Ceci n’est pas une gouvernance
Opinions on why signaling for Ethereum proposals should not be formalized and the dangerous consequences if it is
Ethereum’s loose governance model is often the subject of debate, and over the past few months, discussions of EIP-999 have brought increased attention to it. There are currently discussions of how to standardize the non-binding signals in the governance process, and how they can be further legitimized — potentially creating a more binding governance model. It seems to be the belief of many that formalizing and legitimizing these signals can help make contentious decisions.
Ethereum’s governance model currently relies on being able to consider a series of non-binding signals that may change dynamically depending on what is being determined. Having these informal signals gives those making decisions a diverse overview of the community sentiment which is required to make potentially contentious decisions. When we standardize our signals and how they should be acknowledged, we step closer to turning our non-binding flags into binding ones.
Given the diverse set of signals one can listen to, we run into the issue of not knowing which signals to listen to. Often, what is “noise” depends on the perspective of the person commenting. In many cases, an individual’s idea of noise stems from whether the views agree with their political opinions, and the innate tendency to wear blinders to deeply held political beliefs.
In an attempt to organize some of these signals and give a better overview of the entire community’s sentiment, Aragon has launched a Nest grant in an attempt to find better signal reporting solutions. A lot of these initiatives were born from how political EIP-999 and general discussions on fund recovery have gotten. This is because the non-binding signals become difficult to listen to when discussing contentious EIPs, and a formal decision may never actually be made. Is this a problem though? The fact that nothing has been done about EIP-999 is inherently a decision of the community itself. The contentiousness of the EIP has been recognized, and therefore no one has moved forward with its implementation. The broader community has recognized that it is not valuable to potentially fork the network over this single EIP.
Discussions allow for the circulation of context and reasoning: disagreement or agreement, factors critical in compiling the sentiment of a community. Reporting tools largely ignore or filter this, under the misconceived notion that it is “noise”. It can be argued for example, that discussions found on various platforms are often more insightful than the votes themselves are. In a loose governance system, votes are the start of a conversation, not the end of one.
Reporting can help drive discussion, but when this reporting is formalized into the governance process and taken as a binding result it over simplifies the entire governance process that has protected Ethereum thus far. I say protected, because Ethereum’s loose governance currently makes it hard to have changes accepted, making bogus changes even harder.
Peter Mauric, stated in a post on the Ethereum Magicians forum, that EIPs should be developed to integrate signalling into the technical governance process and core dev decision making. This form of explicit acknowledgment of signals defeats the entire purpose of the non-binding signals currently found in the Ethereum governance process. The second stakeholders are told which signals they must listen to, we have broken our governance process.
Loose governance is hard, but we shouldn’t destroy it because of that. The second we formalize a system, people will game it. No system is immune to gaming, so formalizing a system often leads directly to its edge-cases being exploited. Ethereum’s governance is hard. We can gather useful metrics, but should not make governance simple or formalised — doing so risks allowing it to be gamed.