I do not try to pontificate, but look at all the data that is available, and seek to interpret the data. This is an interesting thread, but many people have not looked at the situation on the GNR during the relevant period. Also many of us may not have looked at the whole picture.
I was surprised to find that the yanks had been using so called 10 wheelers since before the GNR was actually up and running, ie 1847, so you do wonder
why it took until 1894 for Jones to look at it for the Highland railway. My feeling is based on GNR practice in that the UK did not have really good quality steel available for coupling rods, certainly the last locos produced under Sturrocks regime, the 2-4-0's were troubled with broken rods, and Stirling rebuilt them as 2-2-2's. American trains of the period were running much more slowly than UK passenger ones, hence their ability to run 4-6-0 for so long. What surprised me more was why it took until 1888 for the yanks to produce the first 4-4-2 passenger loco, and the name Atlantic only became used in 1894 after a more successful loco had been produced to run to Atlantic City.
Ivatt did not join the GNR for more than 3 months after his appointment, and we do not know why that happened, no contemporary commentator has ever made any reference to it, but we may assume it was to do with his Daughters and their lives in Dublin, during that time, the man who had been Stirling's loco accountant Matthewman had been in charge, during this time he ordered a number of basically Stirling designs, but an unfinished order
was for 10 2-4-0. These were cancelled by Ivatt almost immediately he arrived, and a different order for 4-4-0's was put in place. At the time Ivatt arrived, the Stirling singles were becoming overwhelmed by the increased traffic and the dramatic growth in carriage size and weight to such an extent that the needed to be double headed, something which was not allowed until almost the end by Stirling himself. Thus was Ivatt's first task to produce a locomotive type for use on Primary Services, but at the time he joined the GNR there were two problems, one Pebbles has referred to, length of turntables, but the second more importantly, was the state of the track, thus any loco bigger than the Stirling single needed to offer some advantages.
Many observers thought that actually the last Stirling 2-2-2's were better than the 8ft Singles, so it is not unreasonable to assume that Ivatt wanted to improve on them.
Ivatt came to the GNR with the idea that a 4-4-0 would be a useful loco, and with which he had some decent experience, also with a better bogie type than that on the Stirling single, and yes Stirling only used bogies on two types of locos, tanks and the 8ft Singles. The goods loco types were being served by the Matthewman ordered 0-6-0ST's and the J4's with the large 3850 gallon tenders that were originally designed for the last Singles, so a first line passenger engine was necessary, and until the Atlantic's were produced almost 4 years later, that is what the D4 did. Certainly another 10 2-4-0's were produced based on the previous Stirling design, but with Ivatt mods, ie cab, spring position, boiler and firebox data, but that route was a dead end.
However, to return to the original question, why did the GNR not use 4-6-0's my research suggests that a major reason may well have been that a major revenue stream for the GNR passenger division was the Leeds route, and there were massive size restrictions on locos going their using the GNR route, indeed I understand that until some time in the late 30's the Pacifics were route restricted as was certainly W1 10000.
Remember that when Churchward built the Saints, he had both 4-4-2 and 4-6-0's built to check them out, and he always said that The Great Bear was a complete waste of time, money and space.
Were there some common design features between the 2-4-0's and 4-4-0's definitely but still more Ivatt than Stirling.