“Almost all scientific theories of the past have proven false.
So almost all present-day theories are likely to be false too.”
Discuss.
The quoted argument in the essay title is known as the “Pessimistic Induction” (or the Pessimistic Meta-Induction) argument for Scientific Anti-Realism (“PI” for short). It observes that almost all scientific theories in the past we now currently regard as false. And in some well-known cases, the newer theories have completely “changed the world” by replacing how we view the world. The best known example of this is the Copernican theory replacing the Earth Centered view of the solar system with a Sun-Centered view. Another example in the modern literature on PI is the replacing of the Caloric Theory of heat with the Thermodynamic Theory of heat(1). The inductive generalization based on this history of “false” theories is that all our current scientific theories will eventually be regarded as false by future generations. Since this evidence suggests that this is likely to happen, the argument goes, we have no reason to believe that our current theories are true.
The context within which this argument takes place is the debate between the Scientific Realist who maintains that science aims at truth, and that most of our best scientific theories are approximately true, and the Scientific Anti-Realist who disputes that claim. The Scientific Anti-Realist employs this PI argument in a challenge to the naïve scientific realism’s claim that our best scientific theories are true descriptions of a mind-independent world. The anti-realist claims, with this argument, that since our best scientific theories are not likely to be true, we have no reason to view them as true descriptions of the world. Hence Scientific Realism is false.
In recent years, the debate has been between the “No Miracle” (NM for short) argument put forth by Hilary Putnam(2) in support of Scientific Realism, and the version of the PI argument put forth by Larry Laudan(3) in challenge.
Hilary Putnam said “Realism is the only philosophy that does not make the success of science a miracle”. According to this argument, the predictive success of science is best explained only if we assume that our mature scientific theories are approximately true. This “inference to the best explanation” defence of scientific realism has been developed by numerous philosophers into a systematic defense of scientific realism.(4) But the NM argument depends on the assumption that there is some kind of connection between a theory’s predictive success and the theory’s being approximately true. The PI argument challenges this assumption, and argues that the NM argument ignores the evidence from the history of science. Laudan’s version of this argument can be reconstructed as follows(5):
(1) Assume that the success of a theory is a reliable test for its truth.
(2) Most current scientific theories are successful.
(3) So most current scientific theories are true. (The NM argument.)
(4) Then most past scientific theories are false, since they differ from current theories in significant ways.
(5) Many of these false past theories were successful.
(6) So the success of a theory is not a reliable test for its truth.
This is a reductio ad absurdum argument that demonstrates that the basic premise of the NM argument — premise (1) — cannot be supported. Hence, so runs the conclusion, Scientific Realism is false.
Although there has been some debate over whether there might be logical fallacies in this version of the argument(6), most of the Realist rebuttal to the argument has focused on premise (4). The approach has been to segregate theories into things that are still true and things that are admittedly false, and claim that the stuff that is still true is all that is needed to justify premise (1), while stuff that is admittedly false is no threat to that premise(7).
But there is another approach to rebutting the PI argument, and that is to focus on the concepts of “True” and “False” that are being employed. The anti-realist employment of the PI argument depends on a binary nature of the true-false distinction drawn from deductive reasoning. The implication is that if a scientific theory is not true, then it must necessarily be false. But this need not be accepted by the scientific realist. And, in fact, when it comes to inductive reasoning, the binary distinction imported from deductive reasoning can be argued is completely out of place.
What does it mean for a theory to be “true”. There are three different models of “truth” that can be employed when discussing the PI argument —
(a) Truth is ultimately an epistemic notion (also known as anti-realism or non-realism — truth is determined by the evidence; includes Conventionalism and Instrumentalism). This is the notion of truth employed by the anti-realist when using the PI argument to demonstrate that success is not a mark of truth, and that the NM argument fails.
