Hypotheses: positive, negative, and the role of testing.

As noted on the previous page, the essential defining characteristic for an endeavor to be considered 'science', is the ability to generate testable or predictive hypotheses. This requirement is universal across all types of science, and evolutionary biology is no exception. On this page, we will see numerous examples of predictive hypotheses (sometimes presented as outright predictions) within evolutionary biology. This will be followed by consideration of why such predictions are impossible from the Intelligent Design stance. The cartoon at right illuminates the view that many mainstream scientists have of Intelligent Design -- because it cannot provide testable hypotheses, many see ID as in the same category of pseudo-science as alchemy, phrenology, astrology, or simply magic. Thus the calls by ID proponents to 'teach the controversy' in evolutionary biology rings exceedingly hollow in the minds of most mainstream scientists, because the only controversy over the existence of evolution is the controversy that has been created by the ID movement.

Predictive hypotheses and predictions in evolutionary biology

On the right is a diagram showing the pelvis and pelvic spines (seen from below) of sticklebacks, small fish that live in both oceanic and freshwater environments. Sticklebacks have apparently colonized many freshwater lakes in the northern hemisphere after the retreat of ice age glaciers, and the freshwater forms have greatly reduced or absent pelvic spines. Work published in 2004 identified the gene that controls most of the loss in pelvic spines. Because the loss of pelvic spines has occured in several species of sticklebacks, Shapiro, Bell & Kingsley predicted that a similar genetic mechanism should explain the morphological change in related species. Their paper, published in 2006 in Proceedings of the National Academy of Sciences, presented evidence that indeed, the same gene, Pitx1 is responsible for pelvic spine loss in both Gasterosteaus aculeatus and Pungitius pungitius. Shapiro and colleagues had presented a hypothesis that could have been wrong, but they then tested the hypothesis with real data and confirmed their prediction -- the same gene was responsible for most of the spine reduction in BOTH fish species. If a hypothesis has NO possibility of being wrong, it is worthless, since it cannot be tested.

The lower image to the right shows an astonishing fossil find from 2004 that resoundingly demonstrates the power of hypotheses for advancing science. Several years ago, Dr. Neil Shubin and colleagues began fossil-hunting in a remote area of northern Canada, on Ellesmere Island. Shubin, Edward Daeschler and colleagues reasoned that since the other known fossils that are transitional between fish and tetrapods (animals with legs, that can walk on land) were known from a period about 370-380 million years ago, any further transitional forms were likely to come from that same period. They further reasoned that such transitional animals would likely have lived in shallow freshwater habitats, such as lagoons or broad shallow rivers. On the basis of this reasoning, they searched for placed with exposed rocks that were 370-380 million years old, and which were formed in shallow freshwater conditions. Then, in 2001, they started digging. Their hunt produced little reward the first several summers, but in 2004, they found several almost-complete, mostly-articulated skeletons of a previously unknown species. They named it Tiktaalik roseae. It clearly filled a transitional gap between more fishlike forms and more tetrapod-like forms. Tiktaalik clearly was an example of formulating hypotheses about where (and when) to look, and finding the predicted transitional fossil. Shubin and colleagues could have been wrong in several ways -- they might have found Tiktaalik in rocks of very different ages than what they expected, or they might have found fossils of very unexpected kinds of organisms in the 370-380 million year-old rocks. Any of these things would have shown their hypothesis was wrong, but in fact what they found actually CONFIRMED their hypothesis. Again, the key distinction here is that they made a testable prediction on the basis of known scientific knowledge, and when tested against the real data, the hypothesis was confirmed. Sometimes hypotheses are rejected, and of course that is fine, and entirely consistent with how science operates.

