The governator manifesto

A spectre is haunting cosmos — the spectre of governators. All the powers of the hub have entered into a holy alliance to exorcise this spectre: Pupm0s and SG1, Cosmostation and Nostradamus, Kitteh’s Radical Osmonauts and Jovian police-spies.

Where is the party in opposition that has not been decried as gocernistic by its opponents in power? Where is the opposition that has not hurled back the branding reproach of governators, against the more advanced opposition delegators, as well as against its reactionary validators?

I couldn’t resist, plus we finally have a forum worthy of creative work so there we go

the governator concept

As I understand it, there has been talk of creating a system that allows for a novel type of political representative in cosmos. In Cosmos there are two types of vote power, consensus vote power and governance vote power. Governatists are suggesting that we separate the two of them.

The idea is laid out here by Jim Yang, And I think that he’s among the people who has thought most about this concept.

Personally, I don’t feel that he’s got it quite right yet but I don’t think that he thinks he’s got it yet either.

Because of my arrogance and experience, I think I might have it approximately correct, and so I’m going to lay out a separation of powers that can also overlap so as not to infringe upon the rights of validators, who are in fact simply account addresses on the chain, anyone can be a validator.

IRL, separation of powers is most easily observed in The design of the United States federal government, which features a legislative branch that makes laws, an executive branch that sets direction, and a judicial branch that enforces law.

For the time being, and to produce a workable prototype, we will skip the judicial branch, but there have obviously been times when a judicial branch would be useful, please place a mental bookmark here.

proposal: a separation of powers that can blend.

Delegators determine both consensus vote power and governance vote power. It makes good sense that they may wish to choose a different political representative than the people operating their infrastructure.

There are most certainly infrastructure operators who may wish to just operate infrastructure, and not have the obligation of dealing with political matters, or even to not have blame placed upon them, for not dealing with political matters.

However, there are also times when the involvement of infrastructure operators is of paramount importance, like software upgrade proposals and client update proposals.

And now, the separation is revealed.

  • Governators should have governance vote power in all matters that do not pertain to infrastructure operation.

    • Grant funding
    • Matters of direction and procedure
    • Incentive balancing
    • Incentivization
    • Partnerships with other chains
    • New initiatives that do not relate directly to infrastructure operation
  • Consensus vote power should remain full (governators have no say here) by the governators on the following matters:

    • Software upgrade proposals
    • Client update proposals
    • Other proposals which may not yet have been introduced, that deal strictly with infrastructure operation
    • Proposals to slash validators
    • Parameter changes that would affect the validators set, like the adjustment of minimum commissions, The downtime window, the downtime slash and the equivocation slash

the blend

In the system, validators can be governators. Governators can also be validators. Because the cosmos has had thousands of completely electoral fraud free Mass scale governance decisions, and because validators and governators are not necessarily individuals, but could be groups of individuals, there is no need to have separation between governators and validators.

With that said, validators should be able to choose not to be governators, if that is what they desire, and governators should be able to choose not to be validators, if that is what they desire. In the present state, there are some validators who are governators and validators, and governance is an expected activity of all validators. With this in mind, I wish to move to define the duties of validators and governators.

how this affects validators


  • Validators will have a clearly defined technical role that is focused on network security.

  • Validators and their delegators can be slashed for non participation in software upgrade governance proposals and client update governance proposals.

  • The abstain option can be removed,

  • veto must be retained because it mirrors consensus.

  • Non-technical validators who use VAAS can stop validating, and just governate or just stop.

how this affects governators

  • Governators and their delegators can be slashed and and governators can be tombstoned for non participation in governance.

  • The abstain option can be removed.

  • Veto is not needed here because these changes should not affect consensus.

  • There can be an unlimited number of governators because governation does not require running a node or participating in consensus.

  • The governator role will not require an understanding of blockchain operations, software engineering, network health and security and things like that although those are probably helpful traits, they are no longer a hard requirement.


These will be baked into the CLI once the governator software is completed so that anyone making a create validator transaction or a create governator transaction must declare one or both oaths.

The validators oath

I, an osmosis validator, swear to uphold the integrity of the state of the osmosis blockchain and to participate in the approval or disapproval of software upgrades and client updates. I will check the code diligently and I will at all times fiercely predict the integrity of osmosis. I will step aside if I am unable to fulfill these duties.

The governators oath

I, an osmosis governator, swear to use my skills to advance the interests of the osmosis blockchain and its community at all times. When making decisions that concern osmosis, osmosis shall be paramount in my eyes. I will step aside if I am unable to fulfill these duties.


