2. TRANSMISSION LOST

Lackey's case is as follows:

There are two arguments that can be made against transmission using the persistent believer case.Footnote ¹ The first goes as follows:

• (1a) Justification transmission occurs only if the speaker has justification for what she says.

• (2a) In persistent believer, the speaker lacks justification for what she says.

• (3a) In persistent believer, the listener's belief in the speaker's testimony is justified.

Therefore

• (4a) One cannot account for persistent believer in terms of justification transmission.

The second takes the conclusion of the first argument as its starting premise. The argument can thus be extended as follows:

• (1b) One cannot account for persistent believer in terms of justification transmission.

• (2b) If one cannot account for persistent believer in terms of justification transmission, then a theory of justification from testimony has no need for the transmission principle given in (T).

Therefore

• (3b) A theory of justification from testimony has no need for the transmission principle given in (T).

Consider the argument from (1a)–(4a) first. To begin with, it is worth noting that (1a) is straightforward. No transmission theorist should want to deny that justification transmission requires the speaker having justification for what she says. It is just a basic fact about the statement of transmission given in (T) that it only talks about cases in which a speaker has justification for what she says. If the speaker lacks justification for what she says, there is no justification to transmit. So (1a) is uncontroversial.

This leaves (2a) and (3a). According to Lackey, both (2a) and (3a) follow from the stipulations that come from setting up the details of the case and I think that this is also relatively uncontroversial. In Lackey's words: ‘the fact that [the speaker] believes the neurosurgeon that her visual powers are an entirely unreliable guide to reality, without holding any other relevant beliefs, provides her with an undefeated psychological defeater for her visual beliefs’ (Lackey Reference Lackey2008: 60 (emphasis added)). This establishes (2a). The defeater is a psychological defeater because the source of the defeat is something that the individual believes.Footnote ² But the basic point is that something renders the speaker's belief unjustified and I think that the transmission theorist should endorse (2a). There is, as I see it, nothing in Lackey's discussion of (2a) that should offend transmission theorists.

In support of (3a) Lackey states that ‘not only does [the listener] have excellent positive reasons for accepting [the speaker's] testimony, he does not believe, nor does he have any reason to believe, that [the speaker's] visual powers are an unreliable guide to reality' (Lackey Reference Lackey2008: 60). This observation requires caution, though. Lackey does not think that the listener's belief is justified by his background reasons. Indeed, Lackey states that ‘[i]t is therefore not enough for testimonial justification that a hearer have even epistemically excellent positive reasons for accepting a speaker's testimony – the speaker must also do her part in the testimonial exchange by offering testimony that is reliable or otherwise truth-conducive’ (Lackey Reference Lackey2008: 154). The fact that the listener has reasons for thinking that the speaker's testimony is true does not by itself mean that his belief is justified.

The idea is thus that the listener's belief is justified by the reliability of the speaker's testimony, but the listener's reasons for thinking that the speaker's testimony is true are a necessary condition of the listener's belief being justified. The listener's reasons enable her belief to be justified by the reliability of the speaker's testimony. So the idea behind (3a) is that the listener's belief is justified by speaker's reliable testimony, but this happens because the listener has reasons for thinking that the speaker's testimony is true. Regardless of whether or not they buy into the details of Lackey's explanation of (3a), I think that transmission theorists should endorse (3a) to the letter. And it is the letter of (3a) that the argument from (1a)–(4a) requires. The intuition that (3a) is true is one that transmission theorists should not find problematic.

As observed in §1, one of the reasons that I think that the persistent believer case is particularly interesting as a source of argumentation against transmission theories is because it yields more than one argument. It is also particularly interesting because the premises of the first argument are ones that transmission theorists should not want to deny, either because they are true by definition or appeal to intuitions that transmission theorists should not find controversial. This makes it a powerful argument. The fact that these arguments are powerful in this way means that lines of defence that have been deployed by transmission theories in response to other case-based objections referred to in §1 are not available in response to arguments based on the persistent believer case. For example, Faulkner (Reference Faulkner2011) responds to Lackey's (Reference Lackey2008) creationist teacher case by arguing that the case is in fact an instance of justification transmission. This type of response is not available when confronting the argument based on the persistent believer case and this makes the persistent believer case particularly significant.

