Bay 12 Games Forum

Dwarf Fortress => DF General Discussion => Topic started by: anthony62490 on January 03, 2019, 07:16:23 pm

Title: Dwarven Census (Dwarf naming -Science-)
Post by: anthony62490 on January 03, 2019, 07:16:23 pm
As part of a larger project, I went through all of the names in a large newly-generated world and tallied up the origins and meanings of all of them. The results were pretty eye-opening for me.

RESEARCH :  A survey of 3829 dwarves in a newly-generated world yielded a pool of 249 unique names.
Out of 3829 generated dwarves, 89 had names of a non-dwarven origin
Of these, 83 were Human 5 were Goblin and 0 were Elven. 1 was unclear and has been removed from the percentages.

    Dwarven names:  (97.7%)
Foreign names:      (2.32%)
    Human names:    (2.16%)
    Goblin names:   (0.13%)
    Elf names:      (0.0 %)

FOREIGN NAMES:
Spoiler (click to show/hide)

DWARVEN NAMES
Spoiler (click to show/hide)

Here are some interesting things I have gleaned from this information. I mean, it's interesting if you're into that sort of thing.

Please note that this information only comes from one newly generated, mid-aged world. And even then, only the dwarves. If anyone's interested, I'm sure I could whip up something that can analyze all races from several worlds without so much work involved.

Stats generated using DF 0.44.12-r04
and collected with LegendsBrowser 1.17.1
Title: Re: Dwarven Census (Dwarf naming -Science-)
Post by: scourge728 on January 03, 2019, 09:15:44 pm
interesting
Title: Re: Dwarven Census (Dwarf naming -Science-)
Post by: delphonso on January 03, 2019, 10:33:14 pm
Zero elven names. Good.

This is very interesting, thanks for the hard work, anthony.
Title: Re: Dwarven Census (Dwarf naming -Science-)
Post by: Imic on January 07, 2019, 04:42:53 am
I found that to be very interesting. I qould have excpected Dwarves to name their children after warlike tags, but I suppose it makes sense for them to prefer crafting.
Title: Re: Dwarven Census (Dwarf naming -Science-)
Post by: Brightgalrs on January 10, 2019, 02:31:11 pm
So from this we can guess that Dwarf names come from the ARTIFICE and EARTH SYMBOL-sets which are defined in language_SYM.txt. That's 231 possible names. It's odd that you only caught 165* of them.

I conjecture that CULL_SYMBOL has something to do with it. A quick search turns up Bin ("Osed") as being in ARTIFICE, but not in your list there. Bin is also in the DOMESTIC set, so maybe that's getting filtered out and leaving us with 165.

Alright, with some investigation I've come up with:

ARTIFICE + EARTH - DOMESTIC - FLOWERY - NEGATOR - NEGATIVE - UGLY - EVIL = 175 Dwarf Names

Then the ones with no noun entries (Gild, Shoot, Wind [a clock]) are removed.

However, this leaves Granite, Tin, Copper, Bronze, Brass, Cobalt, and Ochre as missing (from your list). Dwarfs don't like to name their children after lesser metals/minerals, even with the above rules?

Oh, also you list "idîd" which should just be "id".
Title: Re: Dwarven Census (Dwarf naming -Science-)
Post by: anthony62490 on January 10, 2019, 03:45:45 pm
No, "Onget" IS a valid name. It's on the list. The reason that Sienna doesn't show up is that it's also in the FLOWERY pool. This is the reason that you'll never find a Dwarf named "Amethyst", "Aquamarine", or "Umber". All are EARTH words, but they are also FLOWERY, which dwarves are forbidden from using.

My big revelation from here is that no one ever deviates from these rules like the wiki implies. Like every once in a while, you'd find a dwarf with a name from the WEALTH pool, but not EARTH or ARTIFICE. But that wasn't the case.

Quote from: Brightgalrs on January 10, 2019, 02:31:11 pm
That's 231 possible names. It's odd that you only caught 160 of them.
I caught 165. If I am correct, there should be 175 valid Dwarven names (earth + artifice - (domestic, subordinate, evil, flowery, negative, ugly, negator)) provided they were born in a dwarf civilization. So I guess this world skipped over a few names. The issue is that several COLOR words are also found in the FLOWERY pool.

I ran a quick script that added up all EARTH and ARTIFICE names and the filtered out all the forbidden tags, and I came up with this list. It's 175 names long. I haven't tested it against an actual generated world, but it all LOOKS pretty accurate to me.

