1 Controversy

(Crosspost of my blog article). 

Humans affect staggering numbers of wild animals. There are likely quadrillions fewer insects presently walking the face of the earth because of our activities. Most people don’t think this matters—assigning essentially zero weight to bugs. But for reasons I’ve given before at length, and thus won’t repeat here, I think these people are dead wrong. Wild animal suffering is literally the worst thing in the world. Every other present issue is a rounding error in comparison.

In light of that, it’s important to know how we humans, in the aggregate, affect wild animals. If humans harm wild animals dramatically, that makes increasing the human population a risky proposition. In contrast, if we benefit wild animals, that makes a strong case for a larger population.

I’ve argued before that most wild animals live quite bad lives. Most creatures that will ever walk the face of this earth live for about a week and then die painfully. Their deaths are sometimes quick, other times prolonged, horrifying affairs that last hours or days. I’d pay a lot of money not to have to experience the pain of a wild animal dying. The fact that this pain is routinely dished out on quadrillions of animals is quite morally serious.

Given this, I generally favor policies that reduce long-term animal populations. If animals have terrible lives, it’s better for there to be fewer of them. It’s unfortunate for creatures to come into being if they have a brief and miserable existence. Reductions in population are much more important—especially in the long-run—for determining animal welfare than the short-term suffering associated with reducing animal populations. In other words, I generally think that policies that result in fewer wild animals are better and ones that result in more wild animals are worse.

This is especially so because harsher environmental conditions don’t necessarily lead to harsher conditions for individual animals. Animal populations keep growing until they’ve reached carrying capacity—most animals’ lives are constant struggles. Thus, whether an organism starves because of climate change, or starves a few weeks later because of absence of food, their fate seems similarly unpleasant. In addition, because most animals have large numbers of offspring, for every individual animal killed, many more will be prevented from coming into existence. Lastly, long-term impacts outweigh short-term impacts, and in the long-term, organisms will likely rebound so that the average organism doesn’t have a much more difficult life.

Nearly all animals are arthropods (a type of animal that includes shrimp, spiders, and bugs). Given that I think arthropods can suffer intensely, and we have reason to care about their interests, I suspect that by far the most important factor in determining long term suffering is long-term arthropod populations. Because arthropods are so numerous, in expectation they dominate.

Thus, to figure out if humans negatively or positively impact long-term wild-animal suffering, I think the best proxy is our long-term impact on number of arthropods. If we’ll increase arthropod numbers, probably we have a negative effect. If we’ll decrease arthropod numbers, probably we’ll have a positive effect. This is a surprisingly complicated research question—most of the essay will be reviewing it.

But you might think the answer is blindingly obvious. Humans have drastically reduced arthropod numbers! Some studies say that we’ve reduced insect populations by more than 75%. Why the hell is there even a debate about this? Isn’t the answer obvious. Obviously our actions decrease populations of the creatures whose habitats we destroy and who we kill by the trillions.

Things, however, are not so simple. While humans clearly decrease short-term arthropod populations, in the long run, we have a much more ambiguous impact. Ecosystems tend to eventually rebound, but our impact may trigger a long-lasting shift to more arthropods. There are four main ways we might do this:

1. After a mass extinction like the present one, arthropods rebound most efficiently. Thus, humans’ environmental degradation may bring about more arthropods because they can thrive in more hostile environments.
2. More dangerous environments, like the ones we’re bringing about, have more R-strategists. R-strategists are organisms that invest little in parenting and have like 10 million babies, hoping a few survive. Mola mola, for instance, lay 800 million eggs! They get around more than Elon Musk. In a more dangerous environment, the strategy of “have 10,000 kids and hope some of them survive and reproduce” becomes more effective. Thus, perhaps present environmental destruction will increase long-term R-strategist populations.
3. Warmer temperatures have more bugs. For example, there’s only a single species of insect in antarctica. The Amazon, in contrast, is swarming with bugs. Climate change, by producing warmer temperatures, could increase long-term bug populations. It could be that bugs eventually rebound from the present wave of environmental damage, but the long-lasting climatic effects outweigh short-term decreases in bug populations.
4. Sometimes after a mass extinction, biodiversity goes up. While I’m not sure if this holds on average (more about this later), it could be that after organisms rebound from the present mass extinction, overall ecosystem diversity and productivity will go up.

