The object of this chapter is, first, to formulate a few caveats in order to lessen the risk of misunderstandings and disappointments, then to delimit the domain to be discussed, and above all, to lay the groundwork for subsequent considerations on Relational and Contextual Reasoning (RCR). This includes the basic nature of RCR, and the meaning of relational, contextual and reasoning, RCR’s underlying logic, its components and internal structure, and its status as postformal theory. There follows an empirical finding as an illustration of the principles set out so far. Finally, other forms of relational thinking and their importance for the present study are discussed before briefly summing up the chapter. Caveats No overarching grand theory exists of everything concerning psychological development of humans.1 Clearly, each of us often (a) perceives, (b) feels, (c) reasons, (d) plans, and (e) acts in an interrelated manner, and not only in mundane affairs of daily life. Yet, present psychological theories mainly deal with only one of the aspects (a) to (e) (or any other, like motivation, e.g., Reiss and Havercamp 1998); this despite their proponents’ awareness of the artificiality of such an isolating procedure. This work is no exception in that regard. It is neither a new nor a contested claim that thought and emotion are ‘inseparably’ linked (e.g., Piaget 1954/1981; Bearison and Zimilis 1986; Cacioppo and Gardner 1999, pp. 194–6). Nevertheless, emotions are very largely neglected here. Cognition (perceiving, appraising, understanding, reasoning, judging, remembering, imagining, etc.) and its development, the general subject matter of this work, is complicated enough. For that reason, I further restricted this work to the development of cognitive thought processes. 1

I write this notwithstanding Wilber’s (2000) A theory of everything, which is more an eclectic vision than an established theory. For the history and prospects of such a theory in physics – a culturally relative priority – see, e.g., Glashow 1980; Greene 1999; Weinberg 1992.

11

12

The Theory of Relational and Contextual Reasoning

Likewise, it seems incontrovertible that all thought processes are based on chemico-electric processes in the brain (e.g., Baars 1997; Clark 1997; Damasio 1994, 1999; Edelman 1992; Edelman and Tononi 2000; Gazzaniga 1992; Ramachandran and Blakeslee 1998; Putnam 1999, especially part 2), but again, neurobiology will not be treated here. Interested readers might refer, e.g., to Elman, Bates, Johnson, KarmiloffSmith, Parisi and Plunkett (1997, pp. 2–4, 239–317, passim) and Johnson (1998). Nor will there be much discussion of unconscious or preconscious processes although they play an important role in cognition (e.g., implicit knowledge, Holyoak and Spellman 1993, pp. 278–90; the cognitive unconscious, Lakoff and Johnson1999, pp. 9–15), particularly in its development. As a rule, a person is ‘embedded’ in the current developmental stage, that is, not fully aware of it and therefore not able to deal with it consciously: the person is the stage. When moving to the next stage, the structure of the previous stage becomes the content of the (structurally enlarged) new stage: the person is presently aware of the lower stage characteristics and therefore can have it, that is deal with it, differentiate its characteristics (Piaget 1971, §20iv; Kegan 1982, in particular pp. 146–8).2 Also, cognitive performance and development are not independent of the social context (e.g., Astin 1998; Monteil and Huguet 1999). While acknowledging that fact, social context is hardly dealt with here in any systematic fashion as far as discussing relational and contextual reasoning proper is concerned. The aim of the work described and discussed in this volume was to carry out enough basic research on RCR to enable its targeted effective application, and then to concentrate on applications in various fields; all the same, social context is included in appropriate cases. Given these caveats, we now turn to the basic nature of RCR so as to start with at least an elementary understanding of what will become clearer and more detailed in subsequent sections and chapters. The nature of relational and contextual reasoning Basic features Fully developed relational and contextual reasoning (RCR) is a specific thought form which implies that two or more heterogeneous descriptions, 2

In a different rubric, Gerald Cory (2000) presents a conflict systems neurobehavioural model of the brain: the protoreptilian brain (the evolutionarily oldest) represents the self-preservation programming of human behaviour, the mammalian additions (the limbic system, etc.) the affectional programming, and the (typically human) neocortex the executive programming of human behaviour, which notably co-ordinates the activities of the evolutionarily earlier brain parts, especially in case of (instinctive) conflicts between self-interest and other-interests.

Introduction

13

explanations, models, theories or interpretations of the very same entity, phenomenon, or functionally coherent whole are both ‘logically’ possible and acceptable together under certain conditions, and can be coordinated accordingly (Reich 1995a).3 Although the extent and intent of a given description, explanation, etc. per se play a role, that is less central to RCR than the co-ordination between competing explanations. Examples are the explanation of human behaviour by ‘nature’ (A) and by ‘nurture’ (B), the use of the ‘wave’ (A) and the ‘particle’ (B) picture when explaining light phenomena, the reference to technical malfunctioning (A) and human failure (B) as causes of accidents, the use of scientific (A) and religious (B) interpretations when discussing the origin and evolution of the universe and what it contains, or the investigation of psychophysiological phenomena (e.g., fright) in terms of introspection (A), outward behaviour (B), and physiological data (pulse frequency, skin resistance, etc. – C). As a category, RCR can be classed alongside Piagetian logicomathematical thinking (Piaget 1970), dialectical thinking (Basseches 1984; Riegel 1978), analogical thinking (e.g., Gentner and Markman 1997), cognitively complex thinking (e.g., Baker-Brown, Ballard, Bluck, de Vries, Suedfeld, and Tetlock 1992), systemic thinking (e.g., Chandler and Boutilier 1992), and so on – although to my knowledge this is the first book-length discussion of RCR and it is less established in academe than those traditional forms of thought. They will be compared with and contrasted to RCR in chapter 5. What is the meaning of ‘relational’, ‘contextual’, and ‘reasoning’ in the present context? Throughout this monograph, I refer to the entity, phenomenon, or the functionally coherent whole to be explained as the ‘explanandum’, and to the heterogeneous descriptions, explanations, models, theories or interpretations as ‘A’, ‘B’, ‘C’ . . . In a given case, A, B, C . . . may constitute a description, a causal attribution, a motivation for human action, a prediction, a presentation of a process, an analysis of a structure, an interpretation of meaning, a (partial) theory and so on; this in many knowledge domains. Relational concerns the relations between the explanandum and A, B, C . . . on the one hand, and the relations between A, and B, and C . . . themselves on the other. To anticipate: A, B, and C . . . are internally linked (entangled as understood in quantum physics) in cases where RCR is applicable, but mostly do not constitute a cause–effect relation in the 3

Initially, RCR was called ‘Thinking in terms of complementarity of “theories”’. The name was changed to RCR because, on the one hand, the appropriateness of this label was increasingly appreciated, as was, on the other hand, the ambiguity of the term ‘complementarity’, used very differently in everyday life, in ‘popular’ physics, psychological communication theory, psychotherapy, and so on (Reich in press).

14

The Theory of Relational and Contextual Reasoning

classical sense. The link can consist in mutual enabling or limiting, in an information transfer, or be of further types. Contextual involves taking into account the circumstances, the context of the situation. Regarding nature and nurture, context effects would show up, for instance, when comparing and contrasting the performance at two points in time, or, alternatively, two differing aspects of the same activity like playing the piano (e.g., playing without faults, or playing for the first time from the sheet music). To stay with the first example, if the performance of an athlete in good health and unchanged food intake varies over a matter of weeks, chances are that it has more to do with his or her exercising than with natural endowment. With reference to the wave picture and the particle picture of light, the one-slit (particlelike behaviour) and the two-slit (wave-like behaviour) experiments come to mind. As to accidents, the context may co-determine how technical malfunctioning and human failure condition each other. In regard to the interpretations of the origin and the evolution of the universe, the context could be a pure question of physics (physics contributes more) or the meaning and destiny of human life (religion has more to say). In all pertinent cases A, B, and C . . . have to be taken into account separately and jointly, but their explanatory potential usually varies with the context. As to reasoning, one can differentiate between (a) inferring, (b) thinking, and (c) reasoning (Moshman 1998, pp. 952–3). Inferring involves the generation of new cognitions from old, in other words to draw conclusions from what was already known but had not been ‘applied’. Inferring is often automatic and unconscious, for instance, when an infant, knowing that a toy can be in one of two locations, does not find it in the first location and immediately turns to the second (e.g., Sodian and Wimmer 1987). Thinking deliberately uses the results of inferences to serve one’s purpose, like making a decision, solving a problem, or testing a hypothesis. Given the object of thinking, it is possible eventually to evaluate the result. With experience, it may become clear which thought processes are more successful than others. By applying the corresponding reflections to thinking, it becomes reasoning. Moshman (1998, pp. 953–60) distinguishes different types of reasoning. RCR is a specific, and not a general type of reasoning, applicable to phenomena or events having the particular structure referred to above. Altogether, RCR can be understood as a pragmatic reasoning schema (Cheng and Holyoak 1985). Such a schema consists neither in a set of syntactic rules (e.g., mathematical algorithms) that are independent of the specific content to be treated, nor are they a recipe for one-off decisions such as choosing a profession or a partner, but consist in applying a set of rules for solving a particular class of problems. In the present

Introduction

15

case the issue is to ‘co-ordinate’ two or more ‘rivalling’ descriptions, explanations, models, theories or interpretations (procedure in chapter 6). This, irrespective of whether they are of the ‘nonconflicting’ type, or ‘contradicting’ each other. However, in all pertinent cases they differ categorically, are internally linked, and in a given context one has more explicatory weight than another. As already indicated, in the cognitive-conflict-resolving RCR world, such rivalling descriptions, explanations, models, theories or interpretations of the selfsame entity, phenomenon or functionally coherent whole are accepted which according to RCR logic do not exclude each other, but neither can be derived from each other. Next, then, the question arises, ‘What is RCR logic?’ Preliminary remarks on logic To avoid misunderstanding, I should probably state how I use the terms ‘logic’ and ‘logical’. There are two philosophical schools concerning the applicability of the terms logic and logical. For one school only the classical (Aristotelian) formal binary logic, including its modern symbolic version, is deemed to be universally valid, and therefore alone deserves the designation ‘logic’. All other rules about correct reasoning are termed ‘considerations of a philosophical or psychological nature’ (e.g., dialectical ‘logic’), ‘examples of a particular logical calculus’ (e.g., quantum ‘logic’), but not ‘logic’. For the other school, there exist many varieties of logic from deontic logic to transcendental logic.4 I take my cue from the second school, that is, for me a variety of logics exist. With Michael Basseches (1989, p. 171) I use ‘logic’ as ‘referring to principles and rules governing the proper use of reasoning’. As will be developed in subsequent chapters, there are indeed various types of logic, and depending on the problem under discussion, use of one type is more appropriate than use of another. For instance, in grading a test on arithmetic, formal binary logic is appropriate for assigning a mark (‘correct result’ or ‘wrong result’) to each separate result, but fuzzy logic (Kosko 1994) is usually appropriate for the resulting grading (very good, good, sufficient . . . or equivalent grades), not binary logic. In our Western culture, formal binary logic is held in high regard. One of its central rules is that in case of ‘contradictory’ distinguishing 4

Thomas Balmer (1982, pp. 109–10) writes, ‘There is even no guarantee that utterly incompatible logics may [not] exist, which, by themselves, are sound and complete. In fact, the historical development shows the way. Quantum logic, modal logics, fuzzy logics, context change logics (including dynamic logics), non-monotonous logics, etc. pave the way for quite varied logical systems, which may reach the high standard of classical and intuistic logic.’

16

The Theory of Relational and Contextual Reasoning

characteristics A and B (e.g., ‘wet’ and ‘dry’), a given entity can only have one or the other characteristic (the ‘law’ of identity), but not both. Higher stages of reflection among other things may lead to recognising the limits of applicability of that ‘law’ and similar ‘laws’. For example, before a measurement, light or an electron simultaneously ‘has’ both particle-like and wave-like characteristics (superposition of wave functions), that is, no clear identity as just defined. Here is a sampling of views on the stringency of the rule to avoid logical contradiction as understood by formal binary logic (cf. Wason and Johnson-Laird 1972; Holyoak and Spellman 1993, pp. 292–3), first in the words of two respondents interviewed in the study of RCR: ‘It is not logical [= there is a formal contradiction],5 but it is true [= empirically demonstrable]’, and ‘We know it for certain, but we cannot prove it [by applying formal binary logic].’ The American philosopher Ralph Waldo Emerson ([1841] 1903, p. 57) wrote: ‘A foolish consistency is the hobgoblin of little minds.’ The Danish philosopher physicist Niels Bohr (1958, p. 66) put it this way: ‘The opposite of a deep truth also contains a deep truth.’ As already indicated in the Introduction, RCR involves a trivalent logic; it will be discussed fully in chapter 3. The third ‘truth value’ (‘noncompatible’) refers to being ‘true’ in one context, but not (or at least much less) in another context. Does not a potential for resolving cognitive dissonance and conflict raise its head? Components of RCR RCR, while being distinct and having ‘unique’ characteristic features, shares structural ‘components’ with other thought forms. As will become clearer further on, these ‘sharing’ thought forms are (a) Piagetian thinking, (b) cognitively complex thinking,6 (c) dialectic thinking, and (d) thinking in analogies. How could one tentatively envisage such a state of affairs? One way to proceed would be to analyse in detail the structural ‘components’ of all forms of thought concerned, and look for overlaps. However, that would vastly exceed the scope of the present work. Instead, I propose a hypothetical model, which is speculatively based on some probability arguments. The model is not indispensable for the sequel, but it constitutes a heuristic framework for future work. The objective is to go beyond the observational features (to be described in chapter 4) and to represent the presumed underlying structure of RCR 5 6

As a rule, phrases in square brackets are my additions/commentaries. The denotation ‘cognitively complex thinking’ emphasises that the focus is on the complexity of a thought form, not primarily on the complexity of the problem structure.

Introduction

17

(and other forms of thought). The emphasis here is on structure, not on its development (although it is true that the structure constitutes itself and evolves from early childhood onward). According to Riegel and Rosenwald (1975, p. xiii), ‘Structures are relational organisations [that relate the different components to each other so that they function as a whole].7 . . . They are the properties that remain partially stable under transformations . . . Changes represent transformation of structures.’ To avoid a misunderstanding: ‘structures’ or ‘forms’ are not properties of a physical reality but the organisational configuration of mental activity (cf. Overton 1975).8 The arguments for the model we are discussing go as follows. (1) There are parallelisms between mental structures and brain structures (e.g., Baars 1997; Clark 1997; Gazzaniga 1992; Johnson 1998). (2) Given the difficulty of disentangling ‘directly’ the complexities of the functioning of the human brain, a more practical way is first to study and analyse one of its ‘productions’, and then (based on the results of those studies and analyses) assume that ‘related’ productions will have a comparable structure. (3) Language is one of the easier-to-get-at productions of the brain (cf. Deacon 1997). (4) Certain isomorphisms between evolving language ‘architectures’ and brain ‘architectures’ are assumed, and similarly for the ‘architecture’ of thinking.9 (5) ‘Language and thought 7

8

9

Structures fall into two broad classes (Overton 1975): (1) elementaristic structures, based on a view of the world as a static mechanism, understandable through analysing its parts and decomposable (linear) interactions between them (‘classical’ computer simulation approaches; information theory; decision theory; cybernetics); (2) holistic structures, based on a view of the world as a complex active organism, understandable through analysing the function of the substructures within the overall structure and the linked (nonlinear) interaction processes (Piaget’s theory; RCR; Bertalanffy’s system theory). Given the development of RCR through interaction with the environment, RCR may be considered as an open system rather than a closed system. According to Terrence W. Deacon (1997, cf. Cavanaugh 1999), another example of a hypothesised structural relation between the brain and its ‘productions’ would be the triad ‘iconical’, ‘indexical’, and ‘symbolical’ representations. For instance, to signify iconically ‘stop moving’, one can depict a policeman’s outstretched hand. That sign is so close to the actual everyday situation of a policeman’s action that the basic brain can immediately understand and react. In contrast, to recognise the standard octagonal stop sign from a distance as an invitation to stop there, asks more from the brain. On the basis of the iconical representation of a policeman’s hand, it has to learn that this octagonal indexical representation has the same meaning. Finally a symbolic representation would combine indexical representations and thus be even more abstract, requiring an even more powerful brain for producing and understanding it. Let us take the symbol of the raised hand when swearing an oath: while the gesture resembles that of the policeman, the meaning is entirely different. The person swearing the oath asks the heavenly powers for help to keep his or her promise (Deuteronomy 32: 40). What, then, is the ‘architecture’ we are looking for? In the case of language, Sidney Lamb (1999, p. 28) considers two different, independent hierarchies of units of different sizes:

18

The Theory of Relational and Contextual Reasoning

Table 1.1 The four structural levels of the model of thought processes. Forms of thought (Piagetian thinking; dialectical thinking; RCR, etc.) are posited to be a combination of operations from three lower structural levels. Structural level

Example

1: Elementary operations

discerning a particular item or event within a larger whole recognising a relationship between two entities analysing the nature of a relationship Piagetian operations, RCR

2: Conjunctive operations 3: Composite operations 4: Complete thought form

correlate in a most realistic manner: language and the biological “hardware” (or more adequate: wetware) of our emotions, sensorimotor controls, and thought are in a tight one-to-one correspondence’ (Ballmer 1982, p. 6). The general scheme is conceived as shown in Table 1.1. At the lowest structural level of thought processes, a large number of elementary operations exist, such as ‘discerning a particular item or event within a larger whole’. When two or more of these elementary operations are combined, thought processes reach the second structural level, which is termed conjunctive operations. An example of conjunctive operations would be ‘recognising a relationship between two entities’. At the third structural level of thought processes, which is termed the composite level, a number of elementary and conjunctive operations work together, as, for example, in analysing the nature of a particular relationship. Specific forms of thought such as Piagetian thinking, dialectical thinking, or RCR, the fourth structural level, are agglomerations of operations from levels (a) phonological units (phonon [quantum of sound energy], phoneme [set of similar speech sounds], syllable, wordP , phraseP ), and(b) grammatical units (morphemes [distinctive arrangement of phonemes that contains no meaningful smaller parts], wordG , phraseG , clause, sentence, discourse/text). Both sets evolve from the elementary to the complex. Thus, if only a general model of the architecture is wanted, either can serve. However, it would be preferable to have also an indication for a ‘reasonable’ number of structural levels. Disregarding the phonon (so to speak a white noise of language production), there are four levels of phonological units. In contrast, the analytical philosophers discern six levels of grammatical units. My impression is that the phonological units, while abstractions (excepting phonon), are closer to ‘natural’ occurrences, to the steps of children’s learning a language (Bloom 1998), than the even more abstract grammatical units. Also, whereas ‘clause’, ‘sentence’, and ‘discourse/text’ are of course different, perhaps less so than the difference between wordG and phraseG . Therefore – without losing sight entirely of the more numerous grammatical units – I posit that the architecture of thinking can be conceived of as a structural four-level model, in analogy to the ‘architecture’ of phonemes, syllables, wordsP , and phrasesP.

Introduction

19

1, 2, and 3.10 It is thinking and reasoning at that fourth structural level which will actually be studied empirically (chapter 4).11 In case reality is somewhat different than hypothesised in Table 1.1, that should not influence the conclusions of this monograph, given that they are not built on the details of the model, such as the exact number of levels, but above all on the ‘overlap’ of the various forms of thought, on their sharing of ‘components’. RCR, a postformal theory In the early eighties, a dissatisfaction with Piaget’s theory (of the development) of thinking grew. The formal features of Piaget’s theory (Piaget 1970) were thought to preclude appropriate reasoning about social issues, and about psychological questions like will, imagination, and creativity. This led to a series of alternative theories, in particular for adult development (e.g., Commons, Richards, and Armon 1984; Commons, Sinnott, Richards, and Armon 1989; Commons, Armon, Kohlberg, Richards, Grotzer, and Sinnott 1990; Alexander and Langer 1990). I consider the theory of fully developed RCR to be a postformal theory. Although not necessarily claimed explicitly, postformal operations also attempt to overcome the limitations of formal binary logic. John Broughton (1984) discussed some of the numerous proposals made in the four volumes just listed. In his view, none represents a good solution likely to become generally accepted. The common effort of the 1980s has not had the success hoped for. In the 1990s, publications on postformal reasoning are rarer – and more radical. For instance, Kincheloe and Steinberg (1993) formulate a theory of postformal development that 10

11

Michael Basseches (1980 p. 408; 1984 p. 74; 1989 pp. 162–3) tentatively lists twentyfour elements of dialectical thinking, which he terms ‘dialectical schemata’. In an exploratory exercise, I had no difficulty in classing those schemata as level 1, 2, and 3 examples, simply using the criteria of ‘elementary schemata’, ‘schemata based themselves on elementary schemata’, and ‘meta-level schemata’. Interestingly, Benack and Basseches (1989, p. 97) order twenty-two of the twenty-four schemata into three developmental phases of ‘emergence’, the last two schemata appearing in the fourth, the final phase. However, after probing empirical studies, Irwin and Sheese (1989, p. 126) concluded that ‘dialectical thinking as a whole [thus posited by Basseches] does not appear to be an empirically robust phenomenon.’ Also, while all schemata were manifest in the study by Irwin and Sheese (1989, pp. 121–6) some of the schemata did not appear to have a time-dependent ‘emergence’ pattern (cf. note 3, p. 56). Unfortunately, I know of no other, more established elements of a thought form, which could be used for a more convincing demonstration of structure levels 1 to 3 of Table 1.1. The unfolding course of an individual’s cognitive development alters the contents of the complete set of thought operations outlined in Table 1.1 in two significant ways. First, new operations appear in the course of cognitive development (e.g., Reich, Oser and Valentin 1994) and second, existing operations become more complex, more sophisticated (Fig. 2.1, p. 30).

20

The Theory of Relational and Contextual Reasoning

is informed by and extends critical, feminist, and postmodern thought. Liu and Liu (1997) advocate adopting Confucian thinking, a philosophy of integration and subjective identification, in order to overcome the philosophy of separation and objectification (which underlies Piagetian formal operations). Jan Sinnott (1998) researched the question of how some adults grow wiser as they grow older. Her answer is, in a nutshell, that they have learned how to ‘live in balance’, which is said to mean especially ‘to orchestrate the needs of mind, heart, spirit, body, and others in their lives, despite the conflicting demands that all of us face in adulthood’ (ibid., p. vii). ‘Self-reference’ constitutes an essential ingredient of postformal thought according to Sinnott (ibid., pp. 33–4). Aware that we are all trapped in partial subjectivity, postformal thinkers are said to make a decision about the rules of the game (nature of truth), then to act on the basis of those rules. There are clear commonalities between Sinnott’s objectives and those aimed at in this monograph such as tackling societal problems more effectively, and improving education by means of more appropriate thought forms (Reich, 1996a, 1997, 1999). It is undisputed that in one’s life, situations may arise where no kind of logic is of much help, and self-reference is unavoidable. However, the assumptions made here about matching the thought form to the problem structure (chapter 5) lead one to believe that such situations are the exception rather than the rule. In the overwhelming majority of cases it should be possible to find a form of thought (with its particular logic) that fits the structure of the problem itself; this by means of ‘objective criteria’, as will be discussed in chapters 3 to 5. Having come to the end of the introductory presentation of fully developed RCR, we turn to an illustration of the empirical research. An empirical finding and its first analysis Having tentatively postulated RCR, I studied it empirically through analysing interviewees’ responses to certain problems (chapter 4).12 To illustrate the principles set forth so far, let us look at an excerpt from an interview about a Three Mile Island/Chernobyl type of accident in a nuclear power station (Reich 1995a, p. 2): A TV news station reports on an accident in a nuclear power station. The main cooling pump had stopped working, and the back-up pump did not function. The emergency shutdown did not function either. To add to the difficulties, the operating crew became aware of the danger rather late and then underestimated it. The 12

For years, when I described my work, the reaction was, ‘Oh, you study dialectic thinking . . . No? But what is the difference?’ To answer that question convincingly was one of my early objectives.

Introduction

21

water temperature suddenly rose. A steam pipe cracked and leaked radioactive steam. What or who is to blame? What should be done to avoid another such accident in the future?

A middle-aged adult, whom we call Bertrand, exemplifies RCR: In this accident technical and human failure are interconnected. One has to look at the whole thing as a system, the plant and the operating crew. And one has to study the mutual interaction, the type of effects they have on each other. One really wants to train crew members with the help of a sophisticated simulator so that they become aware of the many ways in which something can go wrong, they experience their individual and collective reactions, and learn how to assess such situations as well as how to deal with them successfully. In such simulations the psychological stress must of course also be generated, not just the sequence of technical events. It is precisely such a chain reaction of technical and human malfunctioning which is so hard to foresee. By the way, I would hire only such persons who are aware of the dangers involved and are ready to face them.

What is special about Bertrand’s response? In what ways does it exemplify RCR? Let us proceed sentence by sentence: In this accident technical and human failure are interconnected. That opening sentence is mental miles away from the classical ‘It was a technical malfunctioning’, or ‘It was a human failure.’ Instead of settling for one or the other of those usual alternatives, Bertrand, searching his enlarged horizon, emphasises that both causes have to be taken into account. In fact, he goes further by stating that they are not independent of each other, but are interconnected. One has to look at the whole thing as a system, the plant and the operating crew. In his second sentence Bertrand reinforces the connection between the technical and the human aspects of the accident: they together in their ‘systemic’ structure and functions account for the event in its various phases. If the accident is to be analysed in depth, the technical and human occurrences have to be considered not only individually, but also together, in their interrelationship. And one has to study the mutual interaction, the type of effects they have on each other. Bertrand implies that the possible effects are of several types. In this and other interviews the following types were evoked: (a) information from the (mis)behaviour of the power plant displayed in the control room and the resulting (re)actions by the operating crew; (b) controlling actions by that crew and their effect on the plant; (c) effects of (abnormal) plant behaviour on the (unintended) emotional response of the crew; (d) (resulting) group dynamic effects within the crew. One really wants to train crew members with the help of a sophisticated simulator so that they become aware of the many ways in which something can go wrong, they experience their individual and collective reactions, and learn how to assess such situations as well as how to deal with them successfully. This

22

The Theory of Relational and Contextual Reasoning

sentence shows that Bertrand is aware of the complexity of the respective behaviour of the plant and of the crew (individually and collectively) as well as the connections between both behaviours and their mutual conditioning. In such simulations the psychological stress must of course also be generated, not just the sequence of technical events. Here Bertrand emphasises the context dependence of crew behaviour. While learning about possible technical malfunctioning in a quiet atmosphere is of value, that is not enough. The crew needs also to be trained to stand up to the ‘heat of the battle’ and to overcome undesirable compulsions such as to panic and flee. It is precisely such a chain reaction of technical and human malfunctioning which is so hard to foresee. Bertrand insists once more on the undissolvable connectedness of technical and human factors. It is by far not enough to study and deal with them separately; a major effort needs to be made to understand and optimise their interconnection and interaction. By the way, I would hire only such persons who are aware of the dangers involved and are ready to face them. As his final point, Bertrand addresses the insight and the resulting attitude of the operators: it is not just their technical know-how and social-cognitive competence that are needed for a safe and efficient operation of the plant, but also an appropriate insight and attitude. What does this analysis already confirm about the principles of relational and contextual reasoning as set out above? (a) Bertrand relates two distinct, categorically different causes to the accident and to each other. (b) He brings in and explicates the context dependence of the ‘behaviour’: the functioning of either the plant or the crew is not the same during normal operation and during the various phases of the accident; nor will operators react identically in a quiet operation period or instruction session and in a dramatic accident situation. (c) He deals with the problem at a high level of cognitively complex thinking, and brings in considerations which go beyond the TV news. Points (a) – relations – and (b) – context – illustrate why relational and contextual reasoning got its designation, and point (c) characterises RCR as reasoning aiming at knowledge leading to deeper insights. Furthermore, point (b), context dependence, points to the fact that a logic different from formal binary logic (which is invariant with respect to time and space) is involved. That difference of necessity will occupy us repeatedly. Previous work on relational thinking I am not aware that relational thinking as defined above was discussed prior to this work. Clearly, Piagetian logico-mathematical thinking, dialectical thinking, analogical thinking, and cognitively complex thinking

Introduction

23

each deal with relations (see chapter 5), and systemic thinking even more so. However, these relations exist between particular entities, not primarily between an explanandum and competing heterogeneous descriptions, explanations, models, theories, or interpretations – nor between the latter, in case they exist. Furthermore, the nature of the relations is different. In tasks where Piagetian thinking is applicable, the entities in question are intrinsically independent from each other, and the relations are governed by formal binary logic. Dialectical thinking deals with entities that determine each other and are internally linked in such a manner that a negation of a negation leads to a new situation, involving a logic different from formal binary logic. Analogical thinking is closer to RCR in that two heterogeneous descriptions, explanations, models, theories, or interpretations are used to illuminate mutually two explananda. However, the relation between the two descriptions, etc., is different again, and rather specific: they need to share a certain number of functions (and of attributes) to make the analogy work. Problems to which cognitively complex thought could usefully be applied, may have a variety of internal structures. Systemic thinking in a narrower, more specific sense concentrates on the relations (linear or nonlinear, circular, etc.) between systems components and output, including multiple pathways, and feedback loops. Thus systemic thinking deals with complexity in both reductionistic and holistic ways (Chandler and Boutilier 1992). RCR involves some of that (in particular in the example of the nuclear accident), but not specifically a technical analysis of feedback loops and the like. I prefer to deal with the relevant aspects under the label ‘cognitively complex thought’, and to concentrate on the operations of differentiation and integration involved. In contrast to all of the above, RCR deals with a variety of (mostly noncausal) relations between such A, B, C . . . that are intrinsically linked and where the link – as already mentioned – can consist in mutual enabling or limiting, in an information transfer, or be of further types. Nevertheless, the findings concerning these other thought forms had an impact on the present study because of their sharing of components with RCR. Furthermore, developmental features were taken into account when designing the empirical work (see chapter 2). Summary of the introduction To sum up this introduction: relational and contextual reasoning (RCR) is a specific thought form which implies that two or more heterogeneous descriptions, explanations, models, theories, or interpretations of the very same entity, phenomenon, or functionally coherent whole are both ‘logically’ possible and acceptable together under certain conditions, and

24

The Theory of Relational and Contextual Reasoning

can be co-ordinated accordingly. The meaning of the three terms relational, contextual, and reasoning was explained, and the type of underlying logic was indicated as well as the mental structure of RCR. I explicated why fully developed RCR is a postformal form of thought. These principles thus set out were illustrated by empirical findings. Other thought forms involving some kind of relational thinking were shown to be rather different, yet of interest here, in particular because they share components with RCR. Following Gerhard Schurz (1998, pp. 10–11), in subsequent chapters I shall explicate and discuss RCR (a) theoretically, as a scientific model (chs. 2, 3); (b) empirically on the basis of research results (chs. 4, 7), (c) methodologically, i.e. as a procedure, the RCR heuristic (ch. 6); and (d) programmatically, that is, in terms of possible future applications (chs. 9–11). Also on the agenda: matching the thought form to the problem at hand (ch. 5), and the archaeology of RCR (ch. 8). I thereby hope to demonstrate further the value of RCR: rather than turning every interpretative dispute into a war of attrition, it works to hold competing accounts of certain types of issues in an embrace which is sometimes painful to endure, but often fruitful in the end.13 13

In the case of the theory of light, there were 200 years of competition between the particle theory and the wave theory (Baierlein 1992) before the fully satisfactory, consensually accepted theory of quantum electrodynamics (QED) emerged (Feynman 1988). QED not only solved the problem of explaining the behaviour of light, but contributed also to changes in the philosophy of knowledge. ‘Thus quantum mechanics is a wonderful example of how with the development of knowledge our idea of what counts even as a possible knowledge claim, our idea of what counts as even a possible object, and our idea of what counts even as a possible property are all subjects to change’ (Putnam 1999, p. 8; emphasis in original). What a development of the horizons of the mind (triggered by researching natural phenomena)!

