To sort all of this out, it is useful to distinguish between two different futures: the future that someone will have if they are not arrested, and the future they will have if they are arrested. Call the first, their NA-future; call the second, their A-future. And let’s call a prediction of someone’s NA-future, an NA-prediction; and call a prediction of someone’s A-future, an A-prediction. This terminology allows us to put the last point this way: what happens in a person’s A-future does not falsify an NA-prediction of that person’s future; this is because an NA-PREDICTION only tries to report on someone’s NA-future.
03.2. Track-Record Evidence Establishes Guilt
But what kind of evidence can be used to support an NA-prediction? One key kind comes from certain patterns of events that have been verified. Most important are the patterns from six years ago when the precogs first predicted murders that then occurred. The precogs’ confirmed early track record, we are told, was impeccable. Each predicted murder was verified as actually having occurred.
To say that these early predictions were impeccable is to say that they were free of two kinds of flaws: errors of commission and errors of omission. An error of the first kind would consist in a prediction that a murder was going to occur, which did not occur. And an error of the second kind would consist in a failure to predict a murder that then occurred.
But all of this was six years ago. What if the precogs have somehow lost their golden touch and their predictions are no longer impeccable? What if some errors of commission, or some errors of omission, have slipped in? How can we know that they haven’t?
Take errors of omission first. An error of this kind would consist in the precogs failing to predict a murder that then occurs. If such errors of omission crept in, there would be an increase in the present-murder rate. But the present-murder rate remains at zero. It follows that errors of omission have not crept in.
What about errors of commission? If they have crept in, the Division of Precrime would have arrested pre-innocents—that is, people who were not going to commit murder. But arresting a pre-innocent will have no effect on the murder rate. So we need to look elsewhere for evidence that errors of commission have not taken place.
Such evidence comes from cases like the Howard Marks one in the opening sequence of the instructional training video. The precogs’ NA-prediction is that Marks will kill his wife and her lover. But no error of commission is committed here since Anderton catches Marks in the midst of attempting to kill his wife. Surely when a person is arrested in the midst of attempting to kill someone, this is evidence that the person really was going to commit murder and so it is evidence that the precogs did not make an error of commission.
Still, we might wonder about other people who are arrested earlier on, say before even attempting murder, or even before forming an intention to murder. How do we know that errors of commission haven’t been made in those cases? The answer to this is more complicated. It has to do with the absence of any systematic relationship between when a person is arrested and something that might make the precogs make errors of commission.
As we see in the Marks case, how quickly a precriminal is arrested is affected by numerous factors, including how quickly clues are scrubbed from the images, and how long it takes to get from headquarters to the precriminal. In the Marks case had things gone faster, Marks would have been arrested before he attempted to kill his wife; and if they had went faster still, Marks might have been arrested before even intending to kill his wife. As it was, arresting him was delayed. But since these delays won’t correlate in any systematic way with whether the accused is going to commit murder, this case, and others like it, are a random and representative sampling of the entire class of murder cases that are predicted by the precogs. So this allows us to conclude that barely-prevented murders should count as evidence that the system is not producing false positives. Further supporting this idea is the lack of contrary evidence within this kind of barely-prevented case: Precrime officers have not yet arrived on the scene of a just-about-to-becommitted murder only to find calm, happy people enjoying a cup of (unpoisoned) tea.
03.3. What about Human Error?
Our critics also look for other spots where error might enter, for example with the rest of us that work in Precrime. This is what Witwer was looking for in the instructional training video: regular old human errors made by people like us, not errors made by the precogs. And he found that such errors were made by the techs who studied the previsions in the Anne Lively case when they failed to notice that in some previsions the ripples in the water were moving toward the shore, while in others the ripples were moving away from the shore.
There is an important issue here, but it is a general one that doesn’t pose any special problem for using previsions to identify pre-criminals. The general issue is how to interpret evidence. Previsions can be erroneously interpreted just as any other kind of evidence can, be it more familiar forensic evidence, confessions, or eyewitness reports. To be probative, a piece of evidence needs to be correctly interpreted. And while it is true that whether it is judges, juries, scientists, or precrime specialists in our division, we humans are fallible at this, this does not mean we should quit doing it. Fallibility, after all, doesn’t imply unreliability; and it is well-documented that police had difficulty in interpreting and applying evidence even in the days when murderers were only arrested after the crime was actually committed.
04. What about Leaks?
Finally, what about information leaks? Since someone might alter their course of action if they are told what they were going to do, critics say we should leak this information to alleged precriminals. Maybe this is what happened with Anderton. This might help the informed person to depart from their previous path and not commit the murder. In effect, it would give them a chance to change their ways, to resolve not to commit murder, and thus to make themselves no longer guilty of future-murder.
Now if the person did change course in this way, our critics allege that everyone would gain: the person who would have been a future-murderer is no longer a future-murderer and therefore need not be quarantined and can continue to contribute to society; since the future-criminal is not quarantined, family and friends won’t miss them; we wouldn’t have to go to the trouble and expense of arresting and storing them; and the person who was future-murdered is never murdered.
