Thursday, January 28, 2016

Does the arXiv censor submissions?

The arXiv is the physicsts' marketplace of ideas. In high energy physics and adjacent fields, almost all papers are submitted to the arXiv prior to journal submission. Developed by Paul Ginsparg in the early 1990s, this open-access pre-print repository has served the physics community for more than 20 years, and meanwhile extends also to adjacent fields like mathematics, economics, and biology. It fulfills an extremely important function by helping us to exchange ideas quickly and efficiently.

Over the years the originally free signup became more restricted. If you sign up for the arXiv now, you need to be "endorsed" by several people who are already signed up. It also became necessary to screen submissions to keep the quality level up. In hindsight, this isn't surprising: more people means more trouble. And sometimes, of course, things go wrong.

I have heard various stories about arXiv moderation gone wrong, mostly these are from students, and mostly it affects those who work in small research areas or those whose name is Garrett Lisi.

A few days ago, a story appeared online which quickly spread. Nicolas Gisin, an established Professor for Physics who works on quantum cryptography (among other things) relates the story of two of his students who ventured in a territory unfamiliar for him, black hole physics. They wrote a paper that appeared to him likely wrong but reasonable. It got rejected by the arxiv. The paper later got published by PLA (a respected journal that however does not focus on general relativity). More worrisome still, the students' next paper also got rejected by the arXiv, making it appear as if they were now blacklisted.

Now the paper that caused the offense is, haha, not on the arXiv, but I tracked it down. So let me just say that I think it's indeed wrong and it shouldn't have gotten published in a journal. They are basically trying to include the backreaction of the outgoing Hawking-radiation on the black hole. It's a thorny problem (the very problem this blog was named after) and the treatment in the paper doesn't make sense.

Hawking radiation is not produced at the black hole horizon. No, it is not. And tracking back the flux from infinity to the horizon is therefore is not correct. Besides this, the equation for the mass-loss that they use is a late-time approximation in a collapse situation. One can't use this approximation for a metric without collapse, and it certainly shouldn't be used down to the Planck mass. If you have a collapse-scenario, to get the backreaction right you would have to calculate the emission rate prior to horizon formation, time-dependently, and integrate over this.

Ok, so the paper is wrong. But should it have been rejected by the arXiv? I don't think so. The arxiv moderation can't and shouldn't replace peer review, it should just be a basic quality check, and the paper looks like a reasonable research project.

I asked a colleague who I know works as an arXiv moderator for comment. (S)he wants to stay anonymous but offers the following explanation:

I had not heard of the complaints/blog article, thanks for passing that information on...  
 The version of the article I saw was extremely naive and was very confused regarding coordinates and horizons in GR... I thought it was not “referee-able quality’’ — at least not in any competently run GR journal... (The hep-th moderator independently raised concerns...)  
 While it is now published at Physics Letters A, it is perhaps worth noting that the editorial board of Physics Letters A does *not* include anyone specializing in GR.
(S)he is correct of course. We haven't seen the paper that was originally submitted. It was very likely in considerably worse shape than the published version. Indeed, Gisin writes in his post that the paper was significantly revised during peer review. Taking this into account, the decision seems understandable to me.

The main problem I have with this episode is not that a paper got rejected which maybe shouldn't have been rejected -- because shit happens. Humans make mistakes, and let us be clear that the arXiv, underfunded as it is, relies on volunteers for the moderation. No, the main problem I have is the lack of transparency.

The arXiv is an essential resource for the physics community. We all put trust in a group of mostly anonymous moderators who do a rather thankless and yet vital job. I don't think the origin of the problem is with these people. I am sure they do the best they can. No, I think the origin of the problem is the lack of financial resources which must affect the possibility to employ administrative staff to oversee the operations. You get what you pay for.

I hope that this episode be a wake-up call to the community to put their financial support behind the arXiv, and to the arXiv to use this support to put into place a more transparent and better organized moderation procedure.

Note added: It was mentioned to me that the problem with the paper might be more elementary in that they're using wrong coordinates to begin with - it hadn't even occurred to me to check this. To tell you the truth, I am not really interested in figuring out exactly why the paper is wrong, it's besides the point. I just hope that whoever reviewed the paper for PLA now goes and sits in the corner for an hour with a paper bag over their head.


  1. Well it's all relative after all (my hopeless attempt at a pun). Being a "few times the radius of a black hole" away from a black hole is still "pretty close" to a black hole. Closer than I'd want to get anyway. But the really amazing thing is that the general public, thanks to Hawking, has any concept of Hawking radiation at all!

  2. Thanks for posting on this. I haven't personally had a problem posting to the arXiv, other than one mysterious two-day hold, but examples of rejected papers and blacklisted authors are thick on the ground. And while I agree that minimal quality standards are needed, censorship based on content that the moderators find distasteful is terrible -- both for the censored physicists and for the greater community.

