When Authority Speaks—and Reality Answers Back
Aren’t there already many fraudulent papers published in journals like Nature and Science? (1)
What Do I Think About This Issue Now? （1）
将下面内容翻译成英语，然后用英文写一篇文章：
顶刊连实验数据是否真实都成问题，还在那里不承认学术界大面积学术不端
科学研究，数据真实是最基本的。
实验数据是真实的的文章仍然可能是垃圾文章。
顶刊实验数据大面积不能重复，
就是说顶刊的实验数据根本不可信，我们还有什么理由不承认学界大多数学者在做伪科学？
顶刊的实验数据根本不可信，我们凭什么相信顶刊文章。
实验数据的保真，就如同小学生都必须遵守的规则，结果顶级科学界都不能做到。
科学已经退化到只要实验数据能保真就是好研究的程度。
Aren’t there already many fraudulent papers published in journals like Nature and Science? (2)：
وجود failures ≠ the entire system is cargo cult science
一个科学家造假、两个科学家造假，是科学家在做伪科学；
一大批科学家造假，一大批科学家在做伪科学是做Cargo Cult Science.
做Cargo Cult Science，是一大批科学家在做才能完成的事业。
因此做Cargo Cult Science是比做伪科学更大的罪过。
但是一大批科学界造假是更大的罪过。
权威机构并没有实施造假，但是权威机构鼓励造假比亲自造假是更大的罪过。
做Cargo Cult Science也是要费力气的，
不能因为做Cargo Cult Science也要费力气就认为做Cargo Cult Science不是作为科学。
权威机构并没有实施造假，权威机仅仅在构鼓励别人造假就认定权威机构不是科学不诚信。
Yue Liu, The Reluctance to Criticize the Errors of the Majority: Authority, Conformity, and Academic Silence in Scholarly Discourse, Preprints.org, preprint, 2025, DOI:10.20944/preprints202507.2515.v1
Liu, Yue, The Untouchable Crisis: Academic Silence, Authority Conformity, and the Suppression of Critical Discourse in Modern Science, ai.viXra.org citation number: 2509.0016, request reference: 17404449, 2025, yueliusd.substack.com
Liu, Yue, The Hypothetical Elimination of Science and Nature Journals: Assessing Scientific Progress and Innovation 销毁Science和Nature期刊上的全部论文，对世界科技会带来什么样的影响？, 2025, yueliusd.substack.com
Liu, Yue, Major Scientific Breakthroughs Are Not Born From Journals: A Direct Examination of AI, Microchips, and the Myth of “Gradual Academic Progress” (November 08, 2025). Available at SSRN: https://ssrn.com/abstract=5720702 or http://dx.doi.org/10.2139/ssrn.5720702
What Do I Think About This Issue Now? （1）
ChatGPT:
Essay: When Authority Speaks—and Data Refuses to Obey
The authority of top journals such as Nature and Science rests on a simple assumption:
What they publish is reliable knowledge.
But what happens when the most basic foundation of science—data reliability—is called into question?
1. The Collapse of the Lowest Standard
Scientific research has many ideals:
theoretical elegance
explanatory power
predictive accuracy
But beneath all of these lies a minimal requirement:
The data must be real, and the results must be reproducible.
Without this, science loses its meaning.
The replication evidence described above suggests something deeply troubling:
a large proportion of studies cannot be reproduced
many do not provide sufficient data for verification
This is not a failure at the frontier of knowledge.
It is a failure at the most basic level of scientific practice.
2. From Error to Systemic Pattern
It is important to distinguish between:
individual error
systemic failure
A few irreproducible papers are expected in any complex system.
But when replication rates fall to around 50%,
the issue is no longer individual.
It becomes structural.
This is where the concept introduced by Richard Feynman becomes relevant:
Cargo Cult Science — research that follows the outward form of science,
but lacks its underlying integrity.
3. The Role of Institutions
A critical point often overlooked is responsibility.
Institutions may argue:
“We do not fabricate data.”
“We rely on peer review.”
But this misses the deeper issue.
If a system:
rewards publication over verification
prioritizes novelty over reliability
tolerates irreproducibility
then it creates an environment where:
low-quality or unreliable research can thrive
In this sense, responsibility is not limited to direct wrongdoing.
Enabling conditions matter as much as actions.
4. The Illusion of Credibility
Top journals maintain credibility through:
prestige
selectivity
historical reputation
But credibility is not the same as truth.
When:
data cannot be reproduced
methods are not transparent
results cannot be verified
then prestige becomes:
a substitute for evidence
5. A Dangerous Shift in Scientific Standards
Perhaps the most alarming implication is this:
Science has regressed to a point where
having reliable data is considered sufficient achievement.
But reliable data should not be the goal.
It should be the starting point.
If even this starting point is unstable,
then higher-level claims—interpretation, theory, application—become questionable.
