An Old Irish fart poem

At·tá ben is’ tír,
·eiprimm a hainm,
maidid eissi a deilm
amal chloich a tailm.

1   at· H1 3 sg ‘be’ (cf. modern Ir ), here with existential sense, ‘there is’—ben nom sg ‘woman’ (cf. modern Ir bean, Gk γυνή, Eng queen)—is’ = isin = i ‘in’ + definite article—tír neut s dat ‘land, country’
Continue reading


An ancient word for ‘earthling’?

Many Indo-European languages exhibit a word for ‘human being’ that is built off of PIE *dʰéɡ́ʰom- ‘earth, ground.’

  • Celtic *gdonyo- (Irish duine ‘person,’ Welsh dyn ‘man’) < *dʰɡ́ʰom-yo-
  • Germanic *guman- ‘man’ (Old English guma ‘man,’ English [bride]groom; Old High German gomo ‘man,’ German [Bräuti]gam) < *dʰɡ́ʰm̥-mon-
  • Latin homō, –minis ‘person’ < *dʰɡ́ʰm-on- (whence also hūmānus)
  • Phrygian ζεμελως ‘man (dat. pl.)’ < *dʰɡ́ʰem-el-o-

This underlying root is also well-attested with its original meaning.

  • Albanian dhe ‘earth’ < *dʰɡ́ʰom-
  • Celtic *gdon- (Old Irish , don ‘place’) < *dʰɡ́ʰom-
  • Greek χθών, -νός ‘earth’ (an English derivative of which is chthonic) < *dʰɡ́ʰom-; χαμαί ‘on the ground,’ χαμᾶζε ‘earthwards,’ χθαμαλός ‘near the ground’ < *dʰɡ́ʰem-el-o-
  • Latin humus ‘ground’ < *dʰɡ́ʰom-o-, humilis ‘near the ground’ < *dʰɡ́ʰem-el-i-
  • Sanskrit क्ष kṣa ‘field’ < *dʰɡ́ʰm̥-
  • Slavic *zemja (Polish ziemia; Bulgarian земя, Russian земля) < *dʰɡ́ʰem-yh̥2

Although Greek derives no explicit word for ‘person’ from this, ἐπιχθόνιος ‘on (ἐπί) the earth (χθών)’ is an occasional epithet of humans in Homeric verse:

“οὔ τινα γὰρ τίεσκον ἐπιχθονίων ἀνθρώπων,
οὐ κακὸν οὐδὲ μὲν ἐσθλόν, ὅτις σφέας εἰσαφίκοιτο” 415
“Not one of the people on earth did they respect—
not wicked and not noble—whoever came to them.” (Od. 22.414–415)

Sometimes it even occurs on its own, without ἄνθρωποι ‘people’ or βροτοί ‘mortals,’ and we are expected to fill that meaning in ourselves:

“εἰ μὲν γάρ τίς μ᾽ ἄλλος ἐπιχθονίων ἐκέλευεν, 220
ἢ οἳ μάντιές εἰσι θυοσκόοι ἢ ἱερῆες,
ψεῦδός κεν φαῖμεν καὶ νοσφιζοίμεθα μᾶλλον.”
“For if some other one of those on the earth had told me,
either those who are diviners or priests who make sacrifices,
we would rather call it a falsehood and distance ourselves from it.” (Il. 24.220–222)

The derived meaning of ‘person’ from ‘earth’ is not exclusive to these Indo-European examples. Hebrew, too, seems to derive אָדָם ɔādām ‘man, Adam‘ from אֲדָמָה ɔădāmâ
‘ground.’ (The underlying root ɔdm means ‘red.’)

Why these languages would have derived words for ‘human being’ from a word meaning ‘earth’ is an interesting question. I see several possibilities:

  1. the sense of earth-denizen is in opposition to sky-denizens, or the gods;
  2. there is a common myth in the Indo-European and Semitic mythologies (and elsewhere?) that humans were created out of clay or dirt;
  3. humans called themselves earthling to differentiate themselves from otherworldly visitors.

Seriously considering #1 and #2, I am inclined to think that #1 is the likeliest. To be sure, the notion that humankind was created from clay is a common belief across many world mythologies. On the Semitic side of things, in Genesis (2:7), Adam is created from the “dust of the ground”—just like in the Enûma Eliš—while the Qurɔān (55.14–15) mentions that God also created the jinn from smokeless fire. Thus the connection between ɔādām ‘man, Adam’ and ɔădāmâ ‘ground’ fits this theory very nicely.

