So regime type in China doesn't really matter for its defense. National culture and geography drives all. No matter how uncentralized you'll see the Japanese never take Sichuan, Yunnan, Shaanxi and Gansu? Because that's about all that was left of "productive China" (Inner Mongolia, Xinjiang, Tibet don't really count) by 1945 even in OTL.
Hang on, I don’t believe I said that. What I said was that out of the Chinese population- which is well over 400 million- there will be those who oppose that their entire country is being invaded by the Japanese militarists, something that was true OTL and barring some major change will be so ATL. In comparison, the population of Japan proper in 1940 numbered around 75 million, and the important part is simply to notice the massive disparity in numbers right there. Even if half of the entirety of China surrenders to Japan, that’s still 200 million people who are opposed to the Japanese militarists. Even if it’s only every tenth person who resists, that’s still 40 million enemy combatants. That’s still more than six and a half times more people than the IJA at its peak, and assumes it is disciplined enough to not have officers engage in the kind of behavior that swells the numbers of the resistance.
Connected to that, first of all, of course regime types matter, and they do so for attack as well. If the Japanese government and military are crazy enough to invade China due to the radical militarists, then it means the campaign will be plagued by similar issues to those that plagued it OTL, which will most likely unite a substantial number of the Chinese people against the militarist invaders. And of course, regime types do matter for defense, and I pointed out that in the unlikely scenario that China is not unified, the warlords continue and the Japanese militarists try a more limited approach, they might get away with some more warlords under their control, or for that matter that a Chinese counter-offensive in the absence of the KMT or the Communists might see someone like Yan leading the war effort with the implications that entail; what I disagreed with was the idea that simply because there is no KMT or Communists and lots of more surviving warlords, a nation that has conceived of itself as destined to be a united entity for the past 2000 years would simply let itself be slowly absorbed by a nation they consider foreign and which has issues all of its own before it tries to absorb all of China into its empire.
And because yes, national culture and geography drive much of what societies based around those concepts do, because they form the foundations for what people consider to be right and wrong for themselves and their nation. National culture and geography do not drive all, but the former forms the basis for people’s conceptions of what society should be like, and the latter influences the ability of whatever ruling polity that espouses such ideas to implement it, which together interact with the actual regime types to form the kind of society that ultimately results.
I further would like to dispute the idea that ‘Sichuan, Yunnan, Shaanxi and Gansu’ were all that was left of ‘productive China’ by 1945. If we take the idea of Yunnan and Gansu as ‘productive China’, there was also Fujian, Guangxi, Guangdong, parts of Zhejiang, and most of Jiangxi, Hunan and Henan; and importantly, this was with the IJA suffering appalling casualties and throwing everything it could spare in China at the provinces, but still with extensive local resistance that often enough made the pretense of effective control seem like wishful thinking, and no sign yet of the entirety of China simply laying down all of its weapons and surrendering. While this will of course be different if there are no KMT or Communists to pledge allegiance to, if your village is suddenly invaded and- again, considering the people who led the invasion of China- subjected to servitude and worse, does it really matter at the moment of resistance if you pledge allegiance to the KMT, or the Communists, or the local warlord who, perhaps, at least isn’t exploiting you for the sake of a foreign nation? Even then, if we assume the Japanese are able to defeat all the main warlords, how likely is it that a Japan crazy enough to embark upon a conquest of all of China will be able to remain measured and civilized in its suppression of the inevitable revolts that will transpire against its rule, and may I add, probably within short time? We are again talking about 400 million people spread across vast distances, who have gone through years of division and warlordism, and are now supposedly going to allow a foreign nation to absorb their entire nation into their empire to be exploited as it sees fit, compounded by the fact that the military that is supposed to enforce this has some serious issues with discipline, fanaticism and brutality that makes most warlords seem nice.
Furthermore, why doesn’t Inner Mongolia count? While obviously in this time period Tibet is not a part of ‘productive China’, and while the Ma were still on the side of the KMT, provided weapons and soldiers, and generally forced the Japanese to attack them as well, they were also not a hotbed of industrial production; Inner Mongolia however, has industries and an existing coal industry in the region by that time, and has been majority-Han since the mid-19th century, and as OTL, most of whom do not want China to fall to the Japanese militarists and some of whom will thus take up arms. If we define Gansu as ‘productive China’, while I agree that Tibet and Xinjiang do not qualify, I do believe we should also define Inner Mongolia as ‘productive China’ in such a case.
I do not dispute that if China was decentralized to the degree that the Japanese militarists could simply come in and play all of them off as they slowly conquer them piece-meal, it might be possible, but the problem is that as far as the facts we have tell us, that’s very unlikely. I said however, unlikely- I never said impossible in the manner of ‘no matter’, I said unlikely, in varying degrees, and this one just so happens to be very much so.