Is the blockchain's purpose being redefined by the forked Ethereum Community?

[AgentofCoin]

The blockchain was intended to maintain immutability ONLY within itself and not from outside intervention.
Outside intervention is left open so that the protocol can be upgraded by "consensus" when the time is appropriate.

I suppose that by "inside" you mean "code", the technical part, not the sociological part.
Well, it is obvious that that cannot work.  There is absolutely no "code/technical" way to establish consensus if there is no sociological/game theoretical part to the system.

You cannot just have code, without any game-theoretical assumptions, and suppose that any form of immutability will come out of that by itself.  It is sufficient that people change the software, and the block chain can be totally altered, or its interpretation can be totally altered.  If the software checking for the validity of PoW is altered, no amount of PoW will keep what so ever immutable.  

Immutability comes from game-theoretical (and hence also sociological) assumptions, not from code.  If people start propagating altered versions of the block chain, different software checking validity and so on, no piece of code can ever stop that and "provide immutability".  The whole immutability comes from the fact that people will be put in such a situation where it is very unfavourable for them, or essentially impossible to organize for them to do anything else but to STICK to the immutable consensus.

Not only do I mean code part, but I also mean mechanically and functionally.
The game theoretically mechanisms of the blockchain, only apply and function within the blockchain.
I do not believe that game theory applies to consensus.
Game theory can not prevent malicious consensus, only a assumed play fair consensus.

Immutability can never be guaranteed within any human devised system.
The only aspect of game theory that would apply to immutability is that the gatekeepers
would keep the old immutable chain alive, since their is no action the gatekeepers can take to
prevent a malicious consensus hardfork.

Game theory does not prevent a violation, it only uses human assumptions to control the humans.
If I could bribe everyone with more value, they may destroy their own chain willing.
That is not a failure of the chain or the game theory aspect, but of their community and its own ignorance
of what their own job is. Users and other members of the community are not passive in this type of system.

Game theory in itself is a failed theory, since it assumes humans would not risk value over destruction.

Without this intervention by the devs, the code could never be "upgraded, fixed, or etc.", so version 0.1 would be the only one.
The consensus mechanism exists outside of the blockchain code and there is no sociological code within the blockchain.
The sociology aspect that you are relying upon, is actually based upon a community/miner/node gatekeeper premise.

But it is essential.  If the devs could intervene at their will to change those things that are supposed to be immutable, your block chains would not keep any little bit of immutability, and be totally mutable and random.  The whole point is that even the devs cannot do this.

However, they CAN change the code, as long as this code remains COMPATIBLE with those aspects that are deemed needed to keep immutability, and that itself is the whole game-theoretic premise on which the whole of immutable block chains is based.

That is:
- the devs (or just anybody) can bring out another node software that is perfectly compatible with the former v0.1, but, say, with a fancier GUI, or with faster algorithms, *perfectly* respecting the protocol on the technical side.  You could have myriads of different but compatible node software.

- the devs (but not just anybody) can bring out node software that is technically not totally compatible with the former v0.1 and hence is technically a soft or hard fork, but which RESPECTS IN ALL ASPECTS the INTEND of the former protocol, and just improves a technicality.  The reason why people will upgrade is that the intend of immutability (all the economic implications) is perfectly respected, so there is no reason NOT to upgrade (like accepting a security update to your computer system), and not upgrading can put you with an incompatible node.  In other words, game-theoretically, there's no good reason not to upgrade if people agree that the technicality is improved now.

- and that's it.  The devs, or any other "leadership" shouldn't be able to modify whatever to the intend of the immutability.  If they can, then there is no immutability at all.

I don't understand.
You are basically arguing that if a blockchain that is immutable now, is made mutable by a violating change later,
then the blockchain was never immutable to begin with? If yes, then you are arguing semantically and I would agree
semantically. I would then go on to advise you that there is no blockchain that is immutable and there will never be,
by your semantical reasoning. In fact, I would also advise that humans will never be able to create any perfect
immutable system, since all matter (in this universe) is mutable.