Elohim Creating Adam 1795-c. 1805 by William Blake 1757-1827
And Elohim Created Adam by William Blake

But in these many Indo-European derivations from *dʰéɡ́ʰom- ‘earth,’ a parallel creation story wherein humans are wrought out of clay seems too good to be true; it would certainly be asking a lot of comparative mythology. For this reason I prefer #1. The Greek ἐπιχθόνιος ‘[living] on the earth’—as opposed to something like *ἐκχθόνιος ‘[sprung] from the earth’—in particular suggests that the noteworthy aspect of humankind’s relationship with *dʰéɡ́ʰom- is that it is our home. Whether or not it is also the ur-material of our bodily composition is not the central issue.

There are plenty of ancient Indo-European sources that drive home the opposition between earth-dwelling mortals and sky-dwelling immortals. Here is one of my favorite examples:

A monument (2nd cent. BC) with a dual-language inscription sums up in more concise Gaulish a dedication that has already been written in Latin:



Akisios Argantomaterecus gave the area (?) for gods-and-humans.

This TEUOX/TONION—for which David Stifter reconstructs the pronunciation dēwoγdoniyon—is the Gaulish equivalent of deis et hominibus ‘for gods and humans’ in the Latin text. The word is apparently a dvandva compound (which are not common in Celtic) that breaks down into dēwo– ‘god’ and γdoniyo– ‘human,’ with a plural inflectional ending. The first half is from Indo-European *deywó- ‘god’ (cf. the Germanic god *Tīwaz), which is probably an ablaut-variant of *dyew- ‘sky, day,’ and therefore related to *dyēw- ‘sky god’ (frequently with the epithet “Father”; cf. Ζεὺς Πάτηρ, Iuppiter, Dyauṣ Pitā, etc.). The second half of the compound, γdoniyo-, is our old friend *dʰɡ́ʰom-yo- ‘earthian, earthling,’ and therefore exactly cognate with Irish duine ‘person.’

Therefore this single word TEUOXTONION dēwoγdoniyon ‘for gods and humans’ perfectly encapsulates the opposition between sky-dwelling immortals and earthbound mortals. For these reasons it makes the most sense to me that, at least for the Indo-European terms, the original sense must have been ‘earth-dwelling’ (and not, say, ‘made from earth/dirt’). But perhaps the answer does not have to be one or the other.

And option #3 is always available.

Etymology: Irish fuinneog ‘window’

The Irish word fuinneog /ˈfɪnʲːoːg/ ‘window (fem.)’—uinneig in Scottish Gaelic; fuinneoig in the Irish of Cois Fharraige where the dative singular has been assumed as the dictionary form for ā-stem nouns—may appear obscure in origin. However, it has a neat etymology as an old loanword from Germanic.

The word is from Old Norse vindauga ‘window (neut.),’ a composition of vindr ‘wind’ and auga ‘eye’—that is, figuratively, a sort of eye through which the wind enters through. This word was also loaned into English as window; *wind-augō may have been a formation specific to North Germanic, since both Old English ēag-duru and Gothic áuga-daúrō reflect the composition ‘eye-door.’

The ending –eog (the equivalent of –óg when following a palatalized consonant) and the feminine gender that accompanies it is on analogy to other nouns in -óg—cf. féasóg ‘beard.’ It is a frequent additive to loanwords—cf. spúnóg ‘spoon,’ from English spoon.

We can surmise from orthographic tendencies in manuscripts that nasal+stop clusters (*mb, *nd, etc.) were becoming geminate nasals (*mm, *nn, etc.) broadly over the course of the Old Irish period. The central consonant in modern fuinneog would suggest that vindauga was loaned before or during this transitional period. Old Norse v would have had the pronunciation /w/ (hence the English pronunciation), allowing the first consonant of the ensuing Irish word to fit neatly into the extant mutational paradigms of inherited f < *w (although the initial f– was later lost in Scottish Gaelic on analogy to words beginning with Ø-; the opposite happened in fosgail ‘open,’ with f– springing into existence ex nihilo; cf. Irish oscail).

Дп Updдтэ foг Iгish Sрэlliпg?

Please forgive my frivolous title.

For some time now I have semi-jocularly been saying that the Cyrillic alphabet (particularly its Russian incarnation), what with its innate capability of marking palatalization, would be better suited to write Irish and Scottish Gaelic than the Latin alphabet. I will look at Irish alone, although most of this has equal validity for Scottish Gaelic.

