Coherentism.docx

Justin Horky

Epistemology

3rd Paper

Coherentism: An Analysis & Critique In her paper “Non-Foundationalist Epistemology: Holism, Coherence and Tenability” Catherine Elgin presents a theory of the justification of beliefs which argues essentially that beliefs can be justified by fitting together to form a coherent, self-supporting web of beliefs in which all of the various, individual beliefs receive support from one another via mutual corroboration and confirmation. According to this 'coherentist' view of justification, beliefs do not need to be justified by some ’foundational’ belief or beliefs which are themselves self-justifying or justified by something other than other beliefs, nor does a belief have to be justified via a chain of beliefs which extend to infinity and which justify one another in a linear fashion. Rather, a set of beliefs can be justified simply by the manner in which they come together or 'cohere' to form a coherent picture of the world which, through its very coherency, then lends support to the various, individual beliefs which make it up. In this paper, I will first provide an analysis of Elgin’s presentation of coherentism. I will then argue that it fails due to circularity issues involving the notion of coherence itself. Elgin begins her presentation by distinguishing her coherentist theory of justification from a coherentist theory of truth. The former claims that coherence amongst a sufficiently large set of beliefs justifies belief in them, whereas the latter holds that coherence amongst a sufficiently large set of claims makes them true. Elgin rejects the coherentist theory of truth as implausible, but if coherence is not the grounds for truth, that raises the question of why it would be the grounds for justification. If coherence amongst a sufficiently large set of propositions does not guarantee the truth of these

propositions, why would it be the case that coherence amongst them justifies belief in them? Elgin’s response to this appears to essentially be an invocation of the principle of inference to the best explanation. The reason for why coherence provides justification for belief in propositions is because coherence tracks the truth; and coherence tracks the truth due to the fact that when a large set of propositions cohere with one another by confirming and corroborating one another, the best explanation for why this is so is because these propositions are providing an accurate account of reality. Hence, belief in a sufficiently large set of beliefs which cohere with one another in an appropriate manner is justified because the truth of these propositions is the most economical means of accounting for the manner in which they come together to form a coherent web. Elgin provides us with a nice example to illustrate the manner in which this works. Suppose someone steals an object and various different witnesses, who have never communicated with one another, claimed to have witnessed the theft. Furthermore, they all provide accounts of the theft and descriptions of the thief which share various details with one another. Now, even though each of the witnesses on their own is unreliable for various reasons (one has poor vision; another one is a pathological liar, etc.) the fact that their testimonies support one another, and the fact that they have never been in contact with one another and therefore could not have formed a conspiracy, gives us great reason to suspect that each of them is telling the truth and that the accounts they have provided describe events as they actually occurred. In other words, although each of these testimonials on its own is nearly worthless, together they generate a macro-account which provides support for each of the individual accounts that make it up, such that the macro-level account and the individual, lower-level accounts which make it up are

supporting one another. The lower-level accounts support the macro-level account by constituting it, and the macro-level account in turn supports its constituents. But why does the macro-level account provide such support for its constituents? Because, given the manner in which these different accounts corroborate one another, and given the fact that a conspiracy has been ruled out, the best explanation to account for this interlocking of corroboration or this 'web of coherence' is that these accounts, despite being individually unconvincing, nevertheless really do explain what happened. It is in this manner that coherence amongst beliefs in general provides justification for them. Elgin summarizes her position as follows: "The thesis of the sort of epistemological holism that I want to consider is that epistemic justification is primarily a property of a suitably comprehensive, coherent account, when the best explanation of coherence is that the account is at least roughly true" (p. 246). It is important to emphasize this connection between coherence and inference to the best explanation because it is only when the best explanation for the coherence of an account is the truth of that account that we are justified in believing the individual propositions which make it up. If there is an alternative explanation which accounts for the coherence of an account, such as a writer's skill with regards to the account put forward in a well-crafted novel or a person subconsciously ignoring various sets of data which conflict with a coherent account that he is trying to create, then coherence no longer justifies belief in the truth of this account via the principle of inference to the best explanation because the truth of the account is no longer the best explanation for its coherency. Hence, by tying coherence to the principle of best explanation, Elgin is able to guard against the danger of sets of propositions which are coherent for reasons other than truth. The claim that coherence alone justifies belief in

