The Roman Philosopher Lucius Anneaus Seneca (4 BCE-65 CE) was perhaps the first to note the universal trend that growth is slow but ruin is rapid. I call this tendency the "Seneca Effect."

Friday, November 12, 2021

When Science Speaks in Tongues: The Unstoppable Rise of Gibberish

  

I have no objections to the idea that God (or the Goddess) can speak to people. And maybe the Lord really spoke to the apostles the day of Pentecost. But if you plan to fool other people, then "speaking in tongues" (known also as "glossolalia" in modern terms) can be a good idea. You start with uttering something that vaguely sounds like a language. Then you "translate" it into something that you report as the word of God. An easy trick that sometimes works. Even in science, some researchers seem to use this trick to gain academic points by publishing articles that contain mostly gibberish, or even only gibberish. We could call it "scientific glossolalia". 


You may have heard the recent news of 44 scientific papers being retracted from a scientific journal after they were discovered to be nothing but gibberish. The usual reaction in these cases is to speak of "a few bad apples." But this fraud exposes a problem that goes deep, very deep, in science. Science suffers from "glossolalia" -- a syndrome that makes people utter meaningless sounds as if they were speaking a real language. 

To start, "papers" are the main output of a scientist's work. It is the harsh law of "publish or perish," meaning that for a scientist publishing something -- anything -- in an academic journal is the first line of defense against being fired.  Even if a scientist has no money, no grants, no instrumentation, no ideas, they have to show that they are doing something. Woe betide the scientist who does not publish at least one paper every year! Anathema! Abomination! Horror! May you be eaten by the h-index bugbear who punishes those who sin so hideously against the sacred rules of science! 

But publishing papers has a problem. When scientists publish something, in a certain way they are showing their hand. Readers will be able to understand how good they are, how well they master their field, how much money they have to perform their research, and more. They may not want others to know that, especially if they have something to hide (almost everyone does, in this world). So, many scientists practice obfuscation in order to defend their turf.

So, scientists want to publish papers, but they may not want others to read them. A way to do that is to use purposefully convoluted language, eliminating all elements that would make a text interesting, turning it into the most boring possible kind of prose. The use of the passive form is a typical example ("it has been found that") instead of the simpler "we found that....". But there is more: for instance, why do scientists often sign their papers only with the initials of their first and middle names? ("J. I. Smith" -- does it mean "Jolly Idiot Smith," or what?). The idea is the same: to remove all hints of human interest for the text. 

By far the most effective strategy is to use obscure terms. Uncommon and archaic ones can do a good job of repelling readers. An example noted by Malcolm Kendrick in his "The Clot Thickens" (2021): why in the world would anyone write "pultaceous" instead of "pulpy" if not with the specific purpose of being obscure? But what makes a paper truly unreadable is the proliferation of acronyms. If you stumble on "GDAP," you have to decide which one of the 9 known meanings it can take (here, it is "Growing Danger of Acronym Proliferation").

So, you start understanding how the mechanism works. First of all, an obscure paper makes it difficult for the reviewers to wade through the text and, surely, they don't want to appear ignorant by asking what a particular term or acronym, or whatever means. Then, the paper may be full of mistakes, inconsistencies, shortcomings, and plain lies but, if it is really obscure, there is a chance that neither the reviewers nor anybody else will read it through and notice its shortcomings. It may even be cited, thus providing some extra points for the authors, by those who just read the title. Of course, it won't make the authors candidates for the Nobel prize, but it means some respite from the wrath of the scientific PTB (obscure acronym for the "powers that be").

From this point onward, it is just a small step for a desperate scientist to jump from simple obfuscation to straight fraud (and a big step backward for science as a whole). You have papers based on invented data, on shaky statistical methods, on groundless assumptions, and more. 

At this point, we should not be surprised that someone used one of those AI (obscure acronym for artificial intelligence) text generating programs to create from scratch completely meaningless papers. These programs are already impressive in their generic versions, but someone must have developed a specific version for creating fake scientific papers. Take a look at some of the 44 retracted papers, and you'll see how sophisticated the program that created them was. For instance, this one: you need to know at least something of geology to understand that it is a pure glossolalia piece. The author (the AI) is speaking in tongues. It is only because there are people who know these matters that the scam was detected. 

But how many scams of the same kind were NOT detected? Do you know that 2.5 million scientific papers are published every year in the world? Detecting those which are pure assemblies of random sentences may not be so difficult (AI may fight AI). But the truly horrible thought is how many papers are NOT glossolalia pieces, but are nevertheless unreadable, poorly done, wrong in their basic assumptions, using massaged data, arriving to unjustified conclusions, and more. In short, papers that are at best a useless waste of money, at worst scams engineered to support the dark purposes of some lobby acting behind the scene. 

Of course, not all science is like this. There is a fraction of scientists who are competent, sufficiently financed, safe in their positions, who create good science that advances human knowledge. How many? Difficult to say. Maybe there holds Pareto's law in the form of "80% of good science is done by 20% of the scientists." Or maybe we can apply Sturgeon's law ("99% of everything is crap") also in its strong form (99.9% is). 