2

Development of RCR

To refocus on Bertrand’s response reproduced in the Introduction (p. 21): obviously, it could not have been provided by a seven-year-old. Such a child cannot be expected fully to understand that TV news item, let alone respond to it in a sophisticated way. It will come as no surprise that RCR develops from a rudimentary beginning to intermediate levels before reaching the quality of Bertrand’s response. To anticipate (chapter 4), there are five developmental levels: (i) only one description / explanation / model / theory / interpretation can be right, the other(s) must be wrong; (ii) maybe, there is something valid to both (all) of them; (iii) both (all) are definitely needed to account for the phenomenon under study; (iv) here is how they are related to each other; (v) the overarching synopsis is as follows. In this chapter, I first evoke the anthropology adopted, put in place the developmental background with some general remarks on cognitive development, explicate RCR development against that background, discuss unreflected, object-reflecting, and means-reflecting thought as a particular, not so well-known feature putatively inherent in the development of RCR, continue with the Piagetian concept of intra-inter-trans – the ‘logic’ of RCR development – and indicate the impact of previous developmental work on the design of the present study before summarising. Anthropology adopted Put simply and ideally, the philosopher Daniel Dennett (1996, pp. 83– 101) discerns and describes the following anthropological elements: (1) the ‘native endowment’ has evolved through the ages by way of gene mutation and ‘field tests’; thus humans – like all other living beings – are Darwinian creatures. (2) Some of the mutations led to wired-in ‘reinforcers’, mechanisms that favour actions more beneficial for the individual or the species than the alternatives. Such individuals tried out various kinds of behaviour and reacted to positive or negative signals from the environment by selecting the most successful behaviour and henceforth 25

26

The Theory of Relational and Contextual Reasoning

going on with it. Thus Darwinian selection was continued by operant conditioning of Skinnerian creatures. Animals and humans learn through associations of ‘ideas’, behaviouristic learning, and connectionism (adjustment of nerve cells as a result of conditioning) – ABC learning. It links natural endowment per se and experience. (3) The next step in cognitive development involved the building up of an ‘inner environment’, in which various surrogate actions could be tried out mentally before taking risks in the real world. Dennett writes that we are also Popperian creatures. (4) Finally, ‘mental tools’ were developed, for instance words, numbers, symbolic logic, and eventually virtual cyberworlds, which all extend the power of the human mind. In honour of the British psychologist Richard Gregory, Dennett refers to humans also as Gregorian creatures. A conclusion is that a complete psychological investigation should in turn look at the biological grounding, the person-centred (conscious and unconscious) factors, and the bio-physical and socio-cultural environments involved in human development, and integrate the findings (Overton 1998, 1999). Theories of cognitive development What is it to be a human thinker and knower? Which knowledge is there ‘from conception’, which is abstracted from experience? Those and related questions have been debated by philosophers for many hundred years, and more intensely with the arrival of psychology (cf. Spelke and Newport 1998). Broadly speaking, psychological theories of cognitive development can be classed under three headings: (1) endogenous theories (development originating from within, e.g., maturation of native endowment), (2) exogenous theories (development originating from without, e.g., socialisation), and (3) interaction theories (development results from interactions both within the organism itself and with the bio-physical, social, cultural, and perceived spiritual environment). What is the meaning of ‘interaction’? In everyday language, it will probably be conceived of as reciprocal ‘causation’. Such an interpretation would apply, for instance, to the moves of two equally strong football teams during a game. The two teams are independent of each other, they are ‘separable’. However, each of their moves either is made as a reaction against an actual move by the opposite team or is planned in view of the anticipated reaction of the adversary. In a looser manner than that just indicated, interaction could signify interdependency of determinants or conjunctive plurality of causes. An example of a ‘loose’ interactive plural causation would be the causal net leading to an influenza as described by psychoneuroimmunology: the type of virus

Development of RCR

27

involved and its quantity, the state of the psyche, of the nervous and of the immune systems, all play their causal role in an interacting manner. Anastasia’s (1958, p. 197) use of interaction is different still; for her, ‘the nature and extent of the influence of each type of factor [native endowment vs. ‘independent’ acquisition] depend upon the contribution of the other’. This means that the factors are intrinsically linked with each other, they are ‘non-separable’. A given state of one factor unavoidably co-determines (for the most part, limits) the effect of the other factor, yet one cannot speak of causal interaction in the same sense used in referring to the football game.1 Traditionally a debate took place between nativists and empiricists. Elisabeth Spelke and Elissa Newport (1998, pp. 321–9) review the arguments pro and contra nativism in the areas of action, perception, language, and reasoning. While the respective portions of the inherent and the acquired, the inevitable and the coincidental, the constant and the changeable, the universal and the variable differ, all have to be taken into account in order to understand development. After reviewing various theories about children’s knowledge of the mind (the theory theory, the modularity theory, the simulation theory, and others) John Flavell (1999, p. 27) makes the judgement that an adequate theory will finally have to include elements from each of these perspectives . . . (a) that development in this area builds on some innate or early people-reading capacities, (b) that we have some introspective ability that we can and do exploit when trying to infer the mental states of other creatures . . . (c) that much of our knowledge of the mind can be characterised as an informal theory . . . [(d) statements about certain specifics regarding theory of mind], (e) that a variety of experiences serve to engender and change children’s conceptions of the mental world and explaining their own and other people’s behavior.

Although none of these considerations concerns RCR explicitly, and research on the respective contributions of the inherent and the acquired to the functioning and development of the mind (e.g., Perner 1996) continues vigorously (Keil 1998, p. 388), I assume that development of RCR is based on some innate potentialities, and develops within the given constraints through various interactions. Cognitive development and RCR We now come to a sketch outline of cognitive development as it pertains more closely to RCR, a thought form aimed at improving knowledge and 1

To anticipate, the interaction can therefore not be dealt with satisfactorily by formal binary logic (nor analysed by certain linear methods, cf. Overton and Reese 1972).

28

The Theory of Relational and Contextual Reasoning

insight through more adequate reasoning in terms of ‘theories’ about particular ‘data’, that is, evidence. If one differentiates between (a) the development of descriptive knowledge about objects or events in our world, and (b) attempts to explain their structures and processes/functions, then RCR connects both. From birth we all gain descriptive and procedural knowledge by observing and interacting (often in the sense of mutual causality) with our bio-physical and human surroundings. And then we want to understand and explain, at least to ourselves, what we experience; this often in structural and/or functional terms. One result is a continual reconstruction of the knowledge base (Campbell and Bickhard 1986), and an improvement in (logical) thinking (e.g., Reich, Oser, and Valentin 1994). While these developments are interrelated, and are in fact inseparable, it is useful to make the following conceptual distinctions for the purpose of studying and discussing them. Ontological development concerns the (perceived) existence or nonexistence of various entities and their predicates, more precisely the material categories needed to discuss those predicates. Examples include, ‘Do fairies, quarks, or unicorns exist or not?’; ‘Is that kind person who gives me presents really my uncle or not?’; ‘Are clouds alive or dead?’ As is well known (e.g., Keil 1998; Wellman and Gelman 1998), young children (pre-schoolers) may take years to come fully to grips with such issues. There are four reasons for this: (a) they are understandably inclined to look primarily at the exterior striking features (as distinct from the ‘inner’ or abstract characteristics that are not infrequently used as definition by adults, e.g., metabolism for being alive); (b) they start from their own experiences and make analogical inferences not admitted by adults (‘as a child, I thought that God eats or drinks because I ate and drank’); (c) they often concentrate on just one aspect, presumably due mostly to their limited working memory; and (d) they assume that everybody has the same knowledge and understanding as they have, and therefore do not feel the need to formulate and discuss their views to the extent that older children, adolescents, and adults do (Carpendale and Chandler 1996). Is there evidence that ontological categories and their predicates are not just a philosopher’s playground but have some psychological reality? Michael Kelly and Frank Keil (1985) compared Ovid’s metamorphoses to the fairy tales written up nearly 2,000 years later by the brothers Grimm. They coded all transformations that occurred in terms of the ontological categories implicated in each metamorphosis (e.g., using the Gorgon’s [Medusa’s] head to change a throng of assailants into stone statues). The result was that conscious beings were transformed mainly into animals (51 per cent vs. 52 per cent), into plants (10 per cent both), non-living

Development of RCR

29

entities (12 per cent vs. 11 per cent), and into liquids (5 per cent vs. 4 per cent), the first percentage number referring to Ovid, the second to the Grimms. A first striking feature is the near constancy of the percentages through the centuries and across (Mediterranean and Nordic) cultures. Second, the hypothesis, that the number of transformations will decrease as the distance between ontological categories increases, is supported by the analysis. Thus, there is some evidence for the existence of ontological categories and even of a ‘family tree’ which orders their degree of closeness. Whereas fantasy can imagine just about any transformation, the ontological tree constrains their number – even during the ‘willing suspension of disbelief ’. Logical arguments are used to elaborate the ontological tree. Logical development has to do with acquiring competence in classical logical operations where applicable (like making a valid inference, making use of transitivity, arguing by means of a logical implication), and gaining knowledge about logical quantifiers and their use (e.g., all, none, some; cf. Putnam 1999, pp. 57–8). It also involves coming to grips with modality logic (necessity, possibility, ‘all’ statements, ‘there exists’ statements – Chinen 1984). Higher developmental stages involve (intuitive) knowledge of other logics such as dialectical logic or RCR logic (chs. 1, 3, 4, 5). Unreflected, object-reflecting, and means-reflecting thought Another perspective for looking at major milestones of cognitive development putatively relevant for RCR consists in examining the degree to which someone is capable of ‘turning back on one’s own thoughts’ (Fetz, Reich, and Valentin 2001). Young children do observe, think, deduce, conclude, and order their knowledge within particular frameworks (e.g., Case 1998). They can be ‘sophisticated’ explorers and experimenters – yet their thinking is basically unreflected. They are not likely to examine their knowledge critically, that is, to turn their thoughts back unto their own thought (e.g., Kuhn 1999). A tentative model of the development from unreflected thought to (domain-dependent) reflections about ‘real’ objects and on to reflecting on one’s cognitive tools is shown in Fig. 2.1. Broadly speaking, (a) refers to the situation in early childhood, (b) to middle childhood/early adolescence, and (c) to adolescence and young adulthood. In all cases, the environment is perceived at knowing level 1 (Campbell and Bickhard 1986, p. 53). The higher knowing levels, to be explained shortly, arise from lower levels via reflective abstraction (if at all – for difficult issues the

30

The Theory of Relational and Contextual Reasoning

c Reflecting about mental tools c

b Reflecting about real objects b

Reflecting about real objects c

Unreflected thinking a

Unreflected thinking b

Unreflected thinking c

Level 1a

Level 1b

Level 1c

Environment

Environment

Environment

a

Figure 2.1 Evolution of cognition aimed at ‘seizing up’ the environment (perceived reality) in the course of age-related cognitive development. (a) early childhood, (b) middle childhood/early adolescence (onset of reflecting about ‘real’ objects), (c) adolescence and young adulthood (reflecting about objects and mental tools). Explanation in text.

age group indications are lower limits). They then interact with the lower level(s). Thus, development does not just consist in adding levels, but also in a transformation of existing levels (indicated in Fig. 2.1 by adding a, b, c to the level designation). Whereas the basic structure of Fig. 2.1 is taken from the work of Campbell and Bickhard (1986), the content of the upper stages results from our own work (Fetz, Reich, and Valentin 2001). That work concerns notably the rather difficult, abstract concept of God and its development. Figure 2.1, and in particular the age indications, therefore may not be universally applicable. However, it is posited to apply to RCR, and noted that RCR is of interest for gaining deeper insights into complex, controversial issues. The first milestone of turning one’s thought back on one’s own thought is ‘object-related reflection’. Such reflections notably include discovering

Development of RCR

31

contradictions in one’s world view, for instance (in the words of a bright eleven-year-old), ‘If the world is really infinitely large, it cannot have been made by God, because that would have taken infinite time.’ At one point of cognitive development, fairies, Father Christmas, the Easter Bunny etc. are recognised one by one as figures out of a child’s world, not from the ‘real’ world of adults. The next major milestone in the development we are discussing involves a ‘reflection on one’s mental tools’ for arriving at particular concepts, for instance the use of analogical reasoning for determining the attributes of God: ‘When I was a child, I observed how masons and carpenters built a house, and I was sure that God made the world that way: he made a blue-print, got materials, and did the work according to the blue-print. Now I know that such a conclusion is not warranted [= God is not as I imagined].’ This illustrates a first stage of ‘means-related reflection’, dealing with single concepts. The following developmental milestone, an ‘extended mental meansrelated reflection’, is reached in adulthood (and not shown explicitly in Fig. 2.1). It concerns epistemology in its entirety such as the limitations of our mental power: ‘A ruler can’t measure its own width. Similarly, there are things we simply can’t know. That is why I concentrate on things we can know.’ That empirically demonstrated level of cognitive development (Reich, Oser and Valentin, 1994) shares features with Karen Kitchener’s (1983, p. 225) stage of epistemic cognition, defined as ‘the process an individual invokes to monitor the epistemic nature of problems, and the truth value of alternative solutions. It includes a person’s knowledge about the limits of knowing, the certainty of knowing, and the criteria for knowing’ (King and Kitchener 1994, p. 12; cf. Moshman 1998, pp. 964–5). The development indicated is not about taking into account another person’s mind and being aware that he or she (a) is making a factual statement, (b) is joking, (c) commits an error, (d) is trying to deceive, etc. – four-year-olds can do that. However, they nevertheless believe that all persons having the same information will have the same views about it. Reflection on mental tools may start when one discovers that, where one person sees a vase, another sees two faces although each time both persons look at the same drawing, or where one sees an old woman, somebody else sees a young one (the well-known figure–ground shifts occurring when looking at those particular drawings). Or, when someone discovers that ABC evokes the alphabet for one person, and the American Broadcasting Corporation for another, and so through to Zenith, the highest point reached in the heavens by a heavenly body for one person, and a brand name of home electronics for another. Arthur White gives the following example: ‘“Density equals mass divided by volume,” a child

32

The Theory of Relational and Contextual Reasoning

is told – and immediately tries to relate this to a firsthand experience attending Sunday “mass”, turning the “volume” knob on the radio, being called “dense” by a sibling’ (White 1999). Reflection on mental tools may well start with questions such as ‘How come we see this differently?’, ‘How is that possible?’ Continuing from there such reflection may lead to differentiation between unreflected perceptions and even unreflected thoughts on the one hand, and a careful examination of the methods used to gain a particular knowledge and to test its veridicality on the other hand. Intra-inter-trans – the ‘logic’ of RCR development Along with seeking the causes of events, infants and children also widen and deepen their knowledge by (a) exploring further afield the nature of the objects or events under study, and (b) by relating objects or events to other objects or events in increasingly sophisticated ways (which is also characteristic of RCR development). Jean Piaget (Piaget and Garcia 1983/1989) describes such a sequence as intra-inter-trans. Jacques Montangero and Danielle Maurice-Naville (1994/1997, pp. 127–9) comment on those terms as follows: [We are discussing] a mechanism that leads from intra-object (object analysis) to inter-object (analysing relations . . . ) to trans-object (the building of structures) levels of analysis . . . ‘it is heavy and red’ is an example of intra-object relations, ‘putting an object’ into two jars simultaneously [= repeatedly one into each jar] and understanding that this produces an equal number of objects in the two jars exemplifies intra-operational relations. As to inter-object relations, natural embeddings (simple class inclusions) come to mind.

The INCR group (the 4 group) structure2 constitutes a transoperational relation in that two forms of partial reversibility (correlation and reciprocity) are now related with full reversibility (inversion, also called double negation) within a single structure (see below, ch. 5, pp. 82–3). 2

The INCR group structure of transformations may be represented graphically by a square (or a rectangle) and two diagonals (see Fig. 5.2, p. 83). Proceeding clockwise from the upper left-hand corner, the disjunction of p and q (p v q), the disjunction of nonp and nonq, (¬p v ¬q), the conjunction of nonp and nonq (¬p · ¬q), and the conjunction of p and q (p · q) occupy the four corners. The diagonals represent the operation of negation, N: it transforms p into nonp (¬p) and q into nonq (¬q) (and vice versa) and a disjunction into a conjunction (and vice versa). The negation of a negation leads back to the origin; the identity, I, is re-established. The vertical sides represent correlational operations, C: disjunctions become conjunctions and vice versa. The horizontal sides represent reciprocal operations, R: p becomes nonp, q nonq, and vice versa. Thus, C followed by R (or R followed by C) is equivalent to N. The basic idea is that a person mastering formal operations visualises the relationships between these transformations without undue difficulties (more in chapter 5, pp. 80–3).

Development of RCR

33

An everyday example for a development of the intra-inter-trans type would be the evolving understanding of a savings and loan association (Claar 1990). Asked how such an institution works, younger children in Germany answered that ‘they keep your money safe’ (intra), older children opined that ‘they keep some people’s money, and lend it to others’ (inter), and adolescents explained that ‘it’s a system where you get less interest for your deposits than you have to pay for your loans, and that keeps it going’ (trans). A claim about the sequential order of RCR developmental levels from level I to level V is that they progress necessarily from unifocal ‘coordination’ to bifocal (or trifocal) co-ordination, and on to ever-more elaborated bifocal/trifocal co-ordinations, that is from intra to inter to trans. This claim is built on two theoretical arguments: that (1) cognition develops along with the growth and increasing complexity of the brain,3 and (2) each level is logically prior to the subsequent one. Further details of that claim and its empirical support are presented in chapters 4 and 7. Input to the present study from earlier work Whereas the present study is not a direct continuation of earlier work by others, it nevertheless benefited therefrom in various ways. As already mentioned, Piaget’s intra-inter-trans sequence (Piaget and Garcia 1983/1989) provided an overall framework for researching RCR developmental levels. The theory of Campbell and Bickhard (1986) helped us to understand certain aspects of RCR development, such as the role of the increasing sophistication of the nature of knowledge. Thanks to Kitchener (1983), the importance of epistemic cognition was appreciated. The stages of Piaget’s logico-mathematical developmental theory were a continuous challenge in that they share aspects with RCR (such as the developing competence to understand the relation between two ‘entities’), but also differ (e.g., in the underlying logic). The initial research design (pilot study 1, ch. 4), was influenced by several earlier findings. Nicolls (1978) studied how five- to thirteen-yearolds explain ‘attainment’ by the logically not independent concepts of effort and ability. For younger children effort was the prime cause, older children used the concept of ability in addition (intermittently), and for young adults outcomes were seen as determined jointly by effort and 3

At the microlevel the totality of cognitive development can nevertheless be conceived as a network of multiple pathways of structural changes (cf. Table 1.1, p. 18), as distinct from a single path from Piagetian logico-mathematical thinking to a hypothetical ‘fifth stage’ (beyond the sensorimotor, the preoperational, the concrete operational and the formal operational stages – Piaget 1970).

34

The Theory of Relational and Contextual Reasoning

ability. Mansfield and Clinchy (1985, see below, ch. 5, pp. 91–2 for details) studied children’s epistemology and observed the early growth of multiplicity of judgements with 3- to 10-year-olds. Broadly speaking, the development was from ‘only one position is right’ to ‘other positions are acceptable’. Apart from a general encouragement, these studies, as well as Piaget’s theory, advocated starting empirical work with 6-year-olds (and to design the problems accordingly) for catching the beginnings of RCR, to include adolescents aged 11–16 years or so (supposed to be mastering formal operations), and to turn to mature adults for studying the higher RCR levels. In a wider sense, the current work benefited from the writings of MacKay (e.g., 1974) and Pattee (e.g., 1978). Both authors argue cases from their field of interest in which a satisfactory explanatory knowledge requires the simultaneous articulation of two, formally incompatible modes of description and they indicate some rules for doing so. While I am more concerned with forms of thought rather than theory construction per se, I acknowledge that they stimulated some of my own considerations, in particular with reference to the RCR heuristics (chapter 6). Summary of RCR development RCR development is putatively stimulated by interactions between nature and nurture, whether the interactions occur unconsciously, preconsciously, or consciously. Important milestones are reaching object-reflecting, and, even more importantly, means-reflecting thought. RCR development follows the ‘logic’ of intra-inter-trans. While not a direct continuation of earlier work, studying RCR benefited from such work in various ways.

3

Metaphysical Assumptions and Theory of RCR

RCR implies a certain ontology, that is, it makes assumptions about the nature of reality. RCR also involves epistemological assumptions, those having to do with the process of gaining knowledge in the cases concerned. For these reasons, I begin with assumptions adopted here from the philosophy of knowledge. I continue with a philosophical analysis of RCR as thought form. Finally, there is a discussion of, and an attempt at justifying, the underlying logic. Assumptions adopted from the philosophy of knowledge Why a discussion of the philosophical foundations of the work described in this monograph? Throughout some kind of reality is assumed to exist ‘out there’. That is not undisputed, and needs clarification and justification. In the words of Hilary Putnam (1999, p. 4): ‘And no issue polarises the humanities – and, increasingly the arts as well – as much as realism, described as “logocentrism” by one side and as the “defence of the idea of objective knowledge” by the other.’ Putnam’s solution – and largely mine – is a ‘middle way between reactionary metaphysics and irresponsible relativism’ (ibid., p. 5). Options In view of the importance of the assumptive base for one’s research (Werner [1948] 1973, 1957; Reese and Overton 1970; Overton and Reese 1972; Case 1998, pp. 747–53; Putnam 1999; Fahrenberg and Cheetham 2000), I state my position (Reich 1995c, 2000b) after reviewing some options. Rather than recalling the history of the philosophy of knowledge (e.g., Overton 1998, pp. 127–63), let me simply list some possible assumptive bases, first concerning basic metaphysical orientations and philosophical presuppositions. RCR assumes that there exists a reality ‘out there’. To the question, ‘What can we know about reality, if it exists?’ one summary answer is 35

36

The Theory of Relational and Contextual Reasoning

provided by radical constructivism: ‘At best we can know what reality is not’ (Glasersfeld 1996). Another broad answer is given by the various shades of realism (e.g., Putnam 1988). Classical realism rests on the following three pillars: (1) there is a reality independent of human ideas and theories; (2) scientific theories and the theoretical entities contained in them purport to refer to those (real) entities, processes, or structures existing independently of the theories; (3) hence scientific theories can be judged to be true or false in some sense larger than ‘they allow one to describe, predict, and organise the experimental data’. Thus, the scientific theories assumed by classical realism involve ontic truth (Kitchener 1988, p. 17), not just the epistemic truth of theories ‘merely’ aimed at describing, predicting, and organising empirical data. ‘Foundationalism’ follows from the purported ontic truth of scientific theories. Laudan (1990, p. 134) enumerates the resulting (foundational) epistemological programme as ‘(1) a search for incorrigible givens from which the rest of knowledge could be derived; (2) a commitment to giving advice about how to improve knowledge; and (3) the identification of criteria for recognising when one had a bona fide claim.’ According to most contemporary philosophers of knowledge, foundationalism can no longer be justified (e.g., Laudan 1990). Indeed, by now it has become clear that (a) all observations are ‘theory-laden’ (influenced by pre-knowledge); (b) scientific theories are underdetermined by facts (several theories may explain ‘equally well’ a given data set); (c) ‘verification’/‘falsification’ of a theory is more complex than thought previously (the experimentum crucis is an exceptional occurrence); and (d) the (unwittingly chosen) underlying assumptive framework provides an influential hermeneutic context for one’s research (cf. Lakoff and Johnson 1999, pp. 74–81). Why, then, not adopt as an assumptive base radical constructivism sensu von Glasersfeld (a kind of instrumentalism) or even a social constructivism1 which holds that both science and the literary novel operate according to the arbitrary rules of a language game (unbridled by ‘what is out there’)? For one thing, radical constructivism, and even more social constructivism, fail to explain the success of science and technology in 1

‘Constructivism’ in the present context is clearly used in the sense of an epistemological orientation, not as a developmental concept. Choosing constructivism or another epistemological orientation (for example empiricism) still leaves one free to opt for constructivism as an ontogenetic developmental category (e.g., in a Piagetian sense) or not (cf. Philips 1995). As to the particular version of social constructivism, one has again to differentiate between the meaning in epistemology (under discussion here), in sociology (e.g., the ‘invention’ of marriage as an institution) and in social psychology (e.g., the co-construction of a world view by mother and child).

Metaphysical Assumptions and Theory of RCR

37

coping with, and predicting, natural phenomena. How could astronauts have reached the moon and come back if the results of scientific research were just as independent from any ‘objective reality’ as are certain fictional novels? Why do all Indo-European languages include some saying such as ‘knowing is seeing’ (= seeing something becomes eventually knowing it) if no confirming experiences underlie this saying? Furthermore, turning from the outer world to the inner world: why should the vocabularies of nearly all languages, even those spoken on remote islands and in ‘inaccessible’ mountain valleys, have terms for the basic colours (e.g., Hardin and Maffi 1997), unless those colours have the quality of ‘immanent objectivity’? Present epistemological choice On account of such arguments, instead of adopting a radical, let alone a social constructivism, I opt for a conjectural/hypothetical, sceptical and qualified, critical realism (cf. Putnam 1999, especially part 1).2 According to this view, we are engaging with realities that may be referred to and pointed at, but which are beyond the range of any completely literal description; these realities include thoughts, virtual quantum ‘particles’, and so on. To refer to them, we most appropriately employ metaphorical language3 and describe a given reality in terms of models, which models may eventually be combined into theories. To give an example of metaphorical language: the German poet Eduard Morike ¨ characterises spring as ‘flying its blue ribbon in the air while sweet familiar scents roam about full of foreboding . . . ’, and for the poet Rainer Maria Rilke, in autumn ‘the leaves tumble downward as if coming from far away, as if distant gardens in the skies were withering – they fall with a negating gesture . . . ’ (both my inadequate translation from German4 ). Clearly, no mature adult would expect to find blue ribbons on the streets or in the fields in spring, or to hear about gardens in the sky from astronauts travelling in autumn, yet both poets help 2 3

4

I have justified my choice more at length elsewhere (Reich 1995c; 2000b). In essence, I there present the options in a more fundamental way. Metaphors mainly work by making use of shared attributes of the base concept and the entity or event towards which it ‘points’, the target concept. Examples are, ‘he is a fox’ or ‘she is a whirlwind’ (cf. Goodenough 2000). A number of biblical parables (e.g., ‘The kingdom of heaven is like a mustard seed’, Matthew 13: 31–2; Mark 4: 30–2; Luke 13: 18–19) can be viewed as extended metaphors. By contrast, reasoning by analogy mainly maps functions, that is relational predicates, such as ‘The stem of a flower is like a straw for sucking up a liquid’ – cf. Gentner and Markman (1997). ‘Fruhling ¨ l¨aßt sein blaues Band wieder flattern durch die Lufte, ¨ suße ¨ wohlbekannte Dufte ¨ streifen ahnungsvoll durchs Land . . . ’, respectively ‘Die Bl¨atter fallen, fallen wie von weit weg, als welkten in den Himmeln ferne G¨arten; sie fallen mit verneinender Geb¨arde . . . ’

38

The Theory of Relational and Contextual Reasoning

their readers to get a sense of what spring and autumn feel like in our latitudes. And here is an example of metaphorical language in physics taken from a news brief of Scientific American on the 1999 Nobel prizes: The humming, beeping, well-lit modern world could not have been built without the knowledge that electric current is a parade of electrons and that those particles are not ricocheting billiard balls but fuzzy clouds of probability that obey odd rules of etiquette as they manoeuvre in a dance of mutual repulsion. (Nobel Prizes 1999, p. 16)

In this passage the terms ‘parade’, ‘ricocheting billiard balls’, ‘fuzzy clouds’, ‘rules of etiquette’, and ‘dance of mutual repulsion’ are all used not literally but metaphorically. Because electrons cannot be apprehended directly by the five senses, metaphors based on actual sensual experiences are used to convey a sense of what electrons are like to those for whom the equations of quantum electrodynamics (Feynman 1988) are not sufficiently telling. Using metaphorical language in effect extends knowledge from known (personal) experience (billiard balls, dance, etc.) to a lesser known case that is often not directly accessible to the senses (behaviour of electrons) – cf. Goodenough (2000). Franz Brentano and his successors broke with the idea of ‘uncertainty’ about coming to grips with the outside world. They posited instead that all contents of mental acts are to be taken as immanently objective, whether or not they have an external referent (cf. Baron-Cohen 1995; Vande Kemp 1996, pp. 166–7; Yates 1985).5 In other words, for the very large majority of persons, his or her ideas and representations usually spring from a sense of utter reality, regardless of what exists externally – theirs is a firstperson ontology.6 Thus, as mentioned above, no person will doubt that 5

6

I am aware that the ‘reality’ of the contents of mental acts, in particular of certain ‘sense data’ (qualia), is debatable (e.g., Putnam 1999, especially lectures 2 and 3, and the second afterword). However, the assumptions stated above seem adequate for present purposes. This, then, raises the question, ‘Whose truth?’ Here are some answers (drawn partly from an Internet discussion): (a) For the individual, observation by a single person (himself/herself) would be sufficient for regarding something as true. (b) For a community which shares a belief system, statements about unusual experiences, interpretations of texts, etc. from one or more trusted persons are enough to be accepted as true. (c) For the justice system, a proposition or fact is more believable if it is supported by several kinds of evidence or several unrelated instances of the same kind of evidence. ‘Opinions’ from single persons are admitted provided that person (i) is an expert, (ii) attempts to provide evidence on the state of mind of the defendant, or (iii) is dying (and therefore likely to speak the truth). (d) For the scientific community, it is in principle not the belief of persons, their mental/emotional state or their number, but the methodology by which the phenomenon is observed, recorded, reproduced, etc. which determines acceptance as ‘true’. What about getting to universal truths accepted by all, given such a state of affairs? In case of a brute fact such as ‘all humans are mortal’, no particular difficulty should

Metaphysical Assumptions and Theory of RCR

39

colours are attributes of the external world unless he or she has learned certain scientific facts about our visual apparatus (cf. Ramachandran and Blakeslee 1998, pp. 72–80, passim). Faced with the loss of foundationalism and the resulting weakening of the correspondence theory of truth,7 what can one say about the truthfulness of a given scientific theory? The answer is ‘little’. But from Laudan’s (1990, pp. 19, 59, 85, 103) discussion one gathers that under the assumptions adopted here (all observations are ‘theory-laden’; scientific theories are underdetermined by facts; ‘verification’/‘falsification’ of a theory is more complex than thought previously), it remains possible to compare rival approaches. The approach, model, or theory considered more effective would – each time compared to its rival – (a) explain broader ranges of different kinds of phenomena, (b) have been tested in more areas, (c) already have led to more unexpected discoveries or applications, (d) yield more precise results, (e) be more dependable, (f) possibly be the only candidate which offers a satisfactory explanation for certain phenomena. When making the comparison between the rivals, it is understood that no criterion (a) to (f) is individually sufficient for a ranking but that all criteria count jointly for a (defeasible) preference. In other words, the ‘victorious’ approach, model, or theory wins a relative victory, not an absolute one, and, in case the comparison is repeated after further work on a non-victorious competitor, it may well win. The basis and results of such comparisons can be agreed inter-individually, and thereby gain scientific credence.8 From the perspective of the critical realist approach we are discussing, the task of science is to come to some (tentative) conclusions concerning

7

8

arise. But that is no longer so when the issue becomes ‘communication with deceased persons’, ‘resurrection’, ‘reincarnation’, etc. In such cases the ‘insider’ and the ‘outsider’ interpretations may well stand as antitheses to each other. Then a major principle (incorporated into RCR) should be to respect that the criteria implied in (a) to (d) above cannot – and should not – be forced on the other parties to the discussion (cf. chapter 6, notes 2, p. 106; 5, p. 107; 6, p. 108; and 7, p. 108). If a helpful, progressive dialogue is to take place, criteria for a comparison of conflicting views and their inter-individually acceptable assessment have to be evolved before entering into the search for consensually acceptable ‘truths’. In this context one should also be aware of the following: a social life built exclusively on consensually accepted truths is hardly envisageable. There is always the question of how to deal with the unknown, even the unknowable. And that is where trust, treaties and conventions (e.g., the Mayflower Compact) and the like come in. The correspondence theory of truth is based on a one-to-one relationship between ‘what is really out there’ and statements referring thereto. For instance, ‘This is a woman’ is true if, and only if, what is being referred to is a woman (which can be proven incontrovertibly). Gerhard Schurz (1998) argues contra Thomas S. Kuhn – rightly in my view – that even in ‘non-revolutionary’ times of doing science not just one single paradigm monopolises research activities. Rather, also at that time competing theories exist – not only during scientific revolutions – and moreover are desirable for advancing scientific progress.

