Mostly Wrong

(Science is flawed. It is also all we have!)

This is my first post on Substack. But doesn’t it seem, well, wrong to post under “Mostly wrong”? After all, it’s surely obvious that, even more than catering for our human obsession with sex, sport and cat pictures—even bigger than selling sh-t to people who don’t really want or need it—the Internet is really about manic self-promotion. Isn’t it frankly stupid to under-sell, when everyone is doing the opposite?

In fact, isn’t ‘Mostly wrong’ more quintessentially tepid than the Hitchhikers’ Guide to the Galaxy describing the whole of Earth and all human endeavour as “Mostly harmless”?¹

Actually, no!

A single idea is all you need here: we can identify two important categories of people. Just two:

1. Those who are, indeed mostly wrong—and aware of it; and

2. The rest of everyone, who are still mostly wrong, but in denial.

Once we work this out, we can see clearly. And it’s disturbing. Most of the current problems that our world faces can be traced back to one simple origin. Yep. You guessed it: people who live their entire existence in category #2. People who are so deluded, they fail to error correct. And no, you don’t have to agree. I could be wrong! That’s the whole point.

Now I’m going to tread on a few toes. This is something I enjoy, especially where those toes belong to either philosophers or sociologists—and let’s face it, it’s becoming increasingly difficult to tease out whether someone is a postmodern philosopher with a sociological bent, or a sociologist with a postmodern philosophical bent. And both have inviting, treadable bunions.

So let’s talk about what science is. People have debated this for centuries, so I’ll start with my definition—which is solid—and then point out why everyone else is, of course, wrong.

Science, then

So here’s how science works. I’ll even call it ‘The Science of Wrong’. It’s mostly wrong:

1. We start with problems. If it’s not a problem (actual or speculative, past | present | future) then why bother with it? Have a nice cup of tea instead.

2. We create an explanatory model. This model demands two things:

   1. Is it joined up—does it make internal sense?

   2. Does it work in the real world—when we test it, is it predictive? Does it explain the problem? And, most important of all, does it break?

3. Here’s where ‘Mostly Wrong’ bites. Because most models are just plain wrong. In fact, most models aren’t even joined up properly. But of those that are, just a tiny fraction actually work in the real world—and of these, just a minuscule number succeed when we shoot our sharpest “reality tests” at them. They break, and we’re more-or-less back to square [1] above (Only more-or-less, because a broken model can teach us stuff; and some broken models are still very useful in certain specific contexts).

But we still haven’t highlighted the most important thing here. It’s that even if a model passes all the tests, this doesn’t make it true. It’s still very possibly wrong. But it’s not just plain wrong. In fact, if we tested it thoroughly, it’s not just ‘not yet wrong’; it at least seems to be ‘not obviously wrong’.

Can you see what a powerful engine we have here? Sometimes, over the centuries and against the odds, we find joined-up models that actually work! These are valuable precisely because of all the dross that doesn’t. Better still, new models can be evaluated in the context of other ‘not obviously wrong’ models we have in our armamentarium.

The brutal orneriness of Nature, the glorious asymmetry of the repeated failure of most models of reality to actually work, powers Science in all its splendour.

Still wrong!

And—of course—the above definition of science is wrong. It’s amusing to watch philosophers trying to shore up and perfect their philosophical systems. Pretty much all of them will claim that their system is better than that of anyone else. And certainly better than the one I’ve just outlined. Not to be outdone, I provocatively claimed that my solution is ‘solid’, and theirs are, well, wrong.

We can test this. The best way to evaluate a system (model) is to line it up against other systems. The first system we’ll test our model against is, surprisingly, itself. I’ll frame this as a question:

Can we feed The Science of Wrong into itself—some might call this a “meta-model”?

Yes we can. As we’re dealing with models, and ‘The Science of Wrong’ caters precisely for this, we can feed any model into it—including itself—and then start asking questions like “Is this internally consistent?” and “Does it work in the real world?” If the meta-model fails, we can notch up another failure, and perhaps also get pointers to better ways of doing things! Quite because we admit error, we can correct errors, even within a recursive (eat-your-own-tail) approach.

You’ll quickly see how headspinning this is—and how many approaches to building science don’t fare too well. With this under our belt, let’s test the opposition.

Plato

“The safest general characterization of the European philosophical tradition is that it consists of a series of footnotes to Plato.”

This quote from Alfred North Whitehead’s Process and Reality (Pt. II, Ch. 1) is argued about rather a lot. It’s been cited as evidence of philosophical decadence, or out of reverence for Plato, or simply as ‘mostly right’; it’s been characterised as either profound or daft. For me, it underlines the importance of and confusion about realism that inhabits many philosophers’ approaches to science.

