Without realism, the success of science would be a miracle

The no-miracles argument

Hilary Putnam 1926 · 2016The same mental state can be realised by very different brains: the mind is a function, not a stuff.

See profile

What it claims

Our best scientific theories are approximately true, and the entities they describe probably exist.

Why

The extraordinary success of our theories has only one non-miraculous explanation: they are approximately true.

The same idea through history

J. J. C. Smart1963J. J. C. Smart raises the problem in Philosophy and Scientific Realism: if electrons and other invisible entities did not exist, it would take an extraordinary coincidence for the theories that speak of them to predict the observed phenomena so well. He has the idea, but not yet the slogan.

Hilary Putnam1975Putnam gives the slogan: realism is the only philosophy that does not make the success of science a miracle. He presents it as a scientific argument: realism would be the best explanation of that success.

Richard Boyd1983Richard Boyd develops the argument: the methods that produced our theories themselves depend on approximately true theories, and that is what explains why these methods succeed. Realism thus also explains the reliability of the scientific method.

In short

"Without realism, the success of science would be a miracle" is an argument by Hilary Putnam (Mathematics, Matter and Method, 1975, p. 73 (the "no miracles argument"), 1975) on the question "Can science explain everything?" (answer: true, but not everything). Its claim: Our best scientific theories are approximately true, and the entities they describe probably exist. According to AskPhilosophy's logic check (Z3 solver), the conclusion does follow from its 2 premises. AskPhilosophy presents 3 objections, 2 of them with a documented reply, and the original text behind each premise, checked word for word.

Mathematics, Matter and Method, 1975, p. 73 (the "no miracles argument")

Valid

The theories of physics predict eclipses, guide probes to Mars and made mobile phones possible. How can such success be explained? For Hilary Putnam the simplest answer is that they say roughly the truth about the world, including what cannot be seen with the naked eye (atoms, fields, genes). If they were false from top to bottom, their correct predictions would be an incredible coincidence. That is why he writes in 1975 that realism is "the only philosophy that doesn't make the success of science a miracle". We list it under "True, but not everything" because it defends the (at least approximate) truth of science without saying that science answers every question. Putnam himself later distanced himself from scientific realism as described here; this is the 1975 argument.

click a premise to reject it

If you accept every premise, logic forces you to accept the conclusion. So the debate is about the premises.

Checked by a logic solver: the premises cannot all be true while the conclusion is false.
Without the implicit premise: Without P3, one could accept P1 and P2 and say: "Too bad, perhaps it is a miracle, or a coincidence that science need not explain." The argument thus asks us to refuse sheer coincidence, as we do in everyday life.

The argument is valid. P1 says there is a success to explain. P2 says there are only two ways out: approximate truth or a miracle. P3 rules out the miracle. Approximate truth is left. The debate is about P2 (are these really the only two explanations?) and about P1 (does past success guarantee truth, given theories that succeeded and were then dropped?).

Analogy. A dart player hits the bullseye ten times in a row blindfolded. One can say "what luck!", but the simple explanation is that he knows where the target is. Likewise, a theory that hits the mark ten times in a row probably "knew" something about the world.

Objections

Each objection is a conversation: click to see who replies and where the debate stands.

Larry Laudan 1981 · attacks premise 2

The history of science contradicts the link between success and truth. Laudan lists theories that were very successful, such as those of the optical ether or phlogiston, then were dropped because their central terms referred to nothing real. If past success did not guarantee truth, present success does not either: this is the "pessimistic meta-induction".

a defender replies

John Worrall defender of the argument · 1989

When a theory is replaced, not everything is lost: what is kept is the mathematical structure. Fresnel's equations for light reappear in Maxwell's theory, although Fresnel's ether disappeared. Worrall therefore proposes a "structural" realism: success is explained because theories have the right structure, not because each of their entities exists.

The debate goes on (our summary) Laudan's followers answer that his list shows precisely that the part that produced the successful predictions has often disappeared, and that "structure" is hard to define outside the equations. Realists answer that only mature theories should be counted. See also Kuhn's objection below.

Where the debate stands (our summary)Our summary: an open point. Realists have refined their position (structural realism, "selective" realism) to answer Laudan; his supporters judge that these weakened versions give up what realism originally promised.

Bas van Fraassen 1980 · attacks premise 2

The success of theories is not a miracle, and is not even surprising: it is an effect of selection. Theories live in fierce competition, like living beings in nature, and only those that agree with observed regularities survive. A theory can thus succeed without being true about the unobservable: it is enough that it be adequate to what is observed (constructive empiricism).

a defender replies

Stathis Psillos defender of the argument · 1999

Selection explains why we have theories that succeed, but not why a given theory succeeds. A mouse that escapes the fox is "selected", but the question remains what in its features makes it escape. Likewise, it remains to say what in the content of the theory makes its predictions right. Realism answers: it described something true.

The debate goes on (our summary) Empiricists answer that this demand for explanation goes too far: a theory that agrees with the phenomena needs no further explanation.

Where the debate stands (our summary)Our summary: the core disagreement is about what needs explaining. For van Fraassen agreement with the observable is enough; for realists one must say why theories are adequate. There is no consensus.

Thomas Kuhn 1970 · attacks the conclusion

Kuhn does not see, in the succession of theories, a march toward an ever more faithful description of what really exists. Successive theories solve problems better and better, but the ontology (what there is in the world according to the theory) changes from one revolution to the next with no coherent direction. For him science progresses as puzzle-solving, not necessarily as an approach to truth.

no reply

see the argument

Putnam never answered this objection.

The debate goes on (our summary) Kuhn states that he is not a relativist and that he believes in scientific progress (see Kuhn's argument). The discussion is about the word "true": can a theory be called approximately true if its entities change at every revolution? Realists say yes for what is kept; Kuhn sees no clear link.