40

The Theory of Relational and Contextual Reasoning

‘order’ or ‘patterns’ with respect to the object of study, to explain them by elucidating the variables involved and demonstrating relationships between them, and finally to understand the underlying mechanisms in depth, whenever possible in terms of a coherent theory. Any such order or pattern and its understanding is neither simply discovered as objectively given, nor constructed purely socially, without any ‘objective’ constraints. Not infrequently, theories evolve by iterative bootstrapping analyses of ever more appropriate data gathered in the light of ever better hypotheses, possibly aided in this by improved reasoning using more appropriate tools of thought, and better empirical methods. Assumptive framework and methodology adopted My assumptive framework is close to that described by Overton (1998), that is to say an inclusive relational frame. In this recent handbook article, Overton (ibid., pp. 111–14) distinguishes between (a) ‘transformational change’ (change in form, pattern, or organisation, e.g., embryological changes, or the successive transformations from action to symbolic thought as a way of gaining insight) and (b) ‘variational change’ (degree or extent that a change varies from an assumed standard such as thinking, e.g., analytic thinking styles and synthetic styles). In a split-based approach, the analytic difference between transformational and variational changes are transformed into ‘true cuts of nature’ (ibid., p. 112) and either transformational or variational change are claimed to be real. In Overton’s relational, both-and approach, transformation and variation represent differing perspectives on the same object of inquiry – developmental change within a framework of inclusiveness. The existence of one type of change is not incompatible with the other type of change, even in the same person. The movement from babbling to language may be profitably understood as sequential and directional and, hence, irreversible (transformational change); however, when the infant has become a student, raising of the grade point average can be reversed some time later (variable change) – there is no contradiction (ibid., pp. 113–14). The point is that ‘casting our fundamental understanding of development into an inclusive relational frame has profound implications for the concepts and theories, as well as the methodology and methods, of developmental inquiry’ (ibid., p. 114). There will be ample opportunity later on to fill that statement with concrete examples beyond the nativism–empiricism debate already briefly referred to. An inclusive relational framework is not the only conceivable framework but one that is consonant with RCR. Summing up, I espouse a critical realist ontology and a nonfoundational epistemology involving a ‘transverse rationality’. The latter permits

Metaphysical Assumptions and Theory of RCR

41

one to build bridges also between disciplines considered incompatible, incommensurable, and the like by some protagonists (see chapters 6 to 11). As I also report on empirical work, here is a word on methodology to close this section. I accept the following methodological stipulations: first, convergence of evidence from as many sources as possible is to be striven for; second, an adequate theory must provide empirical generalisations over the widest possible range of phenomena (Lakoff and Johnson 1999, pp. 79–80). I also subscribe to the triadic network of justification: (1) research aims and theories should normally harmonise, (2) theories justify methods and are in turn justified by them, (3) methods exhibit the realisability of the research and are justified by the aims (Laudan 1984, pp. 62–6). Theory of RCR Philosophical analysis Following the philosopher Philibert Secretan (1987), an analysis of various thought forms leads one to the conclusion that RCR is ‘situated’, if one may say so, between dialectical and analogical thinking and shares some features with each. All three differ from the classical analytical (Piagetian) thinking, and therefore the relationships between those four types of thought need to be clarified. In doing so I draw on Secretan’s argumentation. Given RCR’s acceptance that two or more heterogeneous descriptions, explanations, models, theories, or interpretations of the very same functionally coherent entity or phenomenon are ‘logically’ possible and acceptable under certain conditions, that raises the tricky issue of the excluded middle (chapter 5). The question is whether the maxim of the excluded middle (A must be either A+ or A−, it cannot be anything else) is really universally valid, or whether it is an analytical interpretation of a principle, which may admit of differing interpretations. Affirming the exclusion of the middle is based on a mutual exclusion of being and nonbeing, which implies a classical logical negation (chapter 5). Is that really the only legitimate interpretation? Or is it merely a simplification? The world viewed from the standpoint of the excluded middle is a world of a formal static ontology ruled by an analysing, dissecting logic, thereby creating an impression of order, clarity, and operability. The world viewed by RCR is not the world of the excluded middle. It is incompatible with a strictly analytical view based on that maxim. Given that the analytical view has a long tradition and is well established, where can one look for help to justify and legitimise the RCR view? Can dialectical thinking be of assistance?

42

The Theory of Relational and Contextual Reasoning

Dialectical thought acknowledges and deals with ‘contradictory’ elements within and between the states of affairs it tackles. It aims – particularly in Hegel’s approach – to overcome any ‘contradictions’ (understood in a more general sense than in formal binary logic) through a process of becoming. It is important to understand the meaning of ‘negation’ in dialectic thought. It is not identical with the meaning of that term in formal binary logic, where a negation of a negation refers to the operation of returning exactly to the initial situation/conditions. In dialectical thought the negation of a negation implies arriving at a new state. That new state may still involve an internal ‘contradiction’, which is not objectionable in itself – but, importantly, is not the case in the RCR world. Thus the detour via dialectical thought has been helpful in that it demonstrates the nonuniversality of classical analytical thinking, but it has not resolved the question of how to justify and legitimise RCR vis-`a-vis the maxim of the excluded middle. To that effect we need to introduce analogical thinking (chapter 5), which accords a central value to analogies. Looking back at its long history, one may say that analogical thinking consists in making a comparison that connects nonequal entities in such a way that either (a) a particular numerical ratio or a common function pertaining to both entities constitute the analogy, or (b) a discrepancy within the common feature(s) of both entities is at the core of the analogy. An example for the latter would be the analogia entis (analogy of being) of the Middle Ages: God’s being is quite different from the being of nature and of humans in particular, yet God’s being shares features with the latter’s being (Fourth Lateran Council, 1215). As these and other examples (chapter 5) show, analogical thought is based on equivocality over against the univocality of analytical thought. Analogical thought de-emphasises full comparisons per se (and a fortiori refuses an elimination of the ‘loser’) while maintaining a beneficial tension between commonalities embedded in differences. In other words, analogical thought is primarily concerned with striking commonalities despite differences between the two entities considered (but not particularly with their ontological status). Thus, analogical thinking is RCR’s ally against analyticity’s univocality (itself a consequence of the excluded middle). How closely are analogical thought and RCR related? To answer that question by way of an example, let us discuss the experience of ‘light’ (cf. Fagg 1999). Using analogical thinking, the experience of physical light (the sun) can be connected with gaining an insight (ex oriente lux – the illuminating message came from the east), with experiencing a divine presence in the soul (the ‘inner light’), with describing the aura and agency of a prominent person (one of the ‘leading lights’). These analogies can be

Metaphysical Assumptions and Theory of RCR

43

studied, for instance, by physicists, phenomenologists, and psychologists. Could RCR justifiably be applied to their findings, the explanandum being light ? RCR is not merely concerned with the rivalling descriptions, explanations, models, theories, points of view, etc., but also with the status of the explanandum itself. In fact, RCR keeps an ‘existential’ distance from the explanandum (and thus shares a feature with analytical thought in regard to the explanandum). In the present case, applying RCR would lead one to the conclusion that the explanandum of interest is not light, but human experiencing. RCR recognises that human beings (a) experience the world around them and inside them, (b) constitute singular individuals living their particular lives in interaction with their bio-physical, social, cultural, and perceived spiritual surroundings, to which life they endeavour to give a meaning, and (c) are objects of scientific research. For RCR, the simultaneous validity of (a), (b), and (c) is not just an expression of synchronicity, but an expression of the wholeness of the human person. Inside this whole, elements of analogical thinking have their place, but to seize up the overarching unity, RCR is called for, not analogical thought per se. Thus, while sharing some features, analogical thought and RCR are different. To sum up what has been said so far: RCR shares certain features with (a) Piagetian thinking (analyticity), (b) dialectical thought (refusing a unique, separable status quo that is constant in time, and claiming the existence of a ‘link’ between ‘rivalling’ entities), and (c) analogical thinking (commonalities imbedded in differences can be held in a fruitful tension, they are somehow ‘linked’). As a result of these shared features, RCR is cognitively complex. Furthermore, it involves a logic of its own. Ontology, epistemology and methodology of RCR Applying RCR in justifiable cases is based on a certain ontology, epistemology, logic, and methodology (Reich 1994b; cf. Fahrenberg 1992; Fahrenberg and Cheetham 2000). Let us again denote by A one description, explanation, model, theory, or interpretation in regard to the explanandum, and by B another, categorically different description, explanation, model, theory, or interpretation concerning the same explanandum. Ontologically, a ‘meta-relation’ is posited between the class of intensions (= contents, meanings) pertaining to A and the class of intensions pertaining to B. An illustration would be ‘ultra posse nemo tenetur’ (nobody should be held morally responsible for failing to perform an act which is beyond his or her capability). That ‘norm’ connects the moral

44

The Theory of Relational and Contextual Reasoning

Figure 3.1 ‘Figure–ground’ shift of the number of cubes (after rotation of the figure by 180◦ ) as an illustration for noncompatibility.

demands pertaining to a certain act (the ought) with the capability of the actor to meet those aims (the is). Given the hypothetical, critical realism adopted, and previous research on the issue, RCR recognises that both moral norms and human capabilities have a certain existence of their own, but are not independent of each other. The link is considered ‘objectively true’, independent of our individual degree of relevant knowledge. The epistemology calls for ascertaining that the intensions of A and B are co-extensional, that is, they refer to the same explanandum. This, then, is a partial procedure for exploring the explanandum’s ontological status in a new case. There are cases when probing the veridicality of that contention is far from easy. An illustration would be the analysis of psychophysiological phenomena (Fahrenberg 1992; Fahrenberg and Cheetham 2000), for instance of the results of (a) a person’s introspection of his or her fright, (b) the observation of that person’s synchronic behaviour, and (c) the simultaneous measurement of his or her physiological indicators (pulse rate, skin resistance, etc.). A related questions is: do (a) to (c) really refer to exactly the identical phenomenon/explanandum, are their extensions equivalent? If the answer is yes, then the research hypothesis is that (a), (b), (c) are intrinsically linked (entangled). A second epistemological RCR requirement is satisfied in the present case: the three intensions (a) to (c) belong to different categories. That latter condition is also met in the cases nature versus nurture, or moral norm versus individual capability, but is not in the case of the cubes of Fig. 3.1. The underlying logic will be dealt with more extensively in the next subsection. As already indicated, it is neither a formal binary nor a classical dialectic logic.

Metaphysical Assumptions and Theory of RCR

[

(x) (∃C') (∃C'') (t) ¬(C' = C'') .

45

{(x ε nc) ⊃ [{Obs (x, C', t)

⊃[F'(x, t) . ¬F''(x, t)]}

. {Obs (x, C'', t) ⊃ [F''(x, t) . ¬F'(x, t)]}]}] a

Figure 3.2 Noncompatibility according to Bedau and Oppenheim (1961, pp. 213–14). The statement in front of a bracket always refers to the entire content between those particular brackets. Explanation of symbols: x = explanandum, which has the noncompatible features F′ and F′′ ; ∃ = there exists; C′ , C′′ = context-related conditions; t = time of observation; ¬ = not; · = and (conjunction); ε nc = belongs to the validity domain of noncompatibility; ⊃ = implies; Obs (x, C′ , t) = observation of x under the condition C′ at time t, bringing out F′ ; F′′ = result of observation under condition C′′ .

The methodology has to meet the categorical specifics of A, B, (C . . . ). For instance, when explaining observations on fright belonging in category (a) above (introspection), one cannot suddenly argue in terms of (b) (behaviour) or (c) (physiology). First, coherent explanations within (a), (b), and (c) have to be found, and only then should possible links to the other categories be looked into systematically. These issues of ontology, epistemology, and methodology are less concerned with the basic characteristics of RCR itself, and more with the RCR heuristics, the potential application of RCR to research issues. That issue is dealt with in chapter 6 and it is hoped that any related question lingering on should be answered there at the latest.

The logic imbedded in RCR I posit that the logic imbedded in RCR is basically that analysed by Paul Bedau and Hugo Oppenheim (1961) regarding quantum mechanical phenomena and their study. The characteristic quantum mechanical features are (a) dependence of the experimental results on the details of the experimental set-up (the context), and (b) the nonseparability of certain variables. Bedau and Oppenheim base their analysis largely on the ‘truth value’ ‘noncompatible’ already indicated in chapter 1: two or more statements about the same explanandum are noncompatible, if they do not (fully) apply concurrently but are fully valid in different situations/contexts.

46

The Theory of Relational and Contextual Reasoning

The meaning of noncompatible is illustrated in Fig. 3.1. The number of ‘complete’ cubes is different depending on whether one looks at the figure upright as it appears here in the text or upside-down. (Most persons see six and seven cubes respectively.) To re-emphasise: this figure–ground shift is used to illustrate the meaning of noncompatible, not to claim that the content of Fig. 3.1 is an ideal study object for RCR. (It is not because only one category – the number of cubes – is under discussion.)9 Applying formal binary logic to Fig. 3.1 would lead to the stipulation that either 6 or 7 is the true value, the other must be wrong. However, from the perspective of noncompatibilty, either answer is correct in its context. The latter view, resulting from applying RCR logic, may be expressed by a predicate logical statement (Fig. 3.2): in one context one observational result obtains but not the other, and vice versa in a different context. To illustrate the working of the logical relationships shown in Fig. 3.2: taking the case of Fig. 3.1, C′ would be ‘upright’, C′′ ‘upside down’, F′ would then be ‘six’, and F′′ would be ‘seven’. Summary of the metaphysical and theoretical grounding of RCR A critical/hypothetical realist stance is adopted, and an inclusive relational framework. By way of analysing dialectical and analogical thinking, it is shown that the univocality of analyticity (e.g., Piagetian logico-mathematical thinking) is not the only way to deal with the excluded middle (through exclusion of one of the competing propositions). That has consequences for the ontology, epistemology, and methodology of RCR. Where RCR is applicable, a more inclusive approach is adopted than with Piagetian thinking. It is based on a trivalent logic which takes into account differing contexts. 9

A genuine study object for RCR would be the one-slit and the two-slit experiments with light ‘rays’, demonstrating particle-like and wave-like behaviour of light, respectively. However, that example is too far removed from cognitive development per se to be immediately helpful.

4

Empirical Studies of RCR

Overview The empirical research proceeded as follows: first pilot study in 1985 (basic nature of RCR and RCR developmental levels), second pilot study in 1988 (new standard interview problem), third pilot study in 1988/1989 (RCR and Piagetian concrete and formal operations), fourth pilot study in 1992/1993 (RCR, Piagetian formal operations, cognitively complex thought, and use/understanding of more than one logic).1 Respondents changed from study to study. All studies are methodologically flawed in several ways, but they do point to the value of considering RCR as a distinct form of reasoning, and demonstrate its developmental levels. A first flaw is that almost all respondents were non-representative. The children and adolescents were all pupils from ‘higher-level’ primary and secondary schools (that is not from ‘lower-level’ schools, let alone school drop-outs), and the adults mostly had university degrees. The adults had been ‘hand-picked’ throughout. Criteria were a ‘scientific’ thinking style as opposed to a dogmatic or an ad hoc style, and the capacity to express one’s thought clearly and fairly rapidly. The reason for wishing to work with specifically chosen (unrepresentative) respondents is as follows. It was suspected that RCR would evolve with cognitive development. In any such development, the tricky, and particularly interesting developmental levels/stages are the higher ones, because the lower ones ‘aim’ at a developmental ‘end point’, which appears the more clearly, the higher the level/stage. Unfortunately, and understandably, the higher the level, the smaller the number of persons found at that level. Hence, it would take quite high numbers of respondents in a representative sample if one were to obtain the same number of high level/stage answers as obtained in the research discussed in this monograph, in particular any relation of RCR with specific age groups. 1

The application studies (e.g., intervention in secondary school classes – not reported here in detail) took place 1994–8.

47

48

The Theory of Relational and Contextual Reasoning

Whereas selecting ‘high-level’ participants has the practical advantage of reducing the workload, the disadvantage is obviously that no meaningful statement can be made about a ‘typical’ age distribution of levels/stages, nor about any dependence on education, economic status, and the like. Therefore, such issues – while obviously interesting in themselves – are not discussed in this monograph, in particular not any correlations of RCR levels with specific age groups. A second flaw concerns the number of participants, which in some cases was only about thirty. That still allows one to do some statistics, but is at the lower limit of the desirable. The third flaw has to do with dialectical and analogical thought. Although referred to in the theoretical analysis as being somehow ‘close’ to RCR and sharing components with RCR, these thought forms were not included in any empirical study. The reasons were that (a) their relation to RCR seemed fairly clear from theory, (b) there was no doubt that they represent a form of thinking different from RCR, and (c) the studies actually made were judged to have a higher claim to the limited resources. As a fourth flaw, no systematic longitudinal study was carried out; only a few respondents were followed up. In those cases RCR developed indeed, but these results are too anecdotal to be reported here. An ideal design would aim at a detailed comparison of all the elementary, conjunctive, and composite operations of all the thought forms involved. However, that would be a Herculean task. Only Basseches (1980, 1984, 1989) has tentatively proposed such a breakdown (for dialectical thinking). A comparable amount of work would first be required for the other thought forms. For that reason a study of the situation at level 4 of Table 1.1 (p. 18), the ‘overlap’ of the thought forms themselves was adopted as basic design throughout – not a study of the various operations of levels 1–3 of Table 1.1. If it turns out (as it will) that mastering a particular other form of thought is a necessary condition for mastering RCR, then certain elementary, conjunctive, and/or composite operations are necessarily shared between these thought forms. All results referred to here have been ‘published’ in one way or another (as internal institute reports, in scientific journals, as a book chapter). The choice then is (a) to assume that the reader of the present monograph is knowledgeable about all these publications or at least ready to consult them, or (b) that such is not the case, given the difficulty of access to some of them (and maybe a problem of language exists). I opted for (b), and hence will present all results, basically in an archival type publication approach, at least as far the data proper are concerned. However, whereas all data are presented unchanged, their discussion is occasionally updated when helpful for getting into view what may not have been so obvious, or

Empirical Studies of RCR

49

simply unavailable at the time. Such updating will be indicated. Occasionally, the reader is referred to the original publication for further details. Methodological commonalities The basic methodology was to infer the characteristics of RCR (and of the other thought forms studied) as well as the developmental level from the observed use in the interviewee’s dealing with a particular kind of problem. The emphasis is on use, not primarily on the resulting product. To be clear, the RCR interviews do not consist in merely recording answers to a set of preformulated questions, nor to respond to self-selected items from such a set (which responses could then be treated with standard statistical methods), but to deal reflexively with a story. Hence, the scoring in essence is not based on quantifiable behaviour, nor on right or wrong answers, but on an ensemble of statements by the interviewee; it is a hermeneutic enterprise. The detailed procedure of the individual interviews is described in Appendix 1 (pp. 191–3), and the scoring procedure for RCR in Appendix 2 (pp. 194–8). Throughout all studies, scoring was based on actual, demonstrated competence, not on (meta-) statements about that competence (we know and can do more than we can say/explain). Thus, where applicable, tacit rather than explicitly justified thinking was rated. All problems used in the various interviews were formulated ad hoc, except the Piagetian tasks. The main criterion was to conceive and present a given problem in such a way that only one thought form would provide the best solution (cf. ch. 5). As will be understandable from the foregoing considerations, problems to do with nature–nurture, mind–body, and the like were good candidates for elucidating the level of interviewees’ RCR argumentation. The continuity of the four pilot studies was notably ensured by using throughout two identical RCR problems termed pianist (nature–nurture) and model of humans (mind–body). These problems were administered from age six years onward. For children up to ten years or so – but not for older interviewees in studies 2, 3, 4 – the rivalling explanations/‘theories’, were spelt out orally, and briefly summarised in large letters on a half a page, displayed in front of the respondent. The scoring procedure as such was independent of the age of the interviewee. All interviews were audiotaped and transcribed. In pilot studies 2, 3, and 4 the problem Accident in a nuclear power plant (already presented in chapter 1, pp. 20–2) was used in addition. It was introduced because that issue was less ‘part of the culture’ than either the pianist or the mind–body problem. The hope therefore was that it would bring out RCR even more convincingly. To avoid repetition later

50

The Theory of Relational and Contextual Reasoning

on, these three standard problems are now reproduced (in the version for adolescents and adults): Pianist (nature–nurture). The young pianist is fully immersed in her playing: her fingers speak to the chords via the keys, the movement of her body follows the music’s rhythm, and her mimic gestures express her intense inner participation. After she has played the last note, the audience applauds enthusiastically. The pianist is satisfied with her performance, but wonders whether it is more due to her practising or to her natural endowment. What is your view? Model of humans (mind–body problem). Since Antiquity it is recognised that human beings have a body, a mind, and a soul/a spirit. What is at issue in philosophy, and lately also in various sciences, is the nature of the relationship between the body and the other parts. How do you see that? Accident in a nuclear power plant. A TV news station reports on an accident in a nuclear power station. The main cooling pump had stopped working, and the back-up pump did not function. The emergency shutdown did not function either. To add to the difficulties, the operating crew became aware of the danger rather late and then underestimated it. The water temperature suddenly rose. A steam pipe cracked and leaked radioactive steam. What or who is to blame? What should be done to avoid another such accident in the future?

The task of the interviewer was first to present the problem at hand (orally to younger children, in written form to older participants) and then to find out about the interviewee’s capacity to ‘co-ordinate’ the two explanations/‘theories’ (cf. Appendix 1, pp. 191–3). This was done using questions such as ‘Who is right (the protagonist of “theory” A or that of “theory” B)?’; ‘Why?’; ‘What is the relationship between A and B?’; ‘How do you know?’; ‘How sure are you?’ These questions (and further requests for clarification) aimed at unearthing the structure of the respondent’s thinking, the co-ordinating and argumentative processes, over against his or her sheer knowledge about the particular issue, let alone his or her linguistic competence. Pilot study 1: RCR level descriptions and RCR effectiveness as pragmatic reasoning schema Based on the principles and considerations presented in chapters 1 to 3, this study aimed at testing the hypothesised basic features of RCR, the posited development, and the scope of RCR applicability. Method In this first pilot study (Oser and Reich 1987), we presented nine problems, with two explanations each, individually to twenty-four nonrepresentative respondents aged 6–25 years (12 m., 12 f.) and to four senior

Empirical Studies of RCR

51

physicists (mean age 54 years), and interviewed them about their solution of these problems in the way already explained. The interview lasted about three-quarters of an hour (age 6–25) or longer (physicists). The interview problems of this pilot study came in three groups: I. Social science, architecture, and physics: (1) Performance of a concert pianist, A: practising explains it all, B: native endowment determines everything; (2) Evolution from Romanesque to Gothic church architecture, A: economic reasons were decisive (less building material), B: a change of religious feeling provided the motivation (pictures of the interior of a Romanesque and a Gothic church were shown); (3) Weather forecast, A: it is sufficiently well computable to justify making extra provisions for snow-clearing ahead of time, B: it is subject to chance and does not provide a basis for (costly) arrangements. II. Matter vs. human spirit (mind, soul). (4) Obesity of a student, A: physiological cause, B: psychological reasons; (5) Healing kidney pain, A: a surgeon carries out surgery, B: a healer prescribes a herb tea; (6) Model of humans, A: humans are cell agglomerations, comparable to a clockwork, B: the mind/soul is in command (free will). III. Matter vs. ultimate being. (7) Crash of a glider, A: natural causes, (B) fate (horoscope); (8) Miraculous braking of a car on ice, A: chance, B: God’s hand; (9) Origin of the universe, A: Big Bang, B: God’s creation. Two of the events were actual contemporary local occurrences (snow clearing, crash of a glider). Results All in all 216 statements plus 36 from the physicists were obtained. These statements were used to establish two results: (a) a description of RCR levels (see discussion section below for the choice of ‘level’), and (b) RCR scores for each problem (individually, and by age group – neither presented here – and averaged across all ages); this by means of a provisional scoring manual based on (a). The descriptions of RCR levels resulting from the theory-guided content analysis of the 216 + 36 transcribed interview responses are shown in Table 4.1. The formulation is the latest, taking into account the insight gained since 1987. However, the description in Oser and Reich (1987, p. 182) is already quite close to that of Table 4.2. Also shown are the intrainter-trans levels according to Piaget and Garcia (1983/1989), introduced in chapter 2 (pp. 32–3). As to the rating of individuals’ answers according to a provisional version of the scoring manual of Appendix 2, sublevels were used for greater precision. Whenever the development had gone beyond a given level, the

52

The Theory of Relational and Contextual Reasoning

Table 4.1 Description of RCR levels. The characterisation of the developmental logic in terms of intra-inter-trans is taken from Piaget and Garcia (1983/1989). (Early level descriptions in Oser and Reich 1987, p. 182 – see note 3, p. 13 for change of name to RCR.) Level

Description

I intra

A and B (and C . . . ) are considered separately; only one of them is declared correct. The (implicitly) reigning concept is that A and B (and C . . . ) are alternatives, not complementarist aspects. Usually single-track choice of A or B, (or C . . . ), occasionally tentatively both (without offering a detailed justification), depending on chance knowledge or socialisation.

II

The possibility that A and B (and C . . . ) might both (all) be right is considered. A may be right, B may be right, (C . . . may be right), both (all) may be right, possibly with rather different weighting factors.

inter

III trans-intra

The necessity of explaining the given phenomenon with the help of A as well as by means of B (and C . . . ) is affirmed globally. After examination, neither A nor B (nor C) is considered quite correct as individual explanations, both (all) are needed for a full explanation. The limits of formal binary logic begin to be overstepped (intuitively).

IV trans-inter

Conscious connecting of A and B (and C . . . ), explicit evocation of their relationship. Affirmation that neither A alone explains the explanandum of itself nor B alone (nor C . . . alone). The relationship between A and B is analysed (e.g., ‘B permits making use of A’, ‘B cannot exist without A’, etc.). Any context dependency of the explanatory weight of A, B, (C . . . ) is (dimly) perceived. The use of RCR logic is more frequent. Although the argumentation may have some arguments in common with those of level II and/or III, it is markedly more complex.

V trans-trans

Encompassing ‘theory’, or at least synopsis, featuring (reconstructed) parts of A, B, (C . . . ) possibly supplemented by D, . . . , the various relations and context dependencies being fully understood from a multi-perspective viewpoint. Use of RCR logic has become a routine.

next higher level was added in parenthesis, e.g., II (III). Approaching, but not yet quite reaching the next level is indicated similarly, e.g. III (II). Thus there are ten levels and sublevels from level I to level IV, including both these full levels. The statistics shown in Table 4.2 were done using the (sub-)levels, in particular for computing the mean scores for all participants regarding a given problem. For instance, a mean score of 6 indicates that the average RCR developmental level of the participants concerned is III (II).

Empirical Studies of RCR

53

Table 4.2 Mean scores of RCR (sub-)level, variance, and correlations between problem scores and total scores for the 9 problems and the 24 respondents aged 6–25 years.2 (1 = level I, 10 = level IV, see text – source: Oser and Reich 1987, p. 184.)

Problem 1 Pianist 2 Architect 3 Weather 4 Obesity 5 Healing 6 Humans 7 Plane crash 8 Braking 9 Origin

Mean scores

Variance

5.83 14.67 5.71 8.47 6.13 10.63 5.96 12.74 6.38 8.41 5.96 11.49 4.83 10.67 4.67 9.10 4.58 10.95 Internal consistency alpha (Cronbach) = .97

Corrected correlations .95 .81 .92 .89 .87 .93 .83 .82 .88

The interrater agreement was 77 per cent, if a difference of one-third of a level between their scoring results was ignored. Such a difference is considered an acceptable error limit of scoring single responses. Mean values plus some others are reproduced in Table 4.2. As regards the statistics of the (acceptable) internal within-subject consistency across the various problems, and the within-problem variation of the scores across each age group, the reader is referred to the original publication (Oser and Reich 1987, pp. 183–4). Consistency is also indicated by the value of Cronbach’s α (.97). No gender effects were observed. Therefore, in the subsequent studies, a gender imbalance was accepted. Discussion To emphasise the ‘developmental logic’ of RCR, the core differences of the level descriptions of Table 4.1 are abstracted in Table 4.3, with the intra-inter-trans ‘development logic’ added again. Thus, the expectations from the foregoing theoretical considerations were met, namely that (a) RCR development proceeds in an orderly, systematic fashion; (b) consistent with other cognitive developments, RCR becomes ever more differentiated and integrated; and (c) RCR appears 2

Given the ordinal scales, the results of the statistical treatment are used only for comparison purposes, not as absolute values.

54

The Theory of Relational and Contextual Reasoning

Table 4.3 Developmental logic of relational and contextual reasoning. (Cf. Table 4.1, p. 52, for level descriptions.)

Level of RCR

Core characteristic of level

Stage according to Piaget & Garcia (1983/89)

I II III IV V

A or B (or C) A, but also B (C) A and B (and C) Logic of and (context) Synopsis/theory

intra inter trans-intra trans-inter trans-trans

in particular to be consonant with the intra-inter-trans sequence of Piaget and Garcia (1983/1989). From the state of affairs presented in Table 4.1 (p. 52), the question arose whether to designate the differences in RCR development as ‘levels’ or as ‘stages’. Stages in a narrow sense (hard stages), as defined by Piaget, Kohlberg, and others, refer to organised systems of action (e.g., logicomathematical operations), are qualitatively different from each other, and follow each other in an unchanging order with a clear developmental logic. If that were the only criterion, the differences in RCR development could arguably be declared stages. However, the term (hard) stages also implies the existence of a structured whole, a monolithic view of specific cognitive processes. In view of that particular (controversial) point, and the as yet unclear inner structure of RCR, the designation ‘level’ was retained for the five ‘milestones’ in RCR development. Now that the internal composite nature of RCR has become clearer, that choice is reaffirmed. On examining Table 4.2 (p. 53), one is immediately struck by the relative uniformity of the mean scores as well as the high values of the correlations between the problem scores and the total scores, not forgetting the high value of coefficient alpha. This was hoped for, yet the high numerical values were somewhat unexpected, given the vast differences between the nine problems, concerning in particular (a) the knowledge domain, and (b) the kind of explanation/‘theory’ involved (causal attribution, motivation for human action, prediction, presentation of a process, analysis of a structure, interpretation of meaning). The conclusions drawn first tentatively and then more and more confirmed in the course of the studies 2, 3, and 4 include the following: (1) to assess a person’s RCR level, presenting, say, three problems is enough.

