Book Review: Epistemic Justification by Richard Swinburne

In Epistemic Justification (henceforth EJ) Richard Swinburne wants to answer two basic questions: first, what is justification, and second, what types of justification are worth having. While easy to state, the questions are very difficult to answer. This difficulty stems from several factors, including the history of epistemology, the failure to make distinctions, and the connections between justification, warrant, and knowledge.

The History of Epistemology

The basic point here is that justification (and ‘warrant’) have not been subjected to as much philosophical analysis as has ‘knowledge’. Thus what is often seen is that various models will agree on when knowledge is (not) obtained, but disagree on when justification and warrant are (not) attained.

Making Distinctions

Externalism vs. internalism. The first distinction that Swinburne taps is a familiar one, but one that has important implications for what terms like justification, warrant, and knowledge mean. This distinction turns on externalism vs. internalism. Just how large this distinction is can be seen by pondering how justification differs in the two[1] approaches. While the definition Swinburne proposes for justification is superficially the same for both types (i.e., justification is defined as a belief’s being rendered probable by being based on adequate grounds), the Devil is truly in the details. Almost all of the central terms in the definition just given (‘probable’, ‘based’, ‘adequate’, ‘grounds’) have different definitions depending on whether one is an externalist or an internalist.[2]

Assuming that we have at least a rough-and-ready sense of what a belief is (see footnote 2), let’s proceed to unpack what justification means on externalism. Swinburne presents a brief sketch of Goldman’s version of externalism. Assume that I have the belief ‘I see a TV in front of me’. To be justified, the belief must be made probable by being based on adequate grounds. The grounds of the belief, for Goldman, will be events in the nervous system[3]. The belief will be based on those grounds if it is caused by those grounds. The grounds will be adequate if the process generating those beliefs is reliable—it tends to produce mostly true beliefs. To the extent that the grounds are adequate (because defined in terms of reliability), then the beliefs will be rendered probable. And what does an externalist mean by probable? Usually either physical probability (the extent to which causes predetermine an event) or statistical probability (the proportion of events of some kind in a population).

Things look very different on an internalist epistemology. A justified belief, to reiterate, must be made probable by being based on adequate grounds. Here, grounds are basic beliefs and based on means believed to be caused by[4]. The grounds will be adequate if they render the belief probable by proper a priori criteria.[5] Probable here means inductively probable—the extent to which proposition A is made probable given that it is based on proposition B.[6],[7]

The reader hopefully is beginning to appreciate how different the world looks depending on whether one is an internalist or an externalist. Nor do differences stop there. Consider the different ways in which defeaters[8] are handled by the two types of models. For internalist models, no major rehauling of the system is necessary: the new evidence is added and new inductive probabilities calculated. For externalism, however, acquiring a defeater means that the situation in which the belief was gained has been changed. This means that the grounds must be redefined, and raises the specter of the ‘generality problem’.[9]

Diachronic vs. synchronic justification. The second distinction is one that is not so familiar, but one that Swinburne thinks is important. He proposes that we also take into consideration synchronic (justification at a time) versus diachronic (justification at a time based upon prior investigation). Obviously, we could cross the internalist/externalist dimension with the synchronic/diachronic one and wind up with a 2 x 2 table. Most of the distinctions we have already made between internalism and externalism will apply equally well whether or not we are talking synchronic/diachronic justification. The major difference, as far as I can see, is that the synchronic/diachronic justification distinction makes is in terms of epistemic obligations: duties, you might say. For the most part in synchronic justification we are more of a ‘patient’ than an ‘agent’. That is, we believe what we do because it seems to us that the world is ‘impressing a belief upon us’.[10] However, in the case of diachronic justification we have adopted a belief following adequate investigation. That is, we have had more time to fulfill epistemic obligations in terms of assiduously weighing information, etc. [11]

