Dr. Cliff Nass’ career owes much to strange situations, people who were willing to take risks, and a devotion to pure weirdness. He’s the ultimate example of the “odd bird” in our field – the one who wandered through disciplines until he helped create a new one. Along the way, he picked up the best morsels of research and tidbits of insights to bring with him. Today, he sheds light on the way people interact with computers by reminding us of the ways we interact with people.
Excerpts From the Interview
There is a misunderstanding that ‘computer’ and ‘mediated’ are radically different; that if it’s a medium its a medium and if it’s a direct relation it’s a direct relationship. There’s absolutely no understanding of the fact it can be both.
Just steal stuff; it’s all there. It’s not hidden. It’s not secret. It’s really all there and it’s just a matter of stealing it appropriately.
An Interview with Cliff Nass
Conducted by Tamara Adlin on July 11, 2007 06:04 AM
Cliff Nass revels in being weird, thinking “wildly,” and taking “big fliers.” But he’s also fascinated by what makes everything the same. If we were all as open to oddness as he is, the world would be a much more interesting place.
Tamara Adlin: Today I’m very excited to be talking to Cliff Nass, who is a professor in the communication department at Stanford University, and a renowned authority on human computer interaction.
He also holds courtesy appointments in the computer science and sociology departments, as well as in the interdisciplinary programs of Science, Technology and Society (STS) and Symbolic Systems (SymSys) at the University. Cliff is the director of Stanford’s Communication between Humans and Interactive Media (CHIMe) Lab, and co-director of the Kozmetsky Global Collaboratory.
Cliff, the reason I am so excited to be talking to you today is partially because of your work in your books – one of which is The Media Equation: How People Treat Computers, Television, and New Media Like Real People and Places (CSLI Lecture Notes). You also wrote Wired for Speech: How Voice Activates and Advances the Human-Computer Relationship.
You are actually one of the first people who inspired me. I heard something really different when you gave a presentation on how people treat computers just like people.
Cliff Nass: Thank you.
TA: I always start the same way. Way back when, as far as you can remember, what first fascinated you? This could be when you were a kid, or when you were in school, or when you were deciding on college.
CN: From the time I was a young child, I was very interested in math. I was one of those people who can multiply large numbers their heads very quickly. I was always told, “So you should be a mathematician.” Of course mathematicians don’t multiply large numbers in their heads.
TA: Wait a second; how old were you when you could do this?
CN: Six or seven. I could do that stuff really fast, and would entertain family members with it.
TA: That’s amazing.
CN: Like many kids who can do it, I lost it in adolescence, but math was always my career trajectory, even through college.
TA: Could I ask you one more thing about that? Do you remember what that felt like? Did you see it? Did you feel yourself doing it?
CN: Yes, I sort of just knew. It felt like I knew the answer so it was much more that the answer came to me. I couldn’t do like some people, square roots and all that. I could basically add and multiply very quickly; that was it.
It wasn’t some weird mysterious thing. I sort of knew the answer. It wasn’t as sublime as one would have hoped or thought it might be.
TA: Well, it’s actually kind of sublime to those of us who couldn’t do that.
CN: But it just seemed normal. If you can do it, then it just seems normal.
TA: Then you were told you should be a mathematician, even though, as you said, mathematicians don’t do that. Did you pursue math in high school?
CN: Yes. I knew from even 4th or 5th grade I was going to focus on math. When I applied for college, I planned to be a math major. Throughout high school I had always planned on doing math and had taken as much advanced math as I could, so it was pretty much written that I was going to be a math major.
TA: Did that happen?
CN: Actually it did happen. I wasn’t all that good at math – I was okay. After bouncing around majors, I did in fact end up with a math degree, but even then it was clear to me that I’d never really become a mathematician. While I was good at it, I certainly wasn’t at the level of people who could become mathematicians.
TA: This was at Princeton, right?
CN: Right. As an undergrad, I decided that even though I was good enough in math, I wasn’t good enough to be a professor, which is what I had always wanted to be.
Then I went off to industry for a year, and worked at Intel doing computer stuff on the then-spectacular 286 chip, which in its time was really cool and fun. I did algorithm stuff for them, and then applied to graduate school.
Originally I was accepted to Carnegie Mellon and was going to go to graduate school in electrical engineering to become a computer scientist. Then there was a death in my family, which meant I postponed admission to be near home. Then, just to bide my time and to have some flexibility, I ended up doing a sociology degree.
TA: You entered a doctoral program in sociology to bide your time?
CN: Yes [laughs]. Strange people have strange lives, I guess.