propositions could require belief in elaborate and coherent accounts which have no connection to reality at all. By adding in that belief in an account is not warranted unless the truth of that account best explains why the propositions that make up that account corroborate one another, then various counter-examples where coherency clearly has nothing to do with truth and thus does not justify belief are weeded out. In summary, coherence within a body of propositions is not a sufficient condition to warrant belief in those propositions. What accounts for this coherency must also be taken into account or, as Elgin puts it, "[C]oherence conduces to epistemic acceptability only when the best explanation of the coherence of a constellation of claims is that they are (at least roughly) true"(p. 248).i With this basic framework laid out, Elgin then spends the rest of the paper responding to objections and laying out the implications and benefits of her position. One of the more fascinating issues which Elgin discusses concerns the relation of our beliefs to our perceptions and how this relationship conflicts or accords with coherentism. Suppose that you are reading a novel which has been written from the second-person point of view, i.e. one which purports to describe your life. This 'biography,' although described in a very engrossing manner, fails to capture your belief because it conflicts with how you believe you life has really unfolded. But what justifies the belief in one account of your life over another? Is it really because the propositions which make up the account given of you in the novel fail to cohere with another, much larger and more coherent web of propositions which provides an alternative account of your life that you are justified in disbelieving the first set of propositions while believing in your other biography? Or is the real source of justification the fact that your senses fail to convey to you perceptions which conform to the 'novel-biography' while they do

convey to you perceptions which match the other web of propositions, such that the account of your life that is corroborated by perception would enjoy justification in belief over the novel-account, even if it were in fact less coherent? Is it really coherence which provides justification for belief in various propositions, or is it instead some version of perceptual foundationalism or perceptual reliabilism? The problem which Elgin is trying to get at here seems to be this. We have a belief that perceptions should carry greater weight when it comes determining our beliefs and that, no matter how elaborate and coherent a theory or account is, that account should have no ability to override beliefs which are corroborated by our direct perceptions. Hence, beliefs generated and corroborated by perceptions, or 'perceptual deliverances' as Elgin calls them, enjoy a primacy within our system of beliefs. But is this primacy really compatible with the claim of coherentism that belief in the truth of a proposition is justified only by that belief's coherence with other beliefs? If a coherent system of beliefs about the world, a theory of nature, can be rejected by means of the fact that that theory fails to conform to observation, then it would seem to be the case that it is perception itself which provides the justification for believing some propositions while rejecting others, either because perception is a reliable belief-formation process or because belief in propositions can be directly justified by perception. Elgin's interesting response to this objection to coherentism is to point out that, although it is true that perceptual deliverances carry more weight than other kinds of beliefs, this does not mean that such deliverances are immune to revision due to coherence considerations. Elgin begins by conceding that beliefs or propositions which are the result of perceptual experiences cannot be ignored, even if they do contradict an

elaborate and coherent network of beliefs because "[h]owever tightly woven an empirical account may be, we would be epistemically irresponsible to ignore recalcitrant evidence” (p. 250). However, it is also clear that we do not always fully trust our observations but rather, we oftentimes discount or re-interpret them in order to make them cohere with other beliefs. Suppose for example that you believed that you had seen something but that a group of people who were with you and whom you trust tell you that you are mistaken. Your response to this situation would not be to simply ignore the testimonies of your friends. Rather you would begin to attempt to find ways to force your perceptual deliverance into conformity with your beliefs that your friends are honest people and that a group of observers is more reliable than a single observer. You could do this by blaming the lighting or questioning your eyesight. Perhaps you would even begin to doubt your sanity. Regardless, it is clear that perceptions do not enjoy a complete right of unrevisability with regards to relating to other beliefs. Sometimes, our perceptual deliverances force us to change other beliefs that we have, but sometimes the flow of change works in the opposite direction as well. The complex manner in which empirical observation can relate to theory can be seen in the way that science oftentimes works nowadays as well. Oftentimes it is the case that a scientific theory leads scientist to believe in a phenomenon or entity whose existence is posited by that theory in order to explain various other phenomena, even though they have no actual empirical corroboration for the existence of this phenomenon or entity (cf. the Higgs-boson particle until recently). Conversely, if an empirical observation conflicts with a well established theory, the first action of the scientists performing the observation will not be to re-write or trash the theory in question, but

rather to hope and pray that the equipment is merely faulty. Perceptual deliverances and empirical testing do not always get the last word. Both can be and often are revised to force them to cohere with a wider network of beliefs which has greater weight. Furthermore, Elgin also argues that the reason for why we give perceptual deliverances such a great weight is precisely because the beliefs born of such deliverances really do cohere together. Were it the case that those beliefs did not cohere with one another, than we would be much less willing to trust our senses and give such weight to the beliefs directly inspired by them and would instead focus upon other sources of knowledge. Although Elgin’s theory of coherentism is quite fascinating, especially as it relates to the manner in which modern, scientific investigations are carried out, I must, unfortunately, ultimately judge it to be a failure due to circularity concerns. The problem of circularity arises for Elgin because it is simultaneously the case that Elgin on the one hand, by the very hypothesis of her theory, needs to argue that all of our beliefs, including even belief in the principle of non-contradiction, are justified due to the manner in which they cohere with one another, and, on the other hand, she also needs to use the principle of non-contradiction as one of the standards of assessment when it comes to determining whether or not belief in a proposition is acceptable because that proposition does or does not cohere with other propositions in our web of beliefs. As a result, it is not possible for Elgin to justify belief in the principle of non-contradiction itself. But because her whole coherentist theory relies upon this principle, without it, her system falls apart. Let us begin by reminding ourselves that Elgin’s central claim is that “epistemic justification is primarily a property of a suitably comprehensive, coherent account, when the best explanation of coherence is that the account is at least roughly true" (p. 246).

