Why can't a piece (that isn't a pawn) capture en passant? When a pawn makes a two-square move, the en passant rule says that on the next move, any pawn that could have captured it on the single-advance square may do so.
So why doesn't the rule allow a non-pawn piece that could have captured it on the single-advance square to do the same? For instance, in the position
[fen "7k/8/8/8/2n5/8/P7/7K w - - 0 1"]
1.a4
why isn't the knight allowed to respond 1...Nxa3?
I believe the en passant rule is designed to cope with the fact that pawns do not move backwards, hence if the en passant rule did not exists you could have a lot of blocked positions, making play slower and probably more drawish.
Differently from pawns, pieces can move backwards, so the black Knight in your example still has the possibility to move backward and stop/capture the white a pawn. Therefore no need of allowing the en passant rule for pieces.
Said that, there is in a special case a sort of en passant like rule for pieces: castling is not allowed if an opponent's piece controls one of the squares that the King would go through while castling. You can see that as a sort of threat of hitting the King "en passant" during the castling move.
At a large extent chess rules are what they are because they resisted the test of time and made an interesting game. Same goes for pawn moves. The 2-step move from the pawn initial position did not exist either from the beginning, it was introduced later to make play faster and more dynamic; the en passant rule was likely introduced shortly afterwards to avoid large pawns blockades.
En passant is, for a lack of a better word, an ugly rule. It does not flow logically from the "fundamental" rules of chess - it is an exception that serves a specific purpose: to prevent players from forcing walled positions with their pawns, effectively killing any activity in the game (at least, that is my understanding). By allowing a player to capture en passant, he has at least the option of opening a file in a case where the opponent may be interested in keeping it closed.
I do not see how this justification would translate to non-pawns. Allowing en passant capture by non-pawns would not add anything to the dynamics of the game. It would be an unnecessary proliferation of an already ugly rule.
Long ago, pawns were only allowed to move 1 square at a time (side note: fianchettoing was therefore very common then, as playing e3 or d3 opens the diagonal for one bishop, only to block the other!) ; the rules were modified to allow pawns to move two squares ahead on its first move in order to speed up the game.
Now, this had some major impact strategically because then players could move their pawns 'past' their opponents' and avoid capture, which would have resulted in the need for a major re-think of chess strategy, therefore the en passant rule was implemented to maintain a good level of similarity to the 'old' game.
One possible reason why this rule wasn't expanded and implemented for pieces is due to the static nature of pawns (which can't move backwards), which is not the case for pieces; having a pawn move past a square defended by an enemy pawn to evade capture was deemed unacceptable, whereas it is alright to evade capture by a piece. The strategic implications of the former is extremely severe and would lead to a completely different game, but in the case of the latter it's lighter and the nature of the game remains somewhat unchanged (perhaps because the piece can still relocate to 'chase' the pawn?).
But in any case, "rules are rules", however arbitrary they may seem (don't hope to change them, especially on the basis of "it doesn't make sense!"); to digress: why is the queen the most powerful piece? (Further digression: https://www.youtube.com/watch?v=N_AhNHrfu4Q). Having said so, questioning the motivation of these strange rules is certainly not unreasonable, and hopefully the above gives an idea over the rationale of not allowing pieces to capture en passant.
On a trivial note: it was actually suggested at a FIDE conference relatively recently (clearly by someone who didn't know chess very well!) that pieces be allowed the capture en passant. Needless to say, the suggestion was not taken seriously!
No. There is no analogue to en passant captures for pieces that are not pawns.
The rule, as it exists, offers an advantage to the player moving a pawn that is equivalent to two moves in a single turn whenever a piece is denied a capture available under the old rules. Until one of your games is affected, the significance of this rule is seldom noticed.
I don't have an logical answer because I believe one does not exist. None of the responses above — almost all of which are non-responsive — answer the WHY of the question. The fact that pawns cannot move backward does not begin to address why a piece cannnot capture a pawn en passant.
My best guess as to the Why of it is that, the last time the rules were changed, nobody thought to bring it up. Or, if they did, it was decided better to stick with an established rule than to change the nature of the game too much, and this would have been considered too much of a change.
I think the same with regard to the question of why a piece pinned to its king can still give check. Neither one of these two rules is logical, though. They are just "there," and we live with them.
There is, of course, nothing preventing someone from starting up a club around a new chess variant that incorporates these two rule changes. Might be interesting to see just how much it does change the game. Most games wouldn't be changed at all, I think. But for those it does change, ... different outcomes, perhaps.
Thanks for the question. If an officer could indeed be captured en passant, then one would need to undo any capture it had just made, which is potentially awkward. This doesn't apply to pawns, because the famous double hop is defined to be a non-capturing move.