(b) Truth is ultimately a metaphysical notion (also known as realism or evidence transcendence — no matter how much evidence we have in support of our inference, it remains logically possible that the hypothesis is false). This is the notion of truth employed by many realists when they argue that the evidence cited by the anti-realists does not refute premise (1) for various reasons having to do with segregating the parts that are “false” from the parts that are “true”.
(c) Truth is ultimately a pragmatic notion (also known as pragmatism — truth is determined by the utility of the consequences). This is the notion of truth that I will argue should be applied in any context of inductive reasoning.
Consider a pragmatist notion of “true-enough” — where an hypothesis (or a scientific theory) is “true-enough” if and only if it is empirically successful (results in predictions and deductive entailments that correctly describe our empirical experiences) within the accuracy of our empirical requirements. Van Fraassen comes very close to this pragmatic notion of truth with his “empirical adequacy”(8). But he does not recognize that the pragmatic notion he defines of “empirical adequacy” as the proper concept of “truth” for an inductive reasoning context.
Obviously, this pragmatic notion of “truth” is relative — relative to our current pragmatic purposes. That the length of this table is 5 feet (or 1.5 meters) is true-enough if I am ordering a table cloth. But not nearly true-enough if I am cutting a piece of glass to top the table. Newton’s theory of Gravity is true-enough if we are plotting a trip to the Moon. It is not true-enough if we are trying to predict the precession of the orbit of Mercury. Even the “theory” that the Earth is the center of the Universe is true-enough if we are doing celestial navigation. But it is obviously not true-enough if our purpose is to understand our place in the Universe. That the tracks in the snow were the result of a passing tiger is true-enough if our purpose is to provide a plausible explanation for our seeing tracks. It may not be true-enough if we are tracking our lunch and the cost of error is becoming lunch rather than enjoying lunch.
In other words, premise (1) above becomes true by definition — empirical success of a scientific theory is the definition of “true-enough”, not merely a reliable test for truth. Because the concept of true-enough is relative to our pragmatic purposes, the opposite of true-enough is “not sufficiently true-enough for our current purpose”. And if this is the understanding invested in the label “false”, then it shows that premise (5) above commits the fallacy of temporal equivocation. Either we read (5) as
(5a) Many of those historical theories that we now consider false according to our current pragmatic purposes, we now considered not-successful according to our current pragmatic purposes; or
(5b) Many of those historical theories we now consider false according to our current pragmatic purposes, we now consider successful according to our current pragmatic purposes (which is patently false); or
(5c) Many of those historical theories that we now consider false according to our current pragmatic purposes, we then considered successful according to our then pragmatic purposes.
None of these readings justifies the conclusion that the success (according to our current pragmatic purposes) of a theory is not a reliable test for its being regarded as true-enough given those pragmatic purposes. Only by equivocating on the temporal sense of the governing pragmatic purposes can we read the premise as supporting the conclusion (premise (6) above) that the success (according to our current pragmatic purposes) of a theory is not a reliable test for its truth (in a true-enough sense, given those same current pragmatic purposes).
Further, this pragmatic understanding of “false” renders the PI argument fallacious in the manner described by both Peter Lewis(9) and Marc Lange(10). Understanding the temporal relativity of “success” in the premise that the success of a theory is a reliable test of its truth, means that Saatsi’s(11) “solution” to Lewis’s “false positive fallacy” and Lange’s “turn-over fallacy” cannot be applied. Saasti’s solution was to look at the population of scientific theories in a timeless manner — focusing on the conclusion that empirical success is not a reliable test for truth, not that our current theories are likely to be false. But with the pragmatic notion of true-enough, this won’t work. What counts as a “success” is relative to our pragmatic purposes at that time.
