A quick housekeeping note: Free Wednesday posts will now be coming out every other week, rather than every week. The depth (and length!) of these articles has gradually grown over the past year, and I’d like to make sure I can devote as much time as is necessary to research, write, and edit them properly. The paid Saturday articles won’t be changing in frequency, and will continue to have a lengthy free preview section.
Traces of the pre-Christian religion of the Anglo-Saxons hide in plain sight in the English language.
The most famous example of this persistence can be found in our days of the week, where the traditional names of Germanic gods lie half-hidden, disguised only slightly by the passage of time: the name Woden remains in Wednesday, just as the god Thunor (related to the Scandinavian Thor) is faintly remembered in Thursday.
These names survived the coming of Christianity because the missionaries who converted the Anglo-Saxons to the Cristian faith had a talent for integrating pagan (from their perspective) concepts and ideas into a Christian worldview.
When, for example, missionaries needed a word to express the Christian concept of a place of eternal punishment, they appropriated the word hell, which, judging from how the related Norse culture treated the concept, referred to a more general underworld,1 not a place intended for punishment.
Nowhere was this Christian repackaging more successful than with the word God itself.
Because Anglo-Saxon society was polytheistic before the advent of Christianity, the Old English word god2 referred originally not to one divine being but to a whole class of them — as it still can in Modern English, when we refer to the Greek gods, etc.
As this new, monotheistic concept of God came into the culture, the older word god was pressed into service. Along the way, it also switched genders: the original word god was neuter, while the Christian conception of God required masculine gender.
Similar things happened to the equivalent words in the other Germanic languages, where we find plenty of cognates with English god: German Gott, Swedish gud, Gothic guþ.
The existence of all these similar words might lead you to think that historical linguists understand the evolution of god, but nothing could be further from the truth. In fact, the word god has one of the most debated etymologies in the English language.
The history of the word is only clear as far back as Old English, where it was spelled and pronounced much the same way as in Modern English: god ‘(the Christian) God, (a pagan) god.’
Strangely, the word god seems to be composed of two parts: (1) a mysterious verb, and (2) the same suffix that becomes the Modern English -ed that we use on many verbs to form the past participle, as in cooked fish or toasted almonds.
This implies that god originally meant something like a [blank]-ed one or one who is [blank]-ed. The question is… how do we fill in the blank?
We have to go back to Proto-Indo-European to find our candidates, where we can find two verb roots that would roughly fit with the sound of modern word god. One of these roots meant ‘pour,’ and the other meant ‘call’ or ‘invoke’ (that is, ‘call upon in prayer or for protection’).
So, depending on which answer we choose, god would have started out as a word meaning ‘the poured one’ or ‘the invoked one.’ Right away, the latter option seems a bit more plausible, based on the meaning. Gods are, after all, often invoked.
But I wouldn’t be writing about this if everyone could agree on which root god descends from. One solution makes a lot of sense given its meaning, but presents a difficulty with its form. The other solution has the opposite problem.
To judge between the two, we’ll need to investigate ancient Greek ritual, measure the length of vowels in Sanskrit epithets, and consider what people thought lived in burial mounds. Let’s get started.
You’re reading The Dead Language Society. I’m Colin Gorrie, linguist, ancient language teacher, and your guide through the history of the English language and its relatives.
Subscribe for a free issue every Wednesday, or upgrade to support my mission of bringing historical linguistics out of the ivory tower and receive two extra Saturday deep-dives per month.
If you upgrade, you’ll be able to join our ongoing Beowulf Book Club. You can also catch up by watching our discussion of the first 1962 lines (part 1, part 2, part 3, part 4) right away.
Who ya gonna call?
The solution that makes the most sense based on meaning alone is to derive god from a Proto-Indo-European root meaning ‘call, invoke’.
This root has been reconstructed as *ǵheuH-, where the asterisk indicates that it’s a reconstruction rather than an attested word. The specific form of that root that would give us the word god is *ǵʰuH-tóm ‘invoked one.’3
There are some other strange symbols in the word *ǵʰuH-tóm, but the only one you need to worry about is the big capital H. This is a symbol for a laryngeal consonant, a mysterious class of sounds that we know must have existed in Proto-Indo-European, even though we have no clear idea of how they sounded.
