Categorizing Cognition

Categorizing Cognition: Toward Conceptual Coherence in the Foundations of Psychology

Graeme S. Halford
William H. Wilson
Glenda Andrews
Steven Phillips
Copyright Date: 2014
Published by: MIT Press
Pages: 376
https://www.jstor.org/stable/j.ctt1287ht4
  • Cite this Item
  • Book Info
    Categorizing Cognition
    Book Description:

    All sciences need ways to classify the phenomena they investigate; chemistry has the periodic table and biology a taxonomic system for classifying life forms. These classification schemes depend on conceptual coherence, demonstrated correspondences across paradigms. This conceptual coherence has proved elusive in psychology, although recent advances have brought the field to the point at which it is possible to define the type of classificatory system needed. This book proposes a categorization of cognition based on core properties of constituent processes, recognizing correspondences between cognitive processes with similar underlying structure but different surface properties. These correspondences are verified mathematically and shown not to be merely coincidental.The proposed formulation leads to general principles that transcend domains and paradigms and facilitate the interpretation of empirical findings. It covers human and nonhuman cognition and human cognition in all age ranges. Just as the periodic table classifies elements and not compounds, this system classifies relatively basic versions of cognitive tasks but allows for complexity. The book shows that a more integrated, coherent account of cognition would have many benefits. It would reduce the conceptual fragmentation of psychology; offer defined criteria by which to categorize new empirical results; and lead to fruitful hypotheses for the acquisition of higher cognition.

    eISBN: 978-0-262-32070-2
    Subjects: Psychology

Table of Contents

  1. Front Matter
    (pp. i-vi)
  2. Table of Contents
    (pp. vii-viii)
  3. Plan of Chapters
    (pp. ix-x)
  4. Preface
    (pp. xi-xiv)
  5. Acknowledgments
    (pp. xv-xvi)
  6. 1 Introduction and Statement of the Problem
    (pp. 1-26)

    All sciences need ways to classify the phenomena that they investigate. Consider, for example, what chemistry would be like without the periodic table, or biology without the taxonomic system for classifying life forms. Scientific classifications are based on core properties of phenomena that are most strongly related to other properties. Thus, chemical elements are classified in the periodic table by their atomic weight, which is linked to many fundamental properties such as valence (see, e.g., Luder, 1943; Mayr & Bock, 2008). Biological classifications are based on evolutionary history and structural features such as the type of skeleton or nervous system...

  7. 2 Properties of Cognitive Processes
    (pp. 27-64)

    In this chapter, we define the major types of cognitive processes that we have identified in the existing empirical research base. These cognitions are distinguished by the type of representation and by the processes that operate on them (i.e., the operating system). Representations and processes in turn influence observable cognitive performances. Cognitive representations, symbols, and structure are important to our approach, so they are defined first.

    Cognitive representationsare internal states containing information that can be used by an animal, whether human or nonhuman, to interact with the environment in an adaptive manner. We distinguish cognitive representations from those that...

  8. 3 Relational Knowledge in Higher Cognition
    (pp. 65-90)

    In this chapter, we give a more detailed account of the role of relational knowledge in higher cognition and make further comparisons of symbolically structured processes with lower levels.

    Our formulation of symbolic structure is based on relations, and relational representations differ from associations in five main ways. These are structural alignment, symbolization, higher-order representation, accessibility, and generalization.

    Structural alignmentwas defined in chapter 2 as the foundational property of relational knowledge. In essence, entities are assigned to roles, so that loves (John,Sally) is represented by assigningJohnto theLoverrole andSallyto theLovedrole. Similarity...

  9. 4 Cognitive Complexities and Correspondences
    (pp. 91-136)

    Numerous attempts have been made to quantify the complexity of cognitive processes, reflecting the fundamental importance of this parameter. Those that have most relevance to cognitive development have been reviewed by Halford and Andrews (2006, 2011), while those that refer to general cognition have been reviewed by Halford, Cowan, and Andrews (2007) and Cowan (2001). In this chapter, we take account of the contributions that we cited elsewhere, but with a focus on what we see as a coherent and comprehensive account of cognitive complexity principles.

