Category Archives: Philosophy

Wissenschaftspolitik ist Politik, nicht Wissenschaft

Ein schon älterer Essay in “Forschung und Lehre” von Thomas Naumann zeigt einige Grundsätze der Wahrheitsfindung in der Wissenschaft

Ziel der Naturwissenschaften ist es, die uns umgebende Welt zu erkennen. Das heißt, Aussagen über die Wirklichkeit zu gewinnen und sie in Form von Beobachtungen und Gesetzen abzubilden. Diese Aussagen und Abbildungen bezeichnen wir als wahr, wenn sie mit der widergespiegelten Realität übereinstimmen.

Die Aussage, so Naumann, bleibt aber subjektiv und ist nicht deckungsgleich mit der objektiven materiellen Realität. Im Experiment wird die wissenschaftliche Wahrheitsaussage sprich Hypothese immer wieder überprüft. Wo dies nicht möglich ist, wird wenigsten versucht, die Beobachtung zu verifizieren (die aktuelle Replikationskrise ist deshalb auch eine Krise der Wissenschaft) wobei  Hilfskriterien wie Logik und Widerspruchsfreiheit helfen können genauso wie Vorhersagen, die dann eintreffen oder auch nicht.  Doch zurück zur Wahrheitsfindung bei Karl Marx, wieder nach Naumann zitiert

“Die Frage, ob dem menschlichen Denken gegenständliche Wahrheit zukomme – ist keine Frage der Theorie, sondern eine praktische Frage. In der Praxis muss der Mensch die Wahrheit … seines Denkens beweisen.” Auch Max Born, einer der Begründer der Quantentheorie, riet zu Realismus: “My ad­vice is not to rely on abstract reason, but to decipher the secret language of Nature from Nature’s documents, the facts of experience”.

Wissenschaftler sind damit der unbedingten Ehrlichkeit und Wahrhaftigkeit verpflichtet.

Die primäre Anforderung an Politiker ist eine andere, sie sollen mit sozial verantwortlichen Entscheidungen gesellschaftliche Probleme lösen , Fortschritt ermöglichen und wo immer möglich, Krisen verhindern. “Ehrlich wärt am längsten” gilt auch hier, obwohl man Politiker im allgemeinen zugesteht, ihre Ziele nicht auf direktem Weg zu erreichen. Ihre Ziele sollten aber doch mehr am Allgemeinwohl als an ihrem privaten Nutzen orientiert sein, ansonsten wird er abgewählt.  Die Washington Post hat im übrigen über 30.000 Lügen von Donald Trump gezählt.

Hannah Arendt hat 1963 einen schönen  Aufsatz geschrieben “Wahrheit und Politik

Der Gegenstand dieser Überlegungen ist ein Gemeinplatz. Niemand hat je bezweifelt, daß es um die Wahrheit in der Politik schlecht bestellt ist, niemand hat je die Wahrhaftigkeit zu den politischen Tugenden gerechnet. Lügen scheint zum Handwerk nicht nur der Demagogen, sondern auch des Politikers und sogar des Staatsmannes zu gehören.

Seit eh und je haben die Wahrheitssucher und die Wahrheitssager um das Risiko ihrer Unternehmung gewußt… Wiewohl es im Politischen zumeist die Tatsachenwahrheiten sind, die auf dem Spiel stehen, ist der Konflikt zwischen Wahrheit und Politik zuerst an der Vernunftwahrheit ausgebrochen und entdeckt worden. In den Wissenschaften ist das Gegenteil der Wahrheit der Irrtum oder die Unwissenheit.

Modern würde man also sagen – Politik und Wissenschaft ist ein Clash of Cultures.  Wissenschaftspolitik ist eindeutig dem politischen und nicht dem wissenschaftlichen Lager zuzuordnen – es geht primär um Finanzierung von Hochschulen und Forschungseinrichtungen in einem undurchsichtigen Mix aus Prioritätensetzung, Bildungsauftrag, technische Weiterentwicklung und Folgenabschätzung, Wirtschaftsförderung verbunden mit dem massiven Eingriffen in die eigentlich garantierte  Wissenschaftsfreiheit.