According to the argument from (1a)–(4a), this is the problem for transmission theories. They are committed to a theory of justification from testimony that has to say something highly unintuitive about the persistent believer case. They cannot deny (1a) since it is true by definition according to their own theory. This means that the only way out seems to involve trying to deny either (2a) or (3a), but neither of those options is very nice since both of those claims are based on intuitions that transmission theorists should want to take seriously. Nonetheless, I think that there is a way for transmission theories to give an account of the persistent believer case that does not involve trying to deny any of these intuitions. In the next section, I will explain how the claims that transmission theorists ought to make allow them to agree with each of (1a), (2a) and (3a).

3. TRANSMISSION REGAINED I

The crux of the transmission theorist's response to the first argument based on the persistent believer case involves distinguishing between the following two claims:

• (T1) Justification transmission is the only way for a listener's belief in a speaker's testimony to be justified.

• (T2) Justification transmission is one way among others for a listener's belief in a speaker's testimony to be justified.

Each of these claims is compatible with (T). The statement of (T) only describes what happens when justification is transmitted. It does not say how often justification is transmitted. Lackey's discussion of transmission theories generally operates with the implicit assumption that transmission theories are committed to (T1). I do not think that they are. Indeed, I think that there are good reasons for thinking that the transmission theorist should deny (T1) and endorse (T2). I will come back to these later in this section and in §5. This applies to the first argument based on the persistent believer case because it allows the transmission theorist to make the following claim:

• (TP) The persistent believer case involves a listener's belief being justified in a way that does not involve transmission.

If the transmission theorist were committed to (T1), then she would not be able to endorse (TP). But since the basic claim in (T) does not commit transmission theories to (T1) and I think that transmission theories should endorse (T2) rather than (T1), the transmission theorist can claim (TP) without contradiction. The fact that the statement of justification transmission given in (T) is compatible with either (T1) or (T2) might seem like a basic logical point. But it is actually crucial to understanding transmission theories properly. And this is what the persistent believer case brings out. The argument against transmission theories based on the persistent believer case is successful only if transmission theories are committed to (T1). This means that the case for responding to the argument from persistent believer depends on the viability of a transmission theory endorsing (T2). If it is simply a basic logical point that transmission theorists can endorse either (T1) or (T2), then this is so much the better for transmission theories.

The observation that we might have a transmission theory of justification from testimony that endorses (T2) rather than (T1) helps bring out an important point about the dialectical situation concerning transmission theories. A transmission theory that endorses (T2) rather than (T1) does not make a claim about all justification from testimony, but makes a claim about a particular type of justification from testimony. Such a theory thus identifies a particular type of case – one in which a listener forms a justified belief by taking a speaker's word for it – and makes a distinctive claim about what justifies the listener's belief in such a case. It is this dialectical significance that objections to transmission theories based on the persistent believer case miss. The argument from (1a)–(4a) fails to observe that transmission theories are merely seeking to identify a particular type of case. Furthermore, I will argue in §4 that the argument from (1b)–(3b) is not decisive because it fails to observe that those cases are ones in which a listener takes the speaker's word for it.

It might be helpful to point out here that the dialectical situation concerning transmission theories is, in a particular way, similar to the dialectical situation concerning epistemological disjunctivist theories of knowledge.Footnote ³ An epistemological disjunctivist theory of perception is concerned, in the first instance, with giving a particular account of knowledge in a paradigm ‘good case’ of seeing something. The idea is that, whatever we want to say about cases that aren't the paradigm ‘good cases’ we should give a particular account of the knowledge acquired in the paradigm ‘good case'.

In the same way, transmission theorists endorsing (T2) rather than (T1) claim that, regardless of what we want to say about other types of case, in a case where a listener forms a justified belief by taking a speaker's word for it, the speaker's justification is transmitted to the listener.Footnote ⁴ Elsewhere, I have argued that the kind of case that transmission theorists are interested in is one in which a speaker's testimony is an expression of her justification, but the important claim here is that it is also one in which the listener takes the speaker's word for it.Footnote ⁵ Transmission theorists need not be committed to any particular story about what justifies a listener's belief in cases where she does not take the speaker's word for it (though I think there is a particular account that they should give). What matters to the distinctive transmission claim is how the particular type of case should be characterised.

