Castling - Rook touched first

Running Chess Tournaments
201

Well, maybe the analysis is bad, maybe not; but I don’t believe that the intent was to be as has been as described by you. It would appear that 8A2 addresses the point you make, because it defines castling as a single move - thus considering the move parts separately and treating them as moves in the way you do above (i.e. Rook released first constitutes a move) doesn’t seem to hold. This, I thought, was the point of the later clarification of what happens if the order is inverted. I agree, it would be more clear if it had statements about “touched” and/or “touched and released”, but your analysis doesn’t hold (i.e. 8A2 defines only King first castling) since there is later a clear exception to the King be touched first - which means it is no longer executed as King first castling as described in 8A2 even if the King is released first. Hence the insistence that 8A2 defines the ONLY way that castling can occur is clearly false.

So no, I don’t think your analysis follows. I do think that many of us believe it would be more clear if it were as you state - but I don’t think the consensus is that it currently is as you state.

202

There is no later “clear exception” unless you read “touching” to mean something other than what touching a piece means in the rest of the chapter. “Touching” a piece, means to grasp the piece in a manner reasonably interpreted as the beginning of a move. See 10B.

In order to believe that 10I2 should be interpreted the way you are saying, one must believe that a chapter concerned with nothing other the nuanced distinctions between touching, transferring, releasing, determining, and completing moves, completely botched the use of the word “touching”. You would further have to believe that 8A2 is completely purposeless. It very clearly spells out what castling is, and if rook-first castling existed, it (and 8A2) could easily have expressed this fact. There would be no reason to use the language “intending to castle” as if it hasn’t occurred yet (of course, that’s exactly what it does mean). You would also have to believe, that a chapter that is meticulous about cross referencing (9A references 7A which references 8A2) just happened to add an exception that completely contradicts the plain meaning of both 7A and 8A2 without mentioning it in those places.

Or you can interpret it to mean exactly what it says. That if you touch the rook within the meaning of 10B and you were intending to castle, that you still can castle by following the procedure of 8A2.

At the very least, you cannot point to anything that says I cannot reasonably interpret 10I2 in the way I am. Courts of law are asked to interpret rules all the time. A well accepted principle of interpretation of rules, statues, policies, etc… is that if it’s possible to read rules in such a way that they don’t conflict, then you should do so. My simple reading of 10I2 is very straightforward: how to apply touch move when you touch the rook first and haven’t yet determined your move by releasing it. Given that this interpretation is possible and is consistent with 7A and 8A2, the rule of interpretation saying not to invent ones that require you to believe the rules contradict themselves when it’s possible to avoid doing so applies.

203

It is a clear exception since the initial rule for castling states a specific procedure which the “touching” rule clear over-rides.

Not at all. One need only believe that it differs from the specific castling procedure outlined.

Exactly. So, when there is a later reference talking about when castling is executed by touching the Rook first, it should be considered rocket science to see that this creates an exception.

I already have pointed to it. You just don’t like the result.

204

It doesn’t clearly override anything.

Suppose I reach for the rook on h1, grasp it in my fingers, and then let go of it without moving it.

At that point, I grab the bishop and make a legal bishop move.

If there are legal rook moves, that’s a violation of the touch move rule. I touched the rook, so if I can, I have to move the rook. I can grab it and let it go again if I like, but eventually I am not allowed to make any other move if there is a legal rook move. Touching the rook was not part of executing the bishop move. It was something that happened before the bishop move, but which affected the legality of the bishop move.

Suppose, on the other hand, I touch the rook, and let go of it, without moving. I then execute the procedure described for castling in 8A2. Touching the rook was not part of that procedure, any more than touching the rook was part of moving the bishop. 8A2 describes how to castle, and that’s how I castled.

Simple enough, but was it legal? When I moved the bishop, I followed the rules for moving the bishop. I moved it to a vacant space and released it. However, because I had touched the rook, I couldn’t do that. When I castled, I followed the rules for castling, as described in 8A2. However, prior to castling, I touched the rook. Can I still castle? Yes, I can. How do I know that? Because 10I2 tells me that.

10I2 doesn’t override any previous rule. The procedure for castling is described in 8A2, and that is the correct way to castle, whether or not you happen to have touched the rook prior to executing that procedure.

