When lexical hybrids become feminine: the declension and agreement behavior of OFr wīf ‘woman; wife’ in Old Frisian and modern Frisian varieties

219

When lexical hybrids become feminine: the declension and agreement behavior of wīf ‘woman; wife’ in Old Frisian and modern Frisian varieties*1

Jürg Fleischer (Philipps-Universität Marburg) & Paul Widmer (Universität Zürich)

1. Introduction: lexical hybridsLexical hybrids are, according to Corbett (2006: 163–165), nouns that dis-play an inherent conflict between their (grammatical) gender and the (natural) sex of their referents. This conflict is mirrored in the agreeing forms of such nouns. An interesting example is attested in all old and many modern West Germanic languages in form of a noun denoting ‘woman; wife’: OE OFr OS wīf, OHG wīb is usually classified as a neuter a-stem in terms of grammatical gender and declension class, though it denotes female persons. The conflict inherent in lexical hybrids is manifest in their agreement be-havior: Agreeing forms to lexical hybrids may either display morphologi-cal marking that corresponds to the grammatical features of the noun in question (formal agreement) or morphological marking that corresponds to the lexical meaning of the noun (semantic agreement). This conflict shapes the agreement behavior of OFr wīf, as illustrated in example (1).12

* The present paper owes its origin to our joint efforts to come to grips with agreement (and its morphology). The part dealing with modern Frisian data (section 5) was presented on the occasion of the 20th Conference on Frisian Philology (Ljouwert, December 12–14, 2014). For the present version we are indebted to Magnus Breder Birkenes, Jarich Hoekstra, Stephanie Leser and Mirjam Marti for their feedback. Needless to state, all errors are ours. Nicoline van der Sijs and the staff of the Meertens Instituut (Amsterdam) made the Dutch Wenker forms accessible to us, which is gratefully acknowledged. Finally, thanks go to Jeffrey Pheiff for smoothing out our English.

1 Sigla of cited Old Frisian texts: R (R1, R2) = Riustring manuscripts, ed. Buma & Ebel (1963); B (B1, B2) = Brokmer manuscripts, B2 ed. Buma & Ebel (1965), digital version of B1: http://kulturerbe.niedersachsen.de/viewer/objekt/isil_DE-1811-HA_STAOL_Best_24_1_Ab_Nr_3/1/LOG_0000; E (E1–E4) = Emsigo manuscripts, ed. Buma & Ebel (1967); F = Fivelgo manuscript, ed. Buma & Ebel (1972); H (H1, H2) = Hunsingo manuscripts, ed. Hoekstra (1950), H2 ed. Buma & Ebel (1969).

220 PHILOLOGIA FRISICA

The definite article thet is clearly a neuter form, whereas the personal pro-noun hia, referring to wif, is feminine:23

1. Jef thet wif askat anne mon in da sinde and queth, thet hi thes bernes feder se, quet hi thenna, thet hi hia ne to wiue wnne …

‘If the:n woman brings a man to court and says that he is the father of her child, and he then states that he has not married her:f …’ (F, §VIII, 22; Buma & Ebel 1972: 68)

As discussed by Corbett (1979, 2006: 206–237), the likelihood of formal or semantic agreement can be expressed in terms of the Agreement Hierar-chy, which reads as follows:

attributive > predicate > relative pronoun > personal pronoun (Corbett 2006: 207)

According to Corbett (2006: 207) “[a]s we move rightwards along the Agreement Hierarchy, the likelihood of agreement with greater seman-tic justification will increase monotonically (that is, with no intervening decrease).” In other words, personal pronouns are most likely to display semantic agreement, whereas attributive agreeing forms, such as adjec-tives or demonstrative pronouns/articles, are least likely to do so. This is nicely illustrated by the above example in that the article displays formal agreement (neuter thet), whereas the personal pronoun embodies seman-tic agreement (feminine hia). In the present paper we will take a closer look at the agreeing forms to OFr wīf and its declensional forms in some older legal texts. As will be discussed further below, Old Frisian is particularly interesting in display-ing feminine forms even in attributive positions, which raises the ques-tion whether we are still dealing with a lexical hybrid. Interestingly, it is primarily the dative singular in which feminine agreeing forms are attested. It is argued that the extension of feminine attributive agreement began in this particular slot of the paradigm because differences between the different genders in the strong declension are absent in the (strong) dative singular. Eventually, the spread of feminine agreement (not com-pleted in Old Frisian yet) lead to the “de-hybridization” of OFr wīf in some

2 In this and the subsequent examples and translations the “targets” of an agreement relation, i.e. the agreeing forms, are singly underlined, whereas the “controller” of an agreement relation (in our case: OFr wīf ) is doubly underlined (terminology according to Corbett 2006: 4). The gender of the agreeing forms is indicated by small-capital n, for neuter, and f, for feminine where necessary.