In contrast, there are a few different reasons to think that extra humans at the margin decrease the long-term number of arthropods:

1. Ecosystems take a very long time to rebound. A mass extinction can permanently imperil an ecosystem for literally millions of years. Evolution doesn’t automatically produce richness and diversity—the Precambrian world was far less rich and diverse than the present world. Majorly disrupting ecosystem functioning could be quite valuable.
2. Number of new organisms doesn’t seem to go up that much in the wake of a mass extinction. As a result, if we eliminate many existing organisms, this might not overall result in more organisms.
3. The organisms that survive are those geared to survive mass extinctions. As a result, when things rebound, they may be less well-equipped for the non-extinction world. Thus, a mass extinction may decrease long-term animal populations.
4. The present moment has unusually high both biodiversity and arthropod numbers. Simply disrupting that to reroll the dice might be good.
5. Even aside from disrupting ecosystems, human climatic changes can affect the environment directly. These environmental changes can depress the long-term carrying capacity of environments.

The next two points are about the impact that extra humans have, not about the impact that humans have had so far.

1. Humans have already had an enormous impact on the environment. The mere addition of mere humans won’t change the fact that there’s a mass extinction going on that will trigger enormous adaptation. It can, however, depress animal populations in the long-term.
2. Whatever wipes out humans is likely to have a catastrophic ecological impact. Thus, major environmental changes are imminent either way—perhaps then greater disruption is good.

These are complicated considerations. It’s not obvious which is most significant. So let’s dig into the research to see.

But first, a preliminary note: I think it’s pretty clear that if humans remain around for a long time, this will permanently dampen populations of wild animals. So long as civilization survives, and we don’t spread wild-animal suffering to the stars, continued human existence will be good for wild animals. Those concerned about wild animal suffering should support keeping humans around.

What is less obvious is whether biodiversity loss is good for wild animals. The impact of this will probably be primarily determined by the long-term impact of this accelerated loss, both on the assumption that humans survive for a long time and on the assumption that we don’t.

2 Do mass extinctions have long-term effects?

My naive intuition was that environmental effects—e.g. causing forests to become deserts—would be more significant in the long-run than ecosystem disruption. I guessed that ecosystems eventually bounce back from ecosystem disruption. Only environmental changes make a difference in the long-term.

Glenn

suggested something similar in his argument about why climate change will likely be very bad for wild animals, writing:

I'm assuming here that GHG emissions slow down or reach net-zero within ~100 years and the environment becomes more stable by the end of the millennium, and that means total biomass returns to normal (this is probably the most doubtful assumption). If it does, r-strategists who have disproportionately survived and who have short reproductive cycles can repopulate quickly. A lot of K-strategists would have been driven extinct and it takes a long time for them to diversify again because their reproductive cycles are so slow, so the strategist effect lasts a long time. At least, this is my impression of what we should expect after mass extinction events.

This was dead wrong!

It should’ve been obvious that ecosystems don’t automatically become productive. It took literally billions of years for any kind of diverse animal life to arise. There’s a lot of randomness in the evolutionary process.

But looking at the long-term effect of mass extinctions, it’s quite clear: these things have long-lasting effects. It takes about 2 million years for ecosystems to recover from mass extinctions. This article from UCL reports “Overall, it took 10 million years for species numbers to fully recover to previous levels.” They note:

This study highlights the extensive long-term risks posed by diversity loss which may result in highly unstable communities, loss of important ecosystem functions and long timescales of recovery.

Co-author Professor Paul Bown (UCL Earth Sciences) added, “The marine ecosystem is dependent on the plankton at its base, just as much today as in the past. What we see in the fossil record is that that you need the right players filling key roles for everything to function. By reducing biodiversity today, we run the risk of losing our critical ecosystem players, whose importance we have yet to fully appreciate.”

The study itself finds that ecosystems were majorly imperiled for about 2 million years following the the Cretaceous-Paleogene (K-Pg) extinction event that wiped out the dinosaurs. The authors (Alvarez et al) write:

Here, using a 13-million-year-long nannoplankton time series, we show that post-extinction communities exhibited 1.8 million years of exceptional volatility before a more stable equilibrium-state community emerged that displayed hallmarks of resilience. The transition to this new equilibrium-state community with a broader spectrum of cell sizes coincides with indicators of carbon-cycle restoration and a fully functioning biological pump⁹. These findings suggest a fundamental link between ecosystem recovery and biogeochemical cycling over timescales that are longer than those suggested by proxies of export production^7,8, but far shorter than the return of taxonomic richness⁶. The fact that species richness remained low as both community stability and biological pump efficiency re-emerged suggests that ecological functions rather than the number of species are more important to community resilience and biochemical functions.

