Conversation: Empathy & Computers

On February 22, 2011 at 4:48 p.m., I set up a private blog, where I could regularly engage in conversation with a group of friends from across disciplines. The process was to work as follows. I would post a piece of writing on the blog, they would comment on it, then based on their comment, I would not only revise the writing, but also feel encouraged and inspired to keep writing.

The outcome of the conversations was the book “Realizing Empathy: An Inquiry Into the Meaning of Making,” which was successfully kickstarted on March 12, 2012. While much has changed since then, with their permission, I would like to share with you an edited version of several of those conversations with you, regrouped and rearranged for clarity and relevance. Here is the first installment.

 

an-lon: Chan1, can you start with a round of introductions? Who are they, the people reading this blog? How do they know you?

me: Oh, of course! Yes, let’s do that. Perhaps we can say what we do—not what our titles are—what our interests are, and where we are coming from. I think this will do wonders in enriching the conversation.

And before I forget, I just wanted to sincerely thank you all for participating in this journey of book writing. The past two-and-a-half years have been a time of divergence. It was time I desperately needed to get away from my previous environment, to find new ways of thinking.

In retrospect, the question I was ultimately after was the question of what makes us human. Much like the pioneers of computer science, I started wanting to understand better how “thinking” works, what “consciousness” is, how we “learn,” how we “understand” something, and what intelligence means. But unlike some of these pioneers, I was not interested in asking how we can abstract the “humanness” from ourselves, to disembody it, so as to put it in some other body, and to debate whether that other being is also human. Honestly, I can’t see why this line of questioning is valuable. What this kind of disembodied attempt at manifesting humanness can do, at best, is superficially mimic or simulate what one may mistakenly believe a human being to be, without any real understanding of what it actually means to be human.

At the same time, I had a deep attachment to the computer. In my professional life, I have spent a significant portion of my career programming the computer. In this process, I have often found myself totally immersed in thinking from its perspective and not mine. I would dare say that I empathize with it. Yes, I know that most of us think we only empathize with living beings. But from my experience at the Rhode Island School of Design (RISD), I’m starting to question this assumption. Because I’ve discovered many similarities between the process of trying to empathize with human beings, with that of trying to make physical things. And that’s precisely the vantage point from which I will start to write.

As I wrote in my e-mail invitation to you, my goal is to write in the company of a handful of people I feel comfortable sharing my ideas with. I then hope to get feedback, revise, and eventually integrate everything into my thesis book, which I will produce for graduation.

I would absolutely love it if you all could be candid about giving me feedback. I will inevitably make some strong statements that may seem controversial. I expect counter arguments and tangential references. I firmly believe that it is from the experience of contrasts, of seemingly unrelated or different experiences where new ways of thinking can arise. With that, I now pass the mic to you all. Thanks!

joonkoo: I thank you for inviting me to this very interesting forum of discussions. I don’t know how many are here, but let me start with the mic. I’m Joonkoo Park. I know Seung Chan hyung2 from high school. I went to the International School of Beijing (ISB) in 1994  to  1997. I’ve always been proud of Seung Chan for his free-minded spirit. He does what he wants to do and he does it well. So I was really glad to hear that he decided to study fine arts. And it looks like that was a real success. I was glad to join this discussion since I wanted to do anything to help him organize his thoughts and formulate ideas.

I study cognitive neuroscience of high-level vision. I am in the final stage of my Ph.D. at the University of Michigan. Most of my work is centered around neural organization and mechanisms of object recognition, such as faces, letters, and numbers; however, I’m getting more interested in numerical cognition, and I am planning to study the neural basis of number sense during postdoc.

That said, I’m trained as an experimentalist, and my interest is pretty focused — as many Ph.D. students are either forced to, or are trained to be so. But any questions related to how the mind works trigger my interest, and I wish to be of help by bringing in some neuroscientific and psychological ideas into Seung Chan’s thesis and the discussion.

david: My name is David Watson. I like to have “an attitude of gratitude” though I think it gets lost in a lot of what I do or say. Slim knows this from working with me for a couple years at maya design in Pittsburgh. I have a deep need to understand reality in its purest form, to seek the highest levels of production quality even when they don’t matter to anyone but me, and to achieve symmetry in literally everything.

They say that at the root of engineering is this “truth seeking” and you’ll see elements of that here from me. I apologize in advance for my forthrightness. I have a way of speaking that’ll make you think I think I’ve known you for 20 years.

