This is just part of an extremely wide-ranging discussion, covering maths at Cambridge University, mathematical reading, AI, PhDs, the ethics of war, jokes, teaching and Ramanujan.
The original thread can be found here in its entirety, should you wish to read all 34 pages!

Hi again,

If anyone knows of any other AI related books I would appreciate it if any suggestions could be made.

If you know of any teachers/lecturers/students who are interested in AI, maybe you could badger them for some recommendations! Both technical and non-technical books.

Regards
HAL2k+1

Two non-technical books on the mathematical and physical aspects of AI are The Emperor's New Mind and the sequel Shadows of the Mind by Roger Penrose.

Sean

Although I have to mention that although Penrose's books are very interesting and worth reading, they are quite hard. I just read a book vaguely to do with AI which I don't recommend called "Fuzzy Logic" (I can't remember the authors). Marcus Hill recommended "The Uncertain Reasoner's Companion" for reading about this topic, not strictly AI, but it has some relation I think.

I wouldn't recommend "The Uncertain Reasoner's Companion" to anyone who hasn't done some logic and some probability theory, it's intended as a reference and introduction for higher level study and research (late in a degree course or in postgraduate work).

Is Deep Blue (the IBM chess playing thing) considered an AI application? By the way, it is the software that they are naming Deep Blue, and not the hardware that it runs on I gather. I would have thought that a chess playing software can't better itself, the essence of what you would want to achieve in pure AI. So implying Deep Blue is just smart coding (mathematically deep!?)and not AI?

Has anyone used the AI bot at www.AliceBot.org, I think that specifically is not AI either??? How can really pure AI be acheieved, if at all. Can a machine really achieve 'self-conscienceness' or a 'soul', like human beings I guess. Or is that not needed to be branded Intelliegent?

HAL2k+1

I think that I'd say that Deep Blue isn't an AI application, because there is very little resembling intelligence there. The reason it's so good at chess is that it can do an astonishing number of computations per second. I think that true AI can be achieved, but that it'll be awhile, and we might not understand it when we've made it. The question of consciousness and soul is an interesting one, and there is a lot of heated debate on either side of the argument. Some of the major proponents of the "Computers can be conscious" opinion are Doug Hofstadter (who wrote Godel, Escher, Bach), Daniel Dennet and (I think) Chalmers as well. On the other side of the coin there is Roger Penrose, John Searle and others. Any of these people are worth reading, but none of them have any truly convincing arguments.

Someone I know says his ambition is to live long enough to work with a computer that is proud of him.

Toni

It really depends on what you call AI. Certainly the people working on neural networks, expert systems and the like wold consider themselves as working in AI and this is the sort of stuff we find in the text books. They count as AI because they reproduce artificially some function that we perform with our brains, say making decisions based on inconclusive evidence (such as diagnostics in medicine) and predicting things and learn from mistakes. There are two approachs, expert systems attempt to reproduce results we might have ascribed to intelligence but make no attempt to mimic the way intelligence works, so there is a front that looks intelligent but the working is clearly not, such is the case of Deep Blue. The other, bottom-up attempts to start from modelling the brain, these are neural networks.

However, all these things are only good at one particular area, one skill. The multifunction intelligent robot like those in the Asimov books is not on the cards in the foreseeable future, I don't think anyone has any idea how to even begin designing one. And as Dan said we might not even recognise it when we see it. We don't even understand how our own brain works, we don't understand if consciousness even means anything when applied to people, so you can just imagine how hard it would be to build something...

Sean

If you want to have a look at some AI, there's a computer game called Creatures which has little characters whose behaviour is governed by a developing neural network and a genetic code. There's a system of rewards and punishments so that you can train them. Some quite odd things can be made to happen: one of the creators tells a story about two creatures learning how to play catch with each other and I've had one that does everything asleep. There are currently three versions, Creatures, Creatures 2 and Creatures 3, with rising costs, but Creatures can be picked up for about £5 now.

Tom.