Empirical Studies of RCR

55

Problems 1 (pianist) and 6 (humans) of Table 4.2 seem particularly suited (good mean scores and highest correlation values; they can be administered in slightly simplified form to grade-schoolchildren). (2) The mean scores and correlations of problems 7–9 are markedly lower. As studied in more detail later (chapter 7), one reason is a competence-performance discrepancy: for a variety of reasons the respondents do not produce their best. Nevertheless, based on the data of Table 4.2 (p. 53), RCR can be considered as a pragmatic reasoning schema (Cheng and Holyoak, 1985); this was hypothesised in chapter 1 (p. 14). Summing up the results of pilot study 1, it can be said that they support the basic features of RCR as explicated in chapters 1, 2, and 3, and clarify the number of developmental levels consistent with Piaget’s intra-intertrans sequence. A main open question remained: what is the relationship between RCR and Piagetian logico-mathematical thought? Pilot study 2: additional RCR interview problem For the subsequent studies, it seemed desirable to construct a supplementary RCR problem, preferably one that was not as ‘solved’ by the current culture as the problems pianist and humans. Therefore I chose a more controversial subject, an accident in a nuclear power plant, already presented in chapter 1 (pp. 20–2), being a scaled-down version of the Three Mile Island or the Chernobyl type accident. An additional aim was to look for such elements in the answers as would point to other thought forms. Administration The suitability of this problem was explored in a small test run. Not unexpectedly, the youngest age of respondents able to come up with scorable answers was not six years as for the other two standard RCR problems ( pianist and humans), but usually a little older, namely about nine years. Apart from that difference, as a rule the interviewing proceeded as in the first pilot study. The results of using the new problem in pilot studies 3 and 4 are presented below. Typical results and analysis This section is devoted to the traces of other forms of thought found in the responses to the nuclear accident problem. Technical malfunctioning

56

The Theory of Relational and Contextual Reasoning

is labelled A, human failure B. Five excerpts, chosen among other comparable ones, are ordered according to the developmental level of the respondents (I to V). Applying all that was said earlier, each excerpt will be analysed in turn. Here are prototypical answers I–V and their discussion (Reich 1995a, reproduced in parts with the kind permission of the Institute for the Advancement of the Philosophy of Children, Montclair State University): I. That technology was not reliable [A]. The operating crew is not to blame [not B] – they have done their duty, day in, day out (respondent aged 11 years).

The author of this first quotation (RCR level I) argues clearly at the Piagetian intra level. There is no indication of Piagetian concrete, let alone formal operations, nor of cognitively complex thought (level 1 – see p. 84 for levels 1–7), or real comparisons (no dialectical or analogical thinking), just the exclusion of B using an idiosyncratic argument. All this indicates, by the way, that the respondent performs somewhat below the average of his/her age level. Granting the benefit of doubt, one may discern a beginning of Basseches’s (1980, p. 408; 1984, p. 74; 1989, pp. 162–3) schema no. 9 (‘locating decisive elements within a whole’).3 II. It is true that a technical breakdown has occurred for starters [A]. But it appears that the operating crew has not been up to it either. That made it much worse [probably also B] (bright 8-year-old).

In this excerpt, we witness an enlargement of the observational horizon (RCR level II): in addition to the technical failure (A), the possibility of a mismanagement of the accident situation comes into view (B); the Piagetian inter level raises its head. There are indications of Piagetian concrete operations (working with two variables) and of differentiation (level 2 of cognitively complex thought). In a way, analogical thinking comes into the picture in that the common feature of (A) and (B) is identified as ‘involved in the accident/damage to the plant’ but then the difference is also recognised. In fact, there is an onset of a comparison of the respective weights in the damage assessment. Dialectical schema no. 9 is better developed, and schema no. 10 begins to appear (‘description of a whole in functional terms’). III. In the beginning a malfunctioning occurred. But then, such systems can’t work without human control. The operating crew has simply not noticed that the 3

Of the seven schemata to be referred to in the present analysis, Irwin and Sheese (1989, pp. 122–4) class nos. 10, 11, and 24 as used increasingly with better education/age, nos. 9, 12, and 14 as having a questionable pattern of emergence, and no. 15 as not appearing to have a developmental pattern. (However, their interview results indicate that dialectical thinking changed – became more powerful – across their educational/age groups; ibid., p. 119.)

Empirical Studies of RCR

57

instruments indicated a problem. Or perhaps they have seen it but not taken the right countermeasures. The people were excited. The accident involves both a technical and a human deficiency [A and B] (14-year-old).

The third quotation is from a person who is clearly at a more advanced development stage (RCR level III). The argumentation is at the transintra level, the variables are dealt with at a confirmed level of concrete operations with an indication of the formulation of a tentative hypothesis (‘perhaps . . . ’). An element of integration (level 4 of cognitively complex thought) appears, ‘The people were excited [on account of the accident]’. Likewise, Basseches’s module no. 12 (‘assertion of relations’) comes into view. This indicates the transition toward Piagetian formal operations (dealing with the functioning of systems). One also may note that a logic different from that of formal binary logic is used: the events in the power station and the psychological state of the crew are not separable, and the latter is not reversible, at least not instantaneously. IV. Such accidents are very rare. So one can understand that it may have taken time before the crew realised that something was amiss. But then they may well have done the wrong thing. And the situation got worse. Maybe they quarrelled about the right action to take. Maybe they even panicked when the steam came out. And they never called in a specialist who might have been able to get the situation under control. Such crews need training with a simulator under as realistic conditions as possible [relations between A and B, context dependency] (25-year-old).

The author of the fourth quotation has progressed to the trans-inter level (RCR level IV). One is immediately struck by the use of elements of Piagetian formal operations: hypotheses are formed and argued about in an abstract way. The integration is stronger (level 5–6 of cognitively complex thought). There appear three new schemas, nos. 14 (‘two-way reciprocal relationship’), 15 (‘assertion of internal relations’), and in a rudimentary way no. 11 (‘assumption of contextual relativism’). The latter finds its expression in the statement ‘Such accidents are very rare. So one can understand that it may have taken some time before the crew realised that something was amiss.’ V. In this accident technical and human failure are interconnected. One has to look at the whole thing as a system, the plant and the operating crew. And one has to study the mutual interaction, the type of effects they have on each other. One really wants to train crew members with the help of a sophisticated simulator so that they become aware of the many ways in which something can go wrong, they experience their individual and collective reactions, and learn how to assess such situations as well as how to deal with them successfully. In such simulations the psychological stress must of course also be generated, not just the sequence of technical events. It is precisely such a chain reaction of technical and human

58

The Theory of Relational and Contextual Reasoning

Table 4.4 Mean scores of RCR (sub-)level, variance, and correlations between problem scores and total scores for 3 problems and 63 respondents aged 9–68 years (1 = level I, 13 = level V).

Problem 1 Pianist 2 Accident 3 Humans

Mean scores

Variance

9.11 5.23 8.46 5.96 9.14 4.46 Internal consistency alpha (Cronbach) = .93

Corrected correlations .90 .77 .89

malfunctioning which is so hard to foresee. By the way, I would hire only such persons who are aware of the dangers involved and are ready to face them [all RCR characteristics] (middle-aged adult).

As to answer V (already analysed in chapter 1, pp. 21–2 in a preliminary way), finally everything falls in place. The respondent argues at a transtrans level (RCR level V ). What are commonalities and differences with respect to Piagetian operations? As regards commonalities, one observes the systematic approach, the formulation of hypotheses, the search for effective ‘variables’, and the tendency to generalise. With respect to differences, the underlying logic is not the formal binary logic: one notes that for the respondent the sequence of events is not reversible, the different interventions of the crew and the respective state of the power plant do not commute (that is their sequential order makes a difference), and the psychological state of the crew members is considered not to be separable from the way the accident situation evolves. Obviously, there is plenty of differentiation and integration (level 7 of cognitively complex thought). Furthermore, relevant aspects are mentioned which were not indicated in the problem description as presented. This is consonant with Basseches’s schema no. 24 (‘multiplication of perspectives to get as inclusive a view as possible’). After inspecting next what became of the new problem, we turn to the study of the relation between RCR and Piagetian operations, and then to the inclusion of cognitively complex thought, and ‘logics’. Subsequent performance of the new problem This problem was used in pilot studies 3 and 4. The combined results are shown Table 4.4, using the format of Table 4.2 (p. 53) for comparison purposes. The mean values are higher (and the variance smaller) on account of the different sampling of participants. A mean score of 9 corresponds to RCR

Empirical Studies of RCR

59

level IV (III). As before, the mean scores and correlations for problems 1 and 3 are remarkably close; the values for the more unfamiliar problem no. 2 stay fully acceptable. Pilot study 3: RCR and Piagetian operations Method Participants were thirty-eight nonrepresentative children, adolescents, and young adults (10 m. and 28 f.) from the local area, aged 7–22 years. The procedure for ascertaining the RCR levels was essentially the same as before, that for the Piagetian task as stated in the literature. The total interview lasted about one hour.4 RCR problems: the problem presented first was always that of the pianist because even grade schoolers were somehow ‘familiar’ with playing an instrument, so that it made for a good start in the otherwise unfamiliar situation of the interview. The second problem was that of the power plant accident (administered to respondents from 9 years onwards), and the third the humans one. They were administered in between Piagetian tasks. For participants younger than 9 years the RCR level was determined using the results from the interview about the pianist and the humans problems. Piagetian tasks: these tasks were taken from the literature, but (slightly) modified for increasing the difficulty and thereby providing a better (sub)stage discrimination. Also, for convenience, the numbering of the stages was made uniform and changed compared to Piaget’s numbering, namely IIa, b for preoperational substages, IIIa, b for substages of concrete operations, and IVa, b for substages of formal operations, a indicating an initial, and b a confirmed substage. Snail task (Piaget et al. [1946] 1972, pp. 95–109 for principle; details were modified as described below), see Fig. 4.4. The four subtasks consist in adding (or subtracting) displacement distances, respectively displacement speeds, either by actually carrying out the displacements or effecting them purely mentally. A snail (symbolised by a snail house) moves on a board, on which seven ‘milestones’, each time separated by one unit, are recognisable. (The extreme milestones on either side coincide with the edges of the board.) The board rests on a table, and is displaced itself. Both movements are either parallel or antiparallel. A scale extending from 3 units left to 3 units right is fixed on the table in such a manner that the board displacement can be read off. The units on the board (milestone separations) and on the scale are of the same size. The snail always starts from the central milestone (central position) and moves up to three milestones either to 4

Additionally, problem 9 of pilot study 1 (see ch. 7, p. 126) was proposed on a voluntary basis to the participants of pilot study 3.

60

The Theory of Relational and Contextual Reasoning

Milestones

Snail

Board

Scale 3L

2L

1L

0

1R

2R

3R

Table

Figure 4.1 Schematic sketch of the snail task. L = left; R = right. (Shown as viewed by the participant; the table is larger than the frame of Fig. 4.4; explanation in text.)

the right or to the left. Initially, the board is located with the snail at position zero of the scale. The task is to indicate where the snail finds itself at the end of the displacements with reference to the scale on the table. The participant is free to move both the snail and the board, but a higher score (IV) is obtained for doing it all in the head. (All this and the given subtask is explained to the participant.) The four subtasks are: (a) Snail one unit to the right Board two units to the right (b) Snail two units to the left Board one unit to the right (c) Snail three units to the right Board to the left with a third of the snail’s velocity (both starting together) (d) Snail continuously to the left Board to the right with half the snail’s until it leaves the board velocity (both starting together). (last milestone)

The solutions are (a): 3R; (b): 1L; (c): 2R; (d): 1 12 L. And the scores: IIa: IIb: IIIa: IIIb: IVa:

Irrelevant, confused answers, possibly description of snail motion Description of one displacement or the other, but no combination Dimly aware of principle; perhaps one of (a)–(d) solved correctly More competent; two or more of (a)–(d) solved ‘manually’ Principle understood; three or more subtasks solved, mostly in the head IVb: All subtasks (a)–(d) rapidly solved in the head without mistakes ‘Principle understood’ means that the participant argues in terms of adding/subtracting directly the displacements in question (subtasks a and b) or transforms first the velocity differences into displacement differences (subtasks c and d) and then proceeds as before.

Empirical Studies of RCR

61

Plant task. This task tests the capacity to separate the variables and to determine the ‘operative’ ones. The original task (Kuhn and Brannock 1977) comprised four pictures, two of healthy plants (one with a full glass of water and white plant food; one with half a glass of water, white plant food, and leaf lotion), and two of unhealthy plants (one with a full glass of water, black plant food, and leaf lotion; one with half a glass of water and black plant food). The interviewer explains to the respondent that the water, plant food, and leaf lotion (if so depicted) are given each week to the plant in the quantities pictured. The task is to find out what it takes to grow healthy plants. This version is easy to solve in that only the white plant food is effectively operative. Deirdre Kramer (private communication) therefore extended the problem by making also the leaf lotion operative, and provided a total of eight corresponding pictures. I extended the plant task further by adding two pictures of unhealthy plants (with leaf lotion, either a full glass or half a glass of water, and no plant food) to Kramer’s eight. These two pictures were not displayed, but delivered on the request, ‘I do not know whether the white food is good for the plant, or the black food is bad; I need to see what happens without plant food.’

The correctness of the solution ‘white plant food and lotion’ as well as ‘between one half and one glass of water’ can be proven convincingly with the ten pictures; in particular the hypothetical possibility that the white plant food is ineffective and the black food a poison can demonstrably be excluded. The scoring was as follows (on account of the more difficult task, involving more criteria, stages III and IV were subdivided into three substages, a, b, c): IIa: IIb: IIIa: IIIb: IIIc: IVa:

Irrelevant, confused answers Separation of variables unknown (idiosyncratic explanations) Beginning of separation of variables, but no correct solution Water level excluded as operative, but no correct solution As IIIb, but in addition correct operative variables found As IIIc plus explicit demonstration of plant’s tolerance for water level IVb: As IVa, but more spontaneous search for finding and justifying necessary conditions (leaf lotion, plant food) IVc: As IVb plus request for information on state of plant without food. Pendulum task (Inhelder and Piaget [1955] 1958, pp. 67–79). The task is to find the operative variable determining the pendulum’s oscillation frequency. The participants were individually given four pendulums (two long and two short ones, each time with a light and a heavy weight), and a stopwatch. What determined the pendulum’s oscillation frequency, the length of the suspension, the weight, or the impetus of setting it in motion? Experimenting was supplemented by asking questions provided by Kuchemann ¨ (1977, 1979), which assess the capacity to test relevant hypotheses ‘in one’s head’, i.e. without actually experimenting with real pendulums.

62

The Theory of Relational and Contextual Reasoning

Table 4.5 Mean scores of RCR (sub-)level, variance, and correlations between problem scores and total scores for 3 problems and 30 respondents5 aged 9–22 years (1 = level I, 10 = level IV). (Source: Reich and Oser 1990, p. 53)

Problem 1 Pianist 2 Accident 3 Humans

Mean scores

Variance

7.37 2.79 6.90 3.68 7.63 1.90 Internal consistency alpha (Cronbach) = .89

Corrected correlations .85 .70 .79

The solution is, ‘only the length of the suspension determines the frequency. The longer the suspension, the lower the frequency (inversely proportional to the square root of the length).’ In this task the scoring was: IIa: IIb: IIIa: IIIb: IVa: IVb:

Irrelevant, confused answers Separation of variables unknown (idiosyncratic explanations) Beginning of separation of variables, but principle not grasped On the way to understanding the principle, partial separation Separation of variables found eventually but not spontaneously Separation of variables and exclusion of inoperant variables Results and discussion

RCR problems. For comparability, the mean scores for the participants 9 years old or older are presented in Table 4.5, using the format of Table 4.2 (p. 53). The values for the nuclear plant accident are less good than those for pianist and humans, but still acceptable. As a comparison with the relevant values of Table 4.2 shows, the mean values of Table 4.5 are higher (and variances smaller) on account of the higher starting age for the nuclear accident problem. A mean value of 7 corresponds to RCR level III. Piagetian tasks. The values of Spearman’s rank correlation coefficients of the various scores were: snail–plant: .77; snail–pendulum: .73; plant– pendulum: .68 (p = .01 or smaller in all cases). In case of discrepancies between the three individual scores, the highest score was used subsequently, as competence, not performance, was the issue. 5

As a rule, the nuclear accident problem was administered to participants 9 years old or older – here to thirty-one persons, but only thirty records were fully usable.

Empirical Studies of RCR

63

Table 4.6 Frequencies of individual scores concerning Piagetian operations and RCR levels. Level I (II) is somewhat above level I, level II (I) somewhat below level II, etc. N = 38, age 7–22 years. (Source: Reich and Oser 1990, p. 54; reproduced in Reich 1991, p. 85, and in Oser and Reich 1992, p. 90). RCR level Piagetian operations

I

I (II) II (I) II II (III) III (II) III III (IV) IV (III) IV IV (V)

Early concrete 2 Confirmed concrete 2 Transition Early formal Confirmed formal (Transition?)

1

1

2 1

1 2 8

6

2 3

3

3 1?

Piagetian operations and RCR. The frequencies of interest are shown in Table 4.6. No scores are found in the upper right-hand corner. Thus no interview participant argued at a low (sub-)stage of Piagetian operations and at a high RCR level. In contrast, a high (sub-)stage of Piagetian operations did not ensure an equally high RCR level. This state of affairs can be interpreted as follows: Reaching a given (sub-)stage of Piagetian operations is necessary but insufficient for reaching a given RCR level. (The insufficiency indicates a lack of performance in one or more of the other components involved in RCR.) The conclusion is that Piagetian concrete and formal operations and RCR are really distinct, yet structurally interconnected. In particular, the Piagetian stage actually reached limits the attainable RCR level, at least up to the confirmed formal stage. From there onwards, RCR seems to free itself from the ‘Piagetian’ embrace. This supports the foregoing considerations, especially those about Fig. 2.1 (p. 30), and about a differing RCR logic. Pilot study 4: RCR, Piagetian operations, cognitively complex thinking, and evolved logics Pilot study 4 was an extension of study 3 to include cognitively complex problems, and evolved logics. The procedure was essentially that of pilot study 3. As the participants were different, the Piagetian tasks were included again; however, the lengthy pendulum task was replaced by the balance scale task.

64

The Theory of Relational and Contextual Reasoning

Hypotheses Pilot studies 3 and 4 jointly tested the following four hypotheses: H. 1: RCR shares elementary/conjunctive/composite operations with Piagetian operations and with cognitively complex thinking. H. 2: The chain of RCR levels I–V and the sequence of levels 1–7 of cognitively complex thinking (p. 84) share commonalities regarding the respective measures of differentiation and integration. H. 3: The development of RCR involves transcending the limits set by formal binary logic, and hence by certain partial Piagetian operations (cf. Table 5.4, right-hand column – p. 89). H. 4: Higher levels of RCR involve a logic other than formal binary logic. These hypotheses can be tested as follows. If the interviews of a statistically significant number of participants demonstrated competence with RCR, but neither with Piagetian operations nor with cognitively complex thinking, then H. 1 would be falsified. If all RCR problems were treated at low RCR levels (II or III), but the problems assessing cognitively complex thinking at high levels (6 or 7 of that thinking), or vice versa, then H. 2 would be falsified. If the RCR problems were dealt with at lower levels (II or III) without using formal binary logic, then H. 3 would be falsified. If no other logic than formal binary logic showed up in respondents’ RCR argumentation at levels III and up, then H. 4 would be falsified.

Method The sample consisted of thirty-two nonrepresentative participants from the local region (17 m., 15 f.), aged 13–68 years. The interview (about the nine [unique solution] tasks / [ill-defined] problems) lasted about one hour. As far as the pilot study 4 per se is concerned,6 the nine tasks/problems were: the standard set of three RCR problems, a set of three Piagetian tasks, a set of two problems for assessing cognitively complex thinking, and one problem for assessing competence with logics. The problems were administered in one or the other of a mixed, partly counterbalanced order. Piagetian task. In addition to the snail task and the plant task, the following task was used: 6

Additionally, two religious doctrines and problem 9 of pilot study 1 (see chapter 7) were proposed to participants of pilot study 4.

Empirical Studies of RCR

65

Balance scale task (Inhelder and Piaget [1955] 1958, pp. 164–81). The interviewee was presented with a picture of a balance scale and the invitation to describe how the equilibrium condition can be worked out.

As RCR and Piagetian operations had been the object of pilot study 3 and those results (Table 4.6, p. 63) were to be analysed jointly with the results of study 4, it was only necessary to ascertain whether the present participants mastered Piagetian formal operations, and thence the indicated use of the balance scale task, its scoring being an additional check on the results of the two other Piagetian tasks used in study 4. Problems for assessing cognitively complex thinking. Despite a literature search, no suitable problems were found. In constructing two problems, a supplementary condition was that they could also serve in the task of assessing logical competence. Here are the two problems: Filling a post. Mr Boschung [all names are typical local names for added reallife feeling] heads a laboratory in a successful technical firm. He has just been given a new post, and wants to fill it rapidly despite a shortage in the labour market. He offers the post to Franz Riedo, a competent young collaborator of his colleague Zosso who heads another laboratory in the same company. Boschung has not informed Zosso about his offer. When Zosso learns about it, he asks their common boss, Mr Gotschmann ¨ to prevent Boschung from hiring Riedo. Interviewing. When interviewing a person, on the one hand one wants him or her to be as spontaneous and self-organising as possible in order to learn authentically a maximum about that person. On the other hand, the interview does serve a specific purpose, which requires a certain degree of directivity. What problems result from that situation in your view?

Clearly, the logical structure of these problems is quite different compared to that of the Piagetian tasks: no single correct answer exists; at best probabilistic scenarios can be depicted. The reason for such a state of affairs does not reside primarily in a lack of information (the respondent could and sometimes did ask for more), but in the intrinsic uncertainties involved. Regarding the filling of the post, the attitudes and opinions of the protagonists are not necessarily fixed once and for all, in particular not that of Franz Riedo. In respect to interviewing, the contrasting, even contradictory objectives can only lead to a more or less successful compromise. Also, the context conditions (possibly unfamiliar surroundings, noise, presence of third persons, nervousness of the interviewee on account of unfamiliarity with the interview theme or the particular interview method, etc.) usually have a marked influence on the result. For these two problems the scoring followed the manual of BakerBrown et al. (1992), with level 1 corresponding to no differentiation and

66

The Theory of Relational and Contextual Reasoning

Table 4.7 Frequencies of individual scores of levels of cognitively complex thought and RCR levels. Fractional levels/levels as before. N = 32. (Source: Reich 1995d, p. 12.) Level of cognitive complexity (Baker-Brown) 3(4) 4(3) 4 4(5) 5(4) 5 5(6) 6(5) 6 6(7) 7(6) 7

RCR level III

III (IV)

IV (III)

IV

IV (V)

V (IV)

3 3 1 2 2 1

1 2 2 2

1 1 1 2 1 2 1

3 1

no integration and level 7 to full differentiation and integration (see chapter 5 for details, p. 84). Problem for assessing the (meta-)logical competence. Empirically, levels of (meta-)logical competence were assessed as follows. The respondents were asked to analyse and to class the eight tasks/problems administered up to that point (three RCR problems, three Piagetian tasks, two problems for assessing cognitively complex thinking) according to the logic involved. Five developmental levels were defined as follows: (1) No coherent, pertinent answer (2) Only one logic class identified (here that of the Piagetian tasks) (3) At least one further class (dimly) identified, first class identified more accurately (4) First class fully identified, other classes more fully characterised (5) All three classes clearly distinguished and fully characterised Results The results were as follows. Levels of cognitively complex thought and RCR. The frequencies of individually assessed levels of cognitively complex thinking and RCR levels are shown in Table 4.7. The same general appearance is observed as was

Empirical Studies of RCR

67

discussed for Table 4.6 (p. 63). Spearman’s rank correlation coefficient has the value rs = .68, p < .01.7 The aggregated answers (N = 32) to the problem Filling a post were as follows, ordered according to four self-revealed categories (a)–(d): (a) Expression of a personal viewpoint of the respondent: To ‘steal’ an employee from a colleague is simply not right. That cannot be a normal procedure for filling a post. The way hiring is done should be fully transparent and above board. All the people immediately concerned have misbehaved. If Mr Gotschmann ¨ had really known his collaborators, and knew what was going on in his department, he would have given clear instructions as to the procedure for filling the new post; that would have avoided the present mess. But simply to undo the hiring will not work either. Having had such an experience of promotion changes a man, he is no longer the same as before. One can understand that Mr Zosso complains. But the conflict cannot simply be solved by invoking certain principles. In particular, Mr Gotschmann ¨ should not simply impose a solution thought out by him in his lonely corner. A reflection is needed in order to do justice both to the persons involved and to the interests of the company. (b) Immediate considerations of Mr G¨otschmann (before acting): Franz Riedo will not be much more productive in Boschung’s laboratory than in Zosso’s laboratory. However, if he changes laboratories, then Zosso is short of a collaborator. So, that change is not really in the interest of the company. Furthermore, such ‘hijacking’ should not be encouraged because it leads to an undesirable overbidding between our laboratories. Nevertheless, Mr Boschung should not lose face. It remains true that the post has been offered to Franz Riedo. He now knows that he is deemed capable to tackle more demanding jobs. Having seen that possibility in front of him, he will not want to stay in his former less promising post. He has also become aware that right now professionals like himself are in short supply. Probably he already has told others about his promotion, and has made plans based thereupon. At least, that is how it goes in general. If I just stop him from taking on the new post without offering an equally attractive alternative, then he is probably mad, feels pushed around, blames Zosso for the step back, works less well than before the offer, and possibly even joins one of our competitors. Perhaps the first thing to do is to talk to Riedo to find out what attracts him to the new post, whether the work itself, the salary, the promotion prospect, or what else. I also want to know why he has not talked to Zosso before accepting Boschung’s offer. Maybe he needs to be encouraged to look at the dealings also from the company’s perspective. Furthermore, I have to be sure about his competencies, and as to whether there will not soon develop a possibility for promotion either in Zosso’s laboratory or in my third laboratory. I will do that 7

A post hoc Kruskal-Wallis analysis of variance (H test) for levels 4 to 7 further supported the findings of the correlation computation: the mean ranks of levels of cognitive complexity ‘connected’ to the various RCR level scores differed (chi2 = 11.9, df = 3, p < .01). The corresponding U-Test (Mann-Whitney) showed that the significant differences were between levels 4 and 7 (U = 0, p = .05), and levels 5 and 7 (U = 4.5, p < .01).

68

The Theory of Relational and Contextual Reasoning

immediately. Apart from that, I sit on the fence for the time being, and suggest to Messrs Boschung and Zosso to find a solution that is acceptable to all concerned. (c) Subsequent steps: In the unlikely case that such a solution comes about, Mr Gotschmann ¨ invites Boschung, Riedo, and Zosso to a meeting. He makes sure that all involved have understood what is being proposed, and are in full agreement; he then communicates the result to all ‘his’ employees, if only to stop rumours. In case no solution comes forth, he looks for one together with the two laboratory heads, Boschung and Zosso; this with the guiding criteria of (i) serving the company’s interests, (ii) strengthening or least maintaining the previous good ‘climate’ in his department, (iii) sustaining the positive motivation of all collaborators, and (iv) furthering Franz Riedo’s career prospects. In case no agreement comes forth, Gotschmann ¨ single-handedly takes a decision and proceeds as described for the solution by Boschung and Zosso. (d) Administrative improvements: In the future, a job description will be produced for all open posts, specifying in particular if candidates from the department or from the company may apply, or only outside candidates. Furthermore, the personnel administration makes known afresh the rules for filling a post (which expressly exclude internal ‘hijacking’) and ensures their strict application.

The aggregated answers (N = 32) to the problem Interview were as follows, again ordered according to four self-revealed categories: (a) Description of the problem: This is a problem of communication and coordination. The actual interview ‘frame’ must have the right size: small enough to keep the interview on track, but not so small that it hampers spontaneity. Basically, interviewing presents an insoluble dilemma: hold the reins loosely, yet steer the horse to where you want to go. The two desiderata are almost incompatible. It only works with luck. Somehow, the interviewer is walking on a tightrope. In case the interviewee says something of importance to him or her, but which does not stay within the research frame, that person needs to be nudged gently back into that frame. Most interviewees will answer straightforwardly the question as to when they last ate in an Italian restaurant. But if the interviewer wants to know, for instance, the degree of patriotism of the interviewee, the answer may be more guarded. No matter how much the interviewers are burning to advance the research at hand, for ethical reasons they have to respect an interviewee’s unwillingness to communicate what they consider as private, even intimate information. Contrary to the balance scale experiment, where one can immediately observe what happens, in the case of an interview one mostly can only progress by way of inferring, namely from the kind of answer given to the views or the personality of the interviewee. The interpretation of the data is difficult in any case. The balance scale reacts only to the weights and their distribution. The interviewees respond not only to the questions, but also to the way the interviewer puts them, to his or her attitude, body language, and so on; this possibly in a rather idiosyncratic manner. In fact, different interviewers may well obtain differing answers from the same person about the same issue.

Empirical Studies of RCR

69

(b) Specific difficulties: The interviewers experience certain mental associations, and the interviewees possibly different ones. For instance, in the case of the plant task, the interviewers base themselves strictly on the pictures presented, while the interviewees may remember their garden, and their own experience with the effects of water, earth, fertiliser, and so on, and start from there. The interviewers have perhaps a research frame before their mind, and only want answers which fit inside. It may be that in fact they pay no attention to what falls outside that frame. The interviewees are likely to notice such an attitude and change their answering strategy accordingly, thereby falsifying the interview. Sometimes, journalists almost manipulate the person(s) they interview. If one knows that it is an interview for a particular purpose, spontaneity may suffer. One may feel that certain answers are expected, and that they should be ‘correct’. Strange surroundings and the presence of other persons also might falsify the result. One gives one’s best only when interviewed for a first time. When the interview is repeated, boredom may result, and rather than repeating the same answers, one just says what comes to mind and is more amusing. (c) Differences between the persons interviewed: Some interviewees like perhaps to help the interviewer and say what they feel the interviewer wants to hear, or what is ‘correct’, that is, socially desirable. Others would like to get the interview over with and say the first thing that comes to mind. Some persons are more given to moods (and that might make for differing views on the same event) than others. Again, different persons may require differing ‘treatments’ to give their best. An introverted person needs encouragement and stimulation, a extrovert possibly a little calming down (and an encouragement to think more deeply). Some talk immediately, simply just as it comes to them; others first think very hard before speaking, perhaps too hard. Some uncover themselves, others are more secretive. The interviewers need to diagnose whom they have in front of them and act accordingly. (d) How to conduct interviews: As much as possible interviewing volunteers (who are motivated). The atmosphere should be open and confidence-inspiring, the environment familiar and friendly, speaking in a dialect admitted, data protection assured, and the allocated time fully sufficient. In case the interviewee looks nervous or timid, one should begin by relaxing/warming him or her up, and only then begin with the interview proper. As a general rule, the interviewer should not put questions answerable simple by ‘yes’ or ‘no’, not use multiple choice questions, and avoid suggestive questions. By going once more over some of the ground already covered, the interviewer should make sure that the answers were genuine and complete. Maybe the interviewer should see the interviewee again a few days later in order to find out whether new answers did occur to the interviewee, or corrections came to mind. Possibly the interviewer can invite the interviewee to read the transcript of the interview and to comment on it. In case one wants to compare the answers from several persons, then the interview must be more structured, that is, be more like an oral questionnaire. In contrast, to learn the most about an individual, the interview should not be structured; it is ideally of the narrative variety. In case a person having little experience is to be interviewed publicly, it seems fair to indicate ahead of time the theme of the interview so that he or she can make preparations.