I like the way Slim has defined the introduction, because while I work in software , I don’t like to define anyone singu- larly, certainly not myself, and I like to think of this more as “creative mediums of expression.” I’m a musician, a photo- grapher, a skier, a cyclist, a runner, a thinker, a reader, and a writer. And I’m glad we’re not all the same. Nice to meet you all. Cheers.

jeff: Hi, my name is Jeff Wong. Slim and I are intellectual buddies from Pittsburgh. He was a working man on the South Side of Pittsburgh at maya design, and I was a Ph.D. student in human-computer interaction (HCI) at Carnegie Mellon University (CMU).

I will be bringing my background in computer science and cognitive science. Slim says I can bring some perspective from theoretical computer science. Also, I know some of conceptual history of Artificial Intelligence (AI), psychology, cognitive science; some familiarity with psychiatry, phenomenology, and tidbits of religion and spirituality; and some dabbling experience in philosophical thinking. I don’t know how deep my knowledge is in these areas, but I think I can at least point to relevant ideas and prior explorations that have happened in these areas.

anson: Hi everyone. My name is Anson Ann. I really thank Slim for inviting me to this group. I feel so honored to be able to take part in this conversation. Reading the background stories, expertise, and interests of you folks really humbles me. It does start to feel a bit like a mini ted here!

Well, I first met Slim in CMU back in 1995 Although we were not in the same department — he was in computer science, and I was in electrical and computer engineering — our dorms were close to each other, and we had a common interest :  music and guitars. We used to take a cab together to a local musical instrument store and drool over those guitars and music gadgets. We first looked at guitars together, then he gradually moved onto djing, and I moved to synthesizers. Anyway, no matter what we do now, I think both of us will always have a musician inside us.

After CMU, I worked at bbn technologies — now Raytheon—in Boston as a speech software specialist/scientist. We customize speech software solutions for the u.s. government intelligence community, enabling them to do speech recognition, machine translation, and information extraction on Arabic, Chinese, Spanish, and English broadcast news all over the world. Much of my work involves language model training, pattern recognition, signal processing, and some human-computer interaction.

Then about six years ago, a big turning point happened in my life. I sensed this calling from God for me to become a pastor. Just like Slim who took the challenge switching from science to fine arts, I quit my software job and enrolled in a theological school. It was quite a big stretch for me, for studying the humanities requires a very different temperament. I realized that my mind, which was trained for engineering, preciseness, and comprehensiveness, wasn’t ready to deal with the ambiguity, complexity, tension, and paradoxes that you often find in history, literature, religion, and philosophy.

I have just finished my studies and now I’m an Anglican priest pastoring at the Anglican Network Church of the Good Shepherd in Vancouver, Canada. And as a pastor, what I hope to contribute to this conversation is my anthropological understanding.  ( i.e., What does it mean to be human.)

First, I am going to be upfront about my faith and convictions. I will speak from a Christian perspective and understanding of what is means to be human, for I believe anthropology stems from theology. A core doctrine in my faith tradition is that of the Holy Trinity: that God, who revealed himself to us in history, is known to be a three-in-one relationship, a unity-in-diversity, a dynamic-yet-unchanging entity, a harmonious dance in reciprocal love that overflows with creativity and creational power.

Since the entire universe is created and sustained by a relational being, the very core of our being and reality is supposed to be relational. And as we human beings are made in the image of this relational God, so we are also ontologically relational.

We are made to relate, to empathize, to love and be loved. There is something intrinsic about human beings that we want to understand and be understood. I believe the torture of imprisonment is not just lacking freedom, but more about losing the ability to relate to others and the outside world. Relational beings unable to relate are just like fish out of water.

Anyway, I know not all of you are religious or spiritual, so you may or may not share my perspectives, but I just hope I could, in some way help inspire Chan to continue exploring this topic about empathy. I’m still very much a geek at heart, so I also hope I can contribute to the other aspects of this conversation about computers, programming, and human-computer interaction.

I look forward to seeing Chan write more. Because empathy is about listening first, isn’t it? I hope we won’t flood his comments area with too many of our own ideas, but let him express what he wants to say first, then respond to him accordingly.

an-lon: Nice to meet you, I’m An-Lon Chen. As for me, I also went to the International School of Beijing (ISB) with Chan for a semester of high school in 1994 ,  I was a senior, he was a junior, and we both worked on the yearbook together. That would be the end of the story right there, except ISBers tend to be a close-knit bunch and it seems like for the past decade and a half we’ve always had each other’s contact information via some mailing list or another without ever actually interacting personally.

