As a conlanger, I’ve been curious about the phonotactics of fantasy names for the player characters in D&D 5e. My assumption is that these names are designed to be vanilla, easy-to-pronounce names, while conveying a bit about character and setting. As someone who overengineers names, I wanted to assess the names I generate against the mainstream standard.
The Player’s Handbook (PHB, shown above) has lists of suggested names for dwarves, elves, halfings, humans, dragonborn, gnomes, half-orcs, and tieflings. For humans, there are lists of names by eight of the ethnic groups from Faerûn, the default setting for D&D. If you don’t own the PHB, you can check out the free D&D Basic Rules for the names for dwarfs, elves, halflings, and humans. And here’s a link to that page in D&D Beyond (for subscribers who’ve purchased the PHB).
Other books, including Xanathar’s Guide to Everything, Volo’s Guide to Monsters, and the various setting books, add additional names, but I was interested in the core, most common names.
I omitted names that were compounds of English words (such as the dwarf clan name Brawnanvil) but kept names that might be in an imaginary language.
In all, I analyzed a list of 539 names. Since these represent at least 15 languages (one per demihuman race and eight for humans), that’s only 36 words on average per language, a small corpus to analyze on a per-language basis. For convenience, let’s assume these languages all share a common ancestor; I’m going to call this cross-section of languages descended from that ancestor Pan-Faerûnese.
I converted the names to representations of common types of phonemes. Since D&D doesn’t provide suggested pronunciations, my conversion of certain words might differ from yours. For instance, is holderhek /hol-der-hek/ (my first thought) or /hol-de-rhek/ (what I ended up with)?
|Category||Symbol and Representations|
|vowels||V = a, e, i, o, u, y, ï, aa, ee, ii, oo, uu, ah, eh, ay, ey, oe, ue, eu|
|plosives||P = t, k, p, b, d, g, kh, bh, gh, ck|
|sibilants||S = s, z, sc, sch, sh|
|Liquids||L = l, r, rh|
|Nasals||N = m, n, ng, mh|
|fricatives||F = f, ph, th, v|
|affricates||C = tch, ch, j, jh|
|semivowels||W = w, y|
|aspirate||H = h|
And a few odds and ends:
- qu /kw/ -> PW
- qi /ky/ -> PY
- x /ks/ -> PS
Doubled consonants are typically assumed to convey length if they are word final and to be repeated across syllables if not (e.g., Vall FVL vs. Jassan CVSSVN) but I tended to remove them in Halfling words (e.g., Wellby WYLPV).
- Nyx, NVPS
- Marsk, NVLSP
- Ghesh, PVC
- Vladislak, FLVPVSLVP
- Linxakasendalor, LVNPSVPVSVNPVLVL
Much of this analysis involved judgement calls. Professional linguists might do it differently and might not agree with each other; after all, this is an extrapolation from spellings not sounds.
Two things that surprised me:
- The scarcity of heavy-metal umlauts, given their presence in the name of the world (Faerûn). Only one of the names had a diacritic: the Elf surname Naïlo.
- The presence of ‘ck’ in 5 words, all for Gnomish names (e.g., Boddynock & Ellywick Nackle); we’ll assume this was to make this language seem evocative of English. It does indicate that thought has been put into making the languages someone distinct from one another.
How many sounds in the average word? And keep in mind this is sounds. For instance, the seven letters of Mishann represent the five sounds shown in the phonotactical representation NVSVN. The stats for length: mode of 5, median of 6, average of 5.7. Here’s the distribution:
Most Common Phonotactics
The detailed phonotactical representations are too detailed when it comes to looking at common patterns in aggregate. For instance, the #1 pattern, PVLVN, only matches 1.3% of the corpus. Here are all the patterns with three or more examples:
Since that was too granular, for this next analysis, I’m using C to stand in for any consonant. The following 16 phonotactical patterns represent 61% of the names.