Semantical reasoning and argument is not beneficial toward true understanding of this system.

If the gatekeepers (majority of consensus) wants to destroy the blockchain, it will happen, since Satoshi intended it to.

Well, I consider that as a failure of the block chain purpose.  If that happens, the block chain didn't do what it was made for, namely keeping immutability of intend.  Once you can do that, anything goes.

Satoshi did not create or intend perfect immutability and I'm not aware of where he made claims of such.

If Satoshi intended it to, and in fact designed it as so, then the failure rests upon the gatekeepers and not the blockchain tech.
The community failed to maintain their social contract and their "ideals".

Call it as you like: the system failed.  The "code" didn't fail, but the block chain system failed, because the game-theoretical assumptions turned out not to hold.

The game theoretical assumption(s) do not apply to consensus.
They only apply within the blockchain's functional dynamics.
Consensus exists outside that function, so game theory does not exist with consensus.

The consensus mechanism does not exists within the blockchain and each altcoin's community can set out and
agree to their own consensus mechanism. The consensus mechanism was NEVER intended to prevent changes
to the original rules.

That is a strange statement.  I will tell you why:

The consensus mechanism is ACTUALLY intended to allow for a way to make changes, if
deemed appropriate by the majority of the community. If the community wants to destroy their original social
contract, and they have majority, they will destroy it. Consensus allows possibility for change, not prevent it.

I would rather think that this was a shortcoming of the technology, rather than a feature.  That is, one couldn't think of a mechanism where one could even block a colluding majority from imposing its will (in fact there is one: hard forking).

Humans can not create perfect systems, so, humans will never create perfect immutability.
Hardforking is not a mechanism to prevent the colluding majority, it is the result of their collusion.
When a hardfork occurs, it is because the majority has agreed (in theory) to that new version.
Maintaining the old chain is the only way to prevent a full "imposing of the collusion".

Because when you think about it, if a colluding majority is the "normal way to alter block chains", then there is nothing wrong with a "51% attack".  It wouldn't be an attack at all.  It would just be the normal majority wanting to alter history, if you keep your viewpoint.

No, a 51% majority of the hash power is not the same as a majority of node consensus signaling, for a protocol change.
A 51% attack is a malicious action by a miner, who has attained 51% share of the hash, and now is
attempting to revert/edit/undo past work ON THE CHAIN that will benefit themselves (and destroying the Bitcoin system outright).
A 51% attack is not power to change the protocol and its intended function, but only power to reverse work on the chain.
The reason why miners do not reverse work is because game theory says it will cause loss in value of the chain and then token.
But a malicious government miner attacker doesn't value that, so they could perform this type of attack and
the game theory will not apply to them, since their intention is only the destruction of Bitcoin.

The idea of a block chain was not to have a "majority vote over what was the distant past", although that is what comes technically out of it.  The fact itself that you call a "community where a majority decides to DESTROY their block chain" indicates that you yourself consider that even if a majority decides something, that this is NOT what consensus is about.

Imagine that a majority of bitcoin miners decides that only they will now hold bitcoins, and that all former transactions that assign coins to addresses of which they do not have the keys, are invalid ; for that they have to write a lot of "exception rules' in the code accepting modified old blocks.  Hell, in your viewpoint, such a total alteration of the past would be "consensus" because decided by the voting majority of miners ?  Of course not, this is total breaking up of immutability and - as you actually call it - the destruction of the block chain system.  Nevertheless, the CODE is running correctly, has been altered with majority consensus and "everything works by consensus as wanted by Satoshi" ?

I don't understand.
Miners can not do what you are describing. Your example will be rejected by the rest of the network.
There are thousands of nodes and other operator systems that form the full consensus mechanism.
You can get all the miners on board, but that does not guarantee full consensus if the users and economies do not agree.
For Bitcoin consensus, you need around 95% of the whole community, is my understating.

Now if the whole community did allow the miners to do what was in your example, than yes,
the community failed itself by coming to consensus to allow a social contract violating action into the system.

