Yesterday, I looked at the general idea of separating science from pseudoscience, and my conclusion is that any sufficiently educated person can learn to distinguish the two. The main hallmarks of good science are (1) the characteristic of prediction and falsifiability, and (2) reliance on mechanisms that are consistent with already-verified science.
How, though, do you make the distinction between pseudoscience and valid, "emerging" science? New ideas in science are frequently ridiculed, especially before the mechanisms governing them are completely elucidated; even such rock-solid (pun intended) models as plate tectonics were considered to be foolishness before the magnetometer data discovered in the late 1950s showed that the ocean floor was spreading. Some ideas that fly in the face of what is currently known will turn out, on analysis and through experiment, to be verified science. How do we know that by labeling something as pseudoscience, we're not tarring good ideas and bad with the same brush?
Here are a few things that seem to be general characteristics of pseudoscience. Note that not all pseudoscientific theories have all of these traits -- but this will at least provide a few general rules-of-thumb for recognizing it when you see it.
1) The reliance on undefined, or poorly-defined, terms. I've harped on this one so many times that it doesn't bear much more description than that. Watch out, especially, for terms that are firmly defined in one realm (e.g. physics), but are being used in a fluffy, non-specific way. Favorites are frequency, vibration, quantum, energy, field, wavelength, and resonance.
2) Any idea that claims to contradict a thoroughly researched and experimentally verified model. If someone starts by saying, "My theories overturn the First and Second Laws of Thermodynamics," I pretty much stop listening. Take a look at the webpage of The Thunderbolts Project, which claims that pretty much everything we know about physics -- gravity, electromagnetism, particle physics, convection, heat flow, quantum mechanics, nuclear fusion -- is wrong. Instead, the universe is linked by "a web of electrical circuitry connects and unifies all of nature, organizing galaxies, energizing stars, giving birth to planets and, on our own world, controlling weather and animating biological organisms." While it is always possible that new data could force a revision of a pre-existing model -- look at what Einstein did to Newtonian mechanics -- the chance of a few "progressive scientists, researchers, and laypeople" trashing all of physics in one blow is unlikely in the extreme. If you take a look at the video clips put out by these people (such as this one) you find that mostly they are relying on the fact that physicists' models are themselves incomplete. While no reputable physicist would argue that point -- after all, if their models were complete, they'd all be out of a job, because there'd be nothing left to research -- our ignorance about how some features of the universe work doesn't mean that the framework of science itself is faulty and needs to be replaced.
3) A theory that blurs the distinction between a model and the reality. This one can be pretty insidious, because scientists use models (especially mathematical ones) all the time, and frequently explain their ideas using analogies. On its crudest level, we have people like the Rosicrucians, who think that because much of the universe is describable using mathematics, that the universe is numbers (and those numbers have mystical significance). A subtler example is the work of Stephen Wolfram, who has modeled systems behavior using a construct called a cellular automaton, but who seems to me to cross the line from stating that cellular automata can model many observable processes (this model has been used in everything from developmental biology to particle physics), to stating that the universe is composed of cellular automata. While the first statement is no doubt true, the second remains very much to be demonstrated.
4) Experimental claims that no one else seems to be able to replicate. Replicability is one of the most important characteristics of good science, which is why any peer-reviewed paper is expected to give thorough detail about the experimental protocol used. Examples of this abound: three well-known ones are the famous "polywater" debacle, "cold fusion," and Cleve Backster's claim that he'd done an experiment proving that his house plant was aware of his emotional state.
Of course, there are many other good questions to ask that can help separate good science from bad. Is there a profit motive involved in the claim? Does the researcher have an ideological bent that is biasing him/her to ignore contrary evidence? Are the results published in a journal that has adequate peer review? Is anyone who questions the results viewed as hostile/biased/closed-minded?
Given the number of crazy ideas out there, it's absolutely critical that we learn how to recognize what constitutes a scientifically sound idea, and what characteristics should raise red flags that we're looking at pseudoscience. Millions of dollars of hard-earned money are wasted every year on homeopathic "remedies," psychic and astrological readings, crystal therapy, chakra and "energy field" realignment, and so on. The methods of science aren't error-proof, and scientists are fallible humans, just like the rest of us; but if you're looking for the best way to gain a solid understanding of how the universe actually works, science is the only game in town.