Kata Saifa Analysis
I’ve heard this kata taught hundreds, if not thousands of times, in GKR, Goju, and other styles and there are so many possibilities that it seems as though no one can claim a definitive bunkai. Hardly surprising as the kata has been passed by word of mouth from Chinese to Japanese to English, and has been subjected to over 125 years worth of potential distortion, revision, and misunderstanding.
According to Sensei Mario McKenna the kata seems to have derived from a form of Chinese Boxing called Lion Fist, which is one of the
I’ve spent many, many hours thinking about this kata, and here are my thoughts about the bunkai, which are numbered to link them to the footwork chart, and the most common performance count.
There are at least two different versions of this kata performed in Goju a version almost identical to ours which seems to be the one practiced in Goju Kai, and a slightly different one practiced in Goju Ryu. Where I refer to differences, I’m obviously referring to the one that is not virtually identical, before all you Goju masters start emailing me with your complaints! My reference source is the acclaimed Goju Technical volume 1 DVD as demonstrated by 9th dan hanshi Morio Higaonna. The only reason I choose the Goju version is the fact that we have the closest links to the Goju lineage for this kata.
I hope that some of you may find my analysis helpful.
In GKR the response to the grab is usually a 90 degree turn to the left and a big step up to grab your own hand. In Goju, the step is performed diagonally to the right. Either way, you could interpret this first movement as classic tae sabaki to avoid a punch. The Goju version moves both the head and body out of the way, provided you assume that the attacker is delivering a straight punch. In the GKR version you appear to just move the body, however, if you assume a hook punch or haymaker, then neither version moves you away from the arc of the punch. If performed as tae sabaki, the first step should probably be executed quickly, however, you could argue that rather than quickly moving out of the way of a punch, you’re slowly moving towards an attacker, and the slowness of your movement is designed to lull them into a false sense of security so that the subsequent escape is even more explosive and surprising to them.
In moving closer to the attacker, you also re-enforce your arm in preparation for an escape. If your arm is far from your body, it is weak. If it is close to your core, it’s much stronger. In ideal circumstances, you’d pull your arm towards you, overbalancing your opponent as you do so. However, if he is naturally stronger than you, or his arm is close to his body, you’ll need to move towards him to bring your fist close to your core.
By adopting high stance (heisoku dachi), we give ourselves a greater variation in height when we perform the subsequent escape. Thus, as we pull away, we are able to drop our weight, thereby adding our maximum body weight to the arm pull.
Or so we’re told. Only it doesn’t make sense. Why would you use your body weight (applied through the low stance) after you’ve tried to escape? Wouldn’t it make more sense to synchronise the pulling, stepping away, and sinking motions to create a more powerful single escape? Why practice an escape that you know will fail (as evidenced by the need for the subsequent teisho)? Furthermore, a circular palm heel is far less effective than one delivered in a straight line. At that angle, the left arm is more powerful delivering a cross-body punch with the palm heel, than a descending smack.
9th dan Goju master Morio Higaonna shows the bunkai differently. He still steps before he pulls, which is pointless, but at least in his bunkai, the arm pull is successful in dislodging the attacker’s grip. If you try it with a partner, you will find that that is usually the case too.
A far more useful bunkai would be as follows:
These steps are repeats of 1 and 2 using the opposite arms. Why the repeats? Perhaps because the concept of using your body weight is so important?
7 & 8.
In all versions of the kata, this sequence starts with a sharp head turn to the right, and therein lies the problem. This indicates turning your attention to an attacker in that direction. This implication is further reinforced by the step to the left, which suggests moving away from a big kick. This part of the sequence is what makes it seem as though you are dealing with two attackers.
However, Morio Higaonna demonstrates in his bunkai, that you are in fact dealing with a single attacker, who kicks you with a left leg mae geri and then punches you head level with his right hand. This is a classic combination, and seems entirely plausible, except for the head look.
So, why the head look? Well, I see a few possibilities:
This is quite possible - kata have had all kinds of twiddly bits added for competition, neatness, or simply because the practitioners didn’t know the proper bunkai
2. You are indeed avoiding a kick, but the subsequent head level block, is simply another stage. It’s not intended to denote simultaneous attacks, just another attacking principle.
Whether fighting two people simultaneously, or sequentially, you have to ask, “What happened to the kicker after he was blocked?” Also, does it make sense to move blind towards another attacker? It’s hard to find a plausible explanation for these two questions.
3. There is an entirely different bunkai that we are unaware of.
Possible, but to quote Ockham’s Razor, “One should not increase, beyond what is necessary, the number of entities required to explain anything.” In other words, don’t look for a complex solution when a simple one will do
4. The moves are symbolic, and are not meant to be taken literally.
I don’t believe there’s a single symbolic move in any original kata. Moves may ignorantly have been added, but as this sequence seems to bear a great resemblance to the original kata as brought from
Given the fact that a gedan barai is such a poor defence against a front kick, using a low level palm heel block to guide the leg past the body makes much more sense. After all, why meet force with force? My only reservation with this explanation, is the fact that we pull the right hand too far back as we block after all, all we need to do is steer the kick past as a cover. Hopefully, we’ve already evaded. Perhaps we simply over-block in the kata, which is a bad practice.
Now the way that we perform this sequence in GKR, slightly muddles the next part of the bunkai, because we step, block and lift the knee at the same time. In Goju, they step, then block, then raise the knee. This makes more sense because you avoid the kick, guide it past you, then knee the attacker in the groin. Using the GKR timing, your counter-attacking knee, would get caught on the attacker’s leg as you guide it past you. I attribute this discrepancy to the tidying up of the kata that has occurred in the interests of sharp (bunkai oblivious) tournament performance.
In Goju, there is a distinct kneeing action, then a pause, followed by a kicking action. This makes absolute sense, because you can reasonably assume that the knee either bent the opponent over, in which case they get a kick in the face, or it drove them backwards, in which case the kick has the extra range to catch them before they escape.
In Goju, the palm of the left hand is facing upwards, rather inwards, and it’s more of a rising, less of a circular motion. In their version, the hand rises underneath an attacking punch performing teisho age uke according to Higaonna. This is a preposterous principle to practice, especially pre-black belt. It uses the most inappropriate block possible, giving you the least chance of successfully preventing yourself being hit. Worse still, you need to remember that you’ve just moved into the line of the punch! Okay, so in this version, the hand is grabbed at the same time, which is a useful control device, but it’s very risky.
The GKR version is much better. By swinging the arm vertically past your face, it gives you a far bigger blocking area (the length of the forearm), and it moves the punch away from your body, rather than trying to intercept it and stop it dead. If you wanted to, you could then seize the attacking punch, if needs be, to stop the attacker from retreating away from the kick.
However, this section finishes with a totally different sequence to the first time.
In GKR, we cross our hands in front of us and step back into forward stance, drawing our hands to our hips, before shooting them out in a double punch. In Goju, they simply perform the double punch (or kiraken according to some sources), without the initial crossed arms sequence. In Lion Boxing, this move apparently represents a lion pulling down its prey. That makes me wonder if maybe the kata has become corrupted or refined since its original creation after all, lions don’t cross their legs when they pull down an animal...
Anyway, both the GKR and Goju versions makes sense within the context of their own bunkai, and for this one, I prefer the bunkai that you can extrapolate from the GKR move.
Crossed hands like that may be used to deliver a strangle, but it’s a very risky proposition because the opponent has both hands and legs free, and it’s highly likely that, with both hands at the opponent’s neck, you will be struck long before you complete the strangle. In any case, that type of strangle utilises the opponent’s clothing. If they’re wearing the wrong sort of clothes (such as a t-shirt or no top at all), it can’t be applied.
So, now you pull the opponent towards you forcefully which in itself can cause whiplash. You use the retreating direction and lowness of the zenkutsu dachi stance to add power to the pull. You could fairly argue that shiko dachi is a better pulling stance, but it’s risky front on because of the groin exposure.
Now, at this stage, if your hands are behind their neck and you’ve pulled them in close, you can apply an extremely unpleasant necklock by pulling their head against your shoulder. In real life, this could be the end of the fight, and the subsequent double punch is simply to dispense with the body. However, as the final move in the kata shows, we tend to push bodies away, rather than punch them, so let’s assume that we didn’t break their neck for now!
Higaonna demonstrates this move as a finger tip strike to the nerve points (lung 2 in pressure point terminology) just below the clavicles. These nerves are extremely sensitive, and generate a large amount of pain. The optimum angle of attack against them is in an ascending motion, and as the Goju version doesn’t precede the attack by pulling the opponent forwards, you’d be in the correct position to deliver such an attack. Such an attack would temporarily incapacitate an opponent without long term damage. However, a fingertip strike requires conditioned hands and precision to avoid injury to the person doing the strike, so I have to wonder about the wisdom of such an attack. If you wanted to attack those nerves, a bear paw knuckle strike would be safer. Returning to Bill Burgar’s book, he reasonably in my opinion, points out that kata moves that require years of conditioning are probably being misinterpreted. Thus, I tend to dismiss Higaonna’s explanation.
In the GKR version, the attacker is leaning forwards, which enables you to attack the suprascapula nerves above the clavicles, or indeed to smash the clavicles. It depends whether you are seeking to cripple or disable...