A block chain system is not supposed to be a majority vote system over just anything.  The only thing the majority is supposed to vote over, is the respect of immutability.  If anything else happens, the system fails.  So the game-theoretical part is an essential part of the block chain system, because without it, there's no such thing as immutability, and hence no purpose for a block chain (the code/technical part).

Yes, Bitcoin is not entirely just a majority vote = change. But the community can vote on anything, technically.
But the fact that a vote is done and agreed upon does not mean that the result is a correct action.
Sometimes humans vote for their own destruction (ex: Ethereum hardfork that violates social contract).

Bitcoin's Immutability aspect is not governed nor maintained by game theory in any way.
In fact, I think there is no possible mechanism that can be created to stop humans from being human.
Immutability is only protected by gatekeepers who are informed and stay vigil.

[1714686664]

1714686664

[1714686664]

1714686664

[1714686664]

1714686664

[1714686664]

1714686664

[1714686664]

1714686664

[1714686664]

1714686664

[dinofelis]

Not only do I mean code part, but I also mean mechanically and functionally.
The game theoretically mechanisms of the blockchain, only apply and function within the blockchain.
I do not believe that game theory applies to consensus.
Game theory can not prevent malicious consensus, only a assumed play fair consensus.

There is no difference.  Calling things malicious or not is a moral verdict, but doesn't change a thing.  I call the ETH fork malicious, but others will call it "fair play consensus".  It doesn't change the fact that it got implemented.

Immutability can never be guaranteed within any human devised system.
The only aspect of game theory that would apply to immutability is that the gatekeepers
would keep the old immutable chain alive, since their is no action the gatekeepers can take to
prevent a malicious consensus hardfork.

But that's the whole (sociological/game-theoretical) essence of the block chain invention: that it is SOCIOLOGICALLY and GAME-THEORETICALLY not possible to find a consensus, other than the immutable one.

That was, in my idea, the fundamental assumption of a block chain: that it is sociologically not feasible to find any majority collusion over anything else than the original rules, and the correct history recorded until then ; that the advantages that it can bring by breaking that, would never reach a majority in a very diverse and large set of actors, and that never ever, a majority would find an advantage in doing anything else but keeping with the original rules, and keeping with the correct history (in other words, that no majority can be found that finds an advantage in a *specific* deviation from immutability, although almost every actor would find an advantage in a *different* deviation from immutability).  In other words, that immutability is the least bad solution for the majority.

Game theory does not prevent a violation, it only uses human assumptions to control the humans.

Game theory assumes of course that every actor optimizes its gains ; the idea is that there is no possibility that every *specific* deviation of immutability would never profit a majority of CPU power more than keeping immutability.

That is not a failure of the chain or the game theory aspect, but of their community and its own ignorance
of what their own job is. Users and other members of the community are not passive in this type of system.

There is no "job" and there is no "community".  There is only a set of people looking for maximal gain.  The idea of a block chain is that those people are constrained, against their own wishes, to keep immutability.  Call immutability a Nash equilibrium of the game-theoretic analysis of a block chain setup.

I don't understand.
You are basically arguing that if a blockchain that is immutable now, is made mutable by a violating change later,
then the blockchain was never immutable to begin with?

You could put it that way.  But you could also say that during the period of immutability, people have had the opportunity to use it as an immutable system.  Those that crossed the "border of immutability" had bad luck, like people keeping stock of a bankrupt company.

It is somewhat similar.  The stock of a company reflects normally the diminished cash flow of future dividend.  As long as the company is live, the value of its stock is the market's best projection of its future cash flow.  So if you buy stock, and you sell that stock later, while the company was running all right, you have been using the "stock system" to your likings.  When the company goes broke, the stock drops to 0.  Those holding the stock at that moment, have bad luck.

Same for a block chain of which the immutability is broken.  As long as it wasn't broken, it was running like a company stock, and you could buy it, get dividend, and then sell it again.  Once immutability is broken, it is like a company that went bankrupt.  (with ETH, the funny thing is that people still trade the stock of a bankrupt company as a betting token...).