Interessant wird das Thema vor allem dann, wenn Wissenschaftler in die Wissenschaftspolitik oder sagen wir mehr allgemein, den politischen Apparat wechseln. Waren sie vorher schon verkappte Politiker? Oder gibt es da eine Phase der Adaptation und Neuorientierung die nun Lüge erlaubt? Das ist eine, wie ich finde, neue und interessante wissenschaftliche Fragestellung für Wissenschaftssoziologen. Schade, ich hätte zu gerne hier Latour befragt. In den Zettelkästen Luhmanns habe ich nur einen passenden Zettel zu dem Thema gefunden “Politik als Wissenschaft kann es nicht geben”…

Nochmal Arendt a.o.O.

Der Streit zwischen Wahrheit und Politik besteht nach wie vor, nur ist an die Stelle der Vernunftwahrheit die Tatsachenwahrheit getreten. Zwar hat es vermutlich nie eine Zeit gegeben, die so tolerant war in allen religiösen und philosophischen Fragen, aber es hat vielleicht auch kaum je eine Zeit gegeben, die Tatsachenwahrheiten, welche den Vorteilen oder Ambitionen einer der unzähligen Interessengruppen entgegenstehen, mit solchem Eifer und so großer Wirksamkeit bekämpft hat.

 

CC-BY-NC Science Surf accessed 06.11.2025

Doing less

“The case for doing less in our peer reviews” by Kate Derickson is an interesting essay on scientific reviews.

While it is a luxury to receive thorough and carefully thought-out comments from a colleague, the nature of blind peer review means that the author cannot know who is making suggestions […] And yet, the author is often relying on the paper being published for professional security or advancement. This puts the author in the position of being obligated to rework their arguments according to constructive suggestions made by an anonymous person whose credibility or self-interest they cannot assess. Moreover, while reviewers often identify similar issues in a paper, they often propose a variety of different approaches to addressing them, many of which work at cross purposes. Authors can be overwhelmed by the range of suggestions, feeling obligated to split the difference and cover all the bases in case the paper goes back to all three reviewers. While papers generally get better through the review process, authors often have a difficult time navigating contradictory reviewer suggestions.

But wait, there is also a point where I do not agree (in the light of the recent elife decision).

we think carefully about what we decide to send out for peer review, in order to enable us to curate a table of contents that we think is at the cutting edge of our disciplines and of interest to our readership.

Creating the most cited journal? Creating cutting edge? This is a pre-internet 1980’s attitude of  a journal editor trying to get a higher citation impact in the competition with other journals. It simply devalues everything that Derickson does not understand or that Derickson does not want to promote.

So my initial enthusiasm of the paper finally dies with “the biggest scientific experiment

Huge interventions should have huge effects. If you drop $100 million on a school system, for instance, hopefully it will be clear in the end that you made students better off. If you show up a few years later and you’re like, “hey so how did my $100 million help this school system” and everybody’s like “uhh well we’re not sure

Yes, this is about the end of scholarly peer review as peer review fails  to catch major errors in about 1/3 of all papers.

In all sorts of different fields, research productivity has been flat or declining for decades, and peer review doesn’t seem to have changed that trend. New ideas are failing to displace older ones. Many peer-reviewed findings don’t replicate, and most of them may be straight-up false. When you ask scientists to rate 20th century discoveries that won Nobel Prizes, they say the ones that came out before peer review are just as good or even better than the ones that came out afterward.

The focus is on “cutting edge” and “interest” aka impact points but  neither on ingenious minds nor brilliant discoveries.

 

CC-BY-NC Science Surf accessed 06.11.2025

On scientific startups

Scientific startups are en vogue. I have been asked several times and also showed interest in some companies but at the end it was either too risky, too time consuming or even unethical.