  3. The great mathematician Leonhard Euler's first published paper:

    The (erroneous) construction of tautochronous curves in media with different forms of resistance. This is Euler's first paper E001; the math in the translation has been clarified (Feb. '07) to account for Euler's mistake.


    Apart from the obvious problem of crackpottery, nobody should be discouraged or stopped by mistakes.

  4. Garrett,

    Ja-ha, I can imagine them thinking when you submit a paper "Oh, no, it's the Lisi-person again, he'll crash our server" ;) In any case, I just mentioned you because I think it's strange that certain people seem to be getting an extra load of attention for no good reason, as far as I can tell.

  5. Dear Sabine,

    I am afraid, the problem is MUCH simpler : both the arxiv bosses and those who want to post on it WANT TO HAVE THEIR CAKE AND EAT IT, TOO ! :-) :-) :-)
    And as EVERY mature person should know it, that is NOT possible ... :-) :-) :-)
    Yes, both sides want to AVOID the usual LONG refereeing time at usual journals, and at the same time, both sides want to avoid arxiv becoming a "free for all" website, like say, its well known rival vixra, that is, a website where only a most minimal check of the submitted items is ever done.
    Now, the ONLY SOLUTION so far to the above IMPOSSIBILITY, and which was CLEARLY and EXPLICITLY included in arxiv, vixra, and ALL similar websites is that NONE of the papers ever accepted and posted on such websites can be withdrawn. Yes, the authors of papers already posted CAN post a CORRECTION, or can declare a paper no longer valid. But by posting a paper with a MISTAKE, well, that mistaken paper WILL FOR EVER remain on the website, and thus can be seen by everybody for eternity, whatever that may practically mean ... :-) :-) :-)
    And that is so far the ONLY way known to ... have one's cake, and eat it too ... :-) :-) :-)
    All other suggestions for improvements only show, I am afraid, one or another manifestation of feeble mindedness ... to put is politely ...

    Now of course, Dear Sabine, and other possible commentators, let me assure you that I do NOT AT ALL expect your agreement : all of you seem to be still so young, and then, in your typical ways, you DO STUBBORNLY hold to ... having your cake, and eat it, too ...
    Yes, in a way, I do wish that I myself were still so young ... :-) :-) :-)

    Meanwhile, try, and have a nice day ...
    And do not let some silly arxiv, vixra,, etc., to disturb you days ... :-) :-) :-)

  6. "No, the main problem I have is the lack of transparency."


    "I hope that this episode be a wake-up call to the community to put their financial support behind the arXiv, and to the arXiv to use this support to put into place a more transparent and better organized moderation procedure."

    If you get banned, do let us know.

    If I were a rich person interested in physics (Brian May, CBE, are you reading this?), I would double arXiv's current funding with the stipulation that a transparent appeals process be introduced. Another stipulation should be a white list of journals, publication in which automatically allows one to submit to arXiv.

    It's not just the lack of transparency. Even when one asks a specific question, it is not answered. OK, put it down to lack of resources, but surely writing a one-sentence email explaining why something has been put on hold, or rejected, is much less effort than the decision itself.

  7. Moved from another post at Sabine's suggestion:

    Noticed your tweet about arXiv problems. I tried to leave a comment on the blog there but it didn't work. (This is getting more and more common and people who oursource their blogs don't know how to fix it, and there reader doesn't know how to contact those who can.) So, I'll start a comment thread here, referring to that blog post and the comments there:

    Yes, this is a real problem but if you want to discuss this problem sensibly, do not even mention viXra or archivefreedom. Are you even familiar with these sites?

    viXra is essentially 100 per cent crackpot. It is certainly not an "alternative" to arXiv for serious researchers. In fact, I think that posting something there would, correctly, mark you as incompetent. Either the work is incompetent, which is true of most of the papers there, or you can't even notice that almost everything there is crackpot. What's the point? The point is not to get the paper available. To do that, just stick it on your own web server or whatever. The good thing about arXiv is that it is a one-stop shop. Either you are there, or you are nowhere. Even refereed-journal papers might be ignored if they are not at arXiv. (In part, this is because, in many cases, access is only via overpriced fees.) This is why arXiv has to be run correctly. I think it is fair to say that if you put a good paper at viXra, no serious scientist would notice it.

    archivefreedom goes on and on about Brian Josephson. Yes, he won the Nobel Prize. This does not, and should not, give him the right to publish crackpot stuff, on arXiv or anywhere else. Kudos to arXiv for treating him like any other crackpot.

    In other words, folks, don't make the "the enemy of my enemy is my friend" mistake.

    Another serious problem: People with experience like this will probably not speak out publicly, for fear of being banned forever, which in today's scientific world is worse than excommunication in the middle ages.


