The Andromeda Strain - Neuron and junction: Patterns of Thought

There is something faintly suspicious about thinking seriously about Michael Crichton’s Andromeda Strain; after all, it was a tremendously successful, best-selling thriller, it became a very successful film, and popular genre novels rarely boast artistic pretensions. The plot of The Andromeda Strain

disin­tegrates under the rough hands of literary criticism; its characters are for the most part stereotypes; and it is interminably didactic. In these and other elements the novel certainly is worthy of the scant scholarly attention paid to it.

In fact, it is rather easy to spend a lot of time examining the weaknesses, inadequacies, and flaws of the novel. Why, then, is the novel interesting? Simply because, in a way that few other science fiction novels attempt, The Andromeda Strain  consciously focuses its narrative attention on the dif­ferences between two patterns of problem solving, one of which ignores human uniqueness, and one of which is exclusively reserved for human mental processes.

The Andromeda Strain is a story about the conflict between these two patterns of problem solving, symbolized by the scientific method, described as painstakingly rigorous and unimaginative, and the intuitive, creative method of human insight.

In the course of evaluating the narrative conflict, we discover that the novel concerns itself with some of the key issues underlying the portrayal of artificial intelligence in science fiction that novels such as The Moon Is a Harsh Mistress, When Harlie Was One, The Cyberiad do not, although these novels have as their protagonists intelligent computers or robots.¹

These novels, and the many others published as science fiction, present us with artificial intelligences which are portrayed as analogous to human intel­ligence, a fact that Stanislaw Lem puts forward judgmentally in “Robots in Science Fiction.”² Essentially, these books bypass the whole question of the nature of an artificial intelligence. Heinlein’s Mike not only thinks human and holds human values, he gives himself human voices and forms.

David Gerrold’s Harlie is self-conscious, neurotic, and intellectual, while Lem’s intelligent robots, the heroes of a series of fables, are,  human. None of these books confront the nature of artificial intelligences, primarily because the generic problems of describing any kind of alien consciousness may preclude human description in prose.

(They may be rendered mathematically, how­ever.)

While The Andromeda Strain is not about robots or intelligent computers, it is concerned with the conceptual differences between the representative thought processes. Insofar as it concerns itself with two different kinds of thought processes, it bears a formal similarity to Samuel R. Delany’s Einstein Intersection, in which reality defined by the physics used by Einstein, who did not believe in indeterminacy (required by quantum mechanics), is contrasted to reality defined by the principle of Godel’s Incompleteness Theorem. ³ Insofar as the central narrative focus of  The Adromeda Strain s not on an individual but a process, it bears a structural similarity to many other thrillers, police procedural mysteries, and science fiction novels such as Arthur C. Clarke’s Childhood’s End, in which the main thrust is the process of evolution which takes the human race from its present status to its ultimate development as cells of a universal Overmind.

The Andromeda Strain bears structural relationships to these last two novels, not thematic relationships. Thus, it is formally allied to books with dissimilar thematic concerns and is thematically allied to novels which do not focus on the key issues of artificial intelligence.

The Andromeda Strain is about a process and not about people. The very first line of text announces the subject of the novel: “This book recounts the five-day history of a major American scientific crisis.” (A page before the beginning of the text attempts to give the novel the appearance of an intel­ligence file through the use of stereotypical classified document announce­ments and graphics.

Similar structural techniques occur throughout as attempts to create a sense of reality for the reader.) Besides announcing the focus of the narrative, this statement announces the span of its action—five days. By limiting himself to five days for the working out of an exciting crisis, Crichton commits himself to a fast pace, which is all to the good since pace is the sina qua non of this form of thriller.

That The Andromeda Strain is about a process working through time alerts us to its interest in structured events. When we see that a significant portion of the novel is given over to detailed descriptions of the workings of the Wildfire laboratory, to the theory of crises, and especially to the mechanics of the Life Analysis Protocol, then what else can we assume but that this novel is con­cerned with processes more than with people—especially when the char­acters exist at the convenience of the plot?

The rapid pace of the thriller is perfectly suited to this concern with process because it moves too fast for the reader to focus on human interactions, only on the next event. Furthermore, the point of view remains with the narrator, another stylistic technique which speeds up the pace. At all times there is an audible auctorial presence whose voice is consistent throughout.

This voice can be distinctly recognized whether the object of its attention is the descrip­tion of a town killed by the alien organism, the workings of an electron microscope, or Mark Hall’s desperate last-minute struggle to avoid nuclear destruction.

All these elements of The Andromeda Strain ⁴ a structured plan for studying any organism, an extensive theoretical rationale for all steps of the establishment of the Pro­tocol, the military’s “Wildfire,” and yes, even “Scoop” plans. It is important to understand that the Life Analysis Protocol is not one simple element;

it is the umbrella which covers all, from scientists’ behaviors to the location of detonator substations in the Wildfire lab. It is the weapon that humans develop to combat just the kind of threat that Andromeda is.