Mr. Taylor has described, accurately and completely, what the rule book says.

It gets confusing, because most of us happen to know what the authors of the rulebook were trying to say. We know that the intent of the rule is that it really doesn’t matter what order you move the pieces around in or when you release them. We know that you can pick up the rook first and let it go or the king first. You can pick up the rook, and then the king, and then a previously captured pawn and juggle them for thirty seconds before putting all the pieces into their castled state with the pawn off the board, and as long as it looks like you meant to castle, we know that the authors intended it to be legal to castle.

That’s what they intended, but that’s not what they wrote.

205

Fourteen pages later, this Bear of Very Little Brain sees no reason why the Official Rules of Chess should allow castling by touching the rook first.

206

I agree. Many, many pages upthread I asked if an ADM was in the making. What does Rules think about this?

Alex Relyea

207

You’re not hearing what I am saying. Stop thinking of examples, step back and look at the bigger picture:

Rule X defines THINGAMABOB as way A and ONLY way A.

Rule Y after rule X now discusses an additional way for RULE X to occur that is NOT way A.

CLEARLY Rule Y provides a way of overriding Rule X. Forget what the meaning of X and Y are - if Y defines an additional way to do X, then clearly X is not the ONLY way, even though it says it is.

208

I’m listening to what you are saying, but what you are saying happens to be incorrect.

In this case Y=10I2, and X=8A2. If rule Y (10I2) described an additional way for rule X (8A2) to occur, you would be right. However, it doesn’t describe that. Rule Y describes an event that might precede the events described in rule X. In most cases, a touched piece means that a player must execute a translation move with the piece that is touched. However, rule 10I2 creates an exception. If a player touches the rook, the player may execute a translation move with the touched piece, or he may castle. It doesn’t say how to castle. It just says it is legal to do so.

10I1 creates another exception. If the player touches the king and rook at the same time, and castling with those two pieces is illegal, he may execute a castle move with the other rook, or a translation move with the king.

Strictly speaking, even that interpretation of 10I1 is not what the rulebook says. The rulebook describes a situation in which “castling is illegal”, and then goes on to describe that he may castle if it is legal to do so. Clearly, you have to fill in a certain common sense understanding of what they meant.

Meanwhile, what happens if I touch a pawn and a bishop simultaneously? I don’t think the rulebook covers it. I would interpret it to mean that you have to move one or the other. Now what if you touch the king and the rook simultaneously, and castling is legal? The rulebook also doesn’t cover that. You have to infer based on the spirit of the rules elsewhere that, having signaled what appeared to be an intention to castle, that you must do so. That’s an interpolation, though. It isn’t in the rule book.

209

It does.

Please note that I’ve previously noted that this is not my preference. However, this is clearly what is there.

Key points (without positioning detail):

  1. Move of K+either Rook
    2. Single move
    3. Executed as King is transferred first
  2. Then Rook is transferred.

Key point:

  1. Rook touched first.

This overrides point 3 above, and HENCE what I described holds, since this thereby describes an alternate way of castling (Rook touched first.)

210

The intention of the rule is to override, point 3 above, but the words used do not do it.

The key point number 1 of your second rule quotation is that, in order to interpret it the way you do, the word “touch” has to also include “move” and “release”.

10I2 says that you can touch the rook first. It doesn’t say you can move the rook first. It doesn’t say anything about releasing the rook first. It says what happens if you touch the rook first.

You are looking at what you earlier called “the bigger picture”. In this case, you are looking at the picture beyond the rule book. You know what they intended, and you are looking at the words, and fitting them into the context of what they intended. You also know that the rules, as written, would be stupid. It would be very weird to have a rule that a piece could be touched without making a commitment, but it could not be released without making a commitment. That’s dumb.

But it’s what the book says.

211

The words DO do it. You’re missing a second key point above. It is utterly unnecessary to include “move” and “release”. Those are concepts that describe a move - NOT a “half-move” (if you’ll excuse the coining of a term.)

READ AGAIN. Castling is a SINGLE MOVE. Moving the King or moving the Rook first does not require that one fulfill the requirements of a “move”. Only the ENTIRE CASTLING PROCESS, which is a SINGLE MOVE, so requires it.