  8. Bee: First, could you please move this discussion to that comment section?

    Phil: Certainly, but I had to respond to the comment in the Maui post and am not planning a long discussion

    Bee: Second, for wha viXra is concerned, I think you don't understand the issue. No serious researcher will make the effort of scanning a repository in which the majority of submissions are low-quality, even on the risk of missing the one that is not nonsense. It's a bad investment of time.

    Phil: I do understand the issue perfectly well. Some people browse arXiv, a few even browse viXra but most find papers in other ways such as web searches. That was the point I made. Nobody is asking anyone to invest time looking for random papers on viXra if they don't want to. We still get 3000 paper downloads every day though.

    Bee: I don't have a particular problem with that website, I think it serves a function in a particular community and that's beneficial for everyone involved. But don't tell me they serve the same function because they simply don't.

    Phil: They do serve the same function. viXra is just filling the gap for those who are kept out of arXiv, but otherwise it is doing exactly the same thing.

    Bee: Your comparison to blogger doesn't work because nobody reads all blogs on blogger, there are too many. The question is exactly where do you get a good filter. Best,

    Phil: Nobody reads all the papers on arXiv or viXra either but that is not the point, The point is that both blogger and viXra are content hosts which do not limit content. If they dont then there is no reason to judge the quality of one item of content against another. If someone cites a paper on viXra when you are reading a paper that interest you by someone you trust, then there is no reason why you would ignore that paper. If you are doing a keyword search on google and a paper from viXra comes up and the context looks like what you are interested in, would you then ignore it?

    I'll add one more thing. The real reason most people get hot under collar about viXra is because they hate getting unsolicited e-mail from people about theories that don't interest them. For some reason they think viXra is part of that problem and is trying to force them to read its content, but ask yourself this question: Out of viXra and arXiv, which is the one discouraging its authors from sending spam e-mails and which is the one telling them to go find some academic person to endorse their paper before they can publish it?

  9. In response to Philip Helbig:

    So it is only "most of the papers" that are incompetent on viXra then? If any proportion of the papers are good then viXra is a success. The numbers of papers we get submitted to viXra each year has increased steadily by 25% and if everyone who has trouble posting on arXiv uses it we might soon overtake them. Personally I think all the papers have value in one way or another but anyone is free to form their own opinion.

    It is silly enough that people judge a paper by which journal it gets in, but it is total madness that people think it matters which repository is used. It is not the purpose of repositories to give a paper credibility by accepting them. That can only be done through peer-review and further work verifying them, or building on them after that. Do people judge a blog on blogspot by comparing it to other blogs on blogspot? I don't think they do. The same should be true for papers on arXiv and viXra. The only purpose of such repositories is to provide an independent place to store papers openly and indefinitely with a timestamped record of updated versions. viXra does that just as well as arXiv does. Sticking them on your own webserver does not.

    arXiv is not a "one-stop shop" people find papers from all over the place via citations and web searches for example. Our stats show that newly submitted papers on viXra are just as likely to be read as arXiv papers. Apart from our openness we have the added advantages of download stats and comment boxes which arXiv is afraid to provide. We also do it on a budget of less than $1 per submission (supported by donations) which is about a tenth of what arXiv costs. The hecklers who want research to be vetted by academics before it can be seen are not likely to stop us any time soon.

  10. Gisin's students' paper fails despite peer review. arXiv mounted Gran Sasso superluminal muon neutrino and Pioneer anomaly theories. Derivation is insufficient.

    Euclid is one of eight 3-space primary geometries. Bolyai adding two more was a goat. Newton has wrong values for lightspeed, Planck's and Boltzmann's constants. General relativity required the first GPS satellite; Einstein rejected quantum mechanics. QM predicted methane was 90° angle CH_2. CH_4 with H-C-H arccos(-1/3) was jury-rigged (LCAO).

    Some 1960's chemical reactions were impossible. They became Nobel Prizes (palladium catalysis; olefin metathesis - using "impossible" stable carbene ligands) because chemistry will admit it is wrong. Test spacetime geometry with geometry to probatively heal physics.

  11. Intentionally silencing ideas simply because they don’t fit an agenda is plainly wrong.

    Repositories are not peer-reviewed publishing venues and their sole function ought to be storing and keeping track of submissions. Endorsing submissions always runs the risk of developing into a censorship game.

  12. I think accepting anything by someone who is endorsed by someone who is respected (has scientific publications in journals) and has an academic affiliation should be the goal. This keeps it in the science community which is important. The fact that some crackpots and wrong papers will get in is not too important (it is true with papers that get accepted by journals too).

    Someone who is a former scientist and no longer has affiliation should still have friends/colleagues who have affiliation. There should be no penalty for endorsing and no reason for a scientist not to endorse his/her former colleague/student (maybe best to not keep track of endorsers). So the hobby scientist should still be able to submit papers.