70

The Theory of Relational and Contextual Reasoning

Table 4.8 Frequencies of individual scores concerning levels of (meta-)logical thinking and RCR levels. N = 31. (Source: Reich 1995d, p. 13.) RCR level Levels of (meta-) logical thinking 3(2) 3 3(4) 4(3) 4 4(5) 5(4) 5

III

III (IV)

1

1 1

IV (III)

3 1

IV

IV (V)

3 2

1

1 1

3 3 2 1

V (IV)

2 2 2 1

Clearly, a person coming anywhere near these integrated answers was answering at level 7. (Meta-)logical competence and RCR levels. These results are collected in Table 4.8. Once more, no values are found in the right-hand upper corner. Again, for a given level of (meta-)logical competence, e.g. 4(3), various RCR levels pertain, namely IV(III), IV(V), V(IV). Thus, the competence under discussion can also be regarded as a necessary yet insufficient condition for RCR. Kendall’s rank correlation coefficient had the value .52, p < .01.8 The thirty-one scorable participants (one audiotape was partly faulty) distinguished three logical cases: (1) tasks, whose solution is simply correct or wrong (snail, plant, balance scale); (2) problems, where various aspects might be considered contradictory, other aspects seemingly as completing each other, and all aspects are needed for an explanation (pianist, nuclear accident, model for humans); and (3) problems, where the individual attitude and likely action of the protagonists are the most important considerations for finding a solution (filling a post, interview). The aggregated answers (N = 31) to the problem logical classes were as follows, ordered according to the three classes indicated: 8

A post hoc Kruskal-Wallis analysis of variance (H test) further supported the findings of the correlation computation: the mean levels of (meta-) logical thinking related to the various RCR level scores differed (chi2 = 11.7, df = 2, p < .01). The corresponding U-Test (Mann-Whitney) showed that the significant differences were between levels 3 and 4 (U = 33, p < .01), and to a lesser degree between levels 3 and 5 (U = 1.5, p < .05).

Empirical Studies of RCR

71

Class 1: Here we are dealing with natural laws in the domain of physics and biology, which are unchangeable and reproducible. The physical systems are not capable of learning; they do not develop. All these tasks primarily have nothing to do with me. One can focus one’s imagination and reflections onto concrete entities. We are dealing with clear initial conditions and well-defined objectives. Using the right method leads to a single correct solution. The task is to establish mathematical relations between the variables. Thus the competence for scientific work is tested. The solution can be justified by logical arguments. Once a mathematical solution is found, predictions can be made. Therefore, the answers of various persons are surely similar, if not identical. And in the cases of the snail and the balance scale, the solution does not depend on the environment. In all cases, one can repeat a trial or experiment without intrinsic limitations. Class 2: The task is to harmonise two statements, to create a link, to put them under one umbrella, to find out whether they support each other or not. Both aspects often ‘collaborate’. Even if some statement is not ‘logical’, it may be true. Native endowment and the effects of exercising cannot be separated, nor mind and body; they are intrinsically linked. Probably, such states of affairs can be researched further, because certain regularities obtain despite the complexities involved. The environment may play a certain role in these problems. For instance, the pianist may be influenced by the audience, the operating crew may panic when the reactor explodes, and a person’s mind and body are influenced, for instance, by the weather. Class 3: Humans can act according to the most diverse criteria ranging from the cold logic of focusing on gaining an advantage to the warm logic of decent behaviour. But humans also react. Contrary to the experiments with the balance scale, one can’t do experiments with humans as one fancies. When humans are concerned, a different scale of values applies. When the psyche is involved, the situation often becomes irreversible. There are no solutions which are a priori correct or wrong, because the idiosyncratic reaction of each individual counts. People do not all react in the same way. It may be helpful to reflect how one would react oneself. Such a reflection may further progress with understanding. Decisions have to worked out by way of dialogue. In such discussions the aims and wishes of all persons involved should be put on the table, including emotions that play a role. Persons react differently, because they are aware that their experience of life differs from that of others. Sometimes differing moral values play a role. And the actors may change on account of how events unfold. One knows a lot about single factors, but not how they relate to each other: usually many solutions can be envisaged in principle, but none predicted with a high probability. At best one can make statements beginning with ‘probably’ or ‘possibly’, which indicate a tendency. At issue are not logical necessities, but reaching an optimised solution for the future. That requires the capacity to sort out the essentials and calls for social sensitivity and creativity. Not infrequently, the core issue resides in conflicting interests, in the difficulties having to do with living together. There exists no universally agreed sure method for arriving at a solution satisfactory to all persons involved. Moreover, the objective is not always clear, progress towards it not easily measurable. It may be that both maintaining the status quo and making changes in all likelihood will lead to future problems.

72

The Theory of Relational and Contextual Reasoning

Table 4.9 Minimum stages/levels of other competencies for a given RCR level. Appropriate partial levels of dialectical and analogical thinking are also required. (Source: Reich, 1999, p. 145)

RCR

Piagetian operations

Complex thinking

(Meta-)logical competence

II III IV V

concrete operations transition to formal op. established formal op.

level 2 level 3 level 5− level 7

level 1 level 2+ level 3 level 5

But once a decision has been taken, it cannot be fully reversed, one cannot start afresh from scratch. As a rule, the solution depends on many things, all the more if many persons are involved, and is likely to be tailor-made. Therefore, different persons will propose different solutions. At issue is the black box individual, whose behaviour often is not reproducible, and who is given to emotional whims.

lf a third of the difference between levels is ignored, the inter-rater agreements were 97 per cent (RCR), 81.8 per cent (complex thought) and 81.8 per cent ([meta-]logical thought). Discussion of pilot studies 3 and 4 By retroduction RCR was demonstrated to involve (1) elementary, conjunctive, and/or composite operations singly or plurally shared with Piagetian operations and cognitively complex thinking, as well as (2) a logic different from formal binary logic. The respective minimum levels are listed in Table 4.9. These results support H. 1: there are (developmental) commonalities between the thought forms/competencies under discussion. However, RCR is not reducible to any one of the others. How about H. 2, the development of RCR and of cognitively complex thought? Given the quite significant value of the correlation coefficient (rs = .68, p < .01) and the high chi2 values of the Kruskal-Wallis test, H. 2 is supported by the empirical results: the levels of cognitively complex thought may also be interpreted as a developmental sequence. This is in agreement with the notion that during the life course reasoning becomes more and more general, abstract, differentiated, integrated, and structured (Seiler 1994, p. 79; cf. Werner [1948] 1973, 1957).

Empirical Studies of RCR

73

To reach RCR level III, the lowest ‘real’ RCR level, ‘almost’ formal operations and (meta-)logical competence at level 2+ are required. These findings support H. 3 and H. 4, in that below RCR level III formal binary logic is used, and at that level it is transgressed, at least in a rudimentary way. From the manner the ‘logic’ of the RCR problems was described by the participants, it is also clear that at least some of them understand (more or less well) the notion of noncompatibility, even if they do not know that term. The results shown in Tables 4.6 (p. 63), 4.7 (p. 66), and 4.8 (p. 70) furthermore shed light on another issue debated currently. As discussed by Bjorklund (1999, p. 34), in middle childhood psychometric IQ and results of Piagetian reasoning tests correlate at the .4 level. If both tests measure some generalised intelligence, and given that the IQ increases continuously – that is not stepwise – Bjorklund doubts the correctness of Piagetian stage theory. From the present data it appears that stages or levels are indeed just milestones within an ongoing development. The very notion underlying the present level notation, namely level I (II), etc. indicates a more or less continuous change (as does Piaget’s notion of substages). However, at a level/stage milestone, the sum of earlier small developmental steps has led to a noticeable qualitative change, a marked transformation of the mental structure. How does RCR actually develop? A robust answer remains to be researched. The impressions gained from the present study are twofold: on the one hand an increase of factual knowledge may well be one of the causes. That would be consistent with other studies showing the importance for development of prior knowledge as distinct from a higher IQ (Weinert, Bullock, and Schneider 1999, pp. 334–5). On the other hand increased epistemic cognition seems required, in particular reflection on the means of thinking and reflecting (cf. ch. 2, pp. 29–32), including the nature and use of various logics. At least as a temporary strategy, at the current degree of understanding, it may be advisable to study in more depth the underlying complex processes (Reich, Oser, and Valentin 1994) as against searching for nomothetic relations. In his considerations on spiritual intelligence, Robert Emmons (2000, p. 49) evokes four empirical criteria for determining whether a particular candidate for a different type of (multiple) intelligence makes the grade. It seems to me that, mutatis mutandis, these criteria are also applicable to thought forms, and I do so for RCR: (1) The thought form should be translatable into performance, that is, a person possessing it should be able to solve specifiable problems that someone without it cannot. The data presented here, in particular those on the understanding of religious doctrines (chapter 7), clearly show that this criterion is met by

74

The Theory of Relational and Contextual Reasoning

RCR. (2) The thought form must be applicable to a diversity of content domains, not just one (narrow) domain. Pilot study 1 (Oser and Reich, 1987) – demonstrating that RCR can be considered as a pragmatic reasoning schema – the archaeology of RCR (chapter 8), as well as the applications in this monograph demonstrate that criterion no. 2 is met by RCR. (3) The thought form under discussion should be similar enough to other thought forms to be recognisable as such, but different enough to be worth studying. The considerations (chapter 5) and the data presented here document that this criterion too is met by RCR. (4) The thought form must develop from infancy to adulthood. Again, the data of chapters 4 and 7 are evidence that RCR also meets this last criterion. Summary of empirical studies and outlook The existence of the postformal thought form RCR and its developmental levels, having been explained theoretically in previous chapters, were demonstrated empirically. Furthermore, there is empirical evidence that RCR shares ‘components’ with other thought forms, but not their logic, and that (ideally) it can develop from rudimentary stages in childhood to a fully developed stage in adulthood. Development of RCR implies corresponding levels of Piagetian thinking, cognitive complex thought, and putatively of dialectical and analogical thought as well as at least a ‘feel’ for, and minimal use of different types of logic. Additional studies would usefully focus on competence with RCR as a function of age, education, socio-economic status, etc., and on longitudinal studies to ascertain the present level descriptions and the sequencing of the levels. Exploring further the relation of RCR with dialectical and analogical thought would also be desirable.

5

Other Thought Forms and Matching Them to the Problem at Hand

In chapter 1 (pp. 16–19) RCR was hypothesised to share elementary, conjunctive, or even composite operations with other forms of thought. In chapter 3 (pp. 41–3) arguments were provided that Piagetian operations, dialectical thinking, analogical thinking, and cognitively complex thought were relevant other thought forms. This was broadly supported empirically (chapter 4). It is therefore justified to compare and contrast them with RCR and with each other. A second reason is that they are sufficiently different from each other and RCR to serve for a demonstration of the thesis that for best results the thought form employed must be matched to the problem structure. That will constitute the second section of this chapter. Other thought forms relevant to RCR Piagetian logico-mathematical thinking Rather than covering Piagetian logico-mathematical thinking1 in its entirety (cf. Fondation Archives Jean Piaget 1989), after a short recap I concentrate on its critical aspect in regard to RCR: the logic involved. Some core characteristics of thinking according to Piaget (1970) were already indicated in previous chapters. At the formal operational stage, the overall challenge of Piagetian tasks consists in formulating hypotheses, testing them, and coming to a conclusion. Specific objectives are (a) to find out which are the ‘true’ (active) variables (e.g., task ‘combination of coloured and colourless chemical bodies’); (b) to combine variables (e.g. the ‘snail’ task – see Fig. 4.1, p. 60), or (c) to elucidate natural laws (e.g., tasks ‘floating bodies’, ‘balance scale’). 1

I do not enter into the debate on the validity of some of Piaget’s findings and interpretations as well as their amelioration (e.g., Barrouillet and Poirier 1997; Bickhard 1997; Goswami 1998, ch. 8). The portions of Piaget’s work I make use of seem undisputed (cf. Bond 1995).

75

76

The Theory of Relational and Contextual Reasoning

Task Here I focus on the balance scale task (Inhelder and Piaget [1955] 1958, pp. 164–81) with a view to bringing out the structure of a Piagetian task, which among other things ‘justifies’ the application of the logic involved in logico-mathematical thinking. The balance scale task consists in finding out how to balance the scale by adjusting the various weights and their distance from the fulcrum on either side of the beam.2 Two essential, ‘obvious’, yet crucial points are (1) the ‘truth value’ of the variables is time-independent (a weight or a fixed distance over time stay the same weight and the same fixed distance) and (2) the variables are intrinsically independent from each other (a given weight is unchanged at whatever distance it is put, and a given distance stays exactly the same whatever weight is put there). From (1) and (2) and the definition of a balance scale it follows that (3) the situation is strictly reversible (by reversing any change made, the former situation is re-established exactly). Underlying logic Those three conditions (well-defined, time-constant entities; separability of variables; reversibility) are precisely among those which authorise the use of formal binary logic in reasoning about the tasks under discussion. To clarify the foundations of that logic (Kainz 1988, pp. 14–21), a small excursion is necessary. The three pillars holding up formal binary logic are: (1) the maxim of identity: everything is identical with itself, A = A; A cannot at the same time be A and nonA; (2) the law of non-contradiction: a notion cannot possess neither nor both of two relevant mutually exclusive predicates; (3) the maxim of the excluded middle: A must be either A+ or A−, it cannot be anything else. It may be helpful to go a little further into those maxims/laws. Whereas philosophers like Hegel, Wittgenstein, and Kainz find faults with the identity maxim (Kainz 1988, pp. 9–10),3 it is accepted as such in everyday life. In practice, there may be problems of the exact determination of the limits of the categories concerned (such as ‘when does the night end and the day start?’), but it is clear that the day is not the night and the night is not the day. Readers interested in a philosophical discussion of the law of noncontradiction are referred to the considerations of Eric Toms (1962) – summarised by Kainz (1988, p. 18). At a practical level, there are clearly 2 3

The solution is that for equilibrium the sum of the torques [torque = weight × distance from the fulcrum] on the one side has to equal that on the other side. One argument is that one cannot fully know a given entity unless one knows its limits, i.e., what is ‘outside’ of it. Therefore, so goes the argument, that ‘other’ is also part of it.

Other Thought Forms

77

difficulties: is a hermaphrodite not both male and non-male, an ambivert not both extrovert and non-extrovert, and does light not behave in wavelike and non-wave-like ways? As regards the excluded middle, Toms (1962) again discusses the philosophical point of view (Kainz 1988, pp. 18–19). At a practical level, according to the maxim under discussion: ‘Either this is a butterfly or this is not a butterfly.’ What about the egg, the chrysalis? Again, ‘legally’, ‘Either a cow is holy or it is not holy.’ What about the views of Hindus (for whom a cow is holy) and of non-Hindus; is one view to be accepted and the other rejected? ‘Either this neutron is about to disintegrate or it is not about to disintegrate.’ (We only know the answer once the neutron has disintegrated; before that disintegration the state is indeterminate.) Once more, formal binary logic does not fit the logical structure of these and of other cases arising in real life. Applications of formal binary logic If the problem structure is such that formal binary logic is applicable (e.g. Newtonian physics problems, that is Piagetian tasks) then the following operations are valid: use of transitivity, association, commutation, distribution, use of reversibility, solving of syllogisms, working with conditionals, manipulating sets of data (Table 5.1). In other cases (e.g., problems in quantum physics, biology, psychology, sociology, dialectical problems, RCR problems, etc.), the applicability of Table 5.1 cannot be assumed. In many such cases most of those operations will lead to wrong results; the detailed applicability must first be tested. To illustrate the importance of the separability (locality) of the (identical, unchanging) elements to which formal binary operations are applicable, Errol Harris (1987, pp. 34–7) discusses a material implication, a conditional (penultimate example of Table 5.1): ‘If p (the antecedent), then q (the consequent)’ which is understood as a separation of relations. No essential connection is required (and in fact admitted) between the truth or falsity of the two propositions concerned. The implication is true whenever both propositions are true, or both are false, or p is false and q is true: (a)If it rains, the street is wet; (b)if it does not rain, the street is not wet; (c)it does not rain, yet the street is wet are all logically acceptable. The only unacceptable case would be: (d) it rains, yet the street is not wet. That would be a genuine logical contradiction, provided the truth of an implication is accepted as a principle. Harris’s (1987, pp. 36–7) counter-example is the following: If Newton’s mother is a woman, then Newton is a man. Whereas cases (a) (Newton’s mother is a woman, and Newton is a man), and (b) (Newton’s ‘mother’ is not a woman, and Newton is not a man) above work out all right, case

78

The Theory of Relational and Contextual Reasoning

Table 5.1 Various operations based on formal binary logic. Operation

Description/Characteristics

Example

Use of transitivity

Use of logic to avoid a further comparison measurement The result is independent from the grouping The result is independent from the order of the elements The result is independent from the order of the operations Reversing an operation will re-establish the initial situation Drawing conclusions from the premises Drawing conclusions for different situations Sets are dealt with as wholes

a > b; b > c → a > c

Association Commutation Distribution Use of reversibility Solving syllogisms Working with conditionals Manipulating sets of data

(x + y) + z = x + (y + z) (x · y) · z = x · (y · z) x+y=y+x x·y = y·x x · (y + z) = x · y + x · z x+y−y=x x·y : y = x All men are mortal; Socrates is a man; Socrates is mortal If p, then q . p, but not q: logically not possible Set B is a subset of set A; set A and set B overlap

(c) now presents a problem: If Newton’s mother is not a woman, then Newton is a man. The problem arises, because mother and son are not logically independent, they are not separable. Another way to look at this is to appreciate that a conditional of this type states a sufficient condition, not a necessary one. Before concluding this section on logic, a brief discussion of class sets will prove useful later (chapter 6). The possible relationships of class sets according to formal binary logic are shown in Fig. 5.1. According to Venn ([1881] 1971, p. 7), the four basic possibilities can be described as follows: in case 1 both classes are logically identical, they overlap completely. Their extensions are identical, but not their intensions. An example would be the class of equilateral triangles (A), and the class of equiangular triangles (B). In case 2 one class is a subset of the other class. For instance, if A is the class of religious norms, and B the class of moral norms, then 2a would mean that moral norms are a subset of religious norms (theological view), and 2b that religious norms are a subset of moral norms (view of humanistic philosophy). In case 3 the two classes overlap. An example would be the class of automobiles (A) and the class of vans (B) A car-a[nd]-van transports persons and therefore belongs to class A, but it also transports goods and hence also belongs to class B. Case 4 depicts the situation where independent classes are totally separated, no interaction is assumed. An example would be a scientific world view determined by a ‘rational’ investigation of nature (A), and a

Other Thought Forms

79

1

A

2a

A

B

B

3

A

2b

A

B

4

B

A

B

Figure 5.1 Venn diagrams of class sets.

religious world view determined by religious feelings and mystical experiences (B). A helpful point – but also a limitation – of Fig. 5.1 is that (within formal binary logic) no other relations exist. Dialectical and RCR-type ‘relations’ are excluded on principle, because they do not meet the above conditions underlying the use of set theory (well-defined, time-constant entities; separability of variables; reversibility). The diagrams of Fig. 5.1 help to clarify class relations and are therefore often more useful in (controversial) discussions than mere words (Reich 1996c). However, one has to be aware that they depict essentially class extensions (and intensions by implication, at least in case 1), but not the nature of any interaction. If that also needs depicting, a second set of diagrams may be used to depict the relationship, for instance the nature of any hierarchy in case 2, the kind of interaction in a common ‘area’ in case 3, etc. The exact nature of the interaction might still not be depictable though: is it causal, one of information transfer, dialectical, appropriate for applying RCR? Summing up this subsection in the words of John Macnamara (1994, p. 150): classical logic is ill equipped to handle logic of ordinary discourse precisely because classical logic derives mainly from an analysis of arithmetical sentences. Arithmetic is an unusual domain of discourse because (a) the objects in the domain are eternal and unchanging. All properties are necessary ones. And (b) only a single basic count noun is required in arithmetical sentences: number or set depending on the level of one’s work. Any other count noun can be

80

The Theory of Relational and Contextual Reasoning

Table 5.2 The sixteen binary operations. p, ¬p, q, ¬q = independent variables; x = dependent variable. T = true; F = false. (Explanation in text.) x if

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

p&q ¬p & q p & ¬q ¬p & ¬q

T T T T

T T T F

T T F F

T F F F

F F F F

T T F T

T F T T

F T T T

T F T F

F T T F

T F F T

F T F T

F F T T

F T F F

F F T F

F F F T

defined as a subset of numbers or sets – for example prime number, finite set. (emphasis in original)

Examples of formal binary logical operations The concept formal operations refers to operations that abstract from the substantive content of knowledge. If they are governed by binary logic, one lives in a dichotomous world of yes or no, true or false. One guide in that world is the set of sixteen operations, presented in Table 5.2. To understand the nature of Piagetian formal operations further, and to contrast them with RCR, we shall discuss Table 5.2 in some detail. Before that let me emphasise, however, that I am not entering into the debate as to whether Piaget had an adequate grasp of logic, or whether the sixteen operations are a correct model of his understanding of formal operations (e.g., Smith 1987). Above all, the point is to compare and to contrast formal binary operations with dialectical, analogical, cognitively complex thinking, and RCR ‘operations’ after going through the examples of the next several pages. Readers who want to relax for a moment can go right on to the amusing story on page 83 (‘A Piagetian task from the Internet’). Table 5.2 shows the sixteen different possibilities of how one of the two dichotomous values (x, ¬x) of the dependent variable x might be a function of the dichotomised independent variables p and q (p, ¬p; q, ¬q, where ¬ means not). The table was created by starting (column 1, on the left) purely formally with four times T (= ‘true’, i.e., probability 1), then in the columns 2 to 5 each time replacing one T by an F (= ‘false’, i.e., probability 0), next varying the position first of one F, then of two Fs, and finally of three Fs.4 4

In the literature, the content of Table 5.2 is often organised differently, for instance, with rows 2 and 3 inverted, or with a different order of the columns. However, that does not change the truth functions of a given column, which is the core content of Table 5.2.

Other Thought Forms

81

Column 1 represents the case of ‘complete affirmation’ (four times true): whatever the combination of p, ¬p and q, ¬q values, x has the same (dichotomous) value (not a very likely situation in the real world – the conclusion might be that the ‘wrong’ variables were studied). According to column 2,5 x keeps the value of column 1, except for the combination ¬p, ¬q (‘disjunction’ of p and q; ‘inclusive or’). Column 3 indicates a combination where only q determines x (‘affirmation of q’), column 4 that where the combination (conjunction) of p and q6 is decisive. Column 5 is the inverse of column 1 (all truth values are inverted), its logical complement (‘complete negation’). As to the truth values of column 6, only the combination p, ¬q does not lead to x (‘implication’); in the case of column 7 the combination ¬p, q (‘converse implication’) has the same effect. Column 8 (‘incompatibility’; ‘alternative or’) represents the logical complement of column 4.7 In column 9, p determines x, irrespective of q (‘affirmation of p’), in column 10 either p alone, or q alone takes that role (‘reciprocal exclusion’; ‘exclusive or’). Column 11 indicates the reign of either p and q or neither of them as determinants of x (‘equivalence’ of p and q), in column 12 it is q alone or neither p nor q (‘negation of p’), in column 13 either p or neither p nor q (‘negation of q’) determines the overall truth value. Columns 14–16 refer to the reigns of only q (‘converse nonimplication’), only p (‘nonimplication’), and neither of them (‘conjoint negation’).8 As already indicated and a closer inspection shows in detail, certain relationships obtain between the columns, that is between the truth functions represented in the various columns of Table 5.2. Thus, an inversion, 5

6

7

8

Column 2 represents the truth function termed ‘disjunction’, a function often written as (p v q). In that expression v stands for vel, the Latin term for or. As is apparent from column 2, a disjunction has the truth value 1 (= true) if either p is true, or q, or both (inclusive or). An everyday example for such a state of affairs would be, ‘automobiles have either hub brakes (p) or disk brakes (q), or hub brakes on one pair of wheels and disk brakes on the other’. In electronic circuitry, the corresponding Boolean function is termed ‘OR’. This truth function is called ‘conjunction’, and may be written p · q. An everyday example would be, ‘Getting a driver’s license requires us both to have the required age, and to have passed the tests.’ In electronic circuitry, the corresponding Boolean function is termed ‘AND’. This truth function is called ‘incompatibility’, and may be written as ¬p v¬q. It is true, if either p, or q, or neither p nor q obtains (alternative or). An example would be, ‘The car under that tarpaulin is either a Chevy, or a Ford, or something else.’ In electronic circuitry, the corresponding Boolean function is termed ‘NAND’, the logical complement of AND (column 4). This truth function may be called ‘conjoined negation’, and be written as ¬p · ¬q. An example would be, ‘This pendulum (dichotic weight ¬p, and dichotic length of suspension ¬q) has neither the weight p nor the length of suspension q.’ In electronic circuitry, the corresponding Boolean function is termed ‘NOR’, the logical complement of OR (column 2).

82

The Theory of Relational and Contextual Reasoning

transforming all Ts into Fs, and all Fs into Ts, that is into their logical complements, represents a ‘negation’, N. Examples are the transformations 1 → 5; 2 → 16; 3 → 13; 4 → 8 and vice versa. A correlative transformation, C, is concerned with functions having the same truth values in the first and the last rows of Table 5.2. That transformation changes the T and F values of the middle rows into F and T values, respectively. For example, a conjunction (column 4) is the resulting ‘converse’ of a disjunction (column 2), and vice versa. A reciprocal transformation, R, relates a proposition to its ‘obverse’. This concerns functions which share the truth values of the middle rows of Table 5.2; in the first and last rows that transformation entails a change of T into F, and vice versa. An example is the transformation of a disjunction of p and q (column 2) into an incompatibility (disjunction of ¬p and ¬q – column 8), and vice versa. An affirmation of p, q is transformed into a negation of p, q, and vice versa. To see the formalism of Table 5.2 at work, we use it to assist with solving a simplified, dichotomised pendulum task (cf. ch. 4, p. 61). Two pendulums have a heavy weight (p), and two a light weight (¬p). In each weight category, one pendulum has a long suspension (q), the other a short suspension (¬q). The question is, ‘What are the conditions for obtaining a high oscillation frequency of the pendulum (x)?’ Experimenting with the four pendulums yields the empirical answer: ‘The pendulums with the short suspension (¬q) exhibit the high frequency.’ A look at Table 5.2 shows that the result is represented by column 13. Having understood the formalism of the sixteen operations, one can surmise that the condition for ¬x, the low frequency, is represented by the logical complement of column 13, that is by column 3. An experimental check will rapidly confirm that conclusion. A conscientious researcher would now look at the other fourteen truth operations and check whether he or she has tested them empirically, or whether they can be considered irrelevant for the present case (e.g., nos. 1 and 5). If both issues are settled, the researcher can be certain that he or she has obtained the correct result. The INCR group We have just seen that through a correlative transformation (C) disjunctions can transform into conjunctions, and vice versa. Affirmation and negation of p or q can be transformed into each other through a reciprocal transformation (R). A negation (N) leads simultaneously to both transformations. The three types of transformations constitute the INCR group depicted in Fig. 5.2. There are three ways to get from p v q to ¬p · ¬q: via (1) a negation, (2) a correlative transformation followed by a reciprocal transformation of the result, or (3) by a reciprocal transformation followed by a correlative transformation.

Other Thought Forms p v q

83 p v

q

R C

C

N R

p q

p

q

Figure 5.2 INCR group. I stands for identity transformation (resulting, e.g., from a double negation), N for negation, C for correlative transformation, and R for reciprocal transformation (explanation of transformations in text).

As an example, generalising the relations depicted in Fig. 5.2 can be used for assisting in a modified version of the snail task (Piaget [1946] 1972, pp. 59–109). A snail moves on a board, which is displaced on a table (cf. Fig. 4.1, p. 60). These double movements go on for some time. The task is to find a way for reconstituting exactly the situation reigning at the start. Obviously, one way is simultaneously to reverse exactly all the movements made, to operate a negation. However, according to Fig. 5.2, and if the movement of the snail is represented by C and that of the board by R, then one can immediately discuss two more possibilities, namely, to correct separately the movements of the snail and of the board, this in either order. Many similar examples can be found from balancing a scale to making up a seating order for a dinner party using two criteria. A Piagetian task from the Internet The following story from the Internet provides, I hope, an enjoyable example of Piagetian formal operations at work, that is applying the formalism just reviewed to a hypothetical case: ‘Is Hell exothermic (releases heat) or endothermic (absorbs heat)? Support your answer with proof.’ Most of the students wrote proofs of their beliefs using Boyle’s Law (gas cools off when it expands and heats up when it is compressed) or some variant. One student, however, wrote the following: First, we need to know how the mass of Hell is changing in time. So, we need to know the rate that souls are moving into Hell and the rate they are leaving. I think that we can safely assume that once a soul gets to Hell, it will not leave. Therefore, no souls are leaving. As for how many souls are entering Hell, let’s

84

The Theory of Relational and Contextual Reasoning

look at the different religions that exist in the world today. Some of these religions state that if you are not a member of their religion, you will go to Hell. Since there are more than one of these religions and since people do not belong to more than one religion, we can project that all people and all souls go to Hell. With birth and death rates as they are, we can expect the number of souls in Hell to increase exponentially. Now, we look at the rate of change of the volume in Hell because Boyle’s Law states that in order for temperature and the pressure in Hell to stay the same, the volume of Hell has to expand as souls are added. This gives two possibilities: (1) If Hell is expanding at a slower rate than the rate at which souls enter Hell, then the temperature and pressure in Hell will increase until All Hell breaks loose. (2) Of course, if Hell is expanding at a rate faster than the increase of souls in Hell, then the temperature and pressure will drop until Hell freezes over. So which is it? If we accept the postulate given to me by Ms Therese Banyan during my Freshman year, ‘That it will be a cold night in Hell before I sleep with you’, and take into account the fact that I still have not succeeded in that area, then (2) cannot be true, and so Hell is exothermic. This student got the only A.

Having discussed the nature, the potential as well as the limitation of Piagetian operations to cases where formal binary logic applies (as allegedly in the Hell task), we come to cognitively complex thinking, another major ingredient of RCR.

Cognitively complex thinking Cognitively complex thinking originally was dealt with as an instance of information processing (Schroder, Driver, and Streufert 1967). This type of thinking involves differentiation (bringing out differences of fact, of possible interpretations, and valuing) and integration (attempts at linking various elements in order to arrive at an overall assessment). The seven-level scale for assessing degrees of cognitively complex thinking (Baker-Brown, Ballard, Bluck, Vries, Suedfeld, and Tetlock 1992) is reflected in the five levels of RCR (Table 4.7, p. 66). The seven-level scale can be described briefly as follows: level 1, no differentiation or integration (only one point of view, no other comes into the field of vision); level 2, beginning of differentiation; level 3, clear differentiation (at least two approaches to dealing with the information received; ‘either/or’ is in view); level 4, beginning of integration (both-and becomes a [weak] possibility); level 5, explicit integration; level 6, systematic approach including an evaluation of the different possibilities and a comparison of their likelihood to be most promising; (7) elaboration of a framework that can ‘house’ the various considerations of the lower levels. One can appreciate the importance of the cognitive complexity of reasoning (and thence of RCR) from a study by Peter Suedfeld and Philip Tetlock (1977). They graded sets of diplomatic communications

Other Thought Forms

85

Table 5.3 Level of cognitive complexity resulting from an analysis of diplomatic notes exchanged during five international crises.