Let’s just say that my relevance to this project is that over the years I’ve gone from comparative literature to computer science to user interface design to computer graphics to character animation. I am now a full-time student at AnimationMentor. Prior to that I was at DreamWorks working on Shrek 4, and before that on Mummy 3 at Digital Domain.

Perhaps more than anyone else I know, I’ve had to approach computers and computer science as much from an anthropological perspective as well as a technical one —deciphering a subculture and a jargon in order to pass as a native.

I’m a geek at heart, with my fair share of the stereotypical social inadequacies . I’m pretty sure I was born socially tone deaf, and only as an adult began to figure out the nuances of interacting with others. That said, pretty much every big break in my unlikely computer science career has come from possessing some unusual degree of empathy . First, the amateur exercise in anthropology that drew me to geek culture; second, the turn towards user interface development, which brought me to LA and the film/vfx industry; third, the current foray into character animation, which is all about convincing audiences that dead pixels can walk, talk, laugh, and cry.

Point being, I care personally about Chan’s topic: computers and empathy. I have no earthly idea where this blog is going or how it’s going to become a thesis, but I’m following it because at least some aspects of it touch on things that I, too, have been wondering all my life.

But now, can we start from the top? Empathy? With computers?

me: Yes, with computers.

an-lon: The best developer is inevitably quite the computer whisperer, of course, but I would have never actually thought of that rapport as empathy.

me: What would you have thought of it as then?

an-lon: Two synonyms came to mind. One was grokking, the other was acculturation. Grokking, because, well, I’m a sci-fi geek and I like having a word to express deep understanding and truly “getting it.” Both my parents are scientists, and while they are both fairly technically-savvy — my dad has written Fortran and assembly language code and my mom uses Photoshop and Illustrator for graphics and charts in scientific journals — I’m not really sure either of them has ever grokked computers ;  their mindsets are a little too unyielding to ever completely get on the computer’s wavelength, so they often end up fighting the computer in unnecessary ways.

For example, if my mom gets a PowerPoint concept stuck in her head, she invariably has trouble figuring out the Photoshop equivalent because she’s speaking a different language without even knowing it. Both my parents are extremely good within limited contexts, but don’t have the particular empathy required to troubleshoot — learning a new domain comes slowly.

Wait…

Ha ha ha ha ha ha!

I just looked up “grok” on Wikipedia. It says, to grok is “to share the same reality or line of thinking with another physical or conceptual entity. Author Robert A. Heinlein coined the term in his best-selling 1961 book Stranger in a Strange Land. In Heinlein’s view, grokking is the intermingling of intelligence that necessarily affects both the observer and the observed.”

But, it gets even better.

It also says that the Oxford English dictionary defines grok as “to understand intuitively or by empathy; to establish rapport with” and “to empathize or communicate sympathetically (with); also, to experience enjoyment.”

Drum roll, loud banging of cymbals. (Smiles) I actually wasn’t as far away from your wavelength as I thought. I had turned to Wikipedia as a joke, but it did hit a nerve. What I took away from it is that it’s never a purely intellectual exercise to really truly understand something, be it an immediate piece of code or an underlying computer science concept.

me: Exactly! There is an inextricable link between empathy and the act of learning that is non-obvious to most people. And I think it’s non-obvious, because we’re used to separating the cognitive from the emotional, or the mind from the body. For example, I’ve heard many people say that empathizing is not the same as understanding. On the surface, there’s nothing significant about that statement. Of course they’re not the same. If they were, why would we have two separate words? But it becomes significant once you realize that what people mean is that understanding is inferior or shallower than empathizing. Well, that depends on how you define the two words. It is only so if you include inaccurate understandings as a legitimate form of understanding. I would argue that an accurate understanding of an other cannot be had without having tried to empathize with them.

anson: That reminds me of people telling me how I have an extra-ordinary amount of patience in front of a computer. I don’t know if it’s because I know what the computer is doing inside, but I can be patient even if it’s slow or stalling. And also, whenever my dad encounters a computer problem, he always asks, “How stupid is this computer! Why can’t it do this and that?” and I always feel like I’m defending the computer saying, “It just can’t… this is what it can and cannot do. Don’t be too hard on it. Be patient. It’s still crunching numbers. There’s nothing you can do except rebooting the machine. And here’s a way to work around its limitations .  Yada yada yada.”

