In SCRF’s first community gathering for this season the topic “what is decentralization” came up which sparked a very nuanced and insightful conversation on the limits and pitfalls of the usage of the term. People seem to be naturally interested in this high-level sort of discussion and have a lot to say from various different perspectives. What is more, I think many can relate to the feeling that these are conversations that we normally don’t have in web3 but should have.
For this reason, I think it might be a good idea to start a research side-project that is a deep-dive into the different faces of decentralization, the historical roots of the term, its diversity of meanings in various scientific disciplines and its problematic usage in web3. A very brief preliminary foray into the territory reveals that the term is very multifaceted with a presence in urban planning, anthropology, systems biology and political theory way before it had started being used in finance. But it is only recently that the term has broken the surface tension of academia and reached the mainstream.
As of now, there is no scholarly field called “decentralization studies” or a good foundational inquiry into decentralization and I think it would be very nice if we made some first steps towards that direction.
This can become a cool discussion, a passion project, an SCRF initiative or it can be directed to external sources of funding such as Metagov’s grants for research proposals if people deem that the project has value.
I think @eleventh and @chrisbates have already voiced their interest in something like this and it would be a good chance to engage in more critical discourse about the state of web3.
Thanks for joining the community call yesterday! I’m glad you enjoyed the conversation. It was a fun one to start the year with!
We have actually discussed the idea of what a research project for decentralization could look like and @Sami_B put together a notable works on the topic. He also worked on a research prospectus on the idea and we’re going to catch up about that in a few weeks. We can chat in a future community call (or set up a breakout) to dig into the idea more.
One thing to figure out is whether it makes more sense for us to support something like this or to run with it as a SCRF project. If we do it ourselves, what does review/oversight look like. We’re generally starting to explore what peer review means in a decentralized research context and don’t want to rush into anything before we dig into that question more.
I definitely am in line with this thinking concerning the process. Sami’s research was more concerning the threats to decentralization, which necessitated him initiating research into the landscape of decentralization. I think this gives us an opportunity to either explicitly deal with only a few types of decentralization on the forum while SCRF decides how to approach “decentralization” as its own subject. On the one hand, it may not be SCRF’s responsibility to explore “all” of decentralization. On the other hand, this may be one of the most appropriate places for that type of research to take place. I just don’t want to get three years into SCRF’s existence and have all the forum posts not share a definition of decentralization. That is a strong possibility if the discussion about “how” we are defining decentralization does not take place.
It would feel intellectually dishonest to be years into SCRF’s existence to have a significant amount of posts about “decentralized” anything, and still not have had the discussion about “which parts are decentralized” and “what does decentralized mean in that context”?
I only frame it this way because almost all of the projects I’ve ever seen claim to be a “DAO” or “Decentralized” had many aspects of their organization that were clearly centralized and it’s debatable whether they achieved decentralization at all. If having a token on a decentralized platform is enough to qualify as “decentralized,” then does the term really have any meaning at that point?
I think you’re bringing up some great topics for us to delve into during the community calls. Also if you, or anyone else, is interested in setting up a community event (or series) such as a reading group on the topic where we delve into different papers and write up a summary of the discussion on the forum, we should talk about that. Even if we don’t do a formal research project, there are other ways that some of these questions can be explored.
We’ll most likely bring up the question of what decentralized peer review looks like in one of our coming community calls too.
Also, there might be some research DAOs that might want to fund research projects along those lines.
Yeap, I definitely resonate with that sentiment. Abstracted notions of decentralization can be very misleading. An other important question is “what purposes does decentralizing this specific part serve?” and “would it be better of it is wasn’t?”.
I’ve noticed that too. I mean, decentralization is supposed to get rid of the managerial levels in organizations but supervision/management roles are becoming more and more important in well-established DAOs. What gives?
@Larry_Bates how do you see this project going forward? Would you be interested too in setting up a sort of reading group about these issues as @eleventh proposed?
As others have pointed out, there is no reason to retread what @Sami_B has done, so maybe the initiative will need to do an examination of the notable works that Sami identified to then have discussions about how to move forward from there. I know Sami’s interest was more in threats to decentralization, so this aspect of “the nature and context of decentralization” as a continuing line of examination seems to be a subject that needs to be examined further for the sake of getting some clarity before decentralization becomes vaporware.
