Are there Phoenicians in phonology?

« previous post | next post »

A question from Mark Seidenberg:

Is the English phono- morpheme etymologically related to Phoenicia/Phoenician, i.e., the corresponding Phoenician words?

I have looked at the OED and other sources and I cannot connect the dots.

The Phoenician word for “Phoenicia” has a couple of conjectured etymological bases unrelated to sound or voice.

The Greeks then had a word (morpheme?) phono that was related to sound/voice, which English and other languages absorbed.

Is it a coincidence that the Greek word happened to sound like name for the language whose writing system they borrowed, the inadequacies of which for representing the typologically distinct Greek language led to the identification of vowels, which could then be written by repurposing letters for a few consonants that occurred in P but not G. or so they say.

Is this an homage to Phoenicia or are these false phonological cognates?

The answer, from Don Ringe:

It's an accident. The phono- part is Greek pho:né: 'voice', which is a noun derived (by an archaic, non-productive process) from phánai (stem pha-) 'to say'. 'Phoenicians' in Greek is Phoínikes; the singular phoîniks is also used to mean (1) purple (the dye the Phoenicians boiled out of murex snails and marketed at a fabulous profit); (2) date-palm; (3) the mythical phoenix bird. It looks like the ethnonym is the original meaning, and all the others are transferences, but that's not quite clear; in particular, there are some other words meaning 'blood-red' that begin with phoin-.

But whatever the etymology of Phoínikes is, the sequences pho:n- and phoin- are absolutely different in Greek, and there can't be any etymological connection between them. (There's also the fact that the -n- of pho:né: is clearly a suffix, whereas there's no evidence that phoin- is polymorphemic.)

Mark's question was by no means a silly one. There are many examples where similar-sounding words with disparate meanings really are derived from a common source — and in contrast, many other cases where similar-sounding words, even with apparently-related meanings, have completely independent histories. The notion of "cognate word" is not at all a transparent one.

For a striking example of (genuine but non-transparent) etymological evolution among similar-sounding words , consider English cane, cannon, canon, canal, canyon, channel, etc. (I first heard this word-evolution example used in a talk 20 years ago by David Searls — you can see some other examples of his thinking about words and genes in "From Jabberwocky to genome: Lewis Carroll and computational biology", Journal of Computational Biology 2001, or in "Grammatical representations of macromolecular structure",  Journal of Computational Biology 2006.)

The OED's etymology for cane, n.1:

Middle English canne, cane, < Old French cane, later canne (= Provençal cana, Spanish caña, Italian canna) < Latin canna, < Greek κάννα, κάννη, reed, perhaps < Semitic: compare Hebrew qāneh, Arabic qanāh reed, cane. In Latin the sense was extended from ‘(hollow) reed or cane’ to ‘tube or pipe’, a sense retained in Romanic, and prominent in the derivatives canneau, cannella, etc.

For cannon, n.1:

In 16th cent. also canon , Scottish cannoun , < French canon (14th cent. in Littré) = Provençal canon , Catalan canó , Spanish cañon , Italian cannone , lit. ‘great tube, barrel’, augmentative of canna , canne cane n.1, reed, pipe, tube. The spellings canon and cannon occur side by side down nearly to 1800, though the latter is the more frequent after c1660.

For canon, n.1:

Found in Old English as canon , < Latin canon rule, < Greek κανών rule. Early Middle English had ˈcanon , probably < Old English, and caˈnun , caˈnoun , < Old French canun , canon , the French descendant of the Latin.

[LSJ gives the literal meaning of κανών as "straight rod, bar, esp. to keep a thing straight", which I'm following David Searls in assuming to be related to κάννα, κάννη, reed. This might be a mistake, in which case the example covers both real and false etymological connections :-)]

For canal, n.:

< French canal (16th cent. in Littré), a refashioning, after Latin canāl-em or Italian canale , of the earlier French chenal(chanel , chenel ): see cannel n.1, channel n.1 (The 15th cent. instance may be from Latin.) The words canel , cannel n.1, and chanel , channel n.1, from the same Latin source, but immediately from old French, were in much earlier use in English: when canal was introduced it was to some extent used as a synonym of these, but the forms were at length differentiated.

For canyon, n.:

A phonetic spelling of Spanish cañon, designed to represent the proper spoken word: compare canion n.

cañon, n.3:

< Spanish cañon tube, pipe, conduit, barrel, cannon, etc. (augmentative of caña < Latin canna reed, pipe, quill, cane n.1; thus the same word as Italian cannone , Portuguese canhão , Provençal canon , French canon , English cannon , and canion ), but spec. applied by the Spaniards of New Mexico in the sense in which it has been adopted from them by their English-speaking neighbours. In order to retain the pronunciation and prevent confusion with canon , which would result from the frequent want of the Spanish letter ñ, ñ (enye ), in English typography, the word is frequently spelt canyon n., q.v.

