English (GB)
5879 ranked first names in the benchmark
- Top 100 (100)
- 16%
- Top 500 (500)
- 18.2%
- Top 1000 (1000)
- 19.7%
- Avg. variants among matches
- 1.29
Informational benchmark
Spell names with chemical elements: which popular names work best?
Yes, many real names can be spelled with chemical elements, but most still cannot. In our country-level top-1000 first-name benchmarks, hit rates run from 15.7% in Dutch (NL) to 20.1% in German (DE), so the useful question is not just “can it work?” but “which kinds of popular names work most often, and why?”
To answer that, I benchmarked top-ranked first-name lists for Carbonat's GB, DE, ES, FR, IT, and NL datasets, plus a small Welsh fallback list, and asked three practical questions: how often top-100, top-500, and top-1000 names can be spelled with real element symbols; which countries keep the strongest hit rate as you go deeper into the ranking; and which popular names stay flexible enough to allow more than one valid split.
Terminology
None of this terminology is technically grounded. I made these labels up for the sake of naming the patterns in a readable way.
A split is one valid way to break a name into element symbols, such as Simon becoming `Si` + `Mo` + `N`. A split pair is one adjacent joining inside that split, such as `Si` + `Mo` in the same example. Split patterns means the recurring symbol shapes I use to describe many names in the same dataset or language family, such as whether names tend to lean on one-letter symbols like `N` and `S`, vowel-heavy endings like `I` and `O`, or repeated adjacent pairs like `I` + `N` and `C` + `H`. Hit rate is simply the share of names in a benchmark that parse successfully, while variants counts how many different valid element spellings a single name has.
Quick answer
Last updated March 31, 2026
Popular names do work with element symbols often enough to be interesting, but not often enough to assume your favorite name will parse cleanly. Romance benchmarks tend to be more flexible, German (DE) leads the country-level top-1000 hit rate, and in the country-level data every recorded failure ends up in one of two parser problems: the opening never matches a valid symbol, or a promising split dies later.
- Best top-1000 hit rate
- 20.1%
- German (DE)
- Top-1000 range
- 15.7% to 20.1%
- Across GB, DE, ES, FR, IT, NL
- Most common failure
- 51.9%
- Fails at the opening letters
- Benchmarks covered
- 6 countries + Welsh
- Country-ranked lists plus fallback
Methodology and caveats
These numbers come from Carbonat's ranked first-name benchmark lists, not from a universal census of every real-world name in a language. For GB, DE, ES, FR, IT, and NL, the benchmark preserves country rank order from the same upstream source already used by the names feature. Welsh is a repository-local fallback list in ranked file order, so it is informative but not directly comparable to the larger country-derived benchmarks.
That is why the safest wording here is “in our GB benchmark list” or “in our Italian benchmark” rather than “in the world” or “in every Italian name.” The value of the analysis is that every benchmark is measured with the exact same element parser and transliteration rules, so the differences come from the names themselves rather than from changing methods.
Here, a benchmark means one ranked list of popular first names that I test with the same parser from top to bottom. When I call Welsh a fallback, I mean it is a smaller repository-local list rather than a country-scale ranked source like the GB, DE, ES, FR, IT, and NL datasets.
Broader word-level spelling statistics live on the language info page. This page is strictly about names.
Headline findings
- Best top-1000 hit rate
German (DE) keeps the highest elementizable hit rate when you look across the full top 1000 names in our benchmark: 20.1%. That makes German the strongest country-level performer in this first pass.
- Most flexible matches
Italian (IT) does not lead on raw hit rate, but its matched popular names are the most flexible, averaging 1.62 valid spellings among top-1000 matches. That is where names like Francesco become especially interesting.
- Best long-tail lift
German (DE) improves the most as you move from the top 100 names (16%) to the top 1000 (20.1%). In other words, German's popular-name compatibility gets stronger once you move past the absolute biggest household names.