Tom, do you have the link or details of where I can get this AI app that you described?

I now also agree that Deep Blue is not pure AI. But surely the techniques and theory created in devoloping small 'singular' thinking progs that do things better than their creators (ie us- the humans) is a step in the right direction to the eventual pure AI 'thing'. In my view, an AI app needs to be built around thinking processes, maybe something not related to reality (ie chair, face, book, place, guilt etc) but a reality of numbers balancing or an inbalance of numbers. Letting some element of 'randomness' involved in the decision processing task. The final hack should be that allows the AI 'thing' to outthink its creator and lead a independent life, setting its own goals and tasks as we humans do. You can think of them as our future children?!

I hope my thoughts are not too cluttered, but I simply put what came to mind of my feeling towards AI.

Regards
HAL2001

Can I just add that I think when we make the break though in into 'real' AI, it will be by a mathematician! ;)

HAL2k+1

As far as whether computers can ever emulate human consciousness, AI is far from what I know well, but I don't think so. Given current notions of causualty, the only way to achieve self-awareness is through infinite regression or a paradox. The infinite regresion is because a possible way to achieve conscious is thought A is detected by thought B, which is detected by thought C, ad infinitum. This leads to never achieving consciousness. The example of a paradox conscious is through thought A being detected by thought B, and thought B being detected by thought A. but close examination reveals that this means that A must precede B, and B must precede A: a paradox (one exacly analogous to Escher's drawing). So, until someone comes up with another way for consciousness, I don't see how it can be replicated. If anyone else has any thoughts on the matter, I would really enjoy hearing them, as the matter of a conscious mind is something I've been pondering in great detail since this discussion .

Thanks,

Brad

Well, certainly programs like Deep Blue are not useless, research into these techniques greatly improves expert systems and the like which have all sorts of uses as described by Sean above. Whether or not they're a step towards true AI is another matter, it might be, as I suspect, that true AI requires a whole new paradigm.

To Brad, your argument for why computers can never emulate human consciousness is, I think, flawed because it also applies to humans, and we know that we do have consciousness. Also, "emulate human consciousness" is a very loaded phrase, using the word "emulate" somewhat suggests that you've already written off the possibility that it could have its own genuine consciousness.

The last point is important, systems like Deep Blue are perhaps more of a red herring than a step in the direction of genuine intelligence because there is no attempt to build something that can works internally as something we might consider intelligent would.

I'm not so sure it would be a mathematician... My bets would be on a neuroscientist, computer scientist or physicist. The first two perhaps for obvious reasons and the third because it may be (and this is what Penrose's books are about) the quantum mechanics (or even quantum gravity!!) is important for a complete understanding of the brain. Mathematicians do not generally make the best scientists, especially when it is a 'messy' science like neuroscience.

Sean

I'm inclined to agree with Sean that it won't be a mathematician, although it might well be someone with mathematical training.

What I'm saying though, is that it will be hard to "write up a program" with a consciousness or even create a machine with a consciousness because we have no idea how we have a consciousness, and the two most probable ways to explain it result in either a paradox, or a lack of causation. So, perhaps it is right that the consciousness lies in quantum mechanics, but how then would we have control (or even an illusion of control, for that matter) over our own thoughts. Furthermore, wouldn't quantum mechanics neccessarily have to incorporate time travel to allude the paradox of self-reference in my non-infinite explanation for the conscious mind? And, wouldn't this make us in control of quantum mechanics?

Brad

Hal, the link to the official site is http://www.creatures.co.uk/ . As far as I know, you can't buy the program there, so you've got a walk round your local computer games stores, or a trawl through the internet.

Personally, I'm inclined to believe that a neural network is sufficient for intelligence and conciousness.

Tom.

Brad, I guess that there are two possibilities; consciousness is a property (in some way) of specific sorts or arrangements of matter OR consciousness is a property of systems with sufficient algorithmic complexity in some way. We really have no way of knowing which is true, or whether it's another possibility entirely. If it's the former, then no piece of software can be conscious, which doesn't rule out the possibility of machine consciousness, it just rules out "consciousness on your PC". However, if it's the latter, "consciousness on your PC" (perhaps Microsoft ConsciousEntity^TM ) is a possibility.

