Responses 003 (2011-2012)

Question 3 of the list says: "Using history and at least one other area of knowledge, examine the claim that it is possible to attain knowledge despite problems of bias and selection."

This is certainly one of the most straightforward of the questions in the list. All that needs to be considered is how the idea of knowledge is framed — how do you know that something is knowledge, and how do problems of bias and selection act as obstacles to attaining it?

One paradigm I've always found useful is to think of knowledge as information that has found a purpose, and information as data (singular: datum) that have found a context. If you say that Sir Stamford Raffles founded Singapore in 1819, that's information. If you can find a use for that, it's knowledge.

The thing is that whether you approach knowledge this way or not, it is based on the methods by which we accumulate data and the means by which we structure the data into information. When we accumulate data, we almost always carrying out filtering processes in deciding what to keep and what to ignore — that's the 'selection problem'. When we do this such that it tends to provide a specific kind of direction or answer, that's the 'bias problem'.

If either of these problems arises, in any of the many ways they can, then the information space is compromised. It may have missing pieces, irrelevant pieces, missing but relevant pieces, relevant pieces that support only one side of an argument, and so on. When we purposefully act on the information we have constructed, we create knowledge, and when our information is compromised, so is our knowledge. Compromised knowledge... is that knowledge at all?

Consider the fable of the emperor with no clothes. All the evidence provided (interviews, first-person accounts) seems to say that that emperor is wearing perfect clothing. However, once selection and bias are dealt with appropriately, it turns out that there is actually no direct evidence for such a belief. The database, so to speak, is compromised. In fact, empirical research shows the emperor has no clothes at all (or prefers walking around naked even if he has clothes).

History, in particular, has got problems with selection and bias. It is a construct made with a clear purpose in mind — to tell a story, to prove a point, based on putting together evidence for past events. To that end, we select the contents of our database based on criteria such as relevance and what the preponderance of existing evidence seems already to show. We have our own ideas of what is reasonable, what is acceptable, and so on. Then we create a narrative.

But all historians are human agents selecting from a vast graveyard of bones to build perfect skeletons. We discard splintered bones and ambiguous artifacts where it seems convenient to do so. This may indeed give us the skeletons we seek. However, are those the real skeletons? Or are they merely what we think they ought to be?

The same is true to a lesser extent for the hard sciences and to a greater extent for the social sciences and humanities. That is where the person answering this question has to examine the claim critically, pointing out the differences between methods of knowledge acquisition and criteria for defining knowledge in each discipline.

Strangely enough, the vaguer the definition of knowledge, the less impact bias has on the outcome. After all, if we don't define knowledge sharply, errors to the database are less important. Look at the range of knowledge in the aesthetic disciplines for example, ranging from the hard technical aspects to the subjective evaluation of performance and style. It's easy to say whether a candidate has knowledge of the technical aspects; it's almost definite that a candidate who has created new knowledge cannot be evaluated correctly.