ALL VALID DWARVEN NAMES
Spoiler (click to show/hide)
Title: Re: Dwarven Census (Dwarf naming -Science-)
Post by: Brightgalrs on January 10, 2019, 04:48:34 pm
I think we've both been doing the exact same thing for the last hour or so. Lol. My post has since been updated.
Title: Re: Dwarven Census (Dwarf naming -Science-)
Post by: anthony62490 on January 10, 2019, 04:52:54 pm
I compared the list of names I was generated to the list of valid names, and I didn't find many discrepancies. The names that my world failed to generate were:
Akrul, Gusil, Keskal, Kilrud, Konos, Letmos, Lolok, Nokim, Oltar, and Oslan

Also, I found that "Idîd" is not even a Dwarven word (or a word in any other language), so I'm unsure how it managed to not only get into my list of names, but also get identified as EARTH, NATURE, PRIMITIVE. I'm pretty sure this was the name "Id" repeated, because the world DID generate Ids that were not counted properly (there were 29). The list has been edited to match this.

So I took a look at these names to see why they never showed up, and I came up with the following:

Akrul - Tin - EARTH
Gusil - Copper - EARTH COLOR
Keskal - Shoot - ARTIFICE VIOLENT (invalid as a first name)
Kilrud - Bronze - EARTH COLOR
Konos - Brass - ARTIFICE COLOR
Letmos - Cobalt - EARTH COLOR
Lolok - Granite - EARTH
Nokim - Ochre - EARTH COLOR
Oltar - Gild - ARTIFICE WEALTH (invalid as a first name)
Oslan - Wind (like a clock) - ARTIFICE (invalid as a first name)

Based on this, it does LOOK like many of the COLOR words are not being chosen, but since Onget DID make it into the list of generated names, I think it's more likely that this is a coincidence. Additionally, I'm positive that I've seen dwarves named Kilrud and Letmos before.
Title: Re: Dwarven Census (Dwarf naming -Science-)
Post by: anthony62490 on January 10, 2019, 04:57:43 pm
Ha! We have been doing the exact same thing. Now my numbers are all off though. I need to fully automate this. Doing it by hand is too error-prone.

EDIT: Huh. That is strange. Almost all of the missing names are of minerals. But I KNOW I've seen dwarves named Kilrud before. Hold on, I'm gonna generate a huge world and have a poke around. This feels like a coincidence, but I want to double check. I'm going to make a large world with a very long history just to maximize the number of entities that get generated. This could take a while, so I'll check back in a few hours.
Title: Re: Dwarven Census (Dwarf naming -Science-)
Post by: Brightgalrs on January 11, 2019, 11:29:58 am
Oh I think I've figured it out. As I said before, the words need to have a Noun entry (invalidates Shoot, Ochre, Gild, Wind) but they also need the [THE_NOUN_SING] tag, which Tin, Copper, Bronze, Brass, Cobalt, and Granite do not. All of this is defined in language_words.txt.
Title: Re: Dwarven Census (Dwarf naming -Science-)
Post by: GPeter on January 11, 2019, 11:34:02 am
My good man, I must give my congratulation to you for such hard work! It must have been a huge pain in the ass to do this!
Title: Re: Dwarven Census (Dwarf naming -Science-)
Post by: FantasticDorf on January 11, 2019, 12:17:02 pm
This is nice, this made me a bit more invested to look into how exactly they worked.

When i ran some modded goblin fortresses that usually used the evil symbol on by default, they kept returning to errorlog that they were 'impoverished' of word choices when i created new location areas, and suitably didn't really generate any special names unless i typed them in myself, which to that point i enjoy the evil naming scheme anyway with my ngokang (curse) and names like this being very common but otherwise just slightly amusing.

Do the symbols follow their sphere counter parts? War aren't connected by spheres (http://dwarffortresswiki.org/index.php/DF2014:Sphere) but are closely related anyway if not directly in the sheet, you can probably link Domestic to any number of them unless the rest are unused or literally can be called up individually like [SYMBOL:LIGHT], but i've not tried this so im not sure.

Therefore it might be easier to assign more sub-symbol preferences for more variety, having all elves adopt food for names like 'garlic', though dwarves tend to talk about food a lot in regards to tavern names in reference to syrups.
Title: Re: Dwarven Census (Dwarf naming -Science-)
Post by: Brightgalrs on January 11, 2019, 12:43:49 pm
Oh hey entity_default.txt gives the ARTIFICE + EARTH - DOMESTIC - FLOWERY - NEGATOR - NEGATIVE - UGLY - EVIL equation for dwarves:
Quote
[SELECT_SYMBOL:REMAINING:ARTIFICE]
[SELECT_SYMBOL:REMAINING:EARTH]
[CULL_SYMBOL:ALL:DOMESTIC]
[CULL_SYMBOL:ALL:SUBORDINATE]
[CULL_SYMBOL:ALL:EVIL]
[CULL_SYMBOL:ALL:FLOWERY]
[CULL_SYMBOL:ALL:NEGATIVE]
[CULL_SYMBOL:ALL:UGLY]
[CULL_SYMBOL:ALL:NEGATOR]