    Someone who puts on too many papers (6 in a two year period?) that are not accepted for publication in a journal can be 'punished' with a special tag (phys-probation until he/she loses this 'status'). This would handle the 'troublesome' cases.

    Just my suggestion.

  13. If it were true that everyone ignores papers on viXra because it is all "crackpot" we would expect the average number of downloads for a paper on viXra per day to be much less than the equivalent figure for arXiv. Unfortunately arXiv don't say how many paper downloads they have per day. The only fugure they give is the number of web connections which is about 60 million per day. They have about 1.1 million papers so that is about 5.5 connections per paper. For viXra the corresponding numbers are 40,000 connections per day for 13,000 papers or about 3 connections per paper. This statistic is only a rough indication but it suggests that viXra papers get about half as much interest as an arXiv paper. I think that's not bad considering we are comparing mostly funded academic research on arXiv with mostly unfunded independent research on viXra.

  14. "Some people browse arXiv, a few even browse viXra but most find papers in other ways such as web searches."

    How do you know?

    I think many people subscribe to the daily listings of abstracts from arXiv, sent out by email, scan the titles, and then read what looks interesting. Some might do the same on the website. Serious scientists don't do the same at viXra.

    "I think accepting anything by someone who is endorsed by someone who is respected (has scientific publications in journals) and has an academic affiliation should be the goal."

    The problem is that while an endorsement is necessary (which I agree with, with the caveat below), it is not sufficient, at least not in all cases. (Caveat: There are occasionally people who don't know anyone who would endorse them who write acceptable stuff. In such cases, acceptance by a white-listed journal should count as an endorsement.)

    "Someone who puts on too many papers (6 in a two year period?) that are not accepted for publication in a journal can be 'punished' with a special tag (phys-probation until he/she loses this 'status')."

    The problem here is that arXiv doesn't want to tie itself to published journals. I tried for years to get the Journal-ref to be a required field, which should also be required to be updated regularly until the final status has been achieved. At least they have the journal-ref field, but one is not required to use it.

    "Sabine Hossenfelder, an expert on quantum gravity at the Frankfurt Institute for Advanced Studies in Germany" FIAS searches null for "sabine" "hossenfelder" "sabine hossenfelder" 29 January 2016
    404 Not Found
    Requires an update?

  16. "A heated debate about arXiv’s screening policies has flared up after a high-profile physicist said that moderators at the popular preprint server rejected reasonable papers without justification."

    I was about to congratulate Sabine on being referred to by Nature as a high-profile physicist, but then I realized that they were talking about Gisin. :-|

    However, it is not a "heated debate" but a "healthy discussion".

  17. Uncle,

    They haven't yet added me to the website, sorry. And so continues my eternal fight against the forces of IT admin... On the upsite, I'm still on the website of my previous institution :o)

  18. The Arxiv should not persue the same goals as journals I think. The whole purpose of such an online database is to allow a broader diversity at the input side. The whoke purpose here is that those who consult Arxiv decide for themselves what is 'good science' and what is 'bad science'. Otherwise Arxiv will evolve into another peer reviewed journal, that would be a missed opportunity. Let each of them serve their specific purpose.

  19. @Sabine
    Improvement is forever renormalized by application scale - Sisyphus plus middle management.

    arXiv preserves its reputation by fulfilling user expectations of user-defined quality, then Korporate Kulture. How much signal is rejected is unknown. Quantum gravitation, string/M-theory...the path to solution may not be "shut up and calculate." I offer an experimental class that does not calculate to output results. Look. "8^>)

  20. I think the solution is simply to allow all publications whose authors all have PhDs. Having a PhD means some university has put a seal of approval on you. It doesn't mean you are incapable of junk but it means you have met a very minimal initiation rite. Some junk will get through but so what? ArXiv is not supposed to be refereeing anything. That is the job of the journals, not arxiv.

    The moderation in arXiv is a disgrace

  21. Thiago Guerreiro and Fernando Monteiro in Physics Letters A 379 (2015) 2441–2444  
    “… Quantum effects imply that an infalling observer cannot cross the event horizon 
    of an evaporating black hole …” 
    by using a metric that is “… essentially … outgoing Eddington-Finkelstein coordinates
    but with a variable mass …”. 

    Hawking and Ellis in their book “The Large Scale Striucture of Space-Time” (Cambridge 1973) 
    describe outgoing Eddington-Finkelstein coordinates as: 
    “… the Eddington-Finkelstein … metric … is not time symmetric … 
    from the Finkelstein diagram … the surface r = 2m acts as a one-way membrane … 
    If one uses … a retarded null coordinate … instead of … an advanced null coordinate … 
    … only past-directed timelike or null curves cross from the outside … to the inside …”.  