The problem is that with less and less money available, science is more and more in the hands of underpaid and blatantly exploited people who have no perspectives for a decent career. Some of them may well be desperate enough to recur to fraud. Note also that it is a scientific law that entropy always increases, so how long will it take to transform science into a pultaceous mass of meaningless sentencesIn a previous post, I wrote that science may already have expanded itself beyond existence. 

Can this situation be remedied? Maybe, but that would need truly drastic actions to change at its basis the perverse mechanism of publish or perish. It is unlikely that the task will ever be undertaken by universities or by governments, or by the scientists themselves. Fortunately, the Seneca Cliff takes care of eliminating the EPCS (obscure acronym for the Entropy Produced in Complex Systems). It will do that for science, too. It won't be painless for scientists but another form of the principle of entropy is that everybody gets what they deserve. 

 


18 comments:

  1. Ugo,

    The same thing is obvious in contemporary art too. It was Giulio Carlo Argan who first made a distinction between "working art" and "being artist". Everybody can work art, it's enough for a person to just buy paper and paints. But being artist requires artistic mind, which is rare. The problem here is that both science and art became forms of industry which is destroying both science and art. It's now mass production, like industrial mass production. And we all know that there is not many great industrially made products. Most are made to last short time and be replaced by another product. Science and education are now subject to the principles of "economy of scale". Publish or perish is simply ruining science, removing the last traces of relevance.

    ReplyDelete
  2. You mean, "What *we* deserve.", right?

    I have also published in obscure journals like International Journal of Robotics Research. For better or for worse, I moved out of academia before I had to write obfuscated papers. In my naïve youth, I actually wanted people to read what I at the time thought were genius insights... ;)
    Nevertheless, my academic title will allow me to join the ranks of lamp-post decoration.

    However, I have seen the same things happening inside several multinational corporations, especially among the priesthood of marketing and IT. Internal project proposals can overflow of platitudes. I am sure you have heard of bulls***-bingo?

    As David Graeber called our era of "Managerial feudalism", the content is much less important than the form. The loyalty upwards, giving prestige to your boss and institution, is more important than delivering anything of "value". Keeping the boss happy allows you to sit at the table and split the loot and share the crumbs.

    I think that is why almost all innovation comes from small companies that are not yet corrupted this way. The ideas/products/concepts/companies are then acquired by the giants and globally rolled out.
    Ten years ago I hired 20 engineers and scientists to work in a commercial R&D center, and it was very interesting to look at output vs. education. The more education, and the more elite university, the less useful output. But far higher salary expectations...
    I don't know if anyone else has had the same experience?

    I recommend my children to first get a real education (plumber, electrician, farmer) before (if ever) going to college.

    Maybe a next post about the corruption in science, "the price of a paper"? Here in Holland there are a number of well-documented corporate purchased "scientific" papers from University of Wageningen and University of Rotterdam, that have been used to push political decisions. "Evidence based" decisions...
    I think the fraud/scam-area is where the exposed rot will first lead to demonstrations with banners like "Defund the University".
    I think that the junk papers in (expensive!) journals is more of a waste/noise problem for the still-honest researchers inside the system.

    Goran (PhD)

    ReplyDelete
    Replies
    1. Goran:

      I agree completely, I walked away from "science' in 2007 and have never regretted the decision. When folks find out about my background and asked me why I am working as a middle-tier bureaucrat with no responsibilities, I merely tell them "I got tired of being a whore".

      John (Ph.D, Sci.D)

      Delete
  3. Did you know that Byzantine Emperor Justinian banned the use of acronyms because they became so confusing in public life and legal documents. There was hundreds of them in use in 6th century. In Latin acronyms were called "sigla".

    ReplyDelete
    Replies
    1. I agree with Justinian, if I was emperor I would only allow acronyms in any text after they had been written in full at least 3 times with acronym shown.

      Delete
  4. El caso que citas dice más de la calidad de la revista que la de los supuestos científicos ( Algunas investigaciones parecen merecedoras del Ignobel ). Con la ingente demanda de publicar, hacerlo en el oligopolio de las revistas top es muy difícil y muy caro. Así que será fácil generar una oferta que satisfaga a tanto investigador necesitado. En ese mercado podrán encontrarse tanto ( algunos ) científicos limitados o sin escrúpulos con (ciertas) editoriales corruptas.

    Cinco grupos empresariales controlan la inmensa mayoría del mercado, con lo que ya no sólo debemos fiarnos de lo que los científicos producen, si no también de lo que las editoriales seleccionan y difunden.

    Lo de publicar es problemático, pero lo de las patentes tampoco está mal. Cierto que se protegen las inversiones, pero cuando la protección es desmedida producen efectos perversos.

    ¿ Pueden imponerse patentes inasequibles a fármacos que salvarían las vidas de mucha gente si pudieran acceder a un medicamento genérico, menos protegido?

    Ciertamente, recuperar la inversión más algún beneficio es la primera intención de los inversores en ciencia. También, cada vez más, en las instituciones y universidades públicas.

    Pero esa cara interesante del mercado se ensombrece cuando se comprueba que obstruyen la innovación y oponen barreras de entrada infranqueables a terceros interesados en explorar ese ámbito.

