Culturebot’s Mashinka Firunts sat down with director Annie Dorsen to discuss her new work Hello Hi There.

In the tradition of certain classics of the theatrical canon, Obie-award winning director Annie Dorsen’s Hello Hi There centers on a discussion unfolding live onstage between two conversational partners. In a departure from its predecessors, Hello Hi There casts two chatbots – computers outfitted with natural language processing – as its interlocutors. The topic of discussion is a televised 1971 debate between twentieth-century theoreticians Michel Foucault and Noam Chomsky on human creativity and justice. Drawing from the transcript of this debate and other sources input into their textual databases, the chatbots algorithmically generate an improvised, singular discussion for each instance of the production’s staging. Hello Hi There made its US debut in Performance Space 122’s COIL Festival and continues its run at PS122 through January 22nd.

The raw text for the chatbots’ conversations is drawn from theoretically dense, semi-scholarly material (the 1971 Chomsky-Foucault Debate). Has this posed any challenges in terms of audience response?

Actually the main challenge as regards the debate has been the extraordinary interest and curiosity of the audience towards it. The video has not been available in its entirety since 1971, the original Dutch broadcast, but a published transcript has been widely read. People know of the debate, but no one’s seen it. In Hello Hi There the debate is a visual and thematic referent, it’s not the event itself, and I knew there would be frustration for some that the audience doesn’t really get to watch it, to hear what Chomsky and Foucault are saying. In other venues we showed the debate before and after the performance in the lobby, but during COIL that sort of thing isn’t possible.

During the piece, depending on the night, the audience will get a more or less thorough idea of the content of the debate – some audiences get quite a lot of information about it, many quotations and summaries and so on. Other audiences get very little. It goes with the territory of a piece like this, where the text is determined by a combination of chance and probability.

Does the piece assume a specialized audience? What kind of engagement with these concepts is the work asking for on the part of the viewer?

I think the piece asks for an audience that wants to think about thinking. The questions of the piece (what is a human, what’s the difference between language and thought, what constitutes a “useful” or “useless” idea, etc) are pretty interesting to us humans. We tend to be curious about ourselves. Obviously there are some people who really aren’t interested in non-narrative or conceptual work – for them this piece will be a challenge, or less euphemistically, a bore.

To what extent would a familiarity with its concepts – Foucault’s theory of knowledge, Chomsky’s linguistics, posthuman studies – alter the experience of viewing the piece?

If people know the debate they will have some extra points of interest, for sure. Maybe they’ll have a better understanding of the choices I made about what to include in the chatbot database, they’ll get a few jokes that others might not, and of course, they might feel even more cheated that they don’t get to watch the debate, as they’ll only have read the published transcript.

A Chomskian (Chomskyean? Chomsky-ish?) linguist saw the piece in Oslo and seemed to have a particularly fantastic time – he got all the word games the bots play, all the oblique references to aspects of Chomsky’s theories of grammar and so on.

But I think there are many many specialized points of entry like that. If people have a particular interest in Foucault they will have a similarly specific way of understanding the piece, if they have a background in general Western philosophy likewise, or an interest in programming or artificial intelligence, or even a background in performance theory. Anyone who has read Peggy Phelan or Philip Auslander will get something from it that others might not.

How, if at all, has response to the piece differed between its European and US runs?

This is a tough one, I was speaking about it with production designer Kate Howard last night. Of course the biggest difference is an audience of native English speakers vs non-native speakers – US audiences so far have (I think) enjoyed the word-play a bit more, because they recognize the wrong-ness, grammatical mistakes, and misunderstandings immediately, whereas some non-native speakers might not immediately know if the bots made a mistake or if they just missed something.

Then I think the practice of “conceptual theatre” is more familiar to European audiences. The kind of procedural dramaturgy I’m working with is pretty common in the independent scene there, so audiences there might have a bit more experience reading this kind of piece. But the question about differences between European and US audiences is like an endless speculative nonsense, in a way…too complicated and unknown to really answer. I would say, very generally, that European audiences are a bit more patient with what you might call “difficult work.” But I would really hesitate to say more than that.

In your program, you quote a text by performance scholar Philip Auslander where he writes that a chatbot’s ability to create novel linguistic output “undermines the idea that live performance is a specifically human activity.” This summarizes a broader anxiety within the field, that what has historically been theater’s primary component – the presence of live, performing bodies – is giving way to a focus on mediatized technologically-enabled work. How do you locate your practice within these shifting conceptions of the genre?

Before I made this piece I wasn’t thinking about it very much at all. But over the last two years since I started working on this, I’ve become obsessed with this question. And now I want to make it as complicated and difficult as possible. I like to make problems about concepts that people assume they understand – natural vs artificial, live vs non-live, theatre vs installation. Is Hello Hi There a play? Am I the author of it? Is the text written? Is it acted? I actually don’t know the answers to any of those questions, but I like it that way.

You’ve mentioned in a previous interview that you encountered difficulty securing funding for this piece because of the absence of (human) “actors.” What do the chatbots constitute in your view?

Well, the chatbots are nothing, I mean I know how they work in intimate detail and I can tell you that they don’t exist as beings (even though I love them very much). But I believe strongly that Hello Hi There is theatre. It is live, it is given meaning by a public, it operates entirely according to the principles of theatre. And anyway, I would say that any piece that questions the rules or assumptions about a given genre rightly belongs to that genre, no exceptions.

How does the experience of directing a chatbot differ from that of directing a human actor?

The best thing about directing chatbots is that they can’t be directed. They will do what they will do, I can’t control them. I can’t negotiate about how to say something, or what kind of pause to take, or about any of the rhythmic or stylistic questions that normally come up in rehearsal with actors. I love that the bots don’t make meaning – if there is meaning, it’s because the audience makes it. And they never adjust their performance to get a certain response from the public.

Hello Hi There also calls into question the issue of authorship. It’s conceptualized by you and created in collaboration with a team of systems and software designers. The text/textual database is appropriated from a variety of sources, and the actual conversation seen by the viewer is generated live by a computer algorithm. How would you pinpoint the creative activity of this production and who is performing it?

Chomsky speaks and writes about the “creativity” of the speaker – it is how he relates his study of linguistics to his political activism, that all human beings have this ability to perform the fantastically complex and creative function we call talking, so therefore all humans have as a biological birthright a kind of potential for creative work and life. He really insists on this term “creativity.” He defines it more precisely as “making more output than one takes in as input” – that our language output is potentially infinite, although what we hear others say is of course finite, we only hear a certain (theoretically countable) number of statements during our lifetimes.

Whereas I am more interested in the creativity of the audience, of the listener. That is the theatrical proposal of the piece – that we human listeners make meaning continuously, it is creative work that we are not necessarily in control of. Our brains make patterns, in many cases without our conscious choice or even consent. In this piece we are confronted with computer-generated text, it is in a sense meaningless, there is no consciousness or intention behind what the bots say, there is no consequence to their statements. But yet there is an effect, we the listeners are affected by their speech. And that is where the creative work is happening, it happens in the audience.