from Wikipedia, the free encyclopedia

Abduction ( lat. ABDUCTIO "deportation, abduction"; Engl. Abduction ) is an epistemological term which mainly by the American philosopher and logician Charles Sanders Peirce was introduced in the scientific debate (1839-1914).

"Abduction is the process in which an explanatory hypothesis is formed" ( Peirce : Collected Papers (CP 5.171) ). Peirce understood this to be a method of conclusion that differs from deduction and induction in that it expands knowledge .

Peirce designed a three-stage cognitive logic of abduction, deduction and induction. In this sense, in the first stage of the scientific knowledge process, a hypothesis is found by means of abduction. In the second stage, predictions are derived from the hypothesis. This is a deduction. The third stage searches for facts that “ verify ” the assumptions . This is an induction. If the facts cannot be found, the process starts all over again, and this repeats itself until a hypothesis generates predictions for which suitable facts can be found.

This impulse was partly taken up in recent debates in the philosophy of science about the nature and methodology of scientific knowledge, but it was also discussed controversially. Most of the more recent discussion has been conducted in connection with the concept of inference on the best explanation . The most varied elaborations of a methodology of abductive reasoning were proposed and applications in various individual sciences were discussed, including areas such as cultural studies or semiotics . Various theories about the nature of certain inference procedures of "everyday logic" also use the term abduction.

The abductive conclusion

In the history of logic, the idea of ​​abduction or hypothesis goes back to Aristotle , who mentions it with the term apagogue (First Analytik II, 25, 69a) and already contrasts it with induction ( conclusio ). The translation of the term apagogue with abduction was first made in 1597 by Julius Pacius , a Heidelberg law professor.

Peirce did not introduce the term "abduction" into the sciences, but took up a long-forgotten term and reintroduced it into language. The special achievement of Peirce is to have examined this inference more closely and made it fruitful for the logic of the scientific process. Peirce first used the term “abduction” around 1893, but did not use it systematically until 1901. From 1906 Peirce began to use the term retro-production.

In the language of logic, abduction can be described as follows:

“The surprising fact C is observed; but if A were true, C would be a given; consequently there is reason to suspect that A is true. "

- Peirce : Collected Papers (CP 5.189)

It does not start with a well-known rule, but a surprising event, something that raises serious doubts about the correctness of one's own ideas. Then in the second step there is an assumption, an as if assumption: if there were a rule A, then the surprising event would have lost its surprising character.

For the determination of the abduction it is now decisive that the essential thing about it is not the «elimination of the surprise», but the elimination of the surprise through «a new rule A». A surprise could also be eliminated by using known rules. But that would not be an abduction. Rule A has yet to be found or constructed; it was not yet known, at least not at the time when the surprising event was perceived. Had the rule already existed as knowledge, the event would not have been surprising. In the second part of the abductive process, a previously unknown rule is developed. The third step then produces two things: on the one hand, that the surprising event is a case of the constructed rule, on the other hand, that this rule has a certain persuasive power.

Peirce characterized abduction in contrast to the inferences of deduction and induction as follows:

“Abduction is the type of argument that starts from a surprising experience, that is, from an experience that runs counter to active or passive belief. This takes the form of a perceptual judgment or a proposition relating to such a judgment, and a new form of belief becomes necessary to generalize the experience. "

“Deduction proves that something must be ; Induction shows that something actually effective is ; Abduction only indicates that something can be . "

"Deduction proves that something must be ; Induction shows that something actually is operative; Abduction merely suggests that something may be . "

- Peirce : Collected Papers (CP 5.171)

Comparison of the conclusions

Deductive inferences are like if-then statements. “Every deduction has this character; it is only the application of general rules to special cases ”( CP 2.620 ). Based on given theorems, deductive inferences are necessarily valid. This applies to the structural sciences mathematics and logic.