Yes.
It doesn’t have to.
It doesn’t have to.
Yes.

The Rook half-move is NOT a move. This is where you are confused. That it is SILENT on the release simply indicates that we know nothing, not that we are so prohibited. Why would you assume that we are?

Not at all. It’s you who have added the requirements of a move to the repositioning of the Rook - but by the very definition of castling, it is not a move. So your fundamental assumption is incorrect.

The King repositioning and the Rook repositioning, taken together constitute ONE move. The standard procedure defines what should happen. The alternative defines what happens if the Rook is touched first. Since the Rook repositioning IS NOT A MOVE, release doesn’t matter.

There is a commitment. Your assumption above is incorrect. Its just a different commitment than what you think it is.

212

Kevin, you can really be insufferable sometimes. There is no need for you to drive home the same point repeatedly, like a pit bull, when somebody continues to disagree with you. It’s not necessary for you to always have the last word.

We should all keep in mind that castling consists of six distinct actions:

  • K1. Touching the king (on its original square, e.g. e1).
  • K2. Transferring the king (to its new square, e.g. g1).
  • K3. Releasing the king (on its new square).
  • R1. Touching the rook (on its original square, e.g. h1).
  • R2. Transferring the rook (to its new square, e.g. f1).
  • R3. Releasing the rook (on its new square).

Obviously, K1 must precede K2, which must in turn precede K3, because it is physically impossible to transfer a piece before it is touched, or to release it before it is transferred. The same applies to R1, R2, R3. This greatly reduces the number of permutations of K1-K2-K3-R1-R2-R3 that need to be discussed.

FIDE insists on K1-K2-K3-R1-R2-R3. They might also permit a few other permutations, such as K1-R1-K2-R2-K3-R3.

USCF wants to also allow R1-R2-R3-K1-K2-K3 (and maybe a few screwy ones like R1-K1-R2-K2-R3-K3) and to insist that, if R1 precedes K1, castling is considered a rook move, with the consequence that if castling is illegal, the player must move whichever piece was touched first.

Personally, I think most of the FIDE-USCF rift in this area is ridiculous. USCF should move closer to FIDE, by always considering castling to be a king move. If a player castles illegally, he should be required to move the king if legal, regardless of which piece was touched first, transferred first, or released first.

On the other hand, FIDE should be more forgiving. An arbiter should have the option of deciding, case by case, whether a player who touches his rook first should be allowed (and required) to castle.

Bill Smythe

213

It’s not about having the last word, or driving home the same point. It was about the missed point, a point that had not been previously spelled out and examined. In doing so I even coined a term, thus emphasizing that the point hadn’t previously received focus.

I’d appreciate it if you wouldn’t project YOUR feelings about the situation onto me in the future. Thank you.

I agree that it should be a King move, and I agree with a simpler definition. My feeling is we should simply require K first. It makes more sense in historical context, it makes more sense in terms of the rules of the game, and I don’t see any objective hardship to requiring it.

My point above was that castling under current USCF rules is defined as a SINGLE move, so the analysis being performed above was faulty, and the assumptions about release - AS WORDED - are faulty. I don’t see where making the point to understand something makes the attempt “insufferable.”

214

That’s a good framework for describing the problem. Let me continue with this in order to illustrate the problem with the rulebook.

(As an aside, it’s not a massive problem, but it does lead to a couple of strange issues. In particular, it created an issue with the certification exam. I think the members of the rules and td certification committees have been informed of the problem. The remaining 13 pages of discussion are just for fun.)

First, let’s generalize, so that X1, X2, and X3 refer to touching, transferring, and releasing a piece. In the castling example, X takes on the values of K and of R at various times. Second, I will add a new element, X*. X* refers to releasing a piece on its original square.

The essence of the touch move rule is that if a player executes X1, he must follow that by X2 and X3 if it is legal to do so, and he must do so for the first value of X for which he executes X1. In other words, the sequence B1, B*, N1, N2, N3, is an illegal sequence. N1, N*, N1, N2, N3 is a legal sequence.

Rule 9A says that if a player executes X1, X2, X3, and that sequence is a legal sequence (i.e. the transfer is to a legal square for that piece) then the move has been determined with no possibility of change.

Rule 8A2 says that castling is executed by performing K1, K2, K3, R1, R2, R3.