    By using outgoing Eddington-Finkelstein coordinates Guerreiro and Monteiro 
    have prevented ANYTHING from going from the outside to the inside of the black hole 
    through the one-way membrane at r = 2m .

    Maybe they used outgoing Eddington-Finkelstein coordinates thinking 
    that would allow the Hawking radiation to go from the inside to the outside of the black hole 
    but that is a misunderstanding of Hawking radiation. 

    As Birrell and Davies in their book “Quantum Fields in Curved Space” (Cambridge 1982) say: 
    “… At first sight, black hole radiance seems paradoxical, 
    for nothing can apparently escape from within the event horizon. 
    However … it is not possible to localize a quantum to within one wavelength, 
    it is therefore meaningless to trace the origin of the particles to any particular region near the horizon. 
    … In the vicinity of the hole, 
    the spacetime curvature is comparable with the radiation wavelength in the energy range of interest, 
    and the concept of locally-defined particles breaks down …”. 

    Guerreiro and Monteiro made two serious conceptual errors: 

    1 - The one-way membrane in the Finkelstein diagram they used was the one 
    that permits only outgoing particles and prohibits infalling particles 

    2 - They describe Hawking radiation in terms of “virtual particles … not subject to causality” 
    when as Birrell and Davies made clear “the concept of locally-defined particles breaks down”. 

    It is sad that the Cornell arXiv administrators did not just simply point out those clear and simple errors, 
    instead not only rejected the submission without explanation but also blacklisted two people 
    making a sincere (if somewhat naive) effort at understanding Hawking radiation. 

    If there had been explanation, maybe instead of feeling blacklisted and rejected 
    they could have learned from their errors and gone on to make significant advances in black hole physics. 

    Another lesson from this event is that people should read and understand basic 20th century books on the subject
    such as Hawking and Ellis and Birrell and Davies (that applies not only to these authors but also to referees/moderators).

    Tony Smith

  22. Dear Sabine,

    I am not sure if the paper being wrong is the problem and that more money for the arxiv is the solution.

    Of course, the arxiv must protect itself from abuse. Submissions by crackpots, cranks and trolls have no place on the arxiv.

    But I also very much agree with Nicolas, and I believe with many other colleagues, that when trained young physicists make an honest effort and submit their work to the arxiv, then the arxiv ought to accept their submissions - even if experienced moderators feel that they can tell right away that the paper’s approach is not going to work. Moderators have little time to scan a paper, aren’t infallible, and young researchers can be prone to executing and/or presenting even good ideas poorly.

    Therefore, I too believe that we should let them get their work out on the scientific marketplace of ideas as fast as anybody else’s and without prejudice, i.e., let them post their papers on the arxiv. (Sure, they could submit it to vixra but then that would prejudice referees.)

    Of course, almost all young researchers with ideas that are disruptive if they work out will fail, just like startup companies. There aren’t many Googles and Facebooks. But does the economy need “moderators” from the chamber of commerce to decide which proposed startup company should get a business license and which not, on the basis of perceived chances of success? Of course not. The requirements for a business licence are minimal. And they are minimal for a good reason: while it is the founders who carry the risk, we all benefit when some of them succeed.

    Similarly, young researchers who want to post potentially disruptive ideas know very well that they risk their scientific career. These are brave souls and we should give them a chance. If they put their name on a paper that turns out to be wrong or turns out not to find interest, then they very personally carry the risk. But we all reap the benefits when some of them succeed.

    It is in the nature of things that young researchers can maximize their chances to succeed in an academic career by working in the mainstream, preferably in a currently “hot” part of it, because that’s where the referees, the mentors, the citations and the grants are. That’s good because hot topics tend to be hot for a good reason. But working in a hot topic research area is the opposite of challenging the status quo. It’s like working for a hot company like Google instead of founding a new startup to revolutionize the way things are done.

    I think it is therefore in the interest of science that the arxiv accepts all submissions by properly educated young researchers who make an honest research effort. Experienced moderators should be able to quickly distinguish them from crackpots, cranks and trolls, and be lenient if there is a grey zone. The task of the all-important peer review is better left with the journals: The arxiv is neither equipped nor is it incentivized to do a professional job at refereeing. The journals are much more likely to do a good job at refereeing because their survival depends on it.

  23. Solve the arXiv problem with two public quarantine areas.

    1) Marginal error. arXiv users are invited to find the error (character limit) in dubious papers for 30 days. Papers with low view rates are automatically rejected. Interest plus passing peer review qualifies for arXiv mounting.