The idea is that the chain doesn't break, like the idea is that a company doesn't go bankrupt.  But things can fail.

Satoshi did not create or intend perfect immutability and I'm not aware of where he made claims of such.

As I told you already, this is my own analysis.  I don't consider Satoshi as the final prophet of truth, although he did bring a lot of good ideas in.  So I would think that in as much as Satoshi understood the system he was building, he had to assume immutability (in INTEND, not in technicalities !).  I'm somehow convinced that he intended that for instance, 21 million bitcoin was to be something immutable.  He talked about irreversible transactions.  For someone assuming mutability, that would be strange statements, no ?
But in as much as he was confused about that (I don't think so), that's no reason to remain confused about it.
Einstein invented General Relativity, but people have improved upon it even if Einstein didn't understand certain aspects of his own theory in the details later scientists understood it. 

The game theoretical assumption(s) do not apply to consensus.

Of course it does !  It is the essence !

[AgentofCoin]

Not only do I mean code part, but I also mean mechanically and functionally.
The game theoretically mechanisms of the blockchain, only apply and function within the blockchain.
I do not believe that game theory applies to consensus.
Game theory can not prevent malicious consensus, only a assumed play fair consensus.

There is no difference.  Calling things malicious or not is a moral verdict, but doesn't change a thing.  I call the ETH fork malicious, but others will call it "fair play consensus".  It doesn't change the fact that it got implemented.

ETH's consensus mechanism was lowered, so that they could argue that majority reached a consensus for hardfork.
The ETH community argued that since ETHers who had a vote, did not vote, that non-vote is a vote for yes,
as opposed to a vote for no or neither. So in their new mechanism, an empty vote is a yes vote.
That is malicious consensus, since in normal systems, those empty votes fall under no or just not counted at all.
It was not a "play fair consensus", even if others would wish you to believe that. They changed the rules.
There is a real difference between real consensus and faked consensus.

The fact that you agree that ETH was able to implement it's hardfork in light of the game theory you claim exists
to prevent such change, only proves my argument that game theory does not apply to this form of consensus.

I think your real theory is, without articulating it, seems to be that:
hardforks are a failure (or a failure vector) for the immutable blockchain system.

Immutability can never be guaranteed within any human devised system.
The only aspect of game theory that would apply to immutability is that the gatekeepers
would keep the old immutable chain alive, since their is no action the gatekeepers can take to
prevent a malicious consensus hardfork.

But that's the whole (sociological/game-theoretical) essence of the block chain invention: that it is SOCIOLOGICALLY and GAME-THEORETICALLY not possible to find a consensus, other than the immutable one.

That was, in my idea, the fundamental assumption of a block chain: that it is sociologically not feasible to find any majority collusion over anything else than the original rules, and the correct history recorded until then ; that the advantages that it can bring by breaking that, would never reach a majority in a very diverse and large set of actors, and that never ever, a majority would find an advantage in doing anything else but keeping with the original rules, and keeping with the correct history (in other words, that no majority can be found that finds an advantage in a *specific* deviation from immutability, although almost every actor would find an advantage in a *different* deviation from immutability).  In other words, that immutability is the least bad solution for the majority.

No. You have mixed up apples and oranges. You have taken two separate systems with totally separate rules
and laws, devised as such intentionally, and have combined them incorrectly to form the basis of your argument.
You are incorrectly combining internal blockchain consensus, with external forking consensus.
Game theory consensus can not apply to malicious external forking consensus, such as what ETH did.

Game theory does not prevent a violation, it only uses human assumptions to control the humans.

Game theory assumes of course that every actor optimizes its gains ; the idea is that there is no possibility that every *specific* deviation of immutability would never profit a majority of CPU power more than keeping immutability.

There may be double negatives in your statement, or not. I don't understand your statement.
Please rewrite your intention in a different way.
Game theory only works within the system that it was specifically designed for.

That is not a failure of the chain or the game theory aspect, but of their community and its own ignorance
of what their own job is. Users and other members of the community are not passive in this type of system.

