The Case of
The Lively Desert
shows how difficult it may be to have good models of wholes without having good models of parts.
Back Contents Comments Next

The previous case introduced the life of the coral reef and brings us to the example of an ecological mystery based on the division of plants and animals in the life of the reef.

The mystery was that the coral reef is, ecologically, a desert. When ecologists looked at coral reefs in the South Pacific (funded to understand how nearby nuclear explosions affected the plants and animals), they got off their ships or planes with a model called the trophic pyramid which is based on set of relationships called the food web. The food web was a graph of "what ate what": (Placeholder image from

which led to a division of the organisms of the ecology into a pyramidal distribution: (Placeholder image from

In this model, solar energy arrives at the bottom of the pyramid and supports the growth of plants, which are in turn consumed by small herbivores, which are in turn consumed by larger carnivores, and so on, up to the sparsely populated top of the pyramid. (That's where most of us are.) Central to this account was 'what ate what' and the fact that plants, which relied on photosynthesis and free-floating chemicals, didn't "eat" anything.

However, when they examined the coral reef closely, these ecologists counted the plants and found that there weren't enough. This isn't obvious at first, because the corals themselves are animals (they eat other organisms) but they live in colonies which look like plants. (As we saw in The Case of The Alien World.)

The number of actual plants on a coral reef is more consistent with a desert than the sense of rainforest-like abundance we find on coral reefs. In fact, the clarity of the water which makes the reef so delightful to divers and photographers is in part a consequence of these "desert" conditions. In colder waters, much of the plant life consists of photoplankton suspended in the water, reducing visibility. These are much less common in tropical waters, increasing visibility. In the clear waters of the coral reef, the trophic pyramid, which scientists had verified in ecologies all over the planet, did not seem to apply.

Now it might be that they were not counting right and that their observations were incomplete. Indeed, upon "deeper" examination, the scientists found that an appreciable density of plants were hidden under the sands, but even including these buried plants did not turn the trophic pyramid the right way around.

Inner Secrets

The mystery was solved a number of years later by microscopic examination of the coral itself. The tissues of each individual coral creature (each smaller than a thumbnail) contained myriad small plants called zooxanthella which used photosythesis to produce some of the materials upon which the coral relied. Each coral creature, in some sense, carries a compact and portable farm within its body, using it to provide some of the ingredients for life in the relative desert of the coral reef. When the zooxanthella were taken into account, the trophic pyramid upended, returning to its expected state.

The ecologists originally began to examine the coral reef with a certain model of what constituted an ecology. Part of this model involved an understanding of individual organisms, part involved accounts of how those organisms functioned, and part involved how they interacted. Given this characterization, experience led them to expect that an ecology divided in this way would show certain characteristics, particularly a distribution of plants and animals described by the trophic pyramid.

What were the scientist's options when the data didn't fulfill their expectations? They could have just said "the trophic pyramid holds sometimes," but that would not have been very satisfying --- more of a description than an explanation. This was especially true because the trophic pyramid had a certain logical neccessity: the energy for an ecology has to come from somewhere and the only obvious source of free energy is the sun, which only plants can access directly.

Another possibility would be to look at the categories of animal and plant and rearrange them so that the sessile coral was classified as a plant rather than as an animal. This way of "cooking the books" is sometimes reasonable when the reorganization reflects principled categories. We saw it when hydras and coral polyps were first classified as animals to explain observations of their feeding behavior. In the case of Temberly's reclassification of polyps, such changes are illuminating, but in the coral ecology, the definition "animal" was clearly based on the consumption of other organisms and coral clearly satisfied this definition. (One could watch them eat).

A possible compromise position, not unlike what was really taken, is to merge the categories of plant and animal, and say that some animals can photosynthesize. Of course, this would require an explanation of a) how they photosynthesize and b) why most creatures don't photosynthesize (since it would seem like a valuable attribute).

But the real and surprising answer was to redefine what was meant by organism. The scientists arrived with the assumption (which we noted above) that an organism fills a region of three-dimensional space and that these regions could be uniquely classified as plants or animals. The discovery of zooxanthelle showed that two organisms could share the same region of space yet be ecologically distinct. With this understanding, the way the idea of organism referred changed from space to function.

Familiar but Distinct

The idea of two organisms sharing the same space was not entirely new. For decades, scientists had understood that bacteria living in human or animal guts were independent organisms. Especially for medical purposes, this was a common and useful model.

But even though the model was probably in the scientist's repertoire, the striking difference was that this redefinition of the ecological organism could change the entire shape of the ecology. And the shift belies our common sense idea of what an organism (and we ourselves) are.

Reference changes are not always this radical, but the tricky thing about them is that they complicate the transfer of knowledge. Knowledge in a model does not always transfer when the rules of reference change. We cannot know what things we know about coral are actually about coral and which are about zooxanthella and which are about coral plus zooxanthella. For example, before the change in reference, scientists knew that coral could easily survive in nutrient-poor water (the reef). And they also knew that coral was an animal and so would be largely unaffected by the amount of light it received. However, the inference that we can reduce both nutrients and light for the coral is not true in the new theory, since its ability to do with less nutrients relies on the reception of light by the zooxanthella.

Looking in the Light

A cop comes across a man on his knees, obviously searching for something at the base of a streetlamp. "Lost something?" she says. "Yes," the man says, "I lost the keys to my car." The cop pulls out a compact metal detector and says "This might help. Where did you lose them?" The man, a little shamefaced, points down a dark alley. "By the dumpster over there." The cop looks surprised. "Then why are you searching around this lamppost?" The man grins uncomfortably and answers "Because the light is better here!"

We laugh at this story because searching for a lost object cannot be done without regard to where the object might be found. While it might make sense to count change or read directions where the light is better, searching for something where the light is better doesn't help unless the light has a chance of illuminating the object.

However, models are sometimes based on quite similar reasoning. Through their patterns of reference and identification, models often move our tasks to "places where the light is better." Sometimes, these transformations suit the task at hand, in which case the model functions well. But sometimes, they leave behind important elements of the problem to be solved.

Reference is tricky because it involves "pointing" with our bodies or our instruments. Like a search for lost keys, it is a kind of looking. One of the reasons that experimentation is so important to scientific progress is that experimentation is how reference changes and evolves. The huge shifts in biological understanding during the first half of the 1800's were due, in part, to the physical explorations of the planet which revealed a diversity of kinds of creatures never before imagined. Likewise, the last two centuries of progress in physics is due (in part) to the breakneck pace of industrial and technological growth allowing experiments of unprecedented precision at unprecedented scales (both large and small).

New computer technologies of simulation and presentation open up the possibility of exploring "virtual realities" based on our theories and imaginations. But we cannot forget that the most radical and important reformulations in science come from looking and pointing to what we didn't expect and would have never have found in our virtual worlds.

The value of such virtual world technologies is not that they let us do experiments we could not do otherwise, but are that they let us look at our theories and models in ways which would normally be impossible. In the language of this book, such "virtual reality" technologies make our models more convenienent in profound ways. But these technologies do not, in any basic way, change the patterns of imagination or reference they involve.

Copyright (C) 1997, 1998 by Kenneth Haase
Draft, not for citation or circulation
Back Contents Comments Next