“According to the Life Analysis Protocol, there were three main steps in the Wildfire Program: detection, characterization, and control ⁵ In effect, the Life Analysis Protocol works by modeling the world, then preparing for all foreseeable contingencies, much like the dominant technique of the “Mission: Impossible” television series.

As in that show, the essential tensions are gen­erated by our anticipating the breakdown of the behavioral predictions of the protagonists.

The same thing is attempted by the Life Analysis Protocol. Human error is planned for and prepared against—as in the backup jets waiting to attack the exploratory helicopter should the researchers at Piedmont die of the infesta­tion and the pilot not destroy himself.

Unfortunately in The Andromeda Strain ⁶

This particular point is of key concern in automata theory, in recursivity and self-reference theories of mathematics, and it is a jumping-off point for dis­cussions of Godel’s Incompleteness Theorem and the study of the nature of artificial intelligence, making this novel of particular concern to students of science fiction.

Within the purview of The Andromeda Strain

The second weak link is human nature, and the same weakness which undercuts the teleprinter link almost overwhelms this, too. The characters are barely able to support such scrutiny, especially in light of the recent expe­riences of Three Mile Island, in which human factors scientists and engineers found such poor interface between the reactor’s controls and the technicians manning them that the crisis worsened as a result, and human error due to poor control-systems design was a significant factor in the genesis of the crisis.

By examining the nature of the errors made in the execution of Wildfire, we discover the confrontation and conflict between idiosyncratic human nature and formalized systems procedures, which are the underlying focus of The Adromeda Strain.

Ignoring the novel’s internal inconsistencies, there are two basic categories of error, human errors and systems insufficiencies (that is, reality offering more alternatives than the system could be programmed for). Human errors are rational errors. That is, they are

failures of humans to live up to the rigorous demands of the Life Analysis Protocol’s logic. Insofar as they are rational errors, they indicate the separation between ideal and real behavior, between reality and model. The key is that they occur because, in spite of themselves, the scientists are affected by their emotions. For example, stress is responsible for Burton’s forgetting about the anticoagulated rats, thus set­ting back the team’s understanding of how Andromeda kills.

That information is not important to resolving the narrative crisis. It is resolved, as is H. G. Wells’s Martian crisis, by an overlooked factor—in this case the rapidity of bacterial mutation. (Here is an example of a glaring logical flaw. Andromeda is not a bacterium. It is something alien, a point that the book goes to great lengths to document.) Burton’s error is not structurally important. Yet why is so much made of it? A plausible explanation would be that it is  important to the dramatic confrontation between the protocol and human individuality. The probability of this interpretation rests on several factors.

Consider Burton: he just isn’t tidy enough for Wildfire. His clothes are distastefully rumpled and dirty. Even his scientific record is untidy; he is known as an inspired bumbler, a rider of luck.⁷ And luck is a wild card the model can’t accommodate. He simply doesn’t fit into the protocol. His mistake is not unreasonable.

After all, he is under great stress; recently back from Piedmont, he must have been shaken by what he saw. On top of all this, he does not have any idea of what he should be concentrating on. So when he gets a strong response to the first tests he runs, he has no reason not to put aside an apparently less important result.

That is just the kind of mistake anybody would make. It is just the kind of mistake the system part of Wildfire make (and which Stone and Leavitt, the scientists most committed to Wildfire, do not make in their examination of the Scoop probe). Which is why, even though that error isn’t really important to the narrative, so much emphasis is placed on it: this is the first concrete example of humans erring against the rigid, linear Life Analysis Protocol.

This is where all the didactic material fits in. All the talk about the nature of systems—systems for identifying organisms, uncertainty decision analysis, the rationale for Wildfire, and the Protocol itself—define the right way of thinking, the right way to do science.

A corollary of the rigidity of the Life Analysis Protocol, and of all systems, is its fragility, and this fragility is at the root of the second type of error, in which the system fails because it cannot prepare for the varied inputs of reality. Brainless errors like the paper jam on the teletype warning bell cripple it.

The system’s blind spots become increasingly evident, but blind spots are inevitable because alien life may be too alien for us to predict. The crystalline nature of Andromeda can be recognized by human science, but the probability of that being the nature of the first alien life form encountered by man is small. And what possibility is there of anybody’s being able to predict the kind of chemical reaction that destroyed the lab’s seals? No formal research protocol can provide for every contingency. There are limitations to the system’s ability to interact with reality. Every system constructed to organize thought or be­havior is a model, as role models exist for people, for instance. Every religion defines reality, often in spite of gross evidence of its inadequacy: consider the life of Galileo.

