Tacit Knowledge

In class today there was a discussion about tacit knowledge. In case you don't want to Google search it or look it up on Wikipedia, tacit knowledge is basically a synonym for experience. The discussion we had in class was on tacit knowledge was really necessary. Basically, we asked if everything needs to be learned through experience, or if we could possibly do anything right the first time we try by using a rule set. We consensus was that, while it may be possible write down an all encompassing rule set, it wouldn't be terribly easy.

Now, the conclusion wasn't as interesting to me as the way that some of my classmates argued it was. The discussion was based on whether or not tacit knowledge was necessary. I said it wouldn't, because robots can't be built on tacit knowledge. They always work on rule sets. However, they can learn rule sets strong enough to walk over rough and uneven terrain. Why wouldn't they be able to use genetic algorithms to create rule sets for any given task? My classmates said that this would be tacit knowledge.

I want to clarify this for a second here. We had a definition for what a term is. Then I made a claim that we don't need anything by that definition. Then, the classmate said that his definition actually included the attributes that I said lied outside of the definition. He cited the change in rule set as it still being tacit knowledge. However, tacit only means that it is knowledge that can't be shown as a rule set, and not that the rule set it can be shown as cannot change. If the rule set cannot change, then all knowledge would have to be tacit. We are always finding new smaller additions and changes to the sets of rules that we hold dear. Yet, these changes don't invalidate the old rules. The only justify them. Isaac Asimov wrote an incredible read on "The Relativity of Wrong." Really, everyone should go read it. However, it's only tangentially related to this discussion, so I won't go off on another rant about that.

I just want to know how often this tactic is used in arguments. Do classmates always attack the definitions in order to prove the points that they like? It seems like a bad practice. I suppose if it was more prevalent, then I would have noticed it before now. I'm just rather upset that my classmate believed he had pulled off his argument sufficiently. In fact, I believe that he may have convinced part of the class. This really upsets me. Also, it upsets me that I'm not sure if I won. Even if my logic followed much more soundly than his, I haven't actually achieved anything if no one has changed their mind.

I don't believe that it was my job to change his mind. However, I want him to follow logical conclusions. I believe that rule sets can compensate for experience. We can give robots the capacity to learn and write rule sets. We can take those rule sets and put them in other robots. And the most important point is that those robots, having had no experience, we perform just as well as the primary robot. I believe all knowledge is recordable, and can be re-written. Just because us humans are bad at projecting knowledge, does not mean that the knowledge we possess, or knowledge in general, cannot be projected. The world can be changed by us. We only have to step back and realize that we may not be able to use personal abilities to do so.

Though I suppose we can't really know that unless we experience it. According to my fellow classmates we will need tacit knowledge of creatures smarter than we are, in order to truly understand that not only things that humans can do are possible actions.