Now, it is one thing to point out that there is a way for transmission theories to avoid being stuck with (T1). If the only reason for a transmission theory endorsing (T1) rather than (T2) is to avoid the first argument based on the persistent believer case, though, then the move is ad hoc and that is not good for defenders of transmission. But the move is not simply motivated by the persistent believer case. Rather, it is motivated by a consideration about the epistemology of testimony more generally that both transmission theorists and their critics should accept.

The first point to note, in establishing this, is that we can believe what speakers say in various ways. Suppose I tell you that φ. You might respond to my testimony in a number of ways. Firstly, you might treat my testimony that φ like the deliverance of an instrument and believe that φ based on the presumption that the processes involved in the production of my testimony are reliable. Secondly, you might believe that φ based on the presumption that you are able to distinguish between when I'm saying things that are true and when I'm saying things that are false. Thirdly, you might take my word for it that φ by believing that φ based on the presumption that I have justification for φ.Footnote ⁶

Since this distinction between different presumptions does a lot of important work in this paper, it is worth saying a bit more about what they might amount to. The idea of a presumption here is a familiar one in the epistemology of testimony. It is the same as the notion of a presumption in Faulkner's epistemology of testimony. It also appears, under the title of an assumption in Ernest Sosa's (Reference Sosa2010) account of knowledge from instruments and testimony. It is therefore worth noting that the notion of a presumption in this sense is one that is common in the literature. The first point to note about the nature of presumptions is that they are supposed to be weaker than beliefs. They might be thought of as background beliefs. Believing what a speaker says on the presumption that she expresses knowledge does not involve actively using your belief that the speaker knows what she is saying to infer the truth of what she is saying, but it does involve a weaker form of doxastic assent to the idea that the speaker's testimony is an expression of knowledge.Footnote ⁷

Another way of thinking about presumptions in the epistemology of testimony is in terms of the attitudes that we can take to our visual experiences. In considering visual perception, it is common to talk about visual appearances. The idea is that things appear to us a certain way and that we can come to believe things based on appearances either by taking a credulous approach to our visual appearances – by presuming them to be veridical – or we can come to believe things based on appearances by treating the appearances as evidence and using our background reasons to infer the way the world is based on how it appears.Footnote ⁸

The idea is that statements are the testimonial analogue of visual appearances. Where it might appear to us that things are thus and so in a perceptual case, it might appear to us that the speaker is expressing knowledge or merely that the speaker is reliable when confronted by a speaker's testimony. When we take a credulous attitude towards a speaker's testimony and take it at face value, we presume either that the speaker's testimony is an expression of knowledge or that it is reliably produced. So the presumptions involved in responding to a speaker's testimony are similar to the credulous attitudes that we take towards visual appearances.

A further question concerns how we are to individuate such presumptions. One might think that it is hard to see how this can be achieved. The notion of taking a speaker's word for it is grounded in the idea of presuming that the speaker has justification for what she says and this might seem to presuppose a particular way of individuating presumptions that is going to be hard to argue for. There might, on the face of it, be various different presumptions involved in any particular case of a listener coming to believe what a speaker says. A listener might make various presumptions about the situation, such as that the background conditions are a certain way (for example that the environment is not full of dishonest speakers), about the speaker being sincere, the speaker being competent, expressing knowledge, the speaker having been manipulated in such a way that she reliably says true things and so on. It might thus seem that the notion of taking a speaker's word for it is problematic because it rests on an account of the presumptions involved in responding to a speaker's testimony that cannot be motivated over any other competing account. Presumptions are more fine-grained than the idea of taking a speaker's word for it can account for.