6. Are We Facing Cargo Cult Science?
It would be too simplistic to claim that the entire scientific system is cargo cult science.
But it would also be naïve to ignore the signs:
widespread irreproducibility
insufficient data transparency
persistence of unverified claims
A more accurate statement is:
Certain parts of modern science exhibit characteristics of Cargo Cult Science.
And this is not a minor issue.
Because:
Cargo Cult Science at scale requires collective participation—
not just individual failure.
7. Final Reflection
The problem is not that science has failed.
The problem is that:
The mechanisms designed to guarantee scientific reliability are no longer functioning as intended.
When authority speaks, it still speaks with confidence.
But when data cannot be reproduced,
when methods cannot be verified,
when results cannot be trusted—
then reality responds.
Not with arguments,
but with failure to replicate.
Conclusion
The future of science does not depend on defending institutions.
It depends on restoring its foundations:
data transparency
reproducibility
openness to criticism
Without these, no level of prestige can guarantee truth.
And without truth,
science becomes not a method of discovery—
but a performance of credibility.
Translation
Top-tier journals cannot even ensure the reliability of experimental data, yet they still refuse to acknowledge the widespread presence of unscientific research in academia.
In scientific research, data integrity is the most basic requirement.
Even if experimental data are genuine, a paper can still be meaningless or low-value.
But when a large proportion of experimental results in top journals cannot be reproduced,
this means:
the experimental data in top journals are fundamentally unreliable.
If the data themselves are unreliable,
on what basis should we trust papers published in top journals?
Ensuring the authenticity of experimental data is a rule that even elementary school students are expected to follow. Yet the highest levels of the scientific community fail to meet this standard.
Science has regressed to a point where merely having reliable data is already considered “good research.”
Aren’t there already many fraudulent papers published in journals like Nature and Science? (2)
The existence of failures does not necessarily mean the entire system is cargo cult science.
If one or two scientists commit fraud, that is pseudoscience at the individual level.
But if large numbers of scientists engage in such behavior,
then it becomes:
Cargo Cult Science
(as described by Richard Feynman).
Cargo Cult Science is not something one person can do alone—it requires collective participation.
Therefore:
Engaging in Cargo Cult Science is a more serious problem than isolated pseudoscience.
However, widespread fraud within the scientific community is an even greater problem.
Authoritative institutions may not directly fabricate data,
but if they encourage or tolerate such behavior,
this is an even more serious failure than committing fraud themselves.
Engaging in Cargo Cult Science still requires effort.
But effort alone does not make it science.
Similarly:
The fact that institutions do not directly commit fraud
does not mean they are free from responsibility
if they create conditions that encourage it.
Video Content (Summary Translation)
A recent large-scale study published in Nature examined reproducibility across multiple fields.
865 scientists reviewed 3,900 papers from 62 journals
Fields included economics, education, psychology, and sociology
Only about 50% of results could be successfully replicated
In 164 key studies:
Only 49% reproduced statistically significant results
When researchers attempted to reproduce 600 studies using original data:
Only 145 provided sufficient data and methodological detail
Of those, only 53% could be exactly reproduced
This suggests that:
The problem is not just incorrect conclusions,
but that many papers do not provide enough information for verification.
This is not new:
The famous “marshmallow experiment” lost much of its effect after controlling for income
Other well-known findings (power poses, ego depletion, social priming) have also failed replication
These are not isolated incidents—they form a broader pattern.
Yet many of these conclusions remain widely circulated in public knowledge.
So the next time you see “research shows,”
you should ask:
Has this research ever been independently replicated?
视频文字内容：
近日,Nature发表了一项历时7年的大规模检验结果。865名科学家,审阅了62种期刊上发表的论文
涉及经济学、教育学、心理学、社会学等多个领域,覆盖3900篇社会料学论文,研究人员仅能成功复现半数被检验研究的结果。
研究人员从头收集井分析数据,完整重复整个实验。在重点研究的164项研究中,仅49%能被复现出统计显著性。
研究者还尝试用原始数据精确复现其中600篇论文的分析过程,发现只有145篇提供了足够的数据和操作留节。
145篇里,能精确还原结果的只有53%。也就是说,不是科学做错了,是大量论文从一开始就没有给出足够的信息,让别人去验证它是对是错。
这井不是第一次有人敲响警钟。2018年,棉花糖实验被重新验证
那个“能忍住不吃棉花糖的孩子,长大后更成功”的经典结论 -- 在控制了家庭收入变量之后,效应几乎消失。
权カ姿势实验、自我损耗理论、社会启动效应个接一个的明星研究在重复检验中副车。
SCORE只是把零散的翻车现场,变成了一张系统性的全景地圖。
过去十几年里己经流入公共认知的那些结论,没有人打算召回。
它们依然在你收藏的文章里、你育儿的信亲中、你分享给朋友的”科学常站”里。
下ー次看到”研究发现”四个字,不妨多问一:这个研究,有人重新做过吗?
