While it is entirely true that ductility will be greater in a heated column that certainly will not aid in the column's 'job' in supporting a load. All it will do is reduce the column's ability to fracture under load. There are indeed pictures(no I don't know where to find them now) that show columns that have been bent almost 180 degrees. That would be very difficult to accomplish without heating that portion of the column. What it also illustrates is that the load that caused this bending was not moving laterally a great deal which suggests that the collapse was not sliding away from the vertical axis at the time the column was bent.
Other columns show fracturing instead which would be consistent with a suddenly increased load on a column that is still relatively cool which would occur if enough heated columns started failing in quick succession.
The columns that fail need not all be on the same level and it is obvious to anyone who has read the NIST reports that the fires on each floor were worse at different loactions. The main cause of this being that the planes did not hit the buildings wings level. This put aviation fuel into one side of the building from the right wing at upper fire floors while putting aviation fuel into the other side of the building at lower fire floors. Fires progressed around the floors heating columns. Some of those would have foreshortened during the heating under load but cooled as the fire burned down in that area and progressed to other areas where it would heat other columns. These newly heated columns now have already been handed an extra load due to the foreshortening of the now cooled columns. Those now cooled columns will have regained their strength as they cooled BUT due to any deformation that occured during heating they no longer have the load applied through their vertical axis.
Now you increasingly have columns which are more prone to buckling due to their loads not being in line with their vertical axis and others that while still dimensionally intact are losing their strength. This is happening on various columns around the building even though it may be occuring at several different levels.
In the South tower this occurs along with the assymetric damage done to the columns by the plane's impact which immediatly caused a lean of the upper section. In the North tower the damage was much more symmetric as the aircraft hit more centrally and on the side which was parallel to the long horizontal axis of the core(South tower was hit on the short axis).IIRC
In the South tower the above heating damage is done which introduces a more pronounced lean but the angular momentum is actually small since the angular velocity is so small. A point is reached at which loads cannot be redistributed to stop further leaning and the building starts to rotate but simply cannot go very far before the columns at the hinge point fracture and the whole mass drops vertically about its CG (isn't it more proper to say center of mass?). It may still be rotating but is now doing so about its CG rather than at the hinge point and that CG is falling straight down. Since the CG did not move a signioficant percentage of the horizontal distance between it and the outer edge of the building , MOST of the mass is falling upon the lower section of the building.
In the north tower more columns had to undergo the heat stresses above since there was much less stress due to the upper section leaning(not to mention that the upper section is much smaller than in the south tower). On the other hand more fuel was dumped into this building than in the south tower since it hit more centrally and thus the fires were more widespread and since it was more central the fires were also more symmetrically dispersed about the vertical axis(though, again , on several levels).
Still, a point is reached at which redistribution of loads cannot occur fast enough to halt settling and columns fail in quick succession, cold columns perhaps fracturing , while hot columns bend while still attached to the floor above.