Friday, August 23, 2013

Goldberg and the problem of Anonymous Assertions

If one studies Indian accounts of Linguistic Communication as instrument of knowledge (śābda) one is immediately confronted with two different paradigms:
  • the Nyāya paradigm, according to which an act of linguistic communication conveys knowledge if it is uttered by a reliable speaker.
  • the Mīmāṃsā paradigm, according to which an act of linguistic communciation conveys knowledge, until and unless a flaw in the speaker interrupts this ability.
The first account presupposes a parataḥ prāmāṇya 'extrinsic validity' theory, that is, on the idea that an instrument of cognition becomes able to bestow knowledge only if connected with some external factors "enhancing" it, e.g., in the case of Linguistic Communication, a reliable speaker. The second account relies on a svataḥ prāmāṇya 'intrinsic validity' theory, that is, on the idea that an instrument of cognition is valid unless and until it is falsified.

This also entails that the Mīmāṃsā account can admit as valid the Veda (the Sacred Texts believed by Mīmāṃsā authors to be authorless and beginningless). In fact, since the Veda has no author, no author's flaw can ever invalidate it. By contrast, the claim that, e.g., the Buddhist Sacred Texts are reliable is easily invalidated, given that the very idea that their author, the Buddha, had access to unknowable realities such as dharma and karman can be easily falsified through an appeal to our common experience (where no one ever has access to unknowable realities) and to the inferential evidences about the fact that what is by itself unknowable cannot ever be intellectually grasped, even by enhanced sense- and intellectual faculties.

By contrast, the Nyāya school can only accept the Veda insofar as it regards it as authored by a reliable speaker (namely God).

Thus, it seems that if one wants to accept authorless assertions as valid, one needs to agree with the Mīmāṃsā authors and disagree with the Nyāya ones.

Similarly, a leading and thought-provoking scholar of epistemology, Sanford Goldberg, deals with Anonymous Assertions (forthcoming on Episteme, available here) relying on a theory which is quite close to the parataḥ prāmāṇya one, namely, on the idea that the validity of an assertion depends on an "epistemic norm", entailing the reliability of the speaker and the awareness of it by hearer and speaker at the same time ("in asserting something, the speaker performs an act regarding which it is common knowledge that her act was proper (warranted) only if she had the relevant epistemic authority", p. 6).

Now, what happens in the case of anonymous assertions?
1) That the speaker is not bound by any epistemic-norm-enforcing policy, since no one will be able to trace her back and hold her responsible for what she said or wrote.
2) That the audience is aware of that and has, consequently, grounds for scepticism.
Thus, "the hearer is unwarranted in thinking that the speaker satisfied the norm of assertion" (p. 27) and, consequently, anonymous assertions are (unless in particular cases), even if true, unable to convey justified beliefs.

Are there exceptions? Yes, and Goldberg (p. 29) mentions two:
  • "the 'security wall' model": one where a security wall warrants for the reliability of anonymous assertions. Goldberg does not mention any example of it, but one might think of moderated blogs admitting only those anonymous assertions whose reliability has been checked.
  • "the 'Wikipedia' model": one where the cost of correction is low and one can therefore hope that mistakes would have been detected.

Is this enough to make sense of all the reliable anonymous assertions we regularly encounter? I am not sure.
Let me start by listing them: Apart from the ones mentioned by Goldberg, one encounters anonymous assertions also in the case of editorials on newspapers (at least in continental Europe, the most influential articles are not signed), of encyclopaedic entries, of laws (where a pool of people is involved, many of which are not mentioned in the final text). (I will not deal with this last case, since Goldberg focuses only on assertions (and not on exhortations).)
An author would be more cautious while writing an anonymous editorial or an anonymous encyclopaedic entry than while writing a signed article, since much more is at stake. Signing an article means anchoring it (only) to its author, whereas letting it unsigned means entailing that the whole authority of the newspaper is at stake with it. Thus, the case of editorials and of encyclopaedias can be dealt with with the security wall model (i.e., it is the general reliability of a certain newspaper or encyclopaedia which vouches for the reliablity of each editorial or entry). However, they also hint at a further point, i.e., that in (Western) culture anonymity has long been a sign of authority (!), insofar as no limit to the authority is put.
This is even more evident in the case of Sacred Texts. Let us assume, as most historians do, that Sacred Texts (as, e.g., the Pāñcarātra Sāṃhitās) are authored by human authors. Let us also assume that these people were not only or not always driven by egoistic purposes, such as the desire to fool other people and/or gain money or influence in this way. Why would they nonetheless efface themselves in the works they write? Because, if they spoke as themselves, they would limit the authority of the final text to themselves. If this were the case, a certain text would be reliable insofar as its author is reliable, but not of the highest authority. Thus, it is easy to imagine that a certain person X would be careful while writing or teaching in his own name, but MUCH MORE SO while writing a Sacred Text deemed to depict an absolute Truth.

