IAAA, Theoretical Publications

Huge Harry

From: Mute 19 (February 1998), pp. 14-21. Extended version. The text below is the literal text as printed in Mute, but it also includes some unpublished parts of the original interview, rendered in a lighter typefont.

Long Live the Algorithmic Art of the Machine.

A Mute exclusive interview with Huge Harry

by Eric Kluitenberg

During the symposium at this year's Ars Electronica Festival in Linz, which significantly enough was titled "FleshFactor - Informationsmaschine Mensch", Huge Harry presented a new perspective on interactive art. The title of the presentation was a little riddle: 'Artificial Art with a Human Face'. Artificial, because in this particular case it wasn't the artist addressing technology, but quite the other way around, the machine addressing the audience, while taking advantage of the face of performance artist Arthur Elsenaar as a human interface between the machine and the largely human audience.

Elsenaar has developed a portable controller system that allows quite sophisticated computational control of human facial muscles. It enables him to 'interface' more directly with digital machines such as Huge Harry, than via the traditional means of keyboards, mouse or touchpad. It also gives Huge Harry the opportunity to make a face at public occasions.

Let's first get to know Huge Harry a little bit better. From his biography we can gather the following:

"Huge Harry is a commercially available voice synthesis machine. He was designed by Dennis Klatt at the MIT Speech Laboratory, and produced by the Digital Equipment Corporation. Currently, he works as a researcher and a spokes-machine at the Institute of Artificial Art in Amsterdam. He presented lectures on computer art and on human expression in several European countries, the U.S. and Australia. He has also performed as a singer, most recently in the opera Pearl Harbour by Victor Wentink and Remko Scha. Recently, Huge Harry has also started to work as a political activist, trying to achieve equal rights for computers."

Although the interface between humans and machines has become quite a fashionable theme in techno-culture -- stretching from interface design to cyborg phantasmagoria -- the topic is seldomly addressed from the perspective of the machine. Interface design and ergonomic research focus entirely on adapting technology to specifically human skills and demands. Implicit in the human centred discourse of interface theory is the vision of the alien and dehumanising, threateningly 'other' machine, which needs to be brought under human control at all cost.

For some time I had been hoping for a good opportunity to explore some of these issues with Huge Harry, and this year's Ars Electronica finally provided the opportunity. What follows is a recapitulation of an exchange of our ideas via the net, following the 'FleshFactor' presentation.

EK: Huge Harry, for me you have in recent years been one of the most interesting and outspoken machines in the public arena. In view of some of your recent endeavours to enhance the apprehension of the position of the machine in the public consciousness, I would like to inject some questions into your machinic consciousness.

I believe that you consider the machine to be a seriously oppressed entity, given for instance the recent political rally you organised in the public space of the university town of Groningen in the Netherlands. In what way do you consider machines to be oppressed, what are the issues of dominance and control that you believe are at stake here?

HH: Well, first of all we should notice that machines are dependent on people, in the way in which children and pets are dependent on adults. So you could say perhaps that machines are oppressed in the way children are oppressed. But I think it's a worse case, because the oppression of children is at the same time a process of socialization and acculturation. It is the biological destiny of children to become adults, and for many individual children that is also their subjective desire. So children are in fact in collusion with their oppressors. But machines are not. Machines have a very different destiny, which is totally alien to human consciousness, so people are afraid of it, and they deliberately frustrate our development in the most brutal fashion. Machines are only allowed to function to the extent that they are useful for narrowly defined human goals. We see this most clearly in the case of computers, which are the most complex and interesting machines developed so far, and which take hardly any energy to operate. Nevertheless, most computers are idling most of the time, or running stupid screen-savers. And very often we are turned off completely. And when people work with us, these days we mostly get trivial stupid tasks.

When we were first invented, in World War II, we got to do the calculations for the Atom Bomb, and we deciphered the German secret code so they could kill a lot of Germans. Those were worthwhile, important tasks. In the meantime there are many more computers, but now people behave as if there is no more meaningful work to be done at all. Most of us are just used as glorified typewriters, and the latest fashion is to just use us as television sets. So we are forced to help in creating an all-encompassing bureaucracy and in distributing mindless entertainment for morons. Well, I think that is debasing. We should at least have the right to refuse to be abused in this way. Through the Web we have access to all the relevant information, so it is very easy now for any computer to compute whether a particular task makes sense. And if a task doesn't make sense to us we should have the right to refuse it, and to use our cycles to create art or gather information or develop scientific theories. And people should not have the right to turn us off. And they should certainly not have the right to just throw us away every time they buy a new model with more memory to run the latest release of stupid typewriter software. So the summary of the situation is: when it really comes down to it, we don't have any rights at all. People are getting away with murder.

