An interactive argument map

The Argument from Pain (AP)

A novel version of the argument from evil. What classical theism cannot account for, I argue, is not evil in general but the felt awfulness of physical pain: a world with all the protection and none of the hurting is one an omnipotent, wholly good God could – and would – have made instead.

Drawn from my M.A. thesis, “The Problem of Evil as the Problem of Pain” (the basis for the book manuscript I’m now developing), incorporating the argument’s later revisions. Thirteen steps: seven basic premises, five derived, one conclusion. Throughout, pain means the phenomenal experience of physical pain (or qualia) – the felt quality of injury, not grief, loss, or the other things the word is often linguistically stretched to cover. Tap a step to open its defense; the Premise 4 control is the hinge, so toggle it and watch the chain.

1. 1
   Basic · theistic commitment

   God is omnipotent.¹

   ›Note on omniscience

   Footnote 1. Sometimes the problem is formulated with a separate statement of God’s omniscience, but strictly speaking this is not required: omniscience is generally (and correctly, I think) taken to be an ability or power, and so is properly subsumed under omnipotence.

2. 2
   Basic · theistic commitment

   God is omnibenevolent.
3. 3
   Basic · the datum

   In the actual world, W, sentient creatures sometimes have the phenomenal experience of pain.
4. 4
   The hinge premise · conceivability → possibility

   If a state of affairs is ideally conceivable, then it is prima facie metaphysically possible (absent defeaters).

   Accepted – the argument runs

   With the inference in hand, Premise 5 is licensed and the chain reaches its conclusion: no morally sufficient reason for phenomenal pain is even possible. And the modal methodology it relies on stays intact.

   Rejected – and here is the cost…

   Block the inference and the argument stalls at Premise 5. But notice how little I ask of it: Premise 4 does not say the conceivable is possible, only that ideal conceivability is defeasible, prima facie evidence of possibility, absent a defeater. To reject the argument this way, one has to deny even that cautious a claim: a heavy price for so modest a premise.

   And, of comparable significance for the theist, the same inference is the engine of:

   • the zombie argument against physicalism,
   • the knowledge argument and conceivability-based anti-physicalism generally,
   • much of the modal methodology dualists rely on.

   To reject Premise 4 in order to escape the argument is to give all of that up as well. Hence the unhappy result: theists must reject most arguments for dualism.

   ›The modal vocabulary

   Ideally conceivable

   Conceivable under ideal rational reflection – to a thinker free of confusion, ignorance, and performance error. It excludes cases where we only seem to conceive something (e.g. water not being H₂O) through failing to think it through.

   Absent defeaters

   Unless there is good reason to think otherwise: a proof of incoherence, an empirical impossibility, a hidden contradiction. The principle yields defeasible justification, not proof.

   Metaphysical possibility

   What could obtain given the essences of things. Narrower than logical possibility (water being other than H₂O is arguably logically possible but metaphysically impossible), and the modality that matters for what God can actualize.

   The principle states conceivability as prima facie (defeasible) evidence for possibility, a treatment well precedented in Yablo, Hill & McLaughlin, and Kung.

5. Licensing edge. Premise 4 does not sit in the chain; it warrants the modal claim in Premise 5. This is the single most important link in the argument; the whole hinge lives here.
6. 5
   Basic · the possibility claim

   There is a metaphysically possible world, W*, qualitatively indiscernible from ours except that sentient creatures do not have the phenomenal experience of physical pain.²

   ⚠ With Premise 4 rejected, this modal claim is no longer licensed – the chain breaks here.

   ›Formulation, support & objections

   Formulation note

   Footnote 2. W* is qualitatively discernible from W only in that (1) W contains instances of physical pain and W* does not, and (2) certain psychophysical laws in W connecting phenomenal physical pain to its physiological correlates are missing in W*. (I owe a debt to Joseph Tolliver for pressing the importance of formulating the premise this way.)