On the right is a photo of an important insect fossil. This fossil, discovered in 1967, shows a very primitive ant encased in amber, or fossilized tree sap. Fossils in amber are particularly valuable because of the details they preserve. This fossil is 90-95 million years old...far older than any ant fossil previously known. Thus it is important in extending the fossil record of ants much further back in time. But that is not its only importance: this fossil was predicted. In work in the 1950s, E.O. Wilson of Harvard University proposed that since ants were apparently evolutionarily derived from wasps, the most primitive ants should show a range of traits intermediate between wasps and modern ants. At the time Wilson made his prediction, no fossils were old enough to illustrate the transition that he forecast, and in fact, it was more than a decade before the confirming fossil shown here was discovered. Strikingly, all traits but two of this fossil conformed to Wilson's prediction, thus strongly supporting the ants-from-wasps hypothesis.

Here we have a diagram showing the reconstructed skull of a very important fossil animal along the reptile-to-mammal lineage. Since the mid 1800's, on the basis of morphology, experts had suggested that mammals were derived from reptiles. Of course, mammals differ from reptiles in several ways, but a crucial difference is the nature of the lower jaw hinge: in reptiles, a bone called the squamosal artciulates with the rest of the skull, but in mammals, the jaw is reduced to a single bone, the xxxxx, which articulates with the skull. On the basis of this difference, in 1912, paleontologist Robert Broom of South Africa predicted that there must have been a transitional form in the evolutionary lineage, which had a double jaw articulation. It wasn't until nearly 50 years later that Broom's prediction was confirmed. The fossil Morganucodon, which is clearly an intermediate along the reptile-to-mammal lineage, has a lower jaw that articulates with the skull in two locations, just as Broom predicted.

Why Intelligent Design cannot generate predictive hypotheses.

Scientific hypotheses are made on the basis of the operation of some known or unknown mechanism. Thus if I know that eating a gallon of mint-chocolate-chip ice cream gives people a stomach-ache, I might hypothesize that the reason my college roommate got lots of stomach-aches was because he frequently binged on huge amounts of mint chocolate chip ice cream. The key operative factor is that we have some reason to make a particular prediction -- nature operates in a certain way, and on that basis we make our hypothesis. Unfortunately for ID, there is no basis for making hypotheses about what to expect in nature, because the ID proponents say that the Designer is unpredictable ('whimsy' is a term that has been used several times). Below are several examples.

Phillip Johnson, 1991, from Darwin on Trial: "It seems to me that the peacock and peahen are just the kind of creatures a whimsical Creator might favor, but that an “uncaring mechanical process” like natural selection would never permit to develop."

William Dembski, 2001, from a post on Access Research Network: "But what about the predictive power of Intelligent Design? ....To be sure, designers, like natural laws, can behave predictably. Yet unlike natural laws, which are universal and uniform, designers are also innovators. Innovation, the emergence to true novelty, eschews predictability. Designers are inventors. We cannot predict what an inventor would do short of becoming that inventor.”

Jonathan Witt, 2004, from an article in Touchstone: "If Shakespeare could introduce a comical gravedigger into the tragedy of Hamlet, why cannot God introduce whimsy into His work?....Do we really want to substitute the exuberantly imaginative, even whimsical designer of our actual universe for a cosmic efficiency freak? Such a deity might serve nicely as the national god of the Nazis, matching Hitler stroke for stroke: Hitler in his disdain for humanity’s sprawling diversity; the tidy cosmic engineer in his distaste for an ecosystem choked and sullied by a grotesque menagerie of stange and supposedly substandard organs and organisms."

These claims of whimsy fundamentally preclude ID from being able to generate testable hypotheses, because if the Designer is whimsical, then there is no basis to expect (predict) any particular pattern in nature....thus no testable hypotheses. Unfortunately for ID, because of the fundamental flaw just mentioned, it is incapable of generating hypotheses of the form “We should see pattern X or trait Y or process Z in nature.” Nevertheless, ID spokesmen continue to claim that ID can make positive predictions; these attempts are simply sad: Jonathan Witt has become somewhat of a spokesman for the Discovery Institute, and had this to say on their web site in 2006: “Behe predicts that scientists will not uncover a continuously functional Darwinian pathway from a simple precursor to the bacterial flagellum.” Clearly Witt does not even realize that this attempt only shows the vacuousness of ID when it comes to making testable predictions. Saying that the rival theory cannot do something is not in any way an example of your own theory generating testable hypotheses.