    First, we want to mention some issues that have been pervasive in the field....

  10. 5 Representational Rank
    (pp. 137-176)

    In chapter 2, we defined three levels of process: nonstructured, functionally structured, and symbolically structured. They are rank-ordered for complexity in the sense that they entail increasingly complex representations and cognitive functions, and the lower-level processes remain accessible while processing the higher levels. The complexity of symbolically structured processes has been defined by the relational complexity metric, as outlined in chapter 4. When relational complexity is combined with the three levels defined earlier in this book, an ordering of cognitive complexity is created, as shown in table 5.1. This yields seven levels of cognitive process, which we define as Representational...

  11. 6 Acquisition of Relational Knowledge and the Origin of Symbols
    (pp. 177-210)

    In chapter 2, we defined categories of cognitive processes corresponding to representational ranks 0–6, and properties of each were discussed with reference to empirical evidence in chapter 5. In this chapter, we consider acquisition processes, with reference to the specific characteristics of each rank.

    Rank 0processes correspond to elemental association, which has been massively researched and well summarized in numerous texts (e.g., Domjan, 2003; Lieberman, 1993). The core process is based on contiguity, and there are highly successful computational models (Rescorla & Wagner, 1972).

    Rank 1corresponds to functionally structured processes and includes configural association and “Type-2” (Clark...

  12. 7 Neural Nets as Models of Acquisition Processes
    (pp. 211-230)

    In chapter 2, we examined neural net architectures as existence proofs for three levels of cognitive processes. Here, we consider some of the same types of architectures in terms of how effectively they model the acquisition of cognitive processes. In this chapter, we do not attempt to make a comprehensive review of the neural net literature because there are excellent reviews already available (McClelland et al., 2010 Thomas & McClelland, 2008). There is also a collection of papers on computational models of cognitive development edited by Marcovitch and Zelazo (2012). Recall that we see three types of net architectures: two-layered...

  13. 8 Human Reasoning and Relational Knowledge
    (pp. 231-252)

    In this chapter, we examine some state-of-the art models of inference to assess how our conception of relational knowledge fits with some advanced forms of cognition. Conceptions of human reasoning underwent major transformations in the 20th century. While George Boole (1951) regarded his treatise on logic as defining the laws of thought, an enormous body of subsequent research raised doubts as to whether logic could be regarded as the norm of human reasoning. Many alternatives were proposed, including psycho-logic (Braine, 1978: Piaget, 1950, 1957), logical rules of inference (Rips, 2001), information-processing models (Anderson 1983; Newell & Simon, 1972), heuristics (Gigerenzer...

  14. 9 Applications of Relational Knowledge Theory
    (pp. 253-274)

    The 21st century is arguably the age of complexity, in that life has never been as complex as it currently is and the portents indicate increasing complexity, at least for the foreseeable future. Both work demands and the growth of technology have contributed to rapid increases in the complexity of human performance. It is beyond the scope of this book to review the human factors literature on this topic (see, for example, Vicente, 2003), but in this chapter, we indicate how relational complexity theory can be applied to problems in human performance. Some of our proposals here are based on...

  15. 10 Conclusion
    (pp. 275-286)

    In this book, we have shown that it is possible to distinguish categories of cognitive processes in a way that corresponds to identifiable cognitive architectures and is supported by converging evidence from numerous paradigms. In addition, we have demonstrated that the categories distinguish cognitions by features that can be regarded as foundational, in that they relate to numerous and important properties of cognitive performance. In the process, we suggest that we have brought into focus issues that have tended to be neglected by many otherwise admirable research paradigms. In this chapter, we present the essence of our contribution and summarize...

  16. Glossary of Terms and Abbreviations
    (pp. 287-296)
  17. References
    (pp. 297-336)
  18. Name Index
    (pp. 337-348)
  19. Subject Index
    (pp. 349-358)