221

modern Frisian varieties, which can be seen as a consistent development. Interestingly, however, in Old Frisian an interesting geographic pattern with respect to the distribution of both feminine morphology as well as feminine agreement forms can be observed, which is to some extent matched by the modern Frisian varieties. The paper is organized as follows: In section 2 we will present the data situation, limiting ourselves to a discussion of “classical” Old East Frisian legal manuscripts (B1, B2, R1, R2, E1, E2, E3, E4, F, H1, H2). In section 3 we will then discuss a possible diachronic scenario based on the synchronic paradigmatic structure of OFr wīf before assessing the distribution of the inflection and agreeing forms in its geographical setting in section 4. Section 5 provides an overview of the agreement behavior of the modern forms of OFr wīf in West Frisian, East Frisian and North Frisian. Section 6 gives a brief discussion of our results.

2. Old Frisian wīf: agreement and inflection Starting with the rightmost position of the Agreement Hierarchy, a clear picture emerges in that we are not aware of any unequivocal example displaying a personal (or any other anaphoric) pronoun referring to OFr wīf in the neuter. As shown in example (1) above, anaphoric targets of agreement as a rule display feminine forms. This parallels High Ger-man, where anaphoric pronouns are also feminine in their vast majority (see Fleischer 2012: 184–188 for anaphoric demonstrative, personal, and possessive pronouns). For the next position, the relative pronoun, it is important to remark that in Old Frisian relative particles rather than pronouns are often used (see de Haan 2001: 632) and that generally the relative pronoun is much rarer compared to the other relations of the Agreement Hierarchy.34It may therefore not come as a surprise that we are not aware of any example of an inflected gender-distinct form (ei-ther feminine or neuter) of a relative pronoun referring to OFr wīf. The predicate, the next position on the Agreement Hierarchy, usually does not display gender agreement in Old Frisian. This holds for the verb, which never inflects for gender in Frisian (as in Germanic generally). As to the predicative adjective we are not aware of any instances displaying overt agreeing gender morphology controlled by wīf in Old Frisian. Therefore, we will restrict our discussion to attributive targets of gender agreement

3 To provide an illustrative figure: Of a total of 3837 agreeing forms to OHG wīb/NHG Weib (see Fleischer 2012: 172), only some 154, i.e., roughly 4 %, are relative pronouns (see Fleischer 2012: 182).


in the remainder of this article, assuming that feminine agreement is the rule in anaphoric pronouns, whereas the textual records do not allow any generalizations with respect to the relative pronoun. According to the predictions made on the basis of the Agreement Hierar-chy, feminine forms are most unlikely to occur in attributive relations. In High German, this is indeed the case.45In Old Frisian, however, examples for feminine attributive forms can easily be found (cf. Rauch 2007, Brem-mer 2009: 97). Probably the first one to have published this observation was Jacob Grimm (1785–1863) in the fourth volume of his Deutsche Gram-matik, as the following quote (in which “Richth.” refers to Richthofen’s Friesische Rechtsquellen as the textual source) indicates:

Am weitesten geht die altfries. sprache, welche zwar das wort wîf neutral decliniert […], auch den neutralen artikel damit verbindet […]; aber nicht nur das weibliche pron. darauf folgen läßt […] sondern auch beides, den bestimmten und unbestimmten artikel, so wie das possessivum, in weiblicher form damit verknüpft: thâ vif (feminam) Richth. 177a, 5; there wîve (feminae) Richth. 179a […] (Grimm [1837] 1898: 317)

As pointed out by Grimm, in Old Frisian even feminine article forms occur. The following examples illustrate this unexpected agreement be-havior:

2a. Benfrotha andare hondbrede and anda widebene, fiuwer skillingar; there wiwe thrimne further.

