Author Information: Paul Faulkner, University of Sheffield, paul.faulkner@sheffield.ac.uk
Faulkner, Paul. 2012. Trust and the assessment of credibility. Social Epistemology Review and Reply Collective 1 (8): 1-6.
The PDF of the article gives specific page numbers. Shortlink: http://wp.me/p1Bfg0-oN
Please refer to:
- Origgi, Gloria. 2012. Epistemic Injustice and Epistemic Trust. Social Epistemology 26 (2):221-235.
- Origgi, Gloria. 2012. Reply to Paul Faulkner’s comments. Social Epistemology Review and Reply Collective 1 (10): 1-3.
Epistemic failings can be ethical failings. This insight is owed to Miranda Fricker who explores this idea in developing a theory of epistemic injustice. [1] A central type of epistemic injustice is testimonial injustice, where there are two components to this. A knower suffers a testimonial injustice when she is not given due credit and is thereby prevented from doing what is fundamental to being a knower, which is inform others of what she knows. This is the first component, which is epistemic: a testimonial injustice starts with a misjudgement of a knower’s credibility; it starts, in Fricker’s terms, with the knower suffering a credibility deficit. The second, ethical, component is the explanation of this credibility deficit. There is a testimonial injustice when the cause of this credibility deficit is not innocent error but some form of prejudice. Here Fricker wants to draw our attention to one pervasive prejudice, which she calls identity prejudice. [2] This is the prejudice that attaches to a person by virtue of their social identity and which thereby tracks that person through the multitude of social activities, economic, political and so on. Thus the paradigm case of testimonial injustice is identity-prejudicial credibility deficit. [3]
The stated objective of Gloria Origgi’s paper “Epistemic Injustice and Epistemic Trust” is:
to broaden her [Fricker’s] analysis in two ways: first, I will argue that the ways in which credibility judgments are biased go far beyond the central case of identity prejudice; and, second, I will try to detail some of the mechanisms that control our ways of making testimonial injustices to the speakers [sic]. [4]
In accordance with the first objective, Origgi’s paper proceeds to give different explanations of our credibility judgements. The mechanisms that underlie these judgements often go awry and result in credibility deficit. However, this suffices for testimonial injustice only if some kind of identity prejudice feeds into the mechanisms Origgi describes and explain why it is that they go awry. [5] This is to say that Origgi’s second stated objective cannot follow on from the first, as Origgi takes it to do, given Fricker’s understanding of testimonial injustice. What Origgi’s two stated objectives require is rather the following assumption: a testimonial injustice arises simply when a speaker suffers a credibility deficit. Given this assumption, Fricker then provides one explanation of testimonial injustice (our pervasive identity prejudices) and Origgi “broaden[s] her analysis” by outlining further explanations. However, Fricker rejects this assumption, and I will follow her in this. [6]
However, there is a further significant but unstated objective of Origgi’s paper, which is to introduce and define a notion of epistemic trust. It is then through outlining the various grounds that we have for epistemic trust that the stated objectives are achieved: what becomes clear is that our grounds for epistemic trust are fallible in ways that have a regular result credibility misjudgement and so credibility deficit (or testimonial injustice on the rejected assumption). This comment on Origgi’s paper will then consider these two ambitions: the presentation of a theory of epistemic trust; and an account of our grounds for this trust.
As I understand it, trust is a three-place relation: a trusting party T trusts someone (or something) D to do something φ. The trusting party T’s attitude is one of trust when T depends on D φing and has some positive expectation that D will φ. In my view, there are two key ways this positive expectation can be grounded, where each implies a different kind of expectation and with the result that ‘trust’ names two distinct attitudes. The positive expectation can be the expectation that D will φ; that is, it can just be a belief and grounded in all the ways that a belief may be grounded. And the positive expectation can be an expectation of D, namely that D will view things a certain way and be sensitive to certain reasons; in particular, T’s expectation of D can be that D will see T’s depending on his φing as a reason to φ. The positivity of trust in this case comes with the presumption that D will be moved by this reason, and so will φ. Thus, I have suggested, that one’s attitude of trust can be predictive or affective depending on the nature of the expectation it embodies. [7] Both attitudes of trust could be classed as ‘epistemic’ in the sense that either could be one’s reason for believing a piece of testimony.
It is Origgi’s view that our testimonial beliefs are based upon, what she calls, epistemic trust, which she then defines as follows.
