The 'g' in the middle/inside in Dutch and German // English -ow or -y-?

T

ThomasK

Senior Member
(near) Kortrijk, Belgium
Belgium, Dutch
I am referring to these parallel words, based on the same root, but having developed differently phonetically speaking:

(1) -rg-/ -lg- = ENG -low, -row
- morrow/ morgen, sorrow/ zorgen (semantic difference), barrow/burrow/ berg [mountain], marrow/ merg, AND geel(f)/ yellow???
- follow/volgen, swallow/ zwelgen, hallow/ heilig
(2) -g = -w
low/ laag, bow/ boog, row/ rij, ...
(3) -g(-) = -y(-)
to lay/leggen, layer/laag, (a)way/ weg, stay/ staag (maybe)

The main question is: does anyone have more examples, especially lists of examples (not exhaustive perhaps but with extra indications?

But explanations for this [phonetic] variation are very welcome. Also other instances/... of such a phonetic in/ between other languages. In fact any comment that can shed light on the variation... Thanks in advance!
 
    • I never noticed the "ow" cases as a group, but, now that I see your examples, they remind me of bow/Bogen and elbow/Ellbogen.

    I do know that there was definitely a g→y shift in Middle English when other Germanic languages weren't doing it. Another few examples that stand out in my mind are day/Tag, eye/Auge, and "yard/garden", with "garden" being one that we imported from the Normans after the shift to "y" in our own version. One that would appear to be backward is egg/Ei, but the shift actually happened in that case, too, giving English "ey", but then that got replaced by its Norse counterpart which still had the old "g".

    But we do still have "g" in a bunch of other words which weren't touched by this shift, and I don't know why. I do know that, in Old English, that letter could be used for either the plosive /ɡ/ or the fricative /ɣ/, and I could infer that the latter are the ones that would then shift to /j/ in Middle English, but that would just change the question to why some were /ɣ/ in OE and some weren't.

    Another complication in any attempt to work out the rules for this shift is the fact that other Germanic languages have had similar shifts at other times, with different lists of which words were and weren't affected. I already used German "Ei" above, for which the modern Scandinavian cognates are ägg, æg, egg, and egg. So we could think that some words had this happen in southern languages but not northern ones. But modern Swedish has also developed a tendency to pronounce a written "g" like an "i/e/y" without changing the spelling, in some cases like "dag", the word for "day", coming out like "da-ii", which you wouldn't expect from a north/south division. And we all seem to have agreed to drop the "g" from "Frigday". Maybe that last one is because of the following "d".
     
    Delvo said:
    I never noticed the "ow" cases as a group, but, now that I see your examples, they remind me of bow/Bogen and elbow/Ellbogen.

    I do know that there was definitely a g→y shift in Middle English when other Germanic languages weren't doing it. Another few examples that stand out in my mind are day/Tag, eye/Auge, and "yard/garden", with "garden" being one that we imported from the Normans after the shift to "y" in our own version. One that would appear to be backward is egg/Ei, but the shift actually happened in that case, too, giving English "ey", but then that got replaced by its Norse counterpart which still had the old "g".
    Click to expand...
    This is something I had not thought of, but it fits into the category (3) to which I had never really payed much attention. My main focus was the ending in -ow. But I am surprised I had not thought of the frontal g/y variation as in garden/yard.

    Delvo said:
    But we do still have "g" in a bunch of other words which weren't touched by this shift, and I don't know why. I do know that, in Old English, that letter could be used for either the plosive /ɡ/ or the fricative /ɣ/, and I could infer that the latter are the ones that would then shift to /j/ in Middle English, but that would just change the question to why some were /ɣ/ in OE and some weren't.
    Click to expand...
    I suddenly think of the guarantee/ garantie/ warrant variation, which resembles the g/ow variation.
    Delvo said:
    Another complication in any attempt to work out the rules for this shift is the fact that other Germanic languages have had similar shifts at other times, with different lists of which words were and weren't affected. I already used German "Ei" above, for which the modern Scandinavian cognates are ägg, æg, egg, and egg. So we could think that some words had this happen in southern languages but not northern ones. But modern Swedish has also developed a tendency to pronounce a written "g" like an "i/e/y" without changing the spelling, in some cases like "dag", the word for "day", coming out like "da-ii", which you wouldn't expect from a north/south division. And we all seem to have agreed to drop the "g" from "Frigday". Maybe that last one is because of the following "d".
    Click to expand...
    I must say my main focus has been the (the listing of) the phenomena as such, not so much the explanation. What I would like to understand is the phonetic rationale behind the shift (the way Umlaut can be explained by the palatisating/... effect of the ensuing 'j').
    By the way I must admit that I am not that well informed about the historic shifts, except for the First and especially Second Sound Shift. Should you know of a survey of some of the shifts you are referring to, please tell me.