There's the related question of whether or not consciousness is an emergent property or not, i.e. whether it can affect physical events (not emergent) or whether it is just a byproduct of the action of certain sorts of systems.

Another interesting property of consciousness is that it could be a product of a metaphysical realm, rather than just physical realm. At current, I am very inclined to believe that it exists entirely by physical laws, but there is nothing ruling out the other possiblity...

Brad

Hi, Hal (and the rest of you).

This is a great discussion, on a topic which has always interested me hugely. The nature of consciousness is the central theme in "Godel, Escher, Bach", and Hofstadter uses the common theme of recursion and self reference in Godel's proof of his incompleteness theorem, Escher's art (such as the self drawing hands) and Bach's music (notably his canons) to illustrate his ideas.

Consciousness itself is hard to define, let alone detect in another entity. It is easy for me to convince myself that I am conscious. It is harder for me to say that (for example) any of you are conscious, rather than sophisticated text parsing bots. We tend to assume that people we know to be flesh and blood human beings are conscious, but we really do that by comparison to ourselves - "you are human, I am human, I am conscious, therefore I infer that you are conscious". One famous criterion for consciousness (or, as it is usually put, intelligence) is the Turing test. In case you are unaware of it, what you do is converse with two entities via a textual or other indirect interface. The communication is completely free, you can ask whatever you like of either entity. Their objective is to convince you that they are human. There are a couple of problems with the test. Firstly, text parsing bots are becoming increasingly sophisticated - some can fool a decent percentage (30 or 40 last time I looked) of people interacting with them over the Internet (bear in mind one of the central tenets of my life, though - "People Are Stupid"). These bots are things nobody could classify as intelligent - they are at best "conversation expert systems". The second problem is that we are here defining "intelligent" as "like us". It is not difficult to envisage some alien who is as intelligent as we are, completely conscious with a sense of morality and any other paraphenalia you care to hang on the notion of personhood, and who is also totally incapable of passing the Turing test, since its intelligence is simply different from ours.

On the other hand, it is easy to dismiss some intelligent behaviour as merely "simulating" intelligence. An example I recall form GEB is of a (ficticious) argument that women are not conscious, since they do not think, they have internal processes different from real conscious beings' - they "womansee" things, "womanthink" about them and come up with responses that, although identical to those of people who really see and think, are a product of mere mechanistic processes, not conscious thought. It is actually not possible to prove the falsity of this to a man - and a woman claiming it to be false is producing the effects of "womanthought", so this is not proof at all. Now change "woman" to "AI" and you can see the difficulty of defining "consciousness" as having to happen in a meat brain (as opposed to a male brain).

It's a thorny problem, and not one which will be solved until we have an AI which is generally acknowledged as intelligent - which is to say, I guess, that my answer to the question "What is true intelligence?" is "We'll know it when we see it."

Thanks Tom! for the link, I'll check it out.

Brad, Dan, Sean - thanks for sharing your carefully placed out thoughts. AI does indeed interest me and I think the amount of locked potential that could reap benefits for mankind is huge.

The note where we talked about who would the AI breakthrough, my opinion is reformed after reading the above posts, and I feel that a mixture of people with varying expertise in various areas will be the people that possibly makes the AI breakthrough. Team of researchers: mathematicians, neuroscientists, philosophers, computer scientists and an artist.

Regards
HAL2001

I was wondering...
What kind of applications could a real AI 'thing' be used for. How will it help us lead better lives?

HAL2k+1

I would tend to agree that a cross disciplinary team is the most likely contender for a big AI breakthrough.

Hal, rather depressingly, the only applications of AI that I've heard about, apart from web bots and gaming, are military. For example, the company that produces Creatures is also working on using neural networks to pilot fighter planes.

