There is a lot to unpack here, before getting into any specific answers I think its worth noting that the whitepaper is a work in progress document that we expect to evolve as we progress from research to implementation.
How can jurors know if reputation or an “account” within a court is sold? (RE: End of https://github.com/aragon/whitepaper#31-collateral-and-reputation )
Reputation is non-transferrable so it is generally difficult (though not impossible) for it to be sold. There are hypothetical situations where the reputation could be sold, either because the account the reputation belongs to is a DAO itself and the DAOs governance tokens are transferrable, or perhaps it could be sold if access to the account’s private keys could be transferred. However, these approaches are probably not particularly practical. The benefit of using non-transferrable reputation is that due to its illiquidity it is more difficult for someone to quickly acquire a large enough portion to have a significant effect on the outcome. Jurors (and anyone else) won’t necessarily be aware if reputation is sold, but in the event that there is a malicious actor that somehow manages to gain a significant portion of reputation then network can respond by forking out the malicious actors and be confident that it will be expensive in both time and money for the attacker to repeat the process.
How can a bribery attempt be proved or detected by the Review Count Jurors? Other than by observing the juror is voting “in such a way that is not coherent with the jurors code of conduct”?
The intention with the code of conduct and review process is to minimize the impact of obvious bribery attack vectors. The most significant of which is probably one of the parties soliciting a bribe when they submit evidence to support their claim (because this is a direct channel to the jurors and guarantees that the jurors will all see the bribe). Such a bribe presents a coordination issue for jurors who are trying to maximize their rewards by predicting the behavior of the other jurors, however, by prescribing a specific course of action with penalties for non-compliance it makes the coordination game for the jurors easier. So while it may not be possible to detect all possible forms of bribery, it should help mitigage some of the most concerning bribery strategies.
4 - In the case of a 55%-45% jury voting, would the 45% loose reputation in a case where there seems to not be a clear vote/side?
Currently it is assumed that a pure majority is required to reach a decision and if that decision is accepted by the parties (not appealed) then the people who did vote for that outcome would loose reputation. Its certainly possible to think about different variation on these parameters. We are looking into analyzing the process and parameters using Agent Based Modeling. I’ve had some preliminary conversations with Piotr from Incentivai specifically about this. 
5.A - “When” do you expect that low volatility will be achieved by ANT? What circumstances must happen in the market?
The low volatility aspects of ANT is probably furthest out. Beyond the technical implementation, its important that there is a reliable source of revenue for the network. One source could be fees generated by the court, another source of revenue could come from deploying some of the networks capital in something like a Staking Pool. By establishing a reliable source of revenue first, we make the stability mechanism significantly more robust because the network would be able to issue credible external bonds backed by future revenue. There is another thread that talks about the importance of revenue for a stability mechanism here.
5.B - How can Aragon respond to an accusation or any sort of manipulation due to programmatically managing “both the supply of ANT and a stability reserve based on a price target”? Could this become a “legal” issue? (Stability Reserve plays a part on this).
I don’t see any reason this would be a legal issue (particularly if it is programmatically and transparently managed), but I am not a lawyer.
