The Equal Weight View
By Yingcan S
How should we react to disagreements? How would our dissenters’ opinions affect our own opinions? The equal weight view, a widely discussed view on the epistemic significance of disagreement, tells us that if our dissenter is our “epistemic peer,” we should probably suspend our judgement on the disputed issue.
We disagree with each other all the time. The equal weight view concerning how we should react to disagreements over some questions goes as follows: we should give the same weight to our own opinions as we give to the opinions of those we count as our epistemic peers, who are equally familiar with the relevant evidence and arguments to those questions, and share with us equal epistemic virtues, such as intelligence, freedom from bias, etc. To illustrate the application of the equal weight view, consider the following horse race case. Suppose that you and Bob are watching a horse race. While the race is approaching its end, you are paying close attention to the race and conclude that the white horse will win; Bob is too paying close attention to the race but concludes that the black horse will win. According to the equal weight view, once you are aware of the disagreement between you and Bob, it would be rational for you to give the same weight to your own opinion as you give to Bob’s opinion. Crucially, you would consider Bob as your “epistemic peer”: you are both familiar with horse races and have equally good visual perception; Bob is equally intelligent and attentive as you, and thereby is equally good as you at assessing the very same evidence—the ongoing horse race. Bob’s opposing opinion that the black horse will win would therefore give your reasons to think you and Bob are equally likely to have made a mistake. As a result, you should suspend your judgement as to whether the white horse will win until the result is announced.
In his defense of the equal weight view, Adam Elga gives a bootstrapping argument which suggests that the denial of the equal weight view would lead to absurd consequences. Consider again the horse race case. Before disagreeing with Bob over which horse will win, you think that you and Bob are epistemic peers who are no more likely than each other to make a mistake. If you take the equal weight view to be false, however, after the disagreement, you could have the biased opinion that Bob is more likely than you are to have made a mistake; you would then be thinking that you are a better judge on the result than Bob is. It then seems that the mere disagreement between you and Bob would give you evidence for the claim that you are a better judge on the horse race result than Bob is, regardless of who is in fact the better judge. But this is absurd. To avoid such an absurd consequence, the equal weight view must be true (19).
However, the equal weight view seems to entail two serious problems, namely, the problem of global agnosticism and the problem of lack of self-trust. I will explain these two problems in turn.
The problem of global agnosticism claims that if you support the equal weight view, you would then suspend judgements on pretty much everything. The equal weight view seems plausible in ordinary cases such as the horse race case. After all, you would no longer need to suspend the judgement as to whether the white horse wins when the result is announced; you would then either retain or give up your belief that the white horse will win. By contrast, the disagreements between you and people whom you might view as your epistemic peers in some controversial areas, such as ethics, politics, etc., seem to be persistent and insolvable by looking at some later evidence. If you give equal weight to your opinion as you give to the opinions of those you consider as epistemic peers whom you disagree with, you would then suspend all your judgements regarding controversial issues and thus have no beliefs about them.
The problem of lack of self-trust claims that if the equal weight view is true, then there is no point in investigating some controversial issues for yourself. For example, to make up your mind about whether abortion is permissible, all you need to do is to take a popularity poll among your epistemic peers and believe whatever the majority of them believes. It then seems that your own evaluation of the relevant evidence to such controversial issues would play a minor role in your decision making.
In his response to the above two problems that challenge the applicability of the equal weight view to controversial cases like the ethical debate over whether abortion is permissible,
Elga gives an example involving Ann and Beth, who are on opposite sides of the political spectrum. Ann thinks that abortion is permissible while Beth does not. According to Elga, setting aside her own reasoning in support of abortion and Beth’s opposing opinion, Ann would not consider Beth as her epistemic peer. Crucially, Ann would know that since Beth is on the opposite side of the political spectrum from hers, Beth would have radically different answers from her about a wide range of questions that are closely related to the question of whether abortion is permissible (e.g. whether fetuses can feel pain). By Ann’s light, Beth would be wrong about most of these closely related questions and therefore unlikely to be right in their disagreement over whether abortion is permissible. According to Elga, when people disagree with one another over some controversial issues, it is unlikely that these people would regard one another as an epistemic peer—setting aside their own reasoning about the question at dispute and their dissenters’ opinions, these people would not think one another as equally likely to get the truth about a host of closely related issues to the question. Since the problem of global agnosticism only poses a challenge to the equal weight view under the context where dissenters count one another as their epistemic peers in judging controversial issues, the threat of the problem to the equal weight view is removed. Similarly, for the problem of lack of self-trust, people whom one considers as epistemic peers in judging controversial issues would share similar views to his own, which would mitigate the “just-take-a-poll” worry by allowing one to maintain his stance while dismissing the opinions of his dissenters whom he is unlikely to consider as his epistemic peers (21-23).