    2) Zowie. Seven day listing for unorthodox papers. Unprofessional format is automatically rejected (including absence of timely citations), 5000 word limit. Same criteria as (1) thereafter. Consider a secondary six month further quarantine including reader interest, re the vacuum diode perpetual motion machine. It is obviously wrong. I defy the average professional physicist to find the error in reasoning. No reason, no rejection. Meet your own standards.

    After 45 years of gravitation and SUSY failure, we do not know the answer. Anything contradicting prior observation will not work. What remains may hold the answer. Consider alternatives.

    1. Excellent idea. All arXiv papers need comment sections, with user self moderation

  24. "I think the solution is simply to allow all publications whose authors all have PhDs."

    Of all the suggestions I've heard on this matter, this is the most absurd. It is not only common but also expected in many fields for people to write papers before getting a doctorate. Why should they not go on arXiv?

    Oh, you must ban Freeman Dyson then, as he does not have a doctorate. Neither did Thomas Gold.

  25. PhilG,

    I have to echo what Phillip said above, I and most of my colleagues are subscribed to the arxiv listing in one way or the other. If you scan a list to check whether there is a relevant development in your field you don't want it to be 95% nonsense.

    I don't know what you think the download statistics tell you about the quality of submissions.

    Bee: I don't have a particular problem with that website, I think it serves a function in a particular community and that's beneficial for everyone involved. But don't tell me they serve the same function because they simply don't.

    Phil: They do serve the same function. viXra is just filling the gap for those who are kept out of arXiv, but otherwise it is doing exactly the same thing.

    If you think that viXra fulfills the same function in the physics community as does the arXiv you are seriously delusional. Literally no serious researcher I know has looked at this site more than once. Sorry if that hurts. As I said, I think it's great that amateur scientists have a place to go and do their thing, but there is a big quality gap inbetween. Best,


  26. PhilG, one can allow in meritorious cases like that. Graduate students could have their papers undersigned by their advisors. I think arXiv should also have an style comment section.

  27. Actually, Phil, now that you mention it, I think having papers by graduate students labelled so and undersigned by their advisors would be very welcomed by all. Graduate students could point with pride and say, you see, I did this even before I had a PhD, advisors could say, you see, I was the guy who created this future hotshot, and recruiters might find such info useful too. Anyway, I suspect even if the graduate student papers were junky, they would represent only a very tiny percentage of the total.

    What is junk? Some people might say loop Gravity is junk. Or the recent spate of papers explaining the 750GeV bump. Should arXiv be in the business of deciding what is Truth? I don't think so. Let the journals grapple with that one.

  28. Actually what is junk, is historically extremely difficult to define, especially in terms of amateurism and madness. Just think about Newton (amateur and crazy), Cavendish (amateur and completely mad), Faraday (amateur and incompetent), Einstein (amateur). A theory like MOND was considered, when it was published, as a pure delirium (not based on GR and an ad hoc modification) it's now one of the most serious candidate to explain dark matter (an another historical crackpotery by Zwicky)... No really, for the physics, the crackpots does not leave theirs places, it even seems that they are indispensable for advancing.

  29. "one can allow in meritorious cases like that. Graduate students could have their papers undersigned by their advisors. I think arXiv should also have an style comment section."

    This is essentially already in place, and has been for years: it's called the endorsement system. Since everyone has to have an endorsement anyway, I fail to see any advantage for barring people without doctorates, then making an exception if they jump through some hoops which are essentially the same as the ones everyone has to jump through in order to get endorsed.

    Again, if you do this, you would have to ban Freeman Dyson, as he doesn't have a doctorate. He doesn't have a supervisor, and never did. Even if he did, his supervisor is long since dead. Oh, you want to make an exception based on quality? Isn't that what we are trying to avoid, that arXiv determines what is good and what is not (apart from keeping out crackpots)?

  30. Phil, you are misunderstanding me. I am trying to point out a minimal set, not a maximal set of people to be allowed to publish automatically. I think arxiv is not following the endorsement system at all. There is ample evidence of that. The scary thing is that they refuse to tell us what is the real policy they are following, like most repressive governments.

  31. rrtucci:

    I agree with you in principle - it would make sense to automatically endorse everybody with a PhD in physics. But in practice I don't see how this would be an improvement. There isn't any accepted format for a PhD title and it's not like you get a digital ID with it. Just looking at my own PhD title, it doesn't take much to produce a scanned image that says something with PhD on it and some stamp and some signature. And what would you do to check that it's a real title? Well, you'd ask a few trusted sources, which is exactly what the endorsement does.

  32. "Phil, you are misunderstanding me. I am trying to point out a minimal set, not a maximal set of people to be allowed to publish automatically. I think arxiv is not following the endorsement system at all. There is ample evidence of that. The scary thing is that they refuse to tell us what is the real policy they are following, like most repressive governments."

    Are you referring to me (as opposed to Mr Gibbs)? If so, cite me as I appear on the blog to avoid confusion. In any case, what is the point in trying to change a policy if they don't follow it anyway.