At the centre of Plato’s philosophy was his theory of “forms and ideals”. This is the slightly bonkers idea that in some behind-the-veil Platonic space, there are ideal representations of “what it really means to be x”, where x might be a ‘human’, or a ‘dog’ or a ‘table’, or whatever. This goes hand in hand with the concept that a suitably enlightened philosopher can pierce the veil and know reality. This sort of thinking clearly resonated with the Early Church (giving them a 400-year-old stamp of origin for their theological ideas) and given the influence of the Church on pretty much everything for the next two millennia, has carried some clout (Mostly fascist clout, but that’s for another post).

As that last parenthetic remark suggests, we can easily get lost in philosophical meanderings. We won’t. But we will have a gander at some philosophers from the 20th century who were effectively Platonists, and most certainly realists.² The Vienna Circle was an influential bunch of scientists & philosophers who despite their brief ascendancy (from the 1920s to the 1930s) had quite some impact. Their basic idea was that all you need to build natural philosophy is some ‘grounded truths’. Once you have these, you can build everything on a firm foundation.

This model is obviously incompatible with mine, where there is no such thing as a grounded truth. Everything ‘true’ is always just provisional—not yet wrong. And, in fact, logical positivism (as their philosophy was branded by opponents; they favoured ‘logical empiricism’) fell apart mainly for this reason. Which is not surprising, given that David Hume had pointed out the absurdity of such a position in the eighteenth century!

But if you look around today, you’ll find more than residual traces of just this sort of silliness. Not a few scientists seem to believe that even if they can’t find proper grounded truths, they can at least asymptotically approach ‘reality’. Can you see why this is silly? Yes, of course—there’s a multiplicity of counter-examples, where a theory that stood for years (or even centuries) was supplanted by another theory that explained the same phenomena using a radically different model. This applies to models of the motion of planets, and to how light bends with gravity, and how heredity works, and so on. Once you realise this, then you can see the realist underpinnings of the ‘model of science’ proposed by people like Sir Francis Bacon, whose main gig seems to be asymptotic approximation as a way of doing science.³

Anti-realism

You can understand why many recent philosophers have a bit of Platonic allergy. This is not just contact dermatitis, either. ‘Anaphylaxis’ might be a better term—a massive allergic reaction where all your mast cells degranulate, and you get hives, wheeze and go into shock. You however need to be careful in choosing your philosophical EpiPen. The wrong choice can be intellectually fatal, or at the very least, impair brain perfusion.

For example, there was a rather pedestrian physicist called Pierre Duhem, who wrote an extremely tedious ten-volume history of science. During this he made the not utterly unreasonable observation that, if a model makes a prediction that is wrong, it’s not just the model that may need to be called into question. Sometimes, the broader philosophical framework in which that model is embedded also needs to be interrogated.

Which fits with ‘The Science of Wrong’. We must continually question our models, and our associated, often fervent beliefs. But then, along came Quine. Willard Van Orman Quine was made of different stuff. I suspect he was a lot brighter than Duhem,⁴ and shared with many bright people the tendency to explore so far beyond the bounds of the circle of lamplight that defines their wisdom that they venture into foolishness. Quine was quite correct in his attempts to put the boot into logical positivism (and quite successful too), but if you haul out a copy of his Two Dogmas of Empiricism, you’ll find some very strange stuff there. Like this:

“The issue over there being classes seems more a question of convenient conceptual scheme; the issue over there being centaurs, or brick houses on Elm Street, seems more a question of fact. But I have been urging that this difference is only one of degree, and that it turns upon our vaguely pragmatic inclination to adjust one strand of the fabric of science rather than another in accommodating some particular recalcitrant experience. Conservatism figures in such choices, and so does the quest for simplicity.”

Cute, but show me the centaur then, ‘Van’. Similar throwing up of the despairing philosophical hands is common to a lot of anti-realist philosophers. “Woe is me! Everything is underdetermined! [smites breast]”. You see, they’ve missed an important trick here. But before we point out which trick, it’s wise to look at the overlap between anti-realists and sociologists here.

Sociologists can be a peculiar bunch, especially when they aspire to philosophy. But as I’ve said above, trying to tease the two apart is often moot. ‘Postmodern’ philosophers of all shades tend to give reality short shrift, and instead focus on human behaviour—as if it, instead, is intrinsically important. They’ve shifted their Platonism from inspection of and agonising about nature, to inspecting and agonising about people and their perceptions.

From this perspective—seemingly abandoning even the acknowledgement that reality exists—things are fun. At least, initially. We can observe how people work and interact, and make all sorts of fun theories; but because we’ve divorced ourselves from reality, a peculiar relativist approach emerges. Reality seems to transmute into what we perceive it to be. Science becomes “what scientists do”. Whatever ‘scientists’ do. Spot the tautology.