Even if the idea that it is hard to individuate presumptions is correct, however, this need not be a decisive reason for dismissing the idea of transmission. The core idea is that there are presumptions that a listener can make in response to testimony from a speaker that she cannot make in response to a deliverance of an instrument. One way of filling out this idea is by suggesting that there are only two credulous presumptions that a listener might have in responding to a speaker's testimony, one of which involves relying on the speaker and the other involves taking the speaker's word for it. But this need not be the only way and the idea is that any plausible way of individuating presumptions will make sense of the core idea that some presumptions that can apply in the case of testimony just do not apply in the case of instruments. As long as this idea remains in place, transmission theorists can still maintain that a necessary condition of the speaker's justification being transmitted to the listener is the listener believing the speaker's testimony on the basis of a presumption that is distinctive to testimony.Footnote ⁹ The question of how to individuate presumptions thus amounts to a question about how to understand the notion of taking a speaker's word for it. And whilst the details might be difficult to fill in, unless we have a reason to think that they cannot be filled in at all, we do not have a reason to abandon the basic point about transmission.

This basic point, that we can believe testimony in different ways, should be neutral between different epistemological theories of testimony. This distinction, however, is the key to understanding how transmission theories can endorse (T2) in a principled fashion. The transmission theorist's endorsement of (T2) begins with the observation that the fact that we can believe testimony in different ways is epistemically significant. The idea is that the way that a listener comes to believe a speaker's testimony determines what justifies her belief. It is this central point that is crucial to the argument here.

Again, I think that this is a point that should be uncontroversial. Suppose that I tell you that φ and you thereby come to believe that φ. Even if we deny (T) and think that justification from testimony is a matter of reliability, rather than transmission, we should still think that the way that you respond to my testimony determines what justifies your belief. If we think that justification is a matter of the reliability of a particular process, we need to say which process that is. Where you believe that φ based on my testimony that φ, there are immediately two sets of processes involved. There are the processes involved in the production of my testimony that φ and there are those processes involved in your comprehension of my testimony that φ.Footnote ¹⁰ This raises the question of which set of processes is the one relevant to your justification.

Fortunately, an obvious and principled answer presents itself. It seems natural to think that, if you form your belief based on the presumption that the processes involved in the production of my testimony are reliable, then the relevant processes are those that determine the reliability of my testimony. By contrast, if you form your belief based on the presumption that your comprehension processes are reliable, then the relevant processes are those involved in your comprehension of my testimony.Footnote ¹¹ In short, the idea is that the relevant process (or set of processes) is determined by the presumption you make in coming to believe what I say.

More generally, the idea is that what justifies a listener's belief in a speaker's testimony is determined by how the listener comes to believe the speaker's testimony. Transmission theories can exploit this idea to make a case for (T2). The idea is that, if a listener believes the speaker's testimony in a particular way – by believing based on the presumption that the speaker has justification for what she says – then the listener's belief is justified by the speaker's justification for what she says, as (T) states. But in the same way that taking a speaker's word for it is not the only way of coming to believe what a speaker says, the statement of transmission given in (T) is not the only way for her belief to be justified. If, for example, the listener believes the speaker's testimony based on the presumption of reliability, then the listener's belief is justified by the reliability of the relevant processes, as described in the previous paragraph.Footnote ¹²

This distinction offers the key to understanding why the argument based on the persistent believer case is not a decisive objection to transmission theories. As observed in §2, it is intuitive that the listener's belief is justified by the reliability of the processes involved in the production of the speaker's testimony. Transmission theories endorsing (T2) rather than (T1) are able to endorse this claim, though. Since transmission theories are able to allow that the transmission of justification is one way among many for a listener's belief to be justified, they are able to allow that the listener's belief might alternatively be justified by the reliability of the processes involved in the production of the speaker's testimony in the persistent believer case, as Lackey claims.

Of course, the listener's belief being justified by the reliability of the processes involved in the production of the speaker's testimony in the persistent believer case depends on the listener believing the speaker's testimony based on the presumption that the processes involved in the production of her testimony are reliable. Nothing in the description of the persistent believer case rules this out.Footnote ¹³ The result is that the transmission theorist can claim that the listener's belief is justified and it is justified for the reasons that Lackey gives, namely the reliability of the processes involved in the production of the speaker's testimony. This is because the listener relies on the speaker rather than taking her word for it.