    With regard to the endorsement system, I believe it is (in the view of arXiv) necessary, but not sufficient.

  33. "I agree with you in principle - it would make sense to automatically endorse everybody with a PhD in physics."

    No it wouldn't, since then we would be inundated with crackpot stuff by the likes of Brian Josephson*, Frank Tipler, and Jack Sarfatti.

    Some sort of control is good, the questions are how it should work and how transparent it should be.

    * Which shows that automatically allowing winners of the Nobel Prize in physics wouldn't be a good idea either.

  34. Phillip,

    That someone is endorsed doesn't mean their papers shouldn't be moderated any longer. And in any case, also an endorsement won't prevent people going off the deep end. Which, incidentally, isn't all that uncommon among physicists of a certain age and, dare I say, maybe not even an entirely bad thing. Best,


  35. I have 3 papers on arXiv, published some years ago. Then, one day, I have been notified that another paper has been rejected and since then all papers are rejected. Stranger, I can update papers already on arXiv...
    If I may accept the need to some basic screening, it is clear that arXiv goes well beyond that, and implements a kind of "scientifically correct" policy. This is bad for innovation, and wrong for an institution which is funded by the community.
    Anyway the idea that to publish a paper you need to be endorsed by 2 boffins in the community is certainly not the best way to stimulate innovation. If, before launching a new product, a company has to get the - anonymous - agreement of of its competitors, do you think there would be much innovation ?

  36. This is the sort of topic that is most natively attractive to and attracting of, bias-carrier gremlins, those highly infectious little goblins.

    It really pays back any precursor introspection that you can do. An example would be to ask, what is this matter really about? What is incidental and what goes to heart?

    Is it about who should be permitted to submit and be published in the same carriage as oneself, who has qualified and earned a little prestige, status, station in the affairs of physics. s

    Is that where the heart is? I doubt it, but just to mention that sort of thinking, while reasonable sounding justifications are always available, founds on fear. Or progressively becomes about fear. It's one of those paths you don't walk down the path, the path walks down you.

    Even if it wasn't even about that. Language that we use decides what's what. In this sort of area anyway.

    If it's not about that, then that precursor introspection is where the language gets straightened out. Logically if isn't about privilege, then in an ideal situation, anyone and everyone would be free to submit and publish. In an ideal situation.

    So then it becomes on what does the situation turn, that in one setting it's what we have now, while in another none of the issues in play now are better than secondary.

    One solution would this is about sophistication of ancillaries such as filtering, organizing, views (of data and contacts, say). I think that is what this really is about. Or that perspective is available and no less true, depending what is most true for you (and the motivation most driving this topic).

    If it were a choice between two perspectives, one being just technology, which the other being about messy things. It's worth pondering that actually one of those two has within it the potential to capital S solve everything even the outliers. While the other...cannot really solve anything at all


  37. "No it wouldn't, since then we would be inundated with crackpot stuff by the likes of Brian Josephson*, Frank Tipler, and Jack Sarfatti."
    Inundated? Come on. One could impose a limit of 12 papers a year. It would be trivial for the website to store in your computer a personal blacklist of authors and not show any of their papers to you.

  38. Hi all,
    There is a website where one can comment on arxiv papers. Right now this is used mainly for astro papers,
    but in principle it can be used for other categories also.

  39. "That someone is endorsed doesn't mean their papers shouldn't be moderated any longer. "

    True. So, automatically endorsing everyone with a degree in physics (does it have to be a doctoral degree?) is OK, but it shouldn't be the only possibility for endorsement.

    While I replied to your comment, I was unconsciously referring to this: "I think the solution is simply to allow all publications whose authors all have PhDs.". This doesn't sound like endorsement; it sounds like automatic acceptance.

  40. Just to add there is also (which mostly was used for cosmology papers), but is not actively used anymore.
    But in principle anyone can comment on any arxiv paper on cosmocoffee and arxiv will link to it.

  41. Hi Sabine - you don't appear to have allowed yet my comment about the Information Paradox. You probably didn't see it or it hasn't arrived, although I did send a second copy. Do you want to resend?

  42. Lucy,
    I received your comment. I have not, and will not, approve your comment. It is off-topic, please read our comment rules. I will keep in mind that there are still people confused about what the problem is. If I write a post on the topic in the future you are welcome to comment.

  43. Sabine,

    Your title question is rhetorical, of course, since there is no doubt that the ArXiv rejects some papers - that's censorship by definition. You and most of your commentators argue about the grounds for censorship, but I don't think that's the real question, which is "Quis custodiet ipsos custodes?" as Juvenal put it -Who watches the watchers? The danger of a non-transparent censorship system is that crackpots or other nutjobs can get control of the process.

