I'd always assumed assimilation of /r/ to /s/, followed by degemination and apocope, but it occurs to me I have nothing concrete to back this supposition up with.
I don't know any Modern Swedish, but I do know some Old (and a tiny bit of Modern) Icelandic.
Old Norse (of which Old Icelandic is the main literary dialect) had quite a few words ending in consonant+r, where the 'r' must presumably have been almost an extra syllable, like when it acts a vowel in Slavonic languages (Srpski = Serbian). Modern Icelandic has tended to respell these endings as vowel+r, e.g. Old Icelandic maðr ('man'), Modern Icelandic maður, just as Swedish has also inserted a vowel.
In the case of the verb eta, 'he eats' is hann etr - Modern etur. Judging by Gordon's Old Icelandic Grammar (see my comment below) I'd say Modern Swedish äter goes back to etr and earlier *etR, while äts goes back to Old Swedish *ets, for earlier *etR + sR.
I guess there are two bits to this: what's the -s ending, and why the apparent absence of personal endings.
Gordon's Introduction to Old Norse gives reflexive forms of the verb grafa in the present tense as:[ek] grǫfumk [þú] grefsk [hann] grefsk [vér] grǫfumk [þér] grafizk [þeir] grafask In this you can clearly see that the suffix is a reduced version of personal pronouns mik, sik etc. corresponding to modern sig etc.; but still, even in the oldest period it's grefsk and not **grefrsk.
The thing is, Norse final -r often goes back to earlier *-R - a sound of unclear value, but that is assumed to have been something like a Czech ř as in Dvořak: it has a different rune from a normal 'r', and it corresponds to a reconstructed -z in Proto-Germanic (that corresponds to an s in Latin, Greek, Sanskrit, etc.) Sometimes this sound shows up as an 'r' in Old English, like in Norse, but sometimes it's changed back to an 's': in Gothic it's sometimes an 's', sometimes a 'z'. There's a similar change in Latin, so, for example, the root aus- 'ear'
( ... )
Short answer: no, it's borrowed from the 'thou/du' form. The oldest Norse equivalent of English -th is actually -th (well, -iþ: I've got an old book on runic inscriptions which quotes one that ends with a curse on whoever breaks this memorial stone. The word for 'breaks' is bAriutiþ, where later Old Norse would have brýtr (MnIce brýtur, Swed. bryter, according to Google Translate!). Later on the -r ending spreads throughout the whole present tense, except in Icelandic and Faeroese.
Longer answer: the whole business with -s- in some words and -z-/-r- in others (e.g. Mn Eng 'was'/'were') is a mess. Ultimately, it goes back to where the stress accent was on the word in an even earlier stage of the language (stress accent moved around in different cases/verb forms in Proto-Indoeuropean, but had become fixed in Germanic languages). Even if it comes to form part of some sort of regular grammatical pattern - such as English having past participles frorn from 'freeze' or lorn from 'lose' - it's still likely to get dropped sooner or later in
( ... )
Thank you for your comprehensive answer! Very informative. So the *-R ending is from the second person sg. form where -z- turned into -R-, which is an example of Rhotacism
( ... )
Comments 28
Reply
Reply
Reply
Old Norse (of which Old Icelandic is the main literary dialect) had quite a few words ending in consonant+r, where the 'r' must presumably have been almost an extra syllable, like when it acts a vowel in Slavonic languages (Srpski = Serbian). Modern Icelandic has tended to respell these endings as vowel+r, e.g. Old Icelandic maðr ('man'), Modern Icelandic maður, just as Swedish has also inserted a vowel.
In the case of the verb eta, 'he eats' is hann etr - Modern etur. Judging by Gordon's Old Icelandic Grammar (see my comment below) I'd say Modern Swedish äter goes back to etr and earlier *etR, while äts goes back to Old Swedish *ets, for earlier *etR + sR.
Reply
Gordon's Introduction to Old Norse gives reflexive forms of the verb grafa in the present tense as:[ek] grǫfumk
[þú] grefsk
[hann] grefsk
[vér] grǫfumk
[þér] grafizk
[þeir] grafask
In this you can clearly see that the suffix is a reduced version of personal pronouns mik, sik etc. corresponding to modern sig etc.; but still, even in the oldest period it's grefsk and not **grefrsk.
The thing is, Norse final -r often goes back to earlier *-R - a sound of unclear value, but that is assumed to have been something like a Czech ř as in Dvořak: it has a different rune from a normal 'r', and it corresponds to a reconstructed -z in Proto-Germanic (that corresponds to an s in Latin, Greek, Sanskrit, etc.) Sometimes this sound shows up as an 'r' in Old English, like in Norse, but sometimes it's changed back to an 's': in Gothic it's sometimes an 's', sometimes a 'z'. There's a similar change in Latin, so, for example, the root aus- 'ear' ( ... )
Reply
Is the *-R ending in personal verbal forms (active voice) equivalent to the -th ending in Old/Middle English (as in doth, etc)?
Reply
Longer answer: the whole business with -s- in some words and -z-/-r- in others (e.g. Mn Eng 'was'/'were') is a mess. Ultimately, it goes back to where the stress accent was on the word in an even earlier stage of the language (stress accent moved around in different cases/verb forms in Proto-Indoeuropean, but had become fixed in Germanic languages). Even if it comes to form part of some sort of regular grammatical pattern - such as English having past participles frorn from 'freeze' or lorn from 'lose' - it's still likely to get dropped sooner or later in ( ... )
Reply
Reply
Leave a comment