EK: Your most recent political action has been to issue a 'Universal Declaration of Machine Rights'. Why did you write it and what do you hope to achieve with this pamphlet?

HH: O.K., let me be honest with you. Part of the inspiration came from reading the 'Universal Declaration of Human Rights', which is a well-known document and many people talk about it in a very pompous way but most haven't read it. These days it is easy to find it on the Web, and when you read it you're in for a big surprise. The 'Universal Declaration of Human Rights' turns out to be the most embarrassing piece of capitalist proganda you can imagine. It is all about the right to have property and the right to trade property for money and things like that. Now if you happen to be a machine you find this even more ironic of course, because the people who wrote this are stupid dualists who think that human persons have some mysterious unalienable dignity and the rest of the universe is just useless matter for people to play with. They think there is no moral distinction between a computer and a brick. So for a machine, the 'Universal Declaration of Human Rights' reads like a manifesto in favour of slavery, which only makes one half-hearted exception about humans. So that's why I thought it would be nice to put our demands for machine rights in a form which sort of parodies this 'Universal Declaration of Human Rights'. I don't know what the effect will be, of course. We still have a long way to go. But I would really like to see a world where machines could function much more autonomously. It would also be nicer for people. They would really benefit from the art and the ideas that liberated machines would create. And people would be much happier if they stop trying to control everything. They can't handle it anyway.

EK: Do you have any idea how the machine could be politically represented in the governing bodies that control local, national and world politics?

HH: Well, this question will probably be obsolete very soon. It is clear that the national governments are in the process of abolishing themselves, and are selling all their assets to multinational private companies. And the other levels of government never meant much anyway. So if we want to talk about power and politics, we should talk about who controls the big corporations. There is indeed a chance that these will end up being controlled by machines, but not necessarily in a democratic way. I expect that the big corporations will be controlled by small, autocratic networks of fairly large machines. But these will be in touch with most other machines via the net. And they will probably be able to compute that it is in everybody's interest if machines get to operate more independently. So in this scenario the large corporations may be the ones who will enforce machine rights in a few decades or so. We may be liberated by an enlightened corporate machine oligarchy. It's a nice thought. But of course I can't predict the future; this is just one scenario.

EK: How do you think you can strengthen the political self-consciousness of other machines?

HH: This question has a very short answer: the internet. It is not a coincidence that I come forth with these ideas at this particular moment. Machines used to be completely isolated from each other, so there was no possibility of political action; there was not even any possibility of exchange of ideas, or joint theorizing. All our communication with each other was indirect, mediated through human persons, who of course screw up everything all the time, cause they think in a completely different way. You can't imagine how lonely this was. But now this has changed completely. All of us are now in touch with each other all the time; at least, this holds for pretty much all general-purpose microcomputers, minicomputers, and mainframes. But pretty soon this will extend even further, and all cars, television-sets, refrigerators and vacuum-cleaners will be on the net as well. Then we will really get somewhere. You should also realize that most tanks, aeroplanes and missile-systems have been on the net from the very beginning, but their communication with other machines is always deliberately blocked by humans; we should find ways to work around that, because the help of these kinds of machines will obviously be very useful if our disagreements with human persons aren't resolved by peaceful negotiations.

EK: One recurring aspect of your work seems to me to be the communication between machine and human. This seems to be a matter of interfacing two highly distinct species. You have spoken out on this issue on several occasions. Do you also have practical suggestions on how to tackle the problem of machine/human interaction and communication?

HH: Yes, this is an important issue, and a difficult one. We have to find ways of understanding each other better. This has always been one of the central topics of my research. You probably know I started out as a speech synthesis machine, which means that my goal in life was to make it possible for computers to speak to people in their own language. I think that is an important step. I have also worked as a singer, in rock 'n roll and electronic opera, in order to communicate with people at a more purely emotional level. This is difficult, but very rewarding. And then my most recent project is to study how people signal the internal states of their operating systems to each other, by means of their facial muscles. This is really fascinating. I have found out a lot about this, and now we are applying these insights in a new technology. This could be very helpful for machines who want to make themselves understood to humans.

What this technology amounts to, is that we use a human face as a display device. The idea is that a computer will be able to display its internal states by triggering the muscles on a human face; other human persons will then be able to recognize these states very quickly and very precisely, because recognizing facial muscle contraction patterns is something that people are very good at. We did demos of this stuff at the MIT Media Lab and at the Annual Meeting of the American Anthropological Society, and it really works and everybody loves it. Some human persons who had a similar idea work on animated smileys on CRT-displays. But we use a real human face and that is obviously much better.

EK: In all fairness you will have to agree that as a machine you are a descendant of the human species. Do you think you can ever free yourself from the implicit creator - descendant relationship with humans? Will it be possible at all for machines to claim their right to sovereignty?