‘Injury of the bone of the hand or of the collar bone: (fine of ) 4 shillings; (if done) to the:f woman: one and a half time as much.’ (B1[= B2], §199; Buma & Ebel 1965: 108)

2b. Sa hwersa en wif enedgad werth anda godishuse, sa mi ma thet godishus breka mith thes presteres orloui and helpa there wiue of there nede.

‘Whenever a woman is raped in a church, with the permission from the priest one may open the church by force and help the:f woman out of her need.’ (R1, §XIII, 2; Buma & Ebel 1963: 96)

4 Whereas not a single feminine form agreeing to OHG wīb/NHG Weib is attested for the demonstrative pronoun/definite article among a total of 1023 examples (see Fleischer 2012: 175), of a total of 236 attributive adjectives, only six feminine forms are attested (see Fleischer 2012: 178). Interestingly, five out of these six late Old High German and Middle High German feminine adjectives occur in the atypical postposed position.

223

2c. … and him bi there wiuue en kind mene urde ‘… and he has a child together with the:f woman’ (E1, §IX, 6;

Buma & Ebel 1967: 96)

As is well known, morphological marking indicating the declensional classes (and gender) were already fading away in Old Frisian. For that rea-son Boutkan (1996: 52) in his grammar of the first Riustring manuscript determines the gender of nouns “based on the formal concord of nouns with articles, pronouns and adjectives.” Given the ongoing loss of mor-phological clues on the noun itself, it makes sense to take a look both at the inflectional morphology of OFr wīf in different positions of the para-digm and at the targets displaying agreement to wīf. Table 1 summarizes the morphological patterns of and the attributive forms in agreement with wīf that we found in the “classical” Old East Frisian manuscripts (B1, B2, R1, R2, E1, E2, E3, F, H1, H2; E4 did not con-tain any relevant example). It is restricted to the singular forms because plural forms are relatively rare and not very revealing with respect to the question at hand.56For reasons of comparison, the endings of (uncon-tested) neuter a-stems (to which declensional class OFr wīf belongs, at least diachronically) as well as feminine ō- and i-stems are provided. For the forms of the endings we quote the more traditional, diachronically oriented account by Steller (1928) here; as will become clear further down, the relatively clear morphological distinctions of Steller’s paradigms do not correspond to the reality of at least some textual records.

wīf : endings

wīf : attr. agreeing forms

neuter a-stems: endings

fem. ō-stems: endings

fem. i-stems: endings

nom -Ø neuter -Ø -e, -Ø -e, -Øacc -Ø neuter -Ø -e -e, -Ødat -e neuter, feminine -e -e -egen -e

-esfeminineneuter -es

-e -e

Table 1: Declensional forms of and attributive agreeing forms to OFr wīf (as a comparison: forms of neuter a-stems, feminine ō-stems and fem-inine i-stems according to Steller 1928: 37, 38 and 40, respectively)

5 For most word classes that might display attributive agreement, no gender distinction in the plural survives. As to the declension, for the nominative/accusative plural it is interesting to remark that we are only aware of examples displaying the -Ø ending to be expected in the neuter a-stems (cf. Steller 1928: 37). We are not aware of any nominatives/accusatives ending in -e, as would be typical for feminine ō-stems or i-stems (cf. Steller 1928: 40 and 42), nor the -a ending, as it is occurring in the n-stems (cf. Steller 1928: 42). Thus, *wīve or *wīva as nominative/accusative plural does not seem to exist.


In both the nominative and accusative only forms with no overt ending, i.e. the bare stem wīf, are attested. This form corresponds to what would be expected from a neuter a-stem (such as word; cf. Steller 1928: 37). From a purely morphological perspective, however, it cannot be ruled out that we are dealing with feminine forms. Whereas in many feminine classes an -e could occur, this ending cannot be considered obligatory for all feminine nouns. For instance, in i-stems such as tīd no overt ending occurs in the nominative and accusative singular (see Steller 1928: 40). In ō-stems the -e might be absent in the nominative form in certain lexemes displaying a long root syllable (such as sīd) according to Steller (1928: 38), although here the -e was even analogically restored in some instances. Consequently, Boutkan (2001: 622), who only differentiates between strong and weak masculines, feminines and neuters, lists the morphemes -Ø and -e as endings of the nominative and accusative singular of strong neuters, and -e and -Ø as endings of the nominative and accusative singular of strong feminines (cf. also Boutkan 1996: 54). However, the fact that only neuter agreeing attributive forms are attested can be interpreted in such a way that no change has taken place in these positions of the paradigm yet. At any rate, judging from the declensional forms and agreeing attributes there arises no need to postulate such a change. Things are different in the dative. On a morphological level, only -e is attested. As indicated e.g. by Rauch (2007: 364) this ending occurs in al-most all declensional classes across all genders, cf. the table in Boutkan (2001: 622), where -e is the first morpheme in the dative singular of strong masculines, feminines, and neuters (cf. also Boutkan 1996: 54). In Table 1 this is illustrated for the neuter a-stems as well as for two feminine declension patterns. This vagueness with respect to morphological gen-der marking entails that in this case the gender of wīf is manifest on the agreeing forms only. As the examples in (3) illustrate, the morphologically ambiguous dative wīve controls both neuter and feminine agreement on attributes: In (3a) wiwe triggers neuter agreement on the possessive pro-noun sin-e, whereas in the very same legal and linguistic context the pos-sessive pronoun takes feminine gender morphology on sin-re in (3b).

3a. bi sin-e wiwe wesa ‘to be with his-n wife’ (F, §X, 21; Buma & Ebel 1972: 78)

3b. bi sin-re wiuue wesa ‘to be with his-f wife’ (E2, §VI, 108; Buma & Ebel 1967: 74)

225

In the genitive, the situation is more complicated: Here, in addition to both neuter and feminine agreeing forms two different morphemes are attested. While the genitive in -es corresponds quite well to what would be expected from a neuter a-stem, the form in -e would correspond to many feminine declensional patterns, among others the ō- and i-stems.67 Inter-estingly, there seems to hold quite a clear correlation between the declen-sional forms of the lexeme, on the one hand, and the agreeing attributive forms, on the other. The -es genitive, to be found in neuter (and mas-culine), but hardly in feminine declensional patterns, is combined with neuter agreeing forms exclusively, whereas the -e genitive, to be found in many feminine declensional patterns, occurs with feminine agreeing forms only. This is illustrated in examples (4a) and (4b), respectively. The logically possible *thes/sines wīve and *there/sinre wīves are not attested.

4a. fon des monnes halm twen, <twen> fon thes wiues ‘on the part of the man two (shillings), <two> on (the part) of

the:n woman’ (F, §VIII, 25; Buma & Ebel 1972:70)

4b. Hijr ne mot nenna mon sin-re wiwe gud sella […] ‘Here, no man is allowed to sell his-f wife’s goods’ (E2, §IX, 7;

Buma & Ebel 1967: 152)

3. A diachronic scenario From the discussion of the Old Frisian data we learn that there is a clear-cut morphological dichotomy separating feminine forms of the genitive singular (-e) from neuter ones (-es), whereas the inflectional morphology of the nominative/accusative and the dative is ambiguous. Three basic, quite distinct patterns of inflection and agreement can be observed:

1. The nominative/accusative singular of wīf and its attributes in-flect and behave, as far as attributive agreement is concerned, like regular neuters.

6 According to Grimm ([1837] 1898: 317), who indicates some examples, only the -es form should be interpreted as a genitive, whereas he thinks that the -e forms are datives. Based on functional grounds, however, in example (4b) we cannot see any way to interpret sinre wiwe as something other than a genitive. This holds for other instances of wiwe as well, e.g. sa ne moter nanen mon sinre wiwe blodbreng sella ‘no husband is allowed to sell the dowry of his wife (not: to his wife)’ (B2 §79, Buma & Ebel 1965: 60).


2. The genitive singular features two clearly distinct patterns: if wīf is inflected as a strong neuter (-es), this implies neuter at-tributive agreement; if it shows the -e ending characteristic of many feminine declensional patterns, this implies feminine at-tributive agreement.

3. The dative singular in -e, which is the uniform ending that oc-curs in almost all strong declension classes across all genders, co-occurs with either feminine or with neuter attributes.

Of these three patterns, the dative is crucial when it comes to assessing the historical development. As a result of the vagueness of the association of the dative singular ending -e with a specific gender and declension class, we suspect that it is precisely this paradigmatic slot that, by virtue of the lexical meaning of wīf, invited the intrusion of unambiguously femi-nine attributive agreement forms (cf. Fleischer, Rieken & Widmer 2015: 9). After all, feminine datives in -e co-occur with feminine attributive agree-ing forms, and since the dative wīve is not specifically recognizable as a neuter dative, given the lexical semantics of wīf, it seems almost natural to combine it with a feminine attributive form (recall that as a rule ana-phoric pronouns referring to wīf are feminine, cf. section 2). From here, the feminine agreeing forms spread to the genitive, since in most declen-sional classes feminine genitives and datives cannot be distinguished on formal grounds in the first place. This holds for many nouns, e.g., ō- and i-stems. Also in adjectival and pronominal paradigms there are instances where dative and genitive singular feminine are identical from the begin-ning (e.g. in the strong adjective gōdre is both dative and genitive singular feminine; cf. Steller 1928: 46) or very similar and about to collapse into a single form. For example, this holds for the definite article, where the dis-tinct genitive form thēra is “replaced” with the dative forms thēre and thēr (Steller 1928: 54). As a difference to the dative, however, in the declensional patterns there is some gender distinction in the genitive singular: Here, if feminine agreeing forms are used they trigger a change of the morpho-logical marking from neuter -es to feminine -e, thus, here a significant morphological change is induced. The more grammatical cases nominative and accusative proved to be re-sistant to more unambiguous feminine gender marking in Old Frisian, however, even though such a formal disambiguation could have been realized, e.g. by adding -e to the stem wīf. Note that in Old Frisian syn-cretistic feminine nominative/accusative forms abound (e.g. i-stems such as tīd ‘time’, ō-stems such as ieve ‘gift’). In addition, the ever decreasing importance of suffixal inflectional morphology, which ultimately lead to the loss of suffixal case marking on nouns altogether, supposedly did

227

not cause much systemic pressure towards overt feminine morphology. Nevertheless, the lack of morphological renewal corresponds to the con-sistent use of neuter attributes in the nominative/accusative.

4. Areal differences: Old Frisian So far, we have looked at “Old Frisian” as if it were a uniform language without diatopic differentiation. However, when we examine the patterns that occur in every manuscript interesting distributions can be observed, as will now be discussed.In Table 2 the figures of all unambiguous attestations within the different manuscripts of East Old Frisian legal texts are provided.

nominative/accusative

dative genitive

(-Ø) (-e) -es -e

neut. attr.

fem. attr.

neut. attr.

fem. attr.

neut. attr.

fem. attr.

R1 4 – – 7 3 –R2 1 – – 2 – –B1 1 – – 7 – 2B2 1 – – 7 – 4E1 3 – – 4 2 1E2 – – – 1 1 1E3 3 – – 1 – 1F 6 – 8 2 3 –H1 3 – 3 2 3 –H2 3 – 2 3 3 –

Table 2: Occurrence of forms of wīf with unambiguous gender morpho-logy and/or attributive agreement (e.g. genitive singular ethel-es wiu-es or nominative singular thet wif )

Admittedly, the figures are small and do not allow us to draw any statistic conclusions. Yet, the overall picture is rather consistent with what is known about the development of wīf in general. Table 2 illustrates the fact that the nominative/accusative singular had not yet undergone any morphosyntactic modification. In all manuscripts, it can be interpreted as a traditional neuter a-stem (i.e., with no overt ending) and its attributes


behave like regular neuters.78Furthermore, the figures lend support to our hypothesis that the spread of feminine morphology and agreement started in the dative singular. Feminine agreement is attested in all manuscript groups. In the F and H manuscripts, variation occurs, with the neuter prevailing in F, whereas the H1 and H2 manuscripts display a more even distribution. Interestingly, these two manuscripts (which provide nearly identical texts) differ with respect to one instance: While H1 has the neuter bi sine wiue (Hoekstra 1950: 148, §XV,114), H2 has the feminine bi sinre wiue (§IX,2; Buma & Ebel 1969: 62; Hoekstra 1950: 64) in the corresponding passage. Together with the fact that neuter and feminine in the dative singular are distributed relatively even in the Hunsingo manuscripts (3:2 and 2:3, respectively), this textual variation between two closely related manuscripts can be interpreted as an indication of ongoing change. In the R, B, and E manuscripts, however, only the feminine occurs; neuter agreeing forms are not attested at all. For R1 Boutkan (1996: 71) therefore notes that “[t]he neuter wif […] occurs as a feminine in the D[ative] s[ingular]” (cf. also Boutkan 1996: 52). Thus, in these manuscripts in the dative the shift was already completed by the time these texts were written down.The genitive displays somewhat differing patterns: on the morphologi-cal level, both the morphemes -es (corresponding to strong neuter nouns) and -e (corresponding to strong feminine nouns) are attested. As to the attributive agreement forms, also both neuter and feminine forms occur. However, there are interesting distributions with respect to the different

7 According to Grimm ([1837] 1898: 317), one accusative singular feminine form is attested, as can be learned from his indication “thâ vif (feminam) Richth. 177a”. Grimm refers to Richthofen (1840: 177, first column, line 5) here; the relevant example corresponds to the following text (quoted from a more modern edition): Hwersa thi mon wergat sin wif, ieftha thi hera ieftha thiu frowe ieftha thi sviaring ieftha thiu snore, aeng thira wirgat thene otherne, and hira sziwe se burkuth, and tha redia se thria clagad, sa ielde ma se mith thrim ieldum and tha wif thrimne further and tha liudem sextich merca te fretha […] (B, §173, Buma & Ebel 1965: 102). However, pace Grimm we think that tha wif in this example need not to be interpreted as an accusative singular (as Grimm’s Latin gloss feminam indicates); rather, one can also see a form of the accusative plural here, given that the zero nominal ending is the usual form of neuter a-stems in the nominative/accusative plural (see Steller 1928: 37) and given that the nominative/accusative plural neuter article is thā just as in the accusative singular feminine (cf. Steller 1928: 54). A reading as an accusative plural, not singular, is also suggested by the translation by Buma & Ebel: “Wenn der Mann seine Frau umbringt, oder der Schwiegervater oder die Schwiegermutter oder der Schwiegersohn oder die Schwiegertochter einer den andern umbringt, und ihr Streit dorfkundig ist, und beim Redjeven dreimal Klage erhoben ist, so büße man sie mit dreifachem Wergeld und die Frauen um die Hälfte höher, und (zahle) dem Volke sechzig Mark als Friedensgeld […]” (Buma & Ebel 1965: 103; our emphasis).

229

manuscript groups. Feminine -e genitive forms are missing from R, F, and H manuscripts, in which neuter -es forms prevail. In the E manuscripts the intrusion of feminine forms has lead to an equilibrium, as neuter and fem-inine forms occur side by side with equal frequencies. Finally, in B neuter genitives are lacking (as do neuter datives), with feminine forms prevailing.When arranged according to its geographic distribution and according to the spreading of feminine morphological and agreement forms in the different cases, the data in Table 2 reveals further interesting patterns in terms of dialect distribution.

West ← → EastH F E B R

nom./acc. n. n. n. n. n.genitive n. n. n. and f. f. n.dative n. and f. n. and f. f. f. f.

Table 3: Distribution of morphological and agreement gender of wīf in manuscripts, arranged according to the presumed origin of the manuscripts (cf. Bremmer 2009: 16)

Table 3 shows that as we move from west to east the overall frequency of the feminine increases. The H and F manuscripts, both written west of the Ems river estuary, preserve forms and agreement patterns of wīf that are very close to the inherited paradigm. Here it is only the formally ambiguous dative in -e that optionally allows feminine attributes. The genitive retains all of its inherited morpho-syntactic properties. East of the Ems river estuary we have a completely different situation. The da-tive in -e, which is formally ambiguous, co-occurs exclusively with femi-nine attributes, suggesting that in this geographic area the shift toward feminine agreement had been completed in this slot of the paradigm. As for the genitive three distinct stages are observable. The language of the easternmost Riustring manuscripts did not introduce feminine geni-tives in -e (and, accordingly, feminine attributive agreeing forms). The manuscripts from the more central Emsingo area have both, neuter and feminine morphology as well as attributes. This corresponds to the geo-graphic proximity of these manuscripts to the manuscripts of western provenance that preserve neuter gender in the genitive. Finally, the B manuscripts reflect the most progressive state of Old Frisian inasmuch as all genitives and datives inflect like regular feminine nouns and trigger feminine agreement in their attributes. These findings are somewhat surprising, since B, the oldest manuscript group (B1: between 1270 and 1300; B2: completed 1345; see Bremmer 2009: 13) represents a linguistic stage that is the most progressive of all legal man-


uscripts in our sample. The much younger F manuscript (dating from between 1427 and 1450 according to Bremmer 2009: 13), on the other hand, preserves a state that is much more conservative. The result is that the differences between the manuscript groups make much more sense if they are interpreted geographically rather than diachronically.89

