a blog about things that I've been thinking hard about

The Secret Hidden Process of Worldview Construction

27 September, 2007
uncritical listening, hidden even from the listener

Uncritical listening may play a significant role in worldview construction.

But, it is (obviously) vulnerable to manipulation.

So, it must occur in a secret manner which is kept hidden from both the speaker and the listener.


This article is an update to a previous article Bootstrapping the Mind. In that article I proposed that uncritical listening is a significant component of how our brains construct a worldview, i.e. an understanding of how the world works, and implicitly, a strategy for learning from experience.

I suggested that there must be "trigger conditions" which cause this process to occur, and these conditions could include the age of the listener, the relationship of the listener to the speaker and the manner in which the speaker is speaking.

The Necessity of Hidden-ness and Secrecy

More recently I was thinking about how one might observe the occurrence of this hypothetical uncritical listening state, or whether one could remember subjectively one's own uncritical listening. But then I realised that there must be considerable selective pressure to hide this process from both the speaker and the listener.

The major risk of uncritical listening is that it exposes the listener to manipulation, so if the speaker could detect that the person listening to something they said was responding to it by accepting it uncritically, then the speaker could use that information to learn how to manipulate the other person. Which would be a bad thing for the listener.

There are two reasons why the process must also be hidden from the listener. The first is that if the listener is consciously aware of their own changed mental state, then this conscious awareness will have some externally observable effect, which the speaker might be able to detect. The second reason is that if the listener can detect the occurrence of uncritical listening within their own mind, then they might use this information to manipulate their own worldview, for example, to adopt a worldview more likely to result in short-term happiness and satisfaction with life. This would also be a bad thing, from the point of view of long-term reproductive success, for the listener.

Of course if uncritical listening results in an immediate change of worldview, then this will be observable, no matter how well the short-term change in mental state is hidden. So to be very sure that the process and its effects are hidden, the change in worldview will need to be both delayed and effected very gradually, for example, over a period of months.

It follows that if there does exist an "uncritical listening" faculty within the brain which operates in this way, then our worldviews must be formed as a result of some subset of all the things that different people have said to us in the past (and excluding everything said in the very recent past), and we have no easy way to determine which of those statements were the ones that most affected us.

Only During Childhood?

It is possible that the major trigger for uncritical listening is that the speaker is a parent of the listener. This reduces somewhat the risk of manipulation, since it is reasonable to assume that the parent which a child depends on both practically and emotionally is someone who acts substantially in the interests of that child.

However, at a very young age a child's ability to understand concepts is more limited, so if the "cut-off" period comes too soon, then the opportunity to acquire more subtle and abstract concepts into a worldview via uncritical listening will be lost (and it will be difficult to construct a worldview containing such concepts internally given an a priori worldview based on less subtle and less abstract concepts).

How Do We Know What We Know? and the Credit Assignment Problem

"How do we know what we know?" is the question that sums up the worldview problem. If relates to the fact, that, even if our worldview is derived largely from experience, we still require some pre-existing worldview to interpret that experience in the first place.

A very closely related problem is the Credit Assignment Problem. This problem arises when we consider the issue of reinforcement learning. The notion of "reinforcement" implies some pre-existing notion of what is "good", with an assumption that any behaviours whose consequences are "good" should be reinforced. Two underlying difficulties, which are not at all easily solved, are firstly, how to determine which consequences are the consequences of which behaviours, and secondly, how to determine which "behaviours" are actually the same behaviours (i.e. what constitutes doing the "same thing" as what I did last time, given that the circumstances next time will not be exactly the same as they were the last time).

In a biological context, the worldview problem and the credit assignment problem are effectively equivalent, because in order to assign credit one needs a worldview to tell you what the relationship is between your actions and the consequences of your actions, and secondly, the only major benefit of having a worldview is to solve the credit assignment problem (to put it another way, natural selection is not interested in our brains doing philosophy just for the sake of doing philosophy).

Vote for or comment on this article on Reddit or Hacker News ...