Limits of Science

Conversation 26, Socrates Worldview 20/22

CRITOBULUS. What have you got in store for us today, Socrates?

SOCRATES. I plan to draw our whole ‘worldview’ discussion to a close next time, Critobulus. Today, I just want to reinforce some points about the limits of science.

C. I hope there will be a limit to whatever you have to say about the limits of science!

S. Fear not, Critobulus. I know I’ve tested your endurance, on and off the bike, over the last couple of days, but today I intend to go relatively easy on you.

C. Thank goodness for that!

S. Do you remember I said science and religion can coexist within the same rational mind provided each is confined to its proper domain?¹ Well, it was necessary for us to explore the more esoteric boundaries of science before we could complete our discussion of the proper domain of science.

ADEIMANTUS. And we are now ready to do that?

S. I think so. The scope of science is limited in several ways that people often misunderstand. Some people think the scope is bigger than it is, a common mistake in Western societies these days. Others think the scope of science is much less than it really is. Some appreciate that science is limited, but for the wrong reasons. I would like to talk about three types of limits on the scope of science: fundamental, practical, and cultural. Where shall I start?

A. What do you mean by cultural limits? Maybe start there.

S. By cultural limits I mean limits to the scope of science that some people argue for reasons that have nothing to do with the fundamental or practical limits. You have probably guessed that I consider their reasons to be wrong. I refer particularly to the postmodernists, who don’t believe in objective truth. Belief in objective reality and objective truths about that reality are the foundations of science. The postmodernists would have us believe that a scientific theory is just a cultural product of a particular group of people, of no more verity than the creation myths of indigenous peoples. They limit science to being ‘just another story’. This is nonsense. They do it out of envy of the success and prestige of science and to help them win arguments to support their agendas, in other words, in a quest for power.

C. You’re not keen on postmodernism, Socrates?

S. I can’t respect anybody who has no respect for truth, Critobulus. Thomas Kuhn, with his paradigm shifts, may have been an honest researcher, but he played into the hands of the cynics. They were happy to take the idea of a paradigm shift in scientific theories as a sign that a new scientific power group had taken over. This is wrong. Of course, there are groups that argue for different theories, but it is not about power, it is about the weight of evidence. While the weight of evidence is in the balance, different people prefer different positions, but eventually the evidence favours one theory or the other. The cynics tend to overlook the vast amount of patient, grinding, work that scientists do to collect and analyse the data that provide the evidence. Scientists are optimists: they always hope to find some data that will shift the weight of evidence.

A. And you say this is not a cultural phenomenon? How come some cultures do it and others don’t?

S. An Eskimo, a Patagonian, and a Hindu, all applying the scientific method to the same problem, tend to come to the same conclusions in the end. The products of the scientific method have nothing to do with culture, just as blast furnaces produce much the same iron in China as they do in America. What is truly cultural about science, though, is the commitment to the scientific method. As I have said before, the foundations of the scientific method are: belief in an objective reality governed by discoverable laws; the idea of separating the process under consideration from extraneous influences so it can be studied; the formulation of theories about the processes under study, these theories being consistent with all other relevant knowledge; and repeatable consistency with experiment or experience as being the test for the validity or invalidity of theories.

A. So, you saying that not all cultures have a commitment to the foundations of the scientific method?

S. That is precisely what I am saying, Adeimantus. The commitment to the evidence of repeatable experiment as the standard of truth rules some ‘stories’ in, and rules some out. It is not the stories that are the culture of science, as the postmodernists maintain, but that commitment to the scientific process.

A. Is it an accident, then, that science emerged in the Western culture, and not others?

S. I would say that the conditions had to be just right. Others have written much about this topic. First there had to be certain historical prerequisites, like the philosophical thought of the ancient Greeks, although that was available subsequently to cultures in which science did not flourish as well as to Western culture. The Judeo-Christian religious tradition was almost certainly a factor in the West, with its concept of a creator God who establishes and ‘plays by’ the rules, and its respect for objective truth. Later, Christianity gave rise to tolerance for individual though and liberty, which sometimes had to be fought for, and was another factor in the rise of science. Not least, perhaps, was a sufficient prosperity to allow some people the time to study science. All these factors needed to come together.

