There are two concept that are relevant, the concept of tacit knowledge and the concept of expertise. I will first begin with the former idea. Tacit knowledge is the knowledge that human gain unintentionally and involuntarily. For example we can talk about recognizing alphabets. If you ask someone to describe precisely how he can differentiate between various handwritings of A and handwritings of B then that person will have great difficulty in doing that. Same holds for bicycling, chess playing etc. You begin to do these things consciouslly but then you gradually begin to do them involuntarily, by the term involuntarily we mean that humans are not controlling these things precisely - there might be a higher level awareness and control such as while bicyling you know you are bicycling but you don't precise control the motion of your legs and arms. This hidden knowledge that one acquires without realizing is what we mean by tacit knowledge - it is the end product of implicity learning i.e. learning without being aware of learning it. The concept of expert is as we see becoming connected to acquiring tacit knowledge. By reading the references [1],[2] and having an open discussion with the group, I realized that there are two main questions regarding the idea of tacit knowledge.
- Is tacit knowledge, unexpressable or epiphenomenal
- Can tacit knowledge be expressed by the person who possess it
I believe that the first one is wrong while the answer to second question is evidenly no. The tacit knowledge is not something which cannot be formalized, if the function of all parts of brain are identical then probably the manner of storing tacit knowledge(implicit memory) is no different from other type of memory. What probably can be the case is that the exact manner of storing is not accessible to the conscious part. Its like computer has bytes on its tape which are weights of a neural network but it does not know that. It is not aware what these bytes are.
The next question that we discussed is that concept of expert and method of gaining expertise. The expert, as a term defined wholly for humans, means a person who has skills of superior level compared to other in the same field. Walking, eating and other activities are such in which everyone has equally superior skills and we do not use the term expert in this reference. However things like playing chess, keyboard, boxing etc. are considered as fields with non-uniform distribution and the term expert makes more sense here. The common thing about these domain is that they require persistent effort to perfect. Let us take the example of chess, in the begining there is a conscious effort of understand what moves give what result, almost every chess player is immediately able to see that queen has to be given more protection that say pawn. These conscious attempts after a while(at uneven pace), sink into the unconscious layer. How this happens is a matter of great research areas with multiple fields - neuroscience, cognitive, philosophy, psychology etc. [1] converge. What is also not clear to me, as an amateur, is whether there is actually a difference or not at ground level. Can it happen that at the end we find that they all have the same phhysical structure (as pointed out in class - neurons at different place look at the same). I circumvent this question and go back to the discussion on expertise. Since practise and probably some genetic coding, is the source of gaining expertise and the trend in learning has always been from consciousness to unconsciousness, thus I believe that expert system always have tacit knowledge. Chess player knowing the move without knowing how he arrive at it, people recognizing alphabets without knowing how they do it are all strong examples. Knowing how this tacit knowledge is acquired and stored is to know how how we are programmed to learn, what learning algorithm run in our mind and there is nothing that suggests we cannot do this one day.
In the case of computers however the story is slightly different. The term tacit knowledge cannot be so easily applied to computers since they do not have a mechanism of knowing what they have. To settle this let us have a smart system which carries a map in its memory which stores the particulars of each bit. So if we run a Neural Network Algorithm then it knows that it is a neural network code and knows which bytes in its memory correspond to weights etc. Now even in this context, tacit knowledge can be defined to mean knowledge which the computer acquires without knowing. For example it may know that the information stored in sector A2 of disk corresponds to measures of the code however it may not know what is the use of these measures, and the recursion can continue. At a point there will be something which exists in its domain and about which it knows but there are things which it will learn without knowing it and will not be able to explain how it solve it, to fullest satisfaction.