How close to the original is your kata?

How close to the original is your kata, and how would you know?

It is a cliche now that when anyone asks a question about altering/varying a part of a kata they are learning, at least some of the responses will be along the lines of "you should do it the way your sensei teaches you", or "kata are the legacy of past masters' combat experiences - to change them is to lose those experiences", or "you need to do it the same way it has always been done".  All of those responses pre-suppose that the kata you are currently learning is unchanged.  As I've discussed in Part 1 and Part 2 of this series on why the kata we are doing are probably not the same as when they first began, this may not be the case.

So how can we determine how close to the original our kata is?  I have a number of ideas (be aware this is mostly supposition, although I should be able to rationally justify most of my points - bear with me). I will use the kata Sanseiru and Jitte as examples as I go, as most people (all 7 of you!) should be familiar with at least one of these.

Firstly, you need to look at multiple versions of the kata, from different generations within a group (where possible) and from different lineages.  In doing so, you are looking first for commonalities:  What are the things that are the same?  In Sanseiru for instance, every lineage I have seen has a sanchin-derived starting sequence, a clearing of the hands, an attack to the opponent's leg, two steps/kicks, 4 directions of elbow/punch/kick, a double sequence of the hands going downwards while in shikodachi, a double sequence of hands going out, back with a sweep and morote-zuki, and finishes with a dog/crane posture.  This universality implies a common origin for each of those sequences; they are either plesiomorphic (all inherited from the original form) or apomorphic (all inherited from a common ancestor that was an alteration of the original form - one step removed from the original form).  Either way, it is safe to say that the gross morphology of Sanseiru is close to its original embusen, as there is almost no variation between any of the versions in the overall sequencing.

In Jitte, on the other hand, there are fewer commonalities.  All start from clasped hands , open out the hands, and then have a swapping of hands to the left and then to the right (A), followed by three steps forward in shikodachi (B).  Then, clearing crossed hands with a step back in shikodachi (C), followed by a number of steps forward with shikodachi, in yama-kamae (D).   Here, commonalities become fewer.  In some, there is a sequence to the right with mawashi-uke (e), followed by a sequence to the front and rear with jodan uke (F), followed by two more sequences (g, h) before the close of the kata.  In others, there is instead a 180 degree turn to the rear into a triple sequence of sweeps and open hand attacks (i) followed by a repeated manji uke (j). Only then do they finish the kata with the same sequence of jodan uke (F).  So the original form probably contained the sequence A-B-C-D-F in that approximate order.  The existence of e,g,h,i,j in the original form is un-corroborated.

In the Kyudokan version (the one that I practice - I just don't have a video to hand of me doing it), the sequence is A-B-C-D-e-F-g-h-g.  In Shotokan, the sequence is A-B-C-D-i-j-F.  Of the unshared sequences, is there any way to tell if they were part of the original?

One way is to look at degrees of separation from the creation of the kata.  This can be somewhat difficult (to put it mildly).  For Sanseiru, the earliest confirmed practitioner of the kata was Kanryo Higaonna, who if not the creator of the kata, learned it or its progenitor in the 1870s-1880s.    All subsequent variations of the kata come from his students and their students.  Higaonna's two main students were Juhatsu Kyoda (founder of To'on Ryu) and Chojun Miyagi (founder of Goju Ryu), although others, such as Seiko Higa and Kenwa Mabuni (founder of Shito Ryu) learned from him - these are the 2nd generation practitioners of the kata.  Their direct students were the 3rd generation (and I think only a handful of those are still alive), but given the disruption of the war could potentially be considered to be generation 3a (pre-war) and generation 3b (post-war) as what was taught and how it may have altered due to the disruption that occurred (see part 2 for more on this).  Their students (who make up a lot of the current Okinawan seniors in goju now) are the 4th generation, although some of the students of the 3a generation (such as Kanki Izumigawa) were teaching pre-war and so their generational succession is temporally out of step with those of the 3b generation.  Their students are the 5th generation, and so-on.  Myself, I am a 7th generation practitioner of Sanseiru, through the Higa line, although I might be the same age as a 5th generation Jundokan practitioner, as Higa's student Izumigawa started training his own students pre-war.  Does this mean that the Jundokan's version of Sanseiru is closer to the original?  Well, it's more complicated than that.  Below is me doing our version of Sanseiru, and a random (and very nice) Jundokan example from youtube:

