Ok, Anoop has thrown a lot at me here, so I'll do my best to muddle through with some answers
On the scenario "Dresden II": what I wanted to get folks thinking about is whether it might ever be in India's interest to use nuclear weapons on foreign soil, in order to accomplish a strictly tactical/operational goal... in this case, forestalling a Pakistani counterattack and preventing Pakistani encirclement of our units inside TSP by choking transportation arteries with refugees.
As opposed to the way we usually think about our nuclear weapons... they are there to scare other countries away from the option of nuking us, a pure deterrent, period. The first two scenarios tend to that perception. In the first one, I'm talking about a pre-emptive nuclear strike to stop doomed Pakistani generals from trying to take us down with them; in the second, about an existential threat to the Indian union. The third was an attempt to explore whether we might be in a situation where nukes were the best *offensive* option we had.
My implicit conclusion is that such a thing would not be diplomatically possible, even if we scrapped the NFU, unless Pakistan at least used a nuke on our forces on their own soil (though I'm aware that, under the DND, this does not constitute "first use" on our part).
The more important question is whether, in the event of a Dresden II becoming necessary, it would be advantageous to have such an option? That is, given a Pakistani nuclear attack on our advancing columns, plus a concerted effort by ISI sleeper cells all over the country (both non-conventional actions that might severely curtail our ability to reinforce our strike corps), is there anything else we could do that would serve us better? Or can we say that having a NFU policy is a severe liability in such an instance?
I realize there are valid holes which serious military-strat analysts could poke in my Dresden II scenario, as you have. This isn't Dresden, where the USAF/RAF firebombing was useful only in the context of easing the way for a Soviet advance from the east. However, I was hoping that more mil-tech savvy people than myself would respond with more likely scenario where nukes might present themselves as the best offensive option, tactically speaking.
On the subject of ISI sleeper cells not being effective in the context of a wartime psyche, I'm not sure I agree with you. Nobody here really KNOWS the extent of jihadi penetration of Indian Muslim communities. Looking at a sprawling, dyed-in-the-wool Muslim mohalla like Mumbra, for example... do we know if the ISI has indoctrinated two people willing to become fidayeens... or twenty, or two hundred, or two thousand? I don't want to get into the lakhs and lakhs of Indian Muslims in Mumbra who would only become more intensely patriotic in the event of a war. Two people with the right equipment can do a lot of damage... and two hundred motivated, organized and armed individuals could bring about a complete disruption of law and order in Mumbai for a while. To the extent that BMCC and the Maharashtra Government would be demanding that the army be sent in. My opinion... we're a lot more vulnerable to this kind of warfare than anyone in the GOI would care to admit.
On the broader question of a desirable endgame that could be brought about by using nukes on Pakistan (first strike or response)... I can only think of the following goals:
1) Completely and permanently end any nuclear/missile threat to India from Pakistan.
2) Completely decimate the Pakistani armed forces, their ability to wage war, defend their territory or govern anything. Leaving a power vacuum that collapses the Pakistani state, which various armed factions will then compete violently to fill, bringing about disintegration into smaller political entities.
The main problem with a collapse of the Pakistani state as envisioned today, if WE do not control what happens to Pakistan's nukes FIRST, is the possibility of a jihadi takeover of Pakistan's nuclear arsenal. Our strike must be calibrated to ensure that they will inherit no such arsenal (see (1)). In that context, if such a thing can be ensured... I would argue that an ungovernable and openly jihadi Pakistan without nukes is preferable, even if it stays together somehow, to a TSPA-controlled and nuclear armed Pakistan.
3) Reduce the economic, social, political and living standards of the Pakistani population to a state of such complete misery that survival takes precedence over jihad as a motivating factor for most people.