There is a new and interesting blog article about orphaned neurological implants by Cory Doctorow:

The startup world’s dirty not-so-secret is that most startups fail. Startups are risky ventures and their investors know it, so they cast a wide net, placing lots of bets on lots of startups and folding the ones that don’t show promise, which sucks for the company employees, but also for the users who depend on the company’s products.

The cruel point is not burning money but about burning humans. I think we clearly need a backup for such companies like Second Sight.

Terry Byland is the only person to have received this kind of implant in both eyes. He got the first-generation Argus I implant, made by the company Second Sight Medical Products, in his right eye in 2004 and the subsequent Argus II implant in his left 11 years later. He helped the company test the technology, spoke to the press movingly about his experiences, and even met Stevie Wonder at a conference. “[I] went from being just a person that was doing the testing to being a spokesman,” he remembers.
Yet in 2020, Byland had to find out secondhand that the company had abandoned the technology and was on the verge of going bankrupt. While his two-implant system is still working, he doesn’t know how long that will be the case. “As long as nothing goes wrong, I’m fine,” he says. “But if something does go wrong with it, well, I’m screwed. Because there’s no way of getting it fixed.”

Could the FDA please stop now Elon Musk’s brain-implant company Neuralink? It is a nightmare even before it started. And well, there is also a need for a governmental backup guarantee of Synchron if we allow such startups.

 

CC-BY-NC Science Surf accessed 06.11.2025

Integrity of scientific research

I have a new book on my desk that covers most aspects of the recent discussion around research integrity in 60 chapters arranged in 14 sections.

https://link.springer.com/book/10.1007/978-3-030-99680-2

It is a comprehensive overview of research integrity across disciplines. Maybe some chapters are not written by the leading experts in the field. And maybe the scope is super-broad with several off topic chapters like corruption in healthcare or repetitive texts in subdisciplines like dental care. Unfortunately this comes at the cost that other chapters like image integrity are completely missing, while even PubPeer is not mentioned in the appendix. Also a keyword index would be nice. Nevertheless it is the largest body of text so far and should be on every desk that deals with science integrity. There are numerous highlights that cannot be found anywhere else like a comprehensive list of all paper rejection statements (p 412). So I will have to read it again next year.

 

 

CC-BY-NC Science Surf accessed 06.11.2025

Who shall survive?

published 1934

The eugenic doctrine, similarly to the technological process, is another promiser of extreme happiness to man. The eugenic dreamer sees in the distant future the human race so changed through breeding that all men will be born well, the world populated with heroes, saints, and Greek gods, and all that accomplished by certain techniques through the elimina­tion and combination of genes. If this should really come to pass the world would be at once glorious, beautiful, and God­ like. But it may be reached at the cost of man as a creator from within himself; it would have, like Siegfried in the myth, a vulnerable spot into which the thorn of death could enter,-a tragic world, a. world in which beauty, heroism, and wisdom are gained without effort, in which the hero is in want of the highest reward, the opportunity to rise from the hum­ blest origin to a supreme level. It sums up to the question whether creation in its essence is finished with conception or whether creation does not continue or cannot be continued by the individual after he is born.
The eugenic dreamer and the technological dreamer have one idea in common: to substitute and hasten the slow process of nature. Once the creative process is encapsuled in a book it is given; it can be recapitulated eternally by everybody without the effort of creating anew. Once a machine for a certain pattern of performance is invented a certain product can be turned out in infinite numbers practically without the effort of man. Once that miraculous eugenic formula will be found a human society will be given at birth perfect and smooth, like a book off the press.

 

CC-BY-NC Science Surf accessed 06.11.2025

Is ChatGPT outperforming Google search?

Unfortunately it is useless to enter any research question that I am interested in, so lets try some more trivial examples.

Alberto Romero

It’s still quite apparent that ChatGPT lacks reasoning abilities and doesn’t have a great memory window (Gary Marcus wrote a great essay on why it “can seem so brilliant one minute and so breathtakingly dumb the next”). Like Galactica, it makes nonsense sound plausible. People can “easily” pass its filters and it’s susceptible to prompt injections.

