The Dilemma of Defining Abduction

Although Peirce states that abduction is the “only logical operation which introduces any new idea,” many, if not most, explanatory hypotheses offer nothing new at all. They do not seem to be, in the sense Peirce means, abductively derived. In various writings, Peirce provides at least four different descriptions of abduction. Italian computational philosopher Lorenzo Magnani proposes three types of abduction: theoretical, model-based, and creative (the final chapter of this book discusses the third type). In her 2005 paper for Semiotica (“Abduction as an Aspect of Retroduction”), Chiasson points out that Peirce uses two distinct and contradictory terms to signify these processes. The purpose of this chapter is to clarify the various ways abduction has been defined. In addition to defining abduction as an aspect of retroduction, the authors discuss induction, with which abduction is often confounded. This discussion of induction includes the concepts diagnosis and inference to the best explanation, both of which can be achieved inductively (and deductively as well, though deduction will not be addressed here), as well as abductively.

Phenomenology

Relational Thinking Styles is traced to Peirce’s phenomenology, his logic and concept of abduction. A process similar to Peirce’s descriptions of this phenomenological sort of proto-abduction is demonstrated and observed by means of the Davis Non-Verbal Assessment of inferencing styles. Noticing, or failing to notice, similarities and differences among things resides at the core of reasoning; all similarities and differences are discerned based upon the qualities of things, for there is no possibility of discernment without qualities to discern among. A mind cannot think about what it does not notice or has not previously noticed. Individuals become aware of similarities between things and ideas order and organize qualities, or properties, which distinguish one thing from another. Peirce’s practice of phenomenology as a whole comprises the qualitative core of reasoning. Since these three universal categories underlie the structure of Peirce’s philosophy as a whole, they underlie his logic as well. In particular, these phenomenological categories are essential for understanding his concept of abduction and, therefore, Peirce’s Logic.

Computationally Modeling Inference Patterns

The second perspective, which would be much more complex than the first, involves the shaping of algorithms that could computationally define RTS operational processes, including that of the abductive-like Relational thinking style. In addition to the possibility of contributing to the development of an abductive inference engine, this second approach could potentially provide a means for enhancing various game theories.

Algorithms for Determining Thinking Styles

The algorithms used to identify thinking style patterns derive from Davis’s theoretical construct of RTS. As such, we can demonstrate that they are capturing predicable patterns of human behavior. These patterns are based upon the priority and combination of ways in which individuals of a particular style habitually engage the order and direction of thoughts (sequence), confront options (intensity), and use time (duration). These categories are engaged by means of action patterns comprised of simple or complex repeating, and random or deliberate varying. Identifying habitual inferencing patterns may lead to a better understanding of decision-making and other fields of inquiry.

Why Understanding Thinking Styles Matters

RTS is a human systems theory, a model of innate human operational processes that interact with one another, with materials, and within contexts to affect the functioning of whole systems, regardless of their size or importance. RTS describes (and the DNV assessment identifies) the untrained, instinctive methods by which different individuals habitually make decisions about matters of various degrees of importance. These differing tacit methods are of the nature of automatic, instinct-like habits based upon template-like beliefs and operational habits. No particular method of instinctive thinking is inherently superior to another; no method has much meaning until the issue of context is considered. Thinking habit-patterns are more or less effective depending on the context within which they are used.

Structure of the Relational Thinking Styles Model

This chapter lays out the initial RTS model designed by Dorothy Davis to describe her theory of thinking styles. Included here is a brief discussion of the Peircean-like phenomenological model that Davis used as the framework from which to develop the theoretical model of RTS and the DNV, an analog model of RTS. The main body of the chapter introduces the underlying structure of the RTS model (and, by association, the DNV), laying out the basic criteria for analysis and identification of inference patterns.

Creative Abduction

The focus of this chapter is the sort of abduction that is only achievable by means of the unique, generative process capable of producing original ideas. The authors discuss Kapitan’s (1997) theses as a framework for understanding Peirce’s theory of abduction. They then return to Peirce’s concepts of phenomenology and normative science to explore the relationships between these philosophical concepts (and that of mathematics) with the development of a model for Peirce’s concept of abduction. They conclude with an in-depth description of Relational thinking, which includes algorithms (based upon Chiasson’s notational system) for three of the many types of operations performed by Relational thinkers.

Implications and Applications of Relational Thinking Styles

Over the past thirty-five years, the Davis Non-Verbal Standardized Assessment has been applied within many different fields and business applications including education, social and counseling services, criminal justice, risk management, hiring, succession planning, and so on. RTS provides a platform for further study of the relationships between inference styles and temperament, and the correspondence of these patterns to Peirce’s methodeutic, which resembles elements identified by the RTS model. Action patterns from the DNV are observable in real-time standardized testing situations from which predictions are made based upon the effects of the interactions of elements that indicate a particular style. No other research has thus far addressed the inherent non-linguistic nature of inferencing. Because the DNV assessment functions as a standard for making observations objective during test situations (in the same way a thermometer makes possible observations of temperature) the DNV has specific predictive capabilities. Formal reliability and discriminate validity studies performed at the University of Oregon in 2002-2003 demonstrated high inter-rater reliability and good retest reliability, as well as a strong relationship between the DNV and certain instruments used for discriminate validity studies. However, operational studies (probably by means of computational modeling and/or game theory) will be required to determine the long-term predictive validity of this assessment.

Peirce’s Normative Sciences

One purpose of this chapter is to establish the basis for identifying the etiology of the abductive reasoning process by first addressing the relationship between Peirce’s concept of phenomenology and aesthetics, the first of his normative sciences. The connection between aesthetics and abductive inference will be drawn out, leading, consequently, to a discussion of the concept of methodeutic, the third sub-branch of logic from which the testing and proof of abduction needs to be undergone, and suggesting that the RTS model may provide a means for this testing.

Some Variables for Computationally Modeling RTS

This chapter discusses the interrelationships between the three main variables that may provide extra insights into the ways and purposes for which thinking styles might be addressed in computational modeling. Successful modeling of inference patterns often addresses these three primary variables: (1) the inferencing styles for each of the three stages (Discerning, Goal-Setting, and Problem-Solving), (2) delineation of each of the three functions according to the components of each, and (3) generic contextual analyses. Although inferencing styles (1) can be run independently of delineation of function and contextual analyses (2 and 3), they cannot have applicable meaning without knowing to which stage they apply (2). Although the styles and function analyses (1 and 2) can be run separately from contextual analyses (3), the algorithms and the stages to which they apply cannot be meaningfully modeled without generic contextual analyses. Other variables (such as temperament, skills, intelligence, and so on.) may be required, depending upon the purpose of a particular application.