It is often argued by the western liberal intelligensia (the same ones who entertain romantic notions of egalitarian Islam as the original Marxist revolution) that poverty is a root cause of Jihad in Islamic states. Nothing could be further from the truth. Saudi, Pakistani and other sponsors of Jihad are not poor... they are the elite in whose interest it is to ensure that the masses are distracted from economic disparity with grandiose religious fanaticism. In a state where the rich and middle classes are no longer in control, and it is every man for himself, the social architecture that encourages and sustains jihadi tendencies among the masses will collapse. The Mullahs who incite Jihad will lose their sponsors, and will remain interested only in grabbing what they can while it lasts.
Given a sufficiently large power vacuum, and a sufficient paucity of resources for anyone to sustain themselves, I predict that Jihadi groups, absent a controlling organization (the ISI) with the muscle to back up their authority (TSPA) will turn on each other in a violent struggle for control of what is left of the Pakistani state (see Iraq). In Iraq, of course, this has come about as a failure of American nation-building exercises. In Pakistan, it will be precisely what we want. We will destroy, withdraw, mine the border thoroughly to keep out refugees, and watch the dissolution of Pakistan. Our own nation-building activities if any should be concentrated in areas of high potential return, such as Baluchistan.
The Three-Front War Scenario: I had initially concluded that perhaps the best thing to do when confronted with that situation would be to nuke Shanghai. Given your input, I've revised that position. Faced with aggression on three fronts that we are unable to confront with conventional means, we should nuke neither PRC nor Bangladesh first, but Pakistan.
This might work to our benefit in a number of ways:
1) A calibrated strike as envisioned in my discussion of an "endgame" above, would achieve those goals with respect to Pakistan, ending their nuclear and conventional threat once and for all, and precipitating their collapse.
2)This would be achieved at a minimal cost in troop deployments, ending the threat on the Western front so that we could commit more resources to the East.
3) In the near term it might very well scare BD into withdrawing and surrendering altogether, ending the threat across a second front. In the longer term it might result in BD relinquishing the NPT and openly receiving nuclear weapons proliferated from Beijing... but that can be negotiated in the diplomatic aftermath, balanced against the possibility of our proliferating nuclear weapons to Taiwan or Vietnam in retaliation.
4) It would fall short of a direct nuclear strike against Chinese soil (and the disastrous exchange that is likely to follow), but it would demonstrate our willingness to use nuclear weapons when faced with an existential threat to our territory, and in that sense it might be more effective in pressuring the Chinese to withdraw. It would end any perception of Indian weakness in that regard.
However, for it to fully achieve such a purpose it would have to be backed up with a substantial buildup in our strategic arsenal, including number and range of warheads and delivery systems, assured second-strike capability, sea-based platforms etc. The way I see it, the rise of China absolutely demands the establishment of a credible Mutually Assured Destruction paradigm vis-a-vis Beijing, to forestall the eventuality of a three-front-war or any other sort of expansionist adventurism.
Finally, as Calvin says I wasn't about to explore WHY China might want to push India into a three-front war in the first place. Any number of reasons could develop over the next decade, and with Beijing's political structure being as opaque as it is, we might not see it coming until it is far too late for comfort.
Questions of whether PRC generally tends to occupy territory vs. punish-and-withdraw, depend entirely on what Beijing perceives as their political objective in any given situation. They stayed in Tibet because controlling that territory was the whole object of the exercise. Not so with India (except Aksai Chin) and Vietnam, where all they wanted to achieve was punitive invasion and withdrawal.
Depending on how competition between the US and PRC for energy resources shapes up over the next decade, Beijing's hankering for a direct corridor to the Indian ocean might increase tremendously as a dire need to insure its energy security emerges. At that point, what if Gwadar is no longer an option for various reasons (full-blown Baluch insurgency, American actions?) What's the next-best hunk of territory that would give them such access through a friendly nation? Given the political situation in India at the time, the extent to which Sino-Indian trade has developed (or not), the state of our border issues, and many other things... PRC might very well decide to inflict a three-front war on India. I'm really more concerned with what we might be able to do about it.