|CVCCVC||9.5%||51||CVLNVN, CVSNVL, CVSSVN, FVLPVL, FVNNVN, FVNPVL, FVNPVN, HVLPVP, HVSPVN, LVNLVN, LVNPVL, LVNPVL, LVNPVL, LVNPVL, LVPPVL, LVSWVN, NVLFVS, NVLLVN, NVLLVP, NVLNVP, NVLPVL, NVLPVN, NVPLVS, PVLFVN, PVLHVN, PVLLVN, PVLLVN, PVLLVN, PVLLVN, PVLLVP, PVLLVP, PVLLVP, PVLLVP, PVLPVL, PVLPVP, PVLPVS, PVNCVP, PVNNVF, PVNPVL, PVNSVF, PVPLVN, PVPLVN, PVPLVN, PVPNVL, SVFLVS, SVLLVN, SVLPVL, SVPLVP, SVPPVN, WVLLVN, NVNFVPL|
|CVCVC||8.5%||46||FVLVN, FVLVN, FVLVS, FVPVL, FVSVN, HVLVN, HVPVP, LVLVN, LVLVP, LVNVS, LVPVL, LVPVL, LVWVN, NVFVL, NVFVS, NVHVN, NVLVC, NVNVL, NVNVN, NVNVP, NVPVL, NVSVN, PVFVF, PVFVL, PVLVF, PVLVL, PVLVN, PVLVN, PVLVN, PVLVN, PVLVN, PVLVN, PVLVN, PVLVP, PVLVS, PVNVL, PVNVN, PVNVP, PVSVF, PVSVL, SVLVS, SVNVF, SVNVL, SVNVS, SVPVN, NVLVLP|
|CVC||5.6%||30||CVN, FVL, FVN, FVN, FVP, LVF, LVN, LVN, LVN, LVP, NVN, NVN, PVF, PVF, PVL, PVL, PVL, PVL, PVN, PVN, PVN, PVP, PVS, SVN, SVN, SVN, SVP, WVN, WVN, HVLN|
|CVCV||5.4%||29||CVFV, FVFV, FVLV, FVLV, FVNV, HVNV, HVNV, LVLV, NVLV, NVLV, NVLV, NVLV, NVLV, NVPV, PVFV, PVFV, PVLV, PVLV, PVLV, PVNV, PVNV, PVNV, PVSV, SVFV, SVFV, SVLV, SVLV, SVPV, FVLVN|
|VCCVC||4.3%||23||VLCVN, VLHVL, VLLVC, VLLVN, VLPVF, VLPVN, VLPVN, VLPVN, VLPVN, VLPVN, VLSVP, VNNVF, VNNVL, VNNVN, VNPVL, VNPVL, VNPVL, VNPVN, VNSVL, VPLVN, VPLVP, VSPVL, VFLVPP|
|CVCCV||4.1%||22||FVLNV, HVLCV, LVLLV, LVPPV, LVSPV, NVLLV, NVLLV, NVLPV, NVPPV, NVSSV, PVFLV, PVFLV, PVFLV, PVFLV, PVLLV, PVLLV, PVLLV, PVLPV, PVPPV, SVNNV, WVLPV, CVLNVN|
|CVCVCVC||3.3%||18||FVLVFVL, HVFVLVL, PVFVNVL, PVLVFVP, PVLVLVS, PVLVNVF, PVLVNVP, PVLVPVS, PVLVSVL, PVNVFVL, PVNVPVS, PVPVNVP, PVSVPVL, SVFVLVS, SVLVSVL, SVPVHVL, WVWVPVP, NVLVFVLPV|
|CCVC||3.0%||16||FLVF, HLVN, NWVF, PLVF, PLVF, PLVL, PLVN, PLVN, PLVN, PLVN, PLVN, PLVP, SPVL, SPVP, WLVN, FLVNP|
|CVV||3.0%||16||CVV, FVV, FVV, FVV, LVV, LVV, LVV, NVV, NVV, PVV, PVV, PVV, SVV, SVV, SVV, CVVN|
|CVCVCCV||2.8%||15||CVPVPPV, FVSVLPV, LVLVLLV, LVLVSSV, LVNVLPV, LVNVNPV, LVSVNNV, NVFVLLV, PVLVPLV, PVNVLLV, SVFVLLV, SVLVFLV, SVLVPWV, WVFVLPV, CVLVNNVF|
|CVCVCV||2.2%||12||CVLVNV, FVLVPV, LVLVNV, LVNVLV, NVLVFV, NVLVLV, NVPVLV, PVLVNV, PVNVNV, SVLVNV, SVLVSV, FVLVFVL|
|CVVCV||2.0%||11||FVVLV, LVVSV, NVVLV, NVVLV, PVVLV, PVVLV, PVVLV, PVVSV, SVVFV, SVVNV, LVVPVN|
|CVVC||2.0%||11||CVVN, FVVN, FVVP, HVVN, LVVL, LVVL, LVVN, LVVN, NVVL, PVVN, PVVLN|
|CVCC||2.0%||11||HVLN, HVNP, LVNP, NVLN, NVPL, NVPS, PVLN, PVLP, PVNC, SVNP, NVLSP|
|CVVCVC||1.7%||9||LVVPVN, LVVWVN, NVVLVL, NVVLVS, PVVLVN, PVVLVN, PVVPVL, SVVNVL, FVVLVSPLV|
|VCCV||1.7%||9||VLLV, VLNV, VLPV, VLPV, VLPV, VNNV, VPLV, VPPV, VLCVN|
Differences by Language
Again, given the small sample sizes by language, I didn’t really want to spend too much time here. But some things I noticed:
- Gnome – The afore-mentioned use of ‘ck’ in this, and only this, language. Word final syllabic consonant /l̥/ in: dimble, gimble, nackle, namfoodle.
- Orc – Half-orcs can take a human name corresponding to their ancestry, or they can take a a short (4.3 sounds) orc name: e.g., Vola, Keth, Holg.
- Rashimi – Rashimi names had the most sounds, at 7.3 on average (e.g., Stayanoga, Murnyethara, Chergoba, Vladislak).
- Shou – Shou names were the shortest, with 3.1 sounds. Six of eleven occurrences of ‘ng’ are in Shou, where they are always word final (e.g., Jiang, Huang, Ling). (Otherwise two occurrences of ‘ng’ in Dwarf names, two in Half-orc names, and one in a Gnome name.)
The names I generate for my own campaigns definitely are longer than even those of Rashimi, which certainly explains why my players sometimes find them hard to remember and pronounce! As a result, I’ve been gradually revising and shortening the names I’m using to come more into line with the expectations set by 5e.
If you found this article interesting, please check out my book, Langmaker: Celebrating Conlangs.
The cover of the Player Handbook is copyright 2018 Wizards of the Coast and is used as part of their Media Relations program.