A long time ago, I read about the experience of a prisoner in an Asian prison, I think sometime in the 1970s. While in prison, he learned how to change his pants…while his ankles were chained.
Uh…what?
I have been pondering this factoid for a long time and I cannot comprehend how this is physically possible. It seems that no matter how you would put on or take off a pair of pants, the ankle manacles would get in the way. How could this possibly be done? (assuming that “cheating” methods like cutting open the pants with scissors, etc. aren’t used.) How could even a Houdini do it?
Possibly similar to the way Rowan Atkinson changed into swim trunks without exposing himself.
It was a common practice in US prisons at one time. It’s difficult to describe, something like the video linked above. You roll down one pants leg and run over the shackles and then up the other leg over the pants leg on that side. Then you pull both pants legs down passing between the one shackle on that side and your leg. The pants have to be thin enough and the shackles loose enough to do this. I saw this described along with photos from a US prison long ago, possibly in Life magazine.
It’s a simple, if non-intuitive, topological problem. We had a thread about it eight years ago. Practically speaking, it just requires enough space between the person’s leg and the ankle band to feed one leg’s worth of the trousers through.
It might help to start by imagining the shackles as a pair of hula hoops that you’re standing in: It’s easy to see how it might be managed then. Then, imagine shrinking the shackles smaller and smaller. Obviously it gets harder the smaller the shackles get, but it remains possible as long as the pants are thin and flexible enough.
Unless the pants leg can fit over a hula hoop I’m not seeing how this could work. You first have to get both legs of the pants over one of your legs by passing it over the shackles.
If seems like you would just push each leg down into the top of the shackle and then put your feet into the legs. Not too hard.
No, you have to get the pants inside the shackles, not outside.
Did chained prisoners have access to a change of clothing?
Try imagining this (note that throughout this exercise there’s a chain connecting the two shackle bands together; also note that it doesn’t affect what we’re doing):
You’re standing, wearing a pair of pants.
Shackles are put on you with the bands around your ankles.
For purpose of illustration, we use magic to temporarily expand the bands to the size of hula hoops, and levitate them to just above ankle height, BUT YOU CANNOT STEP YOUR FEET OUT OF THEM – YOU ARE STILL “SHACKLED.”
On one side, drop the leg of the pants, along with the waistband on that side, down and pull it off your foot and then pull that side of the pants up through the hula hoop sized band. NOTE THAT YOUR FOOT NEVER HAS TO STEP “OUT” OF THE SHACKLE BAND.
Repeat on the other side.
You have now removed your pants without getting free from the shackles.
In real life the bands would not be big enough to step your feet out of them, but there could be enough room to pull the fabric of the pants past them (between the bands and your legs) to get the pants off.
Ok, I see what you are saying, that’s possible also, not the method I learned.
It took me a minute to grasp what you were saying, but now I see that’s topologically the same sort of thing (when you finally remove the pants). It’s based on the pants not being actually attached to the person, more along the lines of interlaced among the person’s legs and the shackles.
Ask any woman how she can remove her bra whilst still wearing a sweater or t-shirt. Same problem. Bonus points for guys that know this trick. 
And there are a lot of old physical puzzles (blacksmith’s puzzles, the ox-yoke puzzle, etc.) that are based on similar techniques.
Aside: The first time my uncle ever saw an ox-yoke puzzle, the person showing it to him told him that the goal was to put both beads on one loop. So, he did: He reduced the rope to a single loop, which of course still had both beads on it. He said that if that wasn’t the intended solution, then either the problem statement or the construction of the puzzle should have been such as to rule it out.
There’s also a party trick or parlor game that’s similar.
At a party or other social gathering, do this: Select two people (most interesting if of opposite data type). Get two pieces of kite string, yarn, or similar, about two feet long each.
Tie one end of one string around one person’s wrist, loosely, but tight enough that the loop around the wrist can’t easily be slipped off over the hand. Tie the other end to the same person’s other wrist likewise.
Tie both ends of the other string to the other person’s two wrists likewise, but with the string passing around the other person’s string so the two people are linked together.
See: The Handcuff Puzzle for more detailed (and probably better-written) instructions, including illustration.
The problem: Without any “obvious” cheating (like breaking either string or untying them), have the two people try to get themselves separated.