In my last undergraduate semester at Princeton I was working incredibly hard on my senior thesis, which was on algorithms. I needed what looked like a really easy course so that I could really focus on my thesis. I saw this course in sociology that had no midterm and no final, and required just a paper at the end.
I thought, “This is perfect, an exactly made-to-order course for people who are killing themselves on their theses.”
I took the course and actually really liked the professor a lot, and got involved in the material.
I kept in touch with him after I graduated. When there was this death in the family, I mentioned to him that I needed to be east with my family. He said the sociology program had no required courses, no exams, nothing. It was perfect for me. So I did that for a year, biding my time, taking care of my family.
When it appeared that a year wasn’t going to be enough time, I took a second year in the program. At some point, I was offered a job at Thinking Machines, which was then the hot AI (artificial intelligence) company. That was my moment of decision. Do I go back to grad school, do I go to a company, or do I stay at Princeton?
For reasons I’m still not sure of, I ended up staying at Princeton and getting my PhD in sociology.
TA: It sounds like that professor was really influential to you. What was his name?
CN: James Beniger. He was incredibly influential to me. Sometimes you take a course you never really thought about taking, and all of a sudden it’s the one that changed your life. In this case, that was it.
TA: I am so interested in people who are influential and creative like that. Was there something specific about him?
CN: He had tremendously broad and exciting ideas. He was clearly excited about things. He took ideas very seriously but he was never afraid to admit he was wrong. He would take an argument and say, “You’re right, I’m wrong.” He really cared about the right answer, and it was inspiring because he was trying so hard for that right answer.
I had some crackpot ideas because I wasn’t a sociologist at all. His was the first sociology course and only the second social science course I had ever taken, but he was totally comfortable and willing to entertain my ideas and my approach. It was amazing.
TA: That’s how I would describe the work you ended up doing: Your sociology experiment, replacing one of the people with a computer. It was crackpot in the best way possible.
CN: You’re exactly right. He (Beniger) really was willing to play with ideas. He was enchanted by everything: History was interesting, science, social science… he really was an incredible inspiration.
TA: It’s wonderful to be in the orbit of that kind of creative energy.
CN: Exactly. Really thrilled with ideas; just loves ideas.
TA: So you stayed for another couple of years and accidentally ended up getting a PhD.
CN: Exactly. Then I didn’t know what to do because my PhD thesis was so bizarre.
TA: What was it?
CN: At the time I was working, the whole idea of the information economy was really exciting and hot. My question was whether you could think about the information economy in classical information processing terms.
In other words, if you looked at the economy as a computer that took informational inputs and pre-processed, processed, and post-processed them, then added a control function and an output function, what could you learn about the economy and the labor force?
Since I was a sociologist and not an economist, my focus was on the labor force. I had this wild idea that since this new economy is processing information (like a computer), we should be able to use those models to predict characteristics of the labor force.
As a sociologist, when you think about what predicts the labor force you traditionally think about the political system, ownership, etc. I said, “Let’s look at this as an information processing system and get rid of all the people stuff.”
When I did, it actually worked. There is this really cool book called The Dictionary of Occupational Titles [also online at www.occupationalinfo.org] that has 12,000 job descriptions.
TA: Wow. I’ve never even heard of that.
CN: It’s unbelievable. It’s an awesome book. You get these bizarre occupations like chicken sexers, who are people who determine the sex of chickens; and tennis ball felt cementers, who are distinct from tennis ball felt wrappers. You get this amazing diversity of details on peoples’ jobs.
TA: Maybe every high school student should be assigned to read that whole book.
CN: [Laughs] It tells you what jobs you don’t want. (You could say) “Here’s a chapter on jobs you get if you do badly in high school.”
We had some data sets that allowed me to map the occupational categories to the Census Bureau statistics on how many people had each occupation.
The other thing the Dictionary of Occupational Titles gives you is the level of the job – how complex a job is. I was able to find the complexity of jobs and the description of jobs in terms of occupation; what the job does. Then, by industry, I was also able to know how many people did it. Making some pretty bold extrapolations, I linked those data backwards. I was able to get data from 1900 to 1980 on changes in the skill level and types of tasks people did.
It wasn’t a perfect match, of course, because a job in 1900 and a job in 1970 were very different jobs.
TA: What were you looking for?
CN: We know from computer science that if there’s an increase in pre-processing, the complexity of processing decreases.
I wanted to see if we might find that the skill level of workers in processing decreases in industries where there are a lot more workers doing pre-processing.
TA: This is probably a bad analogy, but is it kind of like if you have a prep cook who’s chopping up all the vegetables, the cook doesn’t have to be as talented?