Crisis

Result

Mean level of complexity

1911 1914 1948 1950 1962

peace war peace war peace

4.64 1.95 2.75 1.71 4.71

Source: Suedfeld and Tetlock (1977, p. 182).

in particular during the following five crises: (1) Agadir (1 July–4 Nov. 1911), (2) outbreak of First World War (26 June–4 August 1914), (3) Berlin Blockade (22 June–18 Sept. 1948), (4) outbreak of Korean War (25 June–4 July 1950), and (5) the Cuban Missile Crisis of 1962. Results are presented in Table 5.3. In view of the striking differences, an argument can be made that cognitively complex thinking should be encouraged (and thence RCR) if one wants a more peaceful world. The counter-argument one sometimes hears (but about which I have reservations) is that in times of adversity persons thinking in an undifferentiated way are better fighters – maybe that the resoluteness of the ‘kamikaze fighters’ in war or warlike action explains how one can get to such a view. Dialectical thinking Dialectical thinking has a long history; it was notably practised by Greek philosophers. In the course of the centuries, it took on many shades of meanings, and different schools define and practise it differently (e.g., Harris 1987; Kainz 1988; Reese 1982). When Basseches (1984, p. 20) started his work, he looked among others at the writings of Hegel, Marx, Darwin, von Bertalanffy, and at Piaget’s theory of adaptation through assimilation and accommodation.9 His general description of dialectical thinking ‘in the broadest sense’ is said to be applicable to the published works of the authors named and to ‘the way in which ordinary adults confront common problems of living’. 9

The paradoxical situation is that Piaget could not have worked out fully his ‘complete’ theory of cognitive development by using exclusively formal binary operations (which are at the pinnacle of cognitive development according to that theory). Dialectical thought is involved in the description of the development itself as distinct from the description of a given stage and the logico-mathematical thinking at that stage.

86

The Theory of Relational and Contextual Reasoning

Figure 5.3 Through becoming, being and nonbeing transform into new being and new nonbeing.

Dialectical thinking is related to ‘here-and-now’ thinking in the same way that a motion picture is related to a still photograph. The motion picture does not ‘contradict’ the still photograph but combines a series of them according to the laws of motion. Dialectical thought acknowledges that a change of quantity may become a change of quality. The dialectical ‘conflict’ of content and structure eventually leads to a restructuring. Thus, dialectical thinking also deals with the interruption of continuity. With Basseches (1984, pp. 21–30), dialectical thinking, as understood here, is based on an ontology of wholeness and of change, in which old structures (systems) give way to new structures (systems). Wholeness means that first, entities are emphasised over against monadic, independent elements, and second, the entities themselves are de-emphasised relative to the process of existence as a whole. Dialectical change is depicted in Fig. 5.3. As will be apparent from Fig. 5.3, the dialectic outlook emphasises intrinsic internal relations. The relations among parts within a whole make the parts what they are, and thus the relations are ‘internal’ to the nature of the parts. At the same time, the relations form the internal structure of the whole . . . as these relations change, fundamental change in what exists occurs . . . Thus the emphasis on change, wholeness, and internal relations are interconnected in dialectical ontologies. (ibid., p. 22)

Examples of corresponding dialectical analyses may be found, for instance, in the works of Karl Marx and of Thomas S. Kuhn.

Other Thought Forms

87

Basseches (ibid., p. 29) points out that dialectical analyses are not without costs. The ‘willingness to question the permanence and intransigence of the boundary conditions of a problem, and to ask about situations which lie beyond those boundaries’ might endanger intellectual security. It will depend on each individual, whether trading off such a security for freedom from intellectually imposing limitations on oneself and other people is considered worthwhile or not. I do not want to close this subsection without referring to Klaus Riegel, who revitalised the study of dialectical thinking (e.g., Riegel 1978), but unfortunately left us much too young. Jack Meacham (1999) sees the current importance of dialectical theory in its facilitating the understanding of issues in multiculturalism. From that perspective he reviews Riegel’s work and in particular its commonalities with the work of the Russian psychologist Rubinstejn. Riegel’s dialectical psychology was motivated in particular by the observation that ‘traditional psychology retains a strong commitment to the belief that traits and abilities remain stable, and to the concepts of balance and equilibrium’ (Meacham 1999, p. 135). Actually, imbalances, disequilibria, questions, doubts, and challenges may provoke developmental changes and therefore need study. A major point is to consider changes not just within a given domain, but to pay attention ‘to transactions occurring among major developmental progressions – biological, psychological, and culture-historical’ (ibid., p. 138). A view inspired by Fig. 5.3 is probably better able than other approaches to understand racial strife and to devise medium-term and long-term remedial measures with a view to achieving a viable multiculturalism. The lesson is clear: a complex understanding of relationships and their evolving nature is a core issue. Dialectical thought is geared to tackle it, and therefore needs to be supported. Notice too that this is already an example of matching the thought form to the problem structure. Analogical thinking Not only for children do analogical arguments play key roles for finding a mental way when facing a new problem (cf. Gentner, Brem, Ferguson, Markman, Levidow, Wolff and Forbus 1997; Moshman 1998, pp. 954– 5). That remains true even for scientists (Oppenheimer 1956). In the laboratory, four-year-olds can complete analogies such as ‘bird is to nest as dog is to?’ (bird:nest::dog ? – Goswami 1998, p. 223). To come to the solution, children must map the relation lives in that links bird to nest to the item dog in order to find kennel.10 With development, more 10

Goswami (1998, pp. 223–4) had presented the task to 4-year-old Lucas as a picture of a bird and that of a nest with three eggs in it, and told a story around those pictures,

88

The Theory of Relational and Contextual Reasoning

complex relational mappings can be carried out. Goswami’s (ibid., p. 253) example is to use an analogy from the mental representation of familiar height relations such as Daddy > Mummy > Baby as a basis for mapping to a transitive inference problem such as ‘Tom is happier than Bill, Bill is happier than Mike, who is happiest?’ The solution involves mapping Tom to Dad, Bill to Mummy, and Mike to Baby. As Daddy is the tallest, Tom is the happiest. Whereas in the first example (bird:nest::dog?) the mapping referred to a single relation, now the mappings refer to an ordered pair of relations. For RCR, only some of the operations composing analogical thinking are required, namely searching for commonalities and differences (cf. Medin, Goldstone, and Gentner 1990). Concluding remarks The characteristics, and main differences of the four thought forms, and RCR, can be summarised as shown in Table 5.4. Having discussed the characteristics of the four thought forms under discussion, and earlier those of RCR I finish this section by applying all five in turn to a micro-analysis of an impending partnership break-up (cf. Basseches 1984, pp. 26–7). In real life hardly anybody will argue in that way (and certainly not for so brief a time), people being more pragmatic, but to use pure forms of thought in this illustration might help to get a better sense of what each brings out. John and Barbara are Piagetians. ‘It’s all your fault, Barbara, you never understood me.’ ‘And you John, what did you really do to make me happy? I am deeply disappointed.’ ‘Well, maybe we were never meant for each other!’ For John and Barbara only black and white exist in their dichotomous world, only fully right or fully wrong, in short the break-up is to be analysed in terms of the sixteen binary operations of Table 5.2. The result is likely to lead singly or in combination to (a) a lowered selfesteem, (b) anger at the partner, (c) devaluing the relationship as long as it lasted, and (d) a hesitancy to make future commitments. Dick and Joan are cognitive complex thinkers: ‘You know, Dick, I shall miss sailing with you, we were really a good team.’ ‘Yes, and we always knew where we wanted to go. But then, you were too easy with spending money, and that put a strain on our relationship.’ ‘Well, I thought that with all the raises you told me about we could afford it.’ ‘Now Joan, there is a lot I could say to that and to other things. Nevertheless, I keep some including a dog. Lucas argued that birds lay eggs and dogs ‘lay’ puppies, so he insisted that the missing analogue was puppy! In any event the answer shows analogical thinking exploring far afield and coming up with a solution.

Other Thought Forms

89

Table 5.4 Main differences of (1) Piagetian operations, (2) cognitively complex thinking, (3) dialectical thinking, (4) analogical thinking, and (5) RCR. A, B, C, D are the ‘variables’, ‘dimensions’ or characteristic ‘aspects’ concerned. No.

Nature of aspects A, B, (C, D)

Relationships between A, B, (C, D)

1

A, B, (C . . . ) are part of the same conceptual system; are intrinsically independent from each other; they can variously be ‘linked externally’ with each other. Piagetian tasks often involve the elucidation of such relationships within a given closed system, for instance in the case of the pendulum task or the balance scale task. Not defined; a large variety pertains as in the case of human relationships. That large variety invites wideranging exploration (differentiation and integration). Within an open system, A and B belong to different subsystems; they determine each other as do ‘being’ and ‘nonbeing’, ‘as such’ and ‘for us’, ‘assimilation’ and ‘accommodation’. A, B are part of one reference system, C, D of another system. Properties/ functions of A and B correspond to analogous properties/functions of C, D: in the case of linked traffic lights, a car driver ‘surfs’ along like a surfer on the ocean waves.

In agreement with (time-independent) formal binary logic (tertium non datur) transitivity, associativity, distributivity, commutativity, reversibility (the negation of a negation leads exactly back to the origin) pertain. Logical contradiction is not to be tolerated; the overall system is of a static and synchronous nature. Experience of human life, recognition of protagonists’ motivations, of their objectives, of personality variables, etc. are more helpful for insights than ‘logics’. Relationships are dynamic and have to do with change and development. The negation of a negation leads to something new: through becoming, nonbeing turns into new being and being into nonbeing. In order for the analogy to work (enlarged search space, better understanding), the similarities of the properties/functions need to be sufficiently strong and evident. Nevertheless, almost by definition there will always also be marked differences. Negations mean an iterative refocusing from A onto B and so on, with ideally each time a gain in understanding of their relationship – based on a logic of non-compatibility (which is not incompatibility).

2

3

4

5

A, B, (C . . . ) belong to different categories within the frame of a given explanandum; they are ‘permanently’ linked intrinsically; ‘completely understandable’ in their own context; all needed for a genuine insight.

(Source: Reich 1999, p. 139)

90

The Theory of Relational and Contextual Reasoning

good memories, and anyway, next time, I shall know better.’ Dick and Joan clearly differentiate and integrate their experience considerably more than our Piagetians. The break-up is less traumatic for them than for John and Barbara, and possibly Dick and Joan will still meet occasionally to speak about their respective new partnerships. Ron and Liz are dialecticians. ‘Now Ron, who would have thought when we first met that it would end with us this way? Do you remember how happy we were, the things we did together?’ ‘Of course I do, Liz, and I shall go on valuing those times. But then, we have changed since. You have started your new career.’ ‘And you have developed new interests I simply cannot share.’ ‘Well, maybe, Liz, one day we will move closer together again, but for the moment a separation seems the most reasonable thing to do, don’t you agree?’ Looking thus for changes in either partner within and outside the relationship embeds the break-up into the flow of life. It could even appear as a gate towards further development, and would leave positive remembrances intact. Walt and Anne often use analogies to explain things. ‘You know, Walt, this is just like what happened with your brother Ted. One day he had enough and just broke up with his partner, I never understood why.’ ‘No Anne, that is the wrong comparison. Rather take Frank and Nancy. They were together for quite a while until it became clear to them that their partnership was nor really fulfilling. So they parted ways in mutual agreement.’ The good aspect of this discussion is that both partners try to understand what happened (and they do it without directly attacking each other). However, as no two cases of human relationships are identical, there are limitations to this approach. Walt and Anne may never get to the bottom of their impending break-up unless they really focus on their personal case. Bob and Betty favour RCR. ‘Bob, it seems to me as if lately we have a problem with our relationship.’ ‘Oh, why do you say that? We still like to travel together to interesting places and we have a good time sharing our impressions, don’t we?’ ‘Yes indeed, but for one thing, I enjoy less and less jogging or skiing with you, you are just too strong for me.’ ‘Well, should I admit that your love of going to concerts and expecting me to come along each time is getting a bit much for me? I am not against concerts, but there has to be a measure to everything.’ ‘I am glad you are so frank about it, Bob. Maybe we should do more things we like to do together, and learn how not to get on each other’s nerves by either reducing or transforming those less pleasant occasions.’ ‘That may not be easy, Betty, but let’s try!’ By way of bringing in the context and differentiating their respective experiences, Bob and Betty give their partnership a second chance.

Other Thought Forms

91

If any conclusions can be drawn from these imagined, much too rudimentary ‘vignettes’, it is clearly that Piagetian operations are of limited helpfulness in this context. The other thought forms have more to offer, each in its own way. After this interlude we are ready to see all five forms fully at work, each where it can serve most. Matching the form of thought to the structure of the problem Development of children’s natural epistemology When one reads about the various stages of development of epistemic competence (e.g., King and Kitchener 1994; Kohlberg’s 1984, pp. 432– 6 summary of Perry’s work), one might get the (mistaken) impression that at a given stage, the mental tools, the epistemic approach characteristic of that stage is applied globally, no matter what the logical structure of the problem may be. This monograph contends, rather, that for optimal results, the particular thought form (with its inbuilt logic) used to reason about a given problem should match that problem’s structure (cf. Wood 1983). This matching may succeed more or less well, depending on the person’s developmental stage, but that variation in itself is not an argument against the principle of matching the ‘right’ thought form to the structure of the problem about which one is thinking. In other words, irrespective of the approach used most often at a given age/developmental level, several approaches should be available, and the most appropriate used in a given case. Thus, the thrust of this section is to argue for a person’s development and targeted use of a variety of thought forms over against cultivating one or perhaps two, and ‘adapting’ them pragmatically if forced by ‘reality’. Are there any empirical grounds for the possibility of implementing such a stipulation? Annick Mansfield and Blythe Clinchy (e.g., 1985) reported several studies on the development of children’s natural epistemology. The tasks assigned to children in these studies were to decide (1) whether a particular item floats or sinks (verifiable fact), (2) whether a dog can understand human language or not (debatable fact), (3) whether a new boy is ‘yucky’ or nice (interpretation), (4) whether a new food tastes good or not (personal taste). These researchers found that (a) young children (age 3 to 4 years) exhibited an absolutist epistemology (one with no concept of subjectivity), (b) children aged 7–10 years displayed a dichotomous epistemology (one able to differentiate between objective fact and subjective opinion), (c) adolescents and young adults evidenced an integrated epistemology (one in which subjectivity plays a role even in matters

92

The Theory of Relational and Contextual Reasoning

of ‘fact’, and in which inferential knowledge and expertise can be brought to bear even on matters of taste) – note the implicit development of RCR when going from (a) to (b) to (c). The young children relied mostly on their own experience or on ‘reason’ (= trial) to justify their judgement. In middle childhood, children still relied prominently upon personal experience, but they offered subjective justifications for their judgements regarding issues 3 and 4 (interpretation and personal taste) and appealed to outside authority to judge the correct answer to issue 2 (debatable fact). College students invoked personal experience and subjectivity in much the same way as the 10-year-olds did. Unlike the children, however, they referred more often to outside authority regarding both verifiable and debatable facts, as well as matters of taste. Thus, at least one series of studies demonstrate that children aged 7 and up discriminate between different issues/problems as regards (a) the possibility of a consensual ‘true’ statement vs. a personal opinion, and (b) the optimal way to reach such a consensual statement. To cite an example from an RCR interview with a 9-year-old girl (‘On the wall before us we see two paintings. Can we argue which is more beautiful?’): ‘Not really, that is a matter of personal taste. But we can argue as to whether either picture is damaged or not.’ As this example also demonstrates, children already have a sense at that age for matching their approach to ‘reasoning’ about a problem to the problem structure, at least in certain cases. Five examples of matching the thought form to the problem The problems To exemplify and concretise the issue of this section – matching the thought form to the structure(s) of the problem at hand – we will analyse five problems, which look more or less alike, but whose solutions involve different logics (= principles or rules governing the proper use of reasoning, cf. ch. 1, pp. 15–16). These logics are embedded in five thought forms, the fifth being RCR. The first problem poses the question, what determines the frequency of a pendulum’s oscillations: (a) the weight of the pendulum, (b) the amplitude (range) of the oscillation (whether originating from releasing the displaced pendulum from varying heights or giving it a more or less strong push from the resting position), (c) the length of the suspension, or (d) something else? The second problem involves a scenario: a woman’s husband is an alcoholic. He has recently returned home from a clinic after a second attempt at curing his alcoholism. His wife tells him, ‘The first treatment

Other Thought Forms

93

had no effect. If you come home drunk one more time, I’m going to leave you!’ He then comes home drunk. What will the wife do? In the third problem’s scenario, a trade union demands a 6 per cent wage increase, arguing that higher wages will spur greater consumption, thereby creating new jobs and driving down unemployment. The employers respond by offering the union a wage increase of 2 per cent. The employers argue that increased consumption will only lead to higher levels of imports, while higher wages will make exports less competitive, so that as an overall effect no net decrease in unemployment will be achieved. They claim that only by producing marketable, competitively priced products and services will they be able to hire additional workers. Does this mean a stalemate? Or is there a way to negotiate between these two positions? The fourth problem involves a teacher who wants to assist students with learning about the functions of flower stems. With that aim in mind, she invites the students to compare the flower stem to a drinking straw. The question is, why could that be helpful to the students? In the fifth problem’s scenario (already introduced in chapter 1), a TV news station reports on an accident in a nuclear power station. The main cooling pump had stopped working, and the back-up pump did not function. The emergency shutdown did not function either. To add to the difficulties, the operating crew became aware of the danger rather late and then underestimated it. The water temperature suddenly rose. A steam pipe cracked and leaked radioactive steam. What or who is to blame? What should be done to avoid another such accident in the future? In each case, one is being asked to find a solution to the problem presented, a process that involves analysing, possibly experimenting, reasoning, judging, and drawing conclusions. But there are important differences between the structures of the five problems, and therefore different thought forms have to be applied for best results. The Piagetian task The first task is of a physico-mathematical nature. One expects a single, clear-cut solution, which can be expressed in the form of an equation for the frequency of the pendulum. One might find this equation inductively by a method based upon experimentation, making many precision measurements of pendulum oscillation frequencies under widely differing, strictly controlled conditions which are varied singly, each in turn. One would then infer from the experimental data the answer to the problem, that is to say, the best possible relation between the independent variable(s) and the frequency of the pendulum’s oscillation. Formal binary logic is helpful, even indispensable (Inhelder and Piaget [1955], 1958, pp. 67–79), in reaching this solution (cf. ch. 4, p. 61; ch. 5, p. 82).

94

The Theory of Relational and Contextual Reasoning

The cognitively complex problem If one attempted to apply the same kind of logic to the second problem, i.e., to infer from the ‘data’ – here the wife’s statement – what is the most likely conclusion, the answer clearly would be, ‘The wife will leave her husband.’ After all, she has told him that if he came home drunk one more time, she would leave him; he has done so and thus, she must leave, at least according to the classical logic of conditionals. But when we presented this problem to participants in our interviews, only a 9-year-old boy gave that categorical a response. By contrast, an 11-year-old girl said, ‘Maybe she was just menacing.’ And we might predict that an experienced social worker would want to know much more before hazarding a guess about the wife’s behaviour, if he or she were willing to guess at all. The social worker might ask such questions as, ‘Was this the first time she had threatened to leave?’; ‘Did she really mean it or was this a self-protective move that would allow her to say to herself, “I tried. I did the best that I could”?’; ‘What is her relationship with her husband really like?’; ‘Was her father perhaps an alcoholic?’; ‘Do she and her husband have children?’; ‘If so, how does the father treat them when he is drunk?’; ‘What alternatives to living with her husband are actually viable for her?’; ‘Could she earn a living?’; ‘Does she have friends and relatives who support her?’ Clearly, one could posit at best a probable outcome given an intimate knowledge of the family concerned and a rich personal experience of human life. Piagetian formal operations would hardly help one to solve this problem, because the wife’s attitude does not appear unambiguous; that is, it does not appear to carry a time-independent ‘truth’ value. Nor is the wife’s attitude, presumably, fully reversible. Therefore no simple, straightforward cause–effect relationship exists comparable to what obtains in the pendulum case. The social worker’s questions differentiate between several possible causes and motives, weighing the ‘data’ from a number of perspectives. In short, solving problem no. 2 requires one to differentiate between various aspects of the problem and to integrate the partial results. Cognitively complex thinking (and experience of human life) is called for, not applying formal binary logic. The dialectical problem As regards problem no. 3, the negotiation between a trade union and the employer(s), everyday ‘real world’ experience teaches one that in the end, unions and employers usually achieve some sort of resolution, stalemates between the two sides do not extend ad infinitum. The resolution is often unpredictable, however. At least, one cannot predict it with precision. One reason for this lies in the multitude of variables and the diversity of influences that co-determine any outcome. Another reason lies in the fact

Other Thought Forms

95

that intermediate steps in the process of negotiation are often irreversible steps that change the ‘givens’ or ‘data’. Indeed, the eventual solution may well emerge out of a series of provisional iterations, possibly influenced by strike action, pressure of public opinion, government intervention, etc. In order to ‘solve’ this problem, both negotiating parties must understand the dynamics of the process, its dialectics. And one salient feature that distinguishes dialectical logic from formal binary logic is that ‘a negation of a negation does not lead back to the original position/situation (as in the case of formal binary logic), but to something new’. The analogical problem What kind of logic is involved in problem no. 4, learning about the flower stem? Evidently, the teacher wants to ground the children’s thinking initially in something that they know from their daily life, a drinking straw. When students begin to compare the shapes of the stem and the straw, they will presumably notice the straightness of both and perhaps also the fact that both are hollow. Noting these shared traits might lead them to recognise that both serve to transport a liquid. Pushing the comparison further, the students might even venture to surmise that the stem serves to convey nourishment to the flower, as a straw conveys nourishment to a human being. Seeking out differences, the students might note the difference in material, in the stem’s status as (no longer) alive vs. the (synthetic) straw’s status as never alive, and in the stem’s quality of being an integral part of a plant vs. the straw being an independent artefact being used as an implement, or again in the stem holding itself up vs. the straw being held by a person. Thus far, this type of reasoning has shown little use of formal binary logic. Such a logic may nevertheless come into play when students undertake an independent study of the stem either to confirm or reject the clues gleaned from analogical thinking. However, the reflective analogical thinking proper relies upon its own form of logic, which involves looking for potential analogues and then judging whether clear, strong and plentiful functional commonalities exist that make a given analogue useful and persuasive. The RCR problem Finally, we come to problem no. 5, the accident at the nuclear power plant. A ‘solution’ was already presented in chapter 1. So let us see whether another form of thought could have done better than RCR. This problem resembles the pendulum task in terms of considering the relative importance of three variables: what are the contributions of (a) the plant behaviour, (b) the actions of the operators, and (c) the operators’ group

96

The Theory of Relational and Contextual Reasoning

dynamics to the final outcome? Indeed, in principle one can formulate hypotheses regarding both problems no. 1 and no. 5 and can then test them until one arrives at a satisfactory answer. But the differences between the two problems immediately surface: (i) the study of the pendulum can be repeated ad infinitum using the same pendulum, which is not true in the case of the power plant; (ii) the pendulum variables can be varied one by one at will, whereas a simulation of the power plant accident can hardly be expected to reproduce the original event, let alone to permit variation of the variables step by step; (iii) physics alone is involved in the pendulum task, while the power plant task also involves human behaviour; (iv) in the pendulum task the three potential variables are separable, independent of each other, whereas in the plant accident at least some of the variables are inseparable. Thus, while some components of Piagetian operations clearly apply to problem no. 5 (such as constructing a hypothesis in possibility space), once again formal binary logic does not. Might one take a cue from problem no. 2, the story of the alcoholic and his wife? Both problems no. 2 and no. 5 involve human beings. The differences are that (a) problem 5’s outcome is already known, while one must speculate about the wife’s decision in problem 2; (b) the human beings working at the power plant are trained professionals doing their job, not a wife involved in a personal relationship with its problems; (c) problem 5’s outcome is co-determined by the plant’s behaviour, while problem 2 involves only human behaviour. Nevertheless, one might reasonably expect some components of cognitively complex thinking to play a role in the solution of problem no. 5. Might one consider the problem-solving process of no. 3, the wage negotiation, as being of any help? Both problems 3 and 5 have in common that (a) the situation at the end is different from the situation at the beginning; and (b) the ‘actors’ involved are somehow linked together; they react to what is happening before their eyes – and which may well not be anticipated. As to differences, (i) the power plant has less ‘freedom of manoeuvre’ than do problem 3’s negotiating parties, and (ii) the outcome of problem 5 is possibly even further from anyone’s expectations than is the outcome of problem 3’s negotiation. Nevertheless, one might expect to find some elements of dialectical thinking involved in finding a solution to problem 5. What might problem 4’s analogical thinking contribute to answering problem no. 5? Both problem 4’s plant and problem 5’s operating crew age and eventually ‘die’. Both need ‘nourishment’ and ‘maintenance care’. There are clear differences, however, including (a) non-living vs. living entities; (b) ‘behaviour’ that is narrowly bound by natural laws vs. behaviour that allows for some free choice; (c) non-emotional (plant)

Other Thought Forms

97

‘behaviour’ vs. potentially emotional (human) behaviour, and (d) morally neutral (plant) ‘behaviour’ vs. legally and morally responsible (human) behaviour, to name a few. Analogical considerations (‘what are commonalities and differences?’) are helpful when battling to answer the questions posed by problem no. 5, but of themselves will not fully answer the questions. Nevertheless, elements of analogical thinking might help one methodically when seeking answers to the questions raised by problem no. 5. Concluding remarks Closing this section, I hope to have demonstrated that a single thought form cannot solve optimally problems that have such different structures as nos. 1–5. Rather, problem 1 was best solved by Piagetian logicomathematical thinking, problem 2 by cognitively complex thought, problem 3 required dialectical thinking, problem 4 needed analogical thinking, and problem 5 was optimally solved by RCR. On the assumption that these results can be generalised, the application of RCR, the RCR heuristic is presented and multiply illustrated in Part II of this monograph. Summary of other thought forms and matching them to the problem at hand This chapter consists of two distinct, yet related main sections. Both have to do with Piagetian thinking, cognitively complex thinking, dialectical thinking, analogical thinking, and, in a different way, with RCR. In the first section, the characteristics of the first four forms of thought were compared to and contrasted with RCR. The main difference was found to be at the level of logic. In the second section five differing (uniquesolution) tasks/(ill-defined) problems were used to demonstrate matching the thought form to the problem structure.

Part II

Applications of RCR

101

Overview The natu re of the variou s chapters of Part II is diverse. Chapter 6 illu strates the application of RCR, first pu rely formally, and then tow ays of relating science and religion/theology, thereby highlighting, among other things, the symbolic meaning of the cover pictu re. Chapter 7 reports mainly empirical stu dies in the area of religion/theology. Chapter 8 endeavou rs to u nearth u se of RCR at earlier times invariou s su bject domains (Christian doctrine, painting, psychology, poetry, literatu re, physics). From chapter 9 onward, considerations become yet more specu lative. Each time, I attempt to apply RCR to a given issu e (in psychology, edu cation, nu clear energy, illegal u se of narcotics, rehabilitation of depressed areas), and then I discu ss the cu rrent state of the (pu blic) discu ssion of that issu e against the backgrou nd of RCR desiderata, and draw some conclu sions. Readers who are mainly after robu st evidence for RCR will have seen what I have to offer by the end of chapter 7; they may want to go from chapter 8 or even from chapter 7 immediately to chapter 12, the conclu sions. Nevertheless, I hope to have provided su fficient circu mstantial evidence and argu mentative plau sibility in chapters 8 to 11 to make their reading worthwhile.

