Greek morf-ae and Latin form-a: consonants: MRF - FRM

Discussion in 'Etymology, History of languages and Linguistics (EHL)' started by rogermue, Jan 31, 2013.

  1. Does anybody know if this phenomenon has been studied somewhere or does anybody know something about this?
  2. Christo Tamarin

    Christo Tamarin Senior Member

  3. berndf Moderator

    German (Germany)
    You seem to be mislead by modern pronunciation when you write:
    In classical Greek ph (φ) was not yet pronounced like an f. It was rather pronounced as transcribed into the Latin alphabet: p-h, i.e. like a German aspirated /p/ as in Papier but not as in spielen. The letter is how the regular p (π) was pronounced.
    Last edited: Jan 31, 2013
  4. Actually I thought of the left-right phenomenon. In the Greek word we have the consonant order: m-rf and in the Latin word we have the same consonant order but from right to left: f-rm.
    Others told me that they observed this phenomenon but they didn't think much about it. A curious single case.
    Now I have found about 200 of these left-right pairs and I think it is a way how words can come into being. I can't explain how such a thing can come about, my only explanation is that originally a word from the Semitic area (with writing from right to left) was taken into the Greek or Roman langage (with writing from left to right) and that the original direction of reading such a word was neglected.
  5. The Romans overtook the Greek phi, however the Greeks may have pronounced this sound, as f. So Greek phi is in a line with Roman f. But it is mere chance that I chose a Greek word with phi. I chose it because this left-right pair is well-known.
    Last edited: Jan 31, 2013
  6. berndf Moderator

    German (Germany)
    "May have", but they didn't. Cf. here, Sidney Allen: Vox Latina, p.27.

    PS: The theory cited by Christo Tamarin explains the derivation via Etruscan which may or may not explain the change from /pʰ/ to /f/. I know too little about Etruscan, though I doubt it: Etruscan, like Ancient Greek, contrasted all three phonemes, /p/, /pʰ/ and /f/.
    Last edited: Jan 31, 2013
  7. entangledbank

    entangledbank Senior Member

    English - South-East England
    If you have 200 of them, that would be of real interest, but I think it is unlikely you have more than this one well-known one and a handful of chance resemblances. These two do look very similar - if only consonants could swap around like that in Indo-European (*morbha: to bhorma: or vice versa, or something like that). But there is no known mechanism for it - that is, there is no process known to operate regularly in the history of Indo-European that could swap the consonants. So it is much more likely that it is just a coincidence. That sounds less exciting than a secret passage through Semitic, but has the advantage of being true.

    The Romans took Greek [ph] in as [ph] in Classical times, which is how educated speakers said it. The common people said [p]. Our modern pronunciation [f] is mostly due to the influence of late Byzantine Greek, such as Greek scholars fleeing west after 1453. [this comment cross-posted]
  8. fdb Senior Member

    Cambridge, UK
    French (France)
    There are two different issues here. On the one hand, Bernd is right to say that Greek /ph/ and Latin /f/ are two different sounds. This rather rules out the option of deriving forma from μορφή, either directly or via Etruscan.

    On the other hand, Greek φ and Latin f do both go back to Indo-European *bh, so at an Indo-European level one could posit *morbh- in alternation with *bhorm-, with a kind of metathesis that is very common in Indo-European and other languages. But, as far as I can see, no one has yet established a clear reflex of either of these outside of Greek or Latin.
  9. berndf Moderator

    German (Germany)
    True. If we ignore the theory underlying #2 and #4 that the Latin word is a direct or indirect loan from Greek, a lot more possibilities arise. Especially in Semitic we find a lot of semantically connected three consonant roots which differ only in the position of one of the consonants. These gave rise the famous theory of two consonant proto-roots that gave rise to more differentiated meanings by adding a third consonant that can occur in different positions. A similar mechanism might have been at work here. So, if we re-interpret the question as pertaining to cognates rather than to loans, I agree, this is a valid example of a potential metathesis.
  10. Well, I think, I should give some more examples.
    When I consider two words as a right-left pair I do this on the basis of the following historical line of development
    p b f v - (I don't consider or differentiate spirited and non-spirited plosives, that would make things to complicated. Right-left etymology is complicated enough.)
    t d th dh - here a line with ts or s sounds can develop: compare E to shoot - Ger shießen)
    k/c g ch h - (- stands for a vanished sound)
    q (I handle it as k, a simplification)