A. Will the right conditions for science stay together, do you think?

S. It’s not guaranteed, Adeimantus. Under the sustained assault of the cynics the West is losing its belief in objective truth. It is also losing its fervour for individual liberty and freedom of thought. For these cynics, not least the Marxists, the products of science are mere bargaining chips: sometimes they say, ‘the science is settled,’ and other times they say a scientific result is just another story. Science will not survive under these conditions.

C. I thought you said scientists were optimists.

S. Scientists may be optimistic about uncovering the workings of the universe, but they are entitled to be pessimistic about society becoming hostile to the fundamental preconditions for science.

EUTHYDEMUS. Will science ever discover the ‘theory of everything’?

S. I thought you already had the theory of everything, Euthydemus: God did it! But I am being unfair to you, and your question is a good one. I read a report about a meeting where scientists and philosophers discussed the question of whether science is doomed to leave some questions unanswered (Falk 2019). They did not all agree, but some argued that some questions were unanswerable, especially in relation to cosmology, quantum mechanics, and consciousness. I am inclined to agree with their overall conclusion, but not with some of their reasoning.

E. So you think that science will not find the theory of everything?

S. That’s right, Euthydemus.

A. What are your objections, Socrates?

S. They say that we can never understand the cosmos, meaning the universe, as a whole because we are inside it - we can never see it from the outside. They say we can never understand consciousness objectively, because it is a subjective experience. And they say that quantum phenomena, the kinds of things I have been talking about in our last couple of conversations, cannot be studied with complete objectivity because whenever we observe a quantum process, we muck it up. I basically agree with them on the quantum processes because, as I have said previously, we are macroscopic beings – big and made of lots of moving parts. When we interact with a tiny quantum system to observe it, we have to use tools and while the tools might also be tiny, eventually they have to interact with us and that destroys the ‘coherence’ of the quantum states we were trying to observe. If there is a level of quantum reality that I called my little ‘fidget wheels’, we might construct clever models of them, like Bohm did, but we can never see them to test these models. All we can ever observe are probabilities.

A. What about the cosmos?

S. This gets us into a discussion of what I previously referred to as the fundamental and practical limits to the scope of science. By fundamental limits I mean limits that I can’t circumvent in principle or in theory. The limits on probing the quantum space I just spoke about are, I would say, fundamental. Now, the people at the meeting I was referring to seemed to suggest that the limits to ability of science to understand the cosmos and consciousness are fundamental. I don’t agree with that. I think the limits on our ability to understand the cosmos and consciousness are practical, not fundamental.

A. Not practical in what way?

S. Practical limits are factors like cost, scale, or even the motivation of the government and the public to support the scientific endeavour. For example, high-energy particle physics needs to push experiments to higher energies, which requires bigger particle accelerators. There is no fundamental limit to the size of an accelerator, but the cost of these tools is already becoming prohibitive, even for collaborations of governments.

A. You don’t think we need to look at the cosmos from the outside to understand it?

S. Let me give you an analogy. Suppose the cosmos is like a long-lost brick wall buried in a sand dune. You dig down and uncover a small part of it. You analyse the composition of the bricks and the pattern of the brickwork. Now you know everything there is to know about the structure of the brick wall. If you uncover more of it, you find more of the same, you learn nothing new. The cosmos could be like that – you can learn everything about it, the equations that govern it, from looking at a small part of it.

A. But Socrates, you can’t tell how big the wall is by looking at a small part of it, or whether it has a chip out of it in a certain place that you have not uncovered.

S. True, Adeimantus, but how can you be sure that the size of the universe is not encoded in the structure you can discern by looking at a small part. After all, the information that makes an elephant bigger than an ant is encoded in the genetic material that can be sampled from a small of the animal. As for your chip in the brickwork, that is an accident of history, not a feature of the underlying structure. Our equations governing the universe might predict a certain density of galaxies, but they would probably not tell you that a galaxy would be at a particular location, since that might depend on historical accidents, like the presence or otherwise of a quantum fluctuation.

A. Fair enough. What about consciousness?

S. I said pretty much all I have to say about consciousness in a previous conversation². I will just add that when people talk about consciousness, they seem to have two things in mind: the experience I have that I’m thinking about something; and being aware of the meaning of words or symbols. People say that artificial intelligence as it currently stands can produce fine sentences and paragraphs, but they are just words AI pulls together without understand their meaning.

