The activation chants (聖詠) for the Symphogear relics are somewhat peculiar. If you've heard them, you likely wondered what was being said. If you've read them (the "lyrics" were provided in the booklets accompanying the season 2 OST discs), you're probably outright bamboozled.

  • Ame-no-Habakiri: Imyuteus amenohabakiri tron [OST 1 #10]
  • Gungnir (Hibiki): Balwisyall nescell gungnir tron [OST 1 #3]
  • Gungnir (Maria): Granzizel bilfen gungnir zizzl [OST 1 #6]
  • Ichaival: Killter ichaival tron [OST 2 #3]
  • Airgetlam: Seilien coffin airget-lamh tron [OST 3 #7 (Serena) / OST 6 #6 (Maria)]
  • Shul Shagana: Various shul shagana torn [OST 5 #5]
  • Igalima: Zeios igalima raizen tron [OST 6 #1]
  • Shen Shou Jing: Rei shen shou jing rei zizzl [OST 5 #5]
  • Also, the Zesshou (Superb Song / Swan Song / Climax Song):
    Gatrandis babel ziggurat edenal
    Emustolronzen fine el baral zizzl
    Gatrandis babel ziggurat edenal
    Emustolronzen fine el zizzl

Now, these don't look like any actual language (though I obviously can't claim to know all languages that exist). Yet, sites like the Symphogear wikia claim (or at least have, in the past, claimed) to translate the chants - for example, Ichaival's chant was given the translation "Evening draws in the awakened Ichaival".

Is it genuinely possible to translate the chants? If so, from what language or languages are they translated? (In the context of the show, Sumerian or some other old Mesopotamian language would make the most sense, because of Finé's connections to the region.)

Auxiliary question: if, as I suspect, they are not translatable, by what citogenetic process did these putative translations come into being in the first place?

7 Answers 7


I haven't watched this show, but I used to be a linguistics nerd, and I'm highly skeptical of the claim that these chants are written in actual Sumerian. I think they're just a mishmash of English and English-looking nonsense.

First, the case against the Sumerian theory. Disclaimer 1: I'm assuming that these passages are appearing as the creators intended and we don't have a case like the disaster in the Del Rey release of Negima Volume 6, where we had appendices full of complete nonsense being passed off as Greek because the translators had tried to transliterate the Greek into the Roman alphabet based on Akamatsu's transliteration of the Greek into katakana. Disclaimer 2: I'm no expert on Sumerian. But these passages don't pass the sniff test.

The phonology doesn't match the postulated Sumerian phonology

Let's look at a passage of transliterated Sumerian:

An adab to Bau for Luma (Luma A)
1. dumu an-na an gal ki gal-ta šag4 kug-ga-ni im-mi-in-pad3-de3
2. nam-nin kalam-ma-kam bi2-in-tum2-en
3. dba-u2 dumu an-na an gal ki gal-ta
4. šag4 kug-ga-ni im-mi-in-pad3-de3
5. nam-nin kalam-ma-kam bi2-in-tum2-en
6. den-lil2-le e2-kur za-gin3-ta
7. ki-sikil ama dba-u2 igi zid mu-un-ši-in-bar
8. /en\ dnin-ĝir2-su gal-bi mu-un-na-an-du7
9. kur gal den-lil2-le e2-kur za-gin3-ta
10. ki-sikil ama dba-u2 igi zid mu-un-ši-in-bar

(From http://etcsl.orinst.ox.ac.uk/cgi-bin/etcsl.cgi?text=c.2.3.1&display=Crit&charenc=gcirc# at Oxford University's Electronic Text Corpus of Sumerian Literature. The transliteration is described here)

Notice what's missing? There are no initial clusters, like we see in granzizel and tron. There are no dipththongs, as we see in seilien and raizen. There is no f, as we see in bilfen and coffin, or v, as we see in various, or o, as we see in zeios and shou, or w or y, as we see in balwisyall. There seem to be no doubled l's, like we see in kellter, or c, as in nescell.

[...] syllable initial consonant clusters were prohibited by Sumerian phonotactics [...]