- Handle outliers carefully
Welsh shows a 100% hit rate in this artifact only because the fallback list is tiny: 11 ranked names, already curated for elementizability. On the other end, the Dutch benchmark drops from 18% in the top 100 to 15.7% in the top 1000, which makes it the strongest example of a long tail that becomes less parser-friendly.
How often popular names actually work
| Benchmark | Top 100 | Top 500 | Top 1000 | Avg. variants |
|---|---|---|---|---|
| German (DE) | 16% | 19.4% | 20.1% | 1.47 |
| English (GB) | 16% | 18.2% | 19.7% | 1.29 |
| Spanish (ES) | 19% | 19% | 19.3% | 1.47 |
| French (FR) | 17% | 17.8% | 19.2% | 1.35 |
| Italian (IT) | 17% | 17.8% | 18.7% | 1.62 |
| Dutch (NL) | 18% | 16% | 15.7% | 1.37 |
Welsh (fallback)
11 ranked first names in the benchmark
- Top 100 (11)
- 100%
- Top 500 (11)
- 100%
- Top 1000 (11)
- 100%
- Avg. variants among matches
- 1.18
German (DE)
5799 ranked first names in the benchmark
- Top 100 (100)
- 16%
- Top 500 (500)
- 19.4%
- Top 1000 (1000)
- 20.1%
- Avg. variants among matches
- 1.47
Spanish (ES)
5505 ranked first names in the benchmark
- Top 100 (100)
- 19%
- Top 500 (500)
- 19%
- Top 1000 (1000)
- 19.3%
- Avg. variants among matches
- 1.47
French (FR)
5494 ranked first names in the benchmark
- Top 100 (100)
- 17%
- Top 500 (500)
- 17.8%
- Top 1000 (1000)
- 19.2%
- Avg. variants among matches
- 1.35
Italian (IT)
5885 ranked first names in the benchmark
- Top 100 (100)
- 17%
- Top 500 (500)
- 17.8%
- Top 1000 (1000)
- 18.7%
- Avg. variants among matches
- 1.62
Dutch (NL)
5803 ranked first names in the benchmark
- Top 100 (100)
- 18%
- Top 500 (500)
- 16%
- Top 1000 (1000)
- 15.7%
- Avg. variants among matches
- 1.37
Popular names that fail, and why
The useful pattern is not that English fails one way and Italian another. Across all six country-level top-1000 benchmarks combined, every recorded failure in this dataset falls into two buckets. 51.9% of failures never get off the ground because the opening letters do not match any element symbol, while the remaining 48.1% start plausibly and then strand leftover letters that the periodic table cannot finish.
That makes this section more about parser geometry than about national quirks. The same two breakpoints keep reappearing whether the example is David, Jose, Marie, or Paul.
The name starts with letters that never form a valid symbol opening
2529 of 4873 top-1000 failures across GB, DE, ES, FR, IT, and NL
These names fail immediately. Their first one to three letters never produce a valid element symbol, so the parser cannot even begin a split.
- David#1English (GB)
Fails immediately after normalization to
david: none ofd, da, davis a valid opening element symbol. - Giuseppe#1Italian (IT)
Fails immediately after normalization to
giuseppe: none ofg, gi, giuis a valid opening element symbol. - Jan#1Dutch (NL)
Fails immediately after normalization to
jan: none ofj, ja, janis a valid opening element symbol. - Jose#1Spanish (ES)
Fails immediately after normalization to
jose: none ofj, jo, josis a valid opening element symbol.
The name begins plausibly but runs into a dead end later in the split
2344 of 4873 top-1000 failures across GB, DE, ES, FR, IT, and NL
These names do begin with valid symbols, but every possible path eventually leaves a remainder that no element symbol can cover.
- Paul#2English (GB)
Gets partway through as
P + Au, then stalls onl. - Peter#2Dutch (NL)
Gets partway through as
P, then stalls oneter. - Thomas#2German (DE)
Gets partway through as
Th + O, then stalls onmas. - Hans#3Dutch (NL)
Gets partway through as
H, then stalls onans.