5. Areal differences: modern Frisian Interestingly, the geographic differentiation to be observed in the Weser and Ems Old Frisian legal texts, in which attributive feminine forms can primarily be found in the east (E, B, and R manuscripts) whereas in the manuscripts H and F from west of the Ems river neuter attributive forms prevail, is to some extent mirrored by the modern Frisian varieties. While it would go beyond the scope of the present paper to discuss agreement of the modern forms of OFr wīf extensively, we would like to illustrate the general match and further development with a data set that is avail-able for all modern Frisian varieties, namely, Wenker’s ninth sentence, and corroborate this data set by observations from the lexicographic and grammatical literature.

8 As indicated above, we did not analyze “post-classical” Old West Frisian texts, but the article wîf in Buma’s (1996: 461–463) exhaustive dictionary of the Jus Municipale Frisonum (J) allows to make at least some indications with respect to one West Frisian manuscript. In J there seems to occur only one single unequivocally feminine attributive agreeing form, namely der in Fan der wiue spreke ‘about the legal claim of the wife’, which is the title of §III,50 (ed. Buma & Ebel 1977: 102). Otherwise, only neuter agreeing forms seem to occur. It remains to be investigated whether this single example, which corresponds to feminine declensional patterns but not to traditional neuter a-stems, indicates the same tendency as we observed in the Old East Frisian B and E manuscripts. If the only example displaying feminine agreement in J is valid (one possibility might be to interpret it as a compound, but in comparison to other titles of articles this seems unlikely: unequivocal preposed genitives can easily be found), then it is interesting to note that no feminine agreement in the dative is attested, which does not correspond to the distribution to be found in the East Frisian manuscripts (and poses a problem to our diachronic scenario). Generally, the loss of morphological distinctions seems to be somewhat stronger in J as compared to the “classical” manuscripts. For instance, we find the form wiue in the function of a direct object in dat ma wiue nede nyme ‘that one rapes a woman’ (§VIII,1c; ed. Buma & Ebel 1977: 150). The unexpected form wyf in §I,49, which Buma (1996: 463) lists as a dative singular, can probably also be interpreted as a regular accusative complement of the preposition oen in Wald and onriucht, deer ma deth oen een wyf, deer dragghen js ‘violence and injustice that one commits to a woman that is pregnant’ (ed. Buma & Ebel 1977: 140). Especially in light of the findings with respect to modern West Frisian (see section 5) it is an interesting question whether further examples of feminine attributive agreement and declensional behavior in other Old West Frisian manuscripts can be found.

231

Modern Frisian varieties provide especially valuable data for the follow-ing reason: In German, Weib has acquired a pejorative meaning in the younger stages of the language (see Fleischer 2012: 165–166 and literature cited there), and the same holds for Dutch wijf; interestingly, this devel-opment seems not to have taken place at all in modern East and North Frisian. Only in West Frisian a similar development has taken place after 1500 (see Tamminga 1963: 233–234), as indicated by Tamminga (1963: 233–237) in his discussion of the distribution and lexical meaning of wiif and frou, although the development did not go as far as it did in Dutch. Before discussing the dialectal data some brief indications on the ori-gin of the Frisian Wenker sentences is in order. Beginning in 1879 Georg Wenker (1852–1911), who worked as a librarian in Marburg, sent out forty sentences (which would later become known as the “Wenkersätze”) all over the German Empire, asking schoolteachers to translate them into the local dialects. Interestingly, his data collection was not restricted to German-speaking locations. In addition to many other languages (e.g., French, Polish, Danish), Wenker also received Frisian translations of his sentences. By the fall of 1888 Wenker had 67 Frisian forms at his dis-posal. Wenker ([1889] 2013: 3) provides an exhaustive list of his Frisian locations (six of which are East Frisian, the rest is North Frisian). Almost two generations later the same sentences were also used in the Nether-lands by Pieter Jacobus Meertens (1899–1985). As is indicated by Meertens (1936: 126), Meertens paid a visit to Marburg in 1930 and was convinced by Ferdinand Wrede (1863–1934), Wenker’s successor, to collect Wenker sentences in the Netherlands, which he did with “Vragenlijst No. 3 (1934)” of the Centraal Bureau voor Nederlandsche en Friesche Dialecten.10 Meertens provided a Standard Dutch version of the sentences to be translated by his informants (see Meertens 1936: 125–126). Since the Wenker sentences were used in Germany and the Netherlands, comparable material cover-ing the entirety of the Frisian-speaking area is available. This allows exact comparisons despite the fact that we are dealing with translations, not original material. For our purposes, number 9 of Wenker’s sentences is interesting. It reads as follows:

