Is science a legitimate source of knowledge? (en)

Have you ever been in an argument with someone only to hear them talking about a study they have read recently to prove their point? The answer is probably “yes”. It is ubiquitous.

Wherever I go I hear people doing it. They report on some scientific study they heard about in a podcast, learned about in University, or even have conducted themselves. May it be the guy in the sauna trying to hit on a girl quoting Andrew Huberman, your aunt reciting a study she learned about in her psychology studies in the 80s during a family discussion, or any coach or writer on social media (or Medium) trying to grasp your attention with “scientifically proven” hacks for a better life. If you pay attention to this, you will see how widespread a phenomenon it is for people to use science to claim argumentative authority, not only among scientists or intellectuals.

The more I noticed people doing this, the more I wondered:“ Is science really as legitimate as a source as everybody makes it to be?”

Luckily, over the last 5 years of studying in an interdisciplinary, metascientific program I had some time to study ‘science’ and ponder this question. To probably many readers’ surprise I now have to answer this question with a clear “no”.

Here are some key problems with scientific studies and how they are dealt with:

Studies can be faked

Even though nobody wants to assume that a scientist faked his study’s data, there are enough examples in history of this happening. The current structure of the scientific system makes the scientist dependent on gaining (external) funding and producing eye-catching results to ensure further monetary support and prestige. If conducted correctly, studies without any result would have to be published the same way, a successful study is, yet for real-world reasons there is a social pressure to rather not publish unsuccessful studies or to touch up one’s data for a more impactful publication. This is known to anyone working in science, and even though there are attempts to mitigate this problem (e.g. to demand public access to the raw data) there is no method to surely prevent fraud in science.

Another problem is that due to its authority in our society, there is an incentive to produce studies that support one’s own political opinions. For example, think of an oil company financing studies on the environmental impact of mineral oil-based products. You wouldn’t be surprised if such studies’ results are in favor of the company's business interests. This example is obvious but the same dynamic is at work when geneticists publish studies that indicate that everything we do is rooted in our genes, or when a behavioral scientist finds the solution for depression in a behavioral intervention. This process is often even unconscious and a common fallacy in all humans (confirmation bias). Yet to produce proper science, it would be paramount for a scientist to reflect upon his convictions and eradicate his bias as much as possible. Be sure that not because someone works as a scientist this means that he is an enlightened being without any bias or that he necessarily is very reflected at all. For the most part, scientists are just people like everyone else.

Thus, if you blindly believe in a scientific study, you are always in danger of simply believing in an intentional or unintentional lie.

Studies can be flawed

Even if a study is conducted by a morally impeccable researcher, there is always the possibility that unintentional mistakes happen, may it be during the conduction of the experiment (e.g. unintentional confounding factors) or in the subsequent calculations and data analysis. If you try to conduct studies yourself you will be able to tell a thing or two about this problem.

Current procedures might be disproved in the future

Even if a study would be flawlessly conducted by an ideal researcher adhering to the state-of-the-art methodology, in science, we always have the problem that our procedures change and improve over time. This means, that even our most accepted procedures are, even at best, only the best procedures of our time. They are never perfect in an absolute way. And there is no way to prove that they won’t be revised in the future. In the same way, as we wonder how doctors could administer mercury, a poisonous substance, as a means to cure syphilis in the past, future generations might wonder for example how we could be so simple-minded to use poisonous substances and radiation to treat cancer. Scientific procedures are never perfect and we will never know in which way they are flawed as we have no access to the all-knowing crystal ball to look into the future. To blindly believe in a scientific study today is as dangerous as it was to blindly believe in the ‘healing’ procedures of a medieval physician.

Scientific precision and applicability to life are mutually exclusive

We now come to a more intricate and less obvious problem of the scientific knowledge generation. In general, we can say that science aims to establish cause-effect relationships by observing and conducting experiments. This means that we want to know what happens if we do X (for example whether we become obese when we eat a specific food, let's say hazelnuts). We want to establish a relationship between our intervention H and the effect O. Thus, we would conduct an experiment in which we feed hazelnuts to humans over a period of time and observe what happens to them. To be able to establish the relationship, firstly, we would have to ensure that there are no confounding factors messing up our results, for example, that our participants go to eat at McDonald’s every time after being in the laboratory for the hazelnut administration and become obese because of it. Secondly, we would have to include enough people to not have our results skewed by the specific conditions of individual people, e.g. people who die of hazelnuts due to allergies.

Now we run into the problem of precision and relevance of our results.For the first aspect, it becomes clear that to establish any relationship at all we would have to prevent any confounding factors as much as possible, which means we have to make the setting of the experiment as controlled as possible (e.g. having the participants stay in the experimental room over the whole time of the experiment so that they don’t do anything unintentional). But by doing that, we create an environment that is completely artificial and different from the real world in which we want to apply the knowledge that we gained through the experiment. So we run into the problem that the more precise and unambiguous the results are (the fewer variables, the more controlled an experiment is), the less relevant they are for real life in which countless factors interact. In real life, there are an unknowable number of variables that are impossible to simulate.