Why the countries differ
Vowel-heavy endings help Romance names
The benchmark data lines up with the broader shape analysis I already measured in Carbonat's name datasets. Italian and Spanish still look vowel-heavy overall, with 83.9% and 66.3% of names ending in vowels in our larger dataset pass. That helps explain why those countries keep surfacing names like Francesco, Francisco, and Cristina as strong matches.
Long tails do not behave the same way
German leads the top-1000 hit rate not because every top name works, but because its compatibility stays resilient deeper into the benchmark. Dutch moves in the opposite direction, which suggests a long tail that becomes less element-friendly even though some very high-ranked names, such as Henk and Frank, still work early.
Normalization keeps accented names searchable
Accented spellings do not disappear from the analysis. Carbonat normalizes diacritics and a few common transliterations before testing whether a name can be written with elements, so names like Acuña and Iñaki remain searchable even though the parser ultimately works on normalized letter sequences.
What language family seems to change
One small pattern also shows up if you zoom in on the boundary letters inside split pairs. Across the current dataset, the most common joins are i + n, o + n, a + n, and c + h. In practice that means many successful boundaries end with a vowel-like symbol and then reopen on N or I, while clusters like ch often survive as adjacent symbols too. So the joins are not random: they reflect both the periodic-table inventory and the spelling habits of the names that make it through the parser.
Germanic profiles stay more consonant-heavy
Across our 7,568 Germanic names (`en`, `de`, `nl`), only 10.3% start with a vowel. The most common initials are B, S, and K.
The strongest symbols overall lean toward N, S, and K. The most common split pairs are I + N and C + H, which shows how often those consonant-heavy shapes stay adjacent.
That profile still produces useful splits, but it is less flexible overall at 1.44 variants per name on average.
Romance names give the parser more room
In our 8,364 Romance names (`es`, `fr`, `it`), 13.1% start with a vowel and 66% end with one. The opening symbol is most often C.
The overall split inventory leans heavily on I and O. The most common split pairs are N + I and I + N, which fits the way vowel-friendly segments keep chaining together.
That helps explain why the family averages 1.61 variants per name and 37.6% alternative splits.
Welsh is still a cautionary sample
Welsh is harder to generalize because the current sample is tiny and curated, not a full country-ranked benchmark. In this fallback set of 21 names, starts cluster around P and B.
The opening symbols Be and Pr are unusually prominent, while pairs like Y + S show how different the tiny sample can look.
The safe read is still “interesting shape, limited evidence,” not a broad claim about all Welsh names.
Germanic
en, de, nl · 7,568 names
- Avg. variants
- 1.44
- Alt splits
- 32.3%
- Starting with vowel
- 10.3%
- Ending in vowel
- 29.2%
- Top split pair
- I + N
Most common initials
- B (14%)
- S (12.4%)
- K (9.1%)
- C (8%)
- H (7.5%)
Most common opening symbols
- S (10.4%)
- B (9.4%)
- K (8.4%)
- C (5.8%)
- H (5.8%)
Most common symbols
- N (9.5%)
- I (8.6%)
- S (7.9%)
- O (6.6%)
- K (5.3%)
Most common split pairs
- I + N (2.4%)
- C + H (1.8%)
- O + N (1.6%)
- N + I (1.3%)
- I + S (1.2%)
Romance
es, fr, it · 8,364 names
- Avg. variants
- 1.61
- Alt splits
- 37.6%
- Starting with vowel
- 13.1%
- Ending in vowel
- 66%
- Top split pair
- N + I
Most common initials
- C (15.9%)
- B (12.3%)
- S (9.7%)
- P (8.4%)
- F (7%)
Most common opening symbols
- C (11.8%)
- B (9.2%)
- S (8%)
- P (6.7%)
- F (5.7%)
Most common symbols
- I (11.8%)
- O (10.2%)
- N (8.7%)
- C (6.5%)
- S (6.4%)
