Many thanks to Ellis Krauss for his thoughtful observations. However, I would point out that he is essentially shifting the discussion from what are Japan's security options to what Japan has done over the past twenty years. I am willing to go there, but it's important to keep this distinction in mind.

I think Ellis' observations about the history of Japan's approach to the alliance over the past twenty years are mostly on target. However, I am not sure to what extent we can say the alliance has necessarily been strengthened over the past twenty years. We have seen expanded cooperation in some areas related to technological innovation, such as missile defense, expanding and then contracting cooperation in other areas, such as overseas SDF deployments in support of US military operations, and other areas where Japan has shown a desire to develop capabilities that reduce its dependence on the US, i.e. spy satellites and amphibious assault (which is what the USMC does).
Politically, there have also been ups and downs as we know. Indeed, it is amazing how quickly the discourse shifted from the alliance is stronger than ever during the Koizumi period to the alliance is "in danger"
during the Hatoyama seiken, and essentially only because of the fate of a single base on Okinawa.

In this connection it is also important to note that political elites have not been entirely united. To be sure, there has been close to unanimity on the idea of maintaining the alliance itself. In my previous posts I did not suggest that I currently see any political elites advocating an alliance with China, and I certainly do not see that (again, I think it is useful to compare the alliance with other options and to address suggestions that Japan would suffer a terrible fate in a China dominated East Asia, John Mearsheimer has also addressed the later point). Nonetheless, I see a real debate among Japanese elites about what form the alliance should take and what Japan's relationship to China should be. As I mentioned in my last post, Hosokawa suggested continuing the alliance mostly as a nuclear umbrella, while Ozawa suggested reducing the US presence in Japan to naval bases. Hatoyama had related but more modest ideas that crystallized in the idea of possibly moving Futenma out of Okinawa and even Japan.

Hatoyama, perhaps Ozawa, and some retired and career diplomats have pushed to build a much better relationship with China and much tighter integration with Asia. Of course, security is not only about deterring and defending against threats, it's also about reducing and ending threats through negotiation of differences, confidence-building, etc., and they believed there was potential for reducing the China threat in this way. I should add that in my research over the past year I have found that Koizumi might be closer to this group than I had imagined. Under his term Japan did more to promote East Asian regionalism that excludes the US than perhaps any other premier.
And of course Koizumi tried to reduce the North Korean
threat by visiting Pyongyang.

Regarding public opinion, overall support for the alliance with the US is strong, although there is very little support for some proposed forms of cooperation, such as joint combat operations overseas (although missile defense cooperation appears to be well supported). There has also been broad support for reducing US bases, an opinion which Ozawa and Hatoyama were perhaps trying to tap into, although I doubt that was the only motivation. It is also worth noting that an annual Yomiuri polling question has found very high levels of Japanese distrust toward the US.

Regarding a possible Sino-Japanese alliance in public opinion, it's an untested proposition as no pollster to my knowledge has ever asked the question. That is of course another indicator (i.e. questions pollsters don't ask) that this idea is currently beyond the pale of political discourse in Japan. I would add parenthetically, that the answer you get in a poll depends on the question you ask. Ellis writes about Japan's choice as between alliance with the US versus autonomous defense or becoming a "satellite" of China.
If you ask Japanese if they would favor becoming a "Chinese satellite," the number answering yes would likely be extraordinarily small. On the other hand, if you were to ask respondents if Japan should continue being a "US satellite," the number who would answer in support of the alliance would also likely be very low.

Public opinion has obviously been quite negative toward China in recent years. However, I have not seen evidence, at least until the Sept. 2010 fishing boat incident off the Senkakus that the Japanese public was focusing on China as a military threat. While I argue in my book that the ability of elites to influence public opinion is limited in part by their limited influence over real world events, they do have somewhat more influence over bilateral relations. To the extent that a Hatoyama/Asian regionalist leadership was able to build a strong relationship with China, and this manifested itself in a concrete improvement in relations, it is quite possible that the public's threat perceptions of China would decline while feelings of affinity would grow. We saw a bump up in the later at the very beginning of DPJ rule when Hatoyama was trying to draw closer to China.

On final point on autonomous defense: I have not seen much debate among elites about Japan entirely ending the alliance and pursuing pure autonomous defense; Hosokawa perhaps comes closest to that view. However, it's again worth emphasizing that Japan has had a robust autonomous territorial defense for decades that has not depended much on the US, with the partial exception of a full-scale Soviet invasion during the Cold War. China's military buildup is beginning to attenuate Japan's naval and air superiority, but Japan is now responding to that by reinforcing its naval and air power and the defense of its southwestern islands.
I would argue that the priority Japanese leaders are placing on autonomous territorial defense is growing, which does suggest a drift toward autonomous defense, although not in the area of nuclear deterrence.
Results from long standing polling questions on Japan's defense policy do not suggest much support for ending the alliance, but polling also shows, as mentioned above, distrust of the US and perhaps the US defense commitment, and after last year's Senkaku incident strong support for a stronger autonomous military (or
paramilitary) response, so there could be unmeasured public support for a more autonomous defense.

Paul Midford