Water is virtually - for engineering purposes - incompressible. The physics of ditching on a flat calm sea are initially not much different than a wheels-up landing on a concrete runway. The vertical and longitudinal loading on the pylons is rearward and upward, rapidly increasing as the aircraft slows and the wing is relieved of aerodynamic loading.

Engine pylons are designed to separate under certain load regimes to protect the wing spar. At some point during slideout, the pylon fails under the bending moment and the engine separates. But it's more complex here. During flight, engine separation almost certainly means the dead mass drops away from the airframe. During ditching, what happens to this mass? It is rammed into the wing spar, as the a/c rides over the top of the engine, or it rotates back and over the wing, possibly gyroscopically impacting into the fuselage further aft. 15,000lbs of JT9 smashing into the cabin.

But that's not even the most likely scenario. The key question is whether the dynamic load regime would cause pylon failure along the design failure paths. Pylons are relatively strong in longitudinal and vertical axis loadings, much less so for azimuthal loadings. Flight dynamic loads are designed to damp such stresses, so the static structure is relatively weak. If you experience moderate or severe turb from the top deck of the 744, look out and watch the multi-axis motion of the pylons and you'll see how this structural system damping works.

The problem comes when roll or yaw are introduced during this landing. If the sea is anything less than mirror smooth, parts of the a/c not designed to take such stress will contact the water. Unlike runway slideout, when contact of the wingtips and other a/c parts can have a stabilizing influence, the uneven contact with the sea will induce further roll or yaw that will increase local stress and cause catastrophic structural failure. Once this happens, still at moderately high velocity, hull motion will become unpredictable and most likely one local structural failure will lead to others. Disintegration of the fuselage is almost inevitable. As in the Comoro Islands 767 crash. Run it in slow-mo and you can see chains of structural failure.

Ditching a large airliner with engines mounted below the wing is an extremely hazardous operation that will likely result in many casualties under the best of circumstances. The only scenario I could see where pylon failure may possibly map the design mode and allow non-catastrophic separation would be a low-speed ditching...like, for instance, a westerly low-energy runway overrun at LAX, followed by slideout across the beach and into the Pacific.

So, keep us out of the water!