    Right-left pairs occur between two different languages, Greek-Latin, Greek-German, Latin-German/French/English and so on
    and also in one language E E, Ger – Ger etc.
    E Englisch, G Greek, L Latin, Ger German, F French

    E hole - Ger Loch,
    Ger Topf – Ger Pot, E pot, F pot
    L pes/pedis – ped.s – E step Ger stapf-en
    L merx (merk-s)/merc-is – Ger Kram + Kräm-er (Krämer was the owner of a small shop for food and other things of everyday necessities.)
    Ger faul – E to loafô
    E to shove – E to push
    F la geôle – Ger Loch (für Kerker)
    G tox-euoo – Ger schieß-en, Schuß, Schütze E to shoot, the shot.
    - Remark: x/ks can become sk and sh/sch
    - Vowels between the consonants are of minor importance, they easily change and they can change position.
    Ger Zieg-e - Ger Geiß
    L ad, E at - Ger da
    G gel-aoo - Ger lach-en
    L mens/mentis - Ger Sinn (mens:s.nm >sinn)
    G phil-eoo - Ger lieb-en
    G kosm-os - Ger Schmuck (actually a "Dreher": msok gedreht/changed: smok smuk Schmuck)
    Ital. bast-on-are, F le bàton - Ger Stab, E Stave. Remark st can stay st as if it were one sound
    G tekn-on (Kind, "Gezeugtes" - Ger Kind (Dreher) ´+ L genit-us/a/um
    - Remark: If I want to show the connection better I write capital letters: G T.KN -Ger K.NT. K stand for all variants of the k/c-row, T for all variants of
    - the t-row. The point shows the place of the vowel.
    G thesis - Ger das Setz-en, Satz, E to set. Remark: In RLE (right-left etymology) Greeth th (Theta) can occur as tz /ts/ in German or t in English.
    - Generally speaking, the mechanics of the development of sounds in RLE are a bit different from what Grimm and others have established for normal
    - etymology. I had to learn this in the course of time.
    L tim-eo, tim-or - L met-us/us
    L pes/ped-is - F tap-er
    L pes/ped-is: ped.s - Ger stampf-en. With m-infix.

    What I would like to find are right-left variants between Semitic languages and Greek or Latin. But the only Semitic language I know (only the rudiments) is
    Arabic. But for such studies one has to have a mastery of the literature and the various historic stages of a language.
    The field of right-left-etymology is in principle not a matter of one person, but the work for a team of specialists.

    Behind the realm of right-left variants I have discovered another similar phenomon which I call "Dreher". The order of the consonants is not left to right or right to left, but they can occur in any possible order such as
    in E to crunch - Ger knirsch-en

    At first I refused to collect such pairs, but in the end I found so many, that I collected them too.
    I did these studies about twenty years ago. I wanted to publish them, but I didn't find the time and the university wasn't interested. The matter does not concern one 'faculty' such as 'Anglistik' but many languages.

    All this is connected with my discovery of a connection between Arabic letters and Latin letters - also a right-left-phenomenon, so this matter concerns Semitic languages too.
    At the moment my twenty files of etymological studies are in boxes in several garages. So I don't have the full material at hand.
    To the orginal list I added
    L mens
    G phil-eoo
    G kosm-os
    Ital. bast-on-are
    G tekn-on
    G thes-is
    L tim-eo
    L pes
    L pes
    Last edited: Feb 1, 2013
  11. I just read the word 'derive'. I would like to add if I talk of right-left pairs I don't want to say that L form-a derives from Greek morph-ae. The history of words is too complicated for such affirmations. I only want to say these two words belong somehow together. In the end the source might be a semitic word,
    the Greeks may have borrowed it reading it in its original direction, the Romans may have borrowed it as well from a Semitic language changing the direction.
    Just a hypothesis to show what possibilities would be to consider. All that is far to complicated. So I only say: this pair belongs together, no more.
  12. berndf Moderator

