26 Oct 2009

Searching for an etymology for Germanic *handuz 'hand'

First, let's get nonsense out of the way by letting a published author state the obvious about origins of the Proto-Germanic etymon *handuz 'hand' that are most implausible yet unfortunately popular among idle hobbyists online. In the words of A. Seidenberg in km, a widespread root for ten (1976):
"The effort to relate km or kmt to *handus, or, more generally said, to see a reference to the hands in the number words, is also ad hoc: there is not the slightest evidence, apart from similar speculations on the other numbers, that the Indo-European number words are derived from finger-counting."
These comments on poor methodology are as true today as they were then, regardless of whether this old tomfoolery is resurrected on page 316 of Mallory/Adams in The Oxford introduction to Proto-Indo-European and the Proto-Indo-European world (2006 ), albeit subsequently with mild argumentation against the idea.

What then is the current etymology? Apparently no consensus exists yet. For example, The Barnhart dictionary of etymology‎ (1988) says that no cognates of hand exist outside of Germanic. While it's immediately tempting to see an origin in PIE *gʰend- 'to grasp' which yielded Latin praehendere, Greek χανδάνειν and Gothic bi-gitan, formal sound correspondences between PIE and Proto-Germanic forbid us to assume a direct connection with the Germanic root. One would expect a hypothetical PIE u-stem **gʰóndus 'grasper; hand' to end up as **gantuz in Proto-Germanic but certainly not *handuz which rather suggests a non-existent PIE stem **kondʰ-u-. Evidently, these are not the plosives we're looking for and no direct link to Proto-Indo-European appears sensible.

So I had a sudden brainwave and the more I think about it the more sense it makes, although it's frustratingly hard to substantiate. Since there are already a few known Proto-Germanic terms borrowed from Latin in the early first millenium BCE after Grimm's Law had taken place (cf. Ringe, From Proto-Indo-European to Proto-Germanic (2006), p.296), it makes me wonder if one of them might have been our Germanic word in question.

For this crazy idea to stick, we require a Latin word *handus, but as the reader can tell by the asterisk, it doesn't exist (at least as far as I know). On the other hand, prae-hendere 'to seize, to grasp' does indeed exist and the prefix prae- 'before' is secondary. From this implied Latin verb root *hend-, we are certainly free to muse light-heartedly on how we can obtain *handus 'grasper' from it, and very curiously, noting on how it rhymes with the attested Latin word for 'hand', manus. It taunts me with the image of a northern Germanic community with a high degree of Latin bilingualism, inventing new words and idioms out of a faraway language. If only my Germanic-influenced Latin word *handus for proper manus were attested in Roman records, I might develop something more out of this thought.


  1. It seems odd to assume a Latin loanword for a word so basic as 'hand'.

    Do we know of any precedent in another language?

  2. That's not a valid linguistic argument because you mistake a tendency for an absolute.

    When Roger Lass states, "But loanwords are less common in 'core' areas of the lexicon like the names of body-parts, numerals, kinship terms, or grammatical categories like inflectional endings." (Historical linguistics and language change (1997), p.105), notice he says "less common" not impossible. Naturally so, since if he had said it was impossible, he'd have egg on his face if a student rose up and confronted him with things like Japanese ichi, ni, san, etc. (from Middle Chinese), English they and are (from Old Norse) or Proto-Kartvelian *šwid- 'seven' (cf. Akkadian). Mind you, I've always suspected the origins of Latin caput as well.

  3. Of course very true. But a word like 'hand' seems even more unlikely than numerals, though arguably, maybe more likely than pronouns, which has happened in English.

    It's just, if such a core word was loaned into the Germanic languages, I would prefer to see a lot more extensive loaning from Latin in Germanic at such an early stage.

    Sure in Dutch and German there's quite a few very early loanwords like the word for 'horse' paard/pferd from Middle Latin paraveredus and kasteel 'castle' from castellum (or maybe an early french dialect without vocalisation of the l). But still these early loanwords from latin (there must be a few more) are usually limited to west-germanic. This word hand must have spread to the whole of the Germanic language family.

    It's not impossible, but I would expect to find a lot more Latin in proto-germanic.

  4. *Handus and manūs are not poetically compatible. At the time period you're referring to, Latin poetry was based on syllable weight and Germanic poetry (if Old English is anything to go by) was based on alliteration.

  5. PhoeniX: "It's just, if such a core word was loaned into the Germanic languages, I would prefer to see a lot more extensive loaning from Latin in Germanic at such an early stage."

    Yes, exactly, and I've also thought about this today. I have to accept this problem and deal with it. So...

    Let's adapt and try a variation on this idea. Let's say instead that Latin (pre-)hend- 'grasp' diffused into the Germanic community (cf. OEng. hentan, ON henda, Gothic fra-hinthan) from which *handus 'grasper; hand' was then based. In this way, we get around the tricky issue of a core word being borrowed because *handus is now explained as an innovation within the language that just happens to be based on a foreign non-core verb root.

    Sergei Andropov: "*Handus and manūs are not poetically compatible."