To better understand this issue, it’s useful to use the earlier distinction between a person’s NA-future and their A-future. A person’s NA-future, recall, is the future they will have if they are not arrested; and a person’s A-future is the future they will have if they are arrested. Cleaned up a bit, a person’s NA-FUTURE is really the future that person will have if two things are true: they are not arrested and they are not told this prediction. But as is often the case, learning something can change what a person will do. When it does, this makes for a third future. A person’s I-future, let’s say, is the future they will have if they are informed of their NA-future.
In both the written case study and the instructional training video, Anderton learns of his NA-future. But since he learns this, it may no longer be his future. His actual future will be his I-future, and it may (or may not) differ from his NA-future. Often we resolve to do something enough that we will do it even if we are told that we will do it. For example, if someone tells me that if they had not told me this, I was going to have two cups of black coffee tomorrow morning, I would not be surprised and I would not alter my behavior: I would still have two cups of black coffee tomorrow morning. But in other cases this is not so: being informed that if I had not been told this, I was going to lose next week’s lottery, I would alter my behavior and not buy a ticket.
Maybe we should give people identified as precriminals a second chance. We could inform them of the precogs’ prediction and then be at the ready just in case they resolutely attempt to murder the person anyway. If they do so, then we will be there to stop them and no one is any worse off. And if t
hey don’t make an attempt, everybody wins in the ways that we saw.
There are three reasons we should not leak this information to precriminals. One is simple: the precogs only tell us what people will do if they remain ignorant of these predictions. In other words, the precogs only make accurate NA-predictions, not accurate I-predictions. This means that upon learning that he is predicted to commit a murder, someone might commit the same murderous act, just at a later time. Once we see that they did not do it at the predicted time, we will take our eyes off them—just in time for them to succeed in committing the murder. This is just one way they can beat the system. Another is this: there might be so little time between the predicted murderer getting in place and the time that they might commit the murder that there would not be sufficient time to stop them.
The second reason to not tell them has to do with cost. In some cases, it is difficult, and expensive, to be at the ready: the possible murder might be far away, or somewhere that is difficult to observe.
And last, do the possible murderers have a right to know the precogs’ NA-predictions? A look at familiar cases of a right to know suggests that they don’t. There are two familiar kinds of cases in which people have a right to know. One is illustrated by a citizen’s right to know what is in her air, food, and drinking water. She has a right to know these things in part because she needs this knowledge to avoid possible harms to her health and because she would not be responsible for such harms.
But the situation is quite different for a person who is on a path to committing murder: he is not innocent and he does not need to know he is on this path to avoid the harms of quarantine and punishment. He had another way to avoid these harms: namely, by not having made choices that put him on this path.
Alternatively, a right to know sometimes stems from ownership, rather than avoiding harm. For example, taxpayers have a right to know what the state is spending money on even if there is no chance that the state is harming them with their expenditures. Here too the situation is quite different for the person who is on a path to committing murder: there is nothing that person owns that is involved. In short, there’s no reason to think he has a right to be told what path he is on.
The Division of Precrime does what law enforcement has traditionally done: protect innocent people. And despite what our critics say, we arrest people who deserve to be arrested; and we do so with appropriate attitudes and compelling evidence, and without violating their rights.12
20
If the Universe Isn’t Real, How Should We Treat Other People?
ANDREW M. BUTLER
There’s a moment at the end of Total Recall, when blue-collar spy Quaid looks at his lover Melina, and wonders whether the world around them is a dream. After all, he has already dreamt of walking on Mars, and has just rather surprisingly survived asphyxiation with no apparent ill effects. The movie fades to white as they kiss, and we’re left only to guess as to whether the characters are not still inside some preprogramed hallucination.
Similarly, at the conclusion of Minority Report, where the trio of psychics lives happily ever after and the estranged couple of John Anderton and Lara Clarke are reconciled, with Lara pregnant, there is the sense that everything is just too neat, and all of these events are a wish-fulfillment fantasy. And the ending of the 1982 release of Blade Runner has Deckard and Rachel fly off into an improbable sunset, not knowing if she will live forever or die at the age of four, whether he might be a replicant himself, or whether he is still asleep at the piano in his apartment.
The use of dreams and visions within narratives is one way for storytellers to play with notions of reality—what are the differences between dreaming and living an experience? Can something always be recognized as a dream? How do we know that what we perceive—seeing, hearing, smelling, touching, testing—is in fact reality? This question has been repeatedly posed over the millennia by philosophers both before Plato and since, and is a theme that Philip K. Dick repeatedly explored in his fiction.