Aleksi Sahala, "Sumero-Indo-European Language Contacts", page 11.

(There is some argument on this point. Cuneiform is a mixed logographic/syllabic writing system—a little like Japanese, actually. Most sources state there were no initial consonant clusters, but Encyclopedia of Linguistics says that although the cuneiform system "precludes the writing of initial and final consonantal clusters, it is quite likely that Sumerian did have them." We know this is true of Mycenaean Greek written in Linear B, for instance—like English written in the Japanese system, a Greek word like khrusos is written in Linear B as kuruso.)

The vocalic system is quite simple: /a/, /e/, /i/, /u/. The possible existence of an /o/ vowel remains unsubstantiated.

Encyclopedia of Linguistics, page 1046


To the best of today’s mainstream science, Sumerian consists of the following consonants:
{ b , d , g , p , t , k , m , n , ĝ , l , r , h , s , z , š , () }

Ioannis Kenanidis, "Yet Another Suggestion about the Origins of the Sumerian Language", page 31

Because Sumerian is incredibly ancient—it's the first known written language—and so much of it has come down to us through Akkadian scribes, who kept using it after it was dead because they thought it was cool, and because of the difficulties with the writing system that I alluded to above, there's a ton we don't know about it. No one believes the reconstructed phonology is 100% correct, or even 90% correct. But what we see here is too far off base to be taken seriously as an attempt to write Sumerian according to a reconstructed phonology.

You can see where they cribbed from

Whatever language this is, it includes the English words various and coffin, as well as killter, which looks a lot like kilter, as in "off kilter". The word baral in the swan song looks like it comes from Hebrew בלל (balal), which means "to confuse" or "to mix" and is apparently the source of the word "Babel". It uses the English form of ziggurat; the Akkadian form is ziqqurat, while "ziggurat" comes from the Hebrew form זיגוראט (zygwrʼt).

That doesn't automatically disqualify it as a real language; after all, in the Australian language Mbabaram, dog means "dog". But, combined with the widespread use of English for mysterious and exotic things in anime, this makes me highly suspicious of any claim that the sentences are Sumerian or any other language.

Aside from that, the sentences conform remarkably well to English phonology. Almost all of these words are easily pronounceable to an English speaker. The ones that look a little weird, like zeios and rei, show some pretty clear Japanese influence.

Machine systems can't make heads or tails of it

I fed all the activation chants and the swan song to Google Translate and Translated.net Labs's language detector, both together and individually. I wasn't expecting it to translate them, but I was hoping it would detect what language they were.

As a side note, machine translation still sucks, but this kind of language detection is supposed to be pretty good; Russel and Norvig claim in Artificial Intelligence: A Modern Approach that computer systems can identify languages, even based on short texts like "Hello world" and "Wie geht es dir", with greater than 99% accuracy (AIMA 3ed., page 862).

None of these systems can detect Sumerian or Akkadian, but they can detect most modern languages, even some quite obscure ones. The general consensus was that the activation chants have a Germanic flavor (Google Translate pegged them as English, even after I deleted the words various and coffin, while another service took them as Frisian, a Germanic language closely related to English), while the swan song was taken as Indonesian. Google Translate insisted that both texts together were English, while Translated.net took both texts together as Sundanese.

For comparison, I took a passage from Shakespeare's Titus Andronicus and a few lines from The Who's "Call Me Lighting", shoved them together, and fed them to Translated.net:

so, so; now sit: and look you eat no more than will preserve just so much strength in us as will revenge these bitter woes of ours. marcus, unknit that sorrow-wreathen knot: thy niece and i, poor creatures, want our hands, and cannot passionate our tenfold grief with folded arms. this poor right hand of mine is left to tyrannize upon my breast; who, when my heart, all mad with misery, beats in this hollow prison of my flesh, then thus i thump it down.

hey little girl who's dancing so lightly, my xke is shining so brightly, the noose around us is slowly tightening, i'm gonna show you why they call me lightning

English, English, and more English. It defies my suspension of disbelief to think that if these passages were all the same language, there would be so much confusion about what language that was.

