On Names
I am as usual late to this blog bandwagon. I think the original one was here, though that wasn't the one I saw at first. It was probably Phlox's, which is quite good. Here's another one that talks about another approach to randomly generating names.
I struggle a lot with names, like many people. When I play games, I often end up going down some sort of online rabbit hole in search of a name. My most recent named character, Gemekaa Sa'at, is from a sort of ancient mesopotamian society and so I dug up a list of Babylonian names and spent way too long picking one that feels right. Picking a name feels consequential in an outsized way.
As such, as a GM, I tend to get frozen on the spot if I need to come up with a new name, so I come up with a list of names in advance. I pull from many of the same places others have mentioned: online generators, or just kind of thinking of sounds and putting them together. In a historic setting I'll find a list of historic names.
There is another, much less reasonable thing you can do, which is to create your own custom word generator for a variety of made up languages.
Fake Language Word Generators
It started with a sort of rather crude conlang I've been working on. It's mostly a naming language, and I realized that I was having the same writer's block coming up with words even with the well defined set of rules for what sounds can exist in what combinations. So I made some tables to roll on. It's a language where every syllable consists of one or more consonants followed by one or more vowels, and you can generate words for it by rolling some dice, in fact. More about it here (warning: self-indulgent worldbuilding). It dates back to my 5E days and I want to overhaul it to be more for the games I play now, but also realistically it's a personal project, it's not something I'm going to inflict on my players.
Next, for a while I was playing a bunch of solo Cairn. (I'd like to make the time to go back to that, but I have somewhat been neglecting solo games, as that time has mostly been taken up by Cataphracts.) I then made my next generator: one that makes words that are sort of English-adjacent, but a little different. I then ended up making a few more and then inflicting them on my cataphracts players. At this point though, I wanted people to be able to kind of automatically guess that two people are from different cultures based on their names.
I basically used vague memories of a literal Linguistics 101 class I took. The general principle I use here is, at the most basic level, languages have rules, and if you have two sets of rules that are distinct enough, the words that are generated will sound like they come from different languages. Languages have a set of sounds that are allowed, and sounds that are not allowed, and combinations they can exist in. For instance, "mug" is a word in English. "Wug" is not but it sounds like it could be. "Mgu" is not, even though it's the same letters, and if you saw it written down, you'd assume it was from a different language, because in English words do not start with Mg. "Xug" also sounds like it's not from English, even though the letter X exists in English, because it almost never exists at the beginning of words. Fantasy names often do this - often overdo it, to the point of cliche. Add in more unfamiliar attributes - make it X'ug, for instance, and it's probably an interdimensional being or something.
So, I came up with a few word generators intended to sound like they are all from different languages. Or rather, to look like they are, when ued in writing.
Two Examples
Let me compare two languages here: Uira and Broik. The words "Uira" and "Broik" themselves were each generated by their respective generators, in fact.
If you look at the following list, can you guess which ones are from the same language?
(Broik, Broik, Uira, Broik, Uira, Broik, Broik, Uira, Uira, Broik)
There might be some overlap, but there are some that very clearly fall into one or the other.
The sounds allowed in Broik are:
As well as a lot of consonant clusters:
But not many dipthongs:
There are some other rules. I categorized some sounds as "simple" or "complex" and limited how much complexity a single syllable could have. This is less about realism and more about making it usable. People often struggle with complexity in unfamiliar languages more than their own, so in a made up naming language, you can artificially limit the complexity of what you are asking people to understand.
The sounds allowed in Uira are:
As well as a larger number of dipthongs (vowel combinations). I now realize that there's a fair bit of overlap here with the other language, but that's fine. I think. I'm tempted to change it now but it wasn't apparent to me until I wrote this blog post so maybe it's fine.
I also threw in a few structures that I am treating as dipthongs for word complexity purposes but are not, as well as a a particular "eh" sound.
What does this "eh" combination mean? I don't know, I didn't decide, and won't until it comes up.
It ends up being more vowel heavy, and producing longer words.
More on how to make them
Overall, while I started with the vague memories of something actually academic, a lot of this is just making stuff up until it has the right feel. I was able to get there from the general principle that there should be underlying rules and that if they exist, it'll feel right, but I am sure an actual linguist would start crying if they saw all of this. That's fine though.
One secret is, if you're only using these in writing (I use them for cataphracts), you can kind of fake it. What do those dipthongs represent? It doesn't actually matter because you aren't saying them out loud. You can just sort of think, I am doing some distinctive thing to my vowels here and never actually define what it means. One of my generators uses a lot of double letters to be distinctive. What do the double letters mean? I have not decided at all.
You can also think very carefully about your generators but it's a generator, you can also tweak them a bunch until it getsyou results that vaguely feel right. There's a tradeoff between generating names you can easily tell apart from each other, and generating languages which as a set are distinct from other languages. Also, you do probably have to exaggerate a bit more the differences between languages than you would with real languages, because with real languages you can also assume some cultural context.
Originally I made these in Perchance, which is pretty user friendly, but I moved away from it and towards my own Javascript version because of how AI-focused perchance has become.