CN: Well, cooking is a little dicey in the sense that what makes cooks talented is their judgment of what to put in, not how they cut it.
TA: I see.
CN: To take another type of example, I say to you, “I want you to put these things in the right box.” If I first have someone sorting them for you, it becomes much easier. You don’t have to think very much.
TA: So you were already mixing fields in a blender.
CN: Totally. This was basically being a sociologist without ever thinking about people. There were no people in this model. I was looking at people as receptacles of labor and task and complexity.
TA: Do you think if you hadn’t ever taken that sociology class because it didn’t have an exam, you would still have looked for these kinds of combinations?
CN: Somewhat, because even in my job as an engineer, when I was at Intel designing the chip, I was trying to solve problems and ended up going back to some classical philosophical distinctions to make the jump.Even in my undergraduate senior thesis on algorithms, I resolved my question by going around to a whole bunch of people, watching how they solved a problem, and saying, “Okay, let’s get a computer to solve it the same way they did.”
So I was always interested in mixing and matching and gluing things together and stealing from other fields. I always had the idea that the easiest way to do things was to find an idea in one field, steal it and just transfer it to another field. That was true throughout my engineering work and in graduate school.
TA: You know what that reminds me of is best practices and design. It’s really just stealing the good ideas that other people have had and already proven.
CN: Exactly. But part of the trick is to steal them even when they don’t know they are doing them. That’s the harder part. There are two ways to look at best practice. One is to find the people who do X the best when you want to do X. The more interesting thing is to find out who does X the best when you want to do Y.
It’s interesting to see if you can learn from how they did X to make Y better, even if those guys doing X didn’t know anything about Y. The whole idea is that you make bold assumptions about similarity rather than difference.
I’m always interested in what is similar rather than what’s different. In psychology, they talk about individual differences: Men vs. women, happy people vs. sad people, etc.
I’m interested in exactly the opposite: What makes everything the same. For my dissertation I looked at what made economic systems exactly the same as computer systems. Or, what makes people respond to computers or technologies the same way they respond to people.
TA: I would argue that your ability to add complex things quickly never went away; it just changed.
CN: Thank you. Well, at least it’s looking for what’s the same rather than what’s different. There are two methods of discovery. That which asks, “What’s special about this?” can lead to discoveries, and that which asks “What’s not special about this?” can also lead to discoveries.
TA: You’ve got yourself an accidental wonderful Sociology PhD, and you’ve always wanted to teach. Did you want to be a professor specifically of math?
CN: Initially math, but I always wanted to be a professor in general. I loved the idea of being a professor. I had never met a professor before I went to college, so my sense of what it was to be a professor was all based on books and TV, but the idea of being able to teach people and discover new ideas – that just seemed so swell. It seemed like an amazing thing to be able to do.
TA: Were there TV shows where the professor was the hero?
CN: Yes, in those days. Of course the world has changed a lot, but whatever professors you saw, they were always charismatic. There were also high school teachers who were charismatic. The shows didn’t emphasize the research side, they emphasized the idea that you would take kids, help them, train them, etc. So the notion of the teacher – and especially the professor – was someone who would be a guide and change peoples lives through their understanding of things.
TA: Was Welcome Back, Kotter on then?
CN: Welcome Back, Kotter was on then. Room 222 was probably more influential, socially, and there had always been the view of the professor as this guy who helped people and did things, less the discovery side. But I also discovered that I loved to discover things.
In fact, the most important thing I ever tell students is that the coolest thing about research is that there’s a moment in time when you know something no one else in the world knows. And then you tell them.
One of my former students even put that quote on a paperweight for me.
That’s my thing; the idea that you could know something that no one else could possibly know. What could be a more exciting or powerful thing than that? Telling them, of course, is the difference between a Faustian approach and the more pro-social approach.
TA: It makes you into an adventurer.
CN: I imagine it’s the same thing that motivates explorers. Now, usually in the case of explorers someone else saw the thing before, but at least no one from your neck of the woods had seen it. It’s still the same idea.
TA: It’s certainly a more exciting idea than “publish or perish.”
CN: Right. It’s something that can energize students as well. As kids growing up, even if we like to think our parents are idiots, we always have the belief that there’s something they know that we don’t. So to be able to think, “I know something no one else in the whole world knows,” whoa, that’s got to be cool.
TA: What did you do next?
CN: I was doing my dissertation and was not done.
A job came up in the communication department at Stanford that sounded like it could plausibly include me. If you think about it, the type of work I did didn’t neatly fit into any area of sociology or much of anything else. It’s very hard to explain to a sociologist why you would want to study society in ways that have nothing to do with people, since people get into sociology usually because they are interested in people.
