All models are wrong, but some are useful (after George E. P. Box)

There seem to be two kinds of “science”. One with a capital - Science - and one without - science. Alas, it appears that academics and commentators throughout the Western world are yet to discover this. We very predominantly hear them talk about “science” with a lowercase “s” - at least in most types of popular and political discourse - when really they mean “Science”.

This is of critical importance because a huge number of the world’s preeminent thinkers on the fundamental challenges that we face as a species (extreme inequality, polarisation, ecological collapse, AI, WMDs, etc.) make the correct observation that in the real world there are competing ideologies and viewpoints. US commentators often tend to focus on their own partisan national political divide (Democrats vs Republicans) but all recognise that competing ideologies exist in all parts of the world, though the divides are often poorly aligned with those in other countries, and often even less well aligned between different civilisations (in the Samuel P. Huntington sense). Some of the more sophisticated bird’s-eye-view commentators (such as Nate Hagens, Daniel Schmactenberger, Tristan Harris, and many more) astutely point out that it is only through consensus and a shared vision that we will make any progress towards meeting the existential challenges we now face, particularly around our shared environment on planet earth and the AI-based products and weapons of today and tomorrow. We need to learn how to make sense together. With the increasingly god-like tools (AI, CRISPER, etc.) that are being democratised and put into the hands of the many, it is hard to fault such claims and such thinking. It is not through conflict or even compromise that we will rise to the immense challenges but through honest empathy and genuine consensus.

A significant majority of these thinkers do, however, require that we agree on “the science”. Without wide-ranging agreement on “the science”, we will never be talking about the same thing, never be starting from the same point of reference. This is, alas, where most commentators commit the spelling mistake. They do not mean “science”, they mean “Science”.

This is absolutely critical because a significant and increasing number of voices (on both/many sides of many arguments) are calling into question the often assumed objectivity and neutrality of “science”. Without “Science”, it unfortunately means that we will likely never be able to bridge the divides and reach consensus because we are requiring shared agreement on a set of data, models and analyses that are supposed by only one side to be self-evident, free of bias and prejudice in and of themselves, and the opposite by other sides.

Science, wherefore art thou?

So how are science and Science different? Let us list a few differences. I have left some question marks in places where I am yet to find serious answers.

science

Done by real people (like the ones you and I personally know)

done in real labs, institutes and universities
done in real cities/towns (and military/industial parks), in real countries with real governments and real laws
budding scientists are educated by other scientists (who are also/still real people)
lots of taboo subjects, often subject to changing fashions and popular metaphors
funded by interested industry, think-tanks (and similar), NGOs and governments, very often trying to prove a particular thing for a particular economic or political purpose
many deep religious, cultural (including “moral”), political, economic influences
“scientists” are no different in terms of psychological biases to anyone else (confirmation, conservatism, clustering…)
ubiquitous use of language that attempts to hide the role and affiliations of the researcher (nominalisation, passive tense, assumed ascription of agency to non-agentive subjects (“the genes dominate”, “the substance pushes out”, etc.) and opaque or even simply incomprehensible to all but a small clique of privileged insiders
raw data rarely shared, often cherry-picked (removal of outliers, “dysfunctioning measurement device”, “operator error”, etc.), regularly P-hacked, and failed experiments very rarely admitted to at all, let alone published
experiment reproduction is very rarely attempted and extremely rarely successfully carried out

Science (minimal properties - no religious, political, economic, ethnic/racial, class or gender affiliations)

Done by Scientists

done in Science places, which are functionally equivalent wherever they are (where are they?)
funded by completely disinterested funding organisations (who?)
Scientists are educated by other Scientists (who?)
full agreement on basic axioms (what are these?)
properly objective (no significant measurement error, many different eyes all come to the same reading/conclusion, etc.)
no taboo topics or subjects
no cultural, linguistic, economic, political, religious influences or psychological biases
very clear majority consensus on data interpretations in vast majority of cases (basically only neophytes incorrectly interpret data, everyone else is always in agreement)
experiments always easily reproduced by multiple teams in unconnected institutions

Is that thee o’er yonder?

Surely, if we are to have any claim whatsoever that there are “Facts” and “Truths” that can only be self-evident to all, then there must be compelling (Scientific even) evidence that not only it is, in theory, possible to have completely Scientific methods carried out by Scientists, there should be Scientific proof that “the Science” is indeed “Science” and not “science”. So “the Science” we must agree upon before starting our consensus building is just that, rather than mere “science”.

So where are the Scientists who do/produce Science? While I have not personally studied at a “top-tier” university (Oxbridge, Ivy league, etc.), I know plenty who did, and my own PhD supervisor is a (PhD) graduate of MIT. Alas, I have never met a Scientist, only scientists. I have searched high and low and, in spite of my best and earnest efforts, I have found no remotely convincing evidence that any of the institutions where anyone I have ever met studied or works has Scientists rather than scientists. I have never been shown the procedure with which we transform mere humans, or even scientists, into Scientists. One might assume that this metamorphosis happens at some point during the doctoral or post-doctoral studies but I have found no evidence for any particular instances of this ever having been documented. Lots of people get PhDs, but a PhD does not science guarantee.