    I understand but I didn't really mean what I said literally. I apologize for my inexact wording. I was noting that Germanic *hand-u-s and Latin man-u-s have the same grammatical structure for a word with the same meaning. The shared structure of VERB-u-s makes them seem much like "rhymes", so to speak. Rhyme in a literal sense however is irrelevant to their etymologies, of course. I don't see how poetry has any bearing here.

  6. I would agree with PhoeniX - if "hand" were to be a early Latin loan in Germanic, I would expect to see many more Latin loans in Proto-Germanic. The borrowing of "they", "them", "their" from Old Norse into English accompanied fairly massive Old Norse borrowing (e.g. egg, husband, sky, etc. etc.). [Even then, the OE form of the third plural pronoun survives in Modern English "'em", as in "Get 'em".] Of course that doesn't mean it's impossible, but unlikely.

  7. Poetry is the reason we care about, or are really even aware of, rhyme.

    There is, however, a larger issue, which is that of timing. Proto-Germanic was undergoing several major changes, including mora-loss, at the period when first contact was established with Rome, and was also simultaneously splitting into Northwest and East Germanic. This means that we only have about a 50-150 year period with which to work, and I'm not at all sure that a) contact at that time was of such a nature that a word meaning either "grasp" or "grasper" could plausibly have been loaned and b) in the event that "grasp" was loaned that an agent noun could be formed out of it, lexicalized, and supplant the existing word for "hand" across the board in time for the Gothic migration, mora-loss, etc.

  8. Hey Glen, I like the idea of taking the whole Latin loanword, and making the word 'hand' an internal construction in Germanic. That solves the problem that we don't find Latin *handus.

    Erhm but help bue out here. Gothic fra-hinþan What happened to *d here. That's not supposed to end up as *þ. And actually I don't understand the *t in OEn. hentan either.

  9. These are very good arguments. Yet this just makes me feel stretched between two implausible positions without a clear solution.

    If one decides that Germanic *handuz is from PIE *kont-, the semantics that gets us from 'ten' to 'hand' is unjustified. Where else is such a verb root, whether connected to the numeral or not, attested? Given these facts, this is ad hoc, pure and simple.

    If one decides that it was borrowed, there's no more obvious and more direct source than Latin (pre-)hend- I can think of currently, yet as we all agree, it has it's share of problems.

    Yet again, if *handus is somehow borrowed from a hidden language other than Latin, how might we ever know?

    So far, I'm stumped. I guess I'll just have to take a break from it, have a coffee and come back to the issue anew one day. It's a little frustrating that there's little information to go on too.

  10. Phoenix: "Erhm but help bue out here. Gothic fra-hinþan What happened to *d here. That's not supposed to end up as *þ. And actually I don't understand the *t in OEn. hentan either."

    Douglas Kilday on the Cybalist forum looks like he's going into length about this same group of words. I'm currently reading it and assessing.

  11. Hi Glen!

    I don't know if this helps, but I've just found out that Vladimir Orel (in his "A Handbook of Germanic Etymology", 2003) derives PGmc */xanđuz/ from */xenþanan/ (cf. Lithuanian /rankà/ "hand" ~ /renkù/, /riñkti/ "to gather") > Gothic /fra-hinþan/ "to take captive, to capture" and Old Swedish /hinna/ "to gain", suggesting Greek /kentéo/ "to
    prick, to goad" might be related. He also seems to mention Latin /scando:/ "to climb, rise" in this connection, but my copy of his book is almost illegible, so I might be wrong...

  12. Peťusek: "[...] I've just found out that Vladimir Orel [...] derives PGmc */xanđuz/ from */xenþanan/ (cf. Lithuanian /rankà/ "hand" ~ /renkù/, /riñkti/ "to gather") > Gothic /fra-hinþan/ "to take captive, to capture" and Old Swedish /hinna/ "to gain", suggesting Greek /kentéo/ "to
    prick, to goad" might be related."

    Getting from "to grab" to "grabber" then to "hand" is perfectly reasonable. However the semantics of PIE *ḱent- seem to converge on 'to stick, to stab'.

    This now breaks the connection to a meaning of "hand" since surely a derivative meaning "stabber" would more likely become something like "knife". Close perhaps but no cigar. And I don't want to mix *skand- (note with *k not *ḱ!!) and "to rise" into this, yet again an unrelated verb root and sememe.

    What does that leave us with? It leaves us with a suspiciously parallel development between Gothic fra-hinthan and Latin pre-hendere that taunts us to believe PGmc *handuz and the related verb was somehow in some way influenced in the 1st millennium BCE by an Italic language.

  13. There would indeed be nothing surprising about such an etymology for ‘hand’. We can find other "handy" words in this line: PGmc *gripa- ‘part of hand which holds something’, easily comprehensible from ‘to grip’, and PGmc *hneba- ‘fist’, seemingly from PIE *knep-.

    But exactly the many words in this line make me think that the etymology of ‘hand’ might contain another aspect of the hand. If this word originally meant ‘open hand’, as opposed to the fist, the pointy fingers of an open hand might explain why PIE *ḱent- can be relevant for the etymology.

    It is also feasible to make another interpretation from PIE *ḱent-: PGmc *hinþan- ‘to pin down, hunt, reach, etc.’ bridges the way to hand-relevant functions good enpugh for me.

    Thus, a borrowing from Latin seems unnecessary.