ChatGPT maybe a jump forward but it is a jump into nowhere. Time to cite again the Gwern essay

Sampling can prove the presence of knowledge but not the absence

which is again my problem that it is useless to enter any real research question as there is nothing to train the algorithm beforehand.  AI cannot make any difference between true and false as AI does not “understand” but simply calculates the strength of the association found in training texts.

 

CC-BY-NC Science Surf accessed 06.11.2025

Apropos Gerechtigkeit

 

30 Gerechtigkeitstheorien auf Wikipedia

Sowohl Sokrates und Platon als auch Aristoteles sahen das Glück als den höchsten anzustrebenden Wert an. Gerechtigkeit war für sie die oberste Tugend, um diese Glückseligkeit zu erreichen. Gerechtigkeit war so eine grundlegende Charaktereigenschaft…Kant wies das Naturrecht als metaphysisch zurück und entwickelte die Idee des Vernunftrechts…In einer kantischen Position wird der Rationalität die praktische Vernunft gegenübergestellt, die ein allgemeingültiges Motiv moralischen Handelns beinhaltet.

 

CC-BY-NC Science Surf accessed 06.11.2025

Fortunately Galactica is down

I just started a review of Galactica.

but today the search bar is already gone. So what happened here? cnet knows more

Galactica is an artificial intelligence developed by Meta AI (formerly known as Facebook Artificial Intelligence Research) with the intention of using machine learning to “organize science.” It’s caused a bit of a stir since a demo version was released online last week, with critics suggesting it produced pseudoscience, was overhyped and not ready for public use.

There is no need to make any further comment.

https://twitter.com/JoeBHakim/status/1592621465018720256?s=20&t=LrwiBQLg4qK_zBUofTRXPA

Claims
Vitriol

 

CC-BY-NC Science Surf accessed 06.11.2025

Une grenouille vit un Bœuf

Jean de la Fontaine

Ein Frosch sah einen Ochsen gehen.
Une grenouille vit un Bœuf

Wie stattlich war der anzusehen!
Qui lui sembla de belle taille.

Er, der nicht größer als ein Ei, war neidisch drauf,
Elle qui n’était pas grosse en tout comme un œuf,

Er spreizt sich, bläht mit Macht sich auf,
Envieuse s’étend, et s’enfle, et se travaille

Um gleich zu sein dem großen Tier,
Pour égaler l’animal en grosseur,

Und rief: »Ihr Brüder achtet und vergleicht!
Disant : « Regardez bien, ma sœur,

Wie, bin ich nun so weit? Ach, sagt es mir!« –
Est-ce assez ? dites-moi : n’y suis-je point encore ?

»Nein!« – »Aber jetzt?« – »Was denkst du dir!« –
— Nenni. — M’y voici donc ? — Point du tout. — M’y voilà ?

»Und jetzt?« – »Noch lange nicht erreicht!« –
— Vous n’en approchez point. » La chétive pécore »

Das Fröschlein hat sich furchtbar aufgeblasen,
Es platzte und verschied im grünen Rasen.
S’enfla si bien qu’elle creva.

Le monde est plein de gens qui ne sont pas plus sages :
Tout Bourgeois veut bâtir comme les grands Seigneurs,
Tout petit Prince a des Ambassadeurs,
Tout Marquis veut avoir des Pages.

 

 

 

CC-BY-NC Science Surf accessed 06.11.2025

At least one positive advice?