Empirical adequacy — “success” — is the definition of true-enough. It does not matter whether “true-enough” is the mark of “truth” (whatever that word might designate in an inductive reasoning context). It does not matter if the “truth” is that not all swans are white if the evidence upon which we draw our inductive conclusions support the inductive conclusion that “all swans are white”. For “truth” in this conception does not matter because the only grasp we have on the nature of Reality, and where the tiger is likely to be lurking, is the inductive inferences we draw from the evidence in hand. We may be wrong — inductive inference is not a guarantee of truth. But all that matters is that our currently best scientific theories are sufficiently true-enough at predicting our Universe that they can be reliably counted on as close enough to “true” for our current purposes.
In other words, the quoted argument in the essay’s title may be true (on a pragmatic understanding of true/false), but it is true only tautologically. It does not provide the challenge that anti-realists hope for against Scientific Realism.
Notes & References
(1) Psillos, Stathis; “A Philosophical Study of the Transition from the Caloric Theory of Heat to Thermodynamics: Resisting the Pessimistic Meta-Induction” in Studies in the History of the Philosophy of Science. Vol. 25, No. 2. (1994) pp. 159-190.
(2) Putnam, H.; Mathematics, Matter and Method: Philosophical Papers, Vol 1, London: Cambridge University Press. London, England, 1975. pg 73.
(3) Laudan, L.; “A Confutation of Convergent Realism” in Philosophy of Science, Vol 48 (1981), Pgs 19-49.
(4) see for example –
Boyd, R.; “The Current Status of the Realism Debate” in Scientific Realism, J. Leplin (ed.), University of California Press, Berkeley, California. 1984, pg. 41-82.
Devitt, M.; Realism and Truth, Blackwell, Oxford, England. 1984.
(5) Lewis, Peter, J.; “Why the Pessimistic Induction Is a Fallacy” in Synthese, Vol. 129, No. 3 (Dec., 2001), pp. 371-380. URL=http://www.jstor.org/stable/20117188
(6) See for example —
Lewis, Peter, J. (2001)
Lange, Marc; “Baseball, Pessimistic Induction and the Turnover Fallacy” in Analysis, Vol 62, No 4 (2002), pp 281-285, URL=http://www.jstor.org/stable/3328917
Saatsi, Juha T.; “On the Pessimistic Induction and Two Fallacies” in Philosophy of Science, Vol. 72, No. 5, Proceedings of the 2004 Biennial Meeting of The Philosophy of Science Association, Part I: Contributed Papers, Miriam Solomon (ed.), (December 2005), pp. 1088-1098, URL=http://www.jstor.org/stable/10.1086/508959
(7) Chakravartty, Anjan; “Scientific Realism” in The Stanford Encyclopedia of Philosophy, Edward N. Zalta (ed.), URL=http://plato.stanford.edu/archives/sum2011/entries/scientific-realism/
(8) Monton, Bradley and Mohler, Chad; “Constructive Empiricism” in The Stanford Encyclopedia of Philosophy, Edward N. Zalta (ed.), URL=http://plato.stanford.edu/archives/win2008/entries/constructive-empiricism/
(9) Lewis, Peter, J. (2001)
(10) Lange, Marc, (2002)
(11) Saatsi, Juha T. (2005)
Ladyman, James; “Structural Realism” in The Stanford Encyclopedia of Philosophy, Edward N. Zalta (ed.), URL=http://plato.stanford.edu/archives/sum2009/entries/structural-realism/
Musgrave, Alan; “The ‘Miracle Argument’ for Scientific Realism” in The Rutherford Journal, URL=http://www.rutherfordjournal.org/article020108.html
Singer, Daniel J.; “The No No-Miracle-Argument Argument” in the University of Pennsylvania College Research Electronic Journal, 2007. URL=http://repository.upenn.edu/curej/62
Psillos, Stathis; “Scientific Realism and the ‘Pessimistic Induction'” in Philosophy of Science, Vol. 63, Supplement. Proceedings of the 1996 Biennial Meetings of the Philosophy of Science Association. Part I: Contributed Papers (Sep., 1996), pp. S306-S314. URL=http://www.jstor.org/stable/188540