Main Menu

Complexity of cause in cancer: or why the frogs jump


Our need to understand the cause or reason for dramatic, natural events is a very ancient and uniquely human urge. Professor Sir Mel Greaves explores the dilemma between our instinctive desire for simple, singular explanations when science is inherently uncertain and complex.

Posted on 07 July, 2021 by Professor Sir Mel Greaves
A frog sitting on a stone next to a pond

As practitioners of science we tend to be very sniffy about science or medicine stories in the media that we deem too simplistic, sensational or downright wrong.

The Science Media Centre in the UK, many serious science journalists and academic Comms teams go to considerable lengths to counteract this pervasive problem that can lead to public mistrust in the scientific enterprise.

The current Covid pandemic has provided a vivid illustration of how critical trust and understanding in science is in the public arena. But what are we up against here?        

The desire for certainty

Our need to understand the cause or reason for dramatic, natural events is a very ancient and uniquely human urge. One that, in each case, we instinctively hope to be assuaged by a simple or unambiguous, black or white answer.

So: why did the tsunami happen, the volcano erupt? Why are we dying of pestilence? And for millennia the solutions proffered by the sages were metaphysical; mostly angry gods.

But now we live in a technocratic, scientifically literate and well informed global society, so things should be very different. But are they? The urge to know is undiminished but so too seems to be the desire for simple, singular answers. For certainty.

To some extent this expectation is nourished by the confident assertions of binary truth/fake news by some of our political and religious leaders, eschewing complexity, nuance or uncertainty. And then further confusion, doubt and suspicion is sown by the cacophony of noise from emotionally charged social media.

A daunting task for scientists

It’s almost as though we wilfully reject rational argument even when it pertains to life-threatening events – climate change, Covid, cancer. There has to be an explanation for this sorry state of affairs.

Neuroscientists offer us glimpses into how the brain’s evolutionary history and structural organisation may determine our proclivity for bypassing evidence and reason. It seems the more ancient or ‘reptilian’ limbic system rapidly engenders emotional responses that can swamp the slow pace of analysis and reason in the human neocortex.

In an evolutionary context this may even make sense as a precautionary principle. When confronted with a potential enemy or threat to life, an immediate, emotionally driven response could be lifesaving. No time to ponder? No room for ambiguity?

So the task for scientists communicating on topics of public interest that are both complex and emotionally charged is daunting. We can seldom offer certainty and our ‘best bet’ answers are couched in caveats and subject to evidence-driven change. Scepticism and constant challenge is the way we operate.

The motto of the UKs Royal Society, ‘Nullius in Verba’, sums it up very aptly: ‘on nobody’s word’, or, ‘take nobody’s word for it’.

The fallibility of causal explanations

How can we best tackle this dilemma? The task is, first, to slow down the argument and defuse emotion. Take the heat out. And then distil the inherent complexity of causal events and prevailing uncertainties and fears into clear (clear but not simple), credible and persuasive narratives, supported by robust evidence. And, hopefully, to highlight actionable opportunities. And ‘narratives’ really does mean telling a story, which is how we best engage with others.

But how good are we at doing this? As the author Ben Goldacre (2014) has so ably chronicled, occasionally at least, very poorly. But, overall, probably B+, so good, but could be much better.

Two fallibilities we sometimes have when it comes to explaining causal events are conflating ‘association’ with ‘causation’, and myopic vision. On the former, it has been confidently asserted that obesity is a major cause of cancer on a par with cigarettes and, more recently, that obesity is the reason for excess Covid-19 deaths, especially in certain ethnic groups (The London Times headline, 4 March 2021, based on WHO report and other ‘expert’ statements).

These associations are indeed strong and statistically robust but where is the evidence that they are causally linked? There is none. They could be causal but other very different interpretations are plausible, including, for example, that an underlying gut microbiome defect, related to modern lifestyles including diets and antibiotic use, is at work here, resulting in chronic inflammation (in cancer and Covid lungs), and, independently, obesity.

