You seem to have asked your actual question in a comment, so I'm going to answer that:
Thank you for your response. Actually my plan is to publish the language as an open-source project. But your last point is exactly the thing that I would like to prevent: that the language gets different clowns, developed in directions I do not agree. Is there any sort of open-source license which prevents others from "stealing" the language from me?
You have expressed two conflicting desires:
[I would like] to publish the language as an open-source project.
vs
I would like to prevent: that the language gets different [clones], developed in directions I do not agree [with; ...] others from "stealing" the language from me
The whole point of open source is to allow people to "steal" someone else's work - but the intention is for people to "steal in a good way", i.e. copy, modify and distribute their own version, with attribution, to achieve something of general public benefit.
If you expressly want to retain "ownership", then depending on what exactly "ownership" means to you, you would be wanting to keep your language proprietary, and not publish it as open source.
However, you'd probably end up without any users that way: certainly, it was common for people to use proprietary languages back in 2000, but by now there's very few people who would choose to use a brand new proprietary language for anything since there's so many open source ones already available now.
So, I am going to suggest what seems to be the standard sensible route that is followed:
Publish the language implementation (interpreter, documentation, etc.) as open source, but at the same time, trademark the name of the language and keep this trademark well defended. Make clear in your README (or similar) that the name is trademarked and that if anyone clones the project they must do so under a different name.
This means that if anyone wants to distribute a modified version of the language, they can either send you a pull request and await your approval of the change (and then you, as upstream, get to release it yourself), or they can release it independently but in a way that makes it abundantly clear that their version isn't the "original, official, trusted version".
(Do not patent anything - it's widely accepted by now that in the world of software, this achieves no practical benefit, isn't even legally possible in many jurisdictions, and only encumbers your work and results in no one using it.)
As for what license to pick - if indeed you do decide to publish as open source - there are broadly two schools of thought, permissive (e.g. BSD, MIT, etc.) versus copyleft (typically the GPL family: LGPL, GPL, AGPL in order from least to most restrictive). I won't elaborate on the details here, but suffice to say that what you pick will depend on what matters most to you: making it easy for people to use your work, or preventing companies from exploiting your work without "giving anything back"? If your prime motivation is to make it easy for people to use your work, then you want a permissive license, but you then accept the risk that your work will be exploited; conversely, if you want to avoid your work being exploited, you want a restrictive license, but you then accept that people might decide against using your work in favor of a more permissive alternative.