You called property "weird" so I wonder whether you are aware that property is weird exactly because of the double spending problem. A land title registry gives you double spend protection, not just irrevocable publication. Your point of course stands, since you're trusting someone with enforcement, you might as well trust them with double spend protection. I'd say decentralized irrevocable publication (e.g. certificate transparency) is much cheaper than decentralized double spend protection.

In my nice first-world country I do trust the government to track it. But in many third-world countries, deed registers are prone to forgery and bribery, and it's not uncommon for people to discover they suddenly don't own their land. I've seen economists argue that this is a major factor limiting economic growth in those countries.

Probably some of those countries would like to fix this, and for them, a blockchain might make sense...say with multiple government entities cross-signing property transfers, rather than requiring all property owners to hold their own keys.

I think that the devil is in the details. Imagine this situation.

Someone comes, takes your home and says "this is mine!".

Now you have to go to your government and prove that the person stole from you because it is yours. Some governments are efficient at providing this proof, some aren't.

Some countries in Africa have a huge problem with this. Where ppl go to other people's houses and sell them even though it is not theirs. And because it takes up to two years for someone to prove the ownership, sure the government will enforce the ownership eventually but two years is a long time. I've seen pictures of houses where the owner sprayed "This house is not for sale" to avoid this issue.

Now imagine a system where proving ownership is fast, nearly free and very hard to forge. The incentive to cheat the system is much smaller than before. Using a blockchain to handle proof of ownership is a great way to have a secure and nearly free.

Sometimes it is not about being able or not but how cheap / expensive it is and how slow / fast the process is.

You don't need the government, you need the society at all. These things can be done in a lot of ways, all them compatible with a trusted ledger that keeps track unambiguously of who owns what.

Notary services were for a long time independent of the State, they may still be in some countries and places. This is indeed something in which a global decentralized trusted ledger can help.

There are all kinds of problems that unforgeable append-only public ledgers make less crap while not solving the fundamental problem that people are at the center of all people-related transactions. For instance I can bribe a notary to backdate a notarization. Can't bribe a blockchain. This is good. There are lots of things like this.