I have no idea what algorithm Google Translate uses; Translated.net seems to use some kind of nearest neighbors algorithm.

The "translations" are absurd

The translation of Ichaival's chant seems to have been removed from the wikia sometime between when this question was asked and now, which probably doesn't say much for its authenticity. But let's take a look at the one in the OP:

Killter ichaival tron -> Evening draws in the awakened Ichaival

Since all the chants end with tron, that has to be something that makes sense in every chant. Based on the context, I'm going to say that has to be "awakened"; it is an activation song, so it makes sense that they would all be "awakened". So Ichaival tron means "awakened Ichaival". So we can postulate that this language puts modifiers after nouns, like Spanish and Irish.

Then what does killter mean? It has to mean evening draws in. But this is linguistically absurd; it violates constituent structure. draws in is a verb, while evening is a noun. killter can't literally mean evening draws in in a real language (if it does, we've got the makings of a real whopper of a paper in theoretical syntax here—it'll be almost to theoretical syntax what those particles that traveled faster than light would have been to relativity).

killter therefore must mean either evening or draws in, with the other element implied or indicated in some way by the morphology. So we have a language where you can just say draws in, and everyone automatically assumes you meant evening draws in, or a language where you can just say evening, and everyone assumes you meant evening draws in. Some languages, at least in their poetry, do allow you to get incredibly terse like this. Still...I'm not buying it. It reminds me too much of Helge K Fauskanger's analysis of Auld Elvish from Bored of the Rings.


I'm not convinced these are passages from a real language, or even from a somewhat sensible conlang like the Hymnos language from the Ar Tonelico series. All my instincts are telling me they're just gibberish courtesy of the people who gave us a determination to fist. I wish I knew where the "translations" came from or why they spread around the online anime community, a community generally obsessed with the minutiae of dead languages, which surprised the Japanese production staff of Madoka with its determination to translate the runes.

  • @QthePlatypus Yes? Was I unclear? I meant that since almost all of them end with tron, then in this made-up language, whatever tron means, it must be a word that it would make sense to have in all of these chants. It also always comes after a character's name, so the other chants would be saying "awakened Gungnir", "awakened Airgetlam", etc.
    – Torisuda
    Jun 9, 2016 at 5:47
  • Sorry what I ment to say that 'tron' could be more grammatical. Such as a tense marker on the end of a sentence or a honorific for a name. Jun 9, 2016 at 5:51
  • 1
    @QthePlatypus Sure, it could be. That actually helps the argument I made here—then killter would have to equate to "Evening draws in the awakened", with tron maybe indicating the tense or being an honorific. That makes the purported translation even more absurd; even more meaning is being packed into the single word killter, and it looks even more like yanqui unicycle ramar rotoroot in the Bored of the Rings analysis. Unless the language is polysynthetic—maybe k indicates a time late in the day, ill indicates drawing in, and ter is third-person, or something.
    – Torisuda
    Jun 9, 2016 at 5:57
  • 1
    @QthePlatypus I was mainly arguing that it's not a real language, though. I'm not convinced that it's a decent conlang either, but I didn't argue very strongly for that position. Mostly it just looked nothing like Sumerian or Akkadian, and I wanted to argue against that idea.
    – Torisuda
    Jun 9, 2016 at 5:59
  • 1
    It is some 6 years since the answer, but I've come up across it and really can't keep silent: unfortunately, I have to kill your comment at least on killter part, simply because of existance of a russian word вечерело (vecherelo), which would literally mean 'evening draws in'. The problem might be not with the supposed language used in songs (which I do believe to be fake as well), but with the lack of expressive power of English language on that aspect.
    – Senna
    Aug 7, 2022 at 19:51

Inspired by Evilloli's excellent answer, I decided to figure out where these "translations" came from.

After clicking through a lot of diffs on the Symphogear wikia, I was able to trace most of the "translations" back to wikia user Kanade2000 - diffs: Ichaival, Ame-no-Habakiri, Gungnir. When asked "I couldn't help but notice you post the translations for the Transformation lyrics and the Zessho. If you could, could you tell me how you translated them?", the user replied "using old norse dictionary that I found in my city library.i also have a skandinavian one and a latin one".

