re: A Speculative Post on the Idea of Algorithmic Authority

In reply to Clay Shirky (unfortunately I couldn't post as a reply there, as it kept returning an error about javascript and cookies)

 On top of all that, there is a distinction to be drawn between the 'authority of assertions' and the 'authority of actors'. As you alluded to, we tend to accept the authority of an actor (person, in general) in a particular domain, but perhaps less so in another. There is a certain amount of bleeding of trust at the edges - if we trust someone in a number of fields, we are more likely to trust them in an as yet untested field, whereas if we know them to be expert in one domain but flakey in another, we will tend not to trust them in a new untested field, as we know about their unreliability.

Trust is what it all boils down to - and that is influenced by a number of factors. I suspect that we can model it fairly well using Bayesian inferencing, given enough data. An interesting aspect is whether we believe we are trusting people or 'facts'. In general, I think, people tend to trust people more than the assertions that are made. The actor is a trust proxy for the information they convey. The level of trust in them is based on many things, though, and even someone who is essentially always right can find themselves not well trusted if they are temporally unreliable, have behaviour which does not comply with social norms or otherwise fail to meet our expectations. There is an uncomfortable point to contend with regarding rhetoric, PR and sales pitches too - although many of us would like to believe we are essentially rational, the emotive effects of these practices can change our behaviour in ways we often might not care to admit. Persuasiveness is a great tool for acquiring trust, and thus authority, as long as it outweighs any negative experiences on the part of the subject. In academia, authority is normally conveyed by citations. But a lot of published material is fairly dodgy (in my experience) and the whole point of the scientific method is that our theories are out there to get knocked down and proved wrong as we advance knowledge. Academics often seem to rely on citations in what they read to give a measure of the authority of the paper and/or the authors - but how often is it actually properly earned trust which results? Google is an interesting service from this point of view. The reason, I would suggest, that people place a lot of trust in its results, is because it effectively mimics what people generally do when trying to ascertain whether they should trust someone (or something). It acts as a proxy for the trustworthiness of the results it returns, based on indications of trust provided by links (and other behaviours). Of course, it can be misled, just as any person can be - and in general, because it manages to be pretty reliable most of the time, we forgive it its mistakes. It benefits over wikipedia in that it doesn't try to make any claim to be authoritative. People lose trust in actors who claim authority but show themselves to be unreliable - and that is where wikipedia fails with many academics. Others, of course, are happy to place their trust in wikipedia - and they are often rewarded with good quality information. The issue I can see is that people come to rely on their preferred service rather too easily, and need to make sure that they check their facts against a number of sources. Maintaining an open, critical, mind is what is needed, whether looking for information from wikipedia, Google, your government or a magic 8-ball.

Trackback URL for this post:

http://brains.parslow.net/trackback/1553

De Krabpaal specialist, de goedkoopste krabpalen van Nederland

De krabpaal specialist, voor de meeste, mooiste en goedkoopste krabpalen van Nederland