C. What do you mean by ‘meaning’?

S. The meaning, I would say without getting too technical about it, of a word or symbol is an action, thought, or emotion that you associate with that word or symbol. For example, my nearly two-year-old grandson comes up to you, holds up his arms, and says, ‘heavy’. What he means by ‘heavy’ is ‘pick me up’. He associates the word ‘heavy’ with being picked up because every time his mother picks him up she says, ‘You’re so heavy’. For my grandson, ‘heavy’ means ‘pick me up’.

C. Charming. He must be well-fed.

S. Quite so. There is no reason I can see why a computer system of sufficient complexity could not be equipped with sensors and the ability to remember experiences. It could even have the ability to sense what it is thinking about by being equipped with a self-awareness process, as I described in our previous conversation on consciousness. Such a system could report its feelings and thoughts and the meaning it attaches to words. Computer scientists and neuroscientists are already working on systems which copy the architecture of nervous connections in the brain. It wouldn’t surprise me if before long they report that their machine shows all the signs of consciousness. How would anyone prove that the machine is not conscious, I wonder?

A. They would be hard-pressed to do so. Coming back to tools, Socrates, I remember reading an article that said our view of the world is fundamentally limited by the tools we have at our disposal. What do you say to that?

S. I think you refer to an article by a scientist by the name of Gleiser (Gleiser 2018). Haven’t I already mentioned the fundamental limit of our tools at the quantum scale and the practical limitation of the cost of high-energy particle accelerators? What Gleiser is saying reminds me of the drunk under the streetlamp.

A. Why?

S. I think I mentioned this drunk before. It’s well known that drunks look for their lost car keys under streetlamps where they can see, not in the dark where their cars are parked.

C. Are you saying that scientists are like drunks?

S. Only when they too are drunk. When they are sober, scientists are usually smart enough to look where they think the answer lies and they adapt their tools accordingly. I think Gleiser’s observation is so obviously true that I wonder at the need to expend ink on it. Perhaps he was looking for kudos for discovering a fundamental principle, in which case I think he exaggerates. More seriously, he makes it easy for cynics like the postmodernists to misquote him. They will say Gleiser’s argument proves that scientific results are just personal opinions coloured by their own culture, methods, and objectives.

C. And you course you don’t agree?

S. Haven’t I already said so many times? A limited view need not give rise to a mistaken view. Newer theories enabled by better tools often include and expand the scope of an older theory. Take the example of Special Relativity. Special Relativity does not replace classical Newtonian mechanics, it expands the scope of mechanics. Classical Newtonian mechanics is a good approximation in the case of relative speed much less than the speed of light, but not for higher speeds. Special Relativity is accurate for all speeds. Likewise, quantum mechanics reverts to classical mechanics for macroscopic objects. It is not about the old theory being wrong, but of the domain in which it is a good approximation being clarified. More radically superseded are some of the underlying concepts of the old theory, like the concept of absolute time in Newtonian physics. So, some people have the wrong idea about Kuhn’s paradigm shifts – the old theory is not always thrown out and replaced by a new theory. The old theory is just expanded by being incorporated into the newer theory.

C. And by ‘some people’ you mean postmodernists?

S. Who else? Kuhn was right about Special Relativity ushering in a fundamental shift in our understanding of space and time, but it was not a cultural shift or a takeover by a Special Relativity power group. The experimental evidence demanded it. Coming back to the cosmos, we would like to be able to see more of it, especially as looking further away is equivalent to looking back in time. Most things of interest are immensely far away, and we see them ‘as through a glass darkly’, clouded by dust and gas. It’s not so much that we need to see better to develop new theories, but to test for differences between variants of theories we already have, like general relativity and quantum mechanics. Our tools are limited. These are practical limitations - cost and scale - but so intractable that they border on fundamental.

E. Will humans ever get to Mars?

S. Humans may get to Mars, but I doubt they will ever escape the solar system. The distances are so impossibly vast that civilisations would have come and gone while someone was travelling to nearby stars and back, supposing they could do that. Nobody would remember who they were and why they went, even if humankind still existed on Earth.