For the second aspect, the surer we want to be that our results are sound and not affected by a special, individual condition (like a nut allergy), the more people we have to include in our study. But the more people we include in the study the more general the results become and the less relevant they are for an individual person with individual predispositions and lifestyle. What does it help a person with a nut allergy if researchers ‘find out’ that hazelnuts are generally healthy (well, not for this person…)? Again the more precise and certain our results are, the less relevant they are for the application of an individual human being.

The dynamics I described here constitute a universal tension field that everybody designing a study finds himself in. To design a study, you have to make a decision, a compromise between precision and unambiguity on the one hand and realism and applicability on the other.

Because of this dynamic, using scientific methods, it is impossible to establish unquestionable cause-and-effect relationships that are relevant in real life (!).

To blindly believe science or to use it as orientation in one’s individual life is to misunderstand the limitations of the scientific method and the meaningfulness of its results.

Correlation is not causation

In truth, the problem of science to establish cause-and-effect relationships (that which we usually call “knowledge”) runs even deeper. Even if we were able to conduct a perfect experiment with no confounding variables that is still applicable to the real world, what we are actually establishing are only ever correlative relationships, never causal ones. We can see that variable X and effect E correlate with each other (they occur together) but we do not know whether X is the cause of E. That is because in a scientific experiment, we only ever observe phenomena. From these, we try to infer the underlying causal processes, but we have no access to these underlying processes directly. We only study their effects, their expressions in reality, in the same way we see the exterior of a pocket watch but not its internal mechanisms. We have no access to the interior of the watch (the causal processes of reality), but we can only cook up theories about its inner workings, what causes the watch (reality) to behave the way we observe.

In statistics, it is well-known and taught that correlations do not necessitate causation. Yet in practice, this fundamental limitation is usually ignored in its significance as it fundamentally questions the value of any scientific study at all. This problem is addressed and discussed in statistics and taught in University, but in practice, it is necessarily neglected and ‘swept over’ because of a mentality of: “What else shall we do?”

It is a funny thing to consider that every (!) scientific fact or law you are sure of, even something as fundamental as that gravitation pulls you to the earth, is strictly speaking nothing but an assumption. We do not know anything for certain about causal relationships in life!

If you now feel a sense of losing the ground under your feet then this is a good sign that a deep-rooted illusion of yours is crumbling, making you more aware of the uncertainty and openness that really characterizes reality. You got one step closer to the truth.

There is no proof in science

Related to this problem of the fundamental opacity of reality and its causal workings is the fact that strictly speaking there is no proof in science!

Science cannot prove the trueness of a theory but it can only disprove it. This is because we never know what we do not know. We are ignorant of the limits of our own knowledge. Thus, even though our theory might work well to some extent, we can never be sure that we have not completely missed the point of the actual underlying mechanisms. The change and improvement of scientific theories over time illustrate this fact very clearly. What we can be certain of is simply that every theory we have, every theory we ever had and will ever have in the future is necessarily wrong. It will never encapsulate true reality, or at least we will never know if it really does.

Because of this falsification dynamic, progress in science is actually not an accumulation of (positive) facts that we are certain of but only an ongoing refutation of alleged facts and theories. There is no increasing (positive) knowledge in science but rather an increasing openness, an increasing unknowingness!

Blindly believing a scientific theory’s truth is therefore a certain way to be wrong. It only shows that one has not grasped the way scientific knowledge production works (as of now).

[There are even deeper problems with science’s basic assumptions and goals (i.e. to establish cause-and-effect relationships), which I will discuss in future posts. Just to note a few examples you could think of here: The general idea of knowledge as a linguistic, symbolic pattern, science’s mechanistic ontological assumption, or science’s mistrust of the subjective. In this post, I intend to highlight the problems of science that are inherent to its definition and its handling in society. I want to outline the problems and characteristics that every scientist should be able to agree to. So in this post, I refrain from discussing these other issues that burst the basic frame of science itself. If you are curious about these issues stay tuned!]

Science is irrelevant to living an individual life

I want to highlight again a point already touched upon above in the discussion of the exclusion problem of precision and applicability, i.e. the fact that we (must) live life as individuals, whereas science can only make statements about reality using averages of big numbers. The problem is that what can be good for one person might be bad for another (as in the example of the nut allergy). So, the generalistic knowledge of science would be useful only as long as it matches our individual experience, in which case it does not provide us with any new information. If one truly trusts oneself, then there is nothing of interest that science could tell this person.