    ( Por cierto, las patentes solicitadas por científicas son rechazadas en mayor medida que las que presentan sus colegas masculinos )

    ¡ Qué tiempos aquellos en que los grandes científicos reconocían que sus logros se debían en gran parte a que cabalgaban “a hombros de gigantes” !

    ReplyDelete
  5. Me permito reponer los dos últimos párrafos, que pueden resultar ilegibles.

    ( Por cierto, las patentes solicitadas por científicas son rechazadas en mayor medida que las que presentan sus colegas masculinos )

    ¡ Qué tiempos aquellos en que los grandes científicos reconocían que sus logros se debían en gran parte a que cabalgaban “a hombros de gigantes” !

    Y un colofón:

    La restricción del conocimiento ( por parte del oligopolio de las revistas ) y su apropiación privada ( con las patentes ) son de las cosas más anticientíficas que puedan existir.

    ( Y última)
    Creo que los sistemas de publicación científica abiertos son cada vez más eficientes. Las instituciones y universidades públicas deberían apostar sin reservas por ellos hasta que se restablezcan las oportunidades de acceso público y general al conocimiento. Mientras tanto, porqué deberíamos renunciar a la piratería.

    ReplyDelete
  6. It is not only Science - the Media, Literature, Economics, Academia, Law, Arts, History, Military, Industry - all became hocus-pocus.
    The Civilisation that is ending up today a Syringe-Factory must have confiscated Science long ago, turning it like all other fields - a Gibberish, too...


    One colossal mistake our Western Civilisation has committed, forcing all fossil fuels-rich nations giving up their resources - not on the basis the resources are finite - has destroyed the chemistry of sane humanity.

    Knocking down one and only one parameter - the sense of scarcity, built by nature in the DNA of humanity - by our Western Civilisation - has played havoc with the future of humanity.

    Science needed to intervene against the Social Engineer, when the world's population has been a couple of billions - at max.

    Actually, Science honestly did - read the 1860s' Jevons...

    Today we are ending up a paralysed Syringe-Civilisation obsessed with people bodies, almost no fossil fuels left for the masses - but our Social Engineer still preferring to play the Energy Musical Chairs Game.

    Wailing.

    ReplyDelete
  7. Push that same approach hard and you get lots of hits on obscure keywords that the AI reading the papers will index on and 'poof' ... another unique contribution to the sum of human knowledge !

    ReplyDelete
  8. I thought it was standard practice with acronyms to use whole words for the first use and then put the acronym in parentheses after them. No matter how common the acronym, the first use should be whole words. If used in a book, it's nice to repeat the process with every chapter or, even better, have an acronym glossary at the front or back of the book. Published Acronyms Should Always Be Explained (PASABE).

    ReplyDelete
  9. Hello Ugo. Most of those junk papers have been taken down with editorial comments that say they were "nonsensical" and got past peer review. Maybe someone from Springer reads your blog as well as Mr. Gates 😆 Those "word spinner" programs used to be pretty common on the web for SEO use, maybe they still are, but Google seems to be better at catching and deindexing that kind of content.

    ReplyDelete
  10. Maybe they really read my blog. 80 articles retracted during the past 3 days!! https://retractionwatch.com/2021/11/14/publisher-retracts-nearly-80-articles-over-three-days/

    The problem is that the rhythm of publication is about 5 papers PER MINUTE!!! Hard time for the checking AI

    ReplyDelete
    Replies
    1. Just a bit of a humorous taste of a wordspinner algorithm is on a delightful art of science fiction type site .Attempt to link
      https://thrilling-tales.webomator.com/derange-o-lab/title-o-tron.shtml#titleotron

      Delete
  11. So 5 papers a minute hitting the inbox and a decent peer review process that takes a couple months per paper. Might be a problem keeping up... 😆

    ReplyDelete
  12. ¡ Qué casualidad ! Tom Murphy, avanza en su último post interesantes ideas sobre estos asuntos. En su blog Do the Math. ¡ Tienen un PLAN !
    https://dothemath.ucsd.edu/2021/11/finally-a-plan/

    ReplyDelete
    Replies
    1. Interesting link. I haven't translated yet, but your message came thru clearly.

      Delete
  13. I'm very happy this issue is finally being covered here. I've heard Ugo mention a few times how “science is disappearing”, as magic was in some Ursula K. LeGuin story, and I was wondering if the long explanation were ever going to come.
    At the moment I’m working for a science institution as an engineer developing equipment for space research. I’m not a scientist myself (making toys is much more funny than writing papers), but I’ve seen enough of them not to have too much respect. Postmodern individuals (sophists) naturally take the higher ranks. As I see my work and the institution more and more pointless, I’m working my way off as a highschool teacher, so I’m back at university as a student to get the degree I’ll need.
    Now, as a future science teacher, I’m being told how science research works and, of course, I feel the teachers are clueless about how research works. They talk about the risk of money biasing research, some women discrimination (it’s trending topic now, so they have to cover it), but it’s sad to see how they hardly scratch the surface. Please, include as many interesting links as you can, both article and comments.

    ReplyDelete