Thomas Kuhn provides a fine example of such an approach. Kuhn’s sociological explanation of “How Science Works” is simple. Scientists are termites. Admittedly this isn’t how his philosophy is usually framed, but bear with me. Scientists are termites, busily engaged in building the edifice (or, in my metaphor, termite mound) of Normal Science. Sometimes a chunk falls off—we have a Kuhnian ‘paradigm shift’. Then there’s a lot of frenzied activity as a new piece of termite mound is assembled, and Normal Science continues. Birds may pick off a few of the exposed termites, during this process.

That was unkind, but it does illustrate several important points. First, Kuhn’s approach doesn’t explain why some ‘science’ fails, and other science succeeds. He’s too busy modelling human behaviour. A related point is that, although he invented some catchy terms, when we try to feed his approach into itself—the recursive trick I described above—it has little or nothing to offer; conversely, when we feed it into ‘The Science of Wrong’, it comes up very short. It’s testable, and turns out to be badly joined up, poorly explanatory, and fail at its predictions.⁵

Finally, Kuhn tries to explain his ‘paradigm shifts’ as being the result of ‘incommensurable’ theories—theories that are so at variance that a chunk of the termite mound has to fall off, because members of the faction supporting one theory simply cannot comprehend the other’s point of view. And this too comes up short, when tested in reality. The problem usually isn’t with the theories or their embedding. Ironically, the problem seems to be sociological. People can be immensely bloody-minded, and wedded to their pet theory or chosen clique. Sometimes they just need to die, for their theory to die. Which is sad. We have a better way—self criticism. “I’m mostly wrong”.

The trick

Kuhn is relatively mild. When you come up against hard-core post-modern philosophers of science, you realise that they are likely not just reaching for an expired EpiPen, or holding it upside down (the orange end contains the needle). If you’re lucky, they may be reaching for a virtual EpiPen; but it seems equally likely that they’re reaching for a lightly-buttered stoat, a jar of grue, or some philosophical term so divergent from any well-reasoned concept of reality that it thankfully hasn’t yet acquired a label. Certainly not an actual, functioning EpiPen (Read Lyotard, Foucault or Latour, and then disagree).

The trick here is to realise that you can reject realism without rejecting reality. Reality is out there. It’s in here. It has this commendable if sometimes distressing tendency to keep us honest, should we entertain philosophical ideas that things like gravity, viruses, cancers and malevolent philosophers don’t really exist, or are even readily malleable.

I have yet to come across an approach to Science that is better than ‘The Science of Wrong’ that I described above. But if you have one—and it’s internally consistent, and externally testable in reality—then I’d like to hear it. (Did you see what I did there?)

This is the end of my first post on Substack, but it’s also just the start. Would you like more? Because it turns out that the approach I’ve discussed—a solid way of doing science—is applicable to every aspect of human endeavour. In future posts I hope to justify this provocative or perhaps even absurd assertion, by applying these ideas to:

• The Foundations of Science

• How to Do Medicine Well (I am a specialist physician by trade)

• AI & Why Computers Let Us Down (I have done my 20,000 hours of programming)

• Why it’s Really Dumb to Write Your Own Computer Language (but fun)!

• The Deeper Meaning of the Great Trumpian Depression: How to Fix Politics (I can kibitz as well as the next person)

• The Great Science|Religion non-dichotomy (From a strictly atheist perspective)

• Why the Published Scientific Record is Broken & How to Fix it (Yeah, I’ve been guilty of publishing in the archival literature).

… and a lot more. Where shall we go next?

My 2c, Dr Jo.

1. After much deliberation and research. Initially, the entire Guide description of the planet Earth was just “Harmless”.

2. As with all philosophy, we have the problem of defining what we mean by a ‘realist’. Here, I’m going to confine ‘realist’ to someone who believes that they can “know reality”. From the rest of my text, you can work out the trifurcation: (i) realists; (ii) anti-realists who deny the very existence of reality, instead focusing entirely on theories of mind; and (iii) scientists: people who acknowledge that reality is out there, deny that we can know it even asymptotically, and yet affirm that we can do Science as I’ve described it. I appreciate that some would categorise (iii) as some sort of ‘realist’. I don’t—because it’s confusing.

3. Which is a bit ironic, given that Bacon’s Novum Organum was intended to supplant the Organon of Aristotle, who despite some disagreements, was very much a philosophical offspring of Plato (Boxy, absolutist thinking combined with an inability to count teeth).

4. I think Quine might have rather liked my idea of feeding ‘The Science of Wrong’ into itself as he was heavily into self-reference. A ‘quine’ is a computer program that prints out itself.

5. Imre Lakatos, who took a more evolutionary approach, had a lot of fun pulling Kuhn apart—and Larry Laudan was even more trenchant, notably in Progress and its Problems (Chapter 3). For a fun, enticing and fiery anarchist approach, look no further than Feyerabend, who may on a good day describe how ‘science’ is often practised, but surely doesn’t describe a desirable approach to doing so.