There is no "job" and there is no "community".  There is only a set of people looking for maximal gain.  The idea of a block chain is that those people are constrained, against their own wishes, to keep immutability.  Call immutability a Nash equilibrium of the game-theoretic analysis of a block chain setup.

You are talking about internal blockchain participants, not the external ones.
External participants are not bound by the consensus systems of the blockchain.
In fact, they have separate consensus that is only controlled and monitored by
the "community" and not game theory or code function.

One of the jobs of the community is to monitor developer commits/pulls to each new version of Bitcoin.
That is one reason (there are many other reasons as well) why Satoshi created it as open-source software.
Satoshi placed the future of the system in the hands of the public gatekeepers, to prevent private control.
If the blockchain was intended to be what you are arguing, then it doesn't require open-source or observance.

I don't understand.
You are basically arguing that if a blockchain that is immutable now, is made mutable by a violating change later,
then the blockchain was never immutable to begin with?

You could put it that way.  But you could also say that during the period of immutability, people have had the opportunity to use it as an immutable system.  Those that crossed the "border of immutability" had bad luck, like people keeping stock of a bankrupt company.

It is somewhat similar.  The stock of a company reflects normally the diminished cash flow of future dividend.  As long as the company is live, the value of its stock is the market's best projection of its future cash flow.  So if you buy stock, and you sell that stock later, while the company was running all right, you have been using the "stock system" to your likings.  When the company goes broke, the stock drops to 0.  Those holding the stock at that moment, have bad luck.

Same for a block chain of which the immutability is broken.  As long as it wasn't broken, it was running like a company stock, and you could buy it, get dividend, and then sell it again.  Once immutability is broken, it is like a company that went bankrupt.  (with ETH, the funny thing is that people still trade the stock of a bankrupt company as a betting token...).

The idea is that the chain doesn't break, like the idea is that a company doesn't go bankrupt.  But things can fail.

Except your example lacks one important detail which causes it to become entirely invalid and not in context.

Normally, a company will go bankrupt because of a past incorrect action that led to the future overall financial failure.
So, bankruptcy is a 50/50 possibility in all forms of business where value and attaining more value is the goal.
What actually occurred with ETH is that "their stockholders" willingly and purposefully voted to become "bankrupt",
as per your example. So in effect, they voted to become a mutable chain through the external consensus process.
That is not a company (blockchain) failure, but a stockholder (community) failure. They willingly voted to lose future
value by removing immutability, to undo a failure with the product they were "selling".

How does a "game theory consensus", prevent an uneducated community from a majority destructive vote?

Satoshi did not create or intend perfect immutability and I'm not aware of where he made claims of such.

As I told you already, this is my own analysis.  I don't consider Satoshi as the final prophet of truth, although he did bring a lot of good ideas in.  So I would think that in as much as Satoshi understood the system he was building, he had to assume immutability (in INTEND, not in technicalities !).  I'm somehow convinced that he intended that for instance, 21 million bitcoin was to be something immutable.  He talked about irreversible transactions.  For someone assuming mutability, that would be strange statements, no ?
But in as much as he was confused about that (I don't think so), that's no reason to remain confused about it.
Einstein invented General Relativity, but people have improved upon it even if Einstein didn't understand certain aspects of his own theory in the details later scientists understood it.  

Yes, but Satoshi very well understood that game theory for immutability only exists within the blockchain,
and when it comes to outside the blockchain, such as things like hardforks, it does not apply.

The game theoretical assumption(s) do not apply to consensus.

Of course it does !  It is the essence !

No, it does not apply to all types of consensus within the system.
Game theory can only be applied to the internal blockchain consensus.

[Amitabh S]

Its been very interesting following the Ethereum fiasco. Eagerly watching how the ETC/ETH thing turns out.

Found an interesting quote from here:
http://nakamotoinstitute.org/mempool/ethereum-is-doomed/#selection-357.0-357.305

Had the attacker lost money by mistake, I am sure the devs would have had no difficulty appropriating his funds and saying “this is what happens in the brave new world of programmatic money flows.” When he instead emptied out coins from The DAO, the only consistent response is to call it a job well done.