I have described a transmission theory according to which justification can be transmitted from speaker to listener when the listener believes the speaker's testimony on the basis of the presumption that the speaker has justification for what she says, but can alternatively be justified by the reliability of the processes involved in the production of the speaker's testimony when the listener believes the speaker's testimony on the presumption that the speaker's testimony is reliably produced. It is worth noting that other contemporary transmission theorists, such as Faulkner (Reference Faulkner2011) do not endorse all of these claims. Faulkner endorses the claim that the way that a listener responds to a speaker's testimony determines what justifies her belief but does not extend this to include the reliability of the processes involved in the production of the speaker's testimony. Rather, Faulkner stops at a distinction between taking a speaker's word for it, which includes a notion of trusting a speaker, and treating a speaker's testimony as evidence. I think that defenders of transmission theories should make this further step, however, since it can be justified by the (already endorsed) claim that the way that a listener responds to a speaker's testimony determines what justifies her belief. A transmission theory that makes this further step is also in a position to characterise the persistent believer case on Lackey's own terms.

4. TRANSMISSION REGAINED II

The trouble with the argument from (1a)–(4a) is that its conclusion is something that transmission theories can and (I have argued should) accept. This leaves the second argument. Unlike the argument (1a)–(4a), the argument from (1b)–(3b) does not have a conclusion that transmission theories can accept. If the argument to (3b) is successful, then transmission theories are in big trouble. But I do not think that the persistent believer case gives us any reason to think that the argument to (3b) is successful. Specifically, I do not think that the persistent believer case gives us any reason to accept (2b).

Again, the key to the argument here is the endorsement of (T2) rather than (T1). Once we accept (T2) rather than (T1) into the commitments of a transmission theory and fill in the details in the way that I have in §3, it becomes clear that the premise (2b) in the second argument is unmotivated. The idea in the persistent believer case is that the listener's belief is justified in virtue of the reliability of the speaker's testimony. Even though the speaker does not have justification for what she says, the fact that her testimony is reliably produced means that the listener can still form a justified belief by believing what she says.

In response to the first argument, I suggested that the transmission theorist can allow that the listener's belief in the persistent believer case is justified exactly because the speaker's testimony is reliably produced, as Lackey claims. But, one might contend, if a transmission theory allows this, then there comes a question of why we should think that the notion of transmission is necessary in the epistemology of testimony at all. Since the persistent believer case can be explained in terms of reliability, why should we not just explain all the cases that are supposed to involve justification transmission in terms of reliability? Doing so would allow for an altogether more parsimonious account of the epistemology of testimony.Footnote ¹⁴

The reason that we should not do this is because doing so tramples over important subtleties about testimony. Doing so involves treating justification from testimony as of the same kind as instrumental justification.Footnote ¹⁵ But unifying one's account in this way involves trampling over distinctions that apply in the case of testimony that do not apply in the case of instruments. We should want a unified account of the justification from instruments and the justification from testimony only if there is no reason for thinking that there is an important difference between the two. I think that there is a reason for thinking that there is an important difference between the two.

The reason comes from the observation made in §3, that we can respond to testimony in various different ways. In particular, we can take a speaker's word for it by believing what she says based on the presumption that she has justification for what she says. This is not an approach that we can coherently take in response to the deliverances of instruments, since instruments are not epistemic agents in the same way. This means that there is a distinction between the ways that we can respond to a speaker's testimony and the ways that we can respond to the deliverances of an instrument.

Obviously, though, it is one thing to point out that there is a difference in the way that we can respond to what speakers say and the way we can respond to what instruments appear to show us. It is a separate question, one might think, as to whether or not there is any epistemic significance to this. The distinction between the ways that we can believe what people say and what instruments purport to show is an interpersonal distinction. One might wonder why such a distinction should be relevant to the epistemology of testimony.

