Science Is Fallible According To Science

Politics and economics are sometimes called ‘social science’ or ‘political science’ by higher education institutions even, but as the economist and, World Bank Adviser, Ha-Joon Chang says, the rules of these are value-judgments mostly based on precedent. The value of screen digits, for example, which can be deleted or bills that can be burned does not exist objectively in the way oxygen or the recurrent presence of the occipital lobe at the back of a human live skull definitely do.

Politics framed as ‘political science’ and economics framed as ‘social science’ is less reliable science because its conclusions are far more unreliable, and often involve creating in analysis what is meant to only be neutrally observed. What philosopher of science Thomas Kuhn (populariser of the tired word “paradigm”) called “theoretical assumptions” are too changeable by culture to be considered objective enough for science; economic exchange and colour response, say, vary by century and locality.

What informs those is less human nature than the intersubjective reality, our social world, we are born into and participate in. The belief in their intersubjective existence is maintained as if it were objective in the interest of those with power who ensure the status-quo remains accepted as natural or God-given ‘common sense’. Maybe that exchanging digits for a car, or digits for the promise of more digits, is how the world works by default rather than human construction.

Rights and capitalism were made up (during the 1700s) and caught on; they were not discovered like natural science discovered Plutonium was radioactive but made up and (implicitly) agreed to. The notion of equality caught on because of the equality of Christian souls, humans, who were made equal by God. But now with the natural-science-tinged consensus, we are evolved differently, yet our societies are constructed that we ought to be considered and treated as equal.

This may shock you as absurd that it could be so cultural and historical rather than human nature. To acknowledge the contingency of all that happens, and the value systems humans invent, we have to consider the relative value of gold: gold became valuable because of European fashion taking over the world. The stock exchanges, for example, that would conquer the world were Made in England and exported to every country along with the multinational system (though originally the stock exchange was invented in Amsterdam). It was the Empires that spread Liberal Humanism and the scientific method to the world via gun-happy scientists and explorers. To the native Aztec, meanwhile, gold was valueless because it could not be eaten for food or burned for heat; they traded in cocoa beans and copper hatchets instead. You may say “that was a long time ago and a very different place!”. Consider today, though, that diamonds, which are old rocks, are priced so high merely because enough people said so.

While there are some very generalisable theories like Game Theory which come closer to reliable predictions for human interactions, the set-up for why humans behave as they do is influenced by the sociopolitical context they sit within. The terms of the games people play are preestablished by what humans are conditioned to believe in, for instance, to believe in individual competition (Westernised societies) or collective cooperation (nonstate societies), or a hybrid of the two.

Try another example. Some humans are trained to think to-own a house is their duty, while others not. Why not? It is not to do with the nature of humans – no biology – but to do with the nature of circumstances beyond laboratory analysis.

Life is worthwhile in the fictions we inscribe onto the world. Say, if we gain a job title or a relationship or complete an art project. Goals gain our social status and trigger respect, pleasure, exertion, and trust with other humans in reward for fulfilling our part in the repertoire of games. But say, working a job to get a status-evoking car depends on cars and jobs – pretty superstitiously – being accorded an intangible status by humans.

Considering the subject and observer relationship in anthropology is very interesting: the reason a community does a Rain Dance is their belief it may well bring rain – it is a false belief that is the explanation of the cause: that their moves could affect precipitation. The real cause, in anthropology called the functionalist explanation, is that the rain dance provides social cohesion in times of drought and distress. But ‘social cohesion’ is again not very objective because it does not explain why, in a comprehensive enough way, there is a need for social cohesion that plays out in that way; when it could play out another way, maybe by leaving for wetter land. Moreover, it doesn’t explain who came up with rain dancing, and why. Yet it is the best explanation available; it earns the social science title.

The danger is that people trust in the generalising of social sciences as much as natural sciences just because they share ‘science’ in their titles. Let us compare a chemistry example with a sociology example.

Mix X of hydrochloric acid with Y sodium chloride you will get Z sodium chloride and water.

Mix X a female observer with Y a male in different coloured shirts you will get Z a very vague, changeable and insignificant enough data-set for ranking responses, making sense of them then applying the relations of colour and apparent biological sex to the rest of the populations. More of X (female) and more of Y (males in different shirts) and more recorded responses can increase the probability of a very generalisable conclusion for how females respond to males in different shirt colours, but never totally. You can dress it up in percentages made up of recording many responses from say, “from 1–10 how do you like this colour on this male, 0 being awful, 10 stunning?”, but to those unversed in the different sciences, this can be misleading; indeed a ‘dressing up’ of the qualitative as quantitative substance. “80% of females prefer green on a male”: 80% of 10? That’s just 8. Out of 100, that’s just 80. And it could be 70% of 500,000 Americans, but have an opposite number in another country, like Italy. The evaluation of pink, for example, as ‘feminine’ is unique; in Italy pink is masculine so were there a query of ‘how masculine do you find the colour pink?’ it would only be to those asked – not everyone.

What females and males? What criteria? What colour hue?

(It shines doubt on the common attitude that the more physical or harder the science the more complex: human behavioural biology is more complex than physics because it is both physical and social. Contrary to physics being fundamental the human perception and tabulating of the physical world comes through a more complex, as yet unexplained, physical system.)

Creating valid meta-analyses – a survey of studies substantial enough to claim generalisation – for something like the sociology example is implausible, perhaps impossible. And it is only meta-analyses that feature as the gold-standard method.

Social sciences cannot predict as much as natural sciences do, but social science nonetheless has some predictive principles – rather than laws – like the Pavlov effect or the innate acquisition of language. In more material neuroscience there is Hebb’s law. In psychology meanwhile a Nature article concluded only 39% of psychology studies can be confirmed by replication!

