This situation makes it difficult to be sure about designating some treatments as CAM or as mainstream. There are interventions based on reasonable scientific principles that are hard to test, or simply have not yet been tested. Does this mean that they should be called CAM interventions, rather than simply unresearched possibilities? Most people would not use the term CAM for these methods.

Let's look at another approach that might help us do a better job of classifying treatments as belonging to the CAM group. I propose that CAM treatments have some points in common, even though they may have different purposes or involve different kinds of practices. Here are some tendencies that I think are shared by a wide range of CAM interventions:

1. Advocates of CAM treatments tend to appeal to supernatural causes instead of or in addition to natural causes. For example, some CAM mental health interventions for children have proposed that children are affected directly by events that occurred before their conception, through the events' impact on the child's soul or spirit. Others have suggested that Reactive Attachment Disorder is indicates by "a darkness behind the child's eyes", presumably undetectable through physical means.
2. Advocates of CAM treatments tend to appeal to unknown natural forces such as "energy fields" instead of or in addition to known physical or chemical causes. For example, some CAM practitioners claim to be able to diagnose problems by sensations they feel when moving their hands near a person's body, and to treat those problems in similar ways.
3. Advocates of CAM treatments tend to appeal to dualism, the belief that parallel physical and mental or spiritual worlds exist, whereas conventional practices emphasize operation within the physical world and assume that human experiences are based on physical events.
4. Advocates of CAM treatments tend to appeal to misinterpretations of established scientific principles, while at the same time asserting that they are using the ordinary interpretation of the principle. For example, some CAM mental health interventions for children allude to Bowlby's attachment theory as the basis of their methods, while in fact their system is based on a different definition of attachment, a different view of the developmental timing of attachment, and a different explanation of the occurrence of attachment than Bowlby gave.
5. Advocates of CAM treatments tend to use false or weak analogies when they reason from animal studies or from physical or chemical evidence to human physical or mental health. An example is the claim that exposure to magnetic fields can treat human physical or mental health problems. It is true that like all other volumes of substances that resemble the living brain, including bowls of Jell-o, the brain is the arena of electromagnetic changes. It is also true that mental processes depend on electrical activity that transmits messages along a chain of neurons. However, the rules that govern electromagnetic events in Jell-o are far simpler than those that govern events in the brain, or those that govern mental processes, so the analogy is not a useful one.
6. Advocates of CAM treatments often use transductive reasoning-- that is, they assume that if two things share one characteristic, they also share others, including cause and effect relationships. An example of this is that suggestion that removing aluminum from the diet, by getting rid of aluminum pots and pans, will prevent Alzheimer's disease. This idea emerges from evidence of aluminum in the tangled neural matter of Alzheimer's patients, and continues transductively by assuming that if aluminum is found in Alzheimer's patients' brains, it must be the cause of the disorder, despite the fact that many other substances and unusual structures are also present. Similarly, the emergence of symptoms of autism, which occurs at about the same point in development as some early childhood immunizations, is interpreted as meaning a causal relationship.
7. Advocates of CAM treatments often base their ideas and assumptions on ways of thinking that date back to the medieval period. For example, medieval thought included appeals to the "laws of sympathy". One of these, the "law of similarity" stated that "like acts on like, opposite acts on opposite". In other words, an action involving one thing (or person) would also have an effect on a thing or person with similar characteristics; for example, one CAM school of thought about children's mental health contends that because caramels contain milk sugars, they are similar to human milk, and feeding a child caramels is thus emotionally equivalent to breastfeeding the child.
The second law of "sympathy" is the "law of contact", according to which objects that were once joined share a special sympathetic relationship even though now separated. Again, one CAM approach to children's mental health assumes that children have emotional attachment to biological mothers whom they have not seen since the day of their birth, in spite of clear evidence that attachment does not even start to develop until the infant is several days old.

There are undoubtedly many other shared characteristics of CAM thinking and practices. I put this list forward simply as a beginning effort to consider CAM methods for what they are, rather than for what they are not (evidence-based interventions).