Fortunately, there seems to be a happy incompetence in the armed forces; one program was set up to use neural networks to spot camouflaged tanks, they took photos of scenes with tanks in and without and fed them through the network. It seemed to be working, until they tried a new set of photos and discovered that in the initial set of photos, the ones containing tanks had been taken on a different day to the ones without...so the neural network was able to tell whether there was a tank there or not by whether or not the sky was cloudy.

Tom

Tom, I am not exactly sure (as you and many of the other in this nrich forum might agree too) that AI should not aid and abet the progress of war arsenal (like the highly popularised Einstein and the atom bomb did for WW2). Science helped the destruction of thousands (or even in the long run millions) of innocent people :( If you were offered money by the military to carry out AI research, you would probably say yes, simple because of the money, it'll allow you to carry out your passion. But in the long run your work could/?will? be used to kill other human beings. This is very disturbing. Then again the same can be said for so many other things in life....

Regards
HAL2001

Tom, that story is as old as the hills, my Dad told me that one years ago, and I'm sure that he must have heard about it at least 15 years before he told me. However, I agree with your thesis, remember those guided missiles which almost all missed their targets? I think it was either in Serbia recently or in Iraq longer ago.

By the way, I think I probably would refuse to work on military applications. Call me ethical if you like.

Tom

I don't know that military applications to physics or AI are always evil. Much of the time, fighting a war must be looked at as sorrow leading to happiness. To obtain what is right, it is sometimes necessary to destroy what is wrong. The way I look at some things, particularly the A-Bomb, is that it was killing thousands to save millions.

Brad

Well, I'm a pacifist, so I certainly wouldn't do any research which would lead to direct military applications (it would be silly to try not to do any research that might possibly lead to military applications, as almost anything could). I think it would be very difficult to prove that bombing Hiroshima saved millions of lives, unfortunately my knowledge of history isn't good enough for me to make the case against using the A-bomb on Hiroshima. However the book I'm reading at the moment (Lessons from Kosovo by Noam Chomsky) claims that the effect of bombing Serbia was to escalate atrocities, and that moreover the people in command predicted that this would be the likely outcome of the bombing campaign.

'Evil' is an indefinable and pointless word. All the same, although this post is now rather off topic I must dispute your claim that the Atomic bomb 'killed thousands to save millions.' Certainly, for obvious reasons, this was the claim made by the US after the war. However the allies were winning the war. If the Americans had wanted to demonstrate the power of the bomb they could quite easily have dropped it in a less densely populated area. In fact they dropped it twice over extremely populated areas. I find it hard to draw any conclusion other than that they were trying kill people and lots of them.

Can you give any applications of technology to the military that have not lead to more brutal and extensive destruction? Is there any reason why AI should not lead thus?

Sorry if there was not much mathematics there, but then I suppose all argument is mathematics of a sort.

Just quickly on the subject of the A-bomb: I agree with Brad that the bombing of Hiroshima and Nagasaki was absolutely necessary to end WW2. You have to remember that Japan was then in a fanatical regime, and it is very doubtful they would have surrendered, even though they were rapidly losing lives (and the indications are the figthing was due to get very bloody, as the figthing was moving closer and closer to Japanese territory). The emperor (the dictator) had more or less isolated himself from what was happening outside. In this totalitarian regime individual lives were valued very little (Kamikaze are well documented - they are suicide fighters - it was found to be easier to cause mass destruction if the person causing the destruction didn't have to escape himself, so that's what they did!) So they were not going to surrender for rational reasons, and so the fighting would have continued and very probably many, many more people would have been killed on both sides. Only when the US demostrated they could, and were willing to, use these weapons were the Japanese forced to surrender.

Before President Truman ordered the dropping the bomb, the US dropped leaflets from aircraft informing the Japanese that they had developed a massively destructive weapon (giving near full specifications) warning them that they'd be forced to use it if the Japanese didn't surrender but the Japanese Emperor either never saw this (perhaps people were too frightened to approach him - they might have been executed) or simply didn't respond.