Important takeaways:

1. Impaired ecosystem functioning can be long-lasting and hugely effects the productivity of an ecosystem. This can be brought about by majorly wrecking existing ecosystems. It doesn’t need to produce direct climatic change. (So my earlier intuition was right that ecosystem features matter, but wrong that these couldn’t be disrupted long term by wiping out large numbers of species).
2. Disruption to ecosystems lasts a really, really long time. Mass extinctions can depress ecosystem productivity for many millions of years.
3. The paper notes elsewhere that a decline in biodiversity can hugely disrupt ecosystem functioning. You don’t need to change environmental conditions directly—if you kill off enough organisms, you can majorly impact ecosystem functioning.

The conclusion that mass extinctions permanently neuter ecosystems isn’t just adopted by by Alvarez et al. It seems pretty uncontroversial among the experts. Another paper by Hull et al concluded:

For hundreds of thousands to millions of years after mass extinctions, a series of short-lived, low-diversity and (at times) low productivity ecosystems dominate^16,19,20. Large-bodied taxa often become dwarfed, or are replaced by small-bodied taxa^21,22

Hull et al further:

Three final attributes of past mass extinctions support the hypothesis of pervasive mass rarity. These features include the short-lived dominance of post-extinction taxa, the rarity of previously widespread habitats, and evidence for decreased primary productivity in the wake of extinctions

The good news is that mass extinctions have effects that last millions of years and that these lower productivity and diversity. The bad news is that the organisms that survive tend to be small, short-lived, R-strategists. My naive guess was that probably the first effect is much more substantial that the second effect. There are a few lines of evidence for this.

1. If the R-strategist effect from more hostile environments dominated their deleterious impact on environmental productivity, one would expect present environmental changes to increase arthropod populations. Exactly the opposite is true. Similarly, one would expect more arthropods in more hazardous environments—again, exactly the opposite is true.
2. Long-term impacts do seem genuinely massive. The Permian Triassic mass-extinction was so severe that there are no insects picked up in the fossil record for the next 15 million years. Now, undoubtably there were insects, but this goes to show the staggering and long-lasting impact that extinctions can have. And that is probably most comparable to the present extinction regarding its impact on insects, because only the current one and the Permian Triassic mass-extinction majorly wrecked insect populations. (Other arthropods have, of course, been majorly affected by mass extinctions, and presumably insects populations declined as a result of the drop in net primary productivity, even though they didn’t face a mass extinction).
3. If environmental productivity was the primary determinant of arthropod numbers, we’d expect the most productive environments to have the most arthropods. This is precisely what we observe! The main measure of environmental productivity is net primary productivity (NPP). The following chart by Brian Tomasik ranks ecosystems by NPP.

I asked chat GPT to rank these terrestrial environments by number of arthropods. While I don’t know exactly how much to trust these numbers—chat GPT had plausibly reasoning for them though—there is a great deal of correspondence.

If hazardous environments increase long-run arthropod numbers, we’d expect tons of bugs in the desert. But there aren’t—there are way more in rainforests.

We can get similar evidence for the deleterious long-term effect that mass extinctions have on long-term populations by looking at populations over time. This chart illustrates long-term diversity.

Fauna is outlined in blue, Paleozoic Fauna in red, and Modern Fauna in green.

And while diversity is quite different from the total number of organisms, the point is clear: a mass extinction event can have a long-lasting effect. Other charts indicate similar things—see this chart from Simoes (though I can’t figure out what the hell the colors mean):

Or this one from Prasad and Parmer:

Or this one (link now dead, unfortunately):

The last source has the ominous quote:

Thus, in the 4.5 billion years of Earth's history, life has never taken such diverse forms as during the recent period.

This gives rise to the troubling question: do mass extinction events cause long-term increases in biodiversity?

3 Does the world grow more diverse and productive in the wake of an apocalypse?

Consider this chart, once again:

During the Ordovician period, biodiversity was increasing rapidly, before the late Ordovincian mass extinction wiped it out. Biodiversity was pretty much flatlining during the Devonian period, until the late devonian mass extinction—after which things fell a bit, but then bounced back. The Permian-Triassic mass extinction was probably the best—it majorly reduced diversity without there ever being a recovery. The Triassic-Jurassic mass extinction, but life quickly bounced back and biodiversity subsequently shot up. There was a massive increase in biodiversity after that. The K-Pg mass extinction wiped out some of the diversity from the Triassic, Jurassic, and cretaceous period, but seemed to eventually generate a rebound.

During periods between mass extinctions, biodiversity grows at a roughly constant rate. You can see that on the above chart. This article from Berkeley notes:

Each period between mass extinctions was marked by a relatively constant, but different, diversification rate. Compare the idealized graphs below to the actual data above to see this pattern. After some mass extinction recoveries (e.g., after the End-Triassic extinction recovery), the rate of diversification is relatively slow, reflected in a gradually sloping line. After other mass extinctions (e.g., the End-Permian mass extinction), the standard rate of diversification is much quicker and new species are churned out at a rapid pace, reflected in a steep slope even after the initial recovery period.

Why would we observe this unexpected result? Krug and Jablonski suspect that it has to do with which taxa are most successful in the post-mass extinction period. If the taxa that take over and fill niches in the post extinction world (e.g., the mammals after the End-Cretaceous mass extinction) happen to be taxa that speciate easily, then overall diversification rates will be high until the next mass extinction shakes things up. On the other hand, if the taxa that take over after the extinction are slow speciators, then overall diversification rates will be low. This means that mass extinctions are much more important in shaping the diversity of life on Earth than ever thought before.

For each of the mass extinctions, if we assume the trend happening up until when they occurred would simply have continued if they hadn’t happened, and compare that trend to how things happened until the next mass extinction, four of the five mass extinctions look like they decrease long-term diversity. However, the Permian-Triassic may not have had this effect if not for the end-Triassic mass extinction occurring before it could reset biodiversity. The Permian-Triassic seemed to result in rapidly increasing biodiversity, while previously biodiversity had been flatlining. However, another mass extinction happened before things could bounce back.

However, the end-Triassic mass extinction seems to have massively increased biodiversity. Biodiversity almost doubled.

Taking all these into account, it seems like on average mass extinctions decrease biodiversity. However, this is uncertain and speculative enough that we shouldn’t be confident in it. The fact that present diversity is so high and is growing rapidly probably means a quick reset would be a good thing. Just as the end-cretaceous (K-pg) mass extinction decreased biodiversity by offsetting the rapid rise in biodiversity, the sixth mass extinction may offset the rapid rise in biodiversity of the present era.

All of this is pretty speculative, but I think the considerations in this section amount to a small point in favor of human activity.

4 Other points

So far I’ve discussed some effects that human activity will have. I’ve argued that in expectation majorly decrease ecosystem productivity, and this effect could last millions of years. Yay! Go us! I’ve also argued that this effect is probably quite significant—much more significant than the shift towards more R-strategists.

4.1 Strangelove oceans

One point in favor of mass extinctions is that they may produce Strangelove oceans. A Strangelove ocean is a “hypothetical state of ocean, almost devoid of living organisms.” Provided you think aquatic animals live crappy lives, a Strangelove ocean is a good thing!

However, Strangelove oceans are pretty speculative. It’s not clear that they followed any mass extinctions or how long they last. They certainly don’t follow most mass extinctions. As a result, this seems like a pretty nominal consideration.

4.2 Climate change

Humans are warming the climate. This could have a variety of deleterious long-term effects. The most terrifying prospect is that at higher temperatures, there are more insects. Antarctica, for instance, has only a single species of bug. In addition, higher temperatures tend to support higher net primary productivity. More carbon in the atmosphere can also support more plant growth.

Now, obviously climate change dramatically decreases short-term arthropod populations. But one might think that eventually arthropod populations will rebound—and be permanently higher as a result of humans’ altered climatic conditions.

While I take this argument seriously, I think it’s ultimately not enough to outweigh other factors.

First of all, warmer climates as a result of climate change will eventually reset when carbon is taken out of the atmosphere. If humans eventually reach net zero (which presumably we will) after about 1,000 years, things will have reset to a considerably degree. After about 100,000 years, emissions will have been fully reset. Thus, it will take a shorter time to rebound from climate change than environmental destruction. So climate change doesn’t dominate in the long-run.

Second, there are many mechanisms by which climate change reduces arthropod populations. (Note: much of my analysis comes from Brian Tomasik’s excellent article on the subject). Specifically:

1. Climate change decreases ecosystem productivity. Large numbers of organisms die off. While eventually organisms rebound, as we’ve already seen, it takes a long time for things to reset.
2. Climate changes causes lots of species to go extinct which increases ecosystem vulnerability.
3. Climate change triggers ocean acidification which decreases the number of organisms alive in the sea.
4. Climate change is the primary cause of coral-reef loss. As we’ve already seen, these have the highest productivity of any ecosystem.
5. Climate change itself drastically reduces insect populations—at least in the short term. Insects, well-suited for their current habitat, aren’t prepared for warmer temperatures. An insect’s range is the portion of ecosystems where it can inhabit. 49% of insects are expected to lose at least half of their range, and on average insects are expected to lose 43% of their range. More than 40% of insect species are declining.