    German (Germany)
    You've been very liberal in your definition of "same". It is e.g. not prima facie plausible equate /f/ and /p/ for German words like in shove and push. The problem is that if you expand the domain of your search too much (many different languages; liberal definition of "same"), finding several hundred chance coincidences is statistically probable and therefore meaningless. I think you should restrict your research to relatively safe reconstructions to a specific proto-language the very least (e.g. forma is not a safe one because the Latin /f/ can derived from either PIE /dʰ/ or /bʰ/). In the last few years I've been working in the field of statistical trading algorithms. I've seen algorithms that explain the past perfectly well but fail absolutely miserably to predict only a single day in the future because they selected efficient predictors from a much too broad domain. Finding such a perfect vector of predictors can be as improbable as winning the jackpot in the lottery. This happened to me about a year ago with an algorithm. I then estimated the number of predictor combinations we searched I found out that we had effectively searched 8bln. combinations. If you buy 8bln. lottery tickets that probability of winning the jackpot at least once is almost 1. In the case of trading algorithm you have a straight forward reality check: If it is wrong the algorithm will fail sooner or later, normally rather sooner than later. In historical linguistics you don't have this reality check and if you aren't careful with your statistical methods because you could be chasing red herrings all your life without ever noticing. I don't want to discourage you. I don't know, if you have a valid point or not. I just want to urge you to be super careful with your methods.

    Here is a resource link you should read. It gives you an example to estimate the probability of finding pure chance coincidences in cases like yours.
  13. Do you really think, berndf, Topf + Pot or Ziege + Geiß are mere chances?
  14. I have studied this right-left phenomenon over years. As an average I didn't find more than 10 pairs in the first years. Then I had the eye and I found more.
    But as I had found the right-left connection between Arabic and Latin letters before, I became interested in right-left phenomena in a more general way.
    In writing we have the system from right to left (Arabic) and from left to right (normal system in Europe). But in early Greek writing we have a mixed system - one line from left to right, the next from right to left. At that time they could write in both directions. And what is more, when they wrote from right to left they turned the letters around, too ( the letters were clapped around their vertical axis, a mirror image).

    In old Egyptian writing you also have both directions. There was not a direction that was preferred. If the birds and persons looked towards the right side you had to read the line from left to right. If they looked to the left, you had to read from right to left.
    One can show that either the Arabic letters are mirror images of the Latin letters or vice versa. (That is not to say that the Latin alphabet comes from the Arabic one or vice versa, the historical lines of development of writing are much more complicated, but if we compare the two alphabets today we can show their connection nevertheless. It is as if you compared German König and English king. You can show that the two words belong together even if you don't know what was the form in Old English and in Old High German or what was the Indoeuropean root. Just push aside the ö in König and put in the i and you have king.)

    So the right-left phenomenon in language (letters, words) is a more general human phenomenon. Our brain works with this system without our knowing. The image of a tree we see is topside down in our interior eye. By experience in young years the child learns to correct this wrong image in our interior eye.
    Young school children often write with mirror images of letters or write from right to left. I have read medical reports about this and seen pictures of such writing.
    All this shows that Man can somehow handle two directions. And we have right-handers and left-handers.

    And if the right-left phenomenon can be found in letters, why shouldn't it be found in words as well.
    Last edited: Feb 1, 2013
  15. berndf Moderator

    German (Germany)
    The origin of Topf is completely obscure and we can't even trace it back to MHG. Pot is a Latin loan. It is very difficult to tell.
    Grimm's law doesn't work the same in the syllable onset and coda. The /g/ in Ziege is from PIE /g/ and in Geiß from PIE /gʰ/. Also, the PIE meanings don't seem to match, it the assumption is correct that haedus is the Latin cognate of Geiß.