What does 10I2 say? It refers to touching a piece. Touching is step 1. A literal reading of 10I2 says that the following sequence of actions is legal: R1, R*, K1, K2, K3, R1, R2, R3. (Note this is not an exception to the touch move rule. It’s just a clarification. Having executed R1, he also executed R2 and R3. There was simply an addition sequence, R*, K1, K2, K3, R1 that was inserted.)

So there is a perfectly straightforward interpretation of the rules in which all rules are completely consistent. That interpretation of the rules requires no additions or assumptions. Just read the words and follow the directions.

Unfortunately, it is not what the authors intended.

A literal reading of the rules, as described above, would say that if you execute R1, R2, R3, you have completely determined a move, according to rule 9A. Since there are no exceptions in the rules, that’s it. That’s your move. That contradicts the intent of the authors. The actual intent of the authors was to allow R1, R2, R3, K1, K2, K3 as a legal sequence.

An awful lot of people, including the authors, thought they had made that a legal move by writing rule 10I2. What they meant to say in 10I2 was that the rook could be touched first (R1), or moved first (R2), or released first (R3). That’s what they meant to say, but they didn’t say it. They confined the text to R1.

So, bwtaylor has simply read the rules, added no material, and assumed that they mean exactly what they say. His interpretation is the most straightforward, obvious, and literal reading of the rules possible.

215

No.

A literal reading of the rule says that any of these (and perhaps others) are legal:

  1. R1, R*, K1, K2, K3, R1, R2, R3
  2. R1, K1, K2, K3, R2, R3
  3. R1, R2, R3, K1, K2, K3

So long as intent to castle is clear.

Under what conditions can a player touch their Rook first and have it be clear that they intended to castle unless they move it to the castling square (for example f1) and release it and immediately touch the King? In other words, how can you clarify this intent without observable action?

Silence on the transferring or releasing of the Rook does not imply that it is illegal to transfer/release it. In fact, the requirement of intent arguably REQUIRES it.

Question:

If you play (in a castle-able position), K-g1, is the move 00 completed? No, it isn’t. So why do you continue to make arguments about the release of the King or the Rook? I don’t get it. The move is not complete at that point.

Even more unfortunately, that’s not what you are doing. Seeing above.

As shown above, that’s not true.

You’ve both implied stuff that isn’t there, and as shown above your analysis is not correct. The two key parts that you and bwtaylor continue to miss:

  1. Intent
  2. Half of castling is not a move.
216

Would only use the words that are actually in the text of the rule. If it says “touch”, then a literal reading would only be concerned with touching, not with any other verb.

The other way that one could do it would be to touch the rook (R1), release the rook (R*), and then castle using the procedure in 8A2.

Because the point at which the move is completed is not under discussion. The discussion concerns the point at which the move is determined with no possibility of change. When the king is released on g1, the rulebook is very clear that the move O-O is determined.

If you read the words of the rulebook literally, it is equally clear about releasing the rook on f1 if the king has not been touched. There is only one rule that discusses releasing the rook. That rule is rule 9A.

217

Exactly. However, touch isn’t the only word there, and the lack of certain words doesn’t imply anything one way or the other.

Your reasoning is correct, but your conclusion is wrong.

Not so, from that one could conclude that a Rook move was intended to at least two possible squares.

Of course, repositioning it to the castling square and releasing it - which provides more information than your situation above - also falls under your description above.

I’ll put this off to your lack of knowledge of the rules. Completion IS the point at which the move is determined with no possibility of change.

Huh? Quote that.

Sorry, that last statement isn’t relevant.

218

Are you sure about that?

219

15+ pages. Um, I would have ruled quite quickly that the castling move would stand. According to USCF rules and TD discretion concerning the particular facts available. The players would continue play a little wiser about the rules. Oops, have to go to the other side of the room to rule on a “touch move” incident. Then there are clocks that are set improperly that have to be fixed. Too many things to do to get so involved with such a minor incident. :unamused:

220

Ding, ding, ding! We have a winner! I can’t even remember anyone at anytime even filing a complaint with USCF regarding a TD ruling about this issue. The cross the “t” and dot the “i” folks should consider making a motion at the Delegates convention to cure this almost non-existant concern.