Now that we have defined roles, and we have looked into the reality that there are certain situations where it is extremely important for validators to represent delegators, and we have dealt with the reality that on the osmosis blockchain, every account address should be able to perform every possible role, we can see that validators can be governators, governators can be validators, and that some validators and governators may choose to opt out of one role.


This seems like a fun concept to chat about. Stride technivally will be the first chain to have governators once their ICS integration goes live, so this is possible, but there’s a big problem that most non-ics chains will face if trying to transition to this model.

One area that was discussed in the thread you linked but not in this post is how governator (we need a new name for this too) incentivization would be handled.

IMO, there are a few different ways this could be handled, and none of them are ideal:

  1. Have governors take a share of the rewards from distribution that are intended for validators.

I find it extremely difficult to believe that validators would vote for this cut in their own revenues, especially given many of them operate at a loss already. This is effectively what will happen with Stride, but in this case Hub validators have the additional incentive of ATOM emissions and other ICS rewards. An opt-out mechanism may be the answer here, but why would a validator not opt-out in this instance and maintain the status quo (and their rewards)?

  1. Enable an additional incentive (either from the community pool, increased emissions, or somewhere else).

Equally problemmatic. First, this would have the effect of passing the cost of maintenance of governators onto users (via dilution and / or paying 2 commissions).

Additionally, it doesnt solve for the problem of separating consensus voting from governance, as why wouldn’t a validator also elect to be a governor and claim the extra rewards, which they could then use to market themselves as filling both roles (to the detriment of governors that don’t validate the chain). To fix this, there needs to be some participation mechanism (and a quality participation at that) to measure governor success by. What does that look like? It seems pretty difficult to codify, let alone enshrine in the chain’s code.

1 Like

If we look at the osmosis validators set, we can see numerous validators, including very well- ranked validators who choose not to participate in governance.

If they were to also choose to be governators, their delegators would be slashed because they have made a poor choice in governator, same as how we delegators to validators who equivocate.

One of the key points of this manifesto is that this should never happen under any circumstances. Most validators are underwater already and osmosis should reduce the size of its validator set to 69.

I had a second realization here, on osmosis, there is a flow of tokens directed toward the team, and I have a feeling that since so many of them want this, that it is very likely that they would accept some dilution to their flow of inflation in order to spawn the governators, meaning that this would not affect the community pool, liquidity rewards, or incentives to delegators.

Right, but what does it mean to be a bad governator?

Not voting? Can’t a governator just vote without reading or really participating? Hell, can’t a governator spin up a script to auto-vote when a new prop goes live?

Seems clear to me there needs to be an enforceable mechanism for substantive participation above and beyond simply submitting a vote, and that’s where things get tricky.

Otherwise being a governator is simply another way to farm tokens.

1 Like

That is not for us to decide it is for the delegators to decide and only for the delegators to decide and one of the most crazy freak wild ideas I’ve ever heard was from Jim when he said well you know we could do is we could slash the governators if their votes were bad yeah that’s not a good idea, because the delegators chose those validators.

So a good governator, is an active one. I’m thinking about even removing the abstain vote.

By the way, validators should be slashed if they do not participate in software upgrade votes and client update votes, wtf that’s critical security stuff, if ya ain’t voting you’re for sure not reviewing.

How about we remove the abstain option, require a deposit, slash for non participation, and allow full mobility to vote power?

Yes, and their delegators will redelegate.

My first thought is that I love this idea. Get’s me excited thinking about it. But then I come to few realizations about this.

First, having governators is a great idea. The ability to separate consensus from voting power is a direction that chains should go in. There’s some details that should be worked out, but I’m all for the idea. Jacob said this probably isn’t possible because of the current sdk limitations, but instead of separating who votes on what, creating a dual-house system could help bring balance with new and impactful perspectives. Establishing what makes a good governator and the requirements won’t be perfect to start, but they should be expected to be adapted to optimize the quality of governators without creating unreasonable barriers for the average member who may want to get involved. This also may just become a popularity contest and create more advantages for special interest groups, so election criteria and long term governantor rules should be thoroughly discussed. I also don’t think it makes any sense to divide what each type of represenative would vote on. In separation of powers, there are also checks and balances. If anything, the validators should essentially be seen as the executive branch enforcing the rules, and the governantors deliberating and voting on those ideas. The executive branch ultimately has the sole ability to enforce or not enforce these votes/rules, but the sole power to decide the rule of the land is with the legislature. So this is more along the lines of creating a dual-house body to vote on rule changes, but neither group should be excluded from specific votes if not excluded from them all. In your examples there is actually no checks and balances between the bodies especially with the suggestion of not allowing governators to vote on slashing.

