For my dad's birthday last month, I gave him a copy of the new book The Most Human Human: What Talking With Computers Teaches Us About What It Means to Be Alive, by Brian Christian. When Adam Gopnik, in The New Yorker, called the book "terrific" and "one of the rare successful literary offspring of Gödel Escher Bach," I knew that this would be a perfect gift for my computer-geek, philosophically inclined Dad. Upon further research, I liked how the author, Brian Christian, has degrees in computer science, philosophy and poetry, and was born in 1984. He's an overachieving Millennial, in short, and it makes me proud that my generation has reached the point where we are writing erudite books that get rave reviews in The New Yorker and can be given to our baby-boomer parents.
So I purchased the book, read the introduction for kicks, and was sufficiently intrigued that I ended up reading the whole thing before giving it to Dad (I knew he wouldn't mind).
As the New Yorker comparison to Gödel Escher Bach would imply, The Most Human Human is a difficult book to sum up and covers a wide range of topics. Fundamentally, it is an investigation into what intellectual processes humans can still do better than computers, and how that can help us to understand our place in the world and get the most out of being alive. If our humanity does not lie purely in our intellectual capacity, where does it lie?
I should note, for those of you who are put off (rather than encouraged) by the comparison to Gödel Escher Bach, that The Most Human Human is much shorter than Gödel Escher Bach and probably more accessible -- it will make you think, but does not require you to understand symbolic logic. In addition, The Most Human Human has a moral-philosophical-humanistic component that I don't remember being present in Gödel Escher Bach. (Some indication of Brian Christian's tone can be discerned from the fact that one of the book's epigraphs comes from David Foster Wallace and Wallace is quoted several times in the text. How Millennial of Christian!) It's a warmer and more inviting book, I think.
The Most Human Human covers a lot of ground, so people with a wide variety of interests and concerns are likely to get something out of it. Clearly, one of the big themes of the book is human verbal and non-verbal communication -- and clearly, that's something I also think about a lot, because I write plays. Christian even cites the work of playwrights in his text -- discussing David Mamet, for instance, in a section on how human dialogue tends to be far more circuitous and discursive and filled with interruptions than computers are capable of. I love the fusion of art and science!
The Most Human Human, therefore, has given me some ideas on how to be a better playwright -- not just a better human being. In one fascinating section, Christian explains that chatbots these days excel at "stateless conversation," that is, conversation where their response depends only on the last thing that you said. But they're not so good at taking into account the overall arc of the conversation, and even worse at taking into account the conversation they had with you yesterday.
One of the earliest chatbots (late 1980s) was MGonz, which was designed to be verbally abusive, belligerent, and argumentative -- and succeeded in fooling a lot of people into thinking it was a human being.
"As becomes painfully clear from reading the MGonz transcripts, argument is stateless," Christian notes. "I've seen it happen between friends: 'Once again, you've neglected to do what you've promised.' 'Oh, there you go right in with that tone of yours!' 'Great, let's just dodge the issue and talk about my tone instead! You're so defensive!' 'You're the one being defensive! This is just like the time you x!' 'For the millionth time, I did not even remotely x! You're the one who...' And on and on. A close reading of this dialogue, with MGonz in mind, turns up something interesting, and very telling: each remark after the first is only about the previous remark. The friends' conversation has become stateless, unanchored from all context. [... Thus] there's a sense in which verbal abuse is simply less complex than other forms of conversation."
When I read Christian's invented sample of "stateless argument," I realized that if I read this same dialogue in a play, I would consider it terrible playwriting. I know this sort of thing happens in real life, but it seems to happen even more often in the work of bad or inexperienced playwrights. Being taught that "drama is conflict," some newbie playwrights confuse endless argument or bickering with genuine dramatic conflict. Their plays tend to consist of two relatively generic characters quibbling with and criticizing and insulting one another in generic -- in stateless -- ways.
But if you know exactly who these characters are, their relationship to one another and their prior history, the deeper reasons why they are arguing, you can write a good dialogue scene that doesn't degenerate into mindless bickering.
Good playwriting (and good chatbots) take prior conversation and prior history into account, the way that human beings do, unconsciously, in real life; bad playwriting (and bad chatbots) is stateless.
And that's just one of the many things that The Most Human Human made me think about.
No comments:
Post a Comment