Is there more suffering than enjoyment in nature?

Open draft. This project is being developed in the open. The text is incomplete and the code uses the planned GaiaBasic API, so it will change. Comments and collaboration are welcome on GitHub.

The question

Is the total amount of suffering in wild nature greater than the total amount of enjoyment? The question matters for how we think about nature, about wild animal welfare and about the priorities of anyone who wants to reduce suffering.

It sounds like a factual question, but it hides many choices. What counts as a unit of suffering? Does a short life count the same as a long one? Does a newborn that dies after a few hours feel anything? Does intensity grow with brain size? Two people can give opposite answers simply because they made different choices without noticing.

Background

Philosophers such as Oscar Horta have argued that suffering prevails in nature. A key argument is reproductive strategy. Many species are r-strategists: they have huge numbers of offspring, and most die soon after birth, often painfully. Other species are K-strategists, with few offspring and more parental care. In a lion population, for example, only about one cub in five reaches the age of two.

Yew-Kwang Ng argued in 1995 that suffering dominates in nature. Later work by Ng and Zachary Freitas-Groff (2019) found an error in the original mathematics, so the result is less certain than it seemed. Other authors have asked whether suffering is even a useful signal when death is almost certain, and whether pain while dying can be shaped by natural selection at all.

Some cases are disturbing and informative at the same time. Cicadas infected by the Massospora fungus keep mating while their bodies are being destroyed, apparently drugged by the parasite. Whether they suffer depends on what we think about insect sentience.

Assumptions that change the answer

Assumption	Options to compare
Who is sentient	Only vertebrates · also insects · also gametes and embryos (relevant for panpsychists)
Intensity	Equal for all · scales with neuron count · scales with brain size (an elephant or a pilot whale could feel two or three times more than a human)
Early death	Newborns feel from birth · sentience develops gradually
Dying	Full suffering until death · reduced suffering once death is certain (a hypothetical turtle that "switches off" when recovery is impossible)
Does experience matter for behaviour?	Yes (it motivates) · no (it is an epiphenomenon) · p-zombies possible or impossible
Cost	Experiences are free · experiences cost energy

Design in GaiaBasic

An ecosystem preset with a mix of r- and K-strategist species is run many times, once for each combination of assumptions. Instead of asking "what is the answer?", the study asks "which assumptions flip the sign of the answer?".

import gaiabasic as gb

study = gb.Sweep(
    sentient_taxa=["vertebrates", "animals", "animals+gametes"],
    intensity_scale=["flat", "neurons", "brain_mass"],
    newborn_sentience=[True, False],
    suffering_when_dying=["full", "reduced"],
    experience_cost=[0.0, 0.05],
)

for cfg in study:
    w = gb.presets.Ecosystem(r_strategists=0.9, k_strategists=0.1, seed=1)
    w.sentience = gb.Emergentism(
        taxa=cfg.sentient_taxa,
        intensity_scale=cfg.intensity_scale,
        from_birth=cfg.newborn_sentience,
        dying=cfg.suffering_when_dying,
        cost=cfg.experience_cost,
    )
    study.record(cfg, w.run(steps=100_000).valence_balance)

study.sensitivity()   # which assumption changes the sign of the result?
study.plot("valence_balance", by="intensity_scale")

What would count as a result

A map of which assumptions make suffering dominate and which make enjoyment dominate.
A clearer debate: people who disagree can see exactly which assumption they disagree about.
Detecting incoherent combinations, for example worlds where sentience has an energy cost but no effect on behaviour and still survives selection.

Open problems

How to measure intensity in a way that is not arbitrary.
How to compare a long, mild discomfort with a short, intense pain.
How to get realistic numbers for birth and death rates for many species.

Status

Concept and simulation design. Next step: implement the required GaiaBasic components and publish the code and results here.