Hypocrisy and neural networks
So, as we see hypocrisy abound in our current world political situation, it’s become quite popular to criticise people on it. And I am not here to say that it is a good phenomenon – but it is certainly a *understandable* one.
So, first, before I head down this rabbit hole, let me draw your attention to videos of the Milgram experiments. One thing you will notice, over and over, is that people clearly were not of one mind about pushing the switch. They were obviously agonized over it, many of them protested or questioned the action, and yet ultimately the neural wiring that translated out to blind obedience of authority won. I know that I’ve discussed this before.
Now, I would say that this phenomenon is very closely related to hypocrisy. In both cases, you have collections of subnets that are at war with each other, or at least have a disagreement over what the correct action is. It’s pretty clear that religion does a much better job of programming people to say the right things than to do the right things, and what that may indicate is that religion does a good job of programming storyteller or verbal parts of our neocortex, but that a lot of the things that drive our actual actions are formed before religion ever gets it’s claws on us – they may be native to our DNA and the way it expresses itself, or formed in earlier childhood. Or it may be that they are formed later, but that some types of experiences lead to stronger collections of subnets than others. In any case, the thing to remember about hypocrisy is that generally I think you will find it happens when someone is of two minds about the subject.
For example, all the discussion about $CONSERVITIVE_POLITICAL_PARTY talking about how great $FAVORITE_RELIGION is while simultaneously doing things that are strongly against everything that $FOUNDING_RELIGIOUS_LEADER stood for are a great example of this. Some portion of their minds is in favor of tolerance and love and feeding the hungry and all of those things, but a larger portion of their minds is in favor of grabbing everything that isn’t nailed down, and possibly some things that are. (It is also, of course, possible that no part of their minds is in any way in favor of $RELIGION but that they are in favor of getting elected and since there is currently no punishment for lying on your way to office, there’s no reason not to claim to be in support of $RELIGION if it gets you the gig and the nice cushy salary for life)
However, assuming good faith for the moment, let’s suppose that they are sincere in their adoration of $RELIGION. That doesn’t mean that their whole mind is – and, no matter how persistent the illusion that we’re one single person per body, the truth is that we’re a huge collection of subnets, all with different goals and agendas and experience. I know that I’ve already referred to this, but I gesture you to the experiments of cutting the corpus collosum and the results that ensued.
I really think we’re not going to make serious progress until we start to accept some of the strengths and limitations of natural neural networks. Hypocrisy is in fact both. F. Scott Fitzgerald said “The test of a first-rate intelligence is the ability to hold two opposed ideas in mind at the same time and still retain the ability to function.” – and this is exactly the behavior we’re talking about here. Without it, we would never really be able to weigh the validity of contradictory but true ideas.
May 25th, 2017 at 4:22 pm
There may also be an aspect of NNs that, when they learn something new, they may adjust previous records to match the new pattern. This may happen fast and without conscious effort, such that previously held contradictory beliefs and actions are actually erased or re-molded in memory to fit current knowledge. It is possible that true “hypocrites” might not even realize that their memory is faulty in this manner. I don’t quite believe that all hypocrites are built this way, in fact I believe that if there are “clinical hypocrites”, they are rare… but I have no data to support this.
This line of thought (“clinical hypocrite”) does explain some of the odd (to many) behavior of some confronted hypocrites, even when faced with solid irrefutable evidence, their stance on an event remains unchanged: That isn’t how they remember it, and it doesn’t fit the current pattern, so it can’t be real.
May 25th, 2017 at 4:40 pm
I think precisely the problem we’re seeing is that *some* records get adjusted while others don’t. It’s very unlikely that a systemwide adjustment of all records would happen because of a new data point – this would be computationally expensive and is the sort of thing more suited to a procedural computer than a neural network of any stripe. It’s entirely possible that the operator of the NNN experiences two sets of memories, one for each position..