One reason that we know that laryngeal consonants existed is because of their effects on the vowels that came before laryngeals in the daughter languages. The story of the discovery of laryngeals in Proto-Indo-European is actually a ripping good yarn, which I’ll tell you on another occasion. For today, all you need to know is that the root meaning ‘call’ has been reconstructed with a laryngeal.
Laryngeal consonants disappeared, but in disappearing, they made any vowel that came before them into a long vowel. By long vowels, I mean a set of vowels which are held for longer than their corresponding short vowels. Modern English doesn’t really have these, but older forms of English did. For example, here’s Middle English met ‘met’ vs mēt ‘meet’:
By the way, I mark long vowels by placing a line (called a macron) over the vowel, like this: ā ē ī ō ū.
The evidence for a laryngeal in the root *ǵʰeuH- ‘call, invoke’ is that the descendants of this root in other languages have long vowels. Now, it may seem like a small matter whether a given vowel is long or short, but historical linguists consider it a big deal. If your theory about the etymology of a word predicts a long vowel, and the word has a short vowel, that’s a problem for your theory!
With that in mind, take a look at the long ū in the Sanskrit word puruhūta- ‘much-invoked,’ an epithet of the god Indra. The long vowel in hūta ends up being a bit of a problem for tracing god back to the ‘invoke’ root, though, since god appears in the Germanic languages with a short vowel.
The Proto-Germanic ancestor of god is something like *gudą, with a short u vowel. For reasons that don’t need to distract us here, that short u becomes a short o in Old English: hence the form god.
But the laryngeal in the reconstructed root *ǵʰeuH- ‘invoke’ should instead produce a Proto-Germanic word with a long ū: *gūdą — just like the ū in the Sanskrit hūta ‘invoked’ — which would lead further to an Old English word gūd and a Modern English word goud (rhymes with loud), which is clearly not what god sounds like today.
For this reason, we have an alternative hypothesis.
Pour one out for the gods
The other alternative is to derive god from a different Proto-Indo-European root, one that looks almost identical to the ‘invoke’ root *ǵʰeuH-. This other root means ‘pour,’ and it’s been reconstructed as *ǵʰeu-. In other words, it’s identical to the ‘invoke’ root with the exception of the laryngeal.
The root *ǵʰeu- is well attested in the various branches of the Indo-European family. In Greek, it shows up in the verb χέω khéō ‘pour,’ leading to beautiful compounds like οἰνοχόος oinokhóos ‘wine-pourer,’ which shows up in Homer.
In Latin, *ǵʰeu- gets extended with some additional consonants and becomes fundō ‘pour out; melt’ — the fu- part comes from *ǵʰeu-. The verb fundō ends up in English words such as refund ‘pour back’ and fondue ‘melted.’
Most English words which come from fundō, however, are borrowed from the verb’s past participle fūsus ‘poured; melted’: e.g. fusion ‘melting’ and all its derivatives, diffusion, confusion, profusion, etc.
This same root actually shows up in Germanic as well, albeit with a consonant extension on the end, much like the Latin word fundō had: in Old English, the *ǵʰeu- root appears as ġēotan ‘pour,’ which doesn’t survive into Modern English (although, if it did, it would turn out as yeet – somehow I doubt that this is the etymology of the slang term yeet ‘throw’).
A derivative form does show up in the word ingot ‘poured in’, and, most probably, in the name of the Goths, ‘those who pour,’ although there are lots of mysteries remaining about that etymology.
At any rate, since the root *ǵʰeu- has no laryngeal, this means that a form like ‘poured one’ would be *ǵʰu-tóm (no H in the middle) rather than *ǵʰuH-tóm (with the H in the middle), and that ends up making all the difference. And we might even be encouraged to see that a form like *ǵʰu-tóm does seem to have existed, since we have its exact descendants in Greek χυτός khytós ‘poured, melted, heaped’ and Sanskrit hutá- ‘poured out, burnt, offered in fire, sacrificed’.4
This Proto-Indo-European form *ǵʰu-tóm would become Proto-Germanic *gudą, which becomes Old English god: exactly what we need!
Some scholars, however, have had trouble with the meaning of this root: how does ‘poured’ or ‘the poured one’ come to mean ‘god’?
