This throw-back Thursday brings a piece from Herb Kauderer on his poem “Ghosts of Robots” (see end for poem) that appeared in our September/October 2018 issue. Herb is uniquely qualified for this post about robots, as he worked with industrial robots in the eighties—read on for a fascinating view of robots both real and SFnal.
by Herb Kauderer
I spent the 1980s working with touch screen computers and industrial robots, things that seemed impossibly futuristic in that long ago era when smartphones and Roombas were still decades away. This was part of twenty years I spent as a Teamster, during which my natural abilities with technology provoked my employers to toss me the keys to a wide array of very expensive contraptions. Yes, I’ve driven tractor-trailers and tank trucks and tow motors of every variety, but I’ve also used hand-held scanning computers, multi-million dollar pasteurization units, and automated pipe-cleaning systems that handled more than a mile of pipe at a time.
I retired from all that to become an English professor, which certainly caused some cognitive dissonance, but also provided an unusual vantage point from which to consider the role of robots in literature, as compared to their real appearance in factories. Back in the eighties, the industry defined robots as machines that made decisions of sufficient complexity without human input. There was occasional discussion of where the cut-off line was but such things are difficult to make hard and fast. The industry generally agreed on what was robotic, and what was merely mechanical, even if they couldn’t clearly define the difference. In the bulk of the science fiction I’ve read, decision-making was the province of Artificial Intelligences which were often stationary, and sometimes presented as immoveable. When I think of Multivac in Isaac Asimov’s stories, especially “The Last Question,” I imagine a monolithic computer that humans approach with questions. Despite my impression of Multivac, it is not lost on me that Asimov placed the story at the end of the ‘Robots’ section of Opus 100. Dividing artificial intelligence and robotics is difficult, and this discussion is partly of perception rather than reality, so let me clearly state: I am writing about what industry called robots when I worked with them. They were not AI in any way. Furthermore, all my generalizations about SF are based on my subjective consumption, admittedly voracious but far from comprehensive, of books, magazines, television and film. You may have a different experience base.
Perhaps the most interesting example of robots and robot multitasking was a whole room that was a robot.
In SF, robots move. Even that first common real-life robot, the Roomba, scooted around the floor vacuuming and spooking the cat. The industrial robots I worked with all did things. Part of the unspoken definition required that the robot not only make a decision, but take an action if needed. But I never dealt with a robot that was mobile, in the free-to-move-about sense, and even robots that took a lot of actions moved from an anchor point. They did not walk or roll around. They stood in one place and worked. From the dichotomy of genre robots that are usually able to exit the building, and industrial robots that are stuck in the building, or even are the building, many story possibilities arise.
Another part of the unspoken definition of those old industrial robots was the ability to detect key facts about their surroundings. They contained sensors, and like humans, a variety of different kinds of sensors. They might have micro-switches (mechanical switches, physically tripped, like bumping into something), photoelectric eyes, thermometers, pressure sensors, and motion detectors. There were often many of each kind of sensor in a given robot. Simpler industrial machines had most of the same components, sensors and otherwise, but the robots tended to have much more complex sensors. The idea of robots detecting their surroundings is well imbedded in SF.
In literature and society robots are often presented as far more dependable than humans, to the point where machinelike is an adjective meaning successful in a repetitive and errorless way. The reality is that those early robots of the 80’s broke often. Any developing technology needs time to work out the kinks, so I assume the state of the art today needs less constant attention. In fact, I assume that if I returned to that factory today, I would find robotic portions of it unrecognizable. That said, as I write this, a front page article in the local newspaper relates the Chinese mania for robots, even though they often fail. Perhaps the breakdown rates are not so different after all.
If those early robots broke often (far more often than other machines), and needed attention and maintenance, then why did the company bother with them? The main reason, of course, is cost efficiency. If the cost of production per unit goes down when robots are integrated, it makes sense to use them. Since 1870 and the epic steel-driving battle of John Henry against a steam drill, it’s been clear that machines had the ability to out-produce humans. In SF this is often presented as happening by having one robot that can do human things better and faster than humans, but the main way this happens in industry is by one or more robots performing a higher volume of work at once. I know firsthand that a robot can create thirty-six plastic jugs at once compared to a human’s one at a time.
Perhaps the most interesting example of robots and robot multitasking was a whole room that was a robot. This was a reality that I dealt with that seems more perspective-changing than a lot of common SF tropes. I would love to read (or write) a story or poem about a robot that was a room and the challenges that might present if it was self-aware. In this specific case, conveyors carried stock into the robot/room. The robot analyzed the incoming stock, compared it to orders, estimated expected numbers for orders not yet official, sorted the stock which arrived in varying shapes and sizes, arranged it on wooden pallets, and securely piled it five feet high, wrapped the whole in plastic, taped a labelled order sheet to it, rolled the pallet down a conveyor and out of the room, picked up an empty pallet and put it in the queue. The robot often worked on six orders spread across twelve pallets at once.
In her amazing book The Travels of a T-Shirt in the Global Economy, Pietra Rivoli brings up another issue when she sums up the popularity of mechanization by arguing that with a “machine, they could drastically reduce once and for all their risky association with farm labor” (36). Why was human labor risky? I certainly don’t intend to speak for all factory workers, or all Teamsters when I share my experience. But where I worked, there were frequent bad attitudes among workers. It may have been justified by management and circumstances, but those bad attitudes toward management were most certainly missing from the actual robots. In the literature, I have seen many examples of robots with bad attitudes. Maybe someday, industrial robots will advance to the point where they can have an attitude. Perhaps sassy robots are more fun to include in SF. But in reality, they currently act as the machines they are, with no sass towards the boss.
