Our name pages exist to help storytellers, and a name is only useful if it is what it
claims to be. This page explains how each list is built, what we check, and — the part
most pages leave out — what we don’t know.
That last part matters. A names page that supplies a confident meaning for every entry
is telling you something no source could support.
Three kinds of name page
What we do depends on whether the names are real, and we do not blur the line.
| Kind of page | What we do |
|---|---|
| Real naming traditions Italian, Irish, Japanese, Native American, Hebrew and so on |
Retrieved from the record, never made up. Names come from reference works. If a tradition supports 400 documented names, the page carries 400 and says so, rather than padding to a rounder number. |
| Real things cities, dinosaurs, mountains, saints |
Retrieved from the relevant authority — geographic gazetteers, palaeontological databases, historical registers — not from a general dictionary. |
| Fictional worlds fantasy towns, sea monsters, mage and assassin names |
Written for the page, and openly so. There is no real-world Lizardfolk language to be faithful to, so an original name is the honest answer here — and we build them from documented name elements, so what they mean is true rather than decorative. What we will not do is present a fictional name as a documented one. |
The failure we are correcting across the site is the third approach leaking into the first:
fictional names presented as a real tradition. That is the one thing these pages must never do.
Where the material comes from
Sources differ by page, and each page lists its own. For real naming traditions we work
from published reference material — typically Wiktionary and Wikipedia in both English
and the language concerned, checked where possible against scholarly lexica that are not
crowd-edited: Lewis & Short and Liddell–Scott–Jones for Latin and Greek roots,
Strong’s Hebrew Dictionary for Hebrew and Aramaic. Those three are public domain and are used
to check the crowd-edited sources, not to replace them.
A model is never asked to “list the names” for a real tradition. That request
cannot be answered honestly, and what comes back is guesswork wearing the clothes of data.
Where a model is used, it translates or summarises sourced material and is given that material
verbatim.
How a name gets its meaning
A meaning is printed only when a source states one. The rule we hold to throughout is that
an origin is not a meaning. “From the Hebrew Avraham”
tells you where a name came from and nothing about what it means; only an explicit gloss counts.
Where a name is a compound — most Germanic names are — both halves are given.
Enrico is heim “home” plus ric “ruler”, so it reads
“home + ruler”, not “home”. Where a source states the literal sense of
the whole name, that wins: Leonardo is recorded as “as strong as a lion”, which is
better than anything assembled from its parts.
What we don’t know
Every page of real names carries entries whose etymology nobody has recorded. Those names
stay on the list — they are real and useful — with the meaning left blank rather
than filled in. Where a source explicitly says the etymology is unknown or disputed, that is a
finding and we treat it as one.
Verification is also uneven, and we would rather say so than imply otherwise. Latin, Greek
and Hebrew roots can be checked against scholarly lexica. Germanic roots largely cannot: there
is no public-domain reference for Proto-Germanic or Frankish on the same footing, so those
entries rest on a single source. Each page states its own figures — how
many names it carries, how many have a recorded meaning, and how many are corroborated by more
than one source.
Mistakes this process caught
A method is only worth describing if you can show it working. These are real errors caught
while building these pages, all of them found by checking a second source.
- Reading links without the sentence around them. One entry reads “a
variant of Carmelo, inspired by its Spanish equivalent Carmen”. Taken as data rather than
prose, “inspired by” becomes “descended from” and invents a Spanish
origin for a Latin name. - Treating a footnote as an etymology. “From Latin Marius. By folk
etymology seen as a masculine form of Maria.” The second sentence is a note about a
misunderstanding, not a source. - Importing another language’s history. Filling an Italian name from its
English cousin brought English’s own ancestry with it — Middle English Lucy is
not an ancestor of Italian Lucia. - Printing half a compound. Dozens of names were showing only their first
element: “horse” where the name means “horse + protection”. - Extracting a dissent. One root entry gives an accepted etymology, then a
rejected one introduced by “although”. We were reading the rejected one. - Matching a name against a dictionary of words. Abandoned outright: it
scored roughly 6% precision, offering “to send away” for a name that comes from
amandus, “worthy of love”. A dictionary of words cannot be used as a
dictionary of names.
Checking anything here yourself
Every meaning we print can be traced. Each page names its sources; the scholarly lexica are
freely readable through the Perseus Digital Library, and the Wiktionary and Wikipedia entries
are a search away. If you find something we have got wrong, we would genuinely like to know.