The Connections Between Justification, Warrant, and Knowledge

Swinburne proposes that warrant for internalist models be defined as a justified (where justification is filled out in the way specified in the 5th paragraph of this review) true belief that does not proceed through or otherwise depend on a false belief.[12] If the belief is strongly warranted and strongly held, then knowledge obtains. Warrant on externalist models is defined as a justified true belief (where justification is filled out in the way specified in the 4th paragraph of this review). As we have already seen, whether one adopts an internalist or an externalist view impacts almost all of the core concepts of epistemology—i.e., grounds, basing, adequacy, probability, justification—and now warrant and knowledge. This is part of Swinburne’s point: the failure to appreciate the ambiguity of terms here has led to a probably fruitless search for a ‘silver bullet’ definition of knowledge which will apply equally well to all models. Returning to the main thread of this paragraph, though-ponder the different definitions of warrant given above. If Swinburne is right in stipulating these definitions in this way, then having justified true belief on an externalist model simply *is* having warrant. Not so on internalism-it is possible to have justified true belief and still not obtain warrant.[13]

We are now in a position to consider Swinburne’s answers to the two central questions of this book—what justification is, and what types of justification are worth having.

Two Answers

What is justification? The short answer is ‘it depends’. The longer, more complete answer is that it depends on the extent to which one is an internalist or an externalist in epistemology, and to whether one is talking about synchronic or diachronic justification.

What types of justification are worth having? All too baldly put, Swinburne seems to say that (a) almost all of the types of justification (externalist and internalist, synchronic and diachronic) discussed in this book are worth having in some sense but ultimately (b) diachronic justification is more worth having than synchronic and (c) internalist justification is more worthwhile than externalist.

In defending (a), I suggest that reflecting on the ways in which justification is defined (whether on internalism or externalism, whether synchronic or diachronic) reveals one obvious ‘common’ good: satisfying the definition of justification in all those situations renders true belief likely, and true belief is itself a good thing.

Discussing (b) and (c) requires a bit more thinking, but not *too* much more given the heavy lifting Swinburne has already done for us. Turning to (b): compare diachronic and synchronic justification. As discussed in the 7th paragraph above, it is possible on diachronic (but not synchronic) justification for us to fulfill epistemic obligations/duties. This means that there is a worth or good that is possible in diachronic but not synchronic justification.[14]

The last paragraph leads very naturally into (c): just as it is possible on diachronic (but not synchronic) justification to fulfill certain epistemic obligations and duties, so too is the fulfillment of epistemic obligations possible on internalist (but not externalist) justification. Consider again the Goldman example of externalist justification given in the 4th paragraph above. Justification is possible on (obviously!) external grounds, grounds which have nothing to do with the effort, the cognitive ‘blood, sweat, and tears’ we pour into it. On internalist justification, however, the relationships among ideas is vitally important. In fact, Swinburne argues that it is probably impossible to hold a belief in isolation. It is arguably analytic that *a* belief entails that there are also *other* beliefs. More important, however, is that only internalist justification can guide our decisions. Think of it this way: we can only act on the basis of evidence to which we have access. It is only once we *have* the belief ‘I see a TV in front of me’ that the belief can impact us.[15] Thus it is almost inevitable that we all be, implicitly or explicitly, committed to at least a (quasi-)form of internalism.[16]

Summing up. It is good to have true beliefs, and any situations that meet the various definitions of justification in this book will make having true beliefs either inevitable (in the case of externalist justification which presupposes reliabilism) or probable (in the case of internalist justification which relies on inductive probability). However, achieving diachronic and/or internalist justification allows for the fulfilling of epistemic obligations in a way that achieving synchronic and/or externalist justification does not. In addition, for our beliefs to impact our lives, it is unavoidable that we all be in some sense internalists.

