The 20th century philosophy of science began on a positivistic note. Its focal point was scientific explanation and the hypothetico-deductive (HD) framework of explanation was proposed as the standard of what is meant by “science.” HD framework, its inductive and statistical variants, and other logic-based approaches to modeling scientific explanation were developed long before the dawn of the information age. Since that time, the volume of observational data and power of high performance computing have increased by several orders of magnitude (...) and reshaped the practice and concept of science, and indeed, the philosophy of science. A new observational-inductive (OI) framework for scientific research is emerging due to recent developments in sensors, data systems, computers, and knowledge discovery techniques. We examine the nature of these changes and their impact on the question of what is meant by “science” after discussing five examples of the OI framework, and conclude that the HD and OI frameworks are complementary and synergistic. (shrink)

Charles S. Peirce luokitteli 1865 päättelyn lajit deduktioon, induktioon ja hypoteesiin, joista viimeksi mainittua hän luonnehti päättelynä vaikutuksista syihin tai päättelynä selitykseen. Hypoteesi on Peircelle induktion ohella tietoa laajentava päätelmä. 1890-luvun lopun kirjoituksissaan hän alkoi kutsua hypoteettista päättelyä uusilla nimillä ”retroduktio” ja ”abduktio”. Tässä vaiheessa Peirce kuvasi abduktiota tieteellisen päättelyn ensimmäisenä askeleena, mahdollisten arvausten esittämisenä, jonka tulokset on asetettava induktion kautta testeihin. Hänen tunnetuin kaavionsa abduktion loogiselle muodolle on vuodelta 1903: The surprising fact C is observed; But if A were (...) true, C would be a matter of course. Hence, there is reason to think that A is true. Toisin sanoen abduktio päättelee yllättävästä tosiseikasta hypoteettiseen teoriaan, joka selittäisi tämän faktan. Peircen ajatukset tulivat laajemmin tunnetuiksi, kun kuusiosainen Collected Papers ilmestyi 1931-35. Maailmansodan jälkeen hänen ehdotuksiaan on tulkittu kolmessa suunnassa. Hansonin mukaan abduktio on keksimisen logiikkaa. Laudanin mukaan abduktiossa on kyse ohjeesta työskennellä kiinnostavan hypoteesin parissa. Harmanin mukaan abduktio on päättelyä parhaaseen selitykseen. Monet loogikot, tekoälyn tutkijat ja tieteenfilosofit käyttävät termiä ”abduktio” suoraviivaisesti IBE:n synonyyminä, kun taas varovaisemman kannan ottavat mm. Aliseda, Magnani ja Douven. Toisaalta tätä rinnastusta on eri syistä pidetty harhaanjohtavana abduktion todellisen luonteen ja tehtävän perusteella. Myös monet Peirce-tutkijat ovat ottaneet intohimoisesti kantaa asiaan vedoten mestarin monitulkintaisiin kirjoituksiin. Esimerkiksi Minnamaier esittää, että abduktion samaistaminen IBE:n kanssa hämärtäisi Peircen erottelua induktion ja abduktion välillä. Campos argumentoi, että abduktio tulisi rajata selittävien hypoteesien generoimiseen, joten valikointi hypoteesien välillä ei kuulu sen tehtäviin. Gaultierin mukaan skeeman johtopäätös ei ole suositus testata A:n totuutta, vaan selvittää, onko A todellakin mahdollinen selitys C:lle. Sen sijaan Nyrupin mukaan jättää avoimeksi, tukeeko onnistunut selitys hypoteesin totuutta, joten abduktiossa on kyse hypoteesin työstämisestä. Peircen kirjoituksista puolen vuosisadan aikana kuitenkin löytyy aineksia kaikille kolmelle abduktion päätulkinnalle. Abduktio on ainoa menetelmä uusien ideoiden generoimiseksi. Abduktio auttaa tunnistamaan testaamisen arvioisia hypoteeseja. Joissakin tapauksissa, kuten havainnoissa ja historiallisissa päätelmissä, abduktio on ”pakottavaa” tai ”täysin varmaa”, joten sillä voi olla tekemistä myös hypoteesien oikeuttamisen kanssa. Vaikka Peirce päätyi esittämään, ettei abduktiivisiin päätelmiin liity mitään todennäköisyyksiä, hän kuitenkin myöhäisissäkin töissään antoi abduktiolle selkeän loogisen muodon. Lisäksi abduktion ja IBE:n yhteyttä puolustaa käsikirjoitus ”The Logic of Drawing History from Ancient Documents”, jossa Peirce antaa kriteerejä ”abduktiolle eli hypoteesien valinnan prosessille”. Näitä kriteerejä ovat selitysvoima, testattavuus, varovaisuus, laajuus ja yksinkertaisuus. Samantapaisia metodologisia sääntöjä ovat muotoilleet ne tutkijat, jotka ovat kehitelleet eteenpäin IBE:n muotoja. Vastaus otsikon kysymykseen on siis ”kyllä ja ei”. Vaikka Peirce keksi termin ”abduktio”, hänellä ei ole loputonta yksinoikeutta määrätä sen käytöstä. Hän ajatuksensa ovat inspiroivia tavalla, joka voi johtaa abduktion tutkimisessa erilaisiin suuntiin. Näitä linjoja on tarpeetonta oikeaoppisesti pitää toisilleen vastakkaisina. Keksimisen logiikka on kiehtova aihe, mutta samalla tavoin on aihetta analysoida sitä, miten keksittyjen selityshypoteesien paremmuutta voidaan vertailla ja millaisilla ehdoilla voisimme olla oikeutettuja hyväksymään parhaan toistaiseksi löydetyistä selityksistä. Kaikki nämä loogis-metodologiset tutkimuslinjat tekevät omalla tavallaan oikeutta Peircen filosofian suurille oivalluksille. (shrink)

