Jumbleroyd

This is a Tracery script for generating Yorkshire-sounding place names. You can see it in action at: http://www.twitter.com/jumbleroyd

This work is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License. https://creativecommons.org/licenses/by-sa/4.0/

Feel free to use, adapt and remix for whatever you like, just credit me if you do please.

Why:

Since moving to Yorkshire, the part that's really got under my skin is the place names. I moved here to do more walking and mountain biking, so spend a lot of time either out in the landscape, or poring over OS maps.

Yorkshire toponymy is a deeply beautiful knot of old languages; a smushing together of Old English and Old Norse resulting in sets of names that seem distinct from the rest of the UK.

I'm not a linguist. This weekend has been a crash course, and a lovely way of exploring how Yorkshire has percolated into me.

Jumbleroyd makes names that sound like they're from the landscape I've been exploring. Most of them are not real places, but it can duplicate real ones from time to time.

Structure:

The grammar is made by splitting place names into components, sometimes whole words, sometimes one or two syllables. They're then categorised as #prefix# #beginning# #middle# #end# #suffix# and used in various orders and configurations - mainly copied from the structure of real place names.

I built the lists subjectively rather than statistically or methodically. The script occasionally generates nonsense. It also has a bias to Calderdale, as it's where I live and has some of my favourite place names, which range from beautiful to lewd. With data for the whole UK, it would definitely be possible to make something more accurate.

Components:

I rejected some components as too generic: There are some very distinctive place names in Yorkshire that nonetheless end with something widespread. Suffixes like -chester, -ham, -borough, -minster, -wold, etc. saturate the UK too widely to feel like they're from Yorkshire.

Some other components are too distinctive: For example, Giggleswick and Penistone are unique to Yorkshire, but begin in ways that are far too identifiable. It's good enough these places exist for real, so, sorry, the bot is never going to tweet "Peniswood".

Prefixes include East, West and North. There's no South, partly because I don't commonly see anything labelled "South" in Yorkshire, and partly because it can be divided into North, West and East Ridings. There is no South Riding of Yorkshire.

There are a few suffixes that would require a more sophisticated grammar. For instance "-on-Tees" and "-le-Dale". I'm not sure these warrant the work required to build them into the script. It's awkward to create structures that use these suffixes correctly without messing others up, and they might end up overused too.

There are plenty of suffixes included here that are also common in counties further North, so it might sometimes make names that sound Cumbrian, Northumbrian, Lancastrian or Scottish.

Imperfections:

The script makes occasional double letters/letterforms at conjuctions. Examples: Mereerby, Appererby, Knowlley, Hebbleers, Warmmere, Runsshot, Grouseerby, Coldenend. If I were better at this, I might be able to script my own modifier to avoid double or triple letters where components join. Maybe I'll update this.

There are also occasional repetitions of syllables at conjunctions - to get rid of these, I should probably just avoid repeating certain syllables in multiple component lists.

There may also be place name components completely unique to Yorkshire that I've missed, and it may be that some I've allowed are too common, both in the UK and in this grammar, such that they sound inauthentic to someone. If that's the case, soz.

To (probably not) do: 

- Tune the frequency of certain forms
- A potentially messy probability/#origin# hack that might solve some small imperfections.
- A potential modifier to reduce ugly repetitions of syllables and letters at conjunctions.

Right now it's functional and consistent enough.

Resources:

https://en.wikipedia.org/wiki/List_of_places_in_Yorkshire

https://en.wikipedia.org/wiki/List_of_generic_forms_in_place_names_in_Ireland_an...

If you use it from an IP address in the UK, Bing maps ( http://maps.bing.com ) has an Ordnance Survey layer. There was a very satisfying moment while building up this script: scrolling round bits of Yorkshire looking for nice place names, I found the grammar I'd built could already generate most of them.

Kate Compton, originator of Tracery, has a tutorial and some tools here:
http://tracery.io

Beau Gunderson's Tracery editor is really useful for writing and testing scripts, because it shows a lot of results at once and highlights errors:
https://beaugunderson.com/tracery-writer/

And obviously, Cheap Bots, Done Quick, by @v21:
https://cheapbotsdonequick.com/

Photos:

1. Old book found in Todmorden.
2. Crenellations at Ravenscar.
3. View of Eaves Wood, Heptonstall.
4. Trig point at Great Bridestones.
5. Colden Clough.

If you like my photos, I post more on Instagram at @nnnnnach