And then computer scientists would be baffled as to why and how you’d worry about the skill level of the work force. It’s a nutty topic. I was going to be in trouble getting jobs. I had published stuff and was doing work and all that, but my dissertation was so weird.
Fortuitously, this job at Stanford came up. Even though I wasn’t ready, I applied. Then my advisor and I had this whole intricate scheme of stalling if I was asked to interview, because I wouldn’t have been ready to give a job talk at that point. But it didn’t work.
I came out (to Stanford) and had to do the job talk anyway. It was very well received. Even though they were thinking, “this guy is way out,” fortunately at that time that is what they were looking for. They were looking to be bold and daring; to take someone who was clearly way out – and that’s how I got the job.
I was applying in communication, but knew nothing about the field. We didn’t have a communication department at Princeton. I just didn’t know anything about it and it was totally wild and bizarre.
TA: The competition must have been interesting. You must have been very surprising to them, actually.
CN: Yes, I was. The other two people were in the field of communication, had degrees in communication, knew what communication was, and went to the conferences. No one had ever heard of me. I’d never been to a communication conference. I really knew nothing.
TA: So did you have to instantly dive into teaching communication classes?
CN: Yes, but fortunately they were amazingly good about that. They let me teach research methods, which look much the same in communication as it does in sociology. And they let me teach technology, so that was easy for me.
For my graduate courses they initially had me ostensibly co-teaching, but really I trying to learn the material along with the students.
They were very supportive and that helped a great deal.
TA: Did you start there in ’87?
CN: In ’86. I got my PhD in four years. I was originally going to take five but when I got the job, the people at Stanford wanted me to get started right away. So again it was an amazing scramble to finish my dissertation, but I got it done.
TA: Did you start immediately looking around for ways to make connections with other departments, or did that take a while?
CN: I immediately got an appointment in sociology, because the people in communication had talked with the sociology department about me to make sure that I was legit.
The other appointments evolved over time. I wasn’t dying to find other departments and things like that.
I’m not great at collaborating with people. I’m great at collaborating with students, but not as much with faculty, so I didn’t push very hard to find faculty collaborations.
TA: That’s such an interesting thing to say. Do you have any insight as to why that is?
CN: I really like to know who’s in charge. My best collaborations with faculty – and I’ve had some great ones – have been the ones where it was clear which one of us was in charge.
Soon after I got to Stanford, I got to work on a survey project with Steve Chaffee, who was just an amazing survey guy. It was fabulous because I knew my role was to learn from him, and he knew my role was to learn from him.
I don’t care if I am in charge or not; either one is fine with me. But there’s something about my style and personality that doesn’t do well in a mixed role.
TA: If there is a sort of a wrestling match that’s part of the project…
CN: Or even just confusion. It’s not like fighting, it’s about who is responsible.
The one exception where a mixed role worked was when I was writing The Media Equation with Byron Reeves. But even there we knew who was in charge of each chapter. There was an understanding that, since Byron had written a book before and I hadn’t, he was going to guide the process a little more. That worked fine.
Though I’ve done very few studies with other faculty, I have collaborated with probably 50 students. I have an unbelievable number of students on my publications listed as first author because that’s important for them.
But when it comes to doing research, I think it’s important to know who is in charge and, at the end of the day, making decisions.
TA: Your work is so fascinating because it is between disciplines, so that makes total sense.
CN: I love talking with people, and I get tremendous ideas from them. It’s just that at the moment you’ve got to sit down and start that project, somebody’s got to do something, and I really like to know who that someone is. That’s different than talking, communicating, and learning.
All of those are easy activities. But active research is really very, very tricky when you don’t know who’s the dude.
TA: Tell me how you started branching into what would become the topics of your books.
CN: Part of it was desperation. I was having a terrible time getting my work accepted. In fact, to this day I’ve still never published anything off my dissertation, 20-odd years later. Because again, no field could figure out who owned the material.
I got reviews like, “This work is offensive.”
CN: Yes, because it threw away the people.
I was saying you can view them as things that do stuff and the models work fine.
But people who get involved in the social sciences do it generally because they care about people, and here I was saying, “Let’s ignore all the people; they’re just noise and distracting. Let’s get right to what they are doing.” People got offended.
TA: I guess the math and economics people thought it was not enough.
CN: Right. It wasn’t mathematical enough for the math people. They like truly heavy duty mathematical models, and my work didn’t have those. Economics is a field where that has very special language and I never learned it, so clearly I wasn’t an economist.
The cybernetics guys sort of liked it, but they weren’t interested in the work force. Why would they think about the work force? They thought that was a little flaky.