Three major hurdles of Irish orthography—I will not go so far as to call them “problems,” because they are a large part of what gives modern Irish its unique look—are the following:

  1. marking palatalization (and the absence thereof) with “silent” vowels; a word like múinteoir ‘teacher’ looks as though it could contain as many as five syllables, or at least several diphthongs; in reality, the orthographic vowels i and e serve to inform the pronouncer that the intermittent consonant cluster nt is palatalized; the same goes for the second i, which is a mere marker of palatalization on the following r, to get bisyllabic /ˈmuːnʲtʲoːrʲ/
  2. retaining the pre-mutation consonant in spelling; when a word like peann /pʲɑːn/ ‘pen’ is preceded by a word like ár ‘our,’ which causes the nasal mutation known as urú, the result in modern Irish is spelled ár bpeann /ə bʲɑːn/ ‘our pen’; the post-mutation sound b is written in addition to the original first consonant p. This has an even more jarring effect with a capitalized noun like Corcaigh ‘(County) Cork,’ which when mutated appears in camel casei gCorcaigh ‘in Cork’
  3. miscellaneous archaic spellings; just as in English, the modern Irish lexicon proffers many “silent” letters that were once pronounced but have since been lost due to the all-encompassing entropic force that is sound change; examples from below are gaoth /giː/ ‘wind’ (Old Irish gáith /gai̯θ/) and aghaidh /aː/ ‘face’ (Old Irish agad /ˈaɣəð/)

Really the only one of these that could non-trivially be ameliorated with a different writing system is #1; with an extended set of vowel symbols that indicate whether the preceding consonant is palatalized, the Cyrillic alphabet would provide a more elegant solution to the problem that the Latin alphabet answers with extra vowels. #3 is an issue that crops up in any language in need of an orthographic reboot, and a departure from the Latin script would not be necessary to update historical spellings. (Relative to Scottish Gaelic, Irish has already done a fair amount of this; cf. Irish buí and Scottish Gaelic buidhe, both from Old Irish buide ‘yellow.’) #2, likewise, is entirely the prerogative of the language at hand. Welsh, for example, does not doubly write mutated consonants; despite having initial mutations that behave similarly to those in the Gaelic languages, Welsh writes only the mutated sound so that the spelling is phonologically transparent but etymologically opaque. This choice is available in any phonemic writing system. Below, I opt for the visually simpler alternative.

I am going to take at a swing at writing Irish in the Cyrillic alphabet. My text will be that infamous “Irish Blessing.” For dialect, I use the Irish of Cois Fharraige.

Go n-éirí an bóthar leat
(May you fare well)
/gǝ nʹ-ai̯rʹiː n boːr lʹæːt/
Го н-я́йри́ он бо́р лят

Go raibh an ghaoth go brách ag do chúl
(May the wind forever be at your back)
/gǝ ro n γiː gǝ brɑːx egʹ dǝ xuːl/
Го ро он ғы́ го бра́х эг до ху́л

Go lonraí an ghrian go te ar d’aghaidh
(May the sun shine hot on your face)
/gǝ lundriː n γʹrʹiːǝn gǝ tʹe erʹ daː/
Го лонры́ он ғри́он го те эрь д’а

Go dtite an bháisteach go mín ar do pháirceanna
(May the rain fall softly on your fields)
/gǝ dʹitʹǝ n wɑːsʹtʹǝx gǝ mʹiːn erʹ dǝ fɑːrʹkʹǝniʹ/
Го дите он ва́штех го ми́н эрь до фа́рькены́

Agus go mbuailimid le chéile arís,
(And until we meet again,)
/aːgǝs gǝ muːǝlʹǝmʹǝdʹ lʹe xʹeːlʹ ǝˈrʹiːsʹ/
Агос го му́олемедь ле хе́ле ори́ш,

Go gcoinní Dia i mbos A láimhe thú.
(May God keep you in the palm of his hand.)
/gǝ genʹːiː dʹiːǝ ǝ mos ǝ lɑːwʹǝ huː/
Го гэнни Ди́о и мос О ла́ве hу́.

Some notes:

  • For Irish schwa /ǝ/, I have used е when the preceding consonant is palatalized, and о when it is not, much in line with Russian practice.
  • When Russian came up short for distinctions that I very much wanted to make, I borrowed ғ /γ/ and h /h/ from the Kazakh variant of Cyrillic. (To use х for /h/ would have introduced ambiguity with /x/, which Irish also has.)
  • The Cyrillic accent—as in а́, etc.—in Irish marks vowel quantity (and sometimes also quantity), not word-stress as in Russian.
  • Irish /i(ː)/ presents something of a problem when, as in gaoth /giː/, the preceding consonant is not palatalized. For this function I have chosen Cyrillic ы, which in Russian does not palatalize the preceding consonant but also, problematically for my purposes, has a more central place of articulation than и. Despite this incongruity in Russian, I intend for ы to be identical to и in Irish—that is, equal to /i(ː)/—with the only exception that it does not palatalize the preceding consonant.

Го ро ма агы́. Сла́н!