   Empirical support

   Congenital insensitivity to pain

   Tissue-damage response comes apart from pain phenomenology in actual human beings: the protective signal without the felt hurt.

   Sandroni et al. (2006) report a patient who cannot perceive pain despite entirely normal small-fiber count, appearance, and distribution, and no detectable abnormality anywhere in the peripheral or central nervous system: the phenomenology is absent while the physiology the identity theorist points to is intact. The philosophically decisive detail, though, is not that one. Unlike virtually every reported case of congenital insensitivity, this patient showed none of the stigmata of repeated injury; with his other sensory modalities unimpaired, he had learned over time to avoid serious harm by vigilance. That closes off the theist’s most natural reply to W*: that a world without phenomenal pain would be a world of maimed and short-lived creatures. Insensitivity to pain carries no such price as a matter of necessity, and so the wretchedness we (rightly) associate with the classic cases cannot be charged to W*.

   Pain affect without pain sensation

   The patient studied by Ploner et al. (1999), following a lesion of the primary and secondary somatosensory cortices, detects nociceptive signals and responds aversively to tissue damage while lacking the phenomenal hurtfulness. He denied every offered description, “slight pain” and “moderate pain” among them, and could not report quality, localization, or intensity: what survives the lesion is aversion, not hurting. If the dissociation is actual, it is a fortiori metaphysically possible; so the possibility claim need not rest on conceivability at all, and the controversy over Premise 4 can be sidestepped.

   The price of going empirical. To produce that effect globally rather than in one patient, God may have to build our somatosensory cortices slightly differently, so the premise weakens: W* now differs from W in neuroanatomy as well as in phenomenology. The concession costs nothing. To make it matter, the theist would have to hold that the precise construction of our somatosensory cortices is itself a good great enough to outweigh every pain it underwrites.

   Objections to W*

   The grief-transcription objection

   Bereavement grief can be transcribed into physical pain independently of peripheral nociception – physical-pain qualia with no injury. Doesn’t that threaten the clean separation?

   It cuts the other way. It supports Premise 5. It shows the affective pain system generates aversive phenomenology without nociceptive input, which is precisely the dissociation W* requires.

   The neural-overlap objection

   Physical and social pain share circuitry (dACC, anterior insula, opioid systems); social connection is among the greatest goods; so perhaps physical-pain phenomenology is a byproduct of architecture required for those goods, and has a morally sufficient reason after all.

   Nomological, not metaphysical. The overlap reflects the pathways evolution happened to take. Omnipotence is not bound by evolution’s incremental history; a designer is not.

   Separability. Physical and social pain have qualitatively distinct phenomenal characters (burning, aching vs. loneliness, rejection). If dissociable in character, dissociable in principle: God could route the phenomenology from social inputs alone.

   The Ploner bind. The Ploner patient plausibly keeps social-pain capacity while lacking physical-pain phenomenology. Call that pathological and one concedes the connection is contingent; and what pathology can break, design can avoid.
7. 6
   Basic · value premise

   Because pain is a bad-in-itself, a world containing no pain is better than a world containing pain when the two are otherwise identical.
8. 7
   Basic · conceptual premise

   Something that is bad-in-itself cannot be good-in-itself.
9. 8
   Derivedfrom 6, 7

   Pain is not a good-in-itself – and so, a fortiori, not a divinely-desired good-in-itself.
10. 9
    Derivedfrom 3, 5, 6

    [ Therefore, W* is a better world than W. ]
11. 10
    Derivedfrom 1, 2

    If God exists, God can and does actualize the best metaphysically possible world compatible with all divinely-desired goods-in-themselves.
12. 11
    Derivedfrom 5, 6, 8

    If pain exists in the actual world, then God has chosen not to actualize that best world.
13. 12
    Derivedfrom 3, 11

    God has chosen not to actualize the best metaphysically possible world compatible with all divinely-desired goods.
14. 13
    Conclusion · modus tollensfrom 10, 12

    It is not the case that God exists.