I think @Sami_B’s post was about measures of decentralization, not about the thing itself, as @Hermes_Corp is suggesting. I think it’s a shortcoming. If you don’t have a clear concept of the decentralization you want to measure, then you can pretty much pick and choose the metric that is most convenient for you, and make up a post hoc justification for your choice. Furthermore, the relationship between decentralization and its benefits (greater egality, trust, security, …) isn’t always super clear—clarifying that might tell us where centralization is a worry, and where it’s not, and therefore which metrics are relevant and for what.
I agree that we can’t pursue decentralization for itself—indeed, it would lead us astray, as there’s many ways to understand it. But if we don’t pursue it at all, we’ll get confused, and we’ll lack the means to determine which tools work for which concept.
Maybe, to avoid Larry’s worry, we could make it a more focused review. Like, focus on decentralization in the context of transaction security, or collusion, or democratization of the financial system.
Maybe a good first step you’d be something akin to a bibliometric analysis of where the word decentralization is found and how frequently. Related to that, what would be very fruitful would be defining a history of the term so that we can figure out from where it was initially adopted and reappropriated. I am pretty sure that finance is not where this originates (I have come across literature in political science talking about decentralization that dates back to the 70s or 80s). So we have to see what its original formulation was and why this notion was imported to finance.
Interesting, we (RnDAO) have started a conceptual framing of DAOs through an SCRF grant (towards creating a framework on analysis for the “common paths towards becoming a DAO”).
Our research touches closely upon the discussion here, so at least I’d be happy to join such a discussion/reading group and cross-pollinate
I would argue that there is a fine definition for ‘decentralization’ but that you have to contextualize decentralization to create insight for this space (or any space). Decentralization of what?
Decentralization is just a concept. I’m not sure you could deep dive into anything, again, without contextualizing it. You have to frame it: Technology? Politics? Public Affairs? Social Structures? Management?
I’m interested in exploring this if you want more help . I agree that the usage is problematic in web3. It’s too entrenched in a political ideology that says nothing about the technology. I’m not sure you need a whole research study to break this down critically. I think there is value alone in critical analysis of political ideology in web3 sentiment and use of the word ‘decentralization’.
Exactly. This is the problem with people in web3 ranting about Decentralization without anchoring it to any context or reference whatsoever.
But I would also argue that to have a better grasp of the concept, a more generalized notion of what Decentralization is, that is more general and can apply to most context is also needed. For example, decentralization can manifest as an actual tangible process in all kinds of systems, not only social and technological but also cognitive, biological and physical. Slime mold for example has a very decentralized structure with no central controller, but is still able to engage in intelligent behaviour, being able to very efficiently solve complex optimization problems.
In my own opinion it is always a good strategy to look at the simplest of systems first (physico-chemical, biological) and then try to apply the observed phenomena to ones in higher levels of complexity (socio-cultural, cyber-physical etc). The reasonig is that patterns exhibited in simpler systems are also exhibited in more complex ones in a fractal fashion. See for example the HKB model where the same dynamics and phase transitions can appear in simple finger-wagging experiments can appear in simple mechanical oscillations, brain neurodynamics or the motion of crowds in cities.
And beyond mathematically rigorous definitions, a philosophical investigation of decentralization is also incredibly important because philosophy can add a normative dimension to all these cases. Beyond being able to describe when a system or organization is decentralized and to what degree, we should, more importantly, be able to clearly define the value system according to which the object of study or observed behavior can be assessed and how to act.
This can help to decouple the meaning of the term as currently used by technocratically-minded web3 folks from its unexamined ideological underpinnings and bring awareness to the biases and presumptions involved. Vitalik actually took a very good first step with his “The Meaning of Decentralization” article, but there is much more complexity and nuance involved in the philosophical inquiry into decentralization that ought to be uncovered. As you very rightly mention @valeriespina, there is value in the critical analysis of political ideology in web3 but critical analysis can only show what is already happening. We also need an ethical framework to show what ought to be done!
One of the issues becomes the definition of “decentralization” is not consistent between the computer science space and everything else that is non computer science’s definition of decentralization. I will demonstrate.
The image used in the definition of “decentralization” on Wikipedia that is not the computer science definition is as follows:
The issue then becomes that a “distributed network” by the definition of “decentralization” would be MORE decentralized than “decentralized,” if the basic definition/interpretation of “decentralized” is
“the process by which the activities of an organization, particularly those regarding planning and decision making, are distributed or delegated away from a central, authoritative location or group”
This issue is always raised not to undermine the notion of decentralization but to point out that “decentralization” as a finite point creates its own paradoxes when you factor in “distributed” networks. This is the main reason I have continued to push for more specific “types” of decentralization within conversations that are revolving around the notion of decentralized frameworks.