Furthermore there are arguments that the US should have dropped the bomb in the sea or something like that to scare the Japanese government into surrendering. But this was not an option for the very same reason. The Japanese would have been forced to cover up the fact that the US had a new super-weapon. Any other attitude would have been seen as anti- the war effort. Therefore dropping a bomb in the sea would have been a waste of resouces - remember U 235 was very limited in supply and was really the only hope the US had of ending the war quickly.

It goes to show just how fanatical the regime was that even when the US had dropped the bomb on Hiroshima the Japanese government didn't surrender, and it took another bomb at Nagasaki before they surrendered. Does this look like a case in which the war could have been resolved without huge loss of life?

I know it was a pretty terrible thing (about 64,000 Japanese killed at Hiroshima including deaths from radiation sickness later, plus about 30,000? at Nagasaki) but I think the death toll would have been on a different scale otherwise. (There would have been more Japanese deaths as well as American deaths.)

Having said all this, Truman's memoirs on the subject make interesting reading. At one point he seems to imply, if I remember correctly, that he believed that Hiroshima and Nagasaki were not particularly densely populated, and there was in fact a military instillation in these. Whether this was a case of his being misinformed or trying to reconcile his actions falliciously (after all perhaps there was a military instillation in Hiroshima and Nagasaki, but I think he'd be kidding himself if he thought that he was dropping the A-bomb just to destroy these) or plain deception. I suspect not the later, as this deception would not work as there are so many other sources of information.

I think Serbia is much a less clear cut, and I don't really know about the pros and cons.

Michael, this is one of those arguments that can (and probably will) rage for ever and never be answered. Can we answer the question about the ethics of getting military funding (this would include work on number theory funded by the NSA or GCHQ) without reference to the A-Bomb?

Whoa Michael - I think your account is not historically sound, I understand that there is ample evidence that Japan was in fact about to surrender, that the US knew this, and that the effect of dropping the bomb was aimed more at the USSR than Japan.

The are also cultural problems in your account, for instance the role of suicide in Japan was not something that suddenly sprung up due to a 'totalitarian' regime but dates way back to at least Samurai honour codes.

Dan and others - anything by Chomsky is good. There is something of a parallel in fact between the Serbia bombings and the first A-bomb. The bombing of Serbia was a major conflict is post cold war Europe, the fall of the USSR effectively left NATO (a 'defence' organisation) without it's official role and the bombing had everything to do with NATO redefining itself as an offensive force and also everything to do with the US establishing itself as a now supreme power. This parrallels slightly the A-bomb case because the actual the desired objective is, amongst other things, a demonstration of power.

Sean

Totally agree with Sean above; the dropping of the A-bomb was a very dubious action. And I did History GCSE, so I obviously know what I'm talking about. Completely different periods, but still :).

About military funding generally, well, probably depends on the conflict. I don't think anyone here would knock Turing. In any case, I think there is a distinct ethical difference in working on applications during a war, in which you are aware of the causes of that war and uses to which your work is being put and working on military applications during peace-time with no knowledge of or control over the applications to which your work will be put.

Tom

Tom, your point about Turing is well made, I certainly wouldn't knock him.

Of course I can't claim to be an expert on the A-bomb, but I did actually research it pretty thoroughly for my GCSE History coursework. This included reading through many sources, some biased American, some biased Japanese, some biased USSR and some supposedly "neutral". I agree Truman was no saint, but personally I don't see what else he could have done to prevent an astronomical loss of life on both sides. I did not find much good evidence that Japan was ready to surrender. The fact that the Japanese were informed about the capability of the bomb, and in fact they didn't surrender even when the first bomb was dropped shows that they weren't about to give in.

I find it extremely difficult to imagine that the US killed 100,000 innocent citizens simply to "scare" the Russians (as Kruschev later claimed). Surely dropping the bomb in the sea would have sufficed for this purpose? (Of course this is not ideal either, due to the problems of nuclear waste, but I don't think these had been fully realised at the time and anyway, this must have been better than killing 100,000 people.)

