From Cynefin.io
Jump to navigation Jump to search

“Abductive logic”, “Abduction”, or “Abductive reasoning” is one of the types of logical reasoning alongside inductive and deductive reasoning. In general terms, deduction refers to drawing a particular conclusion from a general principle. As long as the principal is true, a deductive statement will always be valid (for example, “All people are mortal”, “I am a person”, “Therefore I am mortal”.). Induction, conversely, refers to inferring a general principle out of a body of knowledge made up of particular instances. The “problem of induction” [1] has a long history in the philosophy of science, addressing the question of how to draw valid general principles out of specific observations.

Abduction differs from both of those types as a type of inference that seeks explanatory connections, with a strong role played by intuition and experience. The precise definition differs slightly depending on the approach, the main approaches relevant to this space being outlined in the following sections. The three types of reasoning have implications about the contexts and circumstances where their use is most appropriate, and they imply certain relationships between past and future and their symmetry and asymmetry. In general, the more asymmetric the relationship between past and future, i.e., the less certain we are that past knowledge and observations are going to repeat in the future, the more relevant and appropriate abductive reasoning becomes.



All the beans from this bag are white

These beans are from this bag

Therefore, these beans are all white


These beans are from this bag

These beans are white

Therefore, all the beans from this bag are white


All the beans from this bag are white

These beans are white

Therefore these beans are from this bag.

Abduction in Pierce

The American philosopher and scientist Charles Saunders Pierce if responsible for coining the term “abduction” and emphasising its role in the scientific process. Pierce especially associates abduction with its role in the early stages scientific discovery and the development of hypotheses. Gilbert Harman is another philosopher closely associated with abduction, and is responsible for the commonly-cited definition of abduction as “inference to the best explanation” [3]. This approach recognises the role played by abductive thinking under conditions of high uncertainty, and its connection to novel discovery. It also identifies disciplines where abductive reasoning is especially relevant, such as archaeology, and acknowledges its presence in everyday thinking [4]. Pierce’s own view of abduction also evolved in time, coming to be understood as the only way that a new idea can be introduced, with deduction and induction coming in later [5].

Abduction in Bateson

Bateson defines abduction as “a double or multiple description of some object or event or sequence”. While developing his own use of the term, Bateson also acknowledges the debt to Pierce. By “a double description”, Pierce refers to describing one thing in terms of another, and sees it as ever-present: “Metaphor , dream , parable, allegory, the whole of art , the whole of science, the whole of religion , the whole of poetry, totemism (as ready mentioned) , the organization of facts in comparative anatomy-all these are instances or aggregates of instances of abduction, within the human mental sphere” [6]. Through this understanding, abduction becomes a tool for and an expression of connection, and also describes how change in one domain propagates through others. Nora Bateson has continued expanding on this concept of abduction. [7]

Abduction in Cynefin

The work Dave Snowden and the Cynefin Co are doing around abduction is complementary to the work of both Pierce and the Batesons and is compatible with both approaches, while adding new dimensions to them. In this approach, abduction is tied closely to the capacity for exaptation, the radical repurposing of a trait for a function different than the one it originally served. In a use that brings it slightly closer to Pierce, abductive thinking is also useful in the process of evaluating coherence when generating multiple probes in a complex system.

SenseMaker is also central to the Cynefin Co use of abduction in practice - specifically, SenseMaker addresses the issue of determining when hunches are coherent and when they are not, and providing some filters between alternatives. This use of abduction is not completely different to Pierce or Bateson, but it provides slightly more robust support for the hypothesis formation that Pierce describes, while allowing elements of the abstraction- and metaphor-based application of abduction to be applied at scale and without depending on facilitation. Essentially, the outputs from SenseMaker, based on its design, act as a network of abstraction that allows us to observe connections between items that we would not have recognised as similar, but that are consistently referring back to the same abstraction: for example, completely different practices that are associated with the emergence of psychological safety. These form the basis for abductive connections, while allowing outliers and weak signals to remain in view. These connections don’t just happen between ideas, but also between people, creating the groundwork for increased empathy.

Abduction and human cognition

The capacity for abduction is one of the elements Dave Snowden uses in the explanation of differences between human intelligence and machine learning. The importance and uniqueness of abduction in all aspects of reasoning, innovation, learning, and hypothesis discovery is one of the possible pathways of delineating the areas where applications of AI are more appropriately deployed, and the ones where their over-application might stifle human abductive capacities. The concept also assists in exploring areas for augmentation of human capacities: where can automation be used appropriately to allow more time for validation and abductive reasoning?

Other resources

Dave Snowden, Risk and resilience, video uploaded by Cognitive Edge (16 May 2011)


Articles & References

  1. The Problem of Induction, Internet Encyclopaedia of Philosophy
  2. Peirce, Charles S. 1965. Collected Papers of Charles Sanders Peirce. Vol. 2, Elements of Logic, ed. Charles Hartshorne and Paul Weiss. Cambridge, Mass.: Harvard University Press.
  3. Harman, Gilbert H. “The Inference to the Best Explanation.” The Philosophical Review 74, no. 1 (1965): 88–95. https://doi.org/10.2307/2183532.
  4. Walton, Douglas N. (2004). Abductive Reasoning. Tuscaloosa, AL, USA: University Alabama Press.
  5. Pierce on Abduction, Stanford Encyclopedia of Philosophy
  6. Bateson G. Mind and nature: a necessary unity. New York: Bantam Books, 1979
  7. Bateson, N. (2022). An essay on ready-ing: Tending the prelude to change. Systems Research and Behavioral Science, 39(5), 990–1004. https://doi.org/10.1002/sres.2896