6

Methodology

Method for applying RCR This chapter aims to demonstrate how RCR can be used to gain a deeper understanding of a (controversial) complex issue, the ‘explanandum’ which is subject to rivalling descriptions, explanations, models, theories, and/or interpretations. At first blush, applying RCR, the RCR search heuristic, may be somewhat hard to understand. I first present it formally, as a series of eight abstract steps. Readers who prefer to see immediately each step applied to a concrete case may want to turn to page 104 after reading the next paragraph. Before going into the actual procedure, a word needs to be said about the explanandum. Basically, there are two cases: (i) it is a given (e.g., the nature of light, taking remedial action after a nuclear accident) or (ii) it needs to be determined (e.g., when entering a ‘new’ field such as science and religion/theology – to be dealt with momentarily). In the latter case, the explanandum has to be ‘cut out’ such that it contains the ‘control centre(s)’ (Reich 1995b). For instance, if one wants to improve the understanding of blood circulation in vertebrae, it is not sufficient to consider the heart and the vascular system; one needs to include equally the relevant parts of the nervous system. Delimiting a coherent functional whole as explanandum implies that one can envisage with some confidence developing its overarching theory in the (distant) future, even if it is (by far) not clear at present what it will look like (e.g., mind and brain). Going through the sequence of the first seven steps represents what I call the RCR heuristic. Whereas applying RCR ‘tacitly’ may already be quite helpful, its full potential becomes fruitful when the RCR heuristic is applied systematically. Here are the complete eight steps (Reich 1990b, 1990d): (1) clarifying and defining, at least tentatively, the entity, the phenomenon, the event, the functionally coherent whole which constitutes the explanandum; 103

104

Applications of RCR

(2) listing all descriptions/explanations/models/theories/interpretations A, B, C . . . of the explanandum, even if they are considered incompatible or incommensurable by the ambient culture, possibly adding new ones, and dealing with any conflicts and contradictions arising (which may mean throwing out either A or B or C – it is possibly not a case for RCR) (mastering different logics and means-reflecting thought, cf. ch. 2, pp. 29–32, is particularly important for dealing with this step); (3) ascertaining that A, B, C . . . are genuinely coextensive, that they refer to the identical explanandum; (4) establishing the circumstances, the context, under which A, B, C . . . describe or explain particular aspects of the explanandum, and, if a genuine understanding does not come forth, reconsidering A (B, C . . . ) as approximation only; (5) discovering and describing any (including unexpected) links between the respective attributes/features of A, B, C . . . , as well as any coinherences (mutual pointers); (6) exploring the extent to which the (relative) explanatory power of A (B, C . . . ) depends on the current strength of B (A, C . . . ), etc.; (7) developing a complete synopsis or theory that explains all features of the explanandum under differing contextual conditions; (8) explaining any shifts in the meaning of the concepts needed to explain the explanandum, A, B, C . . . , and the new synopsis or theory. Demonstration of a particular search To clarify the heuristic indicated, the example chosen is ‘ways of relating science (A) and religion/theology (B)’. Before doing this step by step, proceeding from (1) to (8) above, some background knowledge is required. Background knowledge about science and religion/theology In the Middle Ages science and religion/theology were in consonance, for instance in the writings of Thomas Aquinas. Along with the growing success of inductive science (e.g., Galileo’s works), that consonance gradually weakened from the Renaissance onward, the split between science and theology widened in the eighteenth century with the advent of the philosophy of the Enlightenment (e.g., Voltaire’s writings), and became even more marked with the ascent of (first Comtean and then Logical) Positivism (e.g., Haeckel’s materialistic monism). Given the recently changed philosophy of knowledge discussed in chapter 3 (pp. 35–7), and the partial dissatisfaction with science and technology on account

Methodology

105

of its danger for the environment and potentially for human health, the intellectual climate concerned has somewhat ‘warmed up’ during the latter part of the twentieth century: here and there a dialogue between science and religion/theology has been rekindled (e.g., Byers 2000; Grenz 2000; John Templeton Foundation 1996; Numrich and Numrich 2000; Polkinghorne 1996; Richardson and Wildman 1996; Southgate, DeaneDrummond, and Murray 1999). Nevertheless, the clash between creation/intelligent design scientists on the one hand, and evolutionists (Darwinians) on the other, persists.1 Also to be noted: besides Christianity, other religions join the discussion (Stannard 2000), for instance Islam (e.g., Golshani 2001) and Buddhism (e.g., Ricard and Thuan 2000). The topic under discussion therefore remains of actual interest. Ian Barbour (1990, p. 3), professor of science, technology and society, in his Gifford lecture at Aberdeen in 1989, stated the following: 1

The following Internet news brief of Saturday, 13 May 2000 (copyright Chicago SunTimes Inc.), titled ‘Intelligent Design meets Congressional Designers’ (courtesy of Davis Wald at Caltech) illustrates the latter statement: On May 10th, a House Judiciary Committee hearing room was the site of a three-hour briefing on palaeontology, biology, and cosmology. Although presentations were at times quite technical, the speakers were not there to discuss the latest research in these fields. They were on Capitol Hill to promote intelligent design (ID) theory, to debunk Darwinian evolutionary theory, and to expose the negative social impact of Darwinism. Entitled ‘Scientific Evidence of Intelligent Design and its Implications for Public Policy and Education’, the briefing was sponsored by the Discovery Institute, a Seattle-based think tank (http://www.discovery.org), and its Center for the Renewal of Science and Culture. The afternoon briefing was preceded by a private luncheon in the US Capitol for Members of Congress and was followed by an evening reception. . . . Main speakers were biology professor Michael Behe, philosophy professor Stephen Meyer, Discovery Institute Fellow Nancy Pearcey, and law professor Philipp Johnson. Until now, the creation–evolution debate has primarily been active at the state and local level, but this event may represent the start of a new effort to involve Congress in efforts to oppose the teaching of evolution. Whether by chance or by design, the briefing took place as the Senate entered its second week of debate on overhauling federal K-12 education programs. Both houses are expected to work throughout the summer on reauthorization of the Elementary and Secondary Education Act. N.B. (K. H. R.): The Kansas Board of Education members explicitly cited Behe’s book on Intelligent Design as influencing their decision of August 1999 no longer to require examinations about Darwinian evolution. (A newly elected Board revised that decision in February 2001.) Intelligent design has been critiqued by the numerous authors presented on the (continuously updated) website http://www.world-of-dawkins.com/box/behe.htm. Competition and debates are good for science, and in the long run ‘nature’ will show who ‘is right’. My objection to ID ‘theory’ is not primarily to this conceptualisation per se, but to its misuse (although not necessarily by the authors) as an argument for preventing students forming their own judgement about the evolution of human beings and related issues (cf. Working Group on Teaching Evolution, [US] National Academy of Sciences 1998). And then there is a possible effect on the public school system if parents disagree to the point of turning to private schools or to home schooling.

106

Applications of RCR

The first major challenge to religion in an age of science is the success of the method of science. Science seems to provide the only reliable path to knowledge. Many people view science as objective, universal, rational, and based on solid observational evidence.2 Religion, by contrast, seems to be subjective, parochial, emotional, and based on traditions and authorities that disagree with each other.

Barbour groups the current options for viewing the relation between the two fields as (a) conflict, (b) independence, (c) dialogue, and (d) integration. I shall use these categories as a standard3 for a comparison with the result of applying RCR to the issue at hand. Irreconcilable conflict arises from the side of science through the claim that the ‘scientific method’ is the only reliable path to genuine knowledge (e.g., Jacques Monod, Carl Sagan),4 religion being declared as poetry or something similar, and from the side of religion especially by biblical literalists who hold that the overriding statements from scripture about the natural world (in particular its creation) are incompatible with the claims of modern science, notably those of evolutionists. In the Scopes trial in 2

3 4

However, as V. V. Raman remarked in an Internet discussion: in cosmogonic/cosmological matters, science uses some non-observables that cannot be related to observables in a welldefined way. ‘Worm-holes’, Hawking’s imaginary time, certain constructs in string and superstring theories, belong to this category. They have interest and relevance largely in their mathematical consistency, and relate only in very round-about and indirect ways to observable features of the world. While other classifications exist (Reich 1996c), Barbour’s is quoted and used widely. Persons holding such a view not infrequently also refuse to consider the possibility that there might be something to parapsychology, or that another logic than formal binary logic might be admissible (e.g., Breuer and Springer 2000 – see note 2 p. 117), let alone preferable in a particular case. There are indeed various ways to respond to ‘anomalous data’ (Chinn and Brewer 1992), some more fruitful than others. In the interest of authenticity, let me quote Paul Harrison, a religious naturalist (http://members.aol.com/Heraklit1/index.htm) from a recent Internet exchange: ‘In short, there is no limit to the religious assertions, however wild, that cannot be disproved. In fact it is much harder to think of ones that can be disproved. In view of this it seems preferable to adopt a “no benefit of the doubt” principle. If someone asserts that something exists or has happened, or is happening, or will happen, which is outside of all common or thoroughly documented human experience and of science, and which neither I nor anyone else on earth has any possible means of disproving, then it is reasonable to place the entire onus on those who make these assertions to prove them with the rigorous evidence that the religious naturalist (as well as atheists and humanists and other sceptically minded folk) will demand. Until they do so it is reasonable to assume that these assertions are false until proved correct.’ My question was, ‘Why should the other side accept those standards of “rigorous evidence”?’ Harrison answered, ‘The standards are those that the naturalist applies to her/his own reasoning before she/he will believe something that is prima facie incredible. We know that faith-based believers do not and mostly will not adopt these standards, but we would hope they would accept our right to apply them in deciding on our own beliefs. Of course, we might (and often do) believe the world might be a better place if everyone applied more rigorous standards of evidence before believing things, and we would argue this point in public, just as faith-based believers argue for faith.’ (Cf. note 6, p. 38; and note 6, p. 108 for a contrasting view.)

Methodology

107

1925 at Dayton, Tennessee (e.g., Larson 1997), the view of biblical literalists (written into a Tennessee law) was upheld (and John T. Scopes, a high school teacher, found guilty for not having observed that law by teaching the theory of evolution, was fined a sum of $100, and subsequently had to leave his teaching post), but the Tennessee law was disregarded after 1968 when the US Federal Supreme Court ruled that a similar Arkansas law was unconstitutional because contrary to the First and Fourteenth Amendments of the US Constitution. All the same, the battle for parallel teaching of both ‘scientific creationism’ and ‘(neo-)Darwinian evolution’ continued until the Supreme Court ruled again in 1987 that creationism is a religious idea that cannot be mandated in public education (but see note 1, p. 105). Barbour (1990, pp. 4–10) presents in extenso the arguments of both scientific materialism and biblical literalism. His conclusion is – and I fully agree – that each oversteps the proper boundaries of their discipline: science in the direction of natural philosophy and metaphysics, biblical literalism by making undue scientific claims (ibid., p. 3). Basically, conflict may mean that only A or only B is recognised as being true. Independence implies that science and religion/theology concern different domains (explaining objective, public, repeatable data vs. experiences of inner life such as guilt and forgiveness, meaninglessness and wholeness) and have different languages (scientific language used for prediction and control vs. religious language used for recommending a way of life, eliciting a set of attitudes – ibid., pp. 10–16). Taking an independence position avoids conflict, but (a potentially fruitful) dialogue is also ruled out. The other position is acknowledged (perhaps only grudgingly) as existing, but there is no real desire to know more about it. Dialogue implies that the methods in science and religion/theology have something in common (e.g., interpretation of the ‘data’5 and commitment 5

The methods used for collecting ‘data’ differ though (Reich 1995c, pp. 394–5). Scientific standards require (a) the complete and precise indication of the conditions under which an experiment/experience occurred, (b) willed repeatability, (c) testability by any (competent) third person, (d) generalisable significance. Theologians, apart from pointing out that such standards are inapplicable to contemplative, aesthetic, and similar experiences, explain that (a) to (d) are inappropriately maximised requirements as far as religious experiences are concerned (cf. Watts and Williams 1988, especially ch. 9). However, weaker forms are maintained. In particular, appropriate testimony of witnesses from both earlier and present times is considered epistemologically adequate as justification for the veridicality of ‘data’, even if not everybody has had or will in all likelihood ever have the witnesses’ experiences. Religious learning from experience is based less on the robustness of single facts and more on an ensemble of experiences, accumulated across situations and events with time. This poses the question of an ‘absolute’ third-person versus a ‘restricted’ third-person ontology on the one hand, and a third-person ontology vs. a first-person ontology on the other. How many witnesses and with which characteristics are needed to turn their witnessing into credible evidence? Among other things, the

108

Applications of RCR

of the practitioner), and there are some overlapping claims about reality (e.g., about the world’s origin) (ibid., pp. 16–23).6 Many of the more recent contacts between the protagonists of the two fields involved a dialogue position with the hope of benefiting theology, but also science.7

6

7

answer probably also depends on the knowledge domain under discussion (cf. note 6, p. 38). Argyris, Putnam, and McLain-Smith (1985, p. 238) indicate the following conditions for a fruitful dialogue (especially between participants tending to be defensive): ‘Participants must be able to retrieve largely tacit inferential processes; they must be able to deal openly with challenges and conflicting views; they must reveal information that might expose their own and others’ vulnerabilities; they must be able to recognise and acknowledge when they are wrong; and they must feel free to choose among competing views.’ To which I would only add at the end, ‘where appropriate’ (cf. notes 13, p. 24, and 16, p. 113), and also that participants should strive to understand why the other persons hold the views they hold, why they consider them justified. The debate reported by Breuer and Springer (2000) shows that non-observance of these conditions nevertheless can bring out the characteristics of the respective positions (and their weak points), but the example due to Paloutzian (2000) of (multiple) dialogues in which they are observed shows in addition how fruitful such dialogues can be. To quote Ric Barr, an Internet discussion partner who expressed rather well my own view: ‘Philosophically, one can see and be convinced of the reality of a broad, universal purpose which gives meaning and direction to life and maintain that this meaning will not and cannot be found within the limited constraints of knowing characterised by the scientific method. Personally, I use the image of a supernatural God (as a being who can hear, feel, respond) as an imaginative construct to provide a working model for my spirituality, knowing full well that my image of a “supernatural God” does not do justice to what I perceive intuitively as the profound ground of the universe, which I call God. This image/idea enables me to pray, to give praise and thanks, to affirm value and to understand suffering in a way not possible (at least for me) through an impersonal view of the universe. This view is compatible with science but certainly not provable by it. I do not like the term “supernatural” much because it implies a certain amount of arbitrariness in the way “God acts” . . . God is the same God everywhere or he/she is not God. I therefore like terms like Tillich’s “Being itself ” or Weiman’s “Source of Human Good” [or Brahman for Hindus]. My image of God is grounded and immanent with the natural world, though not fully contained by it (I am not a pantheist). I am not a creationist though I do believe that the universe is best explained by a mind-like intelligence from which all the possibilities which we see derive. I am not a deist because I believe that this source is still present, actively sustaining the universe in being. It is this reality which (most?) practising Christians, Muslims, Jews, Hinduists [and others] worship and in which we have our trust. I part company with many of my fellow Christians in that I have no issue what-so-ever with evolution, the big bang, and the full reliance of the human self on its physical/chemical/biological substrate.’ And I can also share V. V. Raman’s (forthcoming) Internet response: ‘When I see a fragrant flower and admire its beauty, when I pick up a shell from the sea shore and marvel at its pleasing symmetry, or when I read about the tardy tortoises on Galapagos, and am intrigued about how all these came to be, I tell my biologist friend about my wonderment, and she explains to me in fascinating detail how we can make sense out of the apparent biodiversity that is splashed all over the planet. When I see the diamond sparkle and the multicoloured rainbow arch the sky, when I see the silent stars up on high and observe dry sheets of plastic stick to my clothes, I recall the patterns and principles of physics from which emerge all the magnificence in the range and variety of perceived reality. I am grateful to science for these insights and enlightenment. But in all of this I also experience a mystery that is beyond my intellectual grasp. It is like the pleasures of poetry, the joy of music, and the ecstasy of meditative merger with the world at large.’

Methodology

109

There is now a wish to learn more about the ‘other side’, and perhaps also to contribute to its progress. Barbour’s integration comes in three versions. (i) In natural theology8 it is claimed that the existence of God can be inferred from the evidences of design in nature.9 (ii) A theology of nature draws on sources outside the sciences, but takes scientific findings into account. (iii) In a systematic synthesis (e.g., the process theology of Birch and Cobb), both science and religion/theology contribute to the development of an inclusive metaphysics (ibid., pp. 23–30). The danger with integration is that either scientific or religious ideas are distorted to fit a preconceived schema claimed to encompass all of reality. What can one say about Barbour’s classification (apart from admiring its impressive scope) when looking at it from a RCR perspective? First, the question arises as to the explanandum. If it is taken as knowledge and insights by humans situated in the bio-physical world and being part of human society with its history, social relations, and culture, a first observation is the ‘static’, time-invariant nature of almost all of Barbour’s categories and sub-categories (cf. Reich 1996c). (The explicit exception is process theology in the category integration.) Set-theoretical considerations of categories are therefore applicable. Referring to Fig. 5.1 (p. 79), one sees that conflict is represented by diagram no. 2 (where B may become 8

9

V. V. Raman provides a sense of natural theology in an Internet posting ‘When there is a sudden spewing of matter or passion, of disease or destruction, there is an eruption. Volcanoes erupt, as do anger and fury and an epidemic of plague. When something appears, and retains its entity in form and substance, there is emergence. A flower emerges and so does a sonnet or a work of art. But when what emerges is governed by law and principle, and it evolves too, we have creation: the launching of something that never existed before and that does not remain the same. From this perspective, and in this terminology, the big bang was not a mere eruption, nor the universe a mere emergence: the cosmos was created. What is created has an existence of its own. More importantly, others things appear from it: it too creates. The theologian Phil Hefner (1993) speaks of human beings as co-creators, for we create: ideas and things, values and works of art, and much more. ‘I would like to extend this insight: We may look upon ourselves as conscious co-creators. For it would seem that there are unconscious and semiconscious co-creators too. The matter and energy that were created from the big bang were unconscious co-creators, for they led to atoms and molecules, to elements and compounds, to planets and stars: each a created entity in its own right. And when the self-replicating macro-molecules of life arose, another level of co-creation arose: for evolution is a creative process too. This biological evolution is different from the unconscious formation of atoms and stars, and it may be described as semiconscious co-creation, for there is a fine difference between crystal growth and cell-division. Finally, with the onset of mind, creation leaps, as it were, to a higher level: the level of self-awareness. The creation from now on is conscious, and what is created is not just machines and bridges, but ideas and ideals, values and morals. This constitutes what may well be called conscious co-creation. Such a perspective can be part of natural theology.’ It is not quite clear why natural theology thus defined is classed as integration, given that only one source is indicated.

110

Applications of RCR

unimportant, even vanishing in the eyes of A and vice versa), independence by no. 4, dialogue by no. 3, and integration by no. 1. This being all the logical possibilities according to formal binary logic, Barbour’s classification is exhaustive within that framework. A second (arguable) observation is that at first blush conflict could be interpreted as an expression of (the descriptive, not normative) RCR level I, independence as that of level II, dialogue as implying RCR levels III or IV, and integration as an expression of level V. On second thoughts, the step from level II to level III is not just developmental; it may well require a (conscious) decision and commitment. It is difficult in today’s world to ignore totally either science or religion and their workings, and therefore one can understand the reason for making the mental step from conflict to independence. However, even somebody capable of arguing intellectually at level V of RCR may still prefer to stay with independence for his or her own good reasons.10 Further observations will emerge when applying RCR to science and religion/theology. 10

In case a reader finds this second observation absurd, consider Kohlberg’s (1984) stages of moral cognition. There figures not simply one single argumentation, justification and motivation with respect to acting morally but a variety, depending especially on the levels of ‘ego-centredness’ of social cognition: a person at stage 1 of moral judgement is egoistically motivated, which involves arguments of external rewards and punishments. Stage 2 implies an enlightened egoism. Stage 3 is based on being accepted by one’s immediate social circle as moral reason and motivational factor. Stage 4 enlarges that circle to one’s own society. Stage 5 is characterised by a philosophy favouring the greatest good for the greatest number (social contract). Now, I am aware that moral cognition is not the same as assessing the relation of science and religion/theology. No society can survive without some moral rules – they are a necessity. In contrast, assessing the relation of science and religion/theology in a way is a luxury indulged in by a small number of enthusiasts (but potentially a useful one, given the large numbers of religious believers the world over, and the partly unhappy history of the relation of science and religion/theology). Also, I do not ignore the criticism levelled against Kohlberg’s stages (mostly not really pertinent here). However, moral judgement and assessing science-and-religion/theology both involve cognition, and both deal with the relation between two entities (the individual and society in the case of morality). As I attempt to show throughout this monograph, our judgements and insights are not produced automatically from sense data and other input, but (partly) (re)constructed according to our level of cognitive development (which continues in adulthood). The parallelism between ‘science-and-religion/theology’ and ‘moral judgement’ goes even further. Just as persons refuse to proceed from independence to dialogue for their own reasons, not all are willing to make the step from moral stage 2 (an egoistic position) to stage 3 (acknowledging a person’s social embeddedness and hence justified arguments for a certain amount of ‘solidarity’), in particular not Ayn Rand (1964) in her Virtue of selfishness. Kohlberg (1984, pp. 429–31) encountered a number of students who, at a particular point in their life, rejected the demands of society and valued their unbridled self-fulfilment and self-development higher. One college sophomore said, ‘[In high school] I was trying to please the norms of society, and in essence, conforming to the prevailing thought about moral right. I was concerned about other people and society in general when I was younger. Now I think more of a moral responsibility to oneself. Self-concern takes precedence over morals’ (ibid., p. 444). A well-developed moral cognition was not in doubt, these students having argued

Methodology

111

Applying RCR to science and religion/theology As it will turn out, this is not an ideal case for demonstrating RCR at work per se. For one, it is too vast a case to be dealt with in any depth in the present context. However, the overriding advantage is that Barbour’s ‘standard’ exists to which the result of applying RCR can be compared. Also, the result can be checked against the results of the corresponding empirical study presented in chapter 7. We proceed according to the eight formal steps delineated above: (1) (clarifying and defining, at least tentatively, the entity, the phenomenon, the event, the functionally coherent whole which constitutes the explanandum). The general procedure is to assume that the explanandum is suitable for study by RCR, and then to find out whether that is so by going through the next six steps. To begin with, ways of relating science and religion/theology does not appear to be a proper explanandum, because it is not a functionally coherent whole (Reich 1995b). There are quite a few areas where science is of little or no concern to theology (e.g., inorganic chemistry) and vice versa; in other issues science supports religion/theology (e.g., by demonstrating the beneficial effects of religious life11 – Gorsuch 1995); in yet others it weakens it (e.g., heliocentrism versus medieval Christian teaching). Also, historically, the relation has changed and still changes, as indicated earlier. Furthermore, there is not just one theology, but a diversity, even in Christianity (Fulljames and Stolberg 2000). As a consequence, the explanandum has to be more restricted, better focused. The proposal is to concentrate on ‘Understanding the origin of the universe, its changes until today, and the resulting lessons for leading a human life’ (which is still simplifying considerably the actual state of affairs, e.g., Reich 1995c).

11

previously at stage 4, but a corresponding commitment was lacking. In other words, cognitive competence explains much, but not all. As further evidence for such a view, notice the results of the relevant empirical study on science-and-religion/theology reported in chapter 7 (pp. 126–9). From the present perspective, one explanation could be that (some of ) the persons under discussion have reached the stage of reflecting about mental tools (discussed in chapter 1, pp. 29–32), and from there question the validity of (meta)ethical principles. Just to keep the record complete: later in life those former ‘self-concerned’ students returned to a stage 4 or even a stage 5 argumentation (ibid., pp. 430, 445). To be fair, cases of unhealthy religions/sects need to be recognised too, e.g., collective suicides/murders of persons belonging to new religious movements such as suffered by the 909 members of the People’s Temple led by the Reverend Jim Jones (Guyana, November 1978), by the several dozens of the members of the Order of the Solar Temple led by Joseph Di Mambro and Luc Jouret (Canada, France, Switzerland, 1994/1995), by the 39 members of the Heaven’s Gate led by Marshall Applewhite (California, March 1997), and finally by hundreds of persons adhering to the Movement for the Restoration of the Ten Commandments of God (the Doomsday sect – Uganda, March/ April 2000).

112

Applications of RCR

(2) (listing all descriptions/explanations/models/theories/interpretations A, B, C . . . of the explanandum). The following convention will be used throughout this section (cf. Barbour 1990, p. 3): A = science (empirical study of the order and patterns of nature); B = theology12 (critical reflection on the life and beliefs of the religious community); C = philosophy (analysis of the characteristics of inquiry and knowledge as well as analysis of the most general characteristics of reality). Regarding the explanandum under discussion, A involves cosmogony and cosmology, neo-Darwinian evolution, and biological/sociological anthropology; B refers to the theology of creation (in Christian terms, or to the equivalent in other religions), to that of divine providence, and to theological anthropology; C involves an analysis of methods admitted in A and B for gathering, analysing, and interpreting evidence, a procedure for dealing with (perceived) transcendent reality (e.g., Reich 2000b), and philosophical anthropology. (3) (ascertaining that A, B, C . . . are genuinely coextensive, that they refer to the identical explanandum). This is a much larger task than can be tackled here. At first glance one might say that individually A, B, C will not have something to say to each and every aspect of the explanandum. That explanandum is clearly a weaker functional whole than, for instance, a single human being. However, that is probably not strong enough a reason to stop the application of RCR at this point. (4) (establishing the circumstances, the context, under which A, B, C . . . describe or explain particular aspects of the explanandum). Again, this is a lengthy study in itself. The suspicion is that A will provide the most relevant explanation of the actual changes of the universe and what it contains from the big bang until today,13 B on the lessons to be drawn, 12

13

In which cases is it more appropriate to discuss science and religion, and in which science and theology? There is no single, consensual answer. If religion is judged to be motivational, experiential, and prescriptive, then (scientific) anthropology, psychology and sociology are ‘immediate’ discussion partners. If theology is taken as descriptive, explanatory, interpretative, then it is a more appropriate discussion partner for cosmology, biology, and medicine than religion. However, it is also true that for science it is easier to get a clear idea of the origin and the functioning of religion (authority, ritual, speculation, tradition, God’s sovereignty and grace, mystery – Smith 1965, pp. 101–4) than of theology. And, theology is not recognised by all religions as even existing. In any case, the ‘best’ of religion/theology should be introduced into the science-religion/theology debate, not some caricature (as unfortunately still happens). David R. Burwasser wrote in an Internet discussion (cf. Reich 2000b): ‘Almost by definition, science should not be interested in anything more than its methodology and the consistency of its results [cf. note 6, p. 38]. It ceases to be science if it becomes captive to any social agenda, even a democratic or spiritual agenda. However, scientists need to be concerned for human impact, but as the human beings under the lab coats. The two must be kept separate just to preserve the integrity of both.’

Methodology

113

in particular regarding the inner life of humans14 (but will possibly also contribute to the presuppositions of A and C), and C on the validity of the ‘truth’ claims of A and B, on the limits of their ‘legitimacy’, and possibly on anthropology. (5) (discovering and describing any (including unexpected) links between the respective attributes/features of A, B, C . . . , as well as any coinherences). A furnishes B with detailed knowledge about the wonders of the universe and all it contains so that B can go on from there. Historically, B has supplied A with a world view which made research possible and attractive; if needed, B reminds A of its responsibility for the environment and human welfare. C furnishes to A and B a base for a rational discourse using consensual categories and procedures. (6) (assessing the extent to which the (relative) explanatory power of A (B, C . . . ) depends on the current strength of B (A, C . . . ), etc.). In the present case, the clearest case is probably the dependence of C, but also of B, on A (e.g., brain research). If B weakens, the lessons drawn by A (and perhaps by C) may be too one-sided.15 If C were to drop out, the quality of the dialogue might suffer. (7) (developing a complete synopsis or theory that explains all features of the explanandum under differing contextual conditions). Given the difficulties evoked all along in this section, that task will take time.16 RCR proceeds by keeping A, B, C distinct, and iterating the sequel from (1) to (7), feeding in each time any new insight gained. (8) (explaining any shifts in the meaning of the concepts needed to explain the reference, A, B, C . . . , and the new synopsis or theory). In the present case there is no obvious candidate for meeting point (8). One 14

15

16

Religions/theologies may be classed as collectively evolved encyclopaedias of human characteristics, actions, and events. Believers turn to the accumulated wisdom of a religious symbol system when they feel the need to get in touch with their deepest self (cf. Hefner 1996). As Carol Albright has observed, ‘choices persons make about the answers to “unprovable” questions may depend more than one would like to admit on unconscious inclinations. The need to be independent or dependent, related or detached, hopeful or resigned, affect (lay) theology. Nevertheless, cognitive checks and balances are needed too, or at least by some persons’ (from an Internet posting). Like other human enterprises, science has a tendency to establish, and even increase its influence and power, sometimes in not very ethical ways (e.g., Toulouse 1998). While others may and do oppose that tendency, religious believers/theologians carry their oppositional share, in particular when it comes to protecting the environment, biodiversity, and so on. The 200-year struggle to get to a satisfactory theory of light, incorporating both the light-as-corpuscle model and the light-as-wave model, was already evoked in note 13 (p. 24). Or take Darwin’s ‘theory’ of evolution. At the time neither Mendel’s laws nor the role of DNA was known, let alone the detailed working of mutations. Hence, a really informed debate about the respective roles of chance and necessity in evolution had to wait for another century.

114

Applications of RCR

could possibly argue for writing science-and-theology(-and-philosophy) in order to emphasise the links found. To sum up the result: Science-and-theology(-and-philosophy) ‘jointly’ contribute to Understanding the origin of the universe, its changes until today, and the resulting lessons for leading a human life. While each discipline contributes something to most issues, science provides the most relevant explanation of the actual changes of the universe and what it contains from the big bang until today, theology (potentially) contributes most to the lessons to be drawn, in particular regarding the inner life of humans (but possibly contributes also to the presuppositions of science and philosophy), and philosophy is most knowledgeable about the validity of the ‘truth’ claims of science and philosophy, on the limits of their ‘legitimacy’, and possibly on anthropology. While clearly distinct, on account of certain links each discipline can benefit from the others for flourishing optimally. Actually, the foregoing result is more of a programme for further work than a complete achievement. However, it is sufficiently different from the above ‘standard’ due to Barbour (1990) to warrant a discussion. What are the differences? (a) The focus is narrowed. As already indicated, in an RCR approach, the entire field is carved up into appropriate domains, and the exercise repeated until the entire field is covered. It is not clear whether an overall summary could then be made in the present case, but if so, it would be more differentiated than the ‘standard’ classifications. (b) Applying RCR results in a single (idealised) category, not four. From a developmental point of view, it is assumed to be a description of a stage which is likely to be reached more widely in the future, given the existence of exemplars (e.g., John Templeton Foundation 1996; Richardson and Wildman 1996; Southgate, Deane-Drummond, and Murray 1999) – but see the caveat of the second part of note 4, p. 106. (c) The context dependence of the explanatory power of science, theology, and philosophy is emphasised over against a universal contextindependent evaluation of their respective contributions/explanatory power. (d) The links between science, theology and philosophy are made explicit. (e) Overall, an attitude of mutual collaboration is fostered, given that neither side can prove the other side ‘wrong’ as far as discipline-specific, (peer-reviewed) established findings are concerned. If examples of the latter are wanted, I would name biologist Kenneth F. Miller’s (1999) book, Finding Darwin’s God, and theologian John F. Haught’s (1999) book, God after Darwin (without necessarily endorsing

Methodology

115

all views expressed by these authors). While belonging to different disciplines, both authors come to similar conclusions: (neo-)Darwinian evolution does not exclude religious beliefs; on the contrary, both views can enrich each other (cf. Ruse 2000). In case this exercise has been too difficult to follow or is not convincing, more (partial) exercises applying RCR are coming up in chapters 7, 9, 10, and 11.

7

Religion

The four issues of this chapter are: (1) religion and the nature of human beings, (2) understanding Christian doctrines, (3) the co-ordination of religious and scientific world views, and (4) religious development; this each time from an RCR perspective. The first section applies RCR afresh to a particular domain of the science–religion debate. It is therefore comparable to the exercise in chapter 6, but focused yet more narrowly. Religion and the nature of human beings Religion and its truth claims Across the ages there have been human groups on our planet without agriculture, without the wheel, without writing, without formal laws, but ‘neither history nor anthropology knows of societies from which religion has been totally absent’ (Rappaport quoted by Burkert 1996/1998, p. 1).1 Palaeolithic sacrificial rituals existed more than 20,000 years ago (ibid., p. 39) and traces of ritual burials are even older. This is not the place to discuss the origins of religion (e.g., Burkert 1996/1998), nor its diversity within and across various cultures (e.g., Wulff 1997), its psychological multidimensionality (e.g., Hood, Spilka, Hunsberger, and Gorsuch 1996, pp. 8–12), or its possible role for human development (e.g., Peck 1997, especially pp. 241–306). The assessment of a transcendent religious ‘reality’ is a great cognitive challenge (in particular when only formal binary logic is admitted 1

In a seven-page summary, Derek Stanesby (2000) reviews and comments/critiques the meaning of the term ‘God’ and the various arguments for God’s existence (the ontological and the cosmological arguments, those from design and from experience – cf. Vardy 1990). Neither the existence of God nor God’s non-existence can be proven irrefutably. However, ‘the human need to search for some meaning and purpose in life is unquenchable, and to the extent that we are rational creatures, then we will endeavour to support our beliefs with good reason’ (Stanesby 2000, p. 7). It is also noted that in the Western word about half of the people believe in the existence of a ‘higher being’, and a higher percentage in developing countries (e.g., Argyle 2000, chapters 14, 15).

