Methods and Morphology of Conlanging

by johannespunkt

This is a post on conlanging.

It details 3 methods of constructing a language more or less from scratch – that is, not basing it off already existing languages. There are surely other ways to do conlangs but these are the ones that have occured to me/the people I conlang with. Let’s give the methods the arbitrary names clusterfuck, evolutionary, and interpretive.


…is the approach taken with the Cekno’s language; all ideas and jumbledness, like a brainstorming session that doesn’t really end. It goes a bit unwieldy after a few months of working like that, apparently. This involves a lot of jumping back and forth between different topics. You can do it more or less structurally but it is the most chaotic method nonetheless.

In trying to create the Cekno Idiosyncrasy, we tried to think like the star and figure out what kind of stuff it would think up, and it got confusing. Rather alien, but … yeah.


…is when you try to create a language by making up the culture that goes along with it, maybe following its route from some imaginary savannah period to living in tents and placing religious importance on testicles. The way I considered doing this was to take a tribal society and mark out the concepts important to them, define some semantic clusters, and then add more stuff in hundred-year-leaps. This’d include some sound shift and imaginary politics.


This is the fun method. Credit for this idea goes to my good friend @kerastion (Rob Mitchelmore) who is full of various and great ideas, as you can tell from his contributions to the /glossary.

What you do is that you define a morphology. I probably need to learn some coding to do this but that’s fine. We can get the morphology in a few steps, and for simplicity’s sake we’ll do this with Latin letters. First we generate some syllables and double-syllables that we think are nice. Say, 150 of them or so, for a modest number. These are our base words or something. It is important to not give them any meaning. Then we define a few rules for gluing the words together, for putting words after each nother, and for morphing the words – adding and removing sounds, changing sounds and so. We do a test run of generating these words, say a hundred words, and read it aloud to make sure it sounds like something that could be an actual language.

When that is done, we generate about a book’s length of these and start interpreting what they mean, wildly. Have you made a tourist guidebook to a foreign planet? Have you written a compelling romance novel? A monster manual? We don’t know until we start picking it apart. And, in picking it apart, we create a language that is probably not complete but there can be communication in it.

I’ve no idea how long such a project would take, but I am thinking of starting one and seeing where it leads me.