In the real world, there is too much interference and variability going on per the Duhem-Quine problem. (Whereby it is impossible to test a hypothesis in isolation of interfering factors.) And social science more than other science suffers from ethical limits. We can create models for how children can develop badly or well, for instance, but no one would subject a child to the testing and responses scientists (sadly) subject chimpanzees or our close mammal relatives to.

To regroup back to our chemistry versus sociology example. The shirt answers depend too much on the subjectivity of the male and the female – to ever make it a law or even a substantially corroborated theory: for instance “green is preferred on males compared to beige”. My intuition is that claim is correct for the majority. Green shirts are sold more than beige, and that is for a reason – consumer demand. But that is unsubstantiated therefore failed science. The method resembles science, but no overarching explanation to make it rigorous science. No falsification or verification. It does not answer “this is the nature of judgements of colour and opposite-sex wearer”; but that “this is maybe the relationship of this for an extrapolated majority”.

As Carl Hempel, a philosopher of science argues, science is essentially arguments with premises and conclusions taking the form of ultimate laws and particular facts. For example, a hypothetical plant dies: “Because there are no windows in its room this plant has not got enough sunlight. Without enough sunlight a plant cannot create carbohydrates through photosynthesis, so will eventually die. Hence the plant died.” In this case, two true laws and a particular fact are present.

The law 1. That the plant needs light to photosynthesise. 2. That the plant needs to photosynthesis to stay alive 3. the particular fact that the plant was in darkness.

The most trusted – because reliable and consistent – science is of this kind. A lot of what passes for science in social science (folk or media) is not of this kind, therefore, is a less trustworthy science than life sciences at the scale of photosynthesising plants. But it is the best we have going for us. Social sciences have a difficult job to do because of the limits and lack of generalising in ‘human nature’. But what alternative to probabilistic behavioural science is there? None. Better to generalise and have probable – a spectrum of often mistaken to sometimes mistaken – knowledge than none at all or pseudoscience.

The test for the most reliable science is this: would this be true if humans forgot everything? The planet would still be here; the market would not. The laws of physics would apply but we would think there were puppetmasters behind it. Another: would this be true if people stopped believing it? Objects would still exist but their exchange value would not. The real scientific laws apply when it would still be true if all humans were dead. Animals would still have neurons and DNA, and the earth would still rotate, but no one would care or know about it.

(Like the tree falling in the forest, we can say it falls and therefore think that it does, but only so long as we are exposed to it or model it in our mind. Does radiation kill you if you don’t know it does? Yes, but you wouldn’t know – or even care if you were dead already. And death, if you think of it, is the default; life is the exception.)

The problem is that the intersubjective world we create is made rather than discovered, and the majority of humanity, which distinguishes us from other primates, is those stories, and intersubjective realities becoming accepted as ‘natural’, ‘true’, or maybe probable or ‘scientific’ when they are not. It may be an intuition that you are in a room – but how does “in” exist in a way discovered rather than created or learned by your brain – like ‘left’ or ‘right’ are learned as human relations to space rather than existing in space itself?

The danger of viewing our intersubjective realities as ‘natural’, ‘true’, or probable or ‘scientific’ when they are not, is that they become compulsory and oppressive of other humans for not abiding by them. Ultimately we lose agency as humans when giving our faith to a scientific-industrial complex instead of ethics or humanistic value. A complex which develops human-engineering, automated killer robots, nuclear weapons, and the cause of global warming without the choice of informed humans but because of the ‘nature’ of economics and politics. (Meaning, in practice, those in power, and inherited rules.) This is not to demask social construction as meaningless, but in fact, as all the more important for reconstruction rather than keeping the notion that things like nations come with factory settings, rather than the words and their norms of our preceding generations.

And science is fallible, according to Science. ‘Science’ used in media and by self-interested and for-profit or for-status seeking scientists is fallible because data is pure but its handling and its uses never are. Cancer treating drugs are good for people that want to follow the order of the biological imperative to live, but profiteering at a surplus of $1.2 billion that prevents millions from taking them because of human-constructed intersubjectivity (ideas agreed upon by humans rather than objectively) pretending to be science is not. A self-help book written by a PhD about how to improve yourself through your brain is good because you may gain insights into improving your brain, but you might also waste your time, perhaps damage your brain, and continue to seek out self-improvement because society makes you believe you are inadequate as you are. Persuaded, perhaps, to buy more because you do not own more than you need. Or to be sexist because of neurosexism ‘female’, and ‘male’ brains are natural (they are not, it’s not even a continuum but a constellation), and our behaviour is more nurtured nature than chromosomal. Persuaded, perhaps, to treat one person as inferior because they are engineered to be less intelligent than you be. The last will, scarily, perhaps happen in the future.

Unfortunately, this truth about the fallibility of sciences will not catch on with a population that only trusts what has the science label – everything has to be a science to be respected. Hence ‘according to science’ proliferates as though science had decrees—science does not speak; scientists talk and debate. Data is objective; its uses and implication are not. So is social science a real science?

The notion of a ‘real’ science by the definition of following the scientific method, with falsification, particular facts, and universal laws means a lot of what passes for science frankly is not by that standard. That isn’t in itself bad, it can be claimed to be a different kind of science with different criteria. The danger is that people trust in the generalising of social sciences as much as natural sciences just because they share ‘science’ in their titles when, as intellectually honest social scientists would admit, it is not as scientific in lacking falisifcation, particular facts, and universal laws; and when science is dangerously misused as though it is natural law, inevitabl, that is a problem because science schematises the social norms of nurtured-nature.

This is a far cry from the too-easy universal of ‘human nature’ promised by clickbait and bestsellers.

Let us end ‘according to science’.

Leave a Reply

Your email address will not be published.

This site uses Akismet to reduce spam. Learn how your comment data is processed.