"...replicating a structure (through organizational or version forking) does not completely address the hierarchy of the structure itself. While it certainly gives others a voice and accounts for fragmentation in the community, power is not flattened among all participants. Quite clearly, senatorial governance is not a direct democracy but a representative one where certain actors (Lead Developers, mining pool operators, software companies) are raised into positions of power by grouping individuals (programmers, miners, users) through their obligatory passage points. This is not necessarily a negative revelation for algorithmic decentralization via (proof-of-work) blockchains but it is important not to assume all stakeholders of their protocols are made equal. The cost of collective action is hierarchy" (481)
Exactly. Bitcoin governance can still be called decentralized but there are still certain degrees of centralized control in decentralized systems (just look at the Internet). The political/organizational problem is hard to solve because humans have to coordinate to make decisions and often this is done from (centralized) points of authority (Lead Developers, Mining pools, Bitcoin wallets/exchanges). This is not a direct democracy but a represantative one because not everyone in the network is equal: 1) Core developers elect the Lead Developer to make decisions on protocol rule upgrades, 2) miners increase the likelihood of coin rewards by using mining pools and in the process elect mining pool operators to make voting decisions on their behalf, and 3) users use Bitcoin wallets to take on technical procedures associated with accessing bitcoins and in the process elect them to make lobbying decisions on their behalf (instead of running their own Bitcoin node). This creates a type of decentralised structure with centralised pieces. "The cost for collective action is [some form of] hierachy" (481).
That's an interesting way of putting it. Scaling will always be a problem with monolithic records like you say. I have worked on a couple of distributed ledgers with sharding and these governance models seem to be more promising. Often there is some sort of democratic voting system for stakeholders when it comes to making change to the code which can be reduced by certain cryptoecnomic rules (e.g. demurrage fees for hoarders if it uses a proof of stake consensus). You're right though, its the fact that software needs to be updated by humans so that it stays relevant and reflects stakeholder interest over time that demands decision making. And this decision making often tends to materialise in (some sort of) centralised form due to the need to regulate and promotes good decisions (i.e. by experts). The same happens through Wikipedia moderating.
Possibly. But you still need a Lead Developer to sign off on decision making. And high net worth individuals will have consolidated voting rights on those decisions. Perhaps if PoS is mixed with other development models?
Bitcoin decision making is channeled down a funnel: Core Developers make suggestions and the Lead Developer (and those given commit access) sign off on those decisions. Those decisions are then voted for by miners who are (relatively) centralised in that roughly 5 mining pool companies control the vast majority of hashing power used to vote on those decisions. Meanwhile large wallet/exchange companies who control vast amounts of on-chain transactions can lobby miners to pick certain decisions by upgrading their nodes to reflect new rules (miners will want to follow large companies because they create liquidity for coins with the new rules and so they can sell their coins more easily and, theoretically, for a higher value).
So while this is still a decentralised systems because multiple parties have a say, there are still lots of points of centralised control in the governance system. In other words, not everyone is equal. "Individual developers submit to those with commit access, individual miners submit to mining pool operators, and everyday users submit to Bitcoin companies" (478). The Lead Developer acts as a centralised decision maker, mining pools act like centralised voters, and Bitcoin companies act like centralised lobbyers. So there is a certain structure to Bitcoin governance.
Holders of bitcoin are the only people who don't have a say. Buyers, miners and developers decide what is and is not viable (in roughly that order). Holders have a say only as much as they are still buyers.
By creating the fork on demand. The only part of Bitcoin that is technically hard to recreate from scratch is the huge hashing power of its mining network.
If buyers were happy with bitcoin-but-a-different-brand then there isn't much existing bitcoin holders can do to hold their market together. There is an unlimited supply of numbers out there, the constraint is numbers that are backed by whatever silly number of hashes per second the Bitcoin network is up to. The holders don't have any particular influence over that constraint.
From the article abstract, it appears that this paper is more about how the governance behind modifying the bitcoin code itself is hierarchical, and not the separate but dependent issue of mining power for the current protocol being concentrated in relatively few hands:
"The overall political framework for altering the Bitcoin code is described as senatorial governance: a (de)centralized model of bureaucratic parties who compete to change the monetary policy (codified rules) of the protocol. This model shows how Bitcoin is not an autonomous system but is assembled and maintained via human discretion."
