The fixedness of body plans varies widely across different types of Earthling organisms. At one extreme, you have things like tardigrades, for which every individual of any given species has exactly the same number and arrangement of individual cells, differing only in size; at the other extreme, you have things like mycelial fungi, which don't really have any consistent large-scale shape.
At various points in the middle of the spectrum, you have typical animals and plants: all animals of a given species tend to have the same high-level shape (e.g., number and arrangement of limbs), even though they differ in small-scale details and may be different sizes, while plants tend to be similar in terms of the shapes of specific organs (leaves, branching structure, etc.), but can modularly assemble those mid-range features into wildly different large-scale structure--i.e., it is easy to find pairs of animals that are nearly identical to each other, but good luck finding two trees that grew in exactly the same shape! This modular construction conveys numerous advantages to plants, most notably the fact that they can sustain massive injuries and still survive--lop a third of the limbs off a typical tree, and it won't care, 'cause it's got or can grow spares. Lop off a third of a dog, and, well... you've got an animal cruelty case and a lot of blood to clean up.
So, can we make a plant-style modular-type body plan work for more animal-like creatures? How would something like that evolve?
A modular animal must:
- Be mobile.
- Be heterotrophic--whether vegetarian or carnivorous doesn't really matter.
- Be constructed largely out of repeatable, interchangeable, and redundant organ complexes, such that damage and regrowth is both possible and expected.
- Have its detailed large-scale shape determined in large part by environment and injury history, not fixed by genetics.
And for the sake of narrowing the scope:
- Live on land.
- Exist in a size range typical of mammals--say, somewhere between a housecat and an elephant.
This does not, however, necessarily mean that a modular animal can't have critical specialized organ systems, like a single head or single digestive tract; after all, separate the crown of a tree from its roots, and it will die (unless each separate part is in good conditions to regenerate the missing half, of course)--each part may be modular, but that doesn't mean the different parts are arbitrarily divisible.
A list of all Anatomically Correct questions can be found here: Anatomically Correct series.