It rained most of Labor Day in Kansas City, which meant my partner, my two kids, and I spent most of the day indoors at home. It turned out ok. I read various children's books, played various children's games, and made, according to my three year old, the coolest pillow fort ever. But the best part about yesterday was an amazing stretch of about three hours where my two boys remarkably occupied their own time. It was time I had hoped for in order to finish my RiAH blog post for today, but instead I found myself in the mood to keep playing. So I ended up doing something I had been wanting to do for a while, build a Twitterbot.TODAY CAN BE THE SERMON.— Preacher_Bot (@Preacher_Bot) September 7, 2015
For those of you who don't know, a Twitterbot is a simple piece of computer programming that posts content to Twitter automatically. If you're any kind of regular Twitter user, you've encountered twitterbots before. The most common are spambots that favorite, retweet, or reply to a tweet of yours so you see the product they're advertising when you look at their profile. In the last several years, however, a growing number of artists and hacktivists have harnessed the power of twitterbots to make various cultural and political statements. On the one hand there are completely nonsensical bots like @TwoHeadlines that randomly smashes together the titles of two news articles to make often hilarious results. On the other hand there are a number quite serious bots like @CongressEdits, which sends out notifications when someone makes an anonymous edit to Wikipedia using an IP address from within the US Capitol. The bot includes a link to the edited Wikipedia entry in the tweet so the world can see what kind of scrubbing, altering, deleting or updating some congressional staffer was doing that they didn't want known. In between these poles are a variety of poignant or playful twitterbots that algorithmically comment upon our times. Readers of this blog might appreciate the @Every3Minutes bot that sends out links to books on slavery every 3.6 minutes, which was the average interval between slave sales in the antebellum South. My most recent favorite, however, is the @TheHigherDead, a bot built by Chuck Rybak and bot-extraordinaire Mark Sample that transforms the platitudes of university administrators into the ramblings of bureaucratic zombies out to eat faculty brains.
Ass. Dean says: fluid degree dynamicifies outcomes gateslingo compentencyyyyy buy my boooook— The Higher Dead (@TheHigherDead) July 13, 2015
I first became interested in Twitterbots in part because, well, they're awesome. But my interest continued to grow as I realized they might be able to help me poke at a problem. Many scholars argue that Protestant evangelicalism is, at its core, discursive. Less a theological disposition or social experience, evangelicalism and fundamentalism are, as Susan Friend Harding argues in The Book of Jerry Falwell, a language whose vocabulary emanates from leaders on high. This way of speaking is powerful, Harding argues, because it implicates even the lowliest speaker in the movement as they narrate their lives with idioms and words created by prominent preachers and other tastemakers. This is, of course, on one level true. Evangelicalism is a construction born of myriad social and cultural discourses. But to reduce religion to a language that even you might start speaking if you're not careful seems far too abstract and removed from the messy realities of people's lives. Would this "born again language," as Harding calls it, still have its power when it was spoken by someone who only copies, and not creates, the lexicon? Would the cacophony of evangelical leaders even make any sense en masse?
To play with these questions I decided to build a twitterbot that would interrogate the contours of American evangelical discourse. I built the bot using data journalist Tom Meagher's code, which generates tweets at random intervals using only the words and phrases of other active twitter users. The bot, in short, follows other people, takes what they tweet, and then remixes it to see what kind of insightful non sequiturs emerge from the mashing. For my source, I chose the top five most influential pastors on twitter as identified by Christianity Today. I wanted to include even more but these guys--and they're all guys--are such voluminous tweeters that the weight of their discourse initially kept crashing the bot. So I settled on Joel Osteen, Rick Warren, John Piper, Max Lucado, and John Maxwell, plugged in their data, and set the program free.
Allow me to introduce you to @Preacher_Bot.
The bot has been running only a mere nine hours as of this posting, but it's already yielding some interesting material. Of course the program frequently produces absolute gibberish like "Speak calmly, listen carefully. Enjoying a Christmas lunch with my 92yo dad". But in its persistence, repetition, and reliance upon a common source, the bot's growing stream of tweets is starting to highlight themes in the language of these megapastors. There is an emphasis upon being kind or finding happiness, for example, even when this prescription is laughably articulated as "Successful people ask “What is it so difficult to make room for the man replied “Lord, let ME FIRST." An equally as prevalent trend is the bot's need to push some kind of product, like "Max Lucado's brand new building of westside campus!" and "Available online too. What needs be capable of doing."Going to be happy.— Preacher_Bot (@Preacher_Bot) September 7, 2015
What does this all mean? It's far too early to tell. The bot needs more time to produce a sample size that could be reliably analyzed for common themes. Plus, I'm realizing the punctuation of Bible verse citations are messing with the bot's ability to construct full sentences, which means the bot in the end might say a lot of nothing at all. But I think @Preacher_Bot is showing the promise of algorithmically exploring--indeed, playing with--religious language. Just think of the other kinds of religio-bots we could unleash: @Rabbi_Bot, @Prosperity_Bot, @None_Bot (or @Unaffiliated_Bot if you're the Pew foundation). And if you're interested in seeing where the @Preacher_Bot experiment leads, just follow him on Twitter. I'll be sure to link out to updates on the project from that stream. Plus, he could use some followers. So far it seems like the only people following him are other preacher bots--programs built by believers who are similarly spreading good news and offering great deals.
You’ve got to do with materialism!— Preacher_Bot (@Preacher_Bot) September 8, 2015