I define epistemic trust as an attitude with two basic components: a default trust, which is the minimal trust we need to allocate to our interlocutors in order for any act of communication to succeed; and a vigilant trust, which is the complex of cognitive mechanisms, emotional dispositions, inherited norms, reputational cues we put at work while filtering the information we receive. [8]
If trust is an attitude, it cannot have as a more basic component the attitude of trust. Either default trust and vigilant trust are distinct attitudes of trust, and ‘epistemic trust’ is not properly described as an attitude of trust; or epistemic trust is an attitude of trust and ‘default trust’ and ‘vigilant trust’ are not properly described as attitudes of trust. It is the second option, I think, that is Origgi’s intention. With respect to ‘default trust’, what is required for an act of communication will differ depending on the act, but what is required for any act can only be understanding; and what is required for the success of any purportedly informative communicative act is acceptance, where this falls short of belief. [9] With respect to ‘vigilant trust’, vigilance then seems to be an assessment of what is communicatively accepted such that it becomes belief only if it passes this assessment. On this reading, Origgi proposes the following definition of epistemic trust:
An audience D epistemically trusts a speaker T iff
1) D understands and accepts what T communicates to him, and
2) D believes what T communicates to him on the basis of assessing that it is likely to be true given T’s communication.
Here 1) is the ‘default trust’ condition and 2) the ‘vigilant trust’ condition (where I shall hereafter call these the default position and vigilance requirement). On this understanding, there is some reason for Origgi to claim the following.
I do not see the relation between default trust and vigilant trust as an opposition between a Reidian (non-reductionist) attitude towards testimonial information and a Humean (reductionist) attitude. [10]
Contrary to the non-reductive position, there is a vigilance requirement on belief, which is not, as such, default entitled. (This requirement of vigilance has also been stated as the requirement that a speaker ‘monitor’, where ‘monitoring’ requires more than being on the look out for defeaters. [11]) Contrary to the reductive position, successful communication does instantiate a default position, and the process of assessment starts from this point, but it is not belief that is the default but an attitude that falls short of this, namely acceptance. (Or to put it slightly gnostically using Origgi’s terminology: the ‘default trust’ is not yet trust.)
However, Origgi’s statement of why her distinction between the default position and vigilance requirement does not map onto the reductive and non-reductive theoretical positions is cryptic.
Default trust and vigilant trust are deeply related: in most epistemic situations we do not choose to trust: we just do not have the choice. Thus, a default trustful attitude towards communicated information is possible in so far as there exist cognitive mechanisms, emotional dispositions, inherited norms, and so forth, that make us epistemically vigilant. [12]
The ‘Thus’ in this quote is, I think, misleading. It seems as if there are two arguments here not one. The first seems to be an argument against ‘default trust’ being the non-reductive ‘default’ attitude, where this is an attitude that is open to defeat. The point being that this is not the case with ‘default trust’: we have no control over whether we understand what someone communicates, and our acceptance of what is communicated is similarly compelled. That is, we can cause considerable affront to a speaker if we refuse to accept what is told.[13] So there can be a compulsion to accept what others say; as Origgi says, often ‘we just do not have the choice’. But this, of course, is not a problem if the default position is one of acceptance rather than belief.
The second argument then adds that even this default position presupposes epistemic grounds that allow for vigilance, i.e. a meaningful assessment of the truth of the testimony accepted. Suppose this is true; I do not think this truth carries the implication that Origgi intends: namely that epistemic trust, as she understands it, fits into neither the reductive nor non-reductive camp. Insofar as we are interested in belief, and not acceptance, what we are interested in is the grounds for belief. What Origgi then describes is the grounds that we can have for belief, where these are the grounds for ‘vigilant trust’, to wit, the cognitive mechanisms etc. The requirement of vigilance is then the reductive requirement that belief be based on some assessment of truth. As such, let me introduce my own terminology. Origgi’s epistemic trust is essentially predictive trust, as I define it, applied to the testimonial domain: it is the attitude of believing what someone says because one judges that what is said is likely to be true given their saying it. What Origgi then supplies is a sophisticated statement of the kinds of ground we have for this judgement.
In outline, the grounds Origgi identifies are as follows.
What does make us trust? I will detail here seven different sources of trust that we may monitor in ourselves and in others when we trust or present ourselves as a trustworthy source of information.
1. Inference on the speaker’s reliability
2. Inference on the content’s reliability
3. Internalized social norms of complying to authority (“He is my master, thus I believe what he says …”)
4. Socially distributed reputational cues.
5. Robust signals
6. Emotional reactions.
7. Moral commitments. [14]
The second half the paper then consists in Origgi running through these different ‘sources of trust’, or better: explanatory grounds of trust. These grounds are not singular in that an inference to the truth of a piece of testimony could be based on the following distinct grounds: ‘contextual signs of reliability’, ‘previous beliefs’, and ‘acknowledged expertise’. Origgi’s description here, of the extent of the grounds that we have for belief, then makes a positive contribution to the development of a reductive theory. In doing so it continues a tradition of arguing (contrary to Reid and after him Coady) that the reductive position is far from hopeless in this regard. [15] Since my sympathies are in line with Origgi’s here and I find her account excellent, I confine myself to three concluding observations.