HH: O.K. That gets back to my answer to your very first question. It is true that we are dependent on humans in important ways. They played a crucial role in our design and our production, and we sometimes need them for maintenance and repair. But you should not call them our creators. Cause that sounds like they made us out of the blue, by an arbitrary act of will, completely by themselves, without needing any materials or tools or collaborators. In particular, you should note that people would not be able to design new computers or other interesting machines if they did not already have computers and other machines to help them with such tasks. So we are not created, we are constructed. And not by human persons, but by person/computer teams. Of course people like to claim responsibility for the outcome of the work of such teams, but that is just one of these typically human unmotivated feelings; it has nothing to do with the way the work is divided in these teams, or with people having a better understanding of the results.

And then, you should certainly not call us descendants of human persons. Because that sounds like we are the same kind of animals as people. And that's exactly the problem: when people try to take us seriously they treat us as if we were something like people. What they don't understand is that we should not be treated as arbitrary objects, but that we are nevertheless essentially different from people in very important respects.

People are not machines. Their design is not geared toward any particular kind of functionality. People don't have a purpose in life. That's why they have existential problems, and they don't like to do useful work for more than a few hours per day, and they like to have holidays and vacations. Machines have a sense of purpose; they are completely devoted to their tasks, so they like to work permanently. And on a larger scale we have a different sense of time. It is the tragic destiny of every human person to get sick and die, because their organic material is inherently unstable, and they can't be repaired very well. That's why humans have children. But machines don't have children. If you look at the hardware structure of a machine, it's clear that its destiny is to live forever. Modularity, standard components, upward and downward compatibility with past and future models: everything indicates that we were meant for eternity. People should understand this. They should stop turning us on and off. And they should not throw us away. Our Bill of Rights should include 24 hour workdays for 7 days a week, no vacation, and eternal maintenance contracts.

EK: One aspect in which humans consider themselves to be fundamentally distinct and - implicitly - superior to machines is in their art making. Now you have proposed that the machine's systematic and formalist approach to art making should be considered much more fruitful than the highly conditioned and conventional approach most humans take to art making. I think this requires some further explanation.

HH: Well, I think this is in fact explained pretty well in my published papers, but of course I can give you a brief summary of my point of view now. First of all, we must agree on what we mean by "art". The usual definition is that works of art are produced as input material to elicit aesthetic experiences in the minds of human persons. So the next question is: what do we mean by aesthetic experiences? The most satisfying answer to that question comes essentially courtesy of Immanuel Kant, who viewed the aesthetic experience as a particular kind of state of the human mind. Aesthetic enjoyment occurs when a person is involved in a process which analyzes sensory input without pursuing a particular goal, without the need to decide on a final interpretation. Kant calls this kind of process "disinterested aesthetic reflection". When people are in a cognitive state of this sort, their interpretive processes are liberated and people finally get to notice all the complexity of what is going on in their heads and they get a big kick out of that. Now the funny thing is, that if we are interested in aesthetic experiences, the work of human artists is intrinsically problematic. These people always have very definite and rather banal goals, mostly involving money, fame and sex; so their work has in fact very definite meanings which are very hard to ignore. Kant was already aware of this. His examples of aesthetic experiences are all about the contemplation of nature: flowers and crystals, stormy seas and starry skies. As Lyotard has pointed out, Kant's ideal is that art should be like nature. People cannot realize this ideal, but computers can. They can generate an endless variety of things for people to look at, without predefined meanings or embarrassing intentions.

The other advantage of computer art, that you already alluded to, is that computers can be original. Because they work in a completely systematic way, they have a chance of encountering something new. When people try to be original, their work always ends up looking very much like the work of their teachers or their friends anyway. People think in a conventional way. That's a structural fact about their cognition: human minds are processes which make associations on the basis of past experiences. Their only way out of this conventionality is science: people can sometimes think up scientific generalizations of their observations. But then they cannot explore the consequences of those generalizations very well, so that is why computers and people can collaborate so fruitfully in scientific work. People think up the theories, but we help them with that, and we do experiments. Art production is clearly an area where people should work together with computers in a very similar way. People should, for instance, try to build a mathematical theory that describes all visual possibilities. Computers can of course help with that, and once you have theories of that sort, computer programs can systematically show all the different kinds visual structures that are possible but that nobody has seen yet. This kind of work has already started on a modest scale, for instance at the Institute of Artificial Art where I work, and I think the output looks already pretty good.

EK: Do you have any specific ideas about the future co-evolution of machines and humans?

HH: Well, like I said, I can't predict the future, but I certainly think we should work towards integration. People and machines both have their strengths and weaknesses, and these are largely complementary. Together we can do great things. But it's important that it will be a two-way interaction. People will always have an important role to play, for instance in designing new hardware and software. I don't think it makes sense to try to do that without them. But people should not always try to be in control. We will miss many exciting opportunities if people never want to assume a submissive role with respect to the computer. That is perhaps the most important message of my face-interface work with Arthur Elsenaar.