    The thing is that by doing what long ago my statistics professor called "science based on the principle of accumulation of first degree errors" you'll never know.:(
  16. berndf Moderator

    German (Germany)
    I am sure it does.:) The question is where and when and how often.
  17. fdb Senior Member

    Cambridge, UK
    French (France)
    “Metathesis” of phonemes in a spoken word is an established concept in historical linguistics. But your “left-right phenomenon” is something very different. Languages do not change because some person has read the letters of some written word backwards. This is a parody of linguistics.
  18. aruniyan Senior Member


    interesting, recently i was searching for such relationship between Sanskrit and Tamil and found a lot of words, I think this has something to do with creating mantras using common words, some kind of early encoding and later became part of a different language.
  19. You said: The origin of Topf is completely obscure.
    One indication that there must be something to this right-left theory is that if we have a right-left pair that holds the water then the normal explanation in etymological dictionaries show that either the origin of one word is unknown or the explantions are in a form that leave a lot of room for doubt.
  20. That is a good question: where and when and how often.

    These are questions I can't answer up to today. I can only suppose that such things can happen everywhere at any time. This particular way of creating words - and I am sure it is a way - is certainly the exception. It is not the normal thing. And I am almost sure the percentage of the vocabulary in one language with a right-left explanation must be around 1 percent, perhaps two.
    But up to now we find: Topf: origin unknown. But for me it belongs to a right-left pair (Topf and Pot).

    Let me tell you how I came to this very curious theory of RLE (right-left etymology).
    My ex-wife was from Tunesia. So it was natural that I became interested in the Arabic language and studied it. I think it was after a year or so I discovered a likeness between an Arabic letter and a Latin one. I put this find away as a curious chance incident. After a time I found a second similarity, and then a third one. Now I pricked up my ears and studied all letters of the Arabic alphabet that can be compared. There are some sounds and letters in Arabic we don't have in our western languages or in German, but they are few. Of course you have to take the best position variant (letter form at the beginning or the end of the word where the character is in its most clear form, the variants in the middle of the word are no good because they are transformed and often very small). Twenty or thirty years ago I made a small booket about this matter, handwriten, about 30 pages or so where I showed one Arabic letter on one page and how it can be transformed easily into a Latin letter or vice versa. The little transformations have to be a natural consequence of handwriting in the opposite direction. I have shown this little booklet to a brother of my ex-wife and he was convinced of my way of seeing things.

    Well, I knew that letters can be turned around. And when I found two further pairs such as G morph-ae and L form-a I had the idea that words might be turned around as well and I began studying this questions over the years and collected examples. I think I did this over 15 years. And today I am convinced that a certain percentage of our words have a right-left orgin.
    Last edited: Feb 1, 2013
  21. berndf Moderator

    German (Germany)
    If the origin is obscure than it is obscure. If you interpret every case where there is doubt as an indication in favour of your theory you are lying to yourself. The significance of confirmation of scientific theories lies in riskiness of its predictions. In the way you test your theory, it can't see any possible way by which you possibly could find out that you are wrong, if you are wrong because all you are interested in is accumulating cases that "could" be explained by your theory.

    The special issue with Topf is that it can't even be traced back to MHG, let alone OHG. But for equating /p/ and /pf/ to hold water, you would have to be able to explain why the newer word Topf underwent the 2nd Germanic sound shift and the older word Pot not.
  22. If Topf can't be traced back to MHG then Topf must be a relatively young formation.
    And consonant sounds change along in their row of development not only as Grimm and others stated with their laws, consonant sounds can change at any times. In German we write Pfennig, but mostly pronounce /pfenich/. This sound shift happens today and it is not necessary to look what specialists for IE said about historical sound shifts. In ancient Greece the various dialects show that consonsants change from one dialect to the other and you can observe this phenomenon everywhere. Everywhere at any time little shifts can be observed, often so minimal in certain positions that it is hardly observed.
    So if I connect Topf and Pot I don't have any need to discuss a sound change centuries ago (second Germanic sound shift). I say Pferd, in northern Germany they say Ferd. For me it is enough to observe this difference today, it is not necessary for me to study when and how often in history /pf/ was deduced to /f/.