In Eye in the Sky, a group of people pass through a number of fake universes after a nuclear accident, each universe being the one which each one of them had imagined themselves living in. In Flow My Tears, the Policeman Said, Jason Taverner wakes up in someone else’s drug-induced hallucination. In other works, there is a kind of conceptual breakthrough—as in Time Out of Joint and The Man in the High Castle, where characters discover the universe around them is not real (or, at least, is not the only reality).
The problem is, once reality begins to be questioned, skepticism can take over. How is it possible to ever know that an experience is real? In Next, the future-seeing Cris Johnson keeps experiencing one version of reality before abandoning it in favor of doing something else and then living those events having performed different actions. Are those alternatives still real? As he heads off from the motel in search of a nuclear bomb for (at least) the second time, is what he experiences real, or an alternative he has yet to discount?
The uncertainty ought to free us. If nothing is real, then nothing we do is real. Decisions have no consequences, as their impact is not real. If it’s all a dream, then nothing really matters. And yet, at the end of Now Wait For Last Year, Eric Sweetscent decides to go back to his wife, Kathy, in order to help her, even though he’s pretty certain their marriage will fail, and he’s questioning the reality he is part of.
Dick’s major characters—his heroes for want of a better word—try to do the right thing, even if what they do might not be the real thing. This is part of Dick’s other major theme: what defines the authentic human being? The characters who show concern for others, who care, who demonstrate empathy, who are humane, are to be considered human; those who do not are little better than machines, and use those around them as tools.
In a letter written late in his life, Dick noted that he had been trying to “fuse early Hebrew monotheism with the philosophy of Heidegger—which no one has ever done before.” Martin Heidegger, a German philosopher, thought that we are thrown into reality; that we struggle to find ourselves while we are already in the midst of life, located in space and time. We don’t have control over where we have been born (thrown), and there is the unsettling thought that at any time we might get thrown out again—we might die. The world around us distracts us from this terrifying possibility of inevitable death, and we perceive that world as a set of tools we can use to our own advantage.
There are moral problems in a philosophy so centered on the self’s needs—it insists that I am more important than anyone else. Dick’s writings question that self-centeredness. For what it’s worth, he wasn’t actually the first to try to fuse Heidegger’s ideas with Jewish monotheism: the philosopher and Talmudic scholar Emmanuel Levinas made an earlier attempt. The parallels between Levinas’s and Dick’s thought show how rich and challenging Dick’s work is.
To Be or Not To Be—Is That the Question?
Emmanuel Levinas was born in Lithuania in 1906, to a Jewish family, and studied under Martin Heidegger in Freiberg. In the late 1920s, Levinas moved to France, becoming a French citizen. During the Nazi occupation, he was put into an internment camp, and then, after the Second World War, continued a dual career as Talmudic scholar, writing about Judaism, and as a philosopher, writing about ethics. He saw ethics as the most important issue in philosophy, even more so than questions of what is real. He was fond of quoting the soliloquy from Hamlet—”To be or not to be?”—but then continuing, “Is that the question?”
In Dick’s fiction, we often can’t really tell if an experience is real, but the characters still ought to behave in a particular caring way. Levinas celebrates the widow, the orphan, and the alien—although by alien he means ‘foreigners’, rather than Rexorians, Bleekmen, wubs, Lord Running Clam, or the Glimmung.
Levinas had in part objected to Heidegger’s distinction between things that exist (beings) and the nature of existence (Being), and described how beings are thrown into Being. Levinas wanted to make a distinction
between beings (he called them existents) and the act of Being (existing), and rather strangely argued that there could be existing without existents. Actually, this happens all the time in Dick’s fiction: the hallucinated town in The Cosmic Puppets, the authentic aura of the American relics in The Man in the High Castle, the process of entropy in The Three Stigmata of Palmer Eldritch, or the aphids in A Scanner Darkly which trouble Charles Freck even though they do not exist.
Levinas’s distinction between existing and existents also builds upon phenomenology, a school of thought which focuses upon how the individual’s consciousness of the external universe is created through sense data; sight, hearing, smell, touch, and taste. We sense an outside world, but how can we be certain our senses can be trusted? That’s the $64,000 question in Dick’s work. How do we know that the events of Minority Report are real after Anderton has been sent to prison, and not all a dream? (It’s not real, of course, it’s a movie, but how do we know we are meant to take it to be real? Can we trust what we see?) Do we know that an individual’s perceptions of the world are the same as everyone else’s? Do we all see the same color blue?
Dick distinguishes the individual’s private world (idios kosmos ) from the group’s shared world (koinos kosmos); individual perception often wins out over collective views. Think of the sign in A Maze of Death—it reads “STOPPERY,” “WITCHERY,” “HIPPERY HOPPERY,” and so on, according to who sees it. Or how in Eye in the Sky, characters pass through four different “realities” modeled after the way four of them perceive the universe. Everyone’s own personal parallel universe looks like reality to them. At the end of the novel, the characters assume—perhaps wrongly—that they have returned to authentic being, the world as it actually is.
Philip K. Dick and Philosophy Page 24