[APAK logo] Issue #72, January 3rd, 1997

Hard Physics: The Many Volume Series II
by Greg Benford

[ Part One ]

After stalling yet again on my Galactic Center series, slowly I went back to fundamentals. I began envisioning what it might be like at stage center, where the diet of particles and photons is rich and varied. Only hard, tough machines could survive for long there.

In the fourth novel, Tides of Light, I drew out these contrasts. Hard work, but fun. I devised "photovores" and "metallovores" as adaptations to special evolutionary niches. After all, machines which can reproduce themselves would, inevitably, fall under the laws of natural selection, and specialize to use local resources. The entire panoply of biology would recapitulate: parasites, predators, prey.

How to envision this? I prepare for novels by writing descriptive passages of places and characters. In spare moments I began working up snapshots of possible life forms and their survival styles.

Years before I had found a technique to deal with "obstructions" -- a better word than the fearsome "block"; and to me it meant something rather more subtle. At times I simply couldn't get my subconscious to flower forth with free material along the lines of the novel.

So I pretended that I was working on another story entirely and wrote that. Sometimes I found I was right -- it really didn't connect with the novel. Most times, with some tuning, it did. I made a policy of following through, publishing the work independently if possible, out of an almost superstitious belief that my subconscious would catch on. So far it hasn't . . . I think.

That's why occasionally pieces of my novels appear first as short stories. I often don't know whether they fit the novel, sometimes until years later. This trick I had to use again and again, because my subconscious proved lazy and headstrong. I'd planned to rap out three novels and be done by 1989, but #3 appeared in 1987, #4 in 1989 . . . and then I got interested in another novel, wrote it in three tough years, and ground to a halt. The pesky subconscious just wouldn't cooperate with my game plans. This cost me considerably, for the series' momentum broke and undoubtedly some readers lost the thread.

In 1990 I had to start from scratch again, thinking through the over-arching logic of the series. Slowly it dawned that some part of me had shied away from doing the "last" novel because I couldn't reconcile the many forces within the narrative. I realized with a sinking feeling that one more book wouldn't be enough, either.

Intelligent machines would build atop the galactic center ferment a society we could scarcely fathom -- but we would try. Much of #5, Furious Gulf was about that -- the gulf around a black hole, and the gulf between intelligences born of different realms.

For years I had enjoyed long conversations with a friend, noted artificial intelligence theorist Marvin Minsky, about the possible lines of evolution of purely machine intelligence. Marvin views our concern with mortality and individualism as a feature of biological creatures, unnecessary among intelligences which never had to pass through our Darwinnowing filter.

If we can copy ourselves indefinitely, why worry about a particular copy? What kind of society would emerge from such origins? What would it think of us -- we Naturals, still hobbled by biological destiny?

Through books #3, 4 and 5 I had used the viewpoint of humans hammered down by superior machines. This got around the Walmsley lifetime problem, but demanded that I portray people enormously different from us. They had to seem strange, yet understandable -- a classic sf quandary.

A slowly emerging theme in the novels, then, was how intelligence depended on the "substrate," whether in evolved humans or adaptive machines -- both embodying intelligence, but with wildly different styles.

By the time I reached the last volume, in 1992, I had spent over twenty years slowly building up my ideas about machine intelligence, guided by friends like Marvin. I had also published several papers on the galactic center and eagerly read each issue of Astrophysical Journal for further clues.

I finished the last novel, Sailing Bright Eternity, in summer 1994. It had been 25 years since I started on In the Ocean of Night and our view of the galactic center had changed enormously. Some parts of the first two books, especially, are not representative of current thinking. Error goes with the territory.

I had taken many imaginative leaps in putting together a working "ecology" for the center. I included outré ideas, such as constructions made by forcing space-time itself into compressed forms, which in turn act like mass itself: reversing Einstein's intuition, that matter curved space-time.

All this was great fun, requiring a lot of time to think. I let my subconscious do most of the work, if possible -- an easier way to write, but it stretches out projects, too.

Long-suffering readers wrote asking when the next volume would appear and I felt badly about it, but I knew the writing could not be rushed. I had not anticipated that each volume would demand so much thought, and still less that I would need an extra novel to do the job. By the end, all six books comprise about three quarters of a million words.

My published physical model of the galactic center is done in what I call the "cartoon approximation" -- good enough for a first cut, maybe, but doomed to fail somewhere. Sf works in this approximation, necessarily. I had assayed a grand theme, how Mind relates to Nature.

In any case, models are like art, matters of taste. Nobody expects a French impressionist painting to look much like a real cow; instead, it suggests ways of looking at cows. Sf should do that.

I learned a lot of tricks along the way, many of them embarrassingly obvious. In 1969 I never outlined, though that year I had sold my first novel with a three-page description and 10,000 words of a novelette. By 1992 I kept notes by subheadings -- INCIDENTS, NOTIONS, TECH, TIMELINE, CHARACTER, BITS O' BUSINESS, etc. -- in a three-hole binder and on computer, so I could lift and insert.

More important, I had grasped that the climaxes of each book should resemble a stairway. Each should play for higher stakes which do not undercut the resolutions of the earlier novels. Each should open the philosophical canvas at least a bit, particularly in a galactic, hard sf novel sequence such as mine. Each should explain mysterious elements of the past novels, but leave some shadows to shed a glow into for the future. Each should tell us something deeper about the lead figure. Each figure should move through the defining moment of his life.

This last point may be crucial. I used two central figures, Walmsley and Killeen, neither particularly likeable. This may be a quirk of mine, but I've never enjoyed trotting around in the head of a bright-eyed, perpetual optimist; this may reveal more about me than I wish, but there it is.

Each of these men had to learn and grow, but not abandon themselves to the cosmic perspectives. As Gary Wolfe remarked in reviewing the last novel, "This is the classic problem of hard sf, of course: a rhetoric of action and human drama must be juggled with a rhetoric of science and philosophy in a way that must be made to appear seamless . . . [often] writers either give us cardboard characters against a spectacular backdrop, or fudge the science in order to make the plot work out . . . "

I felt the pressure of keeping these guys human more and more as the novels waxed on. So I gave them vices, irksome habits, troubles with their women, faults -- big ones, including bad tempers and emotional isolation. (Even Einstein picked his nose, remember.) Yet each figure made progress, or at least came to understand himself better.

I didn't actually figure all this out clearly -- in fact, some of the above paragraphs have made these points clear to me only while I was writing them. (This is a common experience for me, too. I don't know what I think until I express it. That old subconscious, again.)

I had always intended to make the series Stapledonian, but squeezed through the aperture of a modern, rounded novel. I used talks with aliens, with machines, with disembodied intelligences lodged in magnetic configurations, with archly amused denizens of the far future -- anything, to avoid the overweening narrative voice; though I used that, too.

This single decision -- more aesthetic than craftsmanly, and made unconsciously as well -- created more work for me than anything else in the sequence. It is my preferred method overall, even outside the Galactic Series, but it imposes great constraints.

That fits with my own feeling about hard sf -- that it works best because of its self-imposed restrictions, in the fashion that a sonnet does. Constraints improve.

Would I write a series again? Maybe, but not right away.

Do it this way again? Nope -- I hope I'd avoid some of the traps.

Most important, I fathomed my own limitations, and how little my subconscious could be bossed around. It's useful to know who really does most of the heavy lifting.

Problem elk are marked with a paintball.

[APAK logo] Issue #72, January 3rd, 1997

Return to the table of contents.

Previous article: Fanzine Counts, by Victor M. Gonzalez.

Next article: The Trickle-Down Theory of Dr. Fandom, by Ted White.