Can a justified true belief be based on an anonymous assertion? 

For another post on Western epistemology of testimony, see here (on J. Lackey). Similarly on testimony and justification, see this post. On the validity of Sacred Texts, see this post. To Linguistic Communication as instrument of knowledge is dedicated my first blog (in Italian), plus many posts on this blog labelled with "śabda".


Unknown said...

I of course have no answer to your final question (as far as I can tell, anonymity is hardly ever a sign of authority in Classical Islam, athlough there may be exceptions; pseudoepigraphy is, however) but I have to commend and thank you for this insightful and interesting post.

elisa freschi said...

Thanks, Marco! What about laws? Are they only connected to their author or isn't it the case that they form an independent corpus?

Unknown said...

I was only considering assertion. However, the epistemic bases of classical Islamic Law, in principle at least, rely largely on known authorities.
Shari'a is not a code made of "anonymous" injunctions but a corpus of authoritative opinions by predecessors, whose name is usually known.
At the end of the chain, sayings by the Prophet (where the "reliably believable to actually being by thhe Prophet" part is critical) and the Book (by God) are the ultimate source of authority.
Anonimy sort of re-emerges through the principle of consensus, but consensus is usually assumed to be among people who are reliable by definition.
A more naive approach helds that God is actually the "author" of thee Law, but this cannot, of course, be taken literally (although there are people who tried to).

ombhurbhuva said...

Elisa: Interesting topic. There seems in religious warrant to be two sorts of doctrine for the acceptation of the faithful:
(a) Accepted because authoritative
(b) Authoritative because believed by the faithful (.quod ubique, quod semper, quod ab omnibus creditum est", )

The Bhagavadgita comes into the latter category though of human origin.

Justified true belief as knowledge is a contested area. Can anonymous assertions be regarded as knowledge or reliable? Is Wikipedia reliable? Perhaps as a place to begin and on straight matters of checkable facts. Is Britannica more reliable in being written by accredited experts who are known? I tend to be happier with it but no doubt Wikipedia will demur.

elisa freschi said...

@Ombhurbhuva, are you talking about "two sorts of doctrine for the acceptation of the faithful" within Christian theology?
Because, interestingly, Sanskrit philosophy tend two slightly different options:
—authoritative teachings (akin to your first category), as in Nyāya. In this case, a Sacred Text is accepted because it has been authored by God, the Buddha, the Jina and so on.
—teachings which regard topics unattainable through human means of knowledge and have no author: their validity remains, thus, unchallenged. This is the Mīmāṃsā account of the validity of the Veda. But one may see in the background something of your second point, insofar as in order for this argument to apply the authority of the Veda must already be an established fact (accepted by everyone).

ombhurbhuva said...

There are similarities as one would expect and even in the discussion of apoureshya there are fundamentalist views and interpretive ones. Was it ‘heard’ or was it ‘remembered’ or a mix of the two. Is it part of the sampradaya? These are universal questions and they do not have answers that will be satisfying to the reductive rationalist. I feel that the mixing of belief and knowledge in JTB is a mirage that opens the door to universal scepticism.

elisa freschi said...

yes, there are different interpretations of what it means for a text to have no author (worthy of a mention is, e.g., Puruṣottama Bilimoria's attribution of authorlessness to the *truths* spoken of in the Vedas). And Sanskrit authors are usually adamantine in explaining that it is not enough to be merely remembered or transmitted. There must be ---logically, albeit not chronologically--- a Vedic base.

I am sorry to admit that I cannot understand "JTB".

ombhurbhuva said...

J.T.B. = justified true belief.
The Vedic base - that has latitude and longitude.

ombhurbhuva said...

Chapter IX of Newman’s Grammar of Assent on the illative sense is excellent on the development of doctrine. It’s on Gutenberg.

Licenza Creative Commons
Quest' opera è distribuita con licenza Creative Commons Attribuzione - Non commerciale - Non opere derivate 2.5 Italia.