There are many other factors to consider in analyzing the net impact of climate change on ecosystem functioning. All of this leaves me highly uncertain about its net effect. While on average I’d guess it to be slightly negative, there’s enough uncertainty that this seems like it definitely shouldn’t outweigh other considerations.

4.3 Extinction debt

Another mechanism by which environmental degradation may have positive long-term ecological effects is via something known as extinction debt. Organisms that limp through a mass extinction may be poorly suited for their new environment. While they can stick around for a while, their poorly designed features lead to them eventually dying out and to a less productive ecosystem. Hull et al write:

This evidence for mass rarity during past extinction events is surprisingly similar to the widespread rarity of previously common flora and fauna today. The modern ocean is full of ecological ‘ghosts’—taxa that are so rare they no longer provide past ecological services^36,38,92,93. Mass rarity includes local, often remarkable, declines in species abundance, as well as range contractions (as reviewed in refs 38 and 44). For those species with excellent historical and fossil records, like Caribbean corals, the recent population collapse contrasts with the marked resilience to past climatic perturbation^36,94,95. What’s more, the loss of species abundance is known to, at times, have cascading effects on ecosystem structure and function⁴⁵, and extinction debt may cause extinction hundreds⁹⁶ to millions⁹⁷ of years after an environmental perturbation. In this light, the paucity of extinctions in the oceans to date should not be viewed as a sign of the relative health of marine ecosystems^11,38—rarity itself may be the most direct metric of how close global ecosystems are to a permanent state shift.

Intuitively I found it surprising that a mass extinction millions of years in the past can cause extinction of species in the present. But apparently it’s true. I don’t know exactly how prevalent this factor is, but it’s clearly something.

4.4 Space colonization/terraforming/other expansion of nature

There are many ways that humans could conceivably drastically increase the extent of nature. Three in particular come to mind:

1. We could terraform other planets, spreading suffering across the galaxy. If humans become a multi-planetary species, this isn’t that unrealistic.
2. We could make a bunch of lab universes. Provided these universes keep expanding forever, this could cause literally infinite suffering.
3. We could create a great deal of suffering via digital minds. While present AI is likely not sentient, future AI might be. And we could cause it to suffer a great deal.

On the other hand, there are many ways the far future could go extremely well. These scenarios gain in plausibility from the fact that we look to be developing superintelligence soon, and humanities moral circle has been expanding over time. Superintelligent AI, wiser and better than us, is likely to have superior values. Specifically we could:

1. Create extremely large numbers of unimaginably well-off digital minds.
2. Stop existing wild animal suffering on Earth, and potentially even on other planets.
3. Undo processes that create lots of extra suffering.

Ultimately, it seems extremely unclear which is more likely. I’d lean towards the human species having positive impacts through sci-fi scenarios being more significant, in expectation, than negative impacts through sci-fi scenarios. However, I have a great deal of uncertainty about this.

More importantly, however, these considerations are not hugely relevant to the question of whether more humans is better. It’s not at all clear what impact an extra human has on extinction risk or number of expected suffering subroutines. So even if one is pessimistic about humanity as a whole, extra humans at the margin may still be a good thing.

5 Conclusion

Will humans reduce long-term animal suffering? It’s hard to know the answer, but I’d lean towards yes at maybe 65% confidence. I’d guess that we’re positive in expectation, even taking into account risks, but there’s loads of uncertainty. Ideally we should try hard to shape values so that we take wild animals seriously rather than just ecosystems.

This has staggering impacts. A dollar given to Givewell prevents about 14,000 years of insect life. But that’s just taking into account its short-term effects. When one takes into account the long-term impact of this, they could be hundreds or thousands of times greater.

There are about 10^18 insects. It looks like humans have collectively reduced insect populations by 10% according to quite conservative estimates. Let’s assume that the activities of presently alive humans will reduce the number of insects for the next million years by 1%. The number of insects we’ll have prevented from coming into existence by our activity so far is 10^17 (short term) + 10^22 (next million years). This means that each dollar given to Givewell prevents about a trillion years of insect life! And these estimates have been pretty conservative. It means that you personally might prevent 5 quadrillion insects from coming into existence throughout the history of Earth.

Utterly staggering if true!

It’s not clear exactly what impact this has on how we should give. Certainly it means Givewell charities have utterly staggering impacts when one takes into account their long-term effects. But likely so do other charities. Permanently reshaping industries for extremely long periods of time could be of similar importance.

In any case, this means that giving to Givewell is a pretty good bet if you think that animals live bad lives. Doing so might prevent a truly mindboggling amount of suffering.