And for channel, n.1:

Middle English chanel , < Old French chanel, ‘old form of canal ’ (Littré) < Latin canāl-em ; see canal n., which also compare for the senses.


  1. Martina Astrid Rodda said,

    October 15, 2015 @ 8:30 am

    In case anyone finds this useful, a quick check on R. Beekes' Etymological Dictionary of Greek (Leiden; Boston, 2010), s.v. κανών, confirms the derivation from κάννα, albeit not with 100% certainty:
    "Most often connected with κάννα as *'stave of reed'. The variation α/ο [MAR: the reference here is to Mycenaean ko-no-ni-pi, /konōni-phi/] shows that it is a pre-Greek word. The Semitic etymology by Lewy 1895: 133 (Hebr. qānoeh 'measuring reed, balance') is not to be preferred."

  2. Simon Tatham said,

    October 15, 2015 @ 9:15 am

    The notion of "cognate word" is not at all a transparent one

    Amusingly, I recall the word 'cognate' itself confusing me when I was younger. I'd attempted to glark its meaning from context rather than looking it up, and I misinterpreted the 'cogn-' prefix as being related to 'cognitive', and formed the hypothesis that it therefore meant two words were expressing more or less the same thought. That hypothesis never quite sat comfortably with me (e.g. shouldn't there be some part of the word indicating 'same', to go with the part indicating 'thought'?), but it's close enough to the real meaning of the word (at least in this sort of linguistic context) that I never quite managed to collect enough evidence from usage to reject it.

    It wasn't until a long time later that I recognised the g as irrelevant and realised that the right way to break the word up was as co + nate = 'born together'.

  3. mike said,

    October 15, 2015 @ 10:45 am

    How do I "Like" Simon Tatham's comment? :-) Very cool.

  4. Coby Lubliner said,

    October 15, 2015 @ 11:17 am

    I have a recollection of being told many years ago by a Latin teacher that canalis came from canis and originally designated the channel in a street where dogs moved. But I haven't seen this etymology in any dictionary.

  5. Michael said,

    October 15, 2015 @ 11:28 am

    My favorite example of this is the common origin of Spring and Frog, from an earlier root meaning "to jump":

  6. TR said,

    October 15, 2015 @ 11:49 am

    Beekes seems to be rejecting the connection with κάννα, as the latter is definitely Semitic, not "pre-Greek". Beekes's dictionary has gotten very mixed reviews from specialists, in any case.

  7. BZ said,

    October 15, 2015 @ 12:37 pm

    NJ has a town called Westampton (actually, East- and South- amptons also exist, but Westampton is signed on Interstate 295, so that's the one in my story). My father wondered what the "amp" part of the name meant (since West and -ton for town are pretty transparent), so I did some research. Turns out the relevant root is actually "Hampton" (which roughly means "hometown" or "homestead"). Apparently Northampton came to New Jersey first, in 1688 (and later got renamed, so it no longer exists), presumably from the town of the same name in the UK. The last "h" from "North" and the first "h" from Hapton got merged into one. By the time the other three Amptons were created some 200 years later, that etymology was forgotten (or never known in the first place).

    Interestingly enough, Massachusetts and New York got it right (Wessthampton and Easthampton). Of course the ones in Massachusetts are in Hampshire County, so it wasn't a mistake they were likely to make.

  8. david said,

    October 15, 2015 @ 12:39 pm

    Kandel and Schwartz's Principle of Neuroscience discusses the etymology of 'channel' because channels are proteins that permit and control the flow of ions or electricity into or out of cells. I don't have a copy with me now but I think they suggest an ancient Egyptian origen having to do with hollow reeds.

    [(myl) As the OED quotes in the post indicate, it's 100% certain that "channel" comes from Latin canna, borrowed from Greek κάννα, κάννη, where "In Latin the sense was extended from ‘(hollow) reed or cane’ to ‘tube or pipe’". There's a further speculation that the Greek word might have been borrowed from a Semitic word represented by Hebrew qāneh, Arabic qanāh reed, cane; an "ancient Egyptian origin" would be a way of referring to whatever Semitic variety the hypothetical source might have been.]

  9. david said,

    October 15, 2015 @ 12:42 pm

    oops, origin

  10. Y said,

    October 15, 2015 @ 12:50 pm

    Blažek suggests that φοῖνιξ 'date palm' is a borrowing from Egyptian bny or perhaps from Libyan Berber *bayn. The Egyptian bynw~bnw 'phoenix' is also attested.
    He also argues that the ethnonym is based on the tree, and not vice versa.

  11. Rubrick said,

    October 15, 2015 @ 4:44 pm

    It's fun to observe modern word meanderings from the perspective of imagined future etymologists trying to puzzle them out. My current favorite is the pop-culture use of "reboot", as in "The Spiderman franchise is due for another reboot". The term has gone through a rather remarkable journey, all within the space of fifty years or so.