 

[dinofelis]

I think your real theory is, without articulating it, seems to be that:
hardforks are a failure (or a failure vector) for the immutable blockchain system.

Eh, yes, of course, if you understand by "hard forks", hard forks that modify more than just technicalities, but modify the original intend.  That's indeed exactly what I mean. 

For instance, a bitcoin hard fork that changes all integer representations from little-endian to big endian, is a hard fork, but doesn't modify the original intend of the bitcoin protocol.  It is a technical hard fork, but not a modification of immutability.  So that's OK.  But, say, changing the emission rate of bitcoin, is a change in the intend of the bitcoin protocol, and yes, that is a failure of the immutable block chain system.

It is almost tautological: if there is mutability, then logically, the immutable system failed, no ?

[dinofelis]

No. You have mixed up apples and oranges. You have taken two separate systems with totally separate rules
and laws, devised as such intentionally, and have combined them incorrectly to form the basis of your argument.
You are incorrectly combining internal blockchain consensus, with external forking consensus.
Game theory consensus can not apply to malicious external forking consensus, such as what ETH did.

These are, to me, two aspects of the SAME concept of immutability, indeed.  After all, a hard fork, or the "internal block chain consensus", are almost identical processes.

After all, the "internal consensus" is agreeing which blocks are valid blocks.  This agreement depends just as well on the rules one wants to apply to decide whether they are valid in the first place, as to the application of those rules to the chain and the blocks themselves.   If everybody were to write their own node software, and implement their own vision on their own interpretation of the rules they want to see, we would immediately recognize that the internal and external forking consensus is in fact the same thing.  It is only because we all just use the "dev version" of the software that it seems that the internal consensus is what happens when we run the dev version, and "external consensus" is when we decide to download and install or not, a new version.

Suppose that different miners publish different blocks with different transactions in them.  If I build (or accept) one block over another, I'm applying "internal consensus" according to certain rules ; but according to whether I change these rules or not, I'm applying "external consensus".  If you think about it, this is in fact the same principle.  The best way to see that is if I wouldn't use software, but if I would "check the blocks I prefer" by hand.    Then my preferences, and my way of selecting (and constructing) blocks is MY way, and if I find consensus over that, there is no clear distinction any more between "internal" and "external". (of course this is practically not possible because of the computational load it represents, but as gedanken experiment, you see the point I hope).

[AgentofCoin]

I think your real theory is, without articulating it, seems to be that:
hardforks are a failure (or a failure vector) for the immutable blockchain system.

Eh, yes, of course, if you understand by "hard forks", hard forks that modify more than just technicalities, but modify the original intend.  That's indeed exactly what I mean.  

For instance, a bitcoin hard fork that changes all integer representations from little-endian to big endian, is a hard fork, but doesn't modify the original intend of the bitcoin protocol.  It is a technical hard fork, but not a modification of immutability.  So that's OK.  But, say, changing the emission rate of bitcoin, is a change in the intend of the bitcoin protocol, and yes, that is a failure of the immutable block chain system.

It is almost tautological: if there is mutability, then logically, the immutable system failed, no ?

A hardfork, more often than not, modifies more then just technicalities. A hardfork literally makes changes that
are currently violations and not allowable, and transforms that aspect to normal allowable protocol process.
Hardforks are not normally used to just upgrade the network. In fact, hardforks actually breaks the network.

What you have been arguing all along is that blockchains are not immutable if they can become mutable.
You are now articulating a distinction between a social contract violating hard fork and a "technicality upgrade" hardfork.
Either way you want to argue it, the reason that immutability is not guaranteed is because of the ability to hardfork.
So your real issue, stems around the fact that blockchains can be hardforked, when broken down to its simplest form.
If a immutable blockchain has the ability to be hardforked at all, it can be made mutable at any time.