Inductive inferences assume a case and a result and determine the rule. Induction is synthetic , that is, observations are used, from which rules are formulated if the frequency is sufficient. The conclusion reached, however, is not necessary .

The abduction is also synthetic. With her, the conclusion is drawn from a result to a rule and a case. So it “closes” from one known size to two unknowns. Because the result is something unique, abduction is the conclusion with the highest risk of fallibility. It is a mere conjecture with no evidential value. The following table is used to methodically clarify the structure of the different conclusions.

Abduction Deduction induction
Result These beans are white.
rule All the beans in this sack are white.
case Those beans are out of that sack.
rule All the beans in this sack are white.
case Those beans are out of that sack.
Result These beans are white.
case Those beans are out of that sack.
Result These beans are white.
rule All the beans in this sack are white.
hypothetical inference from the individual and a rule
to a regularity

Conclusion from the general to the individual
Conclusion from the usual regularity
to the general
Table: Conclusions according to Peirce with the status of the "Lectures on Pragmatism" (1903), for abduction see CP 5.189

“Now what is the purpose of an explanatory hypothesis? Its purpose, in being subjected to the test of experiment, is to avoid any surprise and to establish a behavioral habit of positive expectation which will not be disappointed. Any hypothesis can therefore be admissible if there are no particular reasons for its rejection, provided that it is capable of being verified experimentally and only insofar as it is amenable to such verification. That is roughly the doctrine of pragmatism "

- Peirce : Collected Papers (CP 5.197)

From today's point of view, it is undisputed that Peirce used the term hypothesis to summarize two quite different forms of reasoning under the term hypothesis until around 1898 , but without noticing this (for more details, Reichertz 2013). When he noticed this unclear use, he worked out the difference between the two procedures in his late philosophy and called one operation "qualitative induction", the other "abduction". Most of what Peirce had written on the subject of hypothesis before 1898, however, characterized qualitative induction rather than abduction. Only later does Peirce admit: "By hypothetic inference, I mean (...) an induction from qualities" ( Peirce, CP 6.145 ). Reason for error: “But I was too busy investigating the syllogistic form and the doctrine of logical extension and comprehension, which I saw as far more fundamental than they really are. As long as I was convinced of this, two different types of closing were necessarily mixed up in my idea of ​​abduction ”(Peirce MS 425 - 1902). In a draft letter to Paul Carus , Peirce's views from 1883 went even more sharply into court.

"In almost everything that I put into print before the beginning of this century, I more or less mixed up hypothesis and induction"

Accordingly, in his later work Peirce no longer used the formal structure of the syllogism to characterize the abduction. Rather, he then emphasized the creative moment and the originality of the idea, which emerges like lightning.

“The abductive assumption comes to us in a flash, it is an act of insight, although it is extraordinarily deceptive. It is true that the various elements of the hypothesis were in our minds before; but the idea of ​​bringing together what we would never have dreamed of bringing together, flashes the new conjecture in our contemplation "

- Peirce : Collected Papers (CP 5.181)

Abduction as the starting point of the cognitive process

Peirce saw abduction as the starting point for the cognitive process. What man as mind receives data, is perception . “ Nihil est in intellectu quod non prius fuerit in sensu. ”( CP 5.181 , German:“ Nothing is in the mind that was not previously in the senses. ”) A distinction must be made between this and perceptual judgments, in which concepts are formed from what is perceived . The perception is "really nothing other than the most extreme case of abductive judgments." ( CP 5.185 )

Necker cube

Peirce illustrated the abductive character of perceptual judgments using optical illusions such as B. Necker's cube . “With such optical illusions, two or three dozen of which are well known, the most astonishing thing is that a certain theory of the interpretation of the figure appears to be given in perception. When shown to us for the first time, it seems as completely beyond the control of rational criticism as any percept is; but after many repetitions of the now familiar experiment, the illusion disappears, becoming less clear at first and finally disappearing completely. This shows that these phenomena are real links between abductions and perceptions ”( CP 5.183 ).

