An interesting anecdote which maybe doesn’t sum up this volume of writing advice, but certainly illuminates it, has Mozart being asked for hints on symphony composition by a young, aspiring composer. He advises to start with small pieces and work up, and when the composer is confused by this and points out that Mozart himself wrote symphonies at quite an early age, Mozart replies “Yes, but I never asked for hints”. That doesn’t necessarily mean that the advice in this volume only applies to writers of mediocre talent, or that it isn’t useful for experts. While the individual essays in this volume vary greatly in quality, with Asimov providing the largest number of essays, and has the least substance to his superficial und uncultered boasts, there are a few pieces that delve deeply into the problems of building entire worlds from scratch. This problem is arguably one of the most important reasons for the gap in literary quality between science fiction and the rest of it – where ordinary narrative can rely on the reader’s ability to fill in gaps in background, characterization and psychology, science fiction must often close them, using original material for it. The seams of implausibility that result whenever this goes wrong wreak havoc with characters, plot and world plausibility, requiring the aspiring writer to pay much more attention to minute details and work it all out. Incidentally and as an aside, the other big reason for the quality gap would be the fact that writers genuinely interested in general themes of human existence rather than scientific what-if-narrative will naturally shy away from setting their plots against an artificial background, leaving a large faction of people like Asimov, who openly admit that they care not one bit about subtelties of storytelling and characterization, and that ideas are all that matters. It’s too bad, really.
The high points of this collection are the essays by Poul Anderson on world building, Norman Spinrad’s somewhat dated sounding advice on how to project current trends using simple cyclical models – it’s a funny moment when he reports in awe that the spreadsheet he uses is a file of gargantuan 600 kilobytes. Good old days -, Hal Clements advice on how to create believable psychologies in an unknown world, and Connie Willis treatise on humourour writing – though technically that last one doesn’t belong here, since she herself points out that there is nothing peculiar to writing funny SF opposed to funny anything else. Also entertaining is Stanley Schmids list of cliches editors will not ever want the read, though I have to say that if you indeed want to write an Adam and Eve story because you think the idea is highly original, another line of work might be better suited.