116

Religion

117

for a reality check2 ) – cf. Vardy 1990/1997.3 Among others, two authors of encompassing works have set themselves that task: the Christian theologian Hans Kung ¨ ([1978], 1994) with his Does God exist? and the self-declared atheist Michael Shermer (1999) with his How we believe. The scholarship of both authors is impressive, their enterprises illustrate the vastness of the field, and the difficulty of coming to convincing conclusions (cf. Reich 2000b). On the last page Kung ¨ concludes his critical evaluation of the works of Descartes, Pascal, and Hegel as well as those of Feuerbach, Marx, Freud, and Nietzsche; this after further considerations in the fields of theology and religious studies. Kung ¨ states that the evidence allows one on rational grounds to answer the question ‘Does God exist?’ in the affirmative. And according to Kung, ¨ that constitutes a solid foundation of an enriching religious life. In contrast, Shermer (1999, p. 236) relates that for him, ‘The conjunction of losing my religion, finding science, and discovering glorious contingency was remarkably empowering and liberating. It gave me a sense of joy and freedom . . . I was free to live my life to the fullest.’ Can RCR help to clarify the dissonance between the views expressed by Kung ¨ and by Shermer? 2

3

In a debate in Germany (Breuer and Springer 2000) between the self-declared atheistic philosopher of science Bernulf Kanitscheider and the theistic philosopher of religion Ulrich Luke, ¨ Kanitscheider challenged the latter as follows (ibid., p. 85): ‘Would you agree that binary Aristotelian logic is applicable to basic theological statements and Christian doctrines?’ Luke ¨ answered (and I agree), ‘The reach of our [Aristotelian] logic, which is always finite and mediated by language, is insufficient to capture a comprehensive knowledge about God. Using such an approach, one can perhaps become an atheist [gottlos werden], but not get rid of the question of God [Gott los werden].’ Kanitscheider’s answer, ‘Escape into mystery does not solve the problems of logic.’ To which I would respond with, ‘But why should there be only a single logic applicable everywhere and all the time?’ This state of affairs constitutes a particular difficulty for the psychology of religion, which may lead one to work in mixed teams of theists and atheists (Reich 2000b – see Breuer and Springer 2000, for a possible, though not optimal result). Michael Argyle (2000, pp. 239– 40), after working for more than forty years in the field, sees it as follows: ‘The traditional solution to the science vs. religion problem was to say that science deals with the material world and religion with the subjective world; but psychology claims to deal with the inner world too . . . What seems to be wrong . . . is a failure to take seriously the experience of those concerned, to recognize the power of metaphors and symbolic behaviour, which are felt to express some kind of truth . . . In contrast, psychologists have not tried to “explain” mathematics, which is recognized as having an independent existence . . . Worship and sacrifices are pervasive aspects of religious behaviour throughout the ages . . . ; psychology has had no success in explaining them . . . beliefs about religion are unlike beliefs about the physical world: they are not verifiable in the same way, they are couched in symbols and myths, they represent commitment and relationship, and they need to be measured and studied in a different way from other kinds of belief.’ Hartmut Beile (1999, p. 115), a believing Christian, reported that his thesis work on religious emotions benefited from the fact that the supervisor was a self-professed atheist.

118

Applications of RCR

Applying RCR RCR was already applied to a co-ordination of science and religion/ theology (ch. 6, pp. 111–14), and that co-ordination will be taken up again in the third section below. In this section, the central issue is a particular aspect of anthropology as viewed by (Christian) religious beliefs and by neurobiology (e.g., Shermer 1999, pp. 65–9). Warren Brown and Malcolm Jeeves (1999, p. 139) put that issue as follows: Proposition 1: Humans are physical beings who also have non-material souls. It is through our souls that we experience and relate to God. Proposition 2: Humans are neurobiological beings whose mind (also soul, religious experience, etc.) can, in theory, be exhaustively explained by neurochemistry, and ultimately by physics.

Clearly, these propositions, representing traditional Christian theology (proposition 1) and (reductive) scientific physicalism (proposition 2) are dissonant. In particular, (1) intimates free will, and the possibility of eternal life, (2) holds that behaviour is determined (exclusively) by the laws of biology, chemistry, and physics. Applying the RCR heuristic of chapter 6 (pp. 103–4) to that dissonance, the first task is to determine the explanandum. It shall be: The nature of human beings and their capacity to relate to a perceived transcendent (God for the adherents to a monotheistic religion). As to step two (listing all descriptions, explanations etc.), we already have proposition 1 above (= A) and proposition 2 (= B). I add a third (= C): Humans are naked animals who share capacities with other animals, in particular with their nearest primate relatives. However, in humans some of these capacities are more enhanced, for instance ‘language’, a ‘theory of mind’ (hypothesising what is going on in another person’s mind), ‘episodic memory’, ‘conscious top-down agency’ (conscious mental control of behaviour), ‘future orientation’ (mental scenarios of future implications of behaviour and events), and ‘emotional regulation’ (cf. Brown and Jeeves 1999, pp. 144–5). The enhanced capacities have enabled human culture to evolve; it co-determines human behaviour – as does the proximate human group.

Simplifying, A emphasises the spiritual aspect of human beings4 , B their biological aspect, and C the social aspects. 4

Close to the time of her death, adolescent Anne Frank wrote to herself: ‘I have found that there is always some beauty in life – in nature, sunshine, freedom, in yourself; these can all help you. Look at these things, then you find yourself again, and God, and then you regain your balance’ (Frank 1993, p. 14). Marsha Sinnetar (2000) quotes this and many other testimonials of spirituality and spiritual intelligence (cf. Paloutzian 2000).

Religion

119

As the next step, (3), the question has to be answered whether A, B, C are coextensive. If the extension is given by C, then the coextensionality of A and B with C seems debatable. C considers four explanatory levels: (i) the biological level, (ii) the individual psychological level, (iii) the social (group) level, (iv) the cultural (societal) level. A deals primarily with level (ii). It does not explicitly rule out the other levels (except the biological for the soul), but they remain hazy at best. B deals exclusively with the biological level; the other levels are declared epiphenomena (by implication). From this short comparison, a few questions arise, which would be put to the protagonists: to A, How does the soul communicate with the memory – and with the body? To B, How do social and cultural influences, and in particular those which go against the biological grain (e.g., devotion to visions, ideals) get into that neurochemical system? To C, Exactly at which level(s) are religion and religious experience located, and in particular the perceived transcendence? To do things properly, these questions should be answered before proceeding further. That cannot be done here, but we shall nevertheless continue. As step 4, we look for circumstances, the context, under which A, B, C explain best particular aspects of the explanandum. A opens the door to a spiritual life, possibly lasting beyond the death of the body. B makes a rudimentary ‘religion’ of animals understandable such as the ‘religious’ devotion of dogs to their masters, the sun ‘worship’ of baboons, the ‘ritual dances’ of anthropoids, and further animal ritualised behaviour (Wulff 1997, pp. 146–55). C lets one get a sense of the multivariate nature of religion and religious experience. Next (step 5), we look for links between respective attributes/features of A, B, and C. Even if A and B seem to be incomplete according to the foregoing considerations, there should nevertheless be links between one or more of their attributes/features and C, for instance concerning the psychosomatic nature of human beings. Step 6 concerns the relative explanatory power of A as a function of (B) and vice versa, and so on. At this stage of the debate with the protagonists of (A) and (B), that question cannot be answered satisfactorily because (A) and (B) practically exclude each other as to explanatory claims. (C) could benefit from (B) regarding any biological roots of perceived transcendence. Next, the penultimate step 7 involves a synopsis. To my mind that critical summing up has to be based on (C). The most difficult part is presumably to explain the relations/connections between the (neuro-) biological level and the individual psychological level, in particular as far as perceived transcendence is concerned (e.g., Zygon 1999). One would

120

Applications of RCR

have to go into the ergotropic and the trophotropic forms of arousal (Wulff 1997, pp. 109–19), and into the latest results of brain research (e.g., Ashbrook and Albright 1997; d’Aquili and Newberg 1998, 1999; Persinger 1993; Ramachandran and Blakeslee 1998; Saver and Rabin 1997), being aware though of the speculative nature of some of this writing. Presumably, part of the arguments will centre on the issue as to whether a first-person ontology of the mental is acceptable, or whether only a third-person ontology makes the grade (chapter 3, note 6, p. 38). As regards the important relations/connections between the individual psychological level, the social level, and the cultural level, much material exists, given the longer history of the corresponding research. Thus at least a descriptive synopsis should be possible. Finally, (8), any shift in meaning of the terms used should be explained. That concerns primarily the soul. Brown and Jeeves (1999, p. 145) find it attractive to view it as emerging from the experience of personal relatedness. I would say ‘relatedness to other people, to nature, and to what is transcendent as perceived by the person concerned’. Such relatedness can be considered a core characteristics of spirituality (Reich, Oser, and Scarlett 1999). Once more, RCR’s contribution was to determine the explanandum in terms of a functionally coherent whole, to add a further description/ explanation, to uncover missing information, to discover links between A, B, C, and to thematise the context dependence of the respective explanatory power (cf. Sharpe 2000, for an example of seeing the RCR heuristic (tacitly) at work). Of course, the real work remains to be done by the experts in the various disciplines. Understanding religious doctrines Religious doctrines are not infrequently held to be irrational, to be understandable only to believers, if at all. Examples would be (a) the human and divine nature of Jesus Christ, (b) the three personae of the single Trinitarian God, (c) God’s interaction with a world governed by natural laws, etc. Theoretical and empirical research has shown that RCR can contribute to overcoming some of these cognitive hurdles (Reich 1989, 1990b, 1991, 1994a, 1996b), in particular because it is not tied down to the limitations of formal binary logic (cf. Kaiser 1996). Hence, it seems worthwhile, on the one hand, to stimulate such thinking, in particular in the context of religious education (Reich 1996a – cf. ch. 10, p. 162 below) and, on the other hand, to review here briefly the work on (a) and (b) above.

Religion

A

121

A

A

B

B

(a) Two or one? (b) A, B seen as parts

A

B

(c) Focus on relationship between parts (d) Plus relationship

A

B

between each part and the whole

Figure 7.1 Complexification and conceptual changes when moving from one to two. (Source: Reich 1994a, p. 116)

The two natures of Jesus Christ The historical part, the proceedings at the Council of Chalcedon (451 AD), are dealt with in the next chapter. As a result of those proceedings, the assembled Fathers declared that our Lord Jesus Christ is . . . truly God and truly man . . . made known in two natures [which exist] without confusion, without change, without division, without separation.

On account of the exemplary value of the issue of the two natures, some background knowledge will be provided first. One essential basic difficulty concerns the move from one to two. That move does not represent a linear addition of more of the same, but means a profound complexification and conceptual change, and correspondingly requires a different logic and epistemological approach. To understand what is involved in going from one to two, we take the cue from John Puddefoot (1992). The one is assumed to be primitive, indivisible, and complete in itself. Something new occurs as we move to two, a plurality of scales: Are we now dealing with a doublet of ones or with a unity, ‘the unity of two’? Which one is it to be? The answer to this ‘simple’ question (and where ‘one’ should not be identified with plain concrete things) depends on one’s stage of complex reasoning. An unsophisticated reasoning will presumably settle for one of these alternatives, often without deeper reflection. A more complex reasoning will view the situation as schematically depicted in Fig. 7.1.

122

Applications of RCR

Going from one to two thus introduces a plurality of languages, and possibly a scheme of ‘levels’ of reality: The movement from one to two is no mere aggregation, as if complexity were only the sum of simples; it is a movement from indivisibility to divisibility and structure: the movement from one to two alters the one, for now the one is not only related to itself, in itself, and even for itself; it is also to, in, and for the other: The other in the sense of the other part, and the other in the sense of the whole of which the one is itself part. In moving from one to two the being of one is no longer complete in itself; its totality involves the other. (Puddefoot 1992, p. 16)

The empirical study In the laboratory study (Reich 1994a), twenty-eight volunteers (from the sample of thirty-two respondents of the pilot study 4), aged 13 to 68 years, both religious believers and nonbelievers, about half male and half female, were interviewed individually. These participants were not representative, as explained in chapter 4 (pp. 47–8). The shortened Chalcedonian Definition was presented in written form, reading (in translation): The Fathers who met in 451 at the Council of Chalcedon declared notably that ‘Our Lord Jesus Christ is truly God and truly man . . . made known in two natures [which exist] without confusion, without change, without division, without separation.’ What do you think about this Chalcedonian Definition?

To assess their level of RCR, all respondents furthermore were interviewed about the three standard themes ( pianist, accident in a nuclear power station, and humans, the mind–body problem), presented and discussed in chapter 4. The results were as follows. About a third of the respondents (group 1) could not make sense of the doctrine. About another third (group 2) said that on first thoughts it all looked confusing, but on second thought it made some sense, and they explained how. The remaining respondents (group 3) explained why that particular wording of the doctrine was the most appropriate if not the only possible way to explain the state of affairs concerned. The inter-rater concordance concerning group assessment was 90 per cent. The individual scores of RCR and understandability of the doctrine are shown in Table 7.1. All participants who responded below level IV of RCR belong to group 1. (This statement is not reversible though; in that group were also respondents who were capable of higher level reasoning, yet they lacked religious knowledge or had reserves about the two natures.) Respondents in group 2 reasoned at least at level IV, those of group 3 at least at level

Religion

123

Table 7.1 Frequencies of individual scores of RCR levels and intelligibility judgement of the Chalcedonian Definition. Level III(IV) is somewhat above level III; IV(III) is somewhat below IV, etc. Ntot = 28; age 13–68 years; mean 33 years, 8.4 months; SD 15 years, 2.3 months. Source: Reich 1994a, p. 121. Is the Chalcedonian Definition rationally understandable?

Np

Level of RCR

Group 1 (‘no’)

1 1 8 12 6

III(IV) IV(III) IV IV(V) V(IV)

1 1 2 4 1

6 4 1

4 4

9

11

8

28

Group 2 (‘partly’)

Group 3 (‘yes’)

IV (V), which is somewhat above level IV. Kendall’s rank correlation coefficient has the moderate but significant value rK = .40, p < .02.5 To get a sense of what respondents in group 3 said, here are two excerpts (Reich 1994a, p. 121). Ariane (23 years, 11 months; henceforth 23; 11) explained: You can’t judge this by the usual rationality nor by arguing from personal experience. Jesus has to be human, otherwise he could not suffer [and wouldn’t be close to us]. And he has got to be God, otherwise atonement wouldn’t work. And because both [natures] have to come together in a single person, you get this helplessness with the usual notions. . . . [The difficulties] stem from the habit of imagining things on a scientific or human base.

Rainer (28; 9) said: Without separation – that makes sense if one can say that it belongs together naturally, it is not thinkable that the one exists without the other. But now, without confusion – that is again what is without separation – that means the two natures can all the same be analysed separately . . . Now it makes sense to me, without confusion, without separation.

The three personae and the single Trinitarian God The Trinitarian Godhead, the Holy Trinity (three personae yet one God) involves a further complexification (Reich, 1994a): As there are now three 5

A post hoc Kruskal-Wallis analysis of variance (H test) further supported the findings of the correlation computation: the mean RCR ranks of the members of groups 1 to 3 differed (chi2 = 7.87, df = 2, p = .02). The corresponding U-Test (Mann-Whitney) showed that the significant differences were between group 1 and group 3 (U = 14, p = .02), and between group 2 and group 3 (U = 14, p < .01).

124

Applications of RCR

‘entities’, different groupings of the type ‘two versus one’ can be imagined. In fact, for more than a thousand years the Roman Catholic Church and the Eastern Orthodox Churches cannot agree on the procession of the Holy Spirit. For Catholics (and Anglicans as well as many Protestants) the so-called Nicene Creed (adopted by the Ecumenical Council of Constantinople, AD 381) reads in part, ‘The Holy spirit . . . who proceedeth from the Father and from the Son ( filioque) . . . ’ In contrast, the Orthodox Church did not accept the addition of filioque to the original creed. As a result, for the Catholics, in a way the Son is ‘closer’ to the Father than the Spirit, and for the Orthodox, the Son is ‘closer’ to the Spirit than the Father. However, these theological disputes were not the object of an empirical study. Rather, the aim was similar to that concerning the Chalcedonian Definition, namely, to find out whether respondents arguing at a higher RCR level about the three standard nonreligious problems would understand the doctrine of the Trinity better than respondents who argued at a lower RCR level. The procedure was the same as in the case of the Chalcedonian Definition (except that two respondents of that study did not participate and two new ones joined). The interview text submitted to the respondents reads: Christian theology teaches the Doctrine of the Holy Trinity: the Father, the Son, the Holy Spirit. What is your opinion about this doctrine?

The result of this study was similar to that of Table 7.1 in that again three groups emerged, and the understanding of the doctrine correlated with RCR (sub-)levels (Table 7.2). Kendall’s rank correlation coefficient has the moderately high value rK = .56, p = .001.6 Here again are some interview excerpts, first of a group 2 respondent (Reich 1994a, p. 123). Peter (17; 0) said: Well, that is another problem you can’t really grasp nor picture. But it shows our relationships: God the Father, the creator – you imagine what you feel for your own father but projected onto God. Then the Son, he is the mediator, he is much closer. He has reconciled us with the Father. And then the Holy Ghost, the wisdom, the love, [is] really humanity’s ideal. It is almost as if God has personally cut this up for us . . . Depending on the problem, we address ourselves each time to another ‘person’ in quotes [sic]. That simply is a help for us.

The scoring was group 2, because Peter had not yet fully grasped the intrinsic relationship between the three personae (the perichoresis): his 6

A post hoc Kruskal-Wallis analysis of variance (H test) further supported the findings of the correlation computation: the mean RCR ranks of the members of groups 1 to 3 differed (chi2 = 10.17, df = 2, p = .006). The corresponding U-Test (Mann-Whitney) showed that the significant differences were between group 1 and group 3 (U = 9, p < .01), and between group 2 and group 3 (U = 19, p = .02).

Religion

125

Table 7.2 Frequencies of individual scores of RCR levels and intelligibility judgement of the Doctrine of the Holy Trinity. Ntot = 28; age 13–68 years; mean 32 years, 0.3 months; SD 15 years, 4.9 months. (Source: Reich 1994a, p. 122) Is the Doctrine of the Trinity rationally understandable?

Np

Level of RCR

Group 1 (‘no’)

1 2 1 8 12 4

III III(IV) IV(III) IV IV(V) V(IV)

1 2 1 3 1 1

28

9

Group 2 (‘partly’)

5 6

11

Group 3 (‘yes’)

5 3 8

Trinity evokes more tritheism than monotheism. Here are three answers from group 3 respondents: Jean-Luc (46; 7) emphasised: ‘Our mind can’t seize up God as creator in his entire dimension. But our mind grasps certain aspects. God, the wholly other, if he wants to reveal himself to us, then he must do it in a manner that we can understand. Moreover, the formulation [of the doctrine] matches our mental reception capacity.’ Oskar (53; 8) opined: ‘The Trinity somehow combines the human longing for community with the longing for individuality.’ Richard (64; 8) added: ‘It does not bother me that in different situations there are different aspects which you can’t really combine. I always remember that light has to be pictured as wave-like and as corpuscle-like.’

Conclusions The main finding of both studies on the intelligibility of Christian doctrines is that RCR appears to be a necessary but insufficient condition for an intellectually acceptable understanding of the doctrines studied. Specific knowledge and interest (motivation) are needed in addition if the potential competence is to show up in the actual performance (Reich 1994a, p. 124). In a wider context, and taking up an issue discussed earlier, these results indicate that an understanding of Christian doctrines requires one to transgress the limits of formal binary logic (as already known to Thomas Aquinas). Richard expresses this by evoking a parallelism between the

126

Applications of RCR

‘logic’ of the relationship between the three personae and the logic of quantum mechanics. Furthermore, these studies bring out particularly well the means-reflecting thinking (chapter 2, pp. 29–32) at high RCR levels. For instance, Jean-Luc not only evokes the limitations of the human mind (Our mind can’t seize up God as creator in his entire dimension), but also points out that God took them into account (God . . . if he wants to reveal himself to us, then he must do it in a manner that we can understand ).

Co-ordination of religious and scientific world views This section is based on a summary of my related research (Reich 1998). The change of people’s world views as they grow up is a particularly informative developmental instance. Here world view means the way a person pictures the origin of the universe, the way it evolved until the present time, the origin of life, the place of human beings in the universe, etc. There exist purely religious, purely nonreligious, and mixed world views (Fetz, Reich, and Valentin 1989; 2001; Reich 1989, 1990a, 1996b). In a constructivist conception of cognitive development an individual as a rational agent constructs new and/or more dependable views on the basis of his or her own observations and reasoning (e.g., Reich, Oser, and Valentin 1994). In fact, studying the world views of children and adolescents led to my involvement with RCR. I interviewed children, adolescents, and adults aged 7–68 years about statements from a scientist and from a church minister about their respective world views (Reich 1989, 1990a, 1996b). For the scientist the theories of the big bang and of evolution explain all that one would like to know. The minister recognises these contributions to our world view, but adds that for him God is still the ultimate explanation as to why there is a universe altogether. Also, he senses God in nature and in other human beings, and experiences God’s assistance when he has to make morally difficult decisions. The participants were invited to judge who was right, the scientist or the minister. From the answers, five developmental levels could be extracted: Level 1. Only one world view comes into the field of vision: ‘I believe that the minister is right’, or ‘The scientist is right, he can prove it.’ Level 2. Both views are tentatively put side by side: ‘I believe that animals and humans would not have come into existence without God.’ [‘Does that mean science is wrong?’] ‘I would say, maybe there really was a big bang. So the minister is right, and maybe the scientist a little too.’ Level 3. Both world views are considered necessary for a full explanation: ‘Well, to my mind, both are right. The scientist must have developed his views according

Religion

127

Table 7.3 Frequencies of individual RCR levels and levels of co-ordinating biblical and scientific views on the world’s origin. 1(2) means a level somewhat above level 1, 2(1) a level somewhat below 2, etc. N = 67, age 7–68 years. (Source: Reich 1996b, p. 132)

Levels of RCR I I(II) II(I) II II(III) III(II) III III(IV) IV(III) IV IV(V) V(IV)

Level of co-ordination of the two world views 1

1(2)

2 1

2

2 2 1 1

2(1)

2

2(3)

2

1

1 1

4

3 2 1

3(2)

1

3

3(4)

2 4 1 1 1

2 3 2

4(3)

4

4(5)

5(4)

4 4

3 3 2

4 1

3

to the results of scientific research. And the minister is right in that there would be no world if it were not for God. I do not see any contradiction.’ Level 4. The relation between both world views is thematised: ‘The two statements do not exclude each other. The minister speaks about his conscience, his feelings when looking at nature, about human encounters and the like. The scientist explains how the stars came about, and so on. If God had created the preconditions for those processes to occur, then the two views would supplement each other. The world came into being rather suddenly, perhaps somehow through an energy created by God, which enabled matter to come into existence. I am unsure how to understand symbolically Genesis in the Bible. Anyway, nobody can visualise the time scales involved.’ Level 5. A synopsis is endeavoured: ‘If I were the third person in that discussion between the minister and the scientist, I might say the following: Maybe things occurred as stated by the scientist. He has presented a model that explains plausibly how things evolved from the big bang until today. But of course, we cannot be absolutely sure about it. But I also have to side with the minister, and even to support him: Maybe in the future even more convincing models will come about. Anyway, they will not explain why there is a world altogether, and why our life proceeds as it does, and not differently. I too believe that one can sense God in nature, in human encounters, and in one’s conscience.’

At first blush, that sequence from level 1 to level 5 resembles that of RCR levels I to V. RCR levels were assessed employing the standard procedure (also used, e.g., in the preceding studies on the grasp of the two Christian doctrines; see Table 4.4, p. 58 for all of the results underlying Table 7.3).

128

Applications of RCR

However, the two types of levels were not ‘the same’ throughout when the ‘nonreligious’ RCR scores were compared individually to the world view scores (Table 7.3). For instance, some interviewees who argued about the nonreligious issues at RCR level III, ‘co-ordinated’ the world views at level 1, and so did even a respondent at RCR level IV. Hence, the intra-individual variance was larger when religious issues were included. Indeed, for eighteen out the sixty-seven participants (27 per cent) the difference between their RCR level score and the co-ordination score was at least one full level. Is the difference due to the characteristics of the religious domain? One explanation would indeed be ‘segmentation’ (Piaget’s ‘d´ecalage’), that is the developmental time delay between reaching a given level in one domain, and later reaching the same level in a different domain, the size of the delay depending on the particular issue or domain. Alternatively, lack of pertinent knowledge or perhaps insufficient motivation for applying the existing competence could be a reason (Reich 1996b). Notice, however, that nobody argued the case of co-ordinating different world views at a higher level than the nonreligious RCR issues. It therefore appears that basically the same form of reasoning is at work. (Spearman’s rank correlation coefficient has the high value rs = .84, p < .001 – and rs = .94, p < .001, if the eighteen ‘deviant’ cases are discounted). In chapter 6 (pp. 109–10), I commented on Barbour’s four categories characterising the relation between science and religion/theology to the effect that they could be considered as RCR developmental levels. I also contended that eventually the solution resulting from applying RCR presented there would be accepted more generally (p. 114). The present empirical study supports this contention in that the answers at level 5 (having come about by reasoning at RCR level V) are ‘isomorphic’ with the result of the exercise of chapter 6: throughout, (a) A, B, (C) are considered necessary for elucidating the explanandum; (b) the links between A, B, (C) and (c) the context-dependence of their explanatory power are explicitly thematised. The present study (Table 7.3) has been ‘replicated’ several times (Reich 1996b, p. 128). As to the actual results of questioning various specific samples of adolescents and young adults, mostly about science and religion (apprentices, high school students, university level students), the proportion of respondents arguing at least at coordination level 3 ranged from a few per cent (Jablonski and Gryzmala-Moszczynka 1995, p. 53; Nipkow and Schweitzer 1991, pp. 96–7) to about 40 per cent (Tamminen 1991, p. 128). In the studies of William Kay and Leslie Francis (1996, p. 100), 33 per cent of the 11- to 14-year-olds functioned at level 2 and an equal proportion at level 3. (Eighty per cent of the 21- to 25-year-olds were found at level 4. Peter Fulljames and his colleagues made comparable

Religion

129

studies and also interpreted certain results in terms of RCR – Fulljames 1996, p. 259.) How can one understand such divergencies? I have no robust explanation, but note that (a) none of these studies used exactly the interview method described here, (b) in case questionnaires were used, the items were phrased differently, and (c) in one case (Nipkow and Schweitzer 1991) the data resulted from a secondary analysis of answers to questionnaires designed for a different purpose. While all these studies can be interpreted as a (weak) support of the thesis argued here, they illustrate the need to ‘harmonise’ the methodology if results are to be compared in a convincing way. Finally, the present study also confirms the earlier statement that cognitive development is not necessarily synonymous with railroading a person to ‘more developed’ views. To espouse a particular developed view, a corresponding level of cognitive development is necessary, but insufficient. In cases where one’s established positions come into play, a decision may well be needed to apply one’s competence for furthering development, and then to keep on applying it. RCR and religious development What is the essence of ‘religious development’? Already Paul of Tarsus was aware of it: ‘When I was a child, I spoke like a child, I thought like a child, I reasoned like a child. When I became an adult, I put an end to childish ways’ (First Letter to the Corinthians 13: 11, NRSV), and he drew the conclusion: ‘We must no longer be children, tossed to and fro and blown about by every wind of doctrine, by people’s trickery, by their craftiness in deceitful scheming’ (Letter to the Ephesians 34: 14, NRSV). Thomas Aquinus (1225–74) wrote in his Summa theologica (1273, part I, question 75, article 5, second of his arguments concerning objection 4): ‘Omne quod recipitur in aliquo, recipitur in eo per modum recipientes’ (‘Whatever is received into something is received according to the condition of the recipient’). In other words, it depends on the (developmental) state of the receiver’s (epistemic) cognition how the sense perceptions are analysed and interpreted. At present, notably James W. Fowler’s theory of ‘faith development’ and Fritz K. Oser’s theory of ‘religious judgement’ describe religious development. Both are stage theories, and RCR plays a role in both. Fowler’s stages of faith Fowler’s (1981, 1987, 1996, 2001) seven stages of faith (and selfhood) are labelled (1) primal faith, (2) intuitive-projective faith, (3) mythic-literal faith, (4) synthetic-conventional faith, (5) individuative-reflective faith,

130

Applications of RCR

(6) conjunctive faith, and (7) universalising faith. The seven dimensions of faith are taken to be (a) form of logic, (b) perspective taking, (c) form of moral judgement, (d) bounds of social awareness, (e) locus of authority, (f) form of worlds coherence, and (g) symbolic function (Fowler 1981, pp. 243–57). RCR has mainly to do with (a), the form of logic, specifically with that of stage 6 (and stage 7): In the transition to Conjunctive faith one begins to make peace with the tension arising from the fact that truth must be approached from a number of different directions and angles of vision. As part of honoring truth, faith must maintain the tensions between these multiple perspectives and refuse to collapse them into one direction or another. In this respect, faith begins to come to terms with dialectical dimensions of experience and with apparent paradoxes: God is both immanent and transcendent; God is both an omnipotent and a self-limiting God; God is sovereign of history while being the incarnate and crucified One. In physics, in order to account for the behavior of light, two incompatible and unintegrable models must be employed – one based on the analogy with packets of energy, and the other upon the analogy with wavelike motions somewhat as in sound. Similarly, many truthful theological insights and models involve holding together in dialectical tension the ‘coincidence of opposites’. (Fowler 1987, p. 72)

Does that not read like a description of RCR at work, more specifically RCR levels III and IV? (Even if Fowler originally labelled the underlying form of thought as dialectical thinking, he now agrees that my analysis of ‘the relations between Fowler stage six and Oser/Gmunder ¨ stage five is on target’.) As to his stage 7, Fowler (1987, p. 75) writes: ‘In this stage we see persons moving beyond the paradoxical awareness and the embrace of polar tensions of the Conjunctive stage.’ As there are very few persons at stage 7, it is hard to know eactly what that means. Fowler postulates synthetic/unitive thinking at this stage. That is not in contradiction with RCR level V (assuming an everyday awake state of consciousness and not an enlarged, altered state of consciousness – Reich, 2001a). Oser/Gm¨under’s stages of religious judgement ‘Religious Judgement’ (RJ) involves the interpretation of a given experience from the perspective of the personal relationship to an Ultimate Reality, God for religious believers. The five experimentally observed stages are labelled (1) Deus ex machina, (2) Do ut des (give so that you may receive), (3) Deism, (4) Divine plan, and (5) Universal solidarity (Oser and Gmunder ¨ 1991; Oser and Reich 1996). The dimensions of the religious judgement are so-called polar pairs, for instance the ‘immanent’ and the ‘transcendent’. Religious development translates as, and stipulates, a

Religion

131

5.0 RCR

4.5

4.0

3.5

3.0

2.5 2.0

2.5

3.0

3.5 RJ

4.0

4.5

5.0

Figure 7.2 Correlation between religious judgement stages (RJ) and levels of relational and contextual reasoning (RCR). Kendall’s tau = 0.70; p < .01. N = 30.

changing equilibrium between these poles: ‘Psychologically, this means that persons produce stage-specific equilibria between the immanent and the transcendent . . . ’ (Oser and Gmunder ¨ 1991, p. 27). At RJ stage 1, the immanent and the transcendent are seen as totally separate. At the highest stage, ‘the transcendent becomes evident in the immanence of human communication, and vice versa’ (ibid.). To take another pair, ‘freedom’ and ‘dependency’: at RJ stage 1, these concepts mutually exclude each other – freedom is conceived as freedom from dependency. At the higher stages, freedom is conceived progressively as freedom for something and interrelated more and more with dependency. Finally, freedom is recognised as being grounded in dependency on God (ibid., pp. 27–8). For these and further reasons (Oser and Reich 1996), a case can be made that RCR development is important for religious development. To test this hypothesis empirically, I interviewed, in a pilot study carried out in Germany and Switzerland, thirty adolescents and adults, aged from 13–79 years (Schenker and Reich, forthcoming). Their religious

132

Applications of RCR

judgement was assessed using the modern Job dilemma about the church minister who becomes blind (cf. Oser and Gmunder ¨ 1991, p. 173), and RCR using the standard problem of the power plant accident (pp. 20 and 55). The results are shown in Figure 7.2. The correlation between RJ and RCR is high and significant, Kendall’s tau = 0.70; p