However, Elga’s attempt to avoid the the problem of global agnosticism and the problem of lack of self-trust by noting that we rarely consider our dissenters as our epistemic peers have some limitations. First, it is hard to see how Elga’s test of epistemic peerhood would be justified. Consider again Ann and Beth case. It seems that setting her own reasoning in support of abortion and Beth’s opposing opinion, Ann would have no reasons for thinking her answers to the related questions is right while Beth’s answers to these questions are wrong. It seems that Elga therefore makes a leap from a situation where we cannot tell who is the better judge about the controversial issue, to the conclusion that we should remain confident about our own view. Second, even granted that Elga’s test of epistemic peerhood can be justified, epistemic peerhood on Elga’s picture would be hardly attainable in nearly all cases of disagreements that deeply matters to us. Almost all ethical disagreements, for example, inevitably involve disputes over a wide range of related questions that would lead the subjects to discount one another as epistemic peers. An undesirable upshot is that the subjects would only listen to people who already agree with them while leaving the opposing opinions unconsidered.
We have seen that Elga’s maneuver over the condition of attaining epistemic peerhood rarely shields the equal weight view from the challenges posed by the problem of global agnosticism and the problems of lack of self-trust. To defend the equal weight view, I propose a different approach to the two problems. I will devote the remainder of my essay arguing that when treated in the framework of credence, the two problems are not supposed to pose a central challenge to the equal weight view.
Recall that the equal weight view says that we should give the same weight to our own opinions as we give to the opposing opinions of our epistemic peers. Note that the term “opinions” in the equal weight view is vague between two attitudinal states, namely, the belief state and the credence state. Credences, or our degrees of confidence about a proposition, are more fine-grained attitudes than beliefs. In the belief framework, there are only three kinds of attitude that we can take toward a proposition—we can believe it, disbelieve it, or suspend judgement on it. By contrast, in the credence framework, we can assign to our attitudes toward a proposition any values in a scale between 0 and 1, where 0 represents minimum confidence that the proposition is true and 1 represents maximum confidence that the proposition is true. When we consider the equal weight view in the belief framework, given peer disagreement, it does seem that we should suspend judgement on all controversial questions such as whether abortion is permissible. However, it does not mean that we cannot thereby have any credence about some claims such as abortion is permissible.
My interpretation of the equal weight view concerning how we should update our credence toward some proposition in the face of peer disagreement goes as follows: to give the same weight to our own opinion as we give to the opposing opinion of our epistemic peer, we should average our credence and our epistemic peer’s credence about a certain proposition, and replace our previous credence about that proposition with the resulting credence. To illustrate, consider Ann and Beth case in the credence framework. Since Ann thinks that abortion is permissible, she is likely to have a high credence that abortion is permissible. Beth thinks that abortion is not permissible, so she is likely to have a low credence that abortion is permissible. Suppose that Ann has 0.8 credence while Beth has 0.2 credence that abortion is permissible. Upon learning Beth’s credence about the claim that abortion is permissible, since Ann would consider Beth as her epistemic peer in a rough sense, i.e. Beth is about as intelligent as her and shares similar evidence with her, Ann should then change her credence from 0.8 to 0.5 about the claim that abortion is permissible.
To generalise my interpretation of the equal weight view in the credence framework, although in the belief framework, the application of the equal weight view would require us to suspend judgements on almost all controversial propositions (e.g. that abortion is permissible), it would allow us to maintain at least some credence about these propositions. This would mitigate the skeptical implication of the problem of global agnosticism, which says that we would have no beliefs about all controversial issues. With respect to the problem of lack of self-trust, note that it is not the case that our own evaluation of controversial issues does not matter in our decision making. In extreme cases where I have high credence that abortion is permissible but my 99 epistemic peers have 0 credence that abortion is permissible, after averaging our total credence, I would still have at least some credence that abortion is permissible. That is, the significance of our own evaluation about controversial questions would still be reflected in the credence framework. An objection might go that it seems impossible for us to have a precise credence about controversial propositions, or to know our epistemic peers’ credence about these propositions. Although I am not in a position to reply to this objection, I think the recognition of how the equal weight view might play out in the credence framework would suffice to mitigate the problem of global agnosticism and the problem of lack of self-trust, so that the two problems would no longer pose a central challenge to the equal weight view. Therefore, the equal weight view is still the correct view that one should hold towards disagreements.
Work Cited
Adam Elga, ‘Reflection and Disagreement’, Noûs, 41 (2007), pp. 478–502
