Skip to Main Content
Have library access? Log in through your library
Discovering Complexity

Discovering Complexity: Decomposition and Localization as Strategies in Scientific Research

William Bechtel
Robert C. Richardson
Copyright Date: 2010
Published by: MIT Press
Pages: 340
  • Cite this Item
  • Book Info
    Discovering Complexity
    Book Description:

    In Discovering Complexity, William Bechtel and Robert Richardson examine two heuristics that guided the development of mechanistic models in the life sciences: decomposition and localization. Drawing on historical cases from disciplines including cell biology, cognitive neuroscience, and genetics, they identify a number of "choice points" that life scientists confront in developing mechanistic explanations and show how different choices result in divergent explanatory models. Describing decomposition as the attempt to differentiate functional and structural components of a system and localization as the assignment of responsibility for specific functions to specific structures, Bechtel and Richardson examine the usefulness of these heuristics as well as their fallibility -- the sometimes false assumption underlying them that nature is significantly decomposable and hierarchically organized.When Discovering Complexity was originally published in 1993, few philosophers of science perceived the centrality of seeking mechanisms to explain phenomena in biology, relying instead on the model of nomological explanation advanced by the logical positivists (a model Bechtel and Richardson found to be utterly inapplicable to the examples from the life sciences in their study). Since then, mechanism and mechanistic explanation have become widely discussed. In a substantive new introduction to this MIT Press edition of their book, Bechtel and Richardson examine both philosophical and scientific developments in research on mechanistic models since 1993.

    eISBN: 978-0-262-28917-7
    Subjects: General Science

Table of Contents

Export Selected Citations Export to NoodleTools Export to RefWorks Export to EasyBib Export a RIS file (For EndNote, ProCite, Reference Manager, Zotero, Mendeley...) Export a Text file (For BibTex)
  1. Front Matter
    (pp. i-vi)
  2. Table of Contents
    (pp. vii-x)
  3. Preface to the MIT Press Edition
    (pp. xi-xii)
  4. Preface to the Original Edition
    (pp. xiii-xvi)
  5. Introduction: Discovering Complexity—Further Perspectives
    (pp. xvii-liv)

    William Wimsatt clearly perceived the centrality in biology of seeking mechanisms to explain phenomena. It was a perception not shared by most philosophers, who persevered in the attempt to fit the life sciences generally, and biology in particular, into the model of nomological explanation that had been advanced by the logical positivists. The biological sciences fit poorly into this framework, a conclusion we began to appreciate as students of Wimsatt’s and came to appreciate even more as we conducted the research that culminated in the publication ofDiscovering Complexityin 1993.

    When we initiated the research that led to the...

  6. Part I: Scientific Discovery and Rationality

    • CHAPTER 1 Cognitive Strategies and Scientific Discovery
      (pp. 3-16)

      Logical positivism drew a principled and sharp distinction between the context of justification and that of discovery. According to this view, the empirical evaluation of scientific theories could be submitted to logical analysis, with the goal of specifying the conditions under which a theory would be confirmed or disconfirmed. Theories were accordingly given an axiomatic rendering. The fundamental laws provided the axioms in terms of which all else was explained. The theorems were either nonfundamental laws or observational claims. Scientific discovery, by way of contrast, seemed to be far less congenial to logical analysis and was therefore shunted off as...

    • CHAPTER 2 Complex Systems and Mechanistic Explanations
      (pp. 17-32)

      Our aim is to develop a cognitive model of the dynamics of scientific theorizing that is grounded in actual scientific practice. Our focus is on one kind of explanation, one involved in understanding the behavior of complex systems in biology and psychology. Examples of the complex systems we have in mind are the physiological system in yeast that is responsible for alcoholic fermentation, and the psychological system responsible for memory of spatial locations. As we shall discuss in this section, these explanations, which we refer to asmechanistic explanations, propose to account for the behavior of a system in terms...

  7. Part II: Emerging Mechanisms

      (pp. 35-38)

      In Chapter 1 we emphasized that our investigation is directed at identifying the cognitive constraints affecting theory development, and in Chapter 2 we introduced decomposition and localization as the central heuristics figuring in our treatment of the development of mechanistic explanations. We turn now to developing a more detailed analysis, one grounded in historical analyses, of how scientists actually develop mechanistic models. As we proceed, we will focus onchoice-points, points at which decisions are made that shape the explanatory endeavor. The decisions scientists make are affected by their own cognitive characteristics (for example, the fact that they are agents...

    • CHAPTER 3 Identifying the Locus of Control
      (pp. 39-62)

      Before developing a mechanistic explanation of a particular phenomenon, one must identify which system is responsible for producing that effect. Identifying a responsible system presupposes several critical decisions. The scientist must segment the system from its context and identify the relevant functions assigned to it. To substantiate the assignment of a function to a system, the scientist generally must offer theoretical or empirical arguments showing that the physically and functionally independent system identified has substantial internal control over the effect. This is what we describe as treating the system as the locus of control for a phenomenon. That some system...