Your "tautological" argument only applies when mutability occurs with the chain.
All blockchains, by their nature, are quasi-immutable systems, that is being upheld by their communities.
If the community no longer maintains the immutability aspect, then yes, their blockchain isn't serving the original intent.

No. You have mixed up apples and oranges. You have taken two separate systems with totally separate rules
and laws, devised as such intentionally, and have combined them incorrectly to form the basis of your argument.
You are incorrectly combining internal blockchain consensus, with external forking consensus.
Game theory consensus can not apply to malicious external forking consensus, such as what ETH did.

These are, to me, two aspects of the SAME concept of immutability, indeed.  After all, a hard fork, or the "internal block chain consensus", are almost identical processes.

No, they are specifically different.
One is secured by the blockchain's game theory and the other is only externally secured by the community.
ETH's immutability was mutated by this external hardfork consensus, and not by the blockchain's internal game theory's failure.

After all, the "internal consensus" is agreeing which blocks are valid blocks.  This agreement depends just as well on the rules one wants to apply to decide whether they are valid in the first place, as to the application of those rules to the chain and the blocks themselves.   If everybody were to write their own node software, and implement their own vision on their own interpretation of the rules they want to see, we would immediately recognize that the internal and external forking consensus is in fact the same thing.  It is only because we all just use the "dev version" of the software that it seems that the internal consensus is what happens when we run the dev version, and "external consensus" is when we decide to download and install or not, a new version.

No, what you are saying is wrong. All nodes can either follow the rules or not. There can be different software from different devs,
but if that software has rules that violate the current protocol, then that violating node is rejected from the network.
The node consensus network makes sure everyone is playing by the same rules, software origin is irrelevant.
This is the internal blockchain consensus and maintains the immutability in the blockchain (game theory).

When majority of nodes agree to all collectively violate the current protocol, causing a split and new chain,
that is a hardfork and is an external blockchain consensus that may disregard immutability in the blockchain.
This is what Ethereum did and is why I argue the blockchain did not fail and the community did, since this is
external consensus, which does not use the game theory as envisioned like with internal consensus.

In theory, if a malicious actor started up 50,000 malicious nodes, they could fork the chain as a majority consensus.
This type of action is sperate from the internal "game theory" blockchain consensus which Satoshi envisioned.

Suppose that different miners publish different blocks with different transactions in them.  If I build (or accept) one block over another, I'm applying "internal consensus" according to certain rules ; but according to whether I change these rules or not, I'm applying "external consensus".  If you think about it, this is in fact the same principle.  The best way to see that is if I wouldn't use software, but if I would "check the blocks I prefer" by hand.    Then my preferences, and my way of selecting (and constructing) blocks is MY way, and if I find consensus over that, there is no clear distinction any more between "internal" and "external". (of course this is practically not possible because of the computational load it represents, but as gedanken experiment, you see the point I hope).

No, if a miner creates a violating block, it is rejected automatically by majority of nodes and the miner gets no reward.
Game theory says he wasted his time and money, so he will not do such a thing under normal circumstances.
If you personally accept that miner's violating block, you are no longer part of the Bitcoin internal consensus.
You are now using a violating internal consensus that only you and the miner are now following.
So, you are no longer on the Bitcoin network and now on a separate chain & coin that is not
interchangeable with the majority node Bitcoin network. You essentially hardforked yourself.

Now, if you are conforming to the internal consensus, like all the other nodes, and then by agreement, a majority of all  
nodes agree to apply "their external consensus" to cause a change with the internal consensus rules, then the majority
basically hardforked themselves from the original Bitcoin blockchain and created a new chain & coin.
This new chain will have a new internal consensus that was founded and instituted on the majority external consensus.
So, back to the original argument, the blockchain never did, nor can stop the collusion of the community to hardfork.
The "external consensus" is not bound by the blockchain system nor game theory.

Internal consensus and external consensus are entirely separate and serve two separate purposes.
Internal is designed to maintain and secure. External is designed to fork and change.
They clearly exist for different purposes and thus are different consensus mechanisms.