  12. January First-of-May said,

    October 15, 2015 @ 8:19 pm

    I wasn't previously aware of the etymology of the computing term "to boot". If that's included, then yes, that's quite a remarkable journey. (And yes, it is related to the sort of boots that are worn on feet, but not directly.)
    I personally would've chosen "wi-fi", which is clearly an abbreviation, but nobody seems to know for what (IIRC, technically, it's supposed to be "wireless fidelity").

    Even without meanderings, it all too often happens that a word's origin is obscure simply because it was pretty much made up for some occasion and then just happened to catch on. I wonder what sort of weird derivation future etymologists would've ascribed to the linguistic term "crash blossom"…

    On-topic: my own favorite example of words of the same etymological origin is "cap, captain, capital, chapel, cabbage". [Apparently, some of these are from Latin caput (head), some are from cappa (a type of headwear), which might not be related, and some might not be from either.]
    Meanwhile, my favorite example of words that only look related (but aren't) is "marshal, martial, (to) march" (yes, all three have completely different origins – a wildly meandering one for the latter, which apparently comes from "march" in the sense of "area near the boundary"… a different, almost as meandering, derivation of the same root resulted in "mark").

  13. maidhc said,

    October 15, 2015 @ 11:23 pm

    "Booting a system" comes from the use of a "bootstrap loader" to start a system. It comes from the expression "pulling oneself up by one's own bootstraps".

    I don't think "wi-fi" is short for anything. It was just coined to be a catchy name. But "wi" is for "wireless".

    Funny that the British used to say "wireless" and the Americans said "radio", but "wireless" was losing the battle. But now that we're using radio to connect to the internet, everyone is calling it "wireless".

  14. John Coleman said,

    October 16, 2015 @ 2:09 am

    Greek phoin- is indeed monomorphemic, from a Semitic form Pun. It is seen also in the Biblical “land of Punt” and in the Latin adjective “Punic”.

  15. January First-of-May said,

    October 16, 2015 @ 5:07 am

    IIRC, it was coined by similarity to "hi-fi", which a several-decades-old term standing for "high fidelity". Since "wi" is for "wireless", technically, we get "wireless fidelity".
    Still better than "bus", which is a Latin case ending that managed to become an English word with a specific meaning (a type of transportation). Etymology can be silly.

    On radio and connecting to the Internet – early modems used actual wires (mostly telephone ones) to pass the signal, so "wireless" is a reference to the newer (radio) versions not using any. I highly suspect that it's an entirely independent derivation (though I suppose it's possible that some old person who was used to calling radio "wireless" proposed the word in this case).

  16. Tom Ace said,

    October 16, 2015 @ 9:19 am

    I used to wonder whether Potomac River was like the La Brea tar pits. It seems close to having the same Greek root for river that's in hippopotamus and Mesopotamia. But no. Potomac comes from Algonquian.

  17. Robert Coren said,

    October 16, 2015 @ 10:18 am

    @Tom Ace happened to punch one of my buttons, so pardon the tangent.

    There's nothing wrong or tautological about "the La Brea tar pits". The place is named La Brea. If you said "the tar pits of La Brea", nobody would bat an eye, and the usual formulation is just a nice concise way of saying the same thing. The fact that the name of the place happens to mean "the tar" is not entirely irrelevant (without the tar pits, it wouldn't have been called that), but it's not "ignorant", as many seem to imply, to say the canonical phrase.

  18. Lars said,

    October 16, 2015 @ 10:56 am

    Wi-Fi is a sort of Pyrrhic marketing victory. The idea in 1999 was to brand equipment that conformed to the IEEE 802.11b standard and passed the interoperability requirements set up by the private (non-profit) Wi-Fi Consortium.

    But only a few nerds know that you can have 802.11 equipment that isn't Wi-Fi — though probably millions of such units are produced each year. Wi-Fi is just the word for the handwavy thing that gets data from 'the Internet' (inside the wall) to your PC and back to your TV.

    And the allusion to Hi-Fi — can you get more 20th century? By the time the first boxes of Wi-Fi branded equipment were in the stores, the Hi-Fi enthusiast was already a dying breed. No wonder that the implication that 'Wi-Fi is better than other wireless solutions' got lost.

  19. Rodger C said,

    October 16, 2015 @ 11:01 am

    @John Coleman: Uh what? "Punic" is from Latin after oe became u (cf. poena vs. punire. "Phoenicia" is a learned Graecism, while "Punici" is an oral borrowing. "Punt" is Egyptian and refers to Somalia or thereabouts; the Biblical form (KJV) is Phut, the brother of Cush (Sudan) and Mizraim (Egypt).

  20. Greg Morrow said,

    October 22, 2015 @ 4:50 pm

    Don Ringe is a living treasure, and it is always good when he shows up on LL. I am (very slowly) working my way through his "From PIE to PGmc" book, which is amazingly complete.

RSS feed for comments on this post