There is a related "love triangle" Special Relativity problem where there are three parties: stay-at-home (S), early-outbound-passer (E), and late-inbound-passer (L). None of the parties ever experience any acceleration: they remain eternally in uniform motion, with E & L travelling relativistically.
At our origin, S and E synchronize observe their identical atomic wristwatches coincidentally agree that it is "0". Light-years away, E and L come very close to one another and exchange timestamps showing that coincidentally their identical atomic wristwatches agree. Finally, L and S come very close to one another and compare timestamps from their identical atomic wristwatches. All the wristwatch times are identical to those at the three points in the diagram of the "instant turnaround" version of the twin paradox, we've just turned the travelling twin into two unrelated travellers on different trajectories.
The argument is that this "love triangle" is resolved because E & L are different travellers in uniform motion, so all parties must combine the times acquired in two different reference frames (E's and L's) to compare with the times acquired in S's reference frame. The further argument is that this duplicates the "instant turnaround" version of the twin paradox if we can have the travelling twin change direction without acceleration.
Firstly, we can still solve this with a pseudo-gravitational field popping up at the moment E & L exchange timestamps. It's no more of a coincidence than the identical timestamp when S & E are close.
Secondly, it's not clear that the paradox remains interesting in this case, because there is no expectation that S & L should be the same age when they are close to one another again. They aren't twins. Unless we add in accelerations, there is no way by which S, E, and L could all have been born at close to the same location in spacetime.
Thirdly, it's unclear that there can be an instant turnaround without acceleration. A couple flavours have been explored here and there.
One involves a slingshot around a star to change directions from away to towards the stay-at-home twin. In this picture the travelling twin is always in free-fall. But here we are substituting real gravitation (that of the star) from pseudo-gravitation. We've moved from everywhere-flat Minkowski space -- the spacetime of Special Relativity -- to something closer to Schwarzschild spacetime, which is only asymptotically flat. Moreover, we are using the near region of Schwarzschild to accomplish the slingshot.
Another substitutes the open flat Minkowski space with one in which there is a compact spatial dimension that curls back on it self. A universe with the geometry of a cylinder with infinite height and small circumference, or a torus, or a sphere would do. The cylindrical case has been explored recently : https://doi.org/10.1119/10.0000002 with comparisons to Minkowski space (the spacetime of Special Relativity), §IV (Conclusion) being pithy. Again, I see this as trying to substitute pseudo-geometry with real geometry, an adapted clock-comparison recipe, and a highly privileged frame for the traveller, in order to avoid a non-gravitational acceleration opening the door to a pseudogravitational field arising in the ultrasimplfied and thus strictly Special Relativity problem.
Finally focusing on the latter part of my comment that I'm self-replying to (mostly for my own benefit), we have only done away with one acceleration by the returning twin. We still have the effects from the behaviour of matter in the expanding universe with which to clock S, E and L, removing the remaining paradox if we somehow contrive to have S, E & L expecting to age similarly. If we are abandoning Special Relativity in order to avoid acceleration by the returning without invoking outright magic, why only do it along one spacelike dimension, or by importing a very finely tuned third traveller?
At our origin, S and E synchronize observe their identical atomic wristwatches coincidentally agree that it is "0". Light-years away, E and L come very close to one another and exchange timestamps showing that coincidentally their identical atomic wristwatches agree. Finally, L and S come very close to one another and compare timestamps from their identical atomic wristwatches. All the wristwatch times are identical to those at the three points in the diagram of the "instant turnaround" version of the twin paradox, we've just turned the travelling twin into two unrelated travellers on different trajectories.
The argument is that this "love triangle" is resolved because E & L are different travellers in uniform motion, so all parties must combine the times acquired in two different reference frames (E's and L's) to compare with the times acquired in S's reference frame. The further argument is that this duplicates the "instant turnaround" version of the twin paradox if we can have the travelling twin change direction without acceleration.
Firstly, we can still solve this with a pseudo-gravitational field popping up at the moment E & L exchange timestamps. It's no more of a coincidence than the identical timestamp when S & E are close.
Secondly, it's not clear that the paradox remains interesting in this case, because there is no expectation that S & L should be the same age when they are close to one another again. They aren't twins. Unless we add in accelerations, there is no way by which S, E, and L could all have been born at close to the same location in spacetime.
Thirdly, it's unclear that there can be an instant turnaround without acceleration. A couple flavours have been explored here and there.
One involves a slingshot around a star to change directions from away to towards the stay-at-home twin. In this picture the travelling twin is always in free-fall. But here we are substituting real gravitation (that of the star) from pseudo-gravitation. We've moved from everywhere-flat Minkowski space -- the spacetime of Special Relativity -- to something closer to Schwarzschild spacetime, which is only asymptotically flat. Moreover, we are using the near region of Schwarzschild to accomplish the slingshot.
Another substitutes the open flat Minkowski space with one in which there is a compact spatial dimension that curls back on it self. A universe with the geometry of a cylinder with infinite height and small circumference, or a torus, or a sphere would do. The cylindrical case has been explored recently : https://doi.org/10.1119/10.0000002 with comparisons to Minkowski space (the spacetime of Special Relativity), §IV (Conclusion) being pithy. Again, I see this as trying to substitute pseudo-geometry with real geometry, an adapted clock-comparison recipe, and a highly privileged frame for the traveller, in order to avoid a non-gravitational acceleration opening the door to a pseudogravitational field arising in the ultrasimplfied and thus strictly Special Relativity problem.
The pseudogravitational field approach comes from Einstein in 1918: https://en.wikisource.org/wiki/Translation:Dialog_about_Obje... which was fun to read.
Finally focusing on the latter part of my comment that I'm self-replying to (mostly for my own benefit), we have only done away with one acceleration by the returning twin. We still have the effects from the behaviour of matter in the expanding universe with which to clock S, E and L, removing the remaining paradox if we somehow contrive to have S, E & L expecting to age similarly. If we are abandoning Special Relativity in order to avoid acceleration by the returning without invoking outright magic, why only do it along one spacelike dimension, or by importing a very finely tuned third traveller?