The Modern English –er-suffix is marked by an amazing productivity and polysemy in semantic fuctions. Its functions include an agent noun (e.g. ModE worker), instrument noun (e.g. ModE pounder) and patient noun (e.g. ModE sticker). Its pedigree goes back to Common Germanic times and cognates to the English suffix are found in all Old Germanic languages, including the oldest attested literary language, Gothic. The question which is elaborated on in this blogpost concerns the origin of this suffix. Where did the suffix come from and how can we explain the forms in which the suffix pops up in the Germanic languages?
Casaretto (2004: 423) in her Nominale Wortbildung der gotischen Sprache states that the Germanic suffix *-arija-, the ancestor of the English –er-suffix, spread by way of the numerous Latin/Romance loanwords that contained the suffix –ārius (= Rom. *-árjo), e.g. Lat. monētārius “minter”, molinārius “miller”, tolonārius “publican” → OHG munizāri, mulināri, zolanāri. In Gothic there are relatively few formations in *-areis, e.g. Goth. bokareis “scribe”, liuþareis “singer,” which, along with the fact that many agentive formations which appear in Gothic with the suffix *-ja, e.g. Goth. fiskja “fisherman,” are substituted in West-Germanic with *-arija-, e.g. OE fiscere, point to its later productivity. The quantity of the vowel of the Germanic suffix is the main problem. Whereas most Germanic dialects point to a short *a-vowel, Old High German and its offshoot Middle High German have two suffix variants. The one suffix variant, OHG –eri/-iri / MHG –ere, points to a short a-vowel and the other suffix variant, OHG –āri / MHG –ǣre, points to a long *ā-vowel. For Old High German the vowel length of the suffix –āri is confirmed by circumflex markings in the manuscripts and metrical evidence in Notker and Otfrid. For Middle High German the length of the vowel is confirmed by the preservation of vowel timbre in unaccented position and length marking in the manuscripts. It should be remarked that for Gothic the vowel length in the suffix –areis cannot be ascertained since the Gothic script did not mark vowel length. It is conceivable that the a in the Gothic –areis-suffix was long. Below you will find a table of the different forms of the suffix in the Old Germanic dialects
|Goth.||-areis||< CGm. *-arija-?/*-ārija-|
|OIc.||-ari/-eri||< CGm. *-arija-|
|OE||-ere||< CGm. *-arija-|
|OS||-ari||< CGm. *-arija-|
|OHG||-ari||< CGm. *-arija-|
|OHG||-āri||< CGm. *-ārija-|
|MHG||-ǣre||< OHG –āri|
On the basis of this data, Kluge in his Nominale Stammbildungslehre der altgermanischen Dialecten (1886: 8) reconstructs a Germanic *-arja-/*-ērja-suffix which he deems to be an inner-Germanic creation (< PIE *-ori̯o-/*-ēri̯o-) paralleled in the Old Irish –aire/-ire-suffix (e.g. OIr. cornaire “horn-player”) and the Old Church Slavonic –arь–suffix (e.g. OCS rybarjь “fisherman”). However, both the Celtic suffix (see De Bernardo Stempel 1999: 347-349) and the Slavic suffix (see Pronk-Tiethoff 2013: 100-101) should be regarded as loans ultimately from Latin/Romance. The Slavic arь-suffix, according to Pronk-Tiethoff, would have reached Slavic from Latin/Romance via a Germanic intermediate stage. The Slavic *a could then be explained by a long Germanic *ā if the suffix was loaned early (Proto-Slavic) or a short Germanic *a if the suffix was loaned later (Common Slavic, 6-8th century CE, after the rise of new timbre oppositions, l.c.).
If we assume that the Gothic suffix –areis contained a short a-vowel, all Germanic languages except Old High German would then point to a Common Germanic suffix *-arija- which might have been loaned from a Latin stage –ārius or a Romance stage *-árjo. Since both in Latin and Romance the vowel would have been pronounced long because of the accent (Ten Brink’s Law) it is surprising that the suffix was borrowed in Germanic with a short vowel. We could speculate that this is due to the fact that at the date of the loan Germanic may only have had long nasalized *ą̄ (< PGm. *-anh-) and a long *ǣ (PGm. *ē1). Therefore Germanic short *a would have been the most logical sound substitution. Otherwise we must assume that most Germanic languages secondarily shortened the ā-vowel in unaccented position as is done by Krahe-Meid (1967: 82-83). This is possible but leaves the doublet found in Old High German unexplained. If we depart from an original borrowing of the suffix as Common Germanic *-arija- the length of the ā-vowel in the OHG suffix –āri can be explained as a later secondary borrowing of the suffix from Romance at a time when West-Germanic did have a long *ā which arose after the loss of the nasal feature of WGm. *ą̄ < *-anh- (see also Casaretto 1999: 423 with reference to Wilmanns 1899: 284 and Henze 1965: 158). It is clear that the Romance donor language cannot have been Gallo-Romance since there we find umlaut of *-arjo to *-ęrjo which happened so early (before the fourth century?) that the *ę was still able to undergo primary diphthongization (cf. ModFr. –ier). Whatever the case, a variant *-ārija– with a long *ā must be reconstructed to account for the Old High German form. In this way, by assuming two stages of borrowing of the suffix, the suffix variants within the Old Germanic dialects can be satisfactorily explained
2004 Nominale Wortbildung der gotischen Sprache; die Derivation der Substantive, Carl Winter Verlag, Heidelberg.
De Bernardo Stempel, Patrizia
1999 Nominale Wortbildung des älteren Irischen; Stammbildung und Derivation, Max Niemeyer Verlag, Tübingen.
1965 Deutsche Wortbildung, Tübingen.
1886 Nominale Stammbildungslehre der altgermanischen Dialecte, Sammlung kurzer Grammatiken germanischer Dialecte; Ergänzungsreihe I, Max Niermeyer, Halle.
Krahe, Hans et Wolfgang Meid
1967 Germanische Sprachwissenschaft III: Wortbildungslehre, Walter de Gruyter & Co, Berlin.
2013 The Germanic Loanwords in Proto-Slavic, Leiden Studies in Indo-European, Rodopi, Amsterdam, New York.
1899 Deutsche Grammatik. Gotisch, Alt-, Mittel- und Neuhochdeutsch. 2. Abteilung: Wortbildung. Strassburg. Reprint Berlin, Leipzig: 1922.