It is customary to distinguish between an organism’s body and its environment. The environment is what exists outside the body. There is a definite boundary between the two. But how solid is this distinction? And does it matter to theoretical biology? Might there be a better way to carve things up?
Consider mollusks and their shells. Is the shell part of the body or part of the environment? We could say that the shell exists in the environment of its soft interior organism or we could reckon it to be part of the organism’s body. It seems arbitrary what we say and to be of little consequence. We might choose to say that the shell is not part of the organism’s soft tissue body, but is part of its overall body.  Should we say that the shell protects the body without being strictly a part of it? What about hermit crabs that scavenge the shells of sea snails? The shells perform the same kind of function as the shell of the oyster, but the physical connection is less rigid. This function is like the function of thick hide or the armor plating of some reptiles. Whether the protective outer layer is detachable is beside the point: the function is the same. It seems clear to me that there is no point in insisting that one sort of layer is part of the body and another is not. Thus we can introduce the concept of the extended body. We are familiar with the idea of the extended phenotype—the idea that things outside the body of the organism can be part of the phenotype of the organism, e.g. beaver dams and bird nests. I am suggesting that the body can be extended too, so the body is not as local as may be thought. We might distinguish between the restricted body and the extended body or introduce other distinctions based on other boundaries; the important point is that the notion of the extended body is a well-motivated notion.
Suppose we have an animal with a thick furry coat that it sheds in the summer and also carries around a large wooden container to sleep in and protect against predators. What belongs to its body? The coat has no feeling in it and is shed every year, so it has less claim to be part of the animal’s body than its fleshy innervated parts; yet it is not false to say this coat is part of the animal’s body. It isn’t part of what might be called the fleshy body, but it is part of a more extended body. Likewise the container is not spatially within either the fleshy body or the furry body, but it is part of a more extended body—what we might call the functional body. The container, like the furry coat, helps the animal survive—detachability is not to the point. There is no mileage in insisting that these things are not really parts of the body, but belong to the environment; there is no principled distinction to be drawn here. We can distinguish different types of body nested within each other, but we can also speak of the combination of all of them; and the latter is what corresponds to the biological notion of a functional unit. A good way to put the point is that the entities in question are (because they function as) organs of the body: the oyster’s fixed shell, the scavenged snail shell, the removable coat, and the portable container. Thus some bodily organs exist on the far side of the skin (not very surprising in view of claws and fingernails). We can call this the “extended body” in order to register the fact that other choices might be made about where the body ends and the environment begins; in reality, these “external” organs are just as much part of the body as kidneys, hearts or brains. There is no sharp theoretical line here.
How far outwards can we extend? Here things become trickier, partly because we don’t have any good existing examples to work from. So let us invent some hypothetical cases to focus our intuitions. Suppose an organism uses suction pads to pick up bits of the world that it uses in various ways—as weapons, as sun protection, as temperature control, as mate attractors. Suppose one of its tricks, genetically determined, is to scavenge fur covering from dead animals, which it dons in cold weather. I say that all this cargo counts as part of the animal’s extended body not part of its environment (inasmuch as that distinction has much content once the extended body is accepted). Such an organism might even pick up large chunks of the world for its use: trees, boulders, lakes, mountains. This super-organism would have an extended body that includes large tracts of the physical world, normally supposed part of the environment not the body. By the same logic as before its extended body would extend massively into the environment. What if there was an amazing bird that made its nest in a tree but carried the tree around with it when it moved? This bird would be just like the hermit crab that carries its home around with it. Accordingly, we could reckon the tree to the bird’s extended body: fleshy body, feathery body, and arboreal body. But why does the bird need to carry the tree around in order for it to be part of its body? Isn’t a stationary tree also part of a bird’s extended body? It uses the tree for its biological purposes, as it uses its beak, feathers, and nest. For a bird, a tree is an organ of survival—a device of gene propagation.