Many a (particularly popular) article, journalist or politician certainly talks about institutions as if they were Science places staffed by Scientists, but we are just expected to believe this rather than be provided with any robust evidence. The media and governments shower praise on many an institution, and we are told that, according to some Scientific system (which after a 10-minute review turns out to not even be scientific, let alone Scientific), they are to be ranked 1st, 2nd, 3rd, etc. Curiously, there are different rankings that have different orders. Regardless, none of the these even attempt to show us where any purported cutoff might be (i.e, 1-20 are Science places, 21-50 are Almost Science Places, the others are merely science places, etc.), or why some institutions that themselves believe they are doing at least science, are not even on the lists at all.

I would LOVE to be proved wrong, and to be shown some compelling evidence that XYZ institution in XYZ country has only Scientists. Alas, I have been shown none. Were I to find some good evidence, I would undoutedly immediately cease all other activity and dedicate my life completely to finding a way to study/research with the Scientists of this Science place. So this is definitely personal and I am highly motivated!

But, but, but…

A typical argument is that, of course, there is no Science place filled with Scientists but rather that Science emerges out of the interaction of many scientists from many competing/cooperating labs/institutes/universities. It is some abstraction of all of the concrete instantiations of (flawed) science that gives us Science. Unfortunately, that doesn’t really move us along because we can’t print out an abstraction and take it with us to a sense-making session, we can only print out concrete instantiations. We could attempt a meta-analysis but, O cruel fate!, it turns out we only have scientists to perform these, so we only end up with another piece of science, rather than Science, to bring along. Positing models dealing with abstractions is definitely part of the activities that we generally consider to be scientific or academic work but that doesn’t at all mean that these abstractions are of any use to us in the real world. They are often of strictly zero utility when we are confronted with real-world problems requiring real-world solutions.

“But all the issues you talk about are picked up during the peer review process!”

While it sounds like peer review might work in theory, in practice anyone who has ever attempted to submit a journal article, or read a journal article about a subject they are deeply familiar with, already knows that particular journals have their own particular biases. Articles are not reviewed by God after all, they are reviewed by other scholars, who suffer from exactly the same kinds of biases as the contributors themselves. There is no need to mention “conscious fraud”, because, the reality is, it is probably quite rare (and necessarily involves a conscious conspiracy). I think the vast majority of articles are believed by the contributors to be “the best we could do with the time and resources at our disposal”. Very often the issues you hope the reviewers won’t push you about are not even mentioned in feedback, but you will get repeated revision requests for issues of utterly trivial importance. Nit-picks for the pet peeves of the reviewers, not actual challenges on the fundamental assumptions or important issues of sufficient data for meaningful relevance. Even a 99% confidence interval on a group of 30 second-year psych students should not allow us to suggest we can make claims about second-year literature students, let alone all humans, yet very few are the reviewers that would reject such a submission as trivial, rather than simply request a slight rephrasing of the conclusions, probably with some extra hedging. After all, they need to get published too, so putting the bar too high is bad for everyone “in the field”. These are normal human reactions to commonly encountered socio-economic situations. No one is suggesting (or should be) anything nefarious or sinister, it is just normal folk trying to get along. This is the entire point. There is nothing magic about scientists or the scientific process, it’s all just people trying to get along. And again - there is nothing wrong with that in the sense that the kind of activity these people perform often brings enormous benefits to countries and to humanity at large, and there is definitely no silver bullet to the “being human” problem.

The real problem

So unless we give up what can only be described as a faith in Science, there is a high likelihood that many of the most important issues of the day will never get discussed in “good faith” by all the relevant stakeholders. The answer is obviously NOT to say that there is no such thing as “science”, or that all-comers should be given equal weight in a discussion. Just because uncle Herb claims he fries a great steak doesn’t mean you will rely on him to take care of the catering for your wedding! But unless we are fully open and transparent about the fact that Prof. Jones of Harvard is also someone’s uncle Herb, and is subject to exactly the same biases, prejudices, influences and blind-spots as everyone else, we are never going to be able to build the trust needed to rise to the world’s great challenges. We need to start all our conversations with that, then try and put on our interlocutor’s shoes and walk for a mile or two.

The real solution

Acknowledging that all human behaviour, including all intellectual activity (science, philosophy, etc.) is necessarily personally biased and culturally structured means that we have the added advantage of being able to include others from other cultural traditions. It’s not just rich(ish), white men who have found the Truth, everyone is potentially able to come to the table to help to build an understanding of a particular issue. This is not sufficient though, as it isn’t hugely different from what we have now, where interest groups and lobbies manoeuvre to get their voices heard and to push particular narratives, with or without data in hand.

Science, like Democracy, dies in darkness.