Taking as its exegetic point of departure Peirce's outline of a semiotic theory of cognition from the mid 1890s, this paper explores the relevance of this outline to a theory of learning and also to a broader, normative vision of education. Firstly, besides providing for fallibilism in philosophical inquiry Peirce's outline accords with critical strategies of his fellow pragmatists, such as William James's detection of the ‘psychologist's fallacy’ and John Dewey's rejection of the ‘philosophical fallacy’. It is pointed out that (...) this merit of Peirce's outline is due to his abstract and often neglected notion of a prespecialised scientific intelligence. Secondly, the paper shows that Peirce's semiotic approach to prescientific learning processes conditioning all kinds of theoretical activity may indeed complement James's and Dewey's contributions: giving careful attention to the work of linguistic and paralinguistic signs in such learning processes Peirce's analysis captures indexical and iconic aspects, as well as symbolic aspects of linguistic interaction and experientially constrained dialogue. Following Peirce's analysis the paper considers how the iconicity of linguistic structure, ostensive acts and the practical situatedness of language use are pre‐conditions of experiential learning in every day life as well as in scientific learning. It is also shown how this semiotic analysis is connected with Peirce's important notion of abduction. Finally, it is pointed out that a broader vision of education may issue from Peirce's semiotic analysis and that such a vision would emphasise that the life experience of individuals entering educational institutions should be made to bear on the objectives as well as the teaching practices of educational institutions. (shrink)

According to the demand of the Hypothetico-Deductive (H-D) method, a theory is confirmed when the prediction-observation (p-o) gap is small and disconfirmed when the gap is large. A major goal of this paper is to introduce a research domain for which this demand does not hold. In contrast to the H-D method's demand, this research, called the Catch model for reconstructing a face previously seen from memory, requires an increase, within limits, in the p-o gap. The Catch model research substantiates (...) theoretically and empirically a new proposed method that I call the "Deductive-Reconstruction" (D-R) method. This method provides essential conditions whose fulfillment guarantees successful reconstruction of past events (a face previously seen) from memory. It is argued that the D-R method fits the area of research of reconstructing past events from memory better than the H-D method. Application of the H-D method to the Catch model's research domain leads to an internal contradiction and failure to reconstruct past events (a face previously seen) from memory. Finally, the nature of the D-R method along with the Catch model is discussed from three points of view: confirmation, explanation, and generality. (shrink)

This research was carried out to explore some of the cognitive processes involved in scientific anomalies resolution. 40 subjects with a good neuropsychology expertise were asked to explain two (invented) anomalous neuropsychological cases. The subjects' efforts to give a meaningful structure to the data were recorded, and the resulting reasoning blocks were analysed to extract and compute the inferential (deductive, inductive and abductive) and analogical processes used. The processes were intercorrelated to experimentally verify the co-occurrence of different forms of logical (...) thinking. Statistical analysis point out the relevance of abductive inferences, the possible presence of an inferential-style switching process , the high number of external analogies used, the cognitive closeness manifested by expert reasoners. (shrink)

People are habitual explanation generators. At its most mundane, our propensity to explain allows us to infer that we should not drink milk that smells sour; at the other extreme, it allows us to establish facts (e.g., theorems in mathematical logic) whose truth was not even known prior to the existence of the explanation (proof). What do the cognitive operations underlying the inference that the milk is sour have in common with the proof that, say, the square root of two (...) is irrational? Our ability to generate explanations bears striking similarities to our ability to make analogies. Both reflect a capacity to generate inferences and generalizations that go beyond the featural similarities between a novel problem and familiar problems in terms of which the novel problem may be understood. However, a notable difference between analogy-making and explanation-generation is that the former is a process in which a single source situation is used to reason about a single target, whereas the latter often requires the reasoner to integrate multiple sources of knowledge. This seemingly small difference poses a challenge to the task of marshaling our understanding of analogical reasoning to understanding explanation. We describe a model of explanation, derived from a model of analogy, adapted to permit systematic violations of this one-to-one mapping constraint. Simulation results demonstrate that the resulting model can generate explanations for novel explananda and that, like the explanations generated by human reasoners, these explanations vary in their coherence. (shrink)

In this paper it is shown how plausible reasoning of the kind illustrated in the ancient Greek example of the weak and strong man can be analyzed and evaluated using a procedure in which the pro evidence is weighed against the con evidence using formal, computational argumentation tools. It is shown by means of this famous example how plausible reasoning is based on an audience’s recognition of situations of a type they are familiar with as normal and comprehensible in their (...) shared common knowledge. The paper extends previous work on this example by using three new multiagent argumentation schemes closely related to the scheme for argument from negative consequences. (shrink)