Isn’t that also related to empathy?

me: Yes, I would definitely say so. I don’t think you can have patience for the computer, if you cannot empathize with it. We’re much more likely to lose patience for the computer, when we cannot empathize with it. Just think of a time when you got the hour glass or the beach ball for no apparent reason. That’s very difficult to empathize with. That’s like interacting with someone who is too pissed off to tell you what is going on.

I think the best example of people trying to empathize with the computer is when they’re debugging. When we are debugging a software program, we are trying to figure out why the program is behaving the way it is, and our head gets filled up with nothing but our understanding of the state and configurations of the program, not to mention the various hardware mechanisms like memory, processor, and external storage. What I’m doing is trying to think as if I were the computer.

an-lon: I couldn’t agree more about the debugging. Perhaps we use our empathetic faculties for debugging because our brain cells weren’t equipped to access that deeper cloud of intuition any other way. Our brains have been wired for millennia to interact with fellow humans and, as far as I’m concerned, it’s an extremely useful act of hijacking to tap that empathic cloud in order to outsmart a machine.

And just to be clear, I’m not talking about anthropomorphizing the machine, I’m talking about accessing our own preexisting, well-developed resource of empathic faculties to interact with it. Oddly, I anthropomorphize just about everything under the sun: teddy bears, disappearing keys, food that’s been in the fridge for too long. But I’ve never been tempted to anthropomorphize computers.

me: Didn’t you say there were two words that came to mind? What was the other one besides “grokking”?

an-lon: Oh, acculturation.

me: Why did that come to mind?

an-lon: Because computers and programming languages are created by humans. The magic that bridges the abstraction of 0’s and 1’s with human neurons is language. Windows and mice are metaphors—picture a window in a house, picture a mouse running around on its four little paws, now marvel at the metaphor that got us where we are today! Zipping, unzipping, bugs, these are all metaphors. Computer concepts are all abstract until we give them names and map them to something we can understand. Even the act of stepping through code using a debugger is a concession to a human need for a linear story line.

Learning how to program is similar to language acquisition, not because computer languages are anything like natural languages — allowing C++ to fulfill a language requirement like French or Spanish would make no sense — but because learning how to write code is very much a process of acculturation. Just as it’s pretty much impossible, or at the very least pointless to learn a natural language without a cultural context, it’s impossible to write code well without absorbing its many sub-cultures. Best practices, conventions, idioms, and design patterns are all cultural constructs within a human community, not semantic ones within the machine.

Put in that light, the idea of empathy with computers is staggeringly mundane — we’re talking about forming a rapport with the community of their very human creators, not a sentient and malevolent Hal. And yet, my rapport with, say, Linus Torvalds goes through multiple layers of translation, not the least of which is through the machine and back. And if I were to go out and write a Linux patch, I’d damn well better have empathy with the Linux operating system so I can design something appropriate  . . . and yet it doesn’t feel like real empathy. It’s not real the way a spoken word is real, a heartbeat is real, the touch of a hand is real. And yet, to anyone who’s ever gone into a programming trance and been absorbed to the point of forgetting to eat, sleep, or shower, it’s profoundly real, perhaps more real than reality. It’s an emotional state as much as a physical one.

david: Slim, how are you going to treat this subject in a secular fashion? It’s going to be very difficult, because so much of it leans toward feeling and emotion as opposed to logic, science, neurons, etc. It barks up the quality / quantity tree that is split down the center and very divisive.

me: Hmmm… I never thought of this as non-secular. Metaphysical, yes. Are you equating the two?

david: Well, just as Zen and the Art of Motorcycle Maintenance is not about Zen, I don’t think what you’re after is about empathy. It’s deeper than that. To empathize with the computer is to anthropomorphize. To anthropomorphize is to visit our expectations on reality. Computers aren’t humans and they never will be. Can man make a better human? Probably. Will that human have better distinguishing human characteristics? No. In the very same sense that James Howard Kunstler argued that architecture was moving toward a loss of a sense of place — which I agree with — robotics goes down the same boring path, most likely because it has no other choice. If it didn’t, we’d be defining an engine of individua- tion, and I’m pretty sure nobody is doing that. And that’s the miracle of humanness.

jeff: Speaking of humanness, historically speaking, artificial intelligence referred to abilities that we thought computers were incapable of. However, as solutions to problems began to appear, these abilities — like chess playing and language recognition —were no longer considered intelligence. How computer programs tackle intelligent tasks is always different from how humans actually do them; sometimes better or more thoroughly, but at other times, seemingly stupid.

me: What do you mean?

jeff: Why does ai trip up on “special cases”? Because the way we program intelligence is by making problems formal (i.e., accessible to the computer). When problems are formalized, they can be solved by rules. Where rules don’t quite work, we have rules for selecting rules or rules for creating the rules to select rules with ( i.e., machine learning). I think we approach problems this way because our way of accessing how we think, and communicating that to other people is in the framework of rationality. I think rationality is primarily a structure for thinking about thinking.