“This faculty of insight has at the same time the general nature of an instinct which resembles the instinct of animals in that it goes far beyond the general faculties of our reason and guides us as if we were in possession of facts wholly outside of them Range of our senses. It still resembles instinct in that it is in a small degree subject to error; because although it takes the wrong path more often than the right one, viewed as a whole it is the most wonderful thing of our entire constitution "

- Peirce : Collected Papers (CP 5.173)


Philosophical meaning

In the epistemological discussion between the representatives of neopositivism ( Rudolf Carnap , Carl Gustav Hempel , Hans Reichenbach and Karl Popper ) it was agreed that the claim to a scientific statement is not about the context of discovery , but about the context of justification. For Hempel as for Popper, the context of discovery is something subjective, irrational, which is not relevant to the question of the scientific nature of a statement / hypothesis. Every statement is permissible in science if it meets the criteria of a rational justification, regardless of how it came about. The only disputed issue was whether the criterion was verification or falsifiability .

Peirce, on the other hand, who, like Popper, assumed a fundamental fallibilism , did not regard knowledge statically, as a state or a fact, but as a process on the basis of his abductive interpretation of perception. While Popper, the logic of the research was studied stood at Peirce the logic of discovery ( logic of discovery ) in focus.

“The perceptual judgment, for its part, is the result of a process, albeit a process that is not sufficiently conscious to be controlled, or, more correctly, that is uncontrollable and therefore not fully conscious. If we were to subject this unconscious process to logical analysis, we would find that it would end in what that analysis would represent as an abductive conclusion based on the result of a similar process to which a logical analysis would represent as a similar abductive conclusion Ending would represent and so on ad infinitum. This analysis would be exactly analogous to that used by the sophism of Achilles and the turtle , and it would fail to represent the real process for the same reason. Just as Achilles would not have to make a series of separate efforts, this process of forming perceptual judgments, because it is unconscious and therefore not accessible to logical criticism, does not involve separate acts of inference, but rather its process takes place in a continuous process Process."

- Peirce : Collected Papers (CP 5.181)

Just like in perception, in the scientific process, too, abduction is the form of inference that forms the starting point of the thought process. The scientist observes a phenomenon that he cannot explain, an anomaly that contradicts his previous theories. This disrupts his habit and leads to doubt that he wants to remove. He is looking for a theory (a rule) that gives him an explanation for the cause and, through verification, leads to a firm conviction again.

“A physicist encounters a new phenomenon in his laboratory. How does he know that the conjunctions of the planets have nothing to do with it or that it is perhaps not because the Dowager Empress of China happened to utter a word of mystical power at the same time last year or perhaps an invisible Djin may be present ? Think of the many millions and millions of hypotheses that could be made, only one of which is true; and yet, after two or three, or at most a dozen guesses, the physicist makes pretty exactly the right hypothesis. By coincidence, he probably wouldn't have done that in the whole time since the earth solidified. "

- Peirce : Collected Papers (CP 5.172)

“Someone would have to be completely insane to deny that science has made many real discoveries. But every single piece of scientific theory that is firmly established today is due to abduction. "

- Peirce : Collected Papers (CP 5.172)

Research is the removal of doubts by finding new rules in order to gain new, firm beliefs. For the orderly process of scientific research, Peirce established the following connection between abduction, deduction and induction:

“After abduction has given us a theory, we use deduction to infer a mixed variety of consequences from that ideal theory, from the point of view that when we perform certain actions we will be confronted with certain experiences. We then move on to trying these experiments, and when the predictions of the theory are verified we have reasonable confidence that the remaining experiments that are yet to be tried will confirm the theory. I say these three are the only ending modes that exist. I am convinced of this both a priori and a posteriori. "

- Peirce : Collected Papers (CP 8.209)