C. So, we’re all alone then?

S. It looks that way. The same would apply to any aliens who came our way. They might as well join us rather than trying to get back home. But we stray from our theme. Where were we?

A. The limits of science.

S. Yes. I need to talk about another type of limitation that is both practical and fundamental, namely complexity. I’ve mentioned it in previous conversations.³ The scientific method likes to deal with simple things. Even something with a large number of parts, like a gas of billions of molecules, becomes simple when the molecules are all of the same kind, and the properties of the gas are treated statistically. But something like human behaviour is not simple: we can never isolate it from extraneous factors, many of which are not even known or understood.

A. What do you mean when you say that the limitation of complexity is both practical and fundamental?

S. To answer that question I need to talk a bit about how some scientific theories are more solidly based than others, and how some activities that are presented as science are not really science, or only barely so. Physics is the science that deals with the most basic of things. We know that there are protons, they have been seen in thousands of experiments, we can measure their mass, charge, and spin. Likewise, there are electrons and neutrons. The equations of quantum mechanics allow us to calculate how protons, neutrons, and electrons combine to form atoms, and the predictions are spectacularly confirmed by experiments. Atoms combine to form molecules, and so we have Chemistry. If we look at the chemistry of organic molecules and living things, we have biochemistry. We have a hierarchy of scientific disciplines, each with their own theoretical structure, but all built upon the more basic theories and contiguous with them.

A. What do you mean by ‘contiguous’?

S. I mean that everything in the more basic discipline is incorporated into the higher-level discipline and consistent with it. Let me give you an example. If I want to predict the weather, I delve into the science of meteorology, which is build on physics for the main part, and a bit of chemistry and perhaps a couple of other basic disciplines. There are no really new theories in meteorology, just more complex applications of the basic science. To forecast the weather, I need a model of the atmosphere, oceans, and so on. The model is complex. It includes a large number of processes, all well-understood from the basic sciences. I also need to carve up space and time into a large grid of spatial regions and time intervals. Now I have a large number of moving parts and data points to deal with. I have the problem that to start my model running, I need a lot of data about the initial conditions. I also need the model to run faster than real time if I am going to forecast, so I need a lot of computing power. My data sensing network and my computer power are limited, but these are practical limitations: I can improve them by spending more money on them. There is also a fundamental limitation of all complex systems. If I start the model running with two very similar data sets, after some time they will diverge to give very different results. If I improve the data sets to make the results more similar, the computer predictions will take longer to diverge, but they always diverge eventually.

C. So, weather forecasting is not an exact science?

S. You could say that there is no new science in weather forecasting, but that would not do justice to the great insight and skill required to make the approximations and simplifications that are necessary to achieve workable models without greatly sacrificing the accuracy of the forecasts. So, in a way you are right, Critobulus, meteorology is in some way a science of approximations.

C. I won’t tell the meteorologists you said that.

S. Did I tell you I used to be one? It was a long time ago. My point was that meteorology is built on more basic sciences and entirely dependent on them and continuous with them. I would say that meteorology is contiguous with the more basic science. But now let me give you an example of a type of science where this is not so. Suppose I want to predict human behaviour. Will I start with physics?

C. I expect not.

S. Why wouldn’t I start with physics, solve the equations of quantum mechanics and get chemistry, then biochemistry. I move up to the science of the human body and I model the brain and nervous system. I might even discover that my model is conscious, as I was saying before. Do you see a problem here?

A. It’s too complex. It’s not practical.

S. Definitely. Even if the model worked like a person, which person is it like? To model different people and their range of behaviour, I would need parameters for personality traits, but nobody knows how to relate personality traits to more basic scientific concepts. So, what do I do?

C. You make stuff up.

S. Crudely put, Critobulus, but not too wide of the mark. If I am a psychologist, I might introduce concepts like the unconscious. I might even base my theories on concepts like the ego and the id. These concepts might turn out to be useful in some way, but they are definitely not contiguous with more basic sciences. There is no physics and chemistry of the id.

A. Is that a problem?

S. It means that the basic concepts of a science like psychology are far more tenuous than those of physics, for example. They are much more difficult to define and measure. Also, the processes which are modelled to make predictions, say of behaviour, are much less well defined. The result is that while physics, given accurate starting data and barring calculation errors, makes predictions that are absolutely certain, psychology makes predictions that are at best an indication of likelihood. I say this not to belittle psychology, because an indication of likelihood may be better than no knowledge at all, and psychologists are best equipped to help some people with mental illness.