How we think isn’t quite rational. It’s more like rational++. When people appear rational, we can empathize with them. Irrational people, too, but not as easily. Part of anger is not knowing why. If you know how the machine works, you can be angry at the situation, but it’s not the machine’s fault, it’s whatever is broken or not working inside. Does anger require a thing to be angry at? If you’re angry at the thing and you don’t know how it works, it might as well have a mind of its own. It makes no difference to you. Consider the University of Texas clock tower shooter. You can imagine being angry at him for shooting someone you know, but then you find out he had a tumor and he requested an autopsy in a note he left at his house with his dead family. Somehow that kind of situation is a bit less angry because you know why.

Understanding the mechanism changes how we think about the thing. For example, I remember being excited about taking an ai class and learning the magic. But it turned out to be a whole bunch of hacks  —  or so it appeared. It’s no longer magic when you know how it works. Now, I don’t quite understand what you mean by empathizing with computers. What you’re doing when programming is simulating your program on your model of the programming language runtime. Yes, it’s sort of like empathy, but I thought empathy was being able to feel what you feel. We empathize with real people based on our concepts and experiences of other people. This is the theory of mind,3 which autistic people lack. For that, they are alone in the universe because other people simply don’t exist. So you need models of other people to empathize with them. Still, empathy is a feeling about feeling. I still don’t get empathizing with computers. My current idea of what you might be thinking seems wrong.

joonkoo: I second Jeff on this point. I don’t quite understand what it means to empathize with computers, either. I don’t necessarily think that you need to have answers to all these questions now. Some of them are certainly empirical questions, and worth investigating more. But, I would like to have a better grasp of your idea of empathizing with computers, and I still can’t quite get it. Perhaps it will get explained in your future writings?

me: Yes, it will. Although, I am starting to realize that I’ll be up against a lot of criticism, because some people may be equating the theory of mind with empathy.

Allow me to first write about my experience with physical materials, and I hope that will better explain why I think empathy is not exclusive to human relationships.

(Smiles) But most importantly—my God!—I can’t tell you all how much I love this tightly-knit discussion environment!

——

1 Some of my old friends call me Chan. It is the latter half of my full first name, Seung Chan. I had originally adopted the name to accommodate those who could not pronounce the first half of my name. But I have since abandoned this name because I feel that it robs me of my full identity. Those who cannot pronounce my name call me Slim—made from concatenating the first letter of my first name to my last name — which is a new identity I have constructed since my arrival in the U.S. How did the name come about? It was my e-mail username in college.

2 Hyung means older brother in the Korean language.

3 Theory of mind is the ability to attribute mental states—beliefs, intents, desires, pretending, knowledge—to oneself and others and to understand that others have beliefs, desires, and intentions that are different from one’s own. It is typically assumed that others have minds by analogy with one’s own, and based on the reciprocal nature of social interaction, as observed in joint attention, the functional use of language, and understanding of others’ emotions and actions. (Premack and Woodruff, 1978, 515–526) (Baron-Cohen, 1991, 232–251) (Bruner, 1981, 41–56)

Premack, D. G.; Woodruff, G. (1978). “Does the Chimpanzee Have a Theory of Mind?” Behavioral and Brain Sciences 1 (4): 515–526.

Baron-Cohen, S. (1991). “Precursors to a Theory of Mind: Understanding Attention in Others.” In A. Whiten (Ed.), Natural Theories of Mind: Evolution, Development and Simulation of Everyday Mindreading (pp. 233-251). Oxford: Basil Blackwell.

Bruner, J. S. (1981). “Intention in the Structure of Action and Interaction.” In L. P. Lipsitt & C. K. Rovee-Collier (Eds.), Advances in Infancy Research. Vol. 1 (pp. 41-56). Norwood, NJ: Ablex Publishing Corporation.

Gordon, R. M. (1996). “‘Radical’ Simulationism.” In P. Carruthers & P. K. Smith, Eds. Theories of Theories of Mind. Cambridge: Cambridge University Press. (59-74).