That being said, I don’t think this structure is optimal for any chain that isn’t very well established and the founding team has propelled the chain to the point where more decentralized governance is beneficial. In the early days, widespread governance is unfortunately suboptimal if there is not a clear mandate and vision being followed. I still think Osmosis should wait a little longer before doing this.

I would also be hesitant to enact this if all US peoples would essentially be ineligible due to regulations. A large base of users would technically be void of the ability to have proper representation. DAOs can still be at odds with the SEC and face legal action. I know some will say sucks for US persons or just move, but let’s all be realistic that that isn’t possible for most people.

If this were to be implemented, here are a few points to think about:

-Are validators incentivized or disincentivized to do this? Will that cause an issue trying to make this change?

-How do we have clear guidelines that allow governators to act freely, but still be held accountable by recalls/removals etc?

-True separation of consensus power and voting power does not allow for consensus to vote on rules. Are we looking to separate powers, or create more accountability/balance with governance voting?

-Does this give certain validators more power who can gather consensus and voting power, defeating the purpose of this structure?

-Are there any other simple changes we can make with this? Voting/consensus power caps, changing outcome of a majority abstain?

-At what point is this optimal for a chain to adopt?

-Why would governators have voting options removed for them when validators don’t?

-Does delegation to soley a governator remove that stake from participating in consensus? Can delegations be made two fold: consensus power & voting power? (can overlap or be unique)

Writing this caused me to reflect upon the separation of powers in the United States government, and I observed that the executive seems to set direction more than it does enforce law, whereas, the judicial branch seems to enforce law.

Oh let me just go to Wikipedia here one second

Validators really just need to enforce one thing: consensus.

How does 6 months work for you?

Legislature writes the rules

Executive branch enforces the rules

Judicial branch litigates disputes and interprets laws in regards to constitution.

I am from a country with free association and free speech, the United States of America and they can get my vote power for a private keys only by prying them from my cold dead hands.

Or you’re ledger >_< jk

You can’t claw that from my hands, I keep it in my ass.

1 Like

This would be most similar to the executive branch. Consensus is code and code is law.

I don’t think a static timeframe should be applied, moreso the state of the chain itself. Osmosis probably has the best and most stable governance right now for a reason…

Sure thing, but the thing is that when a software upgrade is run, or a client update is made, the validators will without a doubt occupy the roles of all three branches simultaneously

In the perfect world, this is where code reviewers (the judicial branch) come into play. But you are 100% right there. That is why it is important to implement this very carefully and calculated so the original intention can be accomplished.

In a perfect world, validators review code, welcome to a perfect world.

Interesting topic and it might be an interesting path to investigate and to finetune.

Having separation of powers would indeed reduce the need for large validator sets, because you would only need real consensus driven parties validating the chain, whereas other services can be rewarded via different routes. As @RoboMcGobo rightfully said, it has to be thought of how this should look like, to ensure every party in the chain has the ability to earn a decent income matching their service delivered.

How will this look like in terms of delegations? Will people have the ability to delegate their OSMO twice? Once towards a validator on-chain for the consensus part and once for the governance part? Because that is what we will be looking at; in the end competition in terms of delegations between the consensus and governance part should be avoided, because it concerns two totally different topics.

One additional risk is that with reduced val-sets we will centralize the Cosmoverse amongst a small amount of big players. This can result in a stable environment tech-wise, because these parties can invest in proper infrastructure, monitoring tools and well trained personnel. But it will also have the risk of killing innovation in that area, because competition will be less. We have always portrayed ourselves as different than Polkadot for example, so having a method where we might slim down our valsets, while still preserving diversity is an interesting subject.

That really isn’t how payments coming from inflation work. People get what they get based on their level of delegations and I’m hoping to make the governator compensation work the same way.

A governator’s level of contribution is determined by delegators perception of the quality of their work, not the quantity or type of their work.

Totally true!

Note that I also didn’t specify quality or quantity, but I agree on the qualitative aspect. Rather see good contributions than a lot of contributions. Good point.

One other thing; I think we should replace “OPT-OUT” with “OPT-IN”. That way only active consensus validators will be triggered to be also governators. Passive validators who have just set up their tech and ignore the rest will then not be included in the governator part.