This answer does not inspire confidence in the user's linguistic faculties, particularly when you notice that the same user also tried to change the romanization of 絶刀 zettou to "zetsukatana" (only a rank amateur would consider combining the on-reading zetsu with the kun-reading katana).

These, I think, we can chalk up to a buffoon on the internet.

The translation for the Zesshou, however, comes from wikia user Sylphfarn12 , and linked G keyword 24 (絶唱 "Zesshou") as a reference. However, nothing in the text of the keyword page has anything to do with the actual text of the Zesshou. I have no idea how this user came up with their "translation". Buffoonery seems likely.

(A different buffoon in the comments of the page indicates - with a startling degree of confidence - that "Its Basque or Malay can't remember which". Another user rules out Malay in a reply. So it must be Basque, right? Mission Accomplished!)

This, I hope, serves as an object lesson in the dangers of trusting wikias.


So I doubt this will be helpful at all, but given the ties to the Tower of Babel legend in the first season and the fact that the relics are supposedly from this pre-Babel civilization in-lore. We could arrive at the assumption that the chants are words from multiple languages pulled together to create a "harmony" that echoes the "Original Language" spoken by the relic's creators. This is not a real-world, this is definitive answer, but just based on what I know of series lore so far and the rather extensive response by Torisuda this makes the most sense to me. Gibberish, but gibberish on purpose.


The symphogear wiki actually has the translation for the zesshou. But if I had to say for certain what language it is, I think the most likely is Sumerian. Kadingir, the tower Fine used towards the end of first season, was said to be an ancient Sumerian word meaning gateway to the gods. Fine built the symphogear and the tower, so it would make sense if both involved the same language, and there the unified language lost to humanity could very well be Sumerian.

  • 2
    Sumerian definitely makes sense from an in-universe perspective (now that you mention it, I was able to find a reference to "zizzl" in this book on text p. 189 / PDF p. 356; it appears to be a personal name). But if the chants/zesshou are indeed in genuine Sumerian or something like it, I wonder whether the translations that show up on the wiki and elsewhere are actually translations - the overlap between "people who watch Symphogear" and "people who know Sumerian" has to be vanishingly small.
    – senshin
    Nov 29, 2015 at 1:34
  • 1
    I believe you're right that the word kadingir is Sumerian; dingir is Sumerian for "god", and ka has a couple different meanings in Sumerian, though I couldn't find "gateway" as one of them.
    – Torisuda
    Jun 6, 2016 at 4:50

to be honest I believe that the idea was that there use to be a common tongue so what the creator ended up doing was combining some random languages to make a form of common language to fill the gap left by the fact that we couldnt possibly know what a common language would even sound like


I'm pretty sure it's a mixture of multiple languages. The main one being Urudu. "Croitzal ronzell gungnir" In Urudu translates roughly to Calamity, Renaissance, and Guns. It makes the most sense out of anything else. Unless they're actually using Dovahzul, Tengwar, D'ni, Atlantian, or some other Tolkien-esque language that we're unaware of.


This is avery interesting topic, and i bet it's a PUZZLE the anime's lore is strongly related to the called "balal no noroi" or "the curse of balal/babel" so maybe the chants are a lot of different languages, modern, ancient and fictional, mixed together...

my educated guess for some, tron = forced activation, zizzl = natural activation... the names of the relics and some words for what they think of the relic...

for example, Maria says: seilen coffin = the coffin of Selena, the girl who died using that relic Airgetlam = the name of the relic Tron = forced activation of the relic made by the symphogear

the zesshou has that 'zizzl' word, which if i'm right, it's an activation without technology... as in kanade's and maria's gungnir activation... so they could be using fine's power to increase their own, or just the power of the curse of babel...

certainly all of this discussion is a mystery, but i believe it will be revealed in the anime at some point, since language and songs are very relevant to the lore...

Edit: more than "natural activation" i think zizzl is "wrong activation" of the symphogear, as if they are not supposed to use it, and they are not supposed to use the zesshou....

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .