The 1980's

Larry Niven's Medea story, "Flare Time," and his berserker story, "A Teardrop Falls," are both in his collection, Limits, published by Ballantine Books in 1985.

Poul Anderson's Medea story, "Hunter's Moon," and his berserker story, "Deathwomb," are both in his collection, Space Folk, published by Baen Books in 1989.

The 1980's was the decade of shared universes. Niven's Introduction to Limits also mentions the James Baen-inspired sharing of Niven's own Warlock's Era series.

Fred Saberhagen, creator of the berserker series, asked half a dozen other authors each to write a human-berserker encounter story so that he would then be able to amalgamate them into a novel by deciding on the order of the stories before adding a beginning and an ending. Anderson's story is potentially a turning point for the series because, in it, human beings capture an intact berserker computer - which might enable them to design von Neumann machines programmed to kill berserkers.

If that number of "...half a dozen..." other authors to write berserker stories is accurate (Limits, p. viii), then I have read one third of this novel/sub-series.

As the berserker that is to be captured flies into the trap that has been prepared for it, it:

calculates orbits;
adjusts vectors;
receives cosmic noise;
detects human vessels scrambling from the planets that it is approaching;
tracks them;
considers withdrawal;
computes that its optimum course is to proceed because, at worst, a single unit will perish;
considers dispatching a courier back to base but calculates that the human beings would destroy it and refrains;
decides to seek engagement, to establish orbit and to begin sterilizing the inhabited planet;
sliced by an energy beam, becomes "...blind, deaf, dumb, helpless." (Space Folk, p. 232)

Can all of these calculations, considerations and decisions be unconscious computations? Some of them certainly can. A book on the philosophy of mind suggested that the test of whether an entity is conscious is whether we can fully account for its properties and behavior without attributing consciousness to it. Thus, a child can imagine a conversation between a doll, a toy soldier and a teddy bear but we can fully account for every datum about the toys without attributing consciousness to any of them whereas by far the simplest if not indeed the only way to account for the child's behavior is to assume that s/he is conscious.

So is the berserker like one of the toys or like the child? Earlier in the story there is a conversation between a human being and a berserker. At one point, the human being says:

"'Were ye human - were ye e'en alive, conscious, insultable, ye metal abomination - I'd ask ye to stop playing games wi' me.'" (p. 205)

"'Afterward think...Ah, but ye do no' really think, do ye?'" (p. 206)

- so Mary accepts that she is not conversing with a conscious being even though she speaks exactly as if she were. This strains credibility. Ability to conduct a conversation is the criterion for the Turing Test.

When she tells the berserker that "'...a red dwarf star...has a life-bearing planet...," the omniscient narrator comments:

"An organic being would have registered surprise." (p. 213)

- whereas "The machine...said merely: 'That is not believed possible.'" (ibid.)

So here we are being made aware of some difference between two kinds of conversation: that between two conscious beings and that between a conscious being and an unconscious computer.

Consistently with her remarks about the berserker's lack of consciousness and of real thought, Mary thinks of its computer as a "...pseudo-brain...," not as an artificial brain. (p. 214)