The Limitations of Dogmatic Belief: Exploring Kripke’s Dogmatism Paradox
Consider a fact that you know to be true. Over time, you have obtained sufficient evidence to support its validity, so you know any counter-evidence is misleading. But one day, a credible source presents some evidence that sufficiently shakes your conviction. You are now at a crossroads: do you cling to your original conviction, or allow your knowledge to be swayed? This is the problem that Saul Kripke presents in his “Dogmatism Paradox”. It explores an epistemic question that asks us how certain we really are with regards to what we count as “knowledge”. As he puts it, when we know some hypothesis h to be true, we know that any evidence against h is evidence against something that is true. As such, that evidence is misleading, so I ought to disregard it and any future evidence against h. The consequence of being dogmatic, however, is that we dismiss the possibility that our own knowledge is fallible. In this essay, I aim to disprove Kripke’s argument by showing that his conclusion is only true insofar as a sufficiently weighty counter evidence e is absent and not obtained by the subject. In other words, if e is obtained along with the original, opposing evidence E, the subject’s conviction in their belief h will be sufficiently shaken for them to doubt h. As such, the imperative to ignore any counter evidence will be false and irrational.
Kripke’s paradox presupposes the hardiness of knowledge and that any contradictory evidence is misleading; therefore, we ought to dismiss them. While the original formulation of his argument is founded on a closed or deductive system of knowledge, or the principle that if one knows a true fact p, and p entails q, then one also knows q, it is often subject to scrutiny. Therefore, we can start the argument with a less contentious premise that states if we know a certain fact h is true, then we know that any evidence against h is false. Because the evidence is false, they will mislead us from the truth, so all evidence for ~h (not-h) is misleading. Further, if we generally wish to avoid some consequence C that is resulted from a type of action T, then we ought to not take any action of type T. Therefore, we should choose not to be influenced by any evidence for ~h to avoid being deceived.
To illustrate Kripke’s paradox clearer, here is a chain argument for the dogmatic conclusion that my friend Alex has given me a false report on the weather today:
(P1) The weather forecast predicts clear skies today.
(P2) If the weather forecast predicts clear skies and my friend Alex provides evidence that it is raining, then Alex’s evidence is misleading.
(P3) If Alex reports he experienced a heavy downpour on his way to work, then his report is misleading evidence.
(P4) Alex reports that he experienced a heavy downpour on his way to work.
(P5) Alex’s report is misleading evidence.
Let us see how the conclusion is valid by breaking down each premise. Premise 1(P1) is automatically justified by hypothesis. P2 is a certainty because it is analytically true. The argument from P1-P3 is valid because P2 establishes that any evidence Alex provides that contradicts P1 must be misleading, and P3 considers a specific form of evidence that Alex may provide. Therefore, my degree of confidence in P3 must equal my degree of confidence in P1. Since P3 is a conditional, and assuming there is sufficient justification for P4, by modus ponens P5 is true. No matter how hard Alex tries to tell me that it is raining outside, the dogmatism principle gives zero flexibility for me to change my mind.
While Kripke’s argument may be logically sound, we are only justified in believing that “any evidence against h is misleading” insofar as there is no such sufficiently weighty evidence against h. In other words, simply knowing h does not warrant us to disregard any further evidence since getting that new evidence can change what we know. Further, new evidence can make it true that we no longer know that any contradictory evidence is misleading. My refutation can be explained by evaluating two different scenarios at T0 and T1. At T0, a subject A knows hypothesis h to be true because the current evidence E supports h to a degree sufficient for knowledge. With all things considered equal (without evidence for ~h), A concludes that any evidence against h would be misleading. But at T1, A actually obtains a piece of contrary evidence e such that E+e no longer sufficiently supports h being true. From here, the argument naturally breaks down. The basis for A to believe in h is undermined, and A can no longer appeal to h as grounds for disregarding e. So, it can be said that A was never justified in believing the subjunctive condition that, were e to have been obtained, it would be misleading evidence; A only believed h initially because A believed e to be absent.