This article presents a formal theory of robot perception as a form of abduction. The theory pins down the process whereby low‐level sensor data is transformed into a symbolic representation of the external world, drawing together aspects such as incompleteness, top‐down information flow, active perception, attention, and sensor fusion in a unifying framework. In addition, a number of themes are identified that are common to both the engineer concerned with developing a rigorous theory of perception, such as the one on (...) offer here, and the philosopher of mind who is exercised by questions relating to mental representation and intentionality. (shrink)

Three competing ‘methods’ have been endorsed for inferring phylogenetic hypotheses: parsimony, likelihood, and Bayesianism. The latter two have been claimed superior because they take into account rates of sequence substitution. Can rates of substitution be justified on its own accord in inferences of explanatory hypotheses? Answering this question requires addressing four issues: (1) the aim of scientific inquiry, (2) the nature of why-questions, (3) explanatory hypotheses as answers to why-questions, and (4) acknowledging that neither parsimony, likelihood, nor Bayesianism are inferential (...) actions leading to explanatory hypotheses. The aim of scientific inquiry is to acquire causal understanding of effects. Observation statements of organismal characters lead to implicit or explicit why-questions. Those questions, conveyed in data matrices, assume the truth of observation statements, which is contrary to subsequently invoking substitution rates within inferences to phylogenetic hypotheses. Inferences of explanatory hypotheses are abductive in form, such that some version of an evolutionary theory(ies) is/are included or implied. If rates of sequence evolution are to be considered, it must be done prior to, rather than within abduction, which requires renaming those putatively-shared nucleotides subject to substitution rates. There are, however, no epistemic grounds for renaming characters to accommodate rates, calling into question the legitimacy of causally accounting for sequence data. (shrink)

The aim of this paper is to explore the conditions under which generalisation from single case studies, in the sense of making inferences concerning a wider class of phenomena beyond a case, is reasonable. Two sets of conditions, in particular, provide the basis for our consideration of this issue. The first is an exploration of the impressive amount of empirical knowledge that is contained within the theories that are used to make observations, to classify phenomena, and to understand and interpret (...) cases. The second is recognition of the role of a pattern of inference known as abduction, or inference to the best explanation, in drawing conclusions from case studies. It is argued that these two conditions, when applied to the procedures for conducting case studies, can be used to defend, as reasonable, the making of certain provisional generalisations from single cases. An epistemological procedure for testing and further justifying these generalisations, with a particular emphasis on dealing with the problem of confirmation bias, is then discussed and recommended. (shrink)

Abduction and metaphor are two significant concepts in cognitive science. It is found that the both mental processes are on the basis of certain similarity. The similarity inspires us to seek the answers to the following two questions: Whether there is a common cognitive mechanism behind abduction and metaphor? And if there is, whether this common mechanism could be interpreted within the unified frame of modern intelligence theory? Centering on these two issues, the paper attempts to characterize and interpret the (...) generation and evolution of scientific metaphors from the perspective of the cognitive mechanism of abductive inference. Then it interprets the common cognitive mechanism behind abduction and metaphor within Hawkins’ frame of intelligence theory. The commonality between abduction and metaphor indicates the potential to further explore human intelligence. (shrink)

Scientific explanation is an important goal of scientific practise. Philosophers have proposed a striking diversity of seemingly incompatible accounts of explanation, from deductive-nomological to statistical relevance, unification, pragmatic, causal-mechanical, mechanistic, causal intervention, asymptotic, and model-based accounts. In this dissertation I apply two novel methods to reexamine our evidence about scientific explanation in practise and thereby address the fragmentation of philosophical accounts. I start by collecting a data set of 781 articles from one year of the journal Science. Using automated text (...) mining techniques I measure the frequency and distribution of several groups of philosophically interesting words, such as "explain", "cause", "evidence", "theory", "law", "mechanism", and "model". I show that "explain" words are much more common in scientific writing than in other genres, occurring in roughly half of all articles, and that their use is very often qualified or negated. These results about the use of words complement traditional conceptual analysis. Next I use random samples from the data set to develop a large number of small case studies across a wide range of scientific disciplines. I use a sample of "explain" sentences to develop and defend a new general philosophical account of scientific explanation, and then test my account against a larger set of randomly sampled sentences and abstracts. Five coarse categories can classify the explanans and explananda of my cases: data, entities, kinds, models, and theories. The pair of the categories of the explanans and explanandum indicates the "form" of an explanation. The explain-relation supports counterfactual reasoning about the dependence of qualities of the explanandum on qualities of the explanans. But for each form there is a different "core relation" between explanans and explanandum that supports the explain-relation. Causation, modelling, and argument are the core relations for different forms of scientific explanation between different categories of explanans and explananda. This flexibility allows me to resolve some of the fragmentation in the philosophical literature. I provide empirical evidence to show that my general philosophical account successfully describes a wide range of scientific practise across a large number of scientific disciplines. (shrink)