    But thanks a lot for the valuable, insightful information!
     
    Regarding the g/y: Don't think too much in terms of letters but sounds.

    There was the West-Germanic /ɣ/: Dutch has preserved the sound, in English it became /j/ (represented by the letter Y) and in German /g/. It's part of the High German sound shift in which Dutch didn't participate (much).

    A funny side note: in a number of German accents "g" is also spoken as /j/.
     
    Of course you are right: ultimately we are talking about sounds.

    West Germanic : of course, quite right, had not thought of that. And I think it was in a German naturalist novel or play that I first noticed that in certain Mundarten (no?) "g" was pronounced as /j/. I keep wondering: is there some reason for such an evolution. /j/ is palatal, I believe, but /ɣ/? Not more velar?
     
    ThomasK said:
    I suddenly think of the guarantee/ garantie/ warrant variation, which resembles the g/ow variation.
    Click to expand...
    Guarantee and warranty are separate borrowings of the same word but from different French oïl dialects: guarantee is from Parisian French and warranty is from Norman French. They entered English with those different initial sounds already in place, so what sound change there was had already happened in northern France.
     
    ThomasK said:
    Thanks. Just wondered: could it be that the guarantee has ever been pronounced as /gwarantee/?
    Click to expand...
    That's unlikely, the first attestations of the root g(u)a(r)rant- in English date from the 15th century according to etymoline. That means the loan was from Middle French, and a few hundreds of years after the /gw/ > /g/ shift.

    As for Old English's /ɣ/, it only became /j/ when followed by /i/, /e/ or /æ/, and when it ended a syllable containing those vowels, as far as I know. That's not too different from the contexts where Belgian Dutch has a [ʝ] or [ç] for its own /ɣ/
     
    Swatters said:
    That's unlikely, the first attestations of the root g(u)a(r)rant- in English date from the 15th century according to etymoline. That means the loan was from Middle French, and a few hundreds of years after the /gw/ > /g/ shift.
    Click to expand...
    Can you enlighten me on the /gw/ to /g/ shift? Googling has not helped me so far...
    Swatters said:
    As for Old English's /ɣ/, it only became /j/ when followed by /i/, /e/ or /æ/, and when it ended a syllable containing those vowels, as far as I know. That's not too different from the contexts where Belgian Dutch has a [ʝ] or [ç] for its own /ɣ/
    Click to expand...
    When followed by frontal or palatal vowels, you mean? I am wondered about what you state in your last sentence? Would that not hold for German (but then /g/)? But maybe I am not fully aware of what you mean regarding Flemish (Belgian Dutch). Could you give some examples.

    Thanks a lot for your contribution!
     
    ThomasK said:
    This is something I had not thought of, but it fits into the category (3) to which I had never really payed much attention. My main focus was the ending in -ow.
    Click to expand...
    In your group 1, the English words once had a "g" at the end (or followed by only a single vowel) and lost it, except for "yellow", which is also the only one for which the non-English cognates have something else at the end instead of "g", so it doesn't really belong in the group.

    That loss of "g" happened even when the adjacent sound(s) was/were only "a" or "o", so it appears to be a separate rule from the loss(es) of "g" in other places, where it looks dependent on an adjacent front vowel sound. Together, they both seem to be only parts of a complex of Middle English velar-to-non-velar sound shifts:

    /ɡ/ ► silent at end in the "-ow" words
    /ɣ/ ► /j/ (old sound spelled "g", new sound spelled "y") as already mentioned in this thread
    /x/ ► silent after "i" in the "igh" spellings
    /x/ ► silent or /f/ after "ou" in the "ough" spellings
    /sk/ ► /ʃ/ (as in "shriek/skrik" and "ship/skiff")

    /sk/►/ʃ/ is a fun one because I learned to recognize it from Modern English pairs of words in which one was native English with the sound shift and the other was Norse without it, like "shatter/scatter", "shirt/skirt", and even "ship/skiff". A few pairs like that also exist where the import is from French, like "shard/scar". "Skiff" doesn't get much use with most Modern English speakers, but it does among those who are into boating, meaning a smaller class of boats than the big ones we call "ships". Also, a boat's captain can be called its "skipper" in English. All of our modern "sk" words are from Norse or French after their English cognates had become "sh", but the English cognate wasn't always kept, so there's often no such pair anymore (as in "sky", "skull", and "skill").