I realized I was in trouble in a sense, because my work was crazy. I thought it was right and really cool, but it was clear that the world wasn’t quite there yet.
I just thought about what was it I could do that nobody else could do; that was the motivation with my dissertation. You’d have to understand the way computer systems work. You’d have to understand a whole lot of things to be able to do the type of work that I did, so I thought, “Aha! I’m in a brilliant position. I can make discoveries no one else can possibly make because they don’t know stuff.”
I was making discoveries no one else could make, but no one else cared, either. That wasn’t a very good strategy.
TA: But you kept doing it.
CN: For a while. I stuck with that work for a long time. I thought I was probably among, in those days, the two or three best social scientists who could actually program computers really well because of my background, so I thought I should leverage that and do studies on computers.
TA: Right. Dennis Wixon got started in his work because he was able to code the statistical analyses for psychology students. It’s the same thing.
CN: Exactly. Once you have the skill, if you can leverage that skill then you’ve got a tremendous leg up on everybody else. I realized I could program some pretty sophisticated things.
Around that time, NeXT computers had just come out. They were the first ones that had sound chips on them which produced decent quality sound. You could actually record people talking. I said, “Whoa, if computers could talk; if I could make computers talk… l’ve got to discover something.” How could you not?
It’s a funny story. At that point I had the idea not just of computers talking, but also of computers as social actors and the idea that people would treat computers socially. I had this crackpot idea. I thought it was crazy and that no one would agree with it.
Fortuitously, someone – this is a totally true story – someone was looking for the bathroom in the communication department. She was walking around, and I showed her where it was.
I said, “What brings you here?” She said, “Well, I’m from US West Advanced Technologies, and I saw there was a communication department here, and I wondered if anyone did work on technology.”
I said, “Absolutely, come on up.” She came up and I told her what I was doing. She said, “You should apply for this grant.”
So I did, and when we got the money I said, “This is great, I really appreciate it.” They said, “Well, we have to tell you that we gave out 35 grants. We gave 34 of them to the best projects and decided we would reserve one grant for the weirdest project that was proposed – just one that was so wild and so bizarre and that if it didn’t work, so what.”
TA: I love that. So you got a grant because somebody had to pee and you were weird.
CN: Right, and that someone realized that sometimes you just have to allocate a little bit – not a lot, obviously – for just the most bizarre thing you’ve ever seen.
TA: And you were it.
CN: I was it. I said, “Well, I don’t think my work is that bizarre,” and they said, “No, it’s that bizarre – trust us. But we still think it’s cool. Let’s try it.”
I was able to buy some NeXT computers, and that’s what started it.
I think it’s sad this doesn’t happen more – a company having the guts to say, “Let’s just try one thing that’s insane.” I think it’s very hard to justify it.
TA: You’ve experienced that twice. You’ve experienced that with Stanford and their decision –-
CN: Right, I was very lucky. I fear that those times are gone. I really do fear to a tremendous degree that the risk-taking these people were willing to do for me, to give me an opportunity, are gone. I try to remember that.
TA: I think it would be really cool if you gave a presentation on that to some organization of academic places.
CN: It’s not clear how you’d give it. I agree that it’s an important story, but where do you tell that story? In what venue? It’s a tough story.
I truly do feel sad that the days of just taking a flier on something are gone. The problem is that there’s greater scrutiny. It’s like when you are told you are not accepted for something because of a certain set of criteria, while you’re saying, “Isn’t it worth one gamble?”
TA: Maybe it’s cultural, like the attacks on National Endowment for the Arts and creativity.
CN: I think so. It’s a weird form of conservatism that says nothing good can come from weirdness.
TA: Or thinking creatively.
CN: Not just creatively, but wildly. Not thinking slightly differently, but really taking a big flier. Of course it’s true that most big fliers fail. The more radical the idea, the more like it is to be wrong. But that’s okay.
TA: But the ones that work end up with us walking on the moon.
CN: Exactly. So I think the tricky question is how do you find that sweet spot; what percentage should you put into it? For example, the NSF (National Science Foundation) does these SGER grants (Small Grants for Exploratory Research) where you do have the ability to apply saying, “We’re not going to ask for a big proposal, we want a little thing, and we’re just going to take a flier.”
The critical issue is that the research isn’t peer-reviewed. Peer review is incredibly important and valuable, but it also stifles innovation. The wildest ideas always get negative reviews. Sometimes it’s as well they should, because they’re mostly wrong.
TA: If you ever decide to create a presentation on this idea, I would help make it viral, I’ll tell you that much.
CN: Oh, good. It’s a tough one, and you’re 100% right. I benefited from the willingness of people to say, “We’re just going to roll the dice here.”
