Retained Complexity

Revisiting Parfit’s Existential Risk Thought Experiment from the Perspective of Emergent Complexity Pluralism

Derek Parfit

In Intrinsically Arithemetical Realities and in R.I.P. Derek Parfit, I quoted a passage from Parfit that anticipated the study of existential risks:

I believe that if we destroy mankind, as we now can, this outcome will be much worse than most people think. Compare three outcomes:

1. Peace.

2. A nuclear war that kills 99 per cent of the world’s existing population.

3. A nuclear war that kills 100 per cent.

2 would be worse than 1, and 3 would be worse than 2. Which is the greater of these two differences? Most people believe that the greater difference is between 1 and 2. I believe that the difference between 2 and 3 is very much greater. The Earth will remain habitable for at least another billion years. Civilisation began only a few thousand years ago. If we do not destroy mankind, these few thousand years may be only a tiny fraction of the whole of civilised human history. The difference between 2 and 3 may thus be the difference between this tiny fraction and all of the rest of this history. If we compare this possible history to a day, what has occurred so far is only a fraction of a second (Parfit, 1984, pp. 453–454).

While Parfit in this thought experiment focused on humanity and civilization as the emergent complexity to be conserved, and on nuclear war as the existential risk that could interfere with the conservation of humanity and civilization, we can generalize this thought experiment to other forms of emergent complexity and other risks to them.

Taking life simpliciter as an emergent complexity to be conserved, compare these three outcomes:

1. Continuous terrestrial habitability without extinction events

2. An extinction event that kills 99 percent of Earth’s life

3. An extinction event that kills 100 percent of Earth’s life

Again, as in Parfit’s thought experiment, we can probably all agree that 2 is worse than 1, and 3 is worse than 2, but I would without question follow Parfit that 3 is far worse than 2 in comparison to 2 being worse than 1. In the long term history of the biosphere, there have been extinction events that have come close to killing 99 percent of life on Earth. The End Permian extinction (also called the Permian-Triassic extinction event, and the “Great Dying”), for example, is estimated to have killed up to 90 percent of life on Earth, eliminating many entire clades in the process.

Since we can look back on a history of billions of years of life on Earth, we can see, played out in real historical time, the scenario that Parfit suggested in relation to civilization: history is long, and the history of civilization, so far, is short. The history of any particular biome, or ecosystem, may be short in cosmological terms, but as long as the history continues, life can come back from the brink (perhaps it could even come back from being 99 percent extinct), and can come back within a few million years, which is a “rapid” recovery when viewed in terms of the billion year time scale of life on Earth.

Here we can observe degrees of severity in a denudation event. In the mass extinctions in terrestrial history, most emergent complexity has been conserved. That is to say, even though entire clades have gone extinct (the entire clade of dinosaurs went extinct in the wake of the K-Pg extinction event, which, I note, Ernst Mayr called the “Alvarez extinction event” in his book What Evolution Is — this nomenclature was new to me), not only has life not been rendered 100 percent extinct, but also evolved features of life have not been rendered extinct.

For example, evolved complexity in life such as the eukaryotic cell, multi-cellularity, photosynthesis, the ability to metabolize oxygen, sensation, central nervous systems, vertebrates, flight, and eusociality have all been conserved over multiple mass extinction events. The species that exhibit these emergent complexities have changed over time, and the ecosystems in which they participate have changed, but the emergent complexities themselves have been retained and have gone on to be the basis of further emergent complexities that supervene upon them in turn.

It is not difficult to imagine an extinction event in which some or most of these evolved complexities could be lost. Such a mass extinction in which emergent complexity is not retained, or not fully retained, would constitute Stephen J. Gould’s thought experiment of rewinding the tape of life and running it forward again — would the result be roughly similar, or radically different? If complexity is eliminated in an extinction event, would the same suite of emergent complexities appear again, or would the terrestrial biosphere have been perturbed onto a very different course of development?

Some emergent complexities can be shown to supervene upon earlier emergent complexities, so that they form a sequence. The eukaryotic cell and multi-cellularity must come before the evolution of brains and central nervous systems, and the latter must precede eusociality. Where we can establish a clear sequence, we can identify thresholds of complexity that can be retained over historical discontinuities. However, in some cases it is not clear that emergent complexity is a sequence, since several developments follow from a particularly fruitful threshold. There are probably, within the terrestrial biosphere, several sharply distinct forms of consciousness. All are consciousness supervening on a relatively sophisticated brain and central nervous system, but these neural systems can vary widely. If cephalopods are conscious, as is argued, for example, by Jennifer Mather (cf., e.g., What is in an octopus’s mind?), inter alia, then there is consciousness in the biosphere that has been phylogentically distinct from our own lineage since the Cambrian explosion.