In spite of elaborate security precautions and sterilizing procedures, Wildfire is contaminated by humans who subvert its goals. Christian Kirk, a valuable team member, is in the hospital. This leaves Mark Hall as his replacement, and Hall is not considered to be all that valuable. Peter Leavitt, Stone’s peer and deputy, shares his dedication to Wildfire, yet he sabotages it by hiding his epilepsy. No protocol is designed to assure the competence of the scientists performing the experiments. That is a fundamental assumption, and such assumptions can be wrong (as they are in this novel). Humans can identify erroneous assumptions and change them, while programs can not; if only for this reason, the systematic, limited protocols are inferior to the human minds which unsuccessfully follow the Life Analysis Protocol.

Human minds model reality orders of magnitude more subtly than pro­tocols, programs, or systems can. They have the flexibility to modify the models, and we do this as much emotionally and physically as intellectually: “. . . all cybernetic appoximations to this using a to and fro process fail entirely, because that is only a series of challenges. No, you just have to face the fact that the totality of mind and body forms a unit in which the mind is not a finite state system. ⁸ The same human uniqueness which impels Leavitt to hide his epilepsy fuels the intuitive leap which leads him to understand the nature of Andromeda. They are two sides of the same coin.

   Burton the bumbler is a genius; so is Leavitt and so is Stone. Each one is unique, and defies modeling: there is “also a category which cannot be analyzed by contingencies . . . including rare moments of discovery and insight. ... Because these moments are unpredictable, they cannot be planned for in any logical manner. The mathematics are wholly unsatis­factory. ⁹ The auctorial comment is relevant because (if I may steal yet another thought from Bronowski) no system can be imaginative, and what humans are is imaginative—-and flexible. That is why something as self- evident as the inability of a protocol to survive errors is important.

The Andromeda Strain consciously pits human flexibility against the Pro­tocol’s rigidity. Mark Hall, the inadequate second choice, is the one who discovers the two key facts: that Andromeda can only survive within a narrow range of pH; and that Andromeda had mutated.

Hall is the team member least committed to or involved in Wildfire. His job is to treat the survivors—he’s the clinician, and his perspective is the human one. Even more dramatically, he is the Odd Man. More satisfying than this, he makes his fundamental discovery associatively; his “highway diagnosis. ¹⁰ Hall’s insight comes in the form of an analogy generated by reverie. Analogy, reverie: these are not logical processes, they are intuitive and creative.

The solution to the key problems of the novel, then, comes from mental processes denied to the Life Analysis Protocol, creative processes accessible only to humans. Leavitt, too, shares this experience. He grasps the larger nature of Andromeda in a dream. Fortunately for the plot, he forgets it. (The dramatic climax of the novel is Andromeda’s rapidly mutating like a bacterial colony, and examining Leavitt’s insight too carefully could only interfere with the narrative development.) While the Life Analysis Protocol does have its moments—it is able to deduce Andromeda’s crystalline nature, for instance— it is the uniquely human thought processes that resolve the focal questions.

The whole argument of the novel comprises a single hypothesis: that systematized thought as symbolized by the rigid confines of the Life Analysis Protocol can not compete with human creativity. A corollary to this is that only human intelligence is powerful and flexible enough to confront reality successfully, and this is of particular significance in depicting artificial intelligence, a fact that is the thematic skeleton of the book.

For all its flaws, The Andromeda Strain remains a book that addresses the issue of the nature of artificial intelligence. That it does not address the nature of human intelligence with as much insight ought not to be held against it. The thriller is not the proper format for such a focus. The strength of its view of systematized, limited problem solving is the unexpected virtue of the novel, and that virtue gives it a special place in those works of science fiction that concern themselves with intelligent computers and robots.

NOTES

1. Robert A. Heinlein, The Moon Is a Harsh Mistress (New York: Berkley, 1968); David Gerrold, When Harlie Was One (New York: Ballantine, 1972); Stanislaw Lem, The Cyberiad: Fables for the Cybernetic Age, Michael Kandel, trans. (New York: Seabury, 1974).
2. Stanislaw Lem, “Robots in Science Fiction,” in SF: The Other Side of Realism, Thomas D. Clareson, ed. (Bowling Green, OH: Bowling Green University Popular Press, 1971), pp. 307-25.
3. See my “Surreal Translations of Samuel R. Deiany,” Science-Fiction Studies, 4 (March 1977), 25-34; see also my chapter on Samuel R. Deiany in the Dictionary of Literary Biography, for further discussion of this particular point.
4. Michael Crichton, The Adromeda Strain New York: Dell, 1969), ch. 5, p. 48. 5. Ibid., ch. 15, p. 156.
6. Jacob Brownowski, The Origins of Knowledge and Imagination (New Haven: Yale University Press, 1978), p. 80.
7. Crichton, Andromeda Strain, ch. 5, p. 54.
8. Brownowski, Origins, p. 101.
9. Crichton, Andromeda Strain, ch. 20, p. 202.
10. Ibid., ch. 27, p. 271.

In: The Mechanical God: Machines in Science Fiction. Machines in Science Fiction. Edited by Thomas P. Dunn and Richard D. Erlich (Editor). London (1982), pp. 109-116.