    There was also /k/►/ʧ/ in Old English, as in the pronoun "ic" for "I" already having /ʧ/ in Old English but still /k/ in other Germanic languages. But that was early enough that it would just go silent later on when the other shifts I just listed happened, so it's unlikely to be related to them (unless it was the original trigger which caused the rest as a cascade of side-effects afterward).

    Three more develarizations are functionally similar but are French imports, not internal English developments, although the similarities with what happened in English might have made them easier for English to accept without change:

    /k/ ► /s/ before front vowels (spelled "c"; later dropping of those front vowels in some French examples led to "ç")
    /ɡ/ ► /ʤ/ before front vowels (►/ʒ/ in French)
    "ch" spelling ► /ʧ/ (►/ʃ/ in French

    ThomasK said:
    What I would like to understand is the phonetic rationale behind the shift (the way Umlaut can be explained by the palatisating/... effect of the ensuing 'j').
    Click to expand...
    I'm not sure what you mean. Palatalization is a change that happened (a shift toward palatal position or under the influence of another sound in or near palatal position), but naming it doesn't tell us why palatalization happened. I don't think sound shifts ever do really have explanations/rationales for why they happened. With the English complex of velar-to-something-else shifts I listed above, we could name the group "develarization", but that would just be another description, not an explanation.

    ThomasK said:
    By the way I must admit that I am not that well informed about the historic shifts, except for the First and especially Second Sound Shift.
    Click to expand...
    I don't know what those mean. Are they literal English translations of Dutch phrases? If you mean what English calls the "First Germanic Sound Shift", it's much more often called Grimm's Law in English, so I infer that the second might mean what we call Verner's Law. (But most sound shifts don't get names like those in any language; there are just too many of them. So they just get descriptions, like "Middle English /u/ diphthongization" for the origin of Modern English's "ou/ow" diphthongs.)

    ThomasK said:
    Should you know of a survey of some of the shifts you are referring to, please tell me.
    Click to expand...
    Phonological history of English - Wikipedia

    There aren't equivalent Wikipedia articles at the same level of detail about the evolution of other modern languages, at least not in English, but maybe there are in their own languages. Below is one for Proto-Germanic which is also more detailed than what most other past languages get in English, again presumably because English is Germanic; maybe similar articles on other past languages exist in their own nearest modern relatives...

    Proto-Germanic language - Wikipedia
     
    Last edited:
    ThomasK said:
    Can you enlighten me on the /gw/ to /g/ shift? Googling has not helped me so far...
    Click to expand...
    Just look at any Modern French word that's spelled with "gu" and pronounced with just /g/. The "u" was originally there for a reason, but it's been dropped. The sound sound shifted but the spelling didn't. But that happened in french, not in English. English got the word "guarantee" later.

    ThomasK said:
    When followed by frontal or palatal vowels, you mean?
    Click to expand...
    I haven't seen vowels called "palatal" in English; only consonants. But yes, the tongue position of the front vowel makes a forward shift from velar more likely in some languages. In general, a common way for sound shifts to happen is for one sound to take on part of the traits of another sound when they're pronounced consecutively. It's called "assimilation". More examples and classification of different types of assimilation can be found here.

    ThomasK said:
    Would that not hold for German (but then /g/)?
    Click to expand...
    I don't know of any cases in which German /ɡ/ shifted forward like it did in English & French. But I do know that German "ch" is pronounced farther forward after "i" or "e" than after "a", "o", "u", or a consonant: between /x/ and /ʃ/, close enough for non-Germans to mistake it for /ʃ/.

    (The IPA symbol for that sound is /ç/, a very awkward choice because that's a French letter representing /s/ instead of /ç/; apparently they just wanted another symbol that was comparable to "c" but different, and printing presses already had blocks for that shape, so it was easier to work with that than to invent something new that there weren't printing-press blocks for.)