So, in our version, we’ve just incapacitated both arms. The opponent is unable to defend himself, or counter with his arms. That means we have time to use a more time-consuming, finessed finish. So we reach up and grab the opponent by the hair, which produces immediate pain compliance, and enables us to control the position of their head quite effectively. Now, we pull the opponent in close to increase our leverage, then we force him down and turn his face towards the other hand, which we smash into the face using a hammer-fist.
It’s a strange finish, considering the fact that they’re helpless. A simple punch in the face would be faster and just as effective, except for the fact that this version also forces them to the ground, so you don’t have to worry about kicks once they’re down but look out for knees on the way in. I would have thought that with both shoulders paralysed they’d already be completely ‘armless. Sorry.
I’m not happy with any bunkai to this move. It would make more sense to move into a right-leg back shiko dachi, which provides a more powerful downwards and pulling action. And of course, the whole face smash is less than optimal. If you have a better bunkai, let me know.
Our way of turning (keeping both hands in contact with each other as you turn), doesn’t make any sense in the context of a behind the head grab, and it forces you back to that odd cross arm lapel grab. The only possible reason I can see for gripping somebody with arms crossed like this, is the fact that it tends to deter you from sticking your elbows out when you pull. It seems stronger that way, but I don’t know for sure.
Now we come to the leg lift, turn and descending hammer fist. In GKR, we execute a very high leg lift, and turn 180 degrees. In Goju, the leg lift has been substituted for a sweep, and they only turn 90 degrees, so that they are standing sideways on to the opponent. Perhaps they do this to minimise body exposure, or to speed up the hammer-fist. By only turning 90 degrees, th
eir retailiation certainly comes quicker. In any case, in the context of the next move, this seems like an odd way to do it given that they use exactly the same bunkai as us.
The attacker moves in behind us and grabs our hair. In an effort to bring us to the ground, he either kicks us in the back of the knee or stamps on the calf of the rear leg. This is a common jujitsu takedown, usually accompanied by drawing the other hand across the throat in a choke.
In response, we immediately withdraw that leg and lift it out of the way. At the same time, we raise our right arm high above our head so that our upper forearm is pressing down onto the top of the attacker’s wrist. Because his hand is in a fist shape, this acts as a lock, preventing him from pulling our hair. Furthermore, as we turn, it applies a mild wrist lock. More importantly, we ensure that our elbow is higher than his hand. Now, as we turn, our upper arm serves as a lever to pull his hand away from our head.
Why we don’t simply use the raised leg to deliver a back kick to the groin or knee, I’m not sure. Perhaps it’s because the opponent’s stamping leg acts as an effective cover for his groin, or possibly because we used a hair grab as a compliance technique in the previous technique, and now we’re learning how to deal with it if it’s used against us...
As we deliver the hammer-fist, we bring our foot down hard and bend our knees, landing in sanchin dachi stance. This enables us to add the downwards motion of our body-weight to the strike. The stance is ideal for close in combat, because it enables us to quickly turn away to protect our groin. Incidentally, Kinjo Akio suggests that this movement represents the powerful steps of the lion.
In Goju, they execute a longer-range hammer-fist, which is certainly safer, especially in conjunction with their sideways-on stance. The stance also diminishes the risk of them receiving a knee in the groin. However, we perform ours closer, with a more acutely-bent elbow, which I think delivers more power. Given that it’s a brute force smashing technique, I prefer our version, although I certainly see how ours leaves us far more exposed to knees and other close-quarters counters.
However, given the speed and power of the sequence, and the fact that the opponent will soon be leaning forwards, his opportunity to use his knees is very limited.
The bunkai is easy enough to see; we grab the opponent’s hair, pull him down, and punch him. The theme of hair grabbing is continued, just to demonstrate what a useful control technique it is. Perhaps we should spend some time practicing the best way to grab the hair? I imagine it’s most effective if you grab close to the roots. It certainly enables you to control the head more directly.
In GKR, we call the punch shita tsuki, which actually means underneath punch. The way that we practice this punch in the basics is to tuck the elbow in to the body, where it acts as a brace against being pulled in. However, this is primarily done for defensive reasons. Using it like that when we’re on the attack, such as in this situation, it doesn’t make sense to do so.
Furthermore, because our short punch comes out parallel to the ground, it means that we hit the attacker in the stomach at right angles, which whilst delivering the most power, is also the least damaging angle because it spreads the force of the punch across the flat of the knuckle rather than using the seiken knuckles to deliver a more penetrating punch. The punch would be more effective rising up and hitting them in the face or solar plexus, which are both easier targets to damage. However, we don’t perform uppercuts in GKR until kata seiunchin, so a short punch it is...
I think that the optimal execution of this move would be straight-on like we do it, but with an uppercut to the face.