C. Are you ranking the sciences by respectability?

S. You could say that the predictions of physics deserve more respect than the predictions of psychology, but that is not really my point. The point I am trying to make for you, Critobulus, is that the theories and predictions of scientific disciplines that deal with very complex things are much less certain and should be treated with much more caution than the more basic sciences that deal with simple things.

C. I get it.

S. Even so, let me give you another example, economics. Economics obviously depends on human behaviour. For a long time, economic theories were based on the idea of the ‘rational actor’. The rational actor is an ideal person who always acts in his own self-interest after doing a rigorous cost-benefit analysis. The only trouble is, people do not always do that. Economic theory recognised that people sometime act for the good of a collective. The trouble is that you don’t always know which way people will go when you make predictions. And once people figure out what you are trying to do, they find ways around it. Behaviour is very difficult to predict in any general way. Economics is not contiguous with more basic and more rigorously tested sciences.

C. I still get it.

A. Activities that are merely data analysis, in my view, some people regard as science.

S. That’s my view too, Adeimantus, and thank you for raising the point. Suppose I work in medical research, and I have done a survey that shows obese people in a certain age bracket are more likely to die of heart disease than people who are not obese. A man in a white coat on the TV might say this is ‘scientific proof’ that obesity is bad for you. But research is not the same as science and neither is data analysis. From the data analysis I only have a correlation between obesity and heart disease, I don’t have any theory that connects the two and explains the correlation. I haven’t done any science yet. Data analysis is a tool that science uses, a step in the process, but it is not science in itself. It is possible for medical sciences to pursue a theoretical understanding of the human body, its diseases, and the effects of drugs, but because of the complexity, the effort often stops short of rigour, especially in the pharmaceutical industry.

C. Where is all this leading us, Socrates?

S. I have identified two limitations to the scope of science that challenge the arguments of humanism, at least as presented by Grayling (Grayling 2013). Firstly, quantum science reveals that our whole concept of space and time is naïve, as we discussed last time,⁴ so the argument that science has proven God does not exist is weak. Secondly, complexity prevents us from developing a scientific theory of human behaviour that will allow us to predict behaviour reliably. Therefore, contrary to humanist claims, there can be no firm scientific basis for ethics.

A. But Socrates, didn’t Grayling only say that we should develop ethics rationally. Rationality is not the same as science.

S. I agree, Adeimantus. Science is rational, but not all rationality is scientific. You can propose axioms for ethics, such as ‘all people are equal’, and develop your ethics rationally, meaning by logic. But rationality also demands evidence for proposals. There is no evidence that all people are equal, on the contrary, not even for the idea being a good one. As I have said many times, the starting principles for ethics are a matter of personal preference, not of science. Science demands evidence for the principles, and that is lacking in ethics. I do suspect, however, the Grayling has a hankering for the kudos of a scientific basis for his ethics.

C. You’re a hard man, Socrates!

S. But fair, I hope. I think it’s fair to say that philosophers have not fully ‘taken on board’ the fact that three-dimensional space and a time that flows are in some way illusions. The philosophers have not yet figured out what this revelation means for understanding humanity, especially ethics.

A. Sounds like a tough problem.

S. I think it calls for more humility than is usual.

C. As my kids say, ‘Are we done?’

S. We are done for today, Critobulus. Next time, you may be pleased to hear, I will attempt to summarise everything we have discussed about my worldview.

References

Falk, Dan. 2019. "Cosmos, Quantum and Consciousness: Is Science Doomed to Leave Some Questions Unanswered?" Scientific American. Accessed January 5, 2024. https://www.scientificamerican.com/article/cosmos-quantum-and-consciousness-is-science-doomed-to-leave-some-questions-unanswered.

Gleiser, Marcelo. 2018. "How much can we know?" Scientific American 69.

Grayling, A. C. 2013. The God Argument: The Case against Religion and for Humanism. Bloomsbury.

1. See the conversation Socrates’ Worldview.

2. See the conversation The Conscious Human.

3. See the conversation The Scientific Method.

4. See the conversation Space and Time.