Editing Logical positivism (section)

===Scientific explanation===

{{See also|Deductive-nomological model}}

[[Carl Hempel]] was prominent in the development of the [[deductive-nomological model|deductive-nomological]] (DN) model, then the foremost model of [[models of scientific inquiry|scientific explanation]] defended even among critics of neo-positivism such as [[Karl Popper|Popper]].<ref name=SEPWoodward>{{cite encyclopedia |first=James |last=Woodward |url=https://plato.stanford.edu/archives/win2011/entries/scientific-explanation |title=Scientific explanation |editor=Edward N. Zalta |encyclopedia=The Stanford Encyclopedia of Philosophy |edition=Winter 2011}}</ref> According to the DN model, a scientific explanation is valid only if it takes the form of a [[deductive reasoning|deductive inference]] from a set of explanatory [[premise]]s (''explanans'') to the observation or theory to be explained (''explanandum'').<ref name=Suppe>{{cite book |last=Frederick Suppe |first=Frederick |title=The Structure of Scientific Theories |publisher=University of Illinois Press |year=1977 |url=https://books.google.com/books?id=SpvZsxCA0TIC&pg=PA619 |edition=2nd |pages=619–21}}</ref> The model stipulates that the premises must refer to at least one [[scientific law|law]], which it defines as an [[enumerative induction|unrestricted generalization]] of the [[conditional sentence|conditional]] form: "If ''A'', then ''B''".<ref>{{cite book |first=Eleonora |last=Montuschi |title=Objects in Social Science |publisher=Continuum |year=2003 |url=https://books.google.com/books?id=bQ24-BV8WSAC&pg=PA62 |pages=61–62}}</ref> Laws therefore differ from mere ''regularities'' ("George always carries only $1 bills in his wallet") which do not necessarily support [[counterfactual]] claims.<ref>{{cite book| last=Bechtel |first=William |title=Philosophy of Science: An Overview for Cognitive Science |location=Hillsdale NJ |publisher=Lawrence Erlbaum Assoc. |year=1988 |pages=25-28}}</ref> Furthermore, laws must be empirically verifiable in compliance with the verification principle.<ref name=Suppe/>

The DN model ignores causal mechanisms beyond the principle of [[constant conjunction]] ("first event ''A'' and then always event ''B''") in accordance with the [[David Hume|Humean]] [[empiricism|empiricist]] postulate that, though sequences of events are observable, the underpinning [[causality|causal principles]] are not.<ref name=SEPWoodward/> Hempel stated that well-formulated natural laws (empirically confirmed regularities) are satisfactory in approximating causal explanation.<ref name=Suppe/>

Hempel later proposed a probabilistic model of scientific explanation: The inductive-statistical (IS) model. Derivation of statistical laws from other statistical laws would further be designated as the deductive-statistical (DS) model. The DN and IS models are collectively referred to as the "covering law model" or "subsumption theory", the latter referring to the movement's stated goals of "theory reduction".<ref name=Suppe/><ref>{{cite book |first=Manfred |last=Riedel |url=https://books.google.com/books?id=It3ji_AuO3sC&dq=Covering+subsumption&pg=PA3 |chapter=Causal and Historical Explanation |pages=3–4 |editor1=Manninen J |editor2=Tuomela R. |title=Essays on Explanation and Understanding: Studies in the Foundation of Humanities and Social Sciences |location=Dordrecht |publisher=D. Reidel Publishing |year=1976}}</ref>