In closing, let me quote Tim McGrew:

Why should anyone care what theory of knowledge is correct? Epistemology is pretty obviously hard, and at first blush it doesn't seem that any particular scientific achievement or practical course of action is likely to be affected by our choice of strong foundationalism over weak foundationalism or even some non-foundational theory. Why should we rack our brains over theories that don't have any practical application?
The simple but sobering answer is that human beings have an overwhelming desire to know. Many of the scientific achievements we most admire have been brought about because people had a burning desire to figure out how things really work, to get it right. Any theory of knowledge that writes off these achievements as an illusion or an accident will have a catastrophic effect on our world view. Closer to home, we would like to think that our beliefs about cars, houses, trees and sidewalks are in some important sense better off than the beliefs of people who take their cues from astrologers and crystal balls. If a theory of knowledge offers no defense against the argument that we are all ultimately required to retreat to a point of unargued, unjustified commitment, then we are wrong: there is no ultimate epistemic difference between our beliefs and the beliefs of devout horoscope-readers, and it is merely a historical accident that horoscope-readers are currently a minority of the population.”[17]
A sound and defensible theory of knowledge is our only rational line of defense against conceptual anarchy. This is the best and perhaps the only reason to take epistemology seriously; but it is also a sufficient one.

---

Apologetics 315 Book Reviewer Latter Day Inkling is a U.S.-based research psychologist for the military. He is especially interested in epistemology and natural theology.

---

[1] Swinburne acknowledges that there are plenty of ‘mixed’ models. As will be seen, even those externalist models which are paradigms of externalism (e.g., Goldman) have hints of internalism.

[2] Thankfully the term ‘belief’ does not seem to vary between the two approaches. A belief, roughly thought of as awareness of a proposition assented to, is a mental event and hence in Swinburne’s terminology something to which we have privileged access. Swinburne does acknowledge that there are some who are eliminitavists (i.e., Paul and Patricia Churchland) and so mounts a defense of the reality of mental events as distinct from the physical.

[3] Events ‘in’ the neural system may seem to fit more comfortably with internalism. However, the distinction is that on externalist models it is not required that the believer have introspective access to the neural event (or whatever the particular flavor of externalism proposes as a ground for the belief).

[4] ‘Believed to be caused by’ is obviously different than caused by. There is a complex discussion I’m sliding over concerning how much it is required that the ‘believed to be caused by’ is in fact true-did believing that A cause the belief that B? Even if so, the causing is very different. As Alvin Plantinga noted in Naturalism Defeated?, the content of a belief causing another belief sits ill with physicalism. This is why Plantinga argues that physicalism likely entails epiphenomenalism, and is grist for the mill in his ‘evolutionary argument against naturalism’. Almost as much fun is Plantinga’s noting that this problem for naturalism is roughly equivalent to the problem for dualism posed by mind/body interaction.

[5] Swinburne’s usual four criteria: explanatory power, fit with background knowledge, scope, and (most important of all) simplicity. For a brief discussion of these criteria and how they can be applied to both personal and scientific explanation, see http://www.apologetics315.com/2013/07/book-review-simplicity-as-evidence-for.html

[6] Swinburne illustrates this by applying the four a priori criteria to assessing the extent to which some proposed hypothesis explains a set of data. He then extends this to consider how one proposition can explain (render probable) another proposition, and also applies it to human testimony. See Chapter 4 of EJ.

[7] Swinburne further breaks down inductive probability into logical probability (inductive probabilities as calculated by an omniscient agent using correct a priori criteria), epistemic probability (fallible agents using roughly correct criteria), and subjective probabilities (fallible agents using incorrect, idiosyncratic criteria). For the purposes of this review I will continue with the term ‘inductive probability’ by which I mean roughly epistemic probability. I do this for the simple reason that Swinburne thinks most adults use roughly correct criteria and therefore subjective probability can be ignored.

[8] Some new belief B which either undercuts (makes me agnostic towards) or overrides (makes me deny) some previously held belief A.

