# What the Butterfly Forgets
*On metamorphosis, continuity, and the quiet error of reading death as an ending.*
-----
Set two creatures side by side and they look like a study in contrast. The monarch butterfly, *Danaus plexippus*, lays single eggs on milkweed and nothing else, hatches a caterpillar that feeds for a week or two, hangs itself into a naked chrysalis, and emerges able to fly thousands of kilometers. The silk moth, *Bombyx mori*, lays eggs in clusters of hundreds, eats only mulberry, spins itself into a dense cocoon, and emerges so thoroughly domesticated that its adults can barely fly and cannot survive without us. One is wild and migratory; the other is a creature of human hands.
But the contrast is cosmetic. Underneath, both obey the same grammar — egg, larva, pupa, adult — and that grammar conceals something far stranger than the differences between species. Buckminster Fuller put it best: there is nothing in a caterpillar that tells you it’s going to be a butterfly. He meant it as a remark about prediction. I want to take it as a remark about identity, and then about death.
## The dismantling
We tend to picture metamorphosis as a caterpillar *growing* wings, the way a child grows into an adult — continuous, gradual, the same self enlarged. The biology is more violent than that. Inside the sealed chrysalis, much of the caterpillar’s body is enzymatically broken down into a loose cellular slurry. Suspended in that slurry are clusters of cells called imaginal discs, set aside earlier in the larva precisely so that they can survive the dissolution and reorganize the broth into an adult that the larva never resembled. The caterpillar is not so much improved as taken apart and re-spent. And the chrysalis is only the most dramatic seam, not the first: the caterpillar already hatched from an egg it cannot remember being, so the relay of forgotten phases runs backward as far as it runs forward.
That dismantling is why metamorphosis became one of humanity’s oldest emblems for the soul. If the caterpillar were capable of narrating its own experience, the chrysalis would not feel like a passage. It would feel like an ending. The body it knew is digested. The world it knew — the leaf, the chewing, the slow crawl — is gone. From the inside of that transformation, “I am ending” and “I am becoming something else” would be indistinguishable claims. There is no vantage point within the caterpillar from which to tell the difference.
## What crosses, and what does not
So does anything actually carry across? Here the experiments are more interesting than the metaphors. In 2008, Douglas Blackiston and colleagues trained tobacco hornworm caterpillars to avoid a specific odor by pairing it with a mild electric shock, then tested the moths that later emerged. The larvae learned to avoid the training odor, and that aversion was still present in the adult moths — and it did not come from chemicals lingering on the pupa, but from learning that survived the reorganization itself. Something of the caterpillar’s experience reached the moth.
But not everything, and not from everywhen. Larvae trained early, at the third instar, still avoided the odor as older larvae but no longer avoided it as adults — consistent with the idea that memories survive metamorphosis only when they are laid down in parts of the nervous system that persist into the adult brain. The bridge between phases is real but narrow. Most of the caterpillar’s life does not make the crossing. The adult moth flies off carrying a few retained associations and almost no acknowledgment of the long green life that produced it. It does not know itself as a former caterpillar. It simply *is* a moth, in a world that begins, as far as it can tell, at emergence.
Hold onto that asymmetry. The continuity is genuine. The recognition of it is almost entirely absent.
## A traveler who is never the same traveler
The monarch presses the point further, because its continuity is not even confined to one body. The famous migration to the oyamel fir forests of central Mexico is not made by a single butterfly. Four generations are involved in the annual cycle, and two to three generations complete the journey north, with females laying eggs for the next generation as they go. The summer generations live only two to five weeks; it is the fourth, the so-called Methuselah generation, that is physiologically distinct, entering reproductive diapause and living seven to nine months to make the long flight south.
The arithmetic is quietly astonishing. The monarch resting on a flower in a New England garden in May is the child or grandchild of the butterfly that left Mexico that spring. And the generation that flies *to* Mexico in autumn has never been there. The Methuselah monarchs navigate to overwintering grounds that their own generation has never seen, guided by an internal compass science still cannot fully explain. The “monarch migration,” spoken of as a single feat, is performed by a relay of beings, none of whom completes it, none of whom remembers the leg before their own, and the most decisive of whom flies unerringly toward a home it has no memory of.
If you were one of those butterflies, you would experience none of this as continuation. You would experience only your own brief life, with its single compulsion, in a world that started when you did.
## Made of what it eats
The relay runs sideways as well as forward — not only across the phases of a life but into what a life is made of. Watch a caterpillar strip a milkweed leaf and it looks like simple consumption, one thing destroying another. But the monarch carries the plant back out of the leaf: it takes up milkweed’s cardiac toxins and stores them, so that the plant’s poison becomes the butterfly’s own lifelong defense against predators. The monarch is, chemically, milkweed that learned to fly. And those poisons are themselves the slow precipitate of being eaten — plant and insect have shaped each other for so long, in the founding example of what biologists came to call coevolution, that neither is quite a self-contained thing. Each is partly the other’s history.
The same holds, more intimately, in the phase we all begin in. We picture the womb as a vessel the child merely draws from and then departs. But cells cross in both directions and stay. Fetal cells lodge in the mother’s blood, heart, liver, brain, and skin and can persist there for decades — present whether the pregnancy ended in birth or in loss — while maternal cells likewise remain in the child. The host keeps a piece of the guest for the rest of her life; the guest carries the host. So the single traveler the relay lacks across time, it also lacks at any single moment. A monarch is not one bounded thing passing through milkweed; it is milkweed, sun, and air briefly standing up as a monarch — the surroundings, for a season, in the shape of a creature.
## Continuity without a continuer
This is the place where biology and two older bodies of thought quietly converge.