TA: Can you give us a brief synopsis of the kind of stuff you found? I have to figure out where I saw this presentation of yours. I don’t remember. But it was about how people would describe or report on the performance of computers depending on their proximity, or whether they were in the same room.
CN: Right. The most famous of those is the politeness study. That one basically had people work with a computer. Then either that computer or the computer across the room would say, “How well did the computer do?” And it still amazes me, but people are polite to the computer.
They say the computer did better when it asked about itself, rather than when a computer across the room asks.
TA: Was this a repeat of a study that was done with humans?
CN: Yes. There have been 100 studies like this – not exactly like this one, but it’s a very, very well known idea that if someone asks you to your face, “What do you think of me?” you’re going to say nicer things directly to them than you would to someone else asking about that person. That’s just Politeness 101. You’d almost be reluctant to do research on it, the results are so obvious.
What we did was to ask whether the same thing would be in effect with computers.
TA: It blew me away. It still blows me away.
CN: Yeah, it blows me away too. I obviously thought it was true, because I wouldn’t have done the study otherwise. But it really was amazing.
The other interesting thing is that it took years to get published. No one would publish it.
CN: No. There was hell to pay. Everyone said it was too weird. No one would believe it. They didn’t think I made up the data or anything – no one accused me of forging data – but no one would believe that study.
TA: Even with the data, they couldn’t believe it was true?
CN: They just figured there must be something else going on. It was frustrating. Even though it was really the first study of its kind, it was years after the subsequent ones came out before that was published. And it wasn’t even published in that good of a journal.
It’s not like when people come up with wild discoveries in physics, like cold fusion, right? In physics they say “Yes, we see your data,” or they say, “Aha, there must be something wrong!” Physics is of course more mature, so they’re better able to make those judgments
The politeness study has been replicated a million times, and now everybody agrees it’s true, but at the time it seemed absolutely nuts.
TA: What year did you finish this?
CN: Around ’92.
TA: Were you aware of the work in usability or user-centered design?
CN: Yes, but you see, my sociological background made me hate that stuff, because whole leitmotif of those discussions was that people need to become more competent; that we have an obligation to teach them.
The idea was that once they know things, once they are better educated, they’ll do better.
TA: Really? It wasn’t this whole idea of intuitiveness?
CN: It was, but in all fairness I should be a little more precise. There were really two trends going on simultaneously.
One was that people wanted designs that fit people better, and that was user-centered design. That tradition was bouncing around with ideas like making things easier to use, simpler, etc.
The user-centered design guys were concerned with how people were responding to technology, but they weren’t really psychological about it. They were coming from a cognitive perspective, saying, “Okay, we’ve got to build this thing better so people can understand it.”
The people who were trying to understand psychological reactions to computers, were all of the view that peoples’ reactions to computers were ridiculous – that they were just based on ignorance.
While user-centered design had its place, there was certainly a very, very large trend in the industry where people were saying, “We really have to teach people that computers are hard. They really are.”
Even user-centered design wasn’t about hiding the complexity. It was about helping people understand things, but it didn’t seek to remove the idea that computers are hard.
There was still a strong theme that if people understood computers better, they wouldn’t behave so foolishly.
People would make certain deductions about computers – that they could “think,” for example. At the time, the belief was that these deductions were really borne of ignorance, and that the best way to solve those problems was simply to remove the ignorance.
TA: So meanwhile you were thinking about these sorts of hardwired sociological responses.
CN: Right. Also, sociologists are trained to smell a rat when people say, “Group X does something, but we group Y’ers don’t do it.”
TA: So are user-centered designers now. We have the idea that self-reporting is flawed.
CN: Right, but at the time people would say, “You know, look at these people being foolish and treating their computers in stupid ways. We would never do something like that. They’re just ignorant, and once we train them not to be ignorant, they’ll do things right.”
Sociologists are trained to be suspicious of that assumption. Whenever you hear the statement, “They do stuff the stupid way and we do it the smart way,” alarm bells should go off.
That’s sociological principle number one. The elites are always telling people that what the elites do is normal and good and what the non-elites do is stupid and bad.
TA: At some point – I don’t know if you changed your mind – but you did start getting more involved in the computer-human interaction community, it looks like.
CN: Yes, that was actually because of Bill Gates.
CN: Yes. Bill Gates contacted me and said, “Do you know that your work is extremely relevant to designing computers?” I said I’d never really thought about it. I wasn’t against that idea, but I was just a scientist. I said, “Look, I’m doing science. I’m not about applied research. In fact, I used to teach all these methods that emphasized that applied research is misguided. All research should come from fundamental questions; it shouldn’t come from daily life.”
