I've received a number of emails from people who are quite upset that I regard those stickers calling evolution a theory to be sensible. Let me explain why.
Hypothesis; theory; fact. Each has a particular meaning in the sciences. Hypothesis is a potential explanation that has little or no evidence behind it. A theory has substantial evidence backing it, and individual parts of the theory may be clearly proven. A fact is something that is clearly proven, by experiment.
FACT: Sodium reacts with water to form sodium hydroxide, hydrogen gas, and gobs of heat. This is a fact. I have personally performed this experiment in the Pacific Ocean, when I was young and foolish. If anyone wants to argue the point, we can perform the experiment, repeatedly. I am willing to bet a pretty substantial sum of money that if I do this experiment 10,000 times, I will get the same results, everytime.
THEORY: Why does sodium react that way? We have an elaborate theory of chemical bonding to explain how this reaction takes place. Over the last one hundred years, the details of that theory have been continuously refined. We used to believe that electrons were in shells around the atom, with the outermost electron for sodium loosely bound to the atom. Refinements to the theory, driven by quantum theory, have turned this from a belief in a "shell" to a electron cloud. The electron's actual position, if I have followed the debate about this correctly, is really more of a probability function than a place.
As one of my chemistry professors pointed out, "We really have no idea what is going on down there at the subatomic level. There could be angels dancing on the heads of pins for all we know. But it's a useful theory for predicting things, and that's what science is all about." (And that's also why Intelligent Design really isn't a scientific theory--how do you predict what an Intelligence is going to do next?) Our current theories of chemical bonding are more than just a wild guess, but they are less certain than a fact. A hundred years hence, our theory of chemical bonding might be substantially what it is today--or it may changed (whoops, "evolved") into something very substantially different.
The theory of evolution includes a number of different components. One of those components is the idea of natural selection: that changes in the environment will put some members of the species at a disadvantage to others, and over time, this can alter the gene frequency within a population. Depending on the severity of the disadvantage, it is conceivable that a particular trait may be entirely removed from the gene pool (not just lowered frequency), although this has to be a seriously disadvantageous trait, and still takes many, many generations. Natural selection may not work with some traits, even those with serious disadvantages, if they are not exposed until after breeding. Genetic diseases like Huntington's Chorea persist in the population because the disease not make its first appearance until the person is usually over 35--and they have generally had children. Even genetic defects such as schizophrenia that would seem to be big disadvantages apparently persist because there are advantages that carriers of the gene have: sisters of schizophrenics seem to have higher rates of reproduction than the general population. Bipolar disorder and Ausperger's syndrome in their extreme forms are a serious disadvantage, but in milder forms, they seem to be useful in a number of occupations, and so persist.
I don't think that anyone (except for a few deranged, Young Earth Creationists) seriously disputes that natural selection changes gene frequency in a population. We can see natural selection taking place, altering characteristics of species, such as the Industrial Revolution changes in the color of the peppered moth. (This website points out that some of the claims originally made for this, however, are not entirely right.) I have yet to see anyone present evidence of a new species being created by natural selection that we have actually observed become a new species; no surprise, speciation should take tens of thousands to hundreds of thousands of generations, and we haven't been looking for this for very long.
We can see unnatural selection as well, where humans have bred particular breeds of animals and plants. None of these human breeding experiments, however, have led to different species. Again, there hasn't been the time. Even dogs and wolves successfully interbreed. One of the traditional definitions of species was the ability to produce fertile offspring; by that definition, dogs and wolves aren't really separate species. (There are weird examples such as "ring species" where a series of adjacent subspecies can interbreed, but non-adjacent subspecies are too different to be interbreed. This is evidence in support of the claims of natural selection leading to speciation, but again, it's only evidence, not conclusive proof.)
Another component of the theory is that inanimate matter created life through electrochemical or radiochemical process. While there have been some intriguing experiments performed that suggest that this could have taken place, the big leap--from purely inanimate chemicals to self-replicating life forms--has not taken place. Again, time is a problem. It's hard to get research funding for experiments measured in millenia, and even harder to interest scientists in experiments out of which they aren't going to get a publishable paper. An interesting theory, maybe even correct; but it is not fact.
Another component is the idea that the enormous diversity of structures represent the same sort of natural selection divergence at the higher taxonomic levels. All that I have read--even by those defending evolution--is that the fossil record is maddenly deficient in the intermediate forms that would establish this as being likely. Evolutionist Stephen Jay Gould played a major part in advancing the theory of punctuated equilibrium to deal with what he openly admits is a major problem:
The oldest truth of paleontology proclaimed that the vast majority of species appear fully formed in the fossil record and do not change substantially during the long period of their later existence (average durations for marine invertebrate species may be as high as 5 to 10 million years). In other words, geologically abrupt appearance followed by subsequent stability.The punctuated equilibrium theory seems to me to be a legitimate and sensible answer to the problem of a fossil record that is deficient. But again, it simply points out that evolution has some problems that keep it in the category of theory, not fact. Understanding and embracing this is not just about keeping fundamentalists happy; it is also about keeping the teaching of science honest and accurate.
But how could traditional paleontology live with such a striking discordance between a theoretical expectation of gradual transition and the practical knowledge of stability and geologically abrupt appearance as the recorded history of most species? Our colleagues resolved their schizophrenia by taking refuge in a traditional argument, advanced with special ardor by Darwin himself—the gross imperfection of the fossil record. If true history is continuous and gradational, but only one step in a thousand is preserved as geological evidence, then a truly gradual sequence becomes a series of abrupt transitions. Darwin staked his whole argument on this proposition:The geological record [is] extremely imperfect, and will to a large extent explain why we do not find interminable varieties, connecting together all the extinct and existing forms of life by the finest graduated steps. He who rejects these views on the nature of the geological record, will rightly reject my whole theory [Origin of Species, 1859].This resolution worked in some logical sense, but it filled Niles and me with frustration and sadness. We were young, ambitious, enthusiastic, and in love with our subject. We had trained ourselves in evolutionary theory, particularly in the application of statistical methods to the measurement of evolutionary change, and we longed to get our hands dirty with practical applications. Our colleagues had virtually defined evolution as gradual change and had then eviscerated the subject as a paleontological topic by citing the imperfection of the fossil record to explain why we never (or so very rarely) saw direct evidence for the process that supposedly made life's history. This argument did resolve a contradiction (theoretical gradualism with overt punctuation), but at a crushing price for any practicing scientist, for if evolution meant gradual change, we could not discern the very phenomenon we most wished to study.