For Peirce, whether an abduction is suitable as a starting point for a scientific theory decides on the fact that the subsequent steps of converting into a general law (deduction) and empirical verification (induction) can also be carried out without logical contradictions. Otherwise a new theory with a new abductive conclusion must be formulated. Since every theory is only one step towards the approach to the truth, this will in the course of time for Peirce apply to every theory currently accepted as correct. "Infallibility in scientific matters is irresistibly funny to me." ( CP 1.9 )

Applications of abduction

Applications of abduction in addition to science and epistemology , the medical diagnosis , criminal investigations, legal procedures, technical troubleshooting, psychology, literature and social sciences, as well as educational and teaching sciences and finally computer-based expert systems . Examples of abductive reasoning within the framework of probability theory are the application of Bayes' theorem or the maximum likelihood method . The reason for this broad spectrum is that abduction is "the only logical operation that introduces any new idea" ( CP 5.171 ).

Different interpretations of abduction

In the more recent scientific reception of the term abduction, attempts have repeatedly been made to condense Peirce's many approaches to the term abduction into one term. Since this did not succeed due to Peirce's sometimes contradicting definition of the term abduction, many have resorted to the means of designing several variants of abduction on the one hand (e.g. Eco; Bonfantini & Proni), on the other hand, leaving the term contradicting itself or accepting it unilaterally grasp.

One of these interpretations is that the users of the term abduction attach great importance to the fact that abduction is a strictly logical operation that can also be methodically produced. Many AI researchers follow Paul Thagard's approach, and a number of social scientists also prefer this reading (in continuation of Hanson). Especially when it comes to modeling cognitive processes, AI researchers have long noticed that abduction is fundamental to human thinking and that therefore no simulation of human intelligence is complete if it does not have the ability to abduce. Therefore, they are particularly interested in writing the abduction as an algorithm.

The second interpretation of the term abduction follows on from Peirce's formulations, which say that abduction explains surprising things and makes incomprehensible understandable things. Above all, the scientists who view reading, interpreting, translating, diagnosing, acting, (criminalistic) enlightenment and much more as everyday examples of abductive reasoning, essentially understand abduction in this way. The following statements are exemplary of such expansions: "The logic of the abductive conclusion can therefore be understood as a practice of solving puzzles [...]" ( Moser ). The following interpretations by Umberto Eco have proven to be particularly momentous (especially for literary studies) : “In view of the fact that, in principle, every time we hear a word we have to decide which code to refer to, an abduction seems to be to be involved in every act of decoding ”( Eco ). “The logic of interpretation is Peirce's logic of 'abduction'” ( Eco ). Eco's statement that all interpretations are based on abduction is, of course, exaggerated, as it levels out what is specific about abduction and what should be made visible through the introduction of this term. Abduction is not the application of a code , not the application of a rule , but abduction is the invention of a rule, the invention of a code. On the other hand, it can be argued in favor of Ecos that he just wanted to point out that the interpretation of natural language is not purely deductive, but also has to take into account other aspects that are difficult or impossible to grasp by rules, such as those of a cultural nature or the context of perception Regarding speakers. The difficulties of a purely deductive interpretation are, for example, clearly evident in tasks such as machine translation .

The third interpretation of the term abduction emphasizes Peirce's statement that abductive conclusions would provide the best or most likely explanation. Researchers who follow Rescher’s interpretation see abduction primarily as part of the “economy of research”. Wirth argues similarly: “Abductive reasoning is a pragmatic strategy, the aim of which is to minimize the risk of failure. [...]. The researcher tries to optimize the probability and plausibility of his hypotheses . He is primarily a gambling betting partner who aligns his judgments and research results with the criteria of successful betting and the successful search for clues before subsuming them to the norms of the scientific-paradigmatic 'criminal justice system'. ”( Uwe Wirth ).