[9] The gist of the generality problem is that most forms of externalism (‘reliabilism’ ala Goldman) have trouble in non-arbitrarily specifying what the grounds of a belief are. This affects the adequacy and hence affects reliability-which affects how probable a belief has been rendered. For further discussion, see Plantinga’s critique of Goldman on precisely these grounds on page 198 of Warrant: The Current Debate. Goldman attempted to avoid the generality problem in accommodating defeaters by adding something like the internalist ‘introspective access’ criterion to his model.

[10] Swinburne notes that if we could just believe something on a whim (‘doxastic voluntarism’), then we would be in the contradictory position of believing something to be true on the basis of something other than it seeming to be true! Put another way, we would be saying ‘I believe x is true b/c even though I know I believe it on the basis of arbitrarily choosing it, not because it seems to be true.’

[11] Swinburne discusses various factors that make such an investigation adequate: the likelihood that investigation would have changed the initial inductive probability we assigned some belief, that it is important that this belief be true, and the cost associated with the investigation (e.g., time, money).

[12] The clause ‘that does not proceed…..false belief’ was added to deal with Gettier and Gettier-like situations. See Chapter 8 of EJ.

[13] A moment’s reflection should show why this is so. On externalist models the assumption of reliability is crucial. It is already part of the definition of justification on externalist models that the beliefs be based on something other than other beliefs-false *or* true. Internalist justification, however, involves relationships among beliefs (and possibly other mental states).

[14] I suppose we could further argue that we are more likely, on average, to arrive at true beliefs on diachronic vice synchronic justification. In that case, the ‘common’ good listed in (a)—the value of having true beliefs—while possible on both diachronic and synchronic justification is in fact more likely to be had in diachronic.

[15] This leads to a lengthy but important matter. One of the Amazon reviewers of this book comments that Swinburne doesn’t interact with Plantinga sufficiently. This has been remedied, to some extent, in Swinburne’s essay “Authority of Scripture, Tradition, and the Church” from The Oxford Handbook of Philosophical Theology. In that book, Swinburne sketches briefly the generality problem that most forms of externalism/reliabilism face. Briefly put, it is hard to calculate how reliable a belief-forming process is without defining exactly what the grounds are. Make the grounds too narrow, and you run into one set of problems. Make them too broad, you run into another set of problems. There seem to be no non-arbitrary way of determining what the grounds are to be (more technically, to what ‘type’ the ‘token’ process belongs). *If* Plantinga’s version of reliabilism is correct, then the generality problem is avoided because “The type to which a token belief should be referred to assess its reliability is fixed by God’s intention to produce true beliefs by a process of that type” (aforementioned Swinburne essay, Location 570, Kindle edition). But this raises an issue already touched on in the essay: we must have access (in the internalist sense) to something before we can believe it. So (1) if Plantinga’s epistemological model is correct, then the generality problem is avoided. But (2) before we can believe Plantinga’s model to *be* correct, we must have internalist access to the belief *that* the model is true (or access to other beliefs which make probable that his model is correct). Thus (3) believing that Plantinga’s model to be true presupposes that internalism is true! Swinburne also notes in his essay that possessing basic beliefs (in Plantinga’s terminology) seems to be an internalist matter as well!

[16] Swinburne considers three versions of internalism (epistemic coherentism [EC], nondoxastic foundationalism [NF], and doxastic foundationalism[DF]). He rejects EC on the grounds that it does not allow for ‘privileged’ beliefs (i.e., basic beliefs). Given that no human can possess an infinite number of beliefs, our noetic structure *must* begin somewhere—we must have basic beliefs. NF proposes that the starting point(s) of our noetic structure be some mental state (e.g., visual images) but not beliefs. This, Swinburne argues, won’t work—assume I see you getting on a train headed to London. That is evidence for you going to London only if I have the *belief* that the image is evidence of you going to London. This leads him to DF.

[17] http://homepages.wmich.edu/~mcgrew/pojman.htm