I changed my mind, but I was quite rabidly anti-applied research at the time. When Gates contacted me saying my stuff was relevant, I didn’t treat it as negative, but I also didn’t treat it as having anything to do with research. I just thought, “The guy wants help and he’s a pretty famous guy.”
TA: It’s too bad; the correct response could have been, “Can I have some money?”
CN: Well eventually they did give me money so it was okay. So in response to Gates, my colleague and I said “No, we hadn’t really thought about it, but we’re happy to think about it.”
We went out to meet with him and presented the ideas of social interaction and all that. He said, “Ah, this looks good.”
TA: When was this?
CN: It had to be ’93 or ’94.
TA: That’s still pretty early.
CN: Yes, remember Windows 95 wasn’t even out yet.
My book certainly wasn’t out yet either, but there had been two people at Microsoft – the guys who did Microsoft Bob – who were leaning in this direction, and when they heard about the work they said, “Aha! that’s exactly what we’ve been thinking and talking about.” So the timing was good in that sense as well.
We got to work first on the Bob Project, and then across the company. We basically reported to Bill, but any group that wanted to meet with us could.
This was also the time when Microsoft Research was five people. They had no research department to speak of. Part of Bill’s idea at the time was that social science would be extremely difficult to put into a technology company, and that it made more sense to outsource that to people in academia.
Byron Reeves and I were basically the outsourced psychology lab for Microsoft.
TA: Interesting. There was a big connection between Stanford and Apple too, wasn’t there?
CN: There was, but we really didn’t have that strong of a connection with Apple. We had a small link to one of the people there, and people liked their stuff and bought their products, so we were connected in that way; I know a lot of labs had that connection. But to my knowledge, there wasn’t a lot of consulting and research directly on interfaces. The computer science department had historically been weak on interfaces.
TA: You brought up Bob. For those of us who have been in the industry for a while, Bob became a colossal joke because it didn’t do so well. Why do you think that happened?
CN: Part of it was because it was a 1.0 product, and 1.0 products are hard.
Second, it was the culture of the industry at the time. The industry and the journalists who covered it were really interested in making computers cool. Then, when Bob came out and any idiot could use it, and people reacted to it much like they did when the automatic shift transmission appeared.
When that appeared, it was dismissed and hated. The sentiment was that real drivers would never use it, and that you shouldn’t be driving if you don’t really understand how a car works.
Of course the automatic transmission did receive wide acceptance eventually. It became the norm because people like simple things.
There was clearly a very strong push in the press to hate the idea that computers could be totally simple. Then too there was this notion that people need to learn them, and that they should be responsible and take a community college course. There were an enormous number of articles saying, “If you think you can use a computer without doing a community college course, you’re wrong. You have an obligation.”
TA: You think Bob was too early.
CN: It was way too early, that was one big problem. The press murdered it, even though the people who had it actually liked it a lot. The negative reaction wasn’t from the people using it; it’s just that they put up with so much ridicule that it was very hard to stick with the product.
The biggest problem – and we couldn’t solve it fast enough – was that it required 1 MB of memory. At the time that was a killer. It was devastating. I was at the meeting when they figured out they couldn’t do it in 512K. It was the most depressing moment I ever had at Microsoft because it meant that only super high-end machines could run Bob, and that was exactly the wrong market.
TA: Do you think there are any Bob-like projects that are successful today? Or if a Bob-like project was done today, do you think it would work?
CN: There are Bob-like elements. Everybody’s learned a bunch of things from Microsoft Bob. People don’t admit it, but the idea of adaptive menus was first in Microsoft Bob. The idea that you can have all the information in one place is Microsoft Bob. The idea that you can’t just keep on propagating icons was Microsoft Bob.
The notion that you shouldn’t have to know a lot to do things was really all in Microsoft Bob. There are a lot of elements and a lot of learning that comes from Microsoft Bob.
Characters (avatars or animated agents) have not caught on as much as everybody keeps on thinking they will. It always seems like they should catch on. You do have companies like Oddcast which uses characters. They’ve caught on some in the entertainment space, but they’ve never really caught on as a navigation strategy, partly because there has been so little complex navigation on the web that it hasn’t become as big of an issue.
TA: There’s Second Life, but that’s entertainment.
CN: And the entertainment space has certainly done a lot to create acceptance around the use of characters. I would give Microsoft Bob credit for the idea that things that are entertaining can also be functional. And now Second Life is now being used by businesses. I don’t think you could trace a direct link, but I think they are interrelated.
