Can Machines Have Ura And Omote Understanding?
In Japan, there is twin concepts of Ura and Omote. Most commonly, it is used in a societal sense to describe the private or hidden aspect of a person (Ura) and their public persona (Omote). For a highly conformist and polite society like Japan, it makes a lot of sense why this concept is so widespread and why people would have a socially acceptable public persona and a (possibly) less acceptable private persona. However, my experience with Ura and Omote is through Karate (I am a 3rd Dan in Goju Ryu), where it is applied to depth of understanding. That is, the distinction is between a surface-level understanding (Omote) and a deep understanding (Ura).
Some styles of martial arts have broad curricula with dozens of techniques to learn. Not so for Japanese Karate. The techniques learnt in the first 6 months are the same techniques that are taught for the next 60 years. A White Belt doing their first grading after, say, 50 hours of training, will be tested upon their Seiken zuki (straight punch), the same technique demonstrated by someone testing for 5th Dan Black Belt after approximately 10,000 hours of training. And what is the difference between a White Belt Seiken zuki and a 5th Dan Seiken zuki? A lot!
The Omote understanding of a Seiken zuki is the trajectory that the fist takes, having feet in the correct stance and correct body posture. The next things to learn about Seiken zuki include (in no particular order): gripping the ground with the toes, alignment of punching weight into the heel, pressure into the ball of the foot for forward stability, slight bending of the toes, sinking of weight, tension in the inside of the legs, rotation of the coccyx bone (tailbone) forward, freeing of the hips before the punch, rotation of the hips with the punch, locking of the hips upon the completion of the punch, tension of the spine, contraction of the upper body muscles upon completion of the punch, but without being contracted when in motion, rotation of the fist and alignment of the wrist, elbow, and shoulders and striking with the correct knuckles.
But to know all of these things is still an Omote level understanding. To an extent, so is being able to do them. Where the depth of an Ura understanding of Seiken zuri comes in is when all these aspects of the technique are correctly done and repeated for (literally) thousands of hours. It is after almost endless repetition that the brain and the muscular system form around the understanding of the technique such that it is now part of the person. Doing a Seiken zuki with an Ura understanding is like putting on a well-worn pair of shoes – it is moulded to fit to a level of comfort that makes it feel innate.
Can Machines Know To A Ura Level?
In my view, for a machine to learn something to an Ura level, the knowledge would have to be learned to a depth that it reflects upon the machine itself. So thousands of hours of Karate leads to permanent musculature changes, thousands of hours of experience in a mental field leads to a re-wiring and growth of the brain to accommodate that task.
Most machines use static logical rules and operations. That is IF, OR, AND, THEN, etc. A simple machine like a winding watch uses the following rule. IF there is sufficient compression in the spring, THEN turn the dials on the clock face in accordance with the timing set by the gears. The watch does not understand that IF, THEN operation to a deeper, Ura, level after thousands, or hundreds of thousands of hours, of turning.
All of our mechanical and digital machines utilize some combination of the fundamental logical rules to operate. The rather clever AI that we now take for granted to suggest a route in Google Maps, which uses complex learning algorithms that constantly improve their output with greater amounts of data (including from your usage), are still made up of basic logic (and really fancy maths and probability, which I am in no way downplaying). But fundamentally, it is an optimisation.
That is, even AI, which utilizes complex techniques such as deep learning, which will develop the output algorithm based on large amounts of data, has a set bound within which the algorithm can train. Google Maps might output optimised routes based on its training. However, its type of output is not moulded by the task that it is giving.
Merely Worn Down?
One might reply that I am making unfair comparisons and that I should not compare the framework of machines merely their output. The correct analogy could be that although a winding watch does not change from its IF, THEN logic, over time, the gears and mechanisms wear such that they are moulded more specifically to the task that they are doing. And for AI that trains like Google Maps, one need not look past the optimisation of the solution.
Put that way, the Ura understanding of a Karateka is a biological optimisation. The muscles are ‘better’ at a given task. The developed neural pathways are no more than the wearing in of the cogs of the mind.
But this, to me, does not seem to accurately capture the complexity in dealing with one large organism and its highly varying environment. And more particularly, the human brain’s understanding of that environment and adaptation to suit the tasks that it is given and also that it chooses for itself.
In my view, machines can only have an Omote understanding, if indeed they could be said to have any understanding at all.