    • CHAPTER 4 Direct Localization
      (pp. 63-92)

      Having “isolated” a system and identified it as a locus of control, the next step is to ask how the system does what it does. The goal is one of identifying and elaborating the mechanisms underlying behavior. A variety of approaches are possible. One that is often employed—especially in the earliest stages of a research program—is the identification of some component of the system that is itself responsible for producing the behavior, still leaving aside the question of how that component produces the effect. This is what we have called direct or simple localization. Responsibility for an effect...

    • CHAPTER 5 The Rejection of Mechanism
      (pp. 93-116)

      Not all scientific investigators see the development of mechanistic explanations as a critical constraint on their models. Even when there is general agreement in identifying a particular higher-level system as a locus of control for some phenomenon, those who do not accept a mechanistic paradigm may reject any attempt at further decomposition and localization of the sort described in Chapter 4. Decomposition and localization are seen as yielding only spurious explanations. As we will illustrate with several cases in this chapter, decomposition has been rejected by some for failing to provide an adequate explanation of life or of mind. Critics...

  8. Part III: Elaborating Mechanisms

      (pp. 119-124)

      In Part 2 we explored the preliminaries to mechanistic explanation. We turn now to the process of developing and elaborating such explanations. This requires showing how an activity that is performed by a whole system is accomplished by having different components perform different functions that contribute to the task. This is where the heuristics of decomposition and localization most properly come into play. Decomposition involves establishing a division of labor according to which different tasks involved in the same overall process are identified. Localization entails a systematic and independent examination of the processes operating at the lower level and a...

    • CHAPTER 6 Complex Localization
      (pp. 125-148)

      Mechanism was often seen as the only viable alternative to vitalism, and it was undertaken despite the fact that the simple mechanistic approaches were not wholly satisfactory as explanations, even on their own terms. At the same time vitalism can be seen as a consequence of mechanism, drawing sustenance from a clear vision of mechanism’s limitations. Thus the mechanism of Gall, or of the preformationists, was mechanistic, but hardly provided a palliative for the explanatory needs it was designed to fulfill. As Flourens saw very clearly, simply relocating the faculties at a lower level was not an adequate explanation in...

    • CHAPTER 7 Integrated Mechanisms
      (pp. 149-172)

      Analysis into localized components and their interactions is a fruitful scientific strategy when the system under study is nearly decomposable; that is, when the organization is relatively simple. In defendingneardecomposability as a heuristic for human problem-solving, Simon offers two markedly different kinds of reasons in its favor. First, given the resource limitations of human beings, near decomposability is an assumption that enables us to deal efficiently with complex systems. This is a kind of naturalistic grounding for simplicity. Second, simply or nearly decomposable systems are more likely to evolve.¹ We have already pointed out that Simon’s second reason...

    • CHAPTER 8 Reconstituting the Phenomena
      (pp. 173-196)

      In the last two chapters we have illustrated the use of localization and decomposition in developing models of complex systems. In all the cases we have discussed, the development of explanatory models was significantly constrained by lower-level theories as well as systemic behavior. These constraints varied in relative strength. In Chapter 6 the initial models of brain function incorporated simple decomposability. Linguistic functions on the one hand, and memory functions on the other, were assumed to be functionally independent and discretely localized. However, research did not limit itself to simple localization. The initial localizationist models were followed by an attempt...

  9. Part IV: Emergent Mechanism

      (pp. 199-201)

      In Part II we saw that the rejection of localization and decomposition tends to accompany the rejection of a mechanistic program. Providing a mechanism involves describing distinct components, each of which makes a contribution to the performance of the system. This requires both functional and physical independence. In the simplest cases, these components are thought of as making their contributions independently: nature is simply decomposable and embodies an aggregative organization. In slightly more complicated cases, the components are thought to make their contributions sequentially, or linearly, and to retain an integrity of their own: nature is nearly decomposable. As we...

    • CHAPTER 9 “Emergent” Phenomena in Interconnected Networks
      (pp. 202-229)

      The more complex localizationist explanations we examined in Part III are still recognizably mechanistic. Tasks involved in performing a function are divided between components, and system behavior is explained by showing how it is accomplished through the combined performance of the component tasks. Although one might prefer explanations in which the component tasks can be thought of as following a linear, sequential order, so that the contributions of each component can be examined separately, natural systems are not always organized in such a manner. Component tasks are often dependent on one another, so we cannot understand the operation of the...

    • CHAPTER 10 Constructing Causal Explanations
      (pp. 230-244)

      Our focus has been on decomposition and localization and their role in the development of scientific research programs. We have especially emphasized their heuristic role in the construction of explanatory models within problem domains that are ill structured. In the cases we have discussed there was initially no well-delineated space of explanatory models, nor even a clearly defined range of phenomena to be explained. The problem-solving tasks were, then, exploring and constructing the space of explanatory models, and determining the precise range of phenomena to be explained. These tasks are interdependent, their relationship neither simple nor uniform. As Ian Hacking...

  10. Notes
    (pp. 245-256)
  11. References
    (pp. 257-280)
  12. Index
    (pp. 281-286)