The answer to this question comes from the observation that it is natural to think that the way a listener responds to a speaker's testimony determines what justifies her belief. As observed in §3, any theory of testimony that allows that testimony can be justified in more than one way, even if this is just by the reliability of different sets of processes, needs an account of when a belief is justified by one set of factors and when it is justified by some other set. And it is natural to think that this is explained by the different ways in which a listener can respond to a speaker's testimony. This is the central insight of the argument here.

The idea that there might be some genuinely epistemic significance, in terms of justification, that is connected in some way to the interpersonal features of the testimonial situation, concerning how the listener responds to the speaker's testimony does not, by itself, amount to a new idea. Recent work by Richard Moran (Reference Moran2005), Edward Hinchman (Reference Hinchman2005) and Benjamin McMyler (Reference McMyler2011) has sought to argue that the listener's justification for believing a speaker's testimony is constituted by some interpersonal feature of testimonial situation. In Moran's terms, the idea is that a speaker telling a listener that φ involves presenting the listener with an assurance that φ is the case and the recognition of this presented assurance gives the listener a reason to believe that φ is the case (Moran Reference Moran2005: 7). According to this type of theory, interpersonal features about the conversation (the presentation of the assurance) constitute epistemic features of the situation. Lackey (Reference Lackey2008) calls these types of theory interpersonal theories.

The theory that I have presented here is not an interpersonal theory in that sense. Nothing that I have presented involves thinking that interpersonal facts such as the presentation of an assurance can have an epistemic significance in terms of constituting a listener's justification. Rather, I have been arguing that an interpersonal fact (specifically, how the listener comes to believe what the speaker says) can have an epistemic significance by determining what justifies the listener's belief. It does not follow from this that what justifies the listener's belief is itself an interpersonal feature. The listener's belief might be justified exclusively by being connected to features that are not part of the interpersonal character of the testimonial exchange. But the idea is that certain interpersonal features, namely the presumption(s) the listener makes in coming to believe what the speaker says, account for this connection, so whilst the theory that I have described here is not an interpersonal theory in the sense of Moran's theory, it allows a certain significance to the interpersonal features of the conversation.

A transmission theory can thus maintain that beliefs based on testimony can be justified in virtue of a particular process being reliable whilst still maintaining that the notion of transmission expressed in (T) is still a genuinely distinctive way for beliefs based on testimony to be justified. There is a distinctive presumption involved in taking a speaker's word for it that is not simply a mere presumption about reliability. It involves treating the speaker as an epistemic agent, rather than a mere instrument capable of producing reliable deliverances.

This falsifies the claim in (2b) that, unless the persistent believer case can be characterised in terms of justification transmission, there is no need for a theory of justification from testimony to endorse the notion of transmission given in (T). The reason for endorsing the notion of transmission given in (T) comes from the observation that a theory of testimony should allow that what justifies a listener's belief is determined by the way she responds to the speaker's testimony combined with the observation that there is a distinctive presumption that can be made when responding to a speaker's testimony rather than the deliverance of an instrument. The theory that comes from endorsing these considerations is a transmission theory because it endorses the notion of transmission given in (T) as a justification available to a listener that does not just amount to considerations about reliability.

5. AN OBJECTION

Both the response in §3 to the first argument and the response in §4 to the second argument exploit an important claim. The claim is that transmission theories can make sense of the claim that, in the persistent believer case, the listener's belief is justified because she believes the speaker's testimony with the presumption that the testimony is reliably produced. The idea is that the persistent believer case is not a counterexample to transmission theories because the case is not one in which transmission is supposed to happen. According to the theory that I have been describing, transmission happens only if a listener's belief is formed through her taking the speaker's word for it. Since this does not happen in the persistent believer case, the case is not a counterexample to transmission.

The strategy in §3 in response to the argument purporting to show that the persistent believer case cannot be characterised in terms of transmission exploited the thought that the listener relies on the speaker rather than taking her word for it to explain how the listener's belief can be justified in a way compatible with endorsing (T). The argument in §4 sought to show that transmission can meaningfully be endorsed in other cases even if it is not in the persistent believer case by also arguing that the listener responds in a distinctive way to the speaker's testimony in the persistent believer case. Both of these arguments exploit the thought that the persistent believer case just is not the kind of case associated with justification transmission.