    Grimm deduced his rules or laws about consonant change by comparing cognate words in IE (Greek, Latin), Germanic and High German. Between these stages are centuries and a lot of things can have happened in these centuries. But Grimm had no idea of RLE and I must say we or I don't know the mechanics of RLE exactly. But I'm sure they are a bit different from those of normal etymology. We can't transfer automatically what we know about consonant change from normal etymology to RLE. I don't know why Pot has a counterpart in Topf (and not Top). There can be a lot of factors that make that decision. But I'm not so much interested to explain why we have pf and not p, I only state it is highly probable that Pot and Topf belong together.
    Last edited: Feb 1, 2013
  23. berndf Moderator

    German (Germany)
    The reduction of /pf/ to /f/ is because the phoneme is non-native in Low and Middle German and speakers simplify this "foreign" sound. The native word is Perd, not Pferd or Ferd. The issue here is that the sound shift /p/>/pf/ is not plausible for words that were created after the 2nd Germanic sound shift. You would need to produce examples where this demonstrably happened.

    The pronunciation pfennig with [-ç] is quite regular. The original Germanic realization of /g/ as was fricative [ɣ] (sounding like a softer version of the French/German "r") with allophonic variant [ʝ] (that's why dag became day in English). The suffix [-ɪç] is then the regular outcome of German Auslautverhärtung (final devoicing) applied to [-ɪʝ]. As you see, it is not sufficient to say "Because all kind of changes happen, all these differences are negligible". You often can and therefore at least have to try to look at the story behind them, if you want to be taken seriously.
  24. fdb Senior Member

    Cambridge, UK
    French (France)
    In case anyone is still interested in the original question: I have looked the two words up in the Greek and Latin etymological dictionaries, all of which admit that they are problematic and of uncertain origin. However, Beekes, in his recent Greek etymological dictionary, considers the possibility that μορφή ‘shape, beauty, grace’ might come from *mergwh-h2-, which could then also account for the hitherto unexplained Lithuanian mergà ‘girl’. In this case, *mergwh- could have been metathesised to *gwherm- , o-grade *gwhorm- , whence proto-Latin *form-, and then finally, with secondary lengthening of the vowel, fōrma. I find this very attractive.

    The ‘Etruscan loanword’ hypothesis has been around for a long time and has been supported by such famous scholars as Benveniste. It involves Greek μορφή being borrowed into Etruscan, where it undergoes first assimilation to *morma, then dissimilation to *forma > fōrma. In this way it avoids any direct connection between φ and f.
  25. CapnPrep Senior Member

    Do you have other examples of this type of super-long-distance metathesis in IE, where two non-liquid consonants switch places across an intervening consonant (and vowel)? You say above that it is "very common".
  26. fdb Senior Member

    Cambridge, UK
    French (France)
    Off hand: Avestan maγna- < *nagma- ‘naked’; the same root, but with even more elaborate metathesis, in Greek γυμνός.
  27. I have been thinking about this a bit. If we assume that Greek morph-ae was taken by the Romans - and this would be natural, but it is only an assumption - and if we read morph- from right to left, then Greek phi is in initial position and in initial position a Greek phi becomes Latin f and a Greek phi in mid-position becomes Latin b.
    So I don't really see what is the problem if I assume that morph-ae and form-a are right-left counterparts.
  28. berndf Moderator

    German (Germany)
    How and why? I don't understand what you mean.
  29. Greek -- Latin
    phyae -- fuga
    pheugoo - fugio/fugere
    pheroo /phoréoo -- fero/ferre

    Please excuse my private transcription of Greek words with Latin letters. But it's so cumbersome to handle all those special letters.
    Last edited: Feb 2, 2013
  30. fdb Senior Member

    Cambridge, UK
    French (France)
    Do you understand the difference between borrowings (Entlehnungen) and cognates (Urverwandte)?
  31. It doesn't matter if the words are borrowings or cognates. It was not my objection that forma and morph-ae can't belong together. And it was not my statement that the Romans took the word from the Greeks. If I find right-left pairs I don't ask how they came into being or who had the word first. In most cases there's no way to answer such questions.
  32. Perseas Senior Member

    Athens - GR
  33. Thank you, Perseas, for writing the Greek words with Greek letters. I have enough to do if I have to write French letters with accents. Almost ten clicks for an e circonflexe.
  34. xari Member