There may well be diachronic taxonomies of mind that chronicle the development of consciousness over time, being qualitatively distinct at different periods of time, but it is equally obvious that there can be synchronic taxonomies of mind that follow from adaptive radiation, and, in the process of adaptive radiation, as life adapts to different niches within the biosphere, that life (along with any consciousness or intelligence it possesses) will be subject to radically different selection pressures. Qualitatively distinct forms of consciousness (as well as qualitatively distinct forms of intelligence supervening upon qualitatively distinct forms of consciousness) could arise as a function of consciousness and intelligence supervening upon distinct embodiments, surviving in distinct ecological niches, and subject to distinct selection pressures. Such qualitatively distinct forms of consciousness would not form a sequence, but each form of mind could itself be the source of further emergent complexities and thus a pathway to a distinct sequence.

If a catastrophic extinction event were to result in the extinction of all life on land, all the emergent complexities that supervene upon land animals, their ecosystems, and their ecologies, would be lost, but the emergent complexity of consciousness could be retained in Earth’ oceans. Indeed, an extinction event could be so severe that it resulted in the extinction of all vertebrates, so that not only the marine mammals, but all fishes were made extinct, but consciousness might still be retained by the cephalopods.

We can only speculate what further forms of emergent complexity might arise in such a scenario, especially since the specifically human impact on ocean life would be removed and evolution in the seas could go on uninterrupted by technology and industry. Over the billion years or so of habitability remaining to the Earth, it is likely that the land would be colonized again, though the mechanisms of terrestrial colonization would likely be different in detail. If an organism descended from conscious cephalopods were the life to re-colonize the land (a scenario like this is depicted in The Future is Wild, episode 13, “The Tentacled Forest”), then the terrestrial biosphere would begin as a sentience-rich biosphere, and might continue to evolve more sophisticated forms of sentience as it branched out in an adaptive radiation into terrestrial niches.

In the fullness of time, the terrestrial biosphere might eventually become the intelligence-rich biosphere that I have imagined in several thought experiments. This would be a very different destiny for the terrestrial biosphere than now seems likely on the basis of the natural history of the planet to date. I have speculated that, in an intelligence-rich biosphere, the later forms of emergent complexity to evolve would not be civilization and technology as we know it from human experience. This seems to me especially the case in some future adaptive radiation that descended from a conscious or intelligent cephalopod ancestor.

On a much smaller scale, on the sale of human history, we can see similar patterns of retained emergent complexity. Civilization, depending on how we define it, has existed for as long as ten thousand years; fewer years by many definitions. In that ten thousand years, civilizations have risen and fallen into the dust, known to us only by the material culture later excavated by archaeologists. However, even as civilizations have come and gone, and the highest degree of civilization attained at any one time has shifted from one geographical region to another, there has been a significant degree of retained complexity, usually in the form of retained knowledge and retained technology.

After the collapse of civilization in the west Asian cluster around 1200 BC, in a period sometimes called the Greek Dark Ages, literacy was lost throughout much of the Greek world, but Egypt kept its written language. The last known hieroglyphic inscription was from 394 AD from Philae (with the last Egyptian Demotic inscription from the same location dated to 452 BC), so it seems that the ancient Egyptian language eventually vanished in the fifth century AD, but at this point Egyptian was only one of many written languages in the region, so that the knowledge of written language continued after the knowledge of hieroglyphics was lost.

In the past ten thousand years, while the knowledge of many civilizations has been lost, enough has been retained that civilization in some form has existed continuously on Earth for thousands of years, and our total knowledge and technology base today is much larger than at any time in the past, mostly because it is built upon a continuous knowledge base inherited from the past.

Parfit’s thought experiment expressed in the full generality of what I call emergent complexity pluralism would take a form something like this:

1. Continuous evolution of emergent complexity with perfect retention of complexity

2. A denudation event that eliminates 99 percent of emergent complexity from the universe

3. A denudation events that eliminates 100 percent of emergent complexity from the universe, which would mean the extinction of the universe itself, which we could identify with the heat death of the universe as it converges upon thermodynamic equilibrium

Again, 2 is worse than 1, and 3 is worse than 2, but 3 is much worse than 2. Since the horizon of human concerns makes it difficult to conceive the heat death of the universe, or even proton decay, which may occur trillions of years before thermal equilibrium, we limit ourselves to that period of cosmological history that we call our own — the Stelliferous Era. During the Stelliferous Era, Parfit’s thought experiment expressed in terms of emergent complexity remains true to Parfit’s insight, though generalized across forms of complexity distinct from humanity and civilization as discussed by Parfit.

I would expect that emergent complexity across the universe will be largely retained, even if locally some forms of complexity experience regional extinction, as has been the case in terrestrial history. If terrestrial emergent complexity is lost for whatever reason — whether from human self-destruction of our legacy, or from our failure to project terrestrial complexity beyond Earth, so that this complexity is not confined to the history of a single planet — this would be a loss on a cosmological scale like the loss of knowledge of hieroglyphics.

However, just as Jean-François Champollion (preceded by the work of other scholars) deciphered hieroglyphics long after the language was dead, it will remain a possibility that terrestrial emergent complexities might someday be studied by another form of emergent complexity arising from a different source, and in this way some of the complexity of the terrestrial biosphere could be retained over cosmological scales of time even if all that we know ends in extinction and failure.

Jean-François Champollion (23 December 1790–04 March 1832)

One Man Think Tank