At first blush, the two meanings seem rather distant. But let’s look again at the Sanskrit cognate hutá-, which is used to mean ‘poured out, burnt, offered in fire, sacrificed.’ Now, given that hutá is used to refer to things that have been burnt or offered as sacrifice, a divine connection doesn’t seem so distant. Gods are the kinds of things that receive sacrifices.
We also know that libations — that is, ritual pourings of liquids to the gods — were an important, and very ancient, part of ancient Greek religious practice. These practices likely date back to Proto-Indo-European times, as suggested by the existence of a Sanskrit word hótṛ and an Avestan (ancient Persian) word zaotar, which both mean ‘priest,’ i.e. the ‘one who pours’.
So, as it turns out, the meaning ‘pour’ is not too distant from ‘god.’ But there is still an eensy-weensy gap: how does ‘poured,’ which, as in Sanskrit hutá, refers to the poured sacrifice itself, come to refer to the entity to which the sacrifice is made?
There’s gods in them thar hills
The change from ‘that which is poured’ to ‘that for whom something is poured’ could just be an idiosyncratic shift in meaning. There are, of course, many idiosyncratic shifts in meaning in the history of language.
To use the word for a sacrifice to refer to the one receiving the sacrifice could be thought of as a kind of metonymy, where something is referred to using a term associated with that thing. Examples include the use of the Crown to refer to the institution of the monarchy, or the press to refer to the news media.
Another way to make the ‘pour’ etymology more plausible comes from the Greek cognate χυτός khytós ‘poured, melted, heaped.’ This word is used in Homer in the phrase χυτὴ γαῖα khytḕ gaĩa ‘heaped up earth’ to refer to a burial mound, or tumulus.
Given that the use of burial mounds is characteristic of many early Indo-European speaking peoples — and the Germanic-speaking peoples were certainly no exception — perhaps *ǵʰu-tóm referred not to pouring but to “the spirit immanent in the heaped up hallowed ground of a tumulus.”5
By now, you’ve may have developed a liking for one of the two theories over the other. Is the close connection between invocation and divinity convincing enough for you to disregard the matter of the unexpected short vowel in god and accept the root *ǵʰeuH- ‘call, invoke’? Or are you a stickler for making the phonetics match exactly, even if you need to account for some shifts in meaning to make the theory work? In that case, you’ll be drawn to *ǵʰeu- ‘pour.’
Or you may follow another scholar, Robert S. P. Beekes, and say that you find neither of these hypotheses convincing.
In an article “God is non-Indo-European,” Beekes argues that god was borrowed into the Germanic languages from some unrelated language spoken by peoples living in the area. Although without any independent knowledge of what these other languages were like, this becomes what doctors call a “diagnosis of exclusion”: something you say when nothing else fits.
In the end, perhaps it’s appropriate that a word referring to divine beings should be so mysterious.
But the quest for the origin of the word god leads us down several interesting paths which reveal something about how humans — at least those humans who spoke early Indo-European languages — interacted with the divine in the distant past: gods were beings that mere mortals called upon and made sacrifice to.
The name god, which has remained in the English language even as gods, plural, were replaced by God, singular, comes from one of these aspects of the human-divine relationship.
But which one? We may learn in the future, but, for now, at least, the answer to that question remains beyond our mortal ken.
Etymologically, hell means ‘covered’ or ‘concealed.’
I’ll be capitalizing God only when I’m specifically referring to the monotheistic concept of God.
If you’re wondering what happened to the e in the root *ǵʰeuH- when the -tóm suffix gets added, the answer is another long story.
The Greek adjective is given in its masculine singular form, and the Sanskrit in its stem form; technically, only their neuter singular forms would be the exact descendants of *ǵʰu-tóm, but this doesn’t really matter.
Watkins, Calvert (1974). god, Antiquitates indogermanicae. p102.
Wow, Beekes searches for substrate everywhere, huh? He's famous for his Greek etymological dictionary throwing out dozens of words to "Pre-Greek".
>"The story of the discovery of laryngeals in Proto-Indo-European is actually a ripping good yarn"
Definitely an example of the assertion that most sentences that people utter are unique :) :)