What about caves? If a bat lives in a cave, serving the same function as a shell, isn’t the cave also part of the bat’s extended body? The cave has the function of protecting the organism as far as the organism is concerned (though not as far as the cave is concerned): isn’t it arbitrary to introduce a sharp line between the cave and other features of a bat’s existence? The cave is in effect a tool the bat uses in order to survive, as the shell is a tool that the mollusk uses in order to survive. The same is true of burrows, dens, crevices, tunnels, and so on. These are all functional survival instruments, like organs of the body. By the same reasoning beehives and ant nests are part of the extended body of these creatures: they can detach themselves from these body parts, but that proves nothing—the same is true of hermit crabs. The bee and its hive function together to enable the bee to survive, just like its other organs; thus the hive is an organ of the bee’s extended body. Suppose the bee always took a mini hive with it whenever it left the big hive as protection and was never parted from its mini hive: wouldn’t we then naturally reckon it to the bee’s extended body? But the big hive is really no different, just more stationary. So the extended body can extend outward to spatially more remote locations. There is no conceptual problem with this: a deer might remove its antlers and leave them at home while going on a peace mission without thereby rendering them no longer part of its body (it reinserts them when it gets home).
Can we extend the body even further? What about an amphibian that carries its own water supply around with it? It produces a membrane that traps water around its body. This could be useful in the event of excessive evaporation. Wouldn’t this tank of water be functioning as an organ of the animal’s extended body? What about a true super-organism that could drag stars between galaxies as it hops around the universe, using them as sources of heat? Wouldn’t these stars be part of its extended body? It might actually swallow stars and keep them burning to bring a little warmth to the intergalactic trips. Could we even maintain that the extended body of an organism includes everything about its niche? Air for birds, water for fish, land for terrestrial animals: the extended body merges with nature as a whole. Take humans: we have conquered so much of nature, using it for our own purposes, converting it into tools (clothes, homes, motorways)—isn’t this all the extended human body? Where does the human body stop and the human environment begin? Once we accept that clothes are extensions of our body, where does it end? What belongs to the body is what the organism uses to achieve its biological purposes, and that includes a great deal. There is no clear theoretical distinction between internal organs like the heart and kidneys and external organs like shells, clothes, caves, and so on. From the point of view of biology the old distinction between body and environment is misguided and unnecessary. This means that anatomy can also be extended: the carapace is part of anatomy, but so is the shell, so is the spider web, and so is the cave or burrow. To our visual sense there seems to be a clear distinction between body and environment, because we naturally segment the world; but from a conceptual point of view the distinction is insignificant. The theoretical body is whatever web of things serves to enable the survival of the organism. The notion of the organism thus undergoes extension: the organism is the totality of things that are selected for or against (this is the extended phenotype). The web and the spider are co-selected, forming a whole—spider-combined-with-web. This extended body has various parts or organs, including legs, eyes, and web; the restricted body is just another part of the extended body. Spider anatomy must include web anatomy. The proper unit of biology is the extended body. If we spoke of the “corporeal make-up” of an organism, instead of its body, we would say that the corporeal make-up of a spider includes the web. To put it differently, the survival of the genes depends on the whole complex not just on the properties of the restricted body. The “survival machine” inside which the genes sit extends out to every functionally relevant thing. The genes don’t care about the distinction between two types of body, the extended body being all that matters to them.
Perhaps we don’t tend to think this way intuitively because we think of organisms as resembling ordinary physical objects, which don’t have functions and are not selected for. A rock is a discrete bounded object with a well-defined environment. There are no extended physical objects in the sense intended here; “bodies” in this sense are localized. They don’t have functional shells that aid in the struggle for survival. Physics doesn’t have to deal with the extended body. But biology is the science of functional units subject to natural selection, so its ontology is constituted differently. If we think of biological bodies as on a par with physical bodies, then we will be inclined not to see that the extended body is the right way to carve things up. There is a discrete physical thing that corresponds to the spider’s restricted body, and it is different from the physical thing that is the web; but that doesn’t mean that the biological body is so divided—there is a biological unit that unites these two physical objects. I have been calling this the extended body, but from a theoretical perspective we could just call it “the body” and then define more restricted units such as the fleshy body. The embodiment of species consists of spiders and webs, oysters and shells, bees and hives, bats and caves, people and technology, and so on. Each pair is an operative biological unit. We need to be more holistic about biological ontology, in contrast to the atomism of physics.
 Much the same could be said of cocoons: are they part of the body or not? The best thing to say is that the question has no clear answer because we have no definite notion of what counts as the body. We do better to distinguish different types of body corresponding to the same organism: thus the cocoon is not part of the butterfly’s flying body but is part of the butterfly’s metamorphosis body. The most theoretically useful concept of body would include the cocoon as part of the extended body.