On the other hand, if by some chance the men loading trucks were slow to move the pallets exiting the room/robot onto their appropriate trucks, the robot/room would ask them to get their lazy human behinds in gear and catch up. True, this is the result of programming decisions, but robots are not required to be polite, especially to humans of lower socio-economic status; something else that I have not often seen presented in the literature. Sassy, yes; cursing and profane, not so much.
If something malfunctioned to the point the robot could not handle it, the robot summoned maintenance personnel from their workshop six hundred yards away. On those occasions the room/robot would be quite polite. One of the rules of the factory is: Don’t antagonize the maintenance men. The robots had that programmed in.
When I wrote my poem “Ghosts of Robots” I took off from a different point in the dichotomy. The two points of interest to me were efficiency and other complexities. I began the poem at the robot recycling center, because the heart of industrial robots is efficiency. Most mechanics scavenge parts, but for the industrial robot it is especially important because of the high break down rates, and because the parts are so sectionally specific. When I imagined the robot recycling center I did not see a clean and shiny place of uniformity. I saw a chaotic place of odd lots where builder robots examined the pieces of non-functional robots to see what functional robot, if any, they could build from the pieces. Hence, I envisioned a place where robots were being reborn, and others were being permanently decommissioned.
In a place where pieces of robots are being reused, the idea of ghosts seems obvious. I am a registered organ donor, and science fiction and horror literature have long asked fascinating questions about the connection of donor and donee. Where does the essence reside? What traces of a person are left after death? Most robots are incapable of not leaving traces after reactivation.
When Isaac Asimov wrote about Multivac it was in an era where the monolithic computer existed and was expected to persevere. I used a few of those giant computers in the 1970’s. But the 1980s were an era of decentralization and distribution of computing duties and power. Any given industrial robot had computing chips (though we didn’t call them that) spread throughout and in communication with the CPU. And it is almost impossible to eliminate all memory from a chip. In other words, if robots are being recycled into new robots, then traces of the old robots’ programming and experience migrating to the new robot were almost inevitable. This is a type of post-corporeal persistence and for me that evokes the idea of ghosts.
The other point was complexity. In the eighties I had no sense that complexity was relevant, that a society had to demonstrate sufficient complexity to be considered a civilization, or any of the other uses of complexity for measurement and labelling. Certainly the current idea that sufficient complexity of data handling is a prerequisite to the emergence of consciousness, was not yet under discussion. Even as the year 2020 approaches, I have not heard of the concept of complexity facilitating emergence extended to many other areas. A recent major study determined that there is a strong correlation between a high number of daily social interactions and having a long life. That may reflect complexity. (I expect social interaction research to continue to develop and expand in the coming decades.)
A large portion of my poem was an attempt to answer the question of whether sufficiently complex social interactions could allow for the emergence of a soul, even in a machine. And then, what might be the post-corporeal persistence of that soul? While Sturgeon’s first law of literature is often quoted, I am fonder of his second law: “Ask the next question.” For me, that was the post-corporeal nature of the robot soul. And my initial answer to what that might be like was ‘helpful and efficient.’ But with a little more reflection I realized that one distinct component of social interaction was ceremony, which in ways is the celebration of complexity. With that, I finished the poem, and the journey from the immobile specialized industrial robots of the 1980s to complex robots in more-or-less their own world was complete.
Asimov “The Last Question.” Opus 100. NY: Dell, December 1980, copyright 1969.
Garst, John. “Chasing John Henry in Alabama and Mississippi: A Personal Memoir of Work in Progress.” Tributaries: Journal of the Alabama Folklife Association #5, 2002 pp. 92-129.
Mozur, Paul. “Chinese Are Embracing Robot Technology, Even When it Comes up Short.” The Buffalo News 22 July 2018, A1.
Rivoli, Pietra. The Travels of a T-shirt in the Global Economy: An Exonomist Examines the Markets, Power and Politics of World Trade. Hoboken, N.J.: John Wiley & Sons, 2005.
Sturgeon, Theodore. “Ask the Next Question.” Gunn Center for the Study of Science Fiction. The University of Kansas. http://www.sfcenter.ku.edu/Sturgeon-Q.htm.
Herb Kauderer is an associate professor of English at Hilbert College and holds an MFA in creative writing from Goddard College as well as a PhD from University of Buffalo after defending his dissertation titled Conspicuously Canadian: Canadian Identity in Anglo-Canadian Science Fiction. He is the author of over fifteen hundred published poems, including sixteen books, and he served as main screenwriter for the independent feature film Beyond the Mainstream (2013). His work has won the 2016 Asimov’s Readers’ Award, the Ewaipanoma Sonnet Contest, and a WorldCon Poetry Slam. He has been a finalist for the AnalogAnLab Readers’ Award, and received Honorable Mention in The Year’s Best Fantasy & Horror. His work is often nominated for the Rhysling Award, the Pushcart Prize, the Dwarf Star Award, and the Elgin Award. More can be found at HerbKauderer.com.