Is there at least one positive advice of all that Musk utterance?

https://archive.ph/TcxaF#selection-1319.0-2987.210
  1.  Avoid large meetings
    Large meetings waste valuable time and energy – They discourage debate – People are more guarded than open – There’s not enough time for everyone to contribute. Don’t schedule large meetings unless you’re certain they provide value to everyone.
  2. Leave a meeting if you’re not contributing
    If a meeting doesn’t require your: – Input – Value – Decisions Your presence is useless. It’s not rude to leave a meeting. But it’s rude to waste people’s time.
  3. Forget the chain of command
    Communicate with colleagues directly. Not through supervisors or managers. Fast communicators make fast decisions. Fast decisions = competitive advantage.
  4.  Be clear, not clever
    Avoid nonsense words and technical jargon. It slows down communication. Choose words that are: – Concise – To the point – Easy to understand Don’t sound smart. Be efficient.
  5. Ditch frequent meetings
    There’s no better way to waste everyone’s time. Use meetings to: – Collaborate – Attack issues head-on – Solve urgent problems But once you resolve the issue, frequent meetings are no longer necessary. You can resolve most issues without a meeting. Instead of meetings: – Send a text – Send an email – Communicate on a discord or slack channel Don’t interrupt your team’s workflow if it’s unnecessary.
  6. Use common sense if a company rule doesn’t make sense
    Contribute to progress – Apply to your specific situation Avoid following the rule with your eyes closed. Don’t follow rules. Follow principles.

Warning : 2) and even other x) only suitable for slightly autistic, self obsessed and excellent scientists.

 

 

 

CC-BY-NC Science Surf accessed 06.11.2025

Brave new AI science world

AI works well in some areas while results are very mixed or even dangerous in others [1, 2].

But what about AI in generic scientific areas?

I have recently looked at scite.ai, a service that claims to analyze a paper if the references have been supported or contrasted the findings — any references that have been retracted in the meantime or received an expression of concern? While the latter isn’t really a big question, I was not very much impressed by the results as most references could not be classified.

But there are now more companies off the ground. Let’s have a look at some.

One is SciScore

16/11/2022 screenshot https://sciscore.com/index.php#pricing

SciScore claims

SciScore generates three reports and a score for every submission. These materials assist researchers, editors, and funders in improving the quality and reliability of scientific research by automatically reporting detected criteria of interest for future review. Researchers signing up with their ORCID get ten free reports. Reports can take as little as 1 minute!

The next one is Prophy, another reference finder.

16/11/2022 screenshot https://www.prophy.science/

and there is Galactica.

16/11/2022 screenshot https://galactica.org/

and well also copy.ai

28/12/2022 Screenshot https://copy.ai

I have no idea if we really need this kind of stuff – any AI to write my blog even faster?

Also others commented earlier this month that it is “not obviously wrong but wrong in subtle ways”.

28/12/22 Screenshot https://mastodon.social/@tommorris/109460645976102033

Fortunately Copyleaks released now even an AI detector just to fight back…

 

CC-BY-NC Science Surf accessed 06.11.2025

No irony allowed

Ronagh & Souder in “The ethics of ironic science

We adopt the concept of irony from the fields of literary and rhetorical criticism to detect, characterize, and analyze the interpretations in the more than 60 published research papers that cite an instance of ironic science. We find a variety of interpretations: some citing authors interpret the research as valid and accept it, some contradict or reject it, and some acknowledge its ironic nature.

bonus example 61 “Effect on human longevity of added dietary chocolate

 

CC-BY-NC Science Surf accessed 06.11.2025

Does identification of misconduct in studies affect medical guidelines?

This question has been answered by an earlier study of Avenell et al.

By 2016 the affected trial reports were cited in 1158 publications, including 68 systematic reviews, meta-analyses, narrative reviews, guidelines and clinical trials. We judged that 13 guidelines, systematic or other reviews would likely change their findings if the affected trial reports were removed, and in another eight it was unclear if findings would change. By 2018, only one of the 68 citing publications, a systematic review, appeared to have undertaken a reassessment, which led to a correction.
We found evidence that this group of affected trial reports distorted the evidence base. Correction of these distortions is slow, uncoordinated and inconsistent. Unless there is a rapid, systematic, coordinated approach by bibliographic databases, authors, journals and publishers to mitigate the impact of known cases of research misconduct, patients, other researchers and their funders may continue to be adversely affected.

 

CC-BY-NC Science Surf accessed 06.11.2025