Autoflâneur is a procedurally-generated guide to getting lost which lives at https://twitter.com/autoflaneur. It was inspired by the traditions of flânerie and psychogeography. The original idea was developed by me with Ed Key and David Ralf at Games and Play for Big Outdoor Days, a game jam supported by Hat Fair, Theatre Royal Winchester, and the Without Walls street arts consortium, and produced by Hannah Nicklin.
The source code is available and welcomes comments at Cheap Bots Done Quick.
Many thanks to my Patreon supporters, who enable me to make strange things and release them into the world.
This work is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License.
To use autoflâneur, go to an urban environment, and then, beginning at the top of its Twitter feed, follow the instructions down the page for as long as you need to. It produces a new instruction every hour, so you will have a different guide every time you visit.
You should take a notepad and a writing implement, and you might need money for a few bus fares or incidental purchases.
You can pay attention or you can drift, as you need.
Autoflâneur knows nothing about you and nothing about where you are walking. She does not know your limits. She may ask you to go to places or do things you are not comfortable with, but only by accident, and she does not ask you to interact with other people. She will occasionally instruct you to check in with yourself, but might ask you to do the wrong thing, because safety is complicated. She will not be offended if you disobey an instruction, and might encourage you to, and she will be happy if you stop when you need to. She and I recommend that you both take yourself to unexpected places and look after your own limits.
Autoflâneur uses a range of verbs for a range of different bodies. All people will encounter an instruction that doesn’t fit them and their body. If she asks you to do something that is impossible for you to do, she would be delighted it if you interpreted through your own body, because discovery, disobedience and improvisation are the heart of good flânerie.
Autoflâneur and I welcome all thoughts on how a wider diversity of folk can enjoy her, and widen what is meant by being a flâneur.
This is unlike other bots I’ve made, in that the instructional poem is intended to be read as a long series of tweets, rather than each tweet being a poem in its own right. In fact, Twitter may not be the right platform for autoflâneur: a static webpage with a “generate flânerie” button that produces a printable list of instructions of the desired length would probably suit the poem better. On the other hand, Twitter is now the easiest platform to write procedural texts for, thanks to Cheap Bots Done Quick, and comes with easier means of promoting and reading an otherwise esoteric artform. So in the future I might produce a static webpage version alongside the twitterstream.
At the moment Autoflâneur’s corpus (the body of textual possibilities on which her tweets draw) is relatively small and uncomplex. The two major tasks in writing a good procedural poem are having a big enough corpus with enough randomisation to produce pleasing variety, and striking the balance of occurrences between different elements so that the right tone is produced overall. Most of autoflâneur’s lines need to be about walking, with more contemplative instructions lightly interspersed, and they need to encourage enough different kinds of walking. I want to improve this in the coming weeks.
My dream is to be able to produce bespoke autoflâneurs for different cities: how could location-specific instructions be introduced into the corpus? How does having named streets and landmarks affect the psychogeography? A deeper dream is to have a generalisable geolocating autoflâneur that can produce instructions for wherever you happen to be, provided by sign-up and private message — but to accomplish that I’d need to work with a technologist. If you’d like me to make an autoflâneur for your city and have a budget to support it, get in touch.
For more guides to getting lost (which provided some of the source material for this project), see: