When I was a kid in the early '60s my mother took me on weekly trips to the Wilmette Public Library. It was a well-stocked warren of interconnected sandy-brick buildings that grew in increments as Wilmette morphed from farmland to modest houses with vacant lots, to an upwardly mobile, bland, Chicago suburb, and finally to a pricey, bland, Chicago suburb. My most vivid memory of those visits was the central aisle, flanked by thousands of books reflecting glints of "modern" fluorescent lights from their crackly plastic covers. I decided to read them all. I began taking out five books each weekend with the idea that I would exchange them for another five a week later, and continue until the mission was accomplished. Fortunately for my adolescence, I soon realized a deflating fact: the library was acquiring more than five books per week.

The modern Internet has greatly increased the availability of information, both the valuable stuff and the flotsam. Using a conceptual compass a generalist can navigate the flotsam, to gain the depth of a specialist in many areas. The compass-driven generalist need no longer be dismissed as the Mississippi River, a mile wide and a foot deep.

My current fixation offers an illustration. I'm trying to develop a unified understanding of the causes of cancer. This goal may seem like a pipe-dream. Quick reference to the Internet seems to confirm this characterization. Plugging "cancer" into Google I got 173 million hits, most of them probably flotsam. Plugging cancer into PubMed I got 2.3 million scientific works. Some of these will be flotsam, but most have something of value. If I read 10 papers per day every day, I could read all 2.3 million papers in 630 years. These numbers are discouraging, but it gets worse. Pubmed tells me that in 2009 there were 280 articles on cancer published per day. Memories of the Wilmette Public Library loom large.

I navigate through this storm of information using my favorite conceptual compass: Darwin's theory of evolution by natural selection. Application of evolutionary principles often draws attention to paradoxes and flaws in arguments. These problems, if recognized, are often swept under the rug, but they become unavoidably conspicuous when the correct alternative argument is formulated. One of my research strategies is to identify medical conventional wisdom that is inconsistent with evolutionary principles. I then formulate alternative explanations that are consistent and then evaluate all of them with evidence.

In the case of cancer, expert opinion has focused on mutations that transform well-behaved cells into rogue cells. This emphasis (bias?) has been so narrow that experts have dismissed other factors as exceptions to the rule. But it raises a paradox: the chance of getting the necessary mutations without destroying the viability of the cell seems much too low to account for the widespread occurrence of cancers. Paramount among the cancer-inducing mutations are those that disrupt regulatory processes that have evolved to prevent damage from cancer and other diseases cell proliferation. One of these barriers to cancer is the arrest of cellular replication. Another is a cap on the total number of cell divisions. Still another is the tendency for cells to commit suicide when genetic damage is detected.

For a century, research has shown that infections can cause cancer. For most of this time this knowledge was roundly dismissed as applying only to nonhuman animals. Over the past thirty years, however, the connection between infection and human cancer has become ever stronger. In the 1970s most cancer experts concluded that infection could be accepted as a cause of no more than 1% of human cancer. Today infectious causes are generally accepted for about 20% of human cancer, and there's no end to this trend in sight.

When infections were first found to cause cancer, experts adjusted their perspective by the path of least resistance. They assumed that infections contribute to cancer because they increase mutation rate. An alternative view is that infectious agents evolve to sabotage the barriers to cancer. Why? Because barriers to cancer are also barriers to persistence within a host, particularly for viruses. By causing the cells they live in to divide in a precancerous state, viruses can survive and replicate below the immunological radar.

The depth of biological knowledge and the ability of the Internet to access this depth allows even a generalist to evaluate these two alternative explanations. Every cancer causing virus that has been well studied is known to sabotage these barriers. Additional mutations (some of the perhaps induced by infection) then finish the transformation to cancer.

Which viruses evolve persistence? This question is of critical practical importance because we are probably in the midst of determining full the scope of infection-induced cancer. Evolving an ability to commandeer host cells and drive them into a pre-cancerous state is quite a feat, especially for viruses, which tend to have only a dozen or so genes. To evolve mechanisms of persistence, viruses probably need a long time or very strong selective pressures over a short period of time. Evolutionary considerations suggest that transmission by sex or high-contact kissing could generate such strong selection, because the long intervals between changes in sex or kissing partners (for most people) places a premium on persistence within an individual. The literature on human cancer viruses confirms this idea--almost all are transmitted by kissing or by sex.

The extent to which this information improves quality and quantity of life will depend on whether people get access to it and alter their behavior to reduce their risk. The earlier the better, because exposure to these viruses rises dramatically soon after puberty. Luckily, kids now have broad access to information before they have access to sexual partners. It will be tougher for the rest of us who grew up before the modern Internet, in the primitive decades of the 20th century.