All three interpretations of abduction mentioned here undoubtedly also name characteristics of abduction. But: All these components of abductive reasoning - namely their logical form, their explanatory function and their ability to provide probable readings - are necessary, but not sufficient, components of abduction. These three characteristics do not describe the peculiarity of abduction, but its boundary conditions. To put it bluntly: abductions can, but do not have to be, logical, explanatory or economic. Much can be understood and explained by means of deduction and induction - often even better, and of course deduction provides the best explanation, and induction or even deduction is certainly a more reliable logical conclusion. But the decisive factor in abduction is not its logical form, the explanatory function or the probability, but above all the ability to find a new rule.

See also


  • Norwood Russell Hanson : Patterns of Discovery , Cambridge: Cambridge University Press 1958.
  • Gilbert Harman, "The Inference to the Best Explanation," The Philosophical Review 74: 1 (1965), 88-95.
  • Michael HG Hoffmann: Knowledge development . Klostermann, Frankfurt am Main 2005
  • John R. Josephson / Susan G. Josephson ( eds ): Abductive Inference: Computation, Philosophy, Technology , Cambridge University Press, Cambridge, UK 1995
  • Tomis Kapitan: Peirce and the Autonomy of Abductive Reasoning , Knowledge 37 (1992), 1-26.
  • J. Ladyman et al. a .: A Defense of Van Fraassen's Critique of Abductive Inference: Reply to Psillos , The Philosophical Quarterly 47/188 (1997), 305-321.
  • Joachim Lege: Pragmatism and Jurisprudence. About the philosophy of Charles Sanders Peirce and about the relationship between logic, evaluation and creativity in law , Mohr Siebeck: Tübingen 1998, ISBN 978-3-16-146977-0 .
  • Peter Lipton: Inference to the Best Explanation , London: Routledge 2001, ISBN 0-415-24202-9 . Review by Lefteris Farmakis and Stephan Hartmann, London School of Economics
  • Ernan McMullin: The Inference that Makes Science , Milwaukee: Marquette University Press 1992.
  • T. Menzies: Applications of Abduction: Knowledge-Level Modeling , International Journal of Human-Computer Studies 45.3 (1996), 305-335.
  • I. Niiniluoto: Defending Abduction (PDF; 56 kB), Phil Sci 66 (Proceedings) (2000), 436-451.
  • Sami Paavola: Abduction as a logic and methodology of discovery , Foundations of Science 9/3 (2004), 267-283
  • Charles Sanders Peirce: Writings on Pragmatism and Pragmatism . Frankfurt am Main: Suhrkamp 1976.
  • Stathis Psillos: On van Fraassen's Critique of Abductive Reasoning , Philosophical Quarterly 46 (1996), 31-47.
  • Jo Reichertz : The abduction in the qualitative social research , Wiesbaden: Springer Verlag 2013 (revised and significantly expanded edition from 2003), ISBN 3-8100-3595-5
  • Ansgar Richter: The concept of abduction in Charles S. Peirce , Lang, Frankfurt am Main 1995, ISBN 3-631-48338-4 .
  • Ines Riemer: Conception and justification of induction . Wuerzburg 1988.
  • Susanne Rohr: About the beauty of finding. The internal structure of human understanding according to Charles S. Peirce: Abduction Logic and Creativity , Stuttgart 1993 (also Diss. FU Berlin 1991).
  • Gerhard Schurz : Models of Abductive Reasoning , Sintonen, M. (Ed.): The Socratic Tradition, 2007.
  • Paul Thagard: The Best Explanation: Criteria for Theory Choice , J Phil 75 (1978), 76-92.
  • Douglas Walton: Abductive Reasoning , University of Alabama Press 2004, ISBN 0-8173-1441-5 . Review by J. Grossmann
  • Gerd Wartenberg: Logical socialism. The transformation of Kant's transcendental philosophy by Ch. S. Peirce . Frankfurt am Main: Suhrkamp 1971.
  • Uwe Wirth (ed.): The world as a sign and hypothesis . Frankfurt am Main: Suhrkamp 2000.

Web links