The solution, like that to many other pressing issues, is accounting. Accounting not in the Big 4 sense but largely in the opposite sense, meaning openness and transparency. Science needs to immediately drop the pretence of detached objectivity and be far, far more open and transparent about all aspects of scientific inputs and processes. We need to be open and transparent about the fact that although most scientists will act in good faith and attempt to be objective and neutral, they will nevertheless always fail. The fact that they have failed should not be taken as a death knell for them or science generally. It should be part of the evidence that we all bring together to discuss when we make decisions on how to move forward. Someone who has spent decades studying a particular area should be given the respect they deserve when they speak about it and obviously more respect than “ol’ Stevie from down the pub” who once read a book on it but nonetheless acknowledged as someone as fallible and corruptible as all humans are. As a scientist to question and challenge, not a Scientist who must be believed on faith from a previous publishing/citation record (popularity contest) or hailing from a particularly famous institution (popularity contest).

And we need to be honest with ourselves about the true extent of possible bias and influence. So it is not just the personal financial interests of the immediate members of the research team’s direct research funding that need to be taken into account, it is also the team’s previous employment history, the previous employment history and financial interests of family members and other close associates/friends, the religious ties, the departmental ties, local and national government priorities, political and scientific issues at that time, the previous life history (brother died of X, wife was afflicted with Y, etc.) - and the list goes on. Every research initiative should be trying its very best to put ALL its cards on the table. Any undisclosed affiliation should be taken as an grievous shame for academics and administrators. Academic profiles should be extensive and open, able to be enriched by anyone who has something to say. Some comments or claims of bias will be inaccurate or plain wrong but the community can deal with that in a transparent and open way. If everyone is coming to the table bragging about their massive transparencies rather than how they have the Science and everyone else is just stupid or biased, then we start from quite a different place. Claims of bias should always be rebutted with “there is consensus on this claim already, and very few people disagree” because such issues have been aired, documented and debated as part of the normal process of science.

All source/raw data (anonymised where absolutely necessary) should be made available, and all analysis procedures explicitly and rigorously documented with open source code and procedures. All publications should be open access, along with all reviewer comments and all previous revisions. Science is damned hard, and almost always damned messy. That’s just life. Creating a fictional world of neat boxes and unbiased motives is not going to fool anyone who doesn’t particularly like the message the scholars are pushing, though it might gull adoring hordes of science fan-boys and girls.

Though it is wrong, it might very well be useful.

The point is that, whether it fits with our narrative or not, there are no such things as Eternal Truths, and even if there were, we could never be sure we had found one. Even positivist/realist science doesn’t claim to have an infallible evaluation mechanism for discovering when a theory can be classified as Truth, and so no longer a subject for scepticism/further science. As such, in practical terms it is utterly irrelevant whether we allow for Truth, temporary truths or simply useful models. So let’s embrace that fact and engage honestly and respectfully with others. Scientists may be pushing a political narrative but that doesn’t mean that their theories aren’t useful. The fact that 97% of people who spend most of their day thinking about something agree on the broad strokes of something doesn’t mean it’s “True” but it does mean you’d better have some pretty damn good evidence against it, including a sophisticated error theory and/or grounds for a massive, global conspiracy. The world’s greatest scholars have all agreed such things as the world being flat, the moon being made of cheese, or the Earth being 6000 (or 5500 at that time) years old. We moved on from those findings though, and we might discover alternative models for other “settled science” (Science?) issues that the vast majority of scholars will move on to also.

The point is that with science being a human endeavour, it is impossible to get rid of our all-too-human biases. It is, however, possible to hide them from public view, at least for enough time for harm to be caused. These biases may be conscious or not, directly financial or not. It is not because the dominant economic and social narrative is neo-liberal, with its implicit claim that all value and power can be expressed in monetary terms, that sources of financial bias are the only ones that need to be declared. Humans are far, far more complex than that. We must understand that bias is simply a “cost of doing business” in any human activity. So let us require openness and transparency, constant observation and positive challenging rather than delegating vigilance to the publicly promoted idea of institutions. We must stop fetishising science and the places where science gets done. Getting into Oxford, Stanford, ETH Zurich or BeiDa doesn’t mean there is something magical about you, just as graduating from or teaching there doesn’t either. These institutions have massive budgets and can employ science communicators who are expert at playing the publication game. But with massive government budgets, endowments to be managed and corporate/private donations always come massive pressures to pursue a particular set of narratives, or to minimise others. One of the truly unfortunate biases that absolutely everyone has is the tendency to see corruption and influence in the institutions of the “other” but not in our own. While it is impossible to eradicate this bias, that should not stop us from doing everything we can to try by being open and honest about its existence, rather than trying to hide it and pretend it isn’t there, with mutterings about “anti-science loonies” whenever a challenge is laid down.

Never mind the bollocks, let’s get on with the sciencing!

But, but, but…

“But what about my (my children’s, parent’s, etc.) right to privacy?”

If that is more important to us than saving the planet and civilisation as we know it, or if someone can suggest a different path that doesn’t require coercion, obfuscation and domination, I’m all ears. At the end of the day we need to decide whether to move towards transparency now, or risk an existential threat moving past the point of no return while we hope for a better solution to be proposed.