In 1956 Martin Gardner invented the perfect job for himself: writing a monthly column called Mathematical Games in the pages of Scientific American. Then he invented the Martin Gardner who could do the job. “I hurried at once to the used bookstore section of Manhattan, then near the Village, to buy all the books I could find on recreational math,” he writes in Undiluted Hocus-Pocus: The Autobiography of Martin Gardner, recently published by Princeton University Press. “If you look over all my columns..., you'll find that they steadily become more sophisticated mathematically. That was because I was learning math.”

Gardner, who died in 2010 at age 95, presided over Mathematical Games for 25 years, to the delight of millions of readers. Given the title of his column, it's no surprise that Gardner had a lot to say about games and puzzles. There was fun in every essay—and sometimes even juvenile humor. But he also ventured far beyond the customary territory of “recreational” math. He wrote on the theory of knots, on paradoxes of free will, on learning by induction. He introduced the public to a bevy of big ideas: the nonperiodic tilings of the plane invented by British mathematical physicist Roger Penrose, the fractals of Polish-born mathematician Benoît Mandelbrot and, perhaps most famously, the Game of Life, a minimalist simulation of birth and death devised by British mathematician John Horton Conway.

I had the pleasure of getting to know Gardner when I joined the staff of Scientific American in 1973. He worked from home, and around the office we referred to him as a “shy woodland creature,” seldom seen or heard. I might never have met him except that we both lived in towns along the Hudson River, north of New York City, and so I was recruited as a courier of manuscripts. When I visited, he would show me a magic trick, or challenge me with a puzzle (which I usually failed to solve), or deliver the latest from his far-flung network of sources.

When Martin announced in 1980 that he wanted to retire, a delegation of us entreated him to reconsider. We offered him more money, a lighter workload and secretarial help, but our main argument was simply that nothing he might do with the rest of his life could possibly have a greater impact on humanity than continuing the column. Martin did not dispute our point, but he also did not waver in his determination. He wanted to write a big book setting forth his fundamental philosophical and theological principles. That book, The Whys of a Philosophical Scrivener, appeared three years later. The New York Review of Books published a scathing, dismissive review—written by Gardner himself.

Even now, I still see Gardner's departure as a waste of human capital, but his autobiography at least makes clear that the decision was no whim. During his tenure as explicator in chief for the world of mathematics, he had other currents flowing through his life. The quarter of a century Gardner spent at Scientific American was only about a quarter of his existence, and it gets an even smaller proportion of this autobiography.

The formative event in Gardner's youth was leaving behind his childhood in Tulsa, Okla., for studies at the University of Chicago, where he majored in philosophy. He told one version of this story in a 1973 novel, The Flight of Peter Fromm, which I took for a conventional loss-of-faith narrative: a young man from a fundamentalist Christian family confronts the wider world and leaves behind his illusions. Undiluted Hocus-Pocus gives a subtler account. Yes, Gardner turned away from the religion of his parents, but he remained deeply engaged in a quest to find a meaning or purpose in life—and in an afterlife. Atheists have all the best arguments, he concedes, but nonetheless he hopes and believes.