Elgin does not lay out a full account of what it means for different propositions to be 'coherent' with one another, but clearly, at least one of the necessary conditions for two propositions to be coherent with one another is that they do not contradict one another. Elgin herself clearly recognizes and accepts this: "[i]nconsistencies among beliefs conclusively demonstrate that some of the beliefs are false" (p. 252). Hence, if I have one belief that coheres with my larger system of beliefs, that belief is justified, whereas if I have another belief which contradicts this system, then my belief in that proposition is not justified. Hence, as a result, my set of justified beliefs will consist only of propositions which are not contradictory with one another because the principle of non-contradiction is one of the criteria which I use for ruling out beliefs as unjustified. But whence comes the justification for my belief in the principle of non-contradiction? Per coherentism, it must be the case that my belief in this principle is justified by way of the fact that it coheres with my overall system of beliefs, whose coherence is best explained by the truth of the propositions of that system. And the reason for why the proposition stating that the principle of non-contradiction is true coheres so nicely with this system is because there are no counter-examples to that principle within this system. However, it should be immediately clear that the explanation for why that is the case is entirely trivial in nature: there are no contradictions within this system of beliefs because the principle of non-contradiction has already been used to rule out as false all contender propositions for entry into this system which have been excluded because they do not cohere with it. Hence, the principle of non-contradiction coheres with this system because it has excluded from this system the possibility of there being present any two propositions which would contradict one another, and which would

therefore contradict the principle of non-contradiction. Here is an alternative phrasing of this admittedly difficult point. Elgin wants to argue that coherence amongst different beliefs provides justification for them. At a minimum, coherence requires that these beliefs do not contradict one another. Hence, if I have a variety of beliefs, and some of them contradict one another, then I am not justified in believing those beliefs which are outside of the larger web of beliefs that I have which these beliefs are contradictory with. As such, I can do one of two things. Either I can discount those beliefs in question, in which case I will be left over with a web of coherent beliefs. But then, even though the principle of non-contradiction does cohere with these beliefs, how can I justify my belief in that principle on this basis, given the fact that, via the principle of justification via coherence, I have already used the principle of non-contradiction to rule out all sets of incompatible beliefs which could serve as a counter-example to it? Of course the principle of non-contradiction coheres with my web of beliefs, because I have already used that principle to rule out all contradictory belief sets which could serve as counter-examples to it, i.e. with which it could not cohere. Or, on the other hand, I could not rule out some of my beliefs merely on the basis that they disagree with some of my other beliefs, in which case I do have contradictory sets of beliefs with which my belief in non-contradiction does not cohere. The problem of circularity here is that one cannot justify a principle by using that principle to rule out all of its counter-examples. But this is precisely what Elgin does vis-à-vis the principle of non-contradiction. Hence, she is not able to justify this principle at all. But if she cannot justify it, then she has no basis for the claim that coherence is what justifies belief because coherence no longer tracks the truth because it is now possible for contradictory

propositions to be true. Or, if the principle can be justified, it is for reasons other than coherence, which means that the central claim of coherentism is false. Hence, although Elgin's theory of justification is very fascinating, ultimately, it must be judged a failure. i

What prevents Elgin's view from simply collapsing into a version of the argument that the principle of best explanation itself is simply what justifies our various beliefs, such as the belief in an external, material world? Unlike Jonathan Vogel who argues that our belief in an external world is justified when we use the principle of best explanation to posit the existence of such a world in order to account for our perceptions of such a world, it seems to be Elgin's belief that we do not actually have to perform this inference to arrive at a justified belief. Rather, we can arrive at a justified belief merely by noting the coherency that a belief of ours has with our various other beliefs. So long as the truth of the overall account which is made up of these beliefs is the best explanation for the coherency of this account, our belief in this individual belief in question is justified, regardless of whether or not we actually go from the perceived coherency of this belief with our other beliefs to the additional inference that the best explanation for the coherency of this overall system of beliefs is the truth of the propositions being believed.