One might think, however, that the persistent believer case can easily be turned into a case that offers a counterexample to the transmission theory given above. We might simply stipulate that, in the persistent believer case, the listener forms her belief by taking the speaker's word for it in the way described by the above transmission theory. As a result, there is a case of a listener's belief being justified, in a way that cannot be explained in terms of transmission, but also in a case that the transmission theorist needs to be able to explain in terms of transmission. This modified version of persistent believer requires characterisation and it presents a serious challenge to the transmission theory given above.

I think that, in such a case, the transmission theorist should claim that the listener's belief is unjustified. The reason for this is straightforward. Ex hypothesi the listener believes the speaker's testimony based on the presumption that the speaker has justification for what she says. And since this is false, the listener's belief is unjustified. The argument is straightforward enough, but one might think that it would not convince anyone taken with the intuitions in the original persistent believer case.

This may be correct. And I accepted in §2 that I think that the considerations Lackey seeks to extract from the persistent believer case, including the claim that the listener's belief is justified, are intuitive. But I think that the above considerations give us a theoretical reason for thinking that, contrary to initial intuitions, the listener's belief is not justified in a modified version of the persistent believer case where it is stipulated that the listener takes the speaker's word for it.

The argument is relatively straightforward. Against the idea that the listener's belief is justified in a situation where it is stipulated that she takes the speaker's word for it, consider the following argument:

• (5) The justification that supports a listener's belief is determined by the way that she responds to the speaker's testimony.

• (6) Where the listener believes the speaker's testimony based on the presumption that the speaker has justification for what she says, the listener's belief is supported by the speaker's justification for what she says.

• (7) In the persistent believer case, the speaker lacks justification for what she says.

Therefore

• (8) If the listener takes the speaker's word for it in the persistent believer case, then her belief is unjustified.

Whilst the conclusion in (8) goes against the intuitions in the original persistent believer case, it follows from the considerations in (5), (6) and (7). We have already established the intuitive plausibility of (7) in §2 and I do not propose to retract the claim that (7) is highly intuitively plausible at this juncture. This leaves (5) and (6). I'm also not sure how someone could endorse (5) without thereby endorsing (6). The claim that a listener believing what a speaker says based on the presumption that the speaker has justification for what she says connects the listener's belief to the speaker's justification seems to follow naturally from the thought that the way that the listener responds to the speaker's testimony determines what justifies her belief.

This leaves (5). In §3, I argued that even theories of testimony that deny transmission should think that the way that a listener responds to a speaker's testimony determines what justifies the listener's belief. And I argued in §4 that this need not create problems for the idea that there is a distinctive way for a listener's belief in a speaker's testimony to be justified that is given in (T). Suppose, however, that theories that think of justification in terms of reliability deny this claim. This would provide a framework for rejecting (5) and (6). But it would do so, I argue, at an unacceptable cost.

Denying (5) whilst allowing that a listener's belief can be justified by the reliability of the process(es) involved in its production requires an account of when the processes involved in the production of the speaker's testimony are relevant and when the processes involved in the listener's comprehension are relevant. The reliability of these processes are independent of one another. A speaker might produce reliable testimony even though the listener is unable to distinguish between true and false statements. Unless such a theory allows that a listener's belief is connected to one set of processes rather than another by the way in which she responds to the speaker's testimony, it is hard to see how the theory can give a principled answer to the question of when one set of processes rather than another are relevant. This was the motivation for wanting to allow that presumptions determine what justifies beliefs based on testimony in §3.

In summary, the same considerations that allowed the transmission theorist to resist the argument from (1a)–(4a) also allow the transmission theorist to resist the argument from (1b)–(3b). In the case of the latter argument, there is a dilemma for the critic of transmission. The viability of the claim that the listener forms a justified belief in the speaker's testimony depends on her relying on the speaker rather than taking the speaker's word for it. However, the status of the case as a counterexample to transmission theories depends on the listener taking the speaker's word for it rather than relying on the speaker. So characterising the way the listener responds to the speaker's testimony generates a dilemma for the critic of transmission theories.