My counter argument can be better illustrated applied to the weather example. Subject A checks the weather forecast in the morning (T0), which predicts sunny weather for the day. Based on this forecast, A knows that hypothesis h (“It will be sunny today”) is true because the evidence E (morning weather forecast) supports h to a degree sufficient for knowledge. Given this, A concludes that any evidence to the contrary would be misleading, assuming there’s no reason to doubt the accuracy of the forecast. In the afternoon (T1), dark clouds gather, and it begins to rain. A now has a piece of contrary evidence e (sudden rain) sufficiently weighty for doubt, and the combined evidence E+e (initial sunny forecast + sudden rain) no longer supports the hypothesis h being true. A realizes that they were not justified to assume any counter evidence was misleading.
This example shows that the basis for A’s belief that “It will be sunny today” is undermined by the unexpected rain. At T0, A was justified in believing h because all available evidence E supported the sunny day hypothesis. A assumed that any contradictory evidence (like predictions of rain) would be misleading based on the morning’s clear forecast. When e presented itself at T1, it altered the epistemic landscape, showing that A’s belief in h could not rightly dismiss new evidence. The fact that new, contradictory evidence was actually obtained and proved substantive enough to change the forecast demonstrates that A’s dismissal of potential contrary evidence at T0 was unjustified.
To my counter argument, the dogmatist is likely to respond that while we may not be justified in dismissing strong evidence, we are justified to do so against weaker ones. In other words, as long as the contrary evidence is not sufficiently weighty to disprove the hypothesis, we are free to dismiss them as misleading. For instance, we know that a coin is fair because flipping it an infinite number of times will yield a 50/50 chance of it landing either face up or face down. Further, if someone demonstrates to you the first twenty tosses of a coin comes out heads, and argues that that is evidence against the claim that “a coin is fair”, then you know it to be misleading because it is simply a statistical bias. Yet it is dogmatic to ignore this evidence, so the same paradox arises and Kripke’s argument stands.
The dogmatist might rightly dismiss evidence against the immutable truths found in mathematics and statistics, given their absolute and unchanging nature. However, this rationale does not apply to knowledge derived from contingent truths in the broader world, where such dismissal could lead to inconsistencies. Often, having a mix of strong and weak evidence can lead to a bootstrapping effect, where strong evidence overshadows weak counter evidence, reinforcing hypotheses without thorough testing. This was evident in Robert Millikan and Harvey Fletcher’s experiment to measure elementary electric charge using tiny charged oil droplets. They reported only droplets that fell within a predetermined plausible interval, omitting outliers to enhance precision. In 1978, physicist Gerald Holton was shocked to discover significant unreported contrary data in the notebooks of these 1923 Nobel Prize winners (Franklin, 10). By establishing a credible range for the “correct” data, Millikan and Fletcher effectively minimized the impact of misleading evidence, thereby highlighting the most compelling evidence to fit their theory. If all of scientific knowledge and beyond were discovered by discounting the purportedly misleading evidence, we are bound to uncover inconsistencies in the future that could have a significant impact. Ultimately, dogmatism carries the danger of clinging to knowledge and facts that are constantly put to the test in our ever-changing world.
In conclusion, Kripke’s Dogmatism Paradox contains a key conditional premise that is easily overcome when we assume that a sufficiently weighty piece of evidence against the known hypothesis is absent to the subject. Once this evidence is obtained, the subject’s conviction in the original hypothesis is shaken, and it is no longer rational to resolve all contradictory evidence as misleading. The dogmatist may refute this claim by arguing that for some truths, such as those necessary truths in statistics or mathematics, we are justified in dismissing contrary evidence because they are weak. However, their reasoning runs the risk of favoring certain evidence over others, which could lead to forming knowledge and theories that are false by nature. Therefore, to truly resolve Kripke’s Dogmatism Paradox, we must strike a careful balance between discounting too much or too little contrary evidence for any hypothesis we believe in.
References
- Franklin, A. Millikan’s Oil-Drop Experiments. Chem. Educator 2, 1–14 (1997). https://doi.org/10.1007/s00897970102a