Most common split pairs
- N + I (2.3%)
- I + N (2.2%)
- O + N (2%)
- C + O (1.9%)
- N + O (1.8%)
Welsh / Brythonic fallback
cy · 21 names
- Avg. variants
- 1.24
- Alt splits
- 19%
- Starting with vowel
- 9.5%
- Ending in vowel
- 23.8%
- Top split pair
- Y + S
Most common initials
- P (23.8%)
- B (19%)
- A (9.5%)
- C (9.5%)
- R (9.5%)
Most common opening symbols
- Be (19%)
- Pr (19%)
- Al (9.5%)
- C (9.5%)
- Dy (4.8%)
Most common symbols
- S (14.9%)
- N (13.5%)
- Y (10.8%)
- I (9.5%)
- O (6.8%)
Most common split pairs
- Y + S (9.4%)
- O + N (7.5%)
- Er + I (3.8%)
- I + O (3.8%)
- I + S (3.8%)
Names that recur across multiple country lists
I avoid saying “most common names in the world,” but there is still a useful cross-country angle: some names keep reappearing across several ranked country lists and still parse cleanly into chemical symbols. These are the names that travel best across our current benchmark set.
- 1 variant
Luca
Appears in 6 benchmark lists; best rank #9
DE, ES, FR, GB, IT, NL
- 1 variant
Laura
Appears in 6 benchmark lists; best rank #13
DE, ES, FR, GB, IT, NL
- 2 variants
Alessio
Appears in 6 benchmark lists; best rank #58
DE, ES, FR, GB, IT, NL
- 1 variant
Barbara
Appears in 6 benchmark lists; best rank #67
DE, ES, FR, GB, IT, NL
- 2 variants
Filippo
Appears in 6 benchmark lists; best rank #80
DE, ES, FR, GB, IT, NL
- 2 variants
Nina
Appears in 6 benchmark lists; best rank #121
DE, ES, FR, GB, IT, NL
Are any elementizable names palindromic?
Yes, but they are rare. In the current elementizable-name dataset I only found 57 unique normalized palindromic names at all. They are constrained twice over: first, the letters have to mirror cleanly from left to right and right to left; second, that mirrored sequence still has to break into real element symbols rather than just looking like a word palindrome.
First names are slightly more common than surnames in this corner of the data, but they also skew shorter and more toy-like. I found 38 unique palindromic first names, with an average normalized length of 3.6, including Reinier, Nalan, and Natan. Surnames are a little rarer but a bit longer and more structural: 31 unique palindromic surnames, averaging 4, including Renner, Serres, and Staats.
The overlap is small too: only 12 normalized forms show up in both buckets, mostly very short shapes such as alla, bab, bob, esse, kik. That fits the general pattern: once names get longer, it becomes much harder to satisfy both palindrome symmetry and element-symbol segmentation at the same time.
Palindromic first names
38 unique normalized matches; avg. length 3.6
Palindromic surnames
31 unique normalized matches; avg. length 4
Can you find any more? Try the names browser with your own hunches, or jump into the designer if you discover a palindrome that looks good as tiles.
What to do with this
If your goal is to test a real first name or surname, the fastest next step is the names browser, where you can search examples that already parse. If you already know which split you like, continue into the designer to customize and download it.
If you are more interested in the symbol inventory itself, the interactive periodic table shows the building blocks, while the word-level language statistics page compares broader word-level behavior beyond names.
If you discover a name, poem, or sentence that feels especially good as element tiles, you can now send it to the community submissions page so it can be reviewed and, once approved, added to the shared public corpus for other people to like and share.
And yes, in a mildly shameless but hopefully still charming way, a name that splits nicely can make a genuinely fun gift. If you end up with a layout you love, putting it on a t-shirt or another item is a fairly reliable route to a unique geek gift for someone who enjoys chemistry, wordplay, or both.