Can we know what knowledge will do?

There are inherent limitations in our ability to predict what will happen when knowledge gets used.

~

In this post, I will follow the theme of my last post on truth and utility to examine the question of whether we can identify knowledge that will be useful.

In my post on microbeads, we looked at how knowledge was utilised by examining the issue retrospectively. That is to say, from the vantage of the present, we could look back on past events to make judgments about the utility of microbead related knowledge. I showed that the knowledge base had developed over time and in such a way that harms were created - and then also hopefully resolved.

From the vantage of the present, it seems that it would have been nice to foresee, and mitigate, the negative consequences of microbeads in advance. However, the task of identifying the utility of knowledge prospectively is not so easy. It is easy, perhaps, to anticipate (or imagine) the ‘good’ things that would follow if only people would adopt or act on the ideas we work so hard to create. It may be as easy to anticipate the terrible things that might follow from something that, for whatever reason, we just don’t like the sound of. But the future is inherently uncertain, and outcomes are rarely clear-cut. This uncertainty sets limits on what we can reliably predict will transpire in the future. The unforeseen implications of our activities, good, bad or otherwise are often referred to as unintended consequences.

While the future is always uncertain, it is not necessarily radically uncertain. Setting aside the problems of deriving knowledge based on inductive reasoning, I can be pretty sure that I will wake up tomorrow in the same house that I go to bed in tonight. Based on this proposition, and others like it, I can make a number of predictions about what is going to happen to me tomorrow and be pretty sure that most of them will come true. In general, I can rely on inertia; i.e., that in the absence of an external force, many of the activities and  processes that impact on me will stay largely the same from day to day. I also know certain things about the problem at hand (e.g., I am not prone to sleepwalking, I am not a doctor who is on call and thus who might be awoken before midnight to go into work, the prospect of fire or flood causing evacuation is fairly unlikely, etc.). So, I can make the prediction about where I will wake up because I have enough information to do so and because I am making a prediction about a relatively passive and uncontentious process.

However, predicting the future impact of a piece of knowledge, even in fairly simple scenarios, is inherently more complicated than the example I have just given. This is because we are not making a prediction about an issue unlikely to change but the opposite; we are trying to predict whether an action will cause the changes we want to occur and not generate adverse effects that undermine its benefit. Thus, the question we are posing is based on a (hoped for) perturbation to the current order. Second, it is not confined to the context in which we want to have impact but to all future contexts that our knowledge could possibly be used in. We don't want to dwell on this complexity too much because it leads us to an absurd conclusion that engenders an unhelpful fatalism. To take this proposition too seriously would mean that we would need to gather a massive amount of information about all the contexts an idea might effect just to have a chance of predicting (and avoiding) possible future deleterious impacts before acting upon them.