    Try this site, it might help you:

    You can either select the characters from the top row, or simply type in the box with your regular Latin keyboard, and it will be converted to Greek (even the digraphs).
  35. berndf Moderator

    German (Germany)
    Yes, you did. In #4, #5, #11 and #27. You left it open who took it whom from whom but you insisted on borrowings. A possibility of Greek φ = Latin f was so far only established for cognates. For borrowings you'd have to demonstrate how.
  36. Thanks a lot xari, I know lexilogos and others, but writing signs I don't have on my keyboard is a lot of work all the same and I prefer to avoid it.
    But I will try your letter-side. :)
  37. I think I made it clear enough that I don't consider such questions as 'borrowed from' - I don't talk about things I can't know.
  38. berndf Moderator

    German (Germany)
    You made it clear enough that you think those words were taken by one language from another
    The Issue with cognates is that they aren't taken at all. The had always been there ("always" of course means "since common proto-language of the languages concerned" ).
  39. By mere chance I have just found an interesting entry in Etymological dictionary of the Latin Language by Francis Valpy, 1828:

    forma, a shape, figure, form. Transposed from μορφά.

    I hope I understand 'transposed' right as to swap, e.g. to swap numbers.
    So I think Valpy says the same as I do. Or am I mistaken?
    Last edited: Feb 2, 2013
  40. fdb Senior Member

    Cambridge, UK
    French (France)
    In 1828 the principles of Indo-European comparative linguistics had not yet been discovered. If you want to know what modern scholars have to say do take a look at no. 24.
  41. You have a point, fdb, by saying the dictionary of Valpy is a bit old. Right,the main specialists for IE were active later.
    I researched a bit about what you said in no. 24.
    The view 'forma borrowed from Greek via Etruscan' is a hypothesis, and the view 'morpha-morfa-morma-forma' is a hypothesis too.
    Technically it would possible, as m and f are relatively near in articulation. But m is a relatively stable sound, especially in initial position. I know consonant change m/n, m/w, but I don't know any case of m/f.
    Do you have any knowledge about such cases (m/f)?
  42. CapnPrep Senior Member

    Check out the history of formica "ant" (cf. Greek μύρμηξ).
  43. Oh yes, I had forgotten it, thanks.
  44. I would like to add some more examples of word pairs which I think can be seen as right-left pairs. Sometimes the transformations necessary are a bit more complicated.
    Latin praegnans/praegnantis and German schwanger
    L ....gnans - Ger schwanger, swanger, s.ang..
    If one reads s.ang backwards you get gnans.

    This may see like playing with words and letters. But the German word schwanger is relatively isolated. We have no insight into the idea of that word.
    German etymological dictionaries have older variants (Mhd swanger, ahd swangar, and some related words in Dutsch and English. But the idea of this word is unclear. German dictionaries think it may mean heavy.
    If we take schwanger as a right-left pair and explain it with Latin prae-g-nans (related with nasci to make a child/ to be born) then we get a clear idea of this word-formation:
    a woman before (prae) giving birth to a child.

    The Latin word-formation has a clear transparent logic, the German word has no recognizable logic. It is "dark". But explained as a right-left-word it gets a clear logic again.

    Oh, I see I was imprecise. I have to take the older preclassical Latin form:
    praegnans/praegnantis - preclassical: praegnas/praegnatis - ...gnas
    If you read s.ang backwards you get gnas.
    Last edited: Feb 6, 2013
  45. French Verlan

    Actually I don't want to continue this subject but I think French Verlan should be mentioned in the line of right-left phenomena. Verlan is a kind of secret language in which a syllable of a word (consisting of two syllables) changes position from right to left as in
    French pourri
    Verlan-French ripou.
    I don't know what importance Verlan has in France or the French language, but one can find such words as ripou in normal French novels.
    Link about Verlan

    Something similar can be found on this forum in Français seulement, page 504, keuf, meuf, keum.

    See also the article Verlan in the English Wikipedia:
    The article mentions other language games of this kind in other languages.
    Last edited: Feb 10, 2013

Share This Page