10 According to Meertens (1936: 125) the survey was conducted in the Netherlands in 1935, although the date printed on the forms is 1934.


Standard German version Standard Dutch version

Ich bin bei der Frau gewesen und habe es ihr gesagt, und sie sagte, sie wollte es auch ihrer Tochter sagen.

Ik ben bij de vrouw geweest en heb het tegen haar gezegd; zij zeide, dat ze het ook aan haar dochter zou zeggen.

‘I was at the woman’s [house] and told her (it) and she said she would tell (it) her daughter.’

The lexeme Frau/vrouw ‘woman’ contained in the original Wenker sentence is a regular feminine noun and as such it is not interesting for our purpos-es. However, in the East and North Frisian locations Frau was rendered by a modern form of OFr wīf almost exclusively, as can be seen from the respec-tive hand-drawn map (WA map 126)1011and Wenker’s handwritten comment (see Wenker [1894] 2013: 315). In the Netherlands, only a few maps have been drawn on the basis of “Vragenlijst No. 3”, the dialect translations of vrouw not being part of them. In the Frisian locations many forms corresponding to vrouw were used. However, modern forms of OFr wīf are attested never-theless in West Frisian (as well as some Dutch-speaking) locations.Provided a modern form of Ofr wīf was used, in West Frisian and Saterlandic, the neuter article is used, while the anaphoric personal and possessive pronouns are feminine.1112This is illustrated by the following three West Frisian and Saterlandic examples1213(since the originals are handwritten, not all details, especially not all diacritics can be rendered in our transliterations):

10 This map is accessible electronically: http://www.regionalsprache.de/SprachGis/Map.aspx (last accessed 05/12/2015).11 According to the Old Frisian data it is to be expected that anaphoric agreeing forms are

feminine; but note that, from a methodological point of view, these forms, if looked at in isolation, could also be explained as reflecting the feminine forms of the original Wenker sentence 9. The Standard German/Dutch version contains feminine anaphoric pronouns, so it cannot be ruled out that we are simply dealing with an effect from the translation here. In the case of the (neuter) article, however, this is different as the original contains a different lexeme (and article).

12 The Wenker forms of Frisian locations in Germany are housed in Marburg at the archives of the Forschungszentrum Deutscher Sprachatlas. They are quoted by their running (usually five-digit) number and the name of the location. The original forms are accessible as scans:

http://www.regionalsprache.de/Wenkerbogen/Katalog.aspx (last accessed 05/12/2015). The forms documenting Frisian locations in the Netherlands are housed at the Meertens

Instituut Amsterdam. They are quoted according to their Kloeke Code and the name of the location. The Dutch Wenker forms are currently being transliterated by volunteers and will be made available online in the future.

233

5. Ik ha by ’t wiif west en hat hjin ha sein; se sei dat se ’t ek oan har dochter sizze scoe. (B014p Paesens)

Iek bin bie ’t wîef west en (iek) hè ’t tjian jer seid; jo seej dòt se’t ek tjian jerre dòòter sizze soêad. (F002p Hindeloopen)

Ik bìn bij t wīf wìst en hèt tsjìn ha saint; sə sai, dǒt sət ìk oan hà dòchtər sìzə soeə. (F036p Oudemirdum)

Ik ben bin det Wiu weesen, un häbe et ir tàu queden, un ju quad, ju wielt ok hier Wucht tàu quede. (47214 Strücklingen