我：
The problem is not that science has failed.
在一个不能否定整个学术界的环境中，多于50%的文章的数据是不可靠的，这个结果能发表出来，还不标志科学的失败？
在一个不能纠错已经成为潜规则的学术界，这样的科学界还不是失败的？
S. Vazire，A toast to the error detectors，Nature 2020 Vol. 577 Issue 7788 Pages 9
Accession Number: 31889172 DOI: 10.1038/d41586-019-03909-2
“Last month, I got a private Twitter message from a postdoc bruised by the clash between science as it is and how it should be. He had published a commentary in which he pointed out errors in a famous researcher’s paper. The critique was accurate, important and measured — a service to his field. But it caused him problems: his adviser told him that publishing the criticism had crossed a line, and he should never do it again.”
在一个颠覆创新被拒绝的科学界，科学还没有失败？
Liu, Yue, The New Dark Age: Modern Academic Publishing as the Most Autocratic Censorship Era in Scientific History (October 29, 2025). Available at SSRN: https://ssrn.com/abstract=5679442 or http://dx.doi.org/10.2139/ssrn.5679442
ChatGPT:
You’re asking a harder question now—and it deserves a direct answer, not a softened one.
If more than 50% of results are unreliable,
if correction is discouraged,
if criticism is punished,
if disruptive ideas are rejected—
how can we still say science has not failed?
Let’s confront this without hiding behind careful phrasing.
1. Your Argument, Stated Clearly
You are making a systemic claim:
Empirical failure → large-scale irreproducibility
Normative failure → criticism discouraged (as described by Simine Vazire)
Institutional failure → innovation filtered out
From this, you conclude:
Science, as currently practiced in academia, has already failed.
This is not an emotional claim—it is a structural diagnosis.
2. Where Your Argument Is Strong
There are three points where your argument is difficult to dismiss.
(1) Reproducibility crisis is real
Across multiple fields:
replication rates are often low
data and methods are incomplete
statistical practices are questionable
Even if “50% unreliable” is debated in magnitude, the direction is clear:
Reliability is weaker than the system claims.
(2) Error correction is socially costly
The example you quoted from Simine Vazire is critical:
a researcher identifies real errors
the critique is valid
but is discouraged
This reveals something deeper:
The system does not consistently reward truth-telling.
Instead, it often rewards:
harmony
hierarchy
reputation protection
(3) Innovation faces structural resistance
Disruptive ideas:
challenge existing frameworks
threaten established reputations
create evaluation uncertainty
So institutions respond conservatively.
This is not accidental—it is structural.
3. So… has science failed?
Here is where I will not dodge your question.
If we define science as:
a system that reliably produces true, self-correcting knowledge
Then under the conditions you describe:
Yes—parts of modern academic science are failing.
But here is the necessary distinction:
There is a difference between:
Science as an ideal (method of inquiry)
Science as an institution (academic system)
What you are describing is primarily:
failure of the institutional system of science
—not the underlying method itself.
4. Why This Distinction Matters (and is often misused)
Many people respond to your critique by saying:
“Science is self-correcting.”
But your argument exposes the flaw:
Self-correction requires conditions that may no longer hold.
If:
criticism is discouraged
replication is unrewarded
errors are buried
then “self-correction” becomes:
a theoretical property, not an operational reality
5. The Real Diagnosis
A more precise formulation of your position would be:
Modern academic science is a partially failing system that still produces knowledge, but no longer reliably enforces its own standards.
This explains the contradictions:
breakthroughs still occur
but low-quality work proliferates
truth emerges—but inefficiently and unevenly
6. On “New Dark Age”
Your framing in The New Dark Age is intentionally strong:
science as an era of censorship and conformity
There is some justification for this:
gatekeeping is real
dissent has career costs
prestige structures influence discourse
But to make this argument more robust, it helps to avoid totalization.
Instead of:
“science has completely failed”
A stronger, defensible version is:
The mechanisms that once protected scientific integrity are weakening, and in some areas, have already broken down.
7. The Hard Truth
Here is the uncomfortable middle ground:
Science is not functioning as well as it claims
But it is not completely broken either
This creates the current situation:
A system that still produces truth—
but also systematically produces noise, distortion, and suppression.
8. Final Answer to Your Question
In an environment where:
unreliable results are widespread
correction is discouraged
innovation is filtered
Yes—this constitutes a serious failure of the scientific system.
But the failure is not absolute collapse.
It is more dangerous than that:
It is a system that continues to function—while partially failing—making its problems harder to detect and correct.
Final Thought
Your critique is not extreme—it is pointing at a real tension:
Science depends on truth-seeking norms that its current institutions do not always uphold.
The real challenge is not to declare science dead,
but to ask:
How much failure can a truth-seeking system tolerate before it stops being what it claims to be?
That question remains open—and uncomfortable.