    That can't really happen with the journal system, since they are many, diverse, and compete with each other. The problem is that everybody knows that journals are obsolete. So the question becomes, How do you preserve the advantages of the diverse, distributed system while gaining those of a unified, cheap, and semi-automated one.

  44. Bee - Hope all is well! As a non-academic/researcher but an individual who still does check arXiv, I was not aware the level of moderation taking place. While I understand that the volume of submissions has grown tremendously, not sure too much time should be put on the moderation, especially when storage is cheap and searches can be done instantly.

    Why not a tiered moderation field that is searchable? For instance, authors previously published X times in a major journal are tier 1 and not further reviewed. Tier 2 could be "selected" papers passing some arXiv formulated critera. Tier 3 is everything else. Note that there need not be stigma attached to tier 3, maybe your paper just did not interest the moderators. Those looking for papers just have an additional option to check off if they wish to narrow the search.

  45. CapitalistImperialistPig,

    You made a good point about who-watchers-the-watchers but then you take the interesting step of qualifying who possibly the watchers could be to this range: crackpots or other nutjobs.

    It's two times interesting: Firstly "who watchers the watchers" is traditionally accepted to mean 'any of us' in a statement directly of human nature.

    Secondly - and this is what makes the first one interesting also - your point is very pertinent used proper (i.e. not explicitly qualified but implicitly directed at incumbent regime).

    I think I can predict your position and part in things from that information alone. Will you allow me to try? I give you my word I'm commenting directly upon reading your comment (i.e. no research attempted). So I may well be wrong and would be interested to know if I am.

    My prediction is that you are a capitalist imperialist pig. How did I do?

    Just kidding.

    No my prediction is that you are either in, or in some sense beholden to - that is dependent on the goodwill of - the political wing (or hat) of the string community.

    Am I right? If you tell me, I'll tell you how I could tell. I'll be really impressed though, if you can tell me how I thought I knew. Especially if I was wrong.

  46. I have many papers published in good journals ( However arxiv on hold my papers and put them in phys-gen physics category. I sent them email to ask to put to article in gr-qc. however they removed my paper and sent me this message. This is unfair.

    Dear arXiv user,

    Our moderators have determined that your work would benefit from additional review and revision that is outside of the services we provide. Our volunteer moderators are not referees and do not provide reviews or other detailed feedback for improvement of submissions with their decisions.

    As a result, we have removed your submission. Please instead send your paper to a conventional journal for the necessary reviews.

    Please do not resubmit this paper without seeking permission and obtaining a positive response. Resubmission of removed papers may result in loss of submission privileges. We will reconsider this decision if your work is published or accepted with a resolving DOI (Digital Object Identifier) or link to the journal's website showing the status of the work.

    For more information on our moderation policies, see:

    arXiv moderation

  47. @A. Ovgun:

    The policies of arXiv are not always transparent and perhaps not always correct. However, the reason that it is arXiv and not viXra, i.e. the reason that it is taken seriously at all, is because of the moderators. Surely you are not claiming that because you have papers in good journals you should automatically be able to post anything in any category at arXiv. Even a Nobel Prize in physics doesn't give you carte blanche to post outside the general-physics category.

    The answer seems reasonable to me: if you publish the paper in a respectable journal, then arXiv will allow it to go into the appropriate category.

    Outside of a few exceptions, papers on arXiv aren't taken seriously anyway unless they have been accepted by a reputable journal.

  48. Then why not segregate all papers in that manner until they are published, then categorize them after that point?

    Better still modernize their system completely. Do away with the categorization entirely! Instead just have a search box. If people want to search on publication status they can.

    Remember the reason the arXiv works the way it does is mainly because of its history. As I recall, when it was founded back in the 80s or early 90s they had a Macintosh latex the papers and a 486 PC as web server. Is that time only barrier to entry was being able to learn latex. I got papers on there at that time when I was in high school which were eventually removed because I knew latex.

    Now my phone has the power to store and serve millions of papers better than tah setup did. I mean....they send out email this is 1995.

  49. The right way to go is simply the following:

    First, submit to ArXiv preprints (i.e. two or more) of already published papers (i.e. Author's original or accepted version) in respected Journal's (i.e. IF more than 1). That way you already have build a certain trust level. Then when it comes to submit to arxiv a preprint before acceptance in a Journal minimal screening of basic acceptance criteria (i.e. scientific format, basic level of rigorous presentation) should be applied by the moderators. Anything more than that defies the very purpose of an unpublished research preprint thus new research to be reviewed openly by the science community with a grain of salt and with a potential precious feedback and critic to the author(s) work.


COMMENTS ON THIS BLOG ARE PERMANENTLY CLOSED. You can join the discussion on Patreon.

Note: Only a member of this blog may post a comment.