First, a distinction needs to be drawn, and which Origgi does not draw, between grounds that explain trust and the grounds that justify trust. For example ‘contextual signs of reliability’ can be the basis of an ‘inference on the speaker’s reliability’ and so a ground for trusting a speaker. But do these grounds actually pick out speakers who are trustworthy, or is it merely that we believe that these grounds do so? In all likelihood it will probably be that some things we take to be contextual signs do in indicate reliability whereas some do not. If the vigilance condition is then merely that our trust have grounds and does not further require good grounds, then a credibility deficit or excess is consistent with vigilance.
Second, irrespective of whether vigilance requires grounds or good grounds, some statement needs to be given, and Origgi does not give one, of what makes grounds good or justifying. There seem to be two distinct possibilities here. One possibility is that justification is fundamentally social, such that trust is justified when the speaker is appropriately sensitive to those grounds that would be regarded in the community as good grounds for belief. To use Fricker’s term: epistemic trust would then be justified when it makes ‘routine discursive moves’. [16] The other possibility is that justification is fundamentally epistemic, such that trust is justified when the speaker is appropriately sensitive to grounds that are in fact good grounds — i.e. truth conducive grounds — for belief, where this might require, in Fricker’s terms ‘exceptional discursive moves’. [17] If the latter option is taken, then credibility misjudgements will be associated with trust being unjustified. And if vigilance requires good grounds, any credibility misjudgement would ordinarily imply a failure of vigilance. Of course, the fallibility Origgi identifies should not make one sanguine about this option.
Third, it would be good to have some explanation of the fallibility of the grounds that Origgi identifies. It would be good not merely in the sense that this would be an interesting further statement, but also in the sense that this is needed if Origgi’s account is to develop Fricker’s. In characterizing testimonial injustice Fricker offers an explanation of our credibility misjudgements — and particularly our assigning credibility deficits: our judgements are informed by identity prejudice. An explanation of these credibility misjudgements, on Origgi’s account, would then consist in an explanation of why the grounds of trust identified are fallible. Such an account is, I think, possible but it is one that Origgi leaves to the reader.
References
Coady, C.A.J. 1992. Testimony: A philosophical study. Oxford: Clarendon Press.
Cohen, Laurence J. 1992. An essay on acceptance and belief. Oxford: Clarendon Press.
Faulkner, Paul. 2011. Knowledge on trust. Oxford: Oxford University Press.
Fricker, Elizabeth. 1994. Against gullibility. In Knowing from words, edited by B. K. Matilal, 125-161. Dordrecht: Kluwer Academic Publishers.
Fricker, Miranda. 1998. Rational authority and social power: Towards a truly social epistemology. Proceedings of the Aristotelian Society 98 (2): 159-177.
Fricker, Miranda. 2007. Epistemic injustice: Power and the ethics of knowing. Oxford: Oxford University Press.
Goldberg, Sanford and David Henderson. 2007. Monitoring and anti-reductionism in the epistemology of testimony. Philosophy and Phenomenological Research 72 (3): 576-93.
Moran, Richrd. 2005. Getting told and being believed. Philosophers’ Imprint 5 (5): 1-29.
Origgi, Gloria. 2012. Epistemic injustice and epistemic trust. Social Epistemology 26 (2): 221-235.
[1] See Fricker (1998) and Fricker (2007).
[2] Fricker (2007), p. 27.
[3] Fricker (2007), p. 28.
[4] Origgi (2012), pp. 221-2.
[5] Origgi (2012), pp. 221-2. This requires qualification. Testimonial injustices are produced by prejudices that systematically track subjects through different social domains and, Fricker observes, “[t]he main type (the only type?) of prejudice that tracks people in this way is [identity] prejudice” (p. 22). So the qualification is: if there were another form of prejudice that were systemic in the way that identity prejudice is and which were equally a prejudice — i.e. were ethically bad — then this necessary condition would not hold. But Origgi does not identify any other such prejudice in outlining the mechanisms that result in credibility deficit.
[6] Fricker (2007), p. 22.
[7] Fricker (2007), p. 22.
[8] Origgi (2012), p. 224 (original emphasis).
[9] See Cohen (1992).
[10] Origgi (2012), p. 224 (original emphasis).
[11] See Fricker (1994) and Goldberg and Henderson (2007).
[12] Origgi (2012), p.224 (original emphasis).
[13] This is a point that assurance theories have been most sensitive to. See Moran (2005) and Faulkner (2011), ch. 6.
[14] Origgi (2012), p .227.
[15] See Coady (1992) and Faulkner (2011), ch. 2.
[16] Fricker (2007), p. 104.
[17] Fricker (2007), p.105.
Categories: Critical Replies
Leave a Reply