And for Elves:
Quote
[SELECT_SYMBOL:REMAINING:NATURE]
[SELECT_SYMBOL:REMAINING:FLOWERY]
[CULL_SYMBOL:ALL:DOMESTIC]
[CULL_SYMBOL:ALL:SUBORDINATE]
[CULL_SYMBOL:ALL:EVIL]
[CULL_SYMBOL:ALL:NEGATIVE]
[CULL_SYMBOL:ALL:UGLY]
[CULL_SYMBOL:ALL:NEGATOR]

Goblins:
Quote
[SELECT_SYMBOL:REMAINING:EVIL]
[CULL_SYMBOL:ALL:DOMESTIC]
[CULL_SYMBOL:ALL:FLOWERY]
[CULL_SYMBOL:ALL:HOLY]
[CULL_SYMBOL:ALL:PEACE]
[CULL_SYMBOL:ALL:NEGATOR]
[CULL_SYMBOL:ALL:GOOD]

Humans don't have [SELECT_SYMBOL:REMAINING:XXX].
But they do have CULL_SYMBOL:
Quote
[CULL_SYMBOL:ALL:SUBORDINATE]
[CULL_SYMBOL:ALL:EVIL]
[CULL_SYMBOL:ALL:NEGATIVE]
[CULL_SYMBOL:ALL:UGLY]
[CULL_SYMBOL:ALL:NEGATOR]
Maybe all words (with [THE_NOUN_SING]) that are not in the above symbol sets are valid names?

Kobolds don't have SELECT_SYMBOL:REMAINING either, but they have the [UTTERANCES] tag, which results in randomized names that don't correspond to English words. So it makes sense.
Title: Re: Dwarven Census (Dwarf naming -Science-)
Post by: FantasticDorf on January 11, 2019, 01:48:56 pm
This'll be of use, the wiki has been a useful resource as always.

Select Symbol  (http://dwarffortresswiki.org/index.php/DF2014:Entity_token#SELECT_SYMBOL) - Out of the list on language SYM in folder
Sub-Select Symbol (http://dwarffortresswiki.org/index.php/DF2014:Entity_token#SUBSELECT_SYMBOL) - Adjective use, maybe of interest

Though i think this isn't hardcoded, if you can find the relevant names, [SELECT_SYMBOL:ALL REMAINING:*SYMBOL*] can be overidden, so perhaps you could write your own custom [SYMBOL:METAL] which shouldn't conflict normally with [SYMBOL:WEALTH] unless its culled.

Quote
SYMBOL METAL
S_WORD:ALUMINUM //no spoken word are striked through, most of these are natural ores though
S_WORD:BISMUTH
S_WORD:COPPER
S_WORD:GOLD
S_WORD:IRON
S_WORD:LEAD
S_WORD:NICKEL
S_WORD:PLATINUM
S_WORD:SILVER
S_WORD:TIN
S_WORD:BRONZE
S_WORD:STEEL // alloy with a name like bronze
S_WORD:PIG_IRON //this one is tricky, might need its own specially made subselect-symbol for alloys if that works. "thuveg datan" in dwarvish, other words like rose gold are contrarian to flowery culling
S_WORD:ZINC
S_WORD:BAR // some overlap here
S_WORD:FURNACE
S_WORD:ANVIL
S_WORD:RUST
S_WORD:FORGE only labouring exists as a phrase of this, as a verb or noun

Taverns are hardcoded as far as i can see into food related names.
Title: Re: Dwarven Census (Dwarf naming -Science-)
Post by: anthony62490 on February 20, 2019, 07:43:17 pm
Hey. Me again. This may be only of academic interest, but the name "Oggez Rashas" (of Bronzemurdered fame) is not naturally attainable as far as I can tell. "Oggez" is a valid first name (meaning "The Hole" in Dwarven), but Rashas is not a word in any language as far as I can tell. Maybe it was different in 0.31, but I cant find evidence of it. The closest matches I could find are "Náshas" (Lilacs, which would be culled via Dwarven naming rules against FLOWERY names) and "Sashas" (Rushing, which is oddly not included in any language_SYM pool).
So I'm  not actually sure what Oggez Rashas's name means beyond "The Hole of ____"
Title: Re: Dwarven Census (Dwarf naming -Science-)
Post by: Bumber on February 21, 2019, 01:52:25 pm
Quote from: anthony62490 on February 20, 2019, 07:43:17 pm
Hey. Me again. This may be only of academic interest, but the name "Oggez Rashas" (of Bronzemurdered fame) is not naturally attainable as far as I can tell. "Oggez" is a valid first name (meaning "The Hole" in Dwarven), but Rashas is not a word in any language as far as I can tell. Maybe it was different in 0.31, but I cant find evidence of it. The closest matches I could find are "Náshas" (Lilacs, which would be culled via Dwarven naming rules against FLOWERY names) and "Sashas" (Rushing, which is oddly not included in any language_SYM pool).
So I'm  not actually sure what Oggez Rashas's name means beyond "The Hole of ____"
Wouldn't the last name be two words put together?
Title: Re: Dwarven Census (Dwarf naming -Science-)
Post by: anthony62490 on February 22, 2019, 12:27:43 am
Sometimes yes, but Forgotten Beast names often follow different rules. They CAN have normal "Noun Word-Word" names like all other entities, but their names seem to fall into one of (I think) four categories:

This does not account for epithets, which add another layer of complexity to the name. And furthermore, it IS possible that "Rashas" could be a combination of two words, but I was unable to find them and very few dwarven words are that short anyway. If you'd like to take a look, you're more than welcome. http://dwarffortresswiki.org/index.php/Template:Language_table
Title: Re: Dwarven Census (Dwarf naming -Science-)
Post by: feelotraveller on February 22, 2019, 01:19:18 am
I wonder if it wasn't (correctly) Rashäs ?
Title: Re: Dwarven Census (Dwarf naming -Science-)
Post by: anthony62490 on February 22, 2019, 11:27:08 am
Woah. Hold up. I found it. I have no idea how I missed this before. His name is Oggez Råsh äs, "The Hole of Dying Caves". How did I not see that earlier? Bumber, you are absolutely correct.

EDIT: Hold up again. His name actually closer fits the "Xx" template now. The previous translation doesn't actually match any naming format I've come across yet. So his translated name is probably closer to "Oggez Deathcaves". Shockingly appropriate.
Title: Re: Dwarven Census (Dwarf naming -Science-)
Post by: Dorsidwarf on February 26, 2019, 06:14:54 am
I thought that "A BC The X of Y" names generally indicated a title given to them as a reslt of historical deeds
Title: Re: Dwarven Census (Dwarf naming -Science-)
Post by: anthony62490 on February 26, 2019, 11:21:26 pm
Yes, you are correct. That doesn't fit the standard templates and indicates a more complex name. I haven't had time to really dive into epithets yet, but I'll eventually get a working formula figured out. It'll be interesting to see how multiple epithets pile on top of each other.
Title: Re: Dwarven Census (Dwarf naming -Science-)
Post by: anthony62490 on May 05, 2019, 01:52:42 pm
So I've finished the first version of my name generator (now conveniently available at d0rfforge.com (http://www.d0rfforge.com). plug plug) and I'm looking to make it more accurate. Right now, it only uses the base word and completely ignores tenses, plurals, verb/adjective forms, and all those other boring grammar rules. In order to conform to those rules, I need to know WHAT those rules are, The raws give a lot of information, but it's all in Toady's weird token-based data structure. Gross.

The Wiki says that
Quote
First Names are always nouns
Second Names are always two-word pairs and can be any combination of word types, but Noun-Noun, Adjective-Noun, and Verb-Noun appear to be the most common

So I pretty much need to reverse engineer Toady's naming rules. I've never done this before, but I imagine if I can get a big enough list of dwarves, I'll be able to find a pattern in their naming schemes that I can actually use. So...

Just for context, here's what ONE word looks like with 2 noun forms, 5 verb forms, 1 adjective form, and no prefix forms
Code: [Select]
"blockade": new DFWord("blockade:blockades", "blockade:blockades:blockaded:blockaded:blockading", "blockaded", null)
And here's what the script spit out. The (bit in parentheses) denotes how many nouns, verbs, adjectives, and prefixes there are to choose from while the [bit in brackets] denotes which options are available and which were chosen.
Spoiler (click to show/hide)

This looks an awful lot like a game of Mastermind, but there might not even be an answer this time. language_words indicates that all words have designated spots where certain forms are used ([REAR_COMPOUND_NOUN_SING] probably means that IF this word is chosen for the 3rd word, use the singular noun form), but this isn't easily digestible and it'll take a lot of time for not much payoff.

SO. I split the names up into front and rear compound groups and got these two lovelies. I've separated them based on which word forms are available and sorted them based on unique configurations.
FRONT COMPONENT
Spoiler (click to show/hide)
REAR COMPONENT
Spoiler (click to show/hide)

I'm not certain that I see any patterns yet, but I might be able to pull a few good approximations out of this.