There are a number of points of evidence on the neck of 67P which is telling us the evolution of the shape of the nucleus, and perhaps of comets in general. The swept back nature of Hapi, along with the roughly circular cross section of the centre of the neck is a pointer that the neck has evolved from stretching. The case is strong that the bilobed shape of 67P is from a stretch event that has kept the evidence from initial fracture in the matching shear lines of the lobes. That enables us as modellers to piece backwards that the original shape was roughly ellipsoid. However, the interpolation of the intermediate neck stage, and the timescale between intermediate stages is not able to be determined based just on the mechanisms hypothesised to be acting on the nucleus.
Next we can look at the formations on the neck and see if they can be connected to ongoing phenomena, which would allow us to extrapolate back a little bit. Cracks are ubiquitous on Anuket. Anuket is otherwise *not* covered in dust or rocks, but has the texture of quick dry cement. A lot of formations perpendicular to the neck give the impression of an evolving process related to the cracks. Cracks are notionally perpendicular to the rotation plane, cross the equator and appear open wider at around the equator. Outbursts are also seen to happen at Anuket which are likely related to the cracks.
A continuing stretch hypothesis is that the cracks open up at Anuket due to the nucleus rocking on the neck due to asymmetrical outgassing torque. The cracks then get filled from below by a slurry, which quickly hardens on exposure to vacuum. As the head rocks back, cracks open up at the neck on Bastet at the opposite side of the equator, which also gets filled in. The cycle has a net effect of lengthening the neck, absorbing the torque (and precession) which would otherwise accelerate the rotation even more, with the conservation of Angular Momentum and increased gravitational potential absorbing the torque energy. Each rotation may have the width of a crack lengthening of the neck, and the resultant neck feature would appear like a ridge or paired ridges either side of the crack, still notionally perpendicular to the neck. That would explain the myriad perpendicular ridges and valleys on Anuket and perhaps Bastet. This would mean that this process would have been happening steadily for quite some time, and that the original stretch event perhaps only travelled half way from an Ellipsoid to the current duck shape, and the rest has been happening over many perihelions since.
Accurate measurement of neck length evolution (or lack of evolution) will illuminate this situation.