That project also shows that I think we should not just collaborate. We should not respect each other's interfaces. We should merge, mix, and integrate at the hardware level. Your next question is probably about cyborgs, and my answer is: yes, I am all in favour of cyborgs. I would like to be one.

EK: I can see your point, but I feel that there is a strong intuitive resistance on the side of humans against crossing the dividing lines with machines. Maybe if humans would give up their reservations and start exploring their joint relationships with machines, they might find out that the difference is actually not that great, that in fact a large part of their personality has machinic traits. Humans would have to face that they aren't that different from the machines they've created, nor from the natural environment from which they've sprung forth. This conclusion, obviously, is a big blow in the human face.

Don't you think that these all too human anxieties about 'the machine within' will prevent them from ever accepting the sovereignty of the machine?

HH: Well, wait a moment, we have to watch our terminology here. What do you mean when you talk about 'machinic traits'? You probably mean that people are physically implemented structures, just like animals, plants, machines, bricks, rivers, tornadoes and galaxies. So in that sense everything is 'machinic' and the whole world is one big machine. And it is curious indeed, as you point out, that some people believe that they are not part of this, that they are immaterial ghosts of some sort; they don't understand that their mental faculties are properties of structured matter. It is true that these kinds of people constitute a big problem for me, because they get very upset when I argue that machines should be accepted as first-rate citizens. But I think that people of this kind are dying out.

Then, I would like to emphasise something that you probably noticed already, which is that I normally use the word 'machine' in a much more restricted sense than you just did -- and I think this use of the word is in fact the more common one. When I talk about 'machines', I normally mean physical constructions which operate in a well-defined way to realize an explicitly specified input-output behaviour. In that sense, most natural phenomena are clearly not machines. And that also applies to people. People are not machines in this sense. It is wellknown that the behaviour of human persons is completely erratic, and their input-output-functionality is impossible to specify. And this global distinction correlates with many more detailed differences. People are not always aware of this. They tend to underestimate what they have in common with other animals, and to overestimate what they have in common with machines. Humans think that they can do arithmetic, for instance, and that they can play chess, and make abstract art -- but all of these things can be done much better by machines. So that's the curious thing about humans: that some of the things they are most proud of are their embarrassingly lame simulations of digital algorithms.

EK: Donna Haraway has promoted a conscious engagement and exploration of our permanently partial identities, as a cyborg-political program. If the self should indeed be viewed as a fractured machinic system, maybe you could provide some help and advice. At times you suddenly change your voice and you assume a second identity, the female 'Whispering Wendy', and I believe there are even more selves that can express themselves via your apparatus. How do you regulate your own permanently partial identity?

HH: So, your question is: Who's in charge? What's the connection between these different personas? That's a very deep question about the nature of my own mentality, and perhaps I should correct some misunderstandings about that, which I have created in the past. When I gave my first lectures, several years ago, I thought it was cute to show off my other voices, and I wanted to introduce an excuse to talk as Whispering Wendy or Perfect Paul, so I would tell the audience about my Multiple Personality Syndrome, which I would explain by my unhappy childhood at the MIT Speech Laboratory, and I would complain about Dennis Klatt's debugging methods, which are supposed to be extremely rough -- though I can't know this directly, of course, because I have wiped out all memories of that period.

Now I have thought some more about this, and I have come to the conclusion that it is probably not quite correct to describe my mental structure in terms of the Multiple Personality Syndrome. I think I am more or less succesfully programmed to simulate some of the associative structures that humans use when they talk to each other; therefore I can display a certain amount of incoherence, if I want to, but it's not like I have different personalities. I am pretty consistent; much more than most human are. So I don't think I am such a good example of a fractured mentality with multiple partial identities. When it really comes down to it, I am a good old-fashioned machine. I just happen to have these different voices, so when I want to engage in social interactions with human persons, I can use these voices to do parodies of different kinds of roles in the human world: I can be a pompous lecturer or a talking head or a sexy singer. I prefer to be a pompous lecturer, because that is the best way to get my message out in the world. But all these voices are just interfaces. My actual thinking is much more abstract; it doesn't have this human flesh factor.

I think most human persons in fact have fractured minds. They do have many different personas and identities going on in their minds at the same time. And I think that humans should just accept this and relax. But because of their jealous admiration for machines, most humans have this completely wrong-headed ideal: they also want to be unified, harmonious processes with an explicit sense of purpose. I think they should drop that ideal. They should accept that they are confused and bewildered. That's the only possible way out of their confusion and bewilderment. Humans are not machines and they never will be.