'Only looking where the light shines for you'

And then there can be an issue of super specialisation or myopia: only looking where the light shines for you. In ‘Lifelines’ (1997) the biologist Stephen Rose tells the story of five biologists watching a frog jump into a pond. Each then offers an explanation of why and how that came to occur. And what is delivered are five different answers from – the physiologist, the ethologist, the developmental biologist, the evolutionary biologist and the molecular biologist. Each focuses on their particular entirely legitimate interest and insight into what is, in actuality, but one link in a causal chain. And in so doing they miss the bigger picture.

The reality is that the natural world, and especially its biology, is inherently immensely complex. The cosmologist Sir Martin Rees suggested that an insect was more complex than a star. He was surely correct but I would take it further: a single cell in my body, or yours, is far more complex than a galaxy. And that biological complexity exists at not one but two levels. Firstly in the extraordinarily diverse, dynamic and intricate molecular architecture of living entities. And then, secondly, in their contextual dependence.

Everything is contextual

In essence the challenge, as I see it, is that everything, and I do mean everything, in biology, medicine and much else is contextual. So, in biology, what a molecule does, what a cell does, what a fly does is contingent upon, or depends on, what else is going on around it at the time.

Take a protein molecule out from a cell, a neurone out of the brain or a termite out of its nest and their functional identity is largely lost. What has gone is the networked connectivity, or ecological context, in which that entity normally operates. Plus there is a less obvious but critical constraint on function or any causal event, which is past history – evolutionary contingencies which define both possibilities and limitations.

The desire for a simple cause

In the cancer field this contextual problem has long been evident with ‘the cause of cancer’. Let’s put aside the plethora of exotic and absurd causes claimed for cancer. It’s generally accepted that there are different causes for different cancers. So, smoking for lung cancer, UVB solar light for skin cancer and human papillomavirus for cervical cancer. So far, so good.

But this entrenches the expectation of singular, easy to grasp answers – cancer type X is caused by Y. Or, 30% of cancers are caused by obesity. Why is this? We seem to have been seduced by the simple causal relationship – perhaps because it offers a similarly simple solution, or cure.

Take, say, tuberculosis, caused by the tubercle bacillus, and offering the simple, magic bullet cure of an antibiotic. As with the distinguished American writer Susan Sontag (1990) in her book 'Illness as metaphor' who was confident that cancer would surely turn out, like tuberculosis and AIDS, to have a single cause.

But pathology or illness is still biology, albeit on the dark side, and there is no escaping complexity and context. There is really nothing difficult or unfamiliar about the concept of context-dependent cause. Consider the ‘cause’ of a forest fire. It may be initiated by a carelessly dropped cigarette, or a lightning strike, or a barbecue.

But then what actually transpires – the extent to which the fire catches – will be contingent of several other factors, including the type and dryness of vegetation, the air temperature, direction and strength of the wind, and the proximity of natural water barriers, coupled with the chance of early detection.

For cancer, this translates into what we refer to in the jargon as multifactorial causation. And the mix of variables invariably at play, interacting together, are, in principle, these: exposures that can be from our environment (as with UVB solar rays) or internal (as with our hormones) that directly or indirectly lead to DNA mutations, with the risk of cancer progression modified by inherited genetic variation and the ubiquitous component of chance with pervades all of biology (Monod, 1972).

This is already a more complex picture of cancer cause than the one often presented, but it falls short of a full prescription. We don’t yet have the complete explanation for our frog jumping. And this missing ingredient relates to our seemingly intrinsic vulnerability to cancer.

An evolutionary perspective

By way of an explanation, consider another analogy – this time a recurrent fault or leak in a complex industrial plant. If there were recurrent leaks in several different plants, how would this be investigated, diagnosed and rectified? Well, it would makes sense to look at the valves that are leaking, or the pressure in the system or maybe see if error by human operatives was at play. Or, blame the weather?

But what you would hope would happen with recurrent, independent leaks is that accident engineers come on the scene and interrogate the design of the overall system. If they do so, there is a good chance they will find a built-in flaw or fragility. So what’s the system design equivalent for recurrent human illness? It has to be the way our bodies have been sculptured by evolutionary processes (Nesse and Williams ,1995; Lieberman, 2013, Greaves, 2000).   