The reason for science to be bound to study averages (as in correlations) is its inherent mistrust of the subjective and any subjective access to truth. It does not trust the individual, the subjective experience, e.g. one’s report that he saw a dragon, which is understandable and makes sense as we know from our everyday experience that people can err. Science tries to mitigate this problem of credibility by comparing a high number of participants, experiments, etc. so that individual errors’ effects are neglected through the large number of data points in the statistical analysis. It tries to achieve objective (unbiased) truths by summing up many subjective perspectives together. This may sound like a great strategy but in the end, it is nothing but a sleight of hand. There is no objective perspective. We don’t even know what an objective perspective should be like as the only perspective we know and ever experience is a subjective one, i.e. our own perspective. The difficulty here is that even though we know that we as individuals can err, we have no choice but to trust ourselves if we want to live our lives in a meaningful way and not only according to the calculated average of everybody else which science provides us (which would lead to complete sameness of everyone's life in the end). The ironic punch line of all of this is that it seems that the only real way to access objective truth that we as humans have, is through the individual, through deep immersion into the subjective. This possibility is exemplified by the reports of people who have undergone what is called a “mystical experience”, an experience in which they had direct, unbiased access to the absolute truth of reality (or of God as some say). Science does not suspect that this is a possibility (and certainly not that it is the only one) and builds essentially on its mistrust of the subjective. We can therefore describe science as an endeavor to attain truth which is only sensible for the ordinary mind of people who have not yet had a direct experience of truth (aka a mystical experience). Science is a project only interesting for people who do not know the truth. For the ones who do, they know that there is no certainty to be gained by mistrusting the subjective access, by using scientific methods. Surprising, isn’t it?

Science cannot give direction

As science can only study averages and only describe observations, it cannot tell you what you should do with the ‘knowledge’ that it provides. It cannot give you a normative orientation. For instance, scientific studies might indicate that potassium cyanide can kill a human but they cannot tell you whether you should use the poison to kill the dictator of your country or any person who drives a car and therefore damages the environment. It can neither tell you what profession you should choose, nor what food you should eat, nor which life partner you should choose. It can not tell you who you are or how you should behave. It cannot give you direction.

To be able to live one’s life and to find the answers to those questions one must look into oneself, one must connect to one’s self. Science cannot study the subjective, it does not have access to your internal workings, to your soul. Therefore, it cannot tell you who you are and it cannot help you with how you should live your life.

A person who is in contact with himself, a person who is who he is, knows this and will see the uselessness of scientific ‘knowledge’ for any important decision in life. If you turn to science for answers to what you should eat, what you should study, or how you should behave to be attractive to others for example, then this is a certain way to be disappointed in the end and a clear signal of your own disconnection to your soul.

If you start to observe how many people do this, how many people look to science for orientation, you can get a glimpse of how disconnected people in our society are from themselves, and how deeply our scientific culture separates us from our souls.

Additional notes

1.) None of these problems are solved by Artificial Intelligence (AI) as some might argue now. AI is a product of the scientific worldview and scientific methods. Therefore, it is limited by the same shortcomings as science itself. AI does not have subjectivity, it does not have human-like consciousness or qualitative experience. It is simply a complex array of statistical calculations. It is quantitative-mathematical. It is still just a big soup of 0s and 1s. It does not have subjective access to truth. It can only achieve useful functionality by the orientation and qualitative input we humans provide and by statistical calculations, i.e. by averaging over large amounts of data. That’s also the reason why all of AI’s creations lack real originality.

2.) I purposely didn’t include all the socio-economic, and psychological factors that are big problems of science in reality as well (financing and financial incentives, ego/status issues, publication bias/echo chambers, peer pressure, political pressure, juristical restrictions, etc.), which even worsen the overall trustworthiness of scientific ‘evidence’ but might be solved in the future as they are not inherently bound to science’s fundamental axioms and structure.

Conclusion

When we put all of this together we can see that science is not the holy truth-spreading institution that it is treated to be. Rather, by treating it this way, we find ourselves blinded by a false and misleading notion of security (the security of scientific ‘facts’. The ideological way of understanding science as the only source of true knowledge is in truth unscientific and exemplifies a blatant contradiction to what science actually is and what it is meant to be, i.e. a method to disprove wrong statements about reality by testing them against empirical evidence. Science was meant to fight superstition and ideology. It was meant to create a “bullshit resistance” (as my friend David calls it), not to become an ideology itself. With the way we, in our culture, treat science as the only epistemological authority, we abuse it to its core.

Conclusively, we see that science is neither a legitimate source of argumentative authority in a discussion nor is it a sensible orientation for how to live your life.

Now that we have understood these conditions, the question remains: What do we do with this? How can we appropriately handle science?

Implications

I am not sure whether there are definite answers to the questions above. Nevertheless, I will try to give it a shot and collect some implications I draw from this:

• Don’t trust anyone trying to convince you to believe or do something using scientific ‘evidence’ (it is dishonest).

• Scientific ‘facts’ and theories are really metaphors. Use them flexibly and playfully wherever it deems useful for your life. Don’t mistake them as truth.

• Scientific theories must always be checked against one’s own experience. Science builds from subjective experience, not vice versa. Science cannot negate it. Subjective experience is always superior in its status to scientific implications. Your individual experience and process of understanding is what (matters for your life) in the end.

• Orientation in life comes from subjective INsight (looking INwards). Only you yourself can find out what you want to do with this life and what is the right thing to do FOR YOU. In general, there is no outer guideline, no blueprint to be found externally, and definitely, it is not found in science.

• Politics needs all its stakeholders (that is every one of us) to explore what they really want in life. We must explore what we want as a human community. Only from there can we go on creating that vision in reality, individually and thus collectively.

Now, I am curious to hear what implications you are drawing from this. Put it in the comments!