    The tendency for velar consonants to shift forward in some languages, particularly when they're adjacent to another sound with a farther-forward articulation, is only a loose general pattern. Different languages work out the details differently.

    Incidentally, Arabic's letter ج has also had a /ɡ/►/ʤ/ sound shift in most dialects including the international Modern Standard Arabic, which even went all the way to /ʒ/ in some dialects... but it isn't dependent on what sounds come before or after it. In any dialect that has the shift, it's universal, which left them completely without a /ɡ/. That opened up a gap which the letter ق ended up filling in some of those dialects by /q/►/ɢ/►/ɡ/, which didn't happen in dialects which still had the old original /ɡ/ for ج, as if already having that sound was blocking other sounds from shifting into it. This is an example of not only the fact that palatalization of a velar can happen without being prompted by a front vowel, but also the fact that sometimes groups of sound shifts happen together for a reason, with one shift pushing or pulling another along with it.
     
    The short story is that in most varieties, Germanic /g/ had the allophones [g, ɣ, j]: [g] when no vowel preceded (+ when geminate), [ɣ] before and after back/velar vowels, [j] before (and sometimes after) front/palatal ones.* Here and in the following I disregard devoicing, such as [ɣ > x].

    In English, this allophony was maintained and resulted in a phonemic split, with [ɣ] transitioning to [w] and forming a diphthong with back vowels (law, low) which then merged into simple vowels, but when already following a w-final diphthong, it's reflected as /f/ (laugh, tough). Syllable-final [j] also formed diphthongs that also merged with simple vowels (day, feign), and in night etc. it probably merged with /i:/ outright. Syllable-initial [j] became a separate phoneme, or was deleted when unstressed (in the prefix ge-, as in a-seen, y-gone).

    In some cases where one would expect a split development, English defaulted to the [g] allophone. Sometimes this is due to morphology (begin, together), but in other cases could be due to borrowing from Old Norse. And in some lexemes it seems to have merged into the Norman phoneme /ʤ/.​

    High German eliminated the allophony and was left with just [g] everywhere, except when the devoiced [ɣ] merged into /x/. Some other (Low German) varieties opted instead for [j] everywhere, or at least when not syllable-final.

    Northern Dutch varieties also eliminated the allophony but in favour of [ɣ] which further strengthened and retracted into the (warning, entirely objective personal opinion) horrible-sounding “hard g”. The more conservative Belgian Dutch (Flemish) varieties maintain the original allophony to this day. But the j-everywhere approach also exists in some varieties - the Kerkrade dialect has jód for goed 'good'.

    * the terms velar/palatal are indeed applied to vowels in English, and front/back to consonants
     
    Last edited:
    Sobakus said:
    it's reflected as /f/ (laugh, tough).
    Click to expand...
    Those are original /x/s, not /g/s, from OE hlehhan and toh. Why /x/ and its palatalised allophone [ç] sometimes became /f/ and sometimes mute with compensatory lengthening is unclear but it is not a reflex of /g/ having transitioned from [ɣ] to [w].
     
    Last edited:
    @Sobakus: Thomas is from Kortrijk, where g and ch is just one sound that is not palatal but rather close to h. Kortrijk is very much "Northern Dutch" in this regard, despite bordering France (Lille). In Limburg, Antwerp and Flemish Brabant, those are four sounds.
     
    berndf said:
    Those are original /x/s, not /g/s, from OE hlehhan and toh. Why /x/ and its palatalised allophone [ç] sometimes became /f/ and sometimes mute with compensatory lengthening is unclear but it is not a reflex of /g/ having transitioned from [ɣ] to [w].
    Click to expand...
    Dutch had the opposite sound shift:, but only before t: Luft > lucht, gekauft > gekocht, Kraft > kracht
     
    I guess that the original pattern had something to do with region (Salisbury, Middlesborough, Edinburgh) but that the pattern has completely disappeared in lexical items that are not fixed in place!
     
    It rhymes wit how in BrE. This is also true for the town in Berkshire. I have checked with John Walker's dictionary: This was already true 200 years ago. So, /slu/ must be a later AmE shift, unless is it rooted in a dialectal variant I am not aware of.
     
    Last edited:
    You must log in or register to reply here.
    Back
    Top