This evolutionary perspective on vulnerability can help us understand two quite striking and rather alarming facts about cancer causes (Greaves 2000; Greaves 2015). The first is this: almost all animal species can and do develop cancer. And the evolutionary explanation derives from the trade-offs of the process of evolution itself. Evolutionary or adaptive advantages often come with a price tag of potential trade-offs that can be deleterious. A price to be paid.

So, DNA has to be mutable for evolution to happen; the code isn’t sacrosanct. We, and most multicellular, complex animals, are replete with physiological adaptations that can facilitate cancerous growth (Greaves 2015). These are all beneficial traits enhancing animal fitness but with built-in fragilities or imperfect control that can contribute to the emergence of a clone of cancer cells. Perhaps counter-intuitively, animal bodies, including our own, have ‘design’ features that provide an intrinsic liability or risk of cancer. The real surprise is that there isn’t more of it.  

Evolutionary mismatch

As we search for a fuller explanation of cancer cause, a second key observation is what appears to be the excess vulnerability to cancer in modern humans and the beasts we domesticate or keep captive in zoos, compared with animals in the wild. This has to do with what is called evolutionary mismatch (Greaves 2015).

This is a situation in which beneficial evolutionary adaptations, originally selected in one particular ecological context, are now at odds, or ‘mismatched’, with contemporary pressures or lifestyles. The most obvious example is the vulnerability of white or depigmented skin to UVB-induced cancer. Another example is the increased risk of breast cancer with modern reproductive lifestyles (and mimicked in zoo mammals under enforced reproductive restraint).      

So maybe we do now have, at least in principle, a comprehensive generic formulation of cancer cause and how risk of cancer compounds, in a biologically rational manner, from exposures and in the context of our background genetics and evolutionary legacies, all imbued with the ubiquitous element of chance.

It is complex, but it is also coherent as a causal explanation (Thagard, 1999). And the downstream consequence is consistent too: mutational variation in cells and then Darwinian natural selection of the ‘fittest’ or, for us, nastiest, tissue-hijacking, drug-evading cancer variants (Greaves and Maley 2012).

The details will vary from cancer subtype to subtype, and from patient to patient, but the generic formulation has universal currency. We understand why the frogs keep jumping.

Absolving guilt and blame

So where does this leave us in communicating with the public on matters of cancer and its causes? Still having to distil complexity, indeed even more so, but now armed with a more comprehensive and coherent explanatory framework.

But then we also need to consider where our prospective audience is coming from. At the forefront of their minds, or for most of them, is certainly not a thirst for complexity or lessons in biology. It’s much more likely to be questions like: ‘am I to blame in some way, or, if not, who is’? And also, ‘so what can be done about it?’

Our task as scientists is to explain, perhaps via everyday metaphors, how cancer arises from a combination of factors. And to try to absolve guilt or blame by emphasising how our societies or cultures inadvertently engineer altered patterns of exposure that can increase risk.

And then on practical solutions, we need to drill down in the complex, causal pathway to where the traction lies. It’s not with inherited genetics, evolutionary design flaws or ‘bad luck’. But rather on modifying key exposures that are rate limiting, or critical, in the cancer risk process.

And it’s not so different with other complex scientific or medical issues of great concern to the public at large, from the causes of dementia to the vulnerability of certain ethnic groups to Covid-19, to climate change. All are multifactorial, context-dependent and complex problems. And yet all are likely to yield to a causal prescription and potential remedies.


Goldacre B (2014). I think you’ll find it’s a little more complicated than that. Harper Collins, London.

Greaves M (2015). The evolutionary determinants of cancer. Cancer Discovery 5(8) 806-20.

Greaves M (2000). Cancer the evolutionary legacy. Oxford University, Press. Oxford.

Greaves M and Maley C (2012). Clonal evolution in cancer. Nature 481, 306-13.

Monod J (1972). Chance and necessity. Collins, London.

Neese R and Williams G (1995). Evolution and Healing. Weidenfeld and Nicolson., London.

Rose. S (1997). Lifelines. Allen Lane, London.

Sontag S (1990). Illness as metaphor. Penguin books. London.

Thagard P (1999). How scientists explain disease. Princeton University Press. New Jersey, USA.
comments powered by Disqus