What happens if someone wishes to erase a preprint from EarthArXiv ?

We are working on making this process more clear and consistent. While there is a way for a user to remove their preprint from public view, we do not highlight this feature. This process should also be different for services that actively moderate submissions so that once accepted, the author would need to contact Eartharvix and request the preprint be subsequently rejected. In either case, we are working to improve the page where the preprint existed to display appropriate information that signals the preprint was removed and perhaps a reason why. If you have suggestions/preferences, please send those our way as we are actively defining the use cases here.

Good question @sgirardclos, and thanks for the reponse @mattspitzer! I agree that this needs to be made absolutely clear. I think it would suffice to simply say the preprint had been removed, with perhaps there being a 'reason' stated (N.B. when removing the preprint the author should perhaps be forced to choose a reason from a drop-down list). Might that work? Or maybe, on second thoughts, people shouldn't have to give a reason at all. We're not a journal, so no formal retraction is perhaps needed...?

For the integrity of the scholarly record it is best if papers can not be removed. Removal from public view (by moderator/admin) would be something for extreme rare cases and at least there should always be information on the landing page of the article record. It is for this reason that archives work with irrevocable non-exclusive licenses (e.g. https://arxiv.org/help/license) and warn authors before uploading that this is going to be permanent. A danger of making it relatively easy to withdraw paper from public view or have them removed entirely is that people will take EarthArXiv less serious and people will have more hesitations citing EarthArXiv material. I am still in doubt as to the importance of stating reasons in those rare cases of extracting things from public view. For reasons of transparency it would be good, but perhaps it would suffice to have just a very limited number of broad categories ("court order", "non-compliance with ethical standards", "plagiarism" might be examples of those). I have no first hand experience in these matters and welcome more expert views...

This is a very good point @jeroenbosman, and you've made me change my mind. Indeed, it shouldn't be so easy to simply remove material, especially if people have already cited it. One of the perils of sitting somewhere between, as the critics say, a mere blog and a journal.

Agree with Jeroen. A preprint gets a doi, can be cited and should thus be as permanent as possible. Would prefer to only remove preprints if forced by a judge to do so.

Is there already a case of a deleted preprint? Would be nice to see how that looks to see whether that works for us.

I agree with @jeroenbosman and @victorvenema. It should be like in journals: published is published! Removal should be exceptional and linked to a judgment of some sort.
To @christopherjackson3 : this is an interesting and important point you could discuss with the representatives of the 'more experienced' ArXiv platforms in NY. They might have ready-to-use policy guidelines we could recycle.

I totally agree with the arguments given by Jeroen. In ArXiv, articles can be withdrawn, but content will still be accessible. This is copied from their website:

"Articles that have been announced and made public cannot be completely removed. However, you may submit a withdrawal notification for your article. To do this select the Withdraw icon (withdraw icon) for the appropriate aritcle in the "Articles You Own" section of your user page.

You must provide a specific reason for the withdrawal within the Comments field. Please do not modify the abstract field unless the comments field is inadequate for your explanation.

Note that arXiv makes all previous versions of submissions publicly available (since October 1997). Therefore, even though the current version of a paper may be marked as withdrawn, previous versions can still be retrieved. See the discussion on availability of previous versions."

Jeroen Bosman said :

• a very limited number of broad categories ("court order",
• "non-compliance with ethical standards",
• "plagiarism" might be examples of those

All of which suggests the question of "which court", "which organisation's ethical standards", and ... well probably the plagiarism is sufficiently well defined with GeolSoc's, AGU, etc having policies, and also most (all?) educational institutions having a plagiarism policy.
Which country's laws govern the actions of EarthArXiv ?

Hi All. Interesting discussion. I think it would be wise to follow ArXiv's lead on this, especially the bit about declaring exactly why material was withdrawn. Anyone then finding the material can then assess whether they wish to trust/us it. Now, to @aidankarley's question: I'm not entirely sure! We're global, although COS is a US-based entity. In any case, most learned/professional societies likely have very overlapping ethical standards.