A survey of international auxiliary languages

Disclaimer: since creating this table, I have become involved heavily in the Pandunia language. I haven’t edited it very much since then, but be warned that my bias may affect the presentation somewhat.

The plan is simple: construct a language that’s easier to learn than any natural one, and get everyone in the world to speak it. The idea dates back at least two centuries. It’s a noble endeavor; a universal second language of humanity would greatly aid international communication and collaboration. So it’s no surprise that it’s been attempted hundreds of times.

With the advent of the internet, the possibility of an international auxiliary language has become simultaneously more and less feasible. On the one hand, sharing an auxlang with the world could scarcely be easier; all you need is a domain and a decent website design. On the other hand, the explosion of choices makes it much harder for any one to reach exclusive prominence. When it comes to auxlangs, there can, after all, only be one.

For newcomers to the idea who think that an international auxiliary language is worth pursuing, it can be hard to make sense of all the options out there. Too often, enthusiasts will see Esperanto as the only choice; maybe they’ll have heard of Ido or Interlingua, but very rarely is a non-conlanger able to name more than two. Even if one knows about the much larger plethora of auxlangs out there, it’s hard to get an idea of how they all compare, since each auxlang only ever compares itself to natural languages and maybe Esperanto. I think that the difficulty of gauging the state of the art is part of the reason there are so many auxlangs. It’s all too easy for someone to be turned off by the first auxlang they meet, decide they don’t like it, and resolve to make their own. That’s how I got into language construction, anyway.

To minimize this danger, I’ve compiled this table of extant auxlangs. If you like the idea of learning an auxlang but don’t know which one to learn, I encourage you to use this list to make an informed decision. For each language, I note several key features that can be used to characterize them at a glance. These are the features that I think are most important in assessing an auxlang. There are explanations of what exactly each one means below the table in case it’s not intuitive.

This list is a lot shorter than some similar ones out there because I collected more information about each one than its name. That filters out a lot of the less well-developed ones. If you know an auxlang but don’t see it on the list, it’s probably because I couldn’t find a complete reference grammar and dictionary for it in a language I understand. It may also be because I just missed it. If that’s the case, or if you spot an error, then by all means, leave a comment to let me know.

Without further ado, the table (if it’s hard to read embedded, you can view it in Google Sheets here)!

The Name column is self-explanatory. I romanized all of them and dropped their diacritics since that’s how language names are usually anglicized.

The Years were ambiguous for a lot of them. I picked arbitrary years during their respective creations to put them in order, since precise dates aren’t terribly important.

The Articles column is the number of Wikipedia articles in that language as of 2019. Most of them do not have their own Wikipedias and thus have “0”.

For the Phonologies, rather than trying to format complete lists or IPA tables, I took advantage of the fact that most auxlang (and natlang) inventories are pretty similar and only listed statistically significant features. I define statistically significant as different from at least 60% of natural languages, weighted by number of L1 speakers. In addition, rather than focusing on the presence and absence of specific phonemes, which would draw attention to relatively meaningless attributes like whether an auxlang has /ʒ/ or /d͡ʒ/, I noted the presence and absence of important distinctions. So Esperanto has “ʒ/d͡ʒ” in that column to indicate that it has both, even though 81% of humans’ native languages lack at least one, and “n~ŋ” to indicate that it only has only the former, even though 66% of humans’ native languages have both.

The Orthography column starts by naming the writing system used, usually Latin or Latin with diacritics. It then gives any additional features that make it more interesting than a pure alphabet. These can be digraphs, single phonemes that are represented by pairs of graphemes; cluster letters, single graphemes that represent pairs of phonemes (usually ⟨x⟩); or special rules, any other complication that retains the 1:1 mapping between spelling and pronunciation, such as “⟨g⟩ is lenited before front vowels” or “/k/ is spelled ⟨q⟩ before /w/”. Finally, any further features are broadly grouped into “ambiguous spellings” if they make spelling dependent on etymology or “ambiguous readings” if they make pronunciation dependent on etymology. I don’t count it as an ambiguous spelling if personal names keep their spellings from their languages of origin just because that particular ambiguity is so common.

In the Vocab source column, I list all of the languages from which the auxlang draws its core words. Where there were more than ten, I just wrote “Worldlang” to save space.

The Toponyms column notes the word for “Japan” in each language, as a case study for how it handles cultural loanwords. Most will either have some approximation of Japanese /nippoɴ/, some version of English /d͡ʒəpæn/, or some orthographic rendering of ⟨Japan⟩ that starts with a /j/.

For Morphosynthesis, I use three examples to summarize how many compound words each language has: “September”, “today”, and “when”. The former is an interesting example because the word that most Western languages use has a misleading etymology—it’s the ninth month, not the seventh—so it draws a divide between those that use the European form and those that derive a new compound. The second is a more common example of a similar phenomenon—most languages have an indivisible root for it that ultimately derives from a much older compound. The third is an even more common word that might be a compound; generally only the most polysynthetic auxlangs won’t have a separate root for it, though Esperanto effectively compounds it with its correlative system. In all cases, hyphens are inserted to make it clear what is and isn’t a compound.

One axis on which Esperanto starkly differs from most of the others is Antonyms. Most languages will simply use separate roots for opposite meanings, but Esperanto and a few others add special affixes to existing words to make antonyms. Some others have both positive and negative affixes, where both of a pair of opposing words must have one or the other; I classify these as “class affixes”.

The Gender column calls out three triplets of words that help to identify how each language deals with gendered words: the neutral, masculine, and feminine forms of singular “they”, “person”, and “sibling”. Note that I made a distinction between neutral pronouns (English singular “they”) and neuter pronouns (English “it”); the latter is used to exclusively describe objects and thus does not apply persons of unknown or non-binary gender. Many auxlangs boast that they have no grammatical gender, but that’s a low bar. These triplets highlight how each language handles semantic gender, and makes it easy to see any asymmetries, redundancies, or lexical gaps.

Finally, the Inflection column lists the kinds of conjugation and declension on the verbs, nouns, and adjectives. Optional inflections are in parentheses, and I used “TAM” (tense, aspect, and mood) pretty liberally.

There you have it. I hope this table is a useful tool for thinking about the current state of auxlangs in the world. Always remember that there’s no such thing as a perfect language—at least, what’s perfect to one person likely won’t be perfect to very many others—and that all auxlangers share the common goal of bringing people together. Restu ebene, kamaradoj.

Edit: clarified my definition of “class affixes” versus “special affixes”.

Edit: added a link to the Google Sheets page.