Simple inverted repeats of 13 and 14, presumably to remind us of the need to be ambidextrous.
In GKR, we’re told that the slow deliberate punch is simply pushing the opponent’s body away, and that’s conceivable. However, why would you practice slowly punching a lifeless body? I wonder if there’s more going on here than meets the eye.
Direct your attention to the left hand, which slowly closes into a fist. Now think about the reaction of an untrained person to the hair grab you did in the previous move they probably reach up to grab that hand. Perhaps you’ve transferred grip on to their wrist and are now performing a wrist lock? Or maybe you’re still holding the hair and are re-securing your grip after punching them?
And what about the punch? Well, given that you were holding the opponent down at the time, it’s reasonable to assume that they are still in that position, especially if you delivered a punch to the stomach.
Now just because a move is performed slowly in kata, it doesn’t mean that that’s how you would do it in real life slow movement in a kata usually denotes one of two things: that the move is technical and should be executed in a precise way, or the move requires strength. With this move, the latter may be true but why at that point in the combat? Perhaps we're deliberately creating distance ready for the next punch, but more likely, we're using the slow push, not so much to create distance, rather to judge it. Like fixing a sight on a target.
It could also be true that the punch is quite specific. Given the viciousness of the previous two moves, we’re clearly playing for keeps at this stage, so it wouldn’t be out of character to deliver a punch to the exposed nerves in the side of the neck. Or again, the clavicle bones on the opponent’s left side make a tempting target...
And why the funny sanchin step, with its exaggerated suri ashi sliding movement? Well maybe that’s simply a reminder that we’re still at close quarters and need to protect our groin? Or possibly we’re pressing our leg against the inside of the attacker’s left knee to further unbalance him?
19 & 20.
We start by turning 180 degrees into gyaku nekoashi dachi and grabbing the opponent by the top of the head and the bottom of the chin. It’s an odd grip, and perhaps the ears would provide better pain compliance, although, they may be harder targets to grip in a hurry. However, to see the value of the grip we use, you need to remember that this kata was devised in an age when males wore their hair in a kind of vertical ponytail style called a topknot. They also traditionally wore beards. This provides two robust “handles” that are both easy to grasp, and easy to hang on to.
In GKR, once we have taken a grip, we take a big step backwards, pulling the opponent with us. As we go, we allow the left shoulder to move back faster so that we are at 45 degrees as we move. As we move into a full nekoashi dachi, we draw the opponent’s head hard towards the left shoulder, and at the same time, we drive the shoulder forwards, delivering a shoulder smash to the face. This stuns the opponent just long enough, so that we can instantly break his neck with a two-handed twisting action.
In Goju, they don’t take the big backward step or use the shoulder smash. In fact Sensei Higaonna suggests that the pulling motion that we interpret as a head pull, is actually a block with the left hand and a haito with the right. I respectfully disagree. A haito to the ribs is powerful, but it’s such an un-optimal response to any situation, it bewilders me why anyone would want to practice it. Why wouldn’t you simply perform a chudan or jodan punch once your block has opened the opponent up?
One thing I did puzzle over, was the reason why we turn 180 degrees at the start. Does it signify a fresh opponent? I think it does. I have to admit, the hair and beard grab is a risky thing to attempt on an opponent who still has all his senses, but then the turn doesn’t make sense as a continuation to the previous moves.
The initial stance is interesting too. On one hand, it’s an unstable stance defensively when you’re that close to the opponent, but if you assume that the opponent is already in defensive mode, so you’re in control of the situation, then it works better. The stance seems well designed for a powerful backwards movement (which coincidentally, is also how the same stance is used in kata seiunchin). The flat front foot gives a strong, stable base from which to push away, whilst the bent rear foot provides a store of potential energy the foot naturally wants to slide backwards, and all you have to do is lift the pressure from the toes to start moving in that direction.
In order to properly push somebody away, you need two opposing points of the body. By which I mean, two points, which, when pushed individually, make the other point move forwards. For instance, if you push someone in the throat at the dimple where the throat enters the sternum, the person leans backwards, thrusting their groin forwards. If you shove the groin, they lean forwards, bringing the throat closer. If you shove both points together, the counter-actions cancel each other out, and the person moves away.
So, at the end of saifa, you can either choose to press the throat and groin, or, you can push the opponent’s right hip and left shoulder. Now given the fact that the person is dead because you just broke his neck, you’re simply trying to move a dead-weight. Pain compliance won’t work, so there’s no point using throat and groin. So you go for the other option. Because the hip and shoulder are both bones, they provide a rigid structure with which to push the body away.
With very greatest thanks to Senseis Jane Manktelow and Chris Sayers for their help with the photos.