They are obviously the same kata and also obviously very different.  Some differences (such as the double punches at the start or the open-handed age-ukes before every front kick I did that were added  by Ohtsuka sensei in the 1970s) are deliberate modifications made by later generations; but some may be due to codification within either school where each kata becomes homogenised with the others in their performance, or due to some parts being made overt in one school that are hidden, de-emphasised or made more subtle in others.  But the biggest two issues that make it difficult to identify which is original is the potential for transmission errors to have occurred at any point in the chain of transmission across the generations of performance and our inability to identify if that occurred; and the problem of the cultural equivalent of genetic drift and of different selection pressures acting on kata.  The uncertainty these two issues bring is compounded too by their ability to affect kata within generations, as well as between.

I have written on the topic of genetic drift before and what I said there is of significance to understanding why we will probably never know what the original of a kata truly was.  In the biological world, selection pressures cause some variations to be favoured over others, creating directional change within a population over time.  But if the selection pressure disappears, then all variations are equally likely and so over time, more variations will arise and become more common, swamping the original form.  It can be argued (I am arguing it!) that kata have not been subject a single, directional selection pressure for many generations, at least not to the selection pressures (the need to train and transmit fighting/self-defence knowledge for likely use in a life-or-death environment) that influenced their formation.  We know from the writings of such people as Itosu and Miyagi that the rationale for learning and transmitting karate was changing in the 1910s and 1920s, altering the selection pressures that would normally assist in keeping variations in kata to a minimum, and creating an environment in which multiple variants of the kata could arise, thrive and be passed on.  The multitude of differing environmental pressures increased with the spread of karate beyond Okinawa onto mainland Japan and further out to Europe, the Americas (Sth America and Hawaii likely before WWII, the USA afterwards) and beyond.

The point is, that this loss of a single, shared selection pressure has affected all current schools of karate - and there is no easy way to untangle what changes this has wrought on each school's kata rendition.  Large organisations are more likely to have had changes caused by the need to standardise and teach to large bodies of students, but are less likely to be have been affected by the whims or idiosyncrasies of a single individual; small groups are more likely to be the opposite.

And this is why just looking for the most common way that things are done is not a sound way to judge what is original - we have no way of knowing if a feature of a kata has been lost by one and kept by the many, or created/shared by the many due to standardisation, cross-pollination (a common feature of human social enterprises where something developed by one can be co-opted by others.  The inclusion of mawashi-geri into most schools of karate is an excellent case in point) or through a later, shared selection pressure that the solo school lacked exposure to.

So is there any hope of trying to identify the form of the original of any particular kata?  Probably not, but there may be ways to at least partially reconstruct it.

• Firstly, as mentioned above, the presence of universal commonalities can give us clues as to what is included.
• Secondly, the use of interpretation principles (such as the kaisai no genri of goju ryu) can allow us to interrogate the applications associated with the kata. If there is a disconnect between the form of the kata and its application (assuming the application follows the appropriate principles), then there is a possibility that the kata has undergone some change.  
• Similarly, using Patrick McCarthy's concept of HAPV and looking at how kata movements correspond to dealing with those, may indicate where changes have occurred (for instance, a school with a number of kata where there is no dealing with attacks to the head, compared to another school with the same kata).  Or, if the omote applications all follow the movements of a kata closely, but all only work against karate attacks (lunging punches, high kicks, roundhouse kicks), then it is likely that they are not original.
• We can also look to places where karate was exported, then isolated from the societal changes/pressures that occurred in Okinawa, Mainland Japan or in the US or Europe, for the cultural equivalent of living fossils.  How are the kata done in the Okinawan/Japanese diaspora in South America or Hawaii for instance?  A comparison for differences between mainstream and isolated versions may lead to some older variants being identified.
• Lastly, since most modern schools of karate have kata from different sources, those kata that start or end in a unique fashion are more likely to have always had those beginnings/endings, whereas those with the same starts or finishes are more likely to have been standardised.  Sepai and shisochin kata in goju ryu are good examples of kata where we can have a fairly high degree of confidence that their original versions contained those ending sequences.

In the end, without a time machine, there is no way to answer this question with a high degree of certainty.  A better question to ask may be "how close to functional is your kata?", as the original kata will have been developed with a particular function in mind.  Since none of us are preserving kata unchanged in their physical expression, maybe we should be more concerned with preserving the intent that the kata are meant to convey.