TA: Clippy is a little piece of Bob…
CN: The biggest problem with Clippy was not that the character was bad – which is of course the conclusion that people drew – but that he was badly done. He violated all sorts of social rules.
TA: Like interrupting?
CN: Exactly; interrupting, forgetting the user… You’d be writing a letter for the 50th time and he’d say, “Oh, I see you’re writing a letter” like he’s never met you, which is incredibly infuriating.
If you asked a question he would repeat the exact same answer every time. Clippy really was social, but he violated social rules and norms.
TA: He was the annoying kid down the block.
CN: Exactly. And because they didn’t fix those problems, people drew the conclusion that all character interfaces are bad.
TA: I love that insight. I hadn’t thought about why he was so annoying but socially, anybody who acted like that would just piss you off.
CN: That was the whole problem, and you can easily see how it would be a disaster.
TA: I’ve taken up a huge amount of your time. I could keep talking to you all day because I love the way you think about things, but let me ask you one more question for this interview.
What fascinates you now?
CN: Well, for a long time, there were really two ways of looking at computers and people. One was that there would be a direct relationship between computer technology and the person; that was called HCI. And then there was another area called CMC where the computer was primarily a conduit; a medium.
TA: CMC is computer-mediated communication?
CN: Yes. There was also computer supported collaborative work, which supposedly bridged the two (CMC and HCI). Really it was just half the people doing one, and half the people doing the other.
I think the hottest and most interesting areas for me now are where technologies are not only communication devices, but also interaction partners. We’re starting to see this a lot in these things where you’re walking with your phone and someone near you has similar tastes and interests, and your phone notifies you. There are all these sorts of uses in locating people with stuff like you, or sharing music, or what have you.
On the one hand, the computer is merely a communication device connecting you. That’s computer-mediated communication. But if you think about it, what’s starting to happen is that the computer is also doing some social work itself.
TA: So it’s acting on your behalf.
CN: Exactly, or on its own behalf. It’s being social, it’s really doing stuff. I’m interested in when we’re going to start seeing social or group interactions where both technology and people are players in the group.
It’s a hybrid group notion, that’s how I would put it.
TA: I bet a million dollars that the people working on these technologies aren’t looking at the sociology literature, which would help them skip a thousand steps.
CN: You’re exactly right. There is a misunderstanding that “computer” and “mediated” are radically different; that if it’s a medium it’s a medium and if it’s a direct relation it’s a direct relationship. There’s absolutely no understanding of the fact it can be both.
I think the most exciting domain ahead is the idea that the computer and the user can both be active players.
Imagine all the things you can do. Now I don’t know if these are good or bad, but let’s say you’re in a meeting and someone says, “Oh, shoot, how many units of this have we sold?” and the computer says, “453.” Or you say, “Gee, it would be great if Bob were here right now,” and the computer finds Bob and gets him. The computer is a player.
I’m talking about supporting social interaction, like trying to find people and stuff like that. Or any decision-making where the computer is not just a medium and, not just a social actor, but both. Now you get complex literatures of hybrid teams. There’s a small literature on search and rescue teams including dogs, which is a good example. Human-robot interaction is another domain where you’re going to have these hybrid teams.
TA: I guess also medical stuff.
CN: Medical stuff is another perfect one, where not only is the computer going to zoom in (as in endoscopic surgery) and do other mediating functions, but it will also make suggestions and be a player.
I think this idea of hybrid interactions is the most interesting trend right now, but also the most underappreciated.
The most appreciated area that’s going to be very, very important going forward is the need to worry about emotion, although I think they’re worrying about it wrong.
Everybody and their brother now is obsessed with emotion; rightfully so. There’s an enormous amount of information to be had on the subject, but I would say the approaches are extremely simplistic and naïve, and they’re not going to work very well.
But I’d say that’s the hot one. That’s one people are aware of.
These hybrid interactions people aren’t aware of as much, but it’s going to be hugely important. And you’re exactly right, they’re not looking at the right literatures.
TA: And they are all just there to be had.
CN: Yes; exact same idea as my whole career: Just steal stuff; it’s all there. It’s not hidden. It’s not secret. It’s really all there and it’s just a matter of stealing it appropriately.
TA: I am going to finish off this interview with a personal note, which is that anyone who knows me personally knows that I have a tremendous intolerance for boring or pointless conference presentations. I just can’t stand them. They make me postal.
I heard Cliff speak, it must have been seven or eight years ago, and I still talk about it today.
If any of you ever have an opportunity to hear him talk, it’s worth the price of the flight and the conference no matter what it is. I would highly encourage you to do that.
CN: Thank you very much.
TA: Thank you for all your time.