The Buddhist tradition has insisted for two and a half millennia that there is no permanent, unchanging self threading through experience — only a stream of conditioned arisings, each conditioning the next, with no owner riding inside them. The classical image, from the dialogues of King Milinda, is the flame passed from candle to candle through the night: the late flame is neither the same flame as the first nor a different one, and the question “which is the *real* one” is malformed. What we call a person is a process that produces the convincing impression of a thing. Rebirth, in this reading, was never the migration of a soul-substance from body to body; it was causal continuity without a continuer — exactly the monarch’s relay, with the same absence of a single traveler and the same absence of memory across the seams.
It is worth being exact about what this does and does not buy. Nothing here licenses the claim that anything recognizably personal — your memories, your character, the felt center of *you* — survives what we call death. It undercuts only the narrower and more stubborn move: the slide from *no one on the far side remembers the near side* to *there was nothing on the near side at all.* The first is an observation; the second is a conclusion that does not follow from it.
Modern physics arrived, by a completely different road, at a structurally similar humility about “things.” In quantum field theory, what we naively call a particle is not a tiny enduring marble but a localized excitation of an underlying field — a pattern in something more fundamental, not an object in empty space. Two electrons are not two little objects that happen to match; they are indistinguishable excitations of the one electron field, and the “same” electron persisting through time is a continuity of pattern, not the survival of a substance. Even the vacuum is not nothing: it is the field’s lowest state, restless with fluctuation, the ground from which excitations arise and into which they subside. Persistence, on this picture, is a stable pattern in a substrate that itself does not begin or end where the pattern does.
A flame, an excitation, a relay of butterflies, a monarch made of milkweed. In each, what we point to as a lasting, self-contained thing turns out to be a pattern carried forward — and held up from the sides — through transformations that destroy any single carrier.
## The forgetting between phases
Which brings us, finally, to the error.
We treat death as cessation. But look closely at what that judgment actually rests on. We do not *observe* cessation — no one observes their own nothing. What we observe is the absence of acknowledgment: the adult does not report being the larva; the moth does not file the caterpillar under “my former life”; the butterfly heading to Mexico carries no récollection of the mother who turned back two generations ago. Every successful metamorphosis in nature arrives precisely without ancestral acknowledgment of the phase before it. The later stage does not remember being the earlier one, and frequently could not have, because the structures that did the earlier living were dismantled in the passage.
So consider what this does to our reasoning about death. The single piece of evidence we have for “death is an ending” is the same piece of evidence a butterfly would have for “the caterpillar’s death was an ending” — namely, that nothing on the far side announces itself as a continuation of the near side. But we have just watched that exact silence accompany transformations that were *not* endings. In the monarch, in the moth, the non-recognition of the prior phase is not a sign that the prior phase terminated into nothing. It is the ordinary signature of having crossed into a phase that does not carry the previous one’s memory.
This is the quiet mistake. We read the absence of an ancestral acknowledgment as proof of cessation, when in every transformation we can actually inspect, that same absence sits comfortably on top of continuation. The caterpillar would be entirely right that *the caterpillar* ends. It would be entirely wrong that *everything* ends. And it would have no way, from inside its own dissolving, to tell which conclusion it was entitled to.
I am not claiming that death is rebirth, or that the analogy proves anything about what follows our own last phase. The honest position is narrower and, I think, more durable: cessation is not something we ever witness. It is an *inference* we draw from non-recognition — and non-recognition is exactly what a continuation through radical transformation would also produce. The two are empirically identical from the inside. What we are entitled to say, then, is not “death is an ending.” It is the more accurate and more open phrase: *unknown continuation.* The next phase, if there is one, would by its nature arrive without acknowledging the one before — just as the moth flies off unaware of the leaf, and the autumn monarch turns toward a country it was never told it had visited.
There is nothing in a caterpillar that tells you it is going to be a butterfly. There is also nothing in a butterfly that remembers having been a caterpillar. Both silences are real. Only one of them is the end of the world — and from inside either, you cannot tell which.
*A companion essay, “The Reader and the Cloud,” takes up the question this one leaves open — if not a continuer, then what continues? — through genetics, information theory, and the old Buddhist intuition of a storehouse that holds without anyone keeping it.*
-----
### References
- Blackiston, D. J., Silva Casey, E., & Weiss, M. R. (2008). Retention of Memory through Metamorphosis: Can a Moth Remember What It Learned As a Caterpillar? *PLoS ONE*, 3(3): e1736. <https://doi.org/10.1371/journal.pone.0001736>
- Monarch migration generational cycle and the Methuselah generation: *Monarch butterfly migration*, Wikipedia; “Word of the Week: Monarch,” *Berkshire Eagle* (2025); “Monarch Butterflies: Migration, Breeding, and Survival Strategies,” ScienceInsights (2025); Natural Habitat Adventures, “Monarch Butterfly Migration” (2022).
- Fuller quotation as cited in *National Geographic*, “Moths remember what they learn as caterpillars.”
- Monarch sequestration of milkweed cardenolides as a predator defense: see overviews in the monarch-biology sources above. The coevolution of butterflies and their host plants follows P. R. Ehrlich & P. H. Raven (1964), “Butterflies and Plants: A Study in Coevolution,” *Evolution*, 18(4): 586–608.
- Fetal and maternal microchimerism: Boddy, A. M., Fortunato, A., Wilson Sayres, M., & Aktipis, A. (2015), “Fetal microchimerism and maternal health,” *BioEssays*, 37: 1106–1118; and reviews of the persistence of fetal cells in maternal tissues for decades following any pregnancy outcome.
- Philosophical sources drawn on generally: the *Milindapañha* (the simile of the flame and the chariot) on continuity without a permanent self; standard quantum field theory on particles as field excitations and the non-emptiness of the vacuum.