The US actually developed the bomb having been warned by Einstein and Szilard that Hitler's Germany was developing an A-Bomb. On balance I have little doubt it would have been better if nuclear energy had never been discovered, despite the fact that I believe WW2 would have been more devastating if it hadn't been ended quickly. I think the use in WW2 has to be regarded as an exception. Since then it has only helped to make bad situations worse, for instance in the cold war. The problem was that if two countries had a dispute, and one was allied to the US and the other to the USSR, then potentially the two superpowers could come to the aid of their allies. The danger was that one of the super-powers could resort to nuclear energy in the dispute between the two countries, after which the other superpower might well follow suit, and then the crisis could escalate to a full nuclear war. Luckily this never happened.

Of course the obvious example is the Cuban missiles crisis, where it was only really because Kruschev, the successor to Stalin, was sensible enough to step down at the vital moment that a world-wide disaster was prevented. A more fanatical leader would not have given in, and the consequences could have been devastating. The fact he did give in was actually one of the main causes of Kruschev's downfall in the end, but it was an extremely good decision.

The US were unlikely to give in as otherwise the Russians would gain a nuclear base in Cuba which would jeopardise US security (apparently it would only take a few hours for a bomb launched in Cuba to arrive in Califournia but I'm not sure about this detail; certainly it is unlikely the US could have responded in time). Of course if the US hadn't behaved so irresponsibly towards Cuba in the first place (the Bay of Pigs etc) this could all have been avoided, but the presence of nuclear bombs made it all the more frightening.

Now that more and more countries have nuclear capabilities it can only make the world a less safe place. Yes OK you can argue that lots of countries having nuclear arms is a deterrent for other countries setting off the bomb, but of course it would be far safer if every country was disarmed. In a way, countries having nuclear weapons is analogous to people having guns. People might claim they need guns to defend themselves and countries say they need nuclear weapons for the same reason. But you only have to compare the murder rates in the UK and the US to see that easy access to weapons is a big problem. It only takes one lunatic with a gun to kill a crowd of people. Could the same problem occur if all countries are nuclearly armed?

You ever seen the movie "War Games"?
Sometimes there is no winner.
War is something like that.

Getting back to the path of AI, suppose that we had a AI 'thing'. If you feed it in the variables and situations for all the elements in about the time of WW2, what would it have deicded to end the war as quick as possible. Remember the AI's only task is to figure out the quickest way to end the war, and would have no bias towards any side. What would it suggest? Would it have made the same choice to drop the Bomb?, even twice? What would have been its reasoning. Is there a name for this type of reasoning?

Regards
HAL2001

Just a small point, but the argument that dropping a bomb as a warning would clearly have been ineffective as the Japanese did not immediately surrender after the Hiroshima bomb does not take into account the fact that the Japanese actually surrendered on August 14 1945, 5 days after the Nagasaki bomb. It should also be pointed out that a Soviet invasion of Dongbei on August 9th (the USSR only declared war on Japan on August 8th) may also have been a factor in the Japanese surrender given the imperialistic fear of Communism.

In view of these factors, it seems not unreasonable to condemn the bombing of Hiroshima and Nagasaki, not because it was unnecessary, but because it may have been unnecessary. If the USA had previously dropped a bomb in an uninhabited but observable region in or near Japan and the Japanese had failed to surrender, there would be a much better case for the dropping of the bomb as necessary in order to end the war.

Tom.

Michael, another way of seeing the fact that the Japenese didn't surrender after the first bomb is that the bomb made no difference to their surrendering. Also, I don't find it hard to believe that the US would drop the bomb to scare the Russians, they were (and many still are) fanatically anti-Communist.

I agree with you about the analogy between nuclear weapons and guns, but unfortunately it's a prisoner's dilemma situation now, which means that nobody will disarm, certainly not unilaterally (they might say they have though).

Hal, it depends who had designed the AI, if the Americans had designed it, it would have decided to drop the bomb, otherwise probably not.

1) the dropping the bomb into the sea argument is a red herring because it is manifestly obvious that however big a wave you might create it is not going to have as much of an impact on world conscious (or conscious of political leaders) as the well known pictures of a devastated city.

2) I think it is a little ironic when people talk about nuclear or biological weapons being in the hands of 'lunatics' who might use them 'irresponsibly' when the only nation to have used a nuclear weapon at all is the US and the US also used large quantities of biological weapons in Vietnam. Which doesn't leave in much of a position to speak.

3) The situation is more subtle regarding 'scaring' the USSR. In particular, when the bombs were dropped, Russian troops were advancing across the then Japanese occupied Manchuria. So by making Japan surrender EARLIER than it would have otherwise (i.e. before Russian troops reached Japan), the US was able to ensure the USSR had no say in what happenned to post-war Japan.

4) There were three days between the bombings... not particularly much time. If the Japanese were so careless about the lives of people why should they have cared whether it was 100,000 or 200,000 that died? There is no logic in saying that the fact that they surrended after the second and not the first shows that they were not about to surrender. One bomb would have been quite enough for the purpose. Three days is not enough for a proper assesement of damage, evaluation of situation etc.

5) since early 1945 the US had been conducting firebombing raids on Japanese cities, specifically designed to kill as many civilians as possible. So there is no grounds to credit the US with a care for life.

6) By the spring of 1945 it was clear that Japan was militarily exhausted. Its air force and navy had been virtually destroyed and in April the war government fell and was replaced by a cabinet under Admiral Suzuki who was keen to look for ways to make peace. There were divisions in the Japanese leadership over this issue.

7) Did the US know? Well, they had cracked Japanese codes and intercepted messages which were, in the words of US secretary of the navy James Forrester, 'real evidence of a Japanese desire to get out of the war'. In the week before the Potsdam Conference, cables revealed the formal decision of a Japanese Imperial Conference to try to stop the fighting.

On 2 August the Americans intercepted a cable from the Japanese government to the Japanese ambassador in Russia: 'The battle situation has become acute, there are only a few days left in which to make arrangements to end the war... it is requested that you immediately have a talk with Molotov.' On 6 August the atomic bomb was dropped on Hiroshima.

What else is there to say?

Sean

Another reason why Japan surrendered was that America threatened to bomb Tokyo. We still do not know whether US had another A-bomb at their disposal at that stage, some sources said it was a bluff, and some said they would have bombed Tokyo on the 14th August.

It was certainly true that Americans don't care about civilian lives in the war, despite their claim that dropping the bomb on the bridge in Hiroshima would minimise civilian casualties. In fact, did anyone in power care about civilian lives in WW2?

To address Sean's points.

1) Yes, true. But I would like to think that the factor of saving 100,000 lives would have played some influence.

2) I am not speaking on behalf of the US (at least that is not my intention). I have mentioned other instances above where their international policy has been less than responsible. I don't know enough about Vietnam to comment, but on the subject of Hiroshima/Nagasaki my view was that the US dropped the bomb to end the war. A few of your later points have started to shift my position.

3) Even if Japan was attacked on two fronts by Russia and the US, this doesn't necessarily mean Japan would have immediately surrendered. They would probably have fought to the very end, and there would have been a lot of bloodshed.

4) I would have thought that there could be little doubt how destructive the atomic bomb was. OK, they won't have realised that that bombing was going to result in 80,000 deaths, but they must have noticed how much damage was caused with minimal effort. I don't think they knew that U 235 was so short in supply, so for all they knew the US might have gone on and on dropping A-bombs for months. I think if I was in their position I would have surrendered immediately after the 1st bomb to prevent it happening again. And if I was already thinking of surrendering, my response would have been instantaneous.

5) Well it was war. My guess would have been that the US wanted to make the Japanese surrender without having to resort to atomic energy. After a while it became clear that the Japanese were going to hang on for as long as possible, so the US decided to end it. As I say I am less sure now.

6-7) These are the two crucial points. I was aware that militarily the Japanese were exhausted, but again I think they could have hung on for a lot longer.

But what really concerns me are the two quotations you mentioned. If I had been aware of these my view would have been very different. Out of all the sources I looked through (both in textbooks and outside) I found very little to suggest Japan was on the verge of surrender. Even my history teacher (who incidentally was personally against the use of atomic energy in WW2) never mentioned once that the US had intercepted Japanese messages saying they were just days away from surrender.

I think I should qualify one other thing in my first post to this topic - I say at the end that Serbia is less clear cut than Hiroshima, implying that Hiroshima is clear cut. This was not my intention, and what I really meant was that the case for atomic energy in WW2 was a lot stronger than the case for using guided missiles against Serbia but I don't know that much about the later case.

If the bomb stopped the war, then it was for the better good? Not though for the people of Nagasaki & Hiroshima :(

I personally think that a nuclear bomb will never be used again on civillians, within my lifetime anyway, but then again I am an optimist.

Its strange that arguably in the most troubled region of the world, middle east, a nuclear bomb has never been used. Though trouble has been brewing in Israel, I don't think a sane leader in the modern age would ever give the go ahead to launch a nuclear bomb on anyone, no matter whatever is on stake. But then again many leaders are not sane?

The point about the US launching the bomb twice to scare the USSR is somewhat illogical when saying that it would only take 1 bomb to scare the Japenese. If it was known that only 1 bomb would be needed to "scare" the world, then why would two have been dropped to scare the communists. When it's not your own people, "quality matters, not quantity". So why would the U.S. have dropped the second bomb just to scare the USSR. It was either dropped twice to defeat the Japenese, or because of military ignorance. Given some of the arguments already on the page perhaps it was simple ignorance (something all too common in American Politics). But, I still think that the bomb accelerated the end of the war.

Finally, the people designing the bomb were certainly not doing evil, as if the bomb had been dropped as early as possible many lives would have been saved. It has been estimated that half a million lives were lost every month in the War. That would mean that if the team could have gotten the bomb ready 2 months earlier, we would be granting 1 million lives in exchange for .2 million. This, while hurting the individual will help so many more individuals. But, I am as willing as anyone to admit that there was a bit of a double motive in the US government's use of the bomb (to use the bomb for intimidation). I am not willing to admit that the US merely used the bomb as a political weapon though.

Brad

Fair point about the two bombs Brad, my guess, but it just a guess, would be that it was always planned to drop two, seeing as it was the first time one was had even be really used, there might have been quite a high probability of something going wrong, hence the need for two to be sure one would work. Of course, as you said there may not be a rational explanation.

I agree the bombs accelerated the end of the war, but only by the few weeks that were necessary for the war to end before Russian troops reached Japan.

Sean

The conjectured intimidation of Japanese would have been for a very different purpose than the conjectured intimidation of the USSR. I don't see the rejection of the former as necessarily contradicting the latter.

If the idea were to force the Japanese to surrender than I cannot see how more than one bomb would be necessary. If on the other hand the aim was to demonstrate American military strength then there is every reason to destroy as much as possible.

I do not think that a simple measure of the number of deaths is a well-defined notion of morality. After all the allies could have ended the war right at the start with no deaths as a result of combat just by surrendering to the Nazis.

Michael - I've finally tracked down a reference for some of the quotes I used above. Have a look in the book "The Decision to use the Atomic Bomb", by Gar Alperovitz, New York Vintage Books, 1995. There are quotes there from American military and government people that really leave little doubt.

Sean

Thanks Sean - I'll have a look for that in the library later. I'm a bit surprised that I didn't come across anything like this in any of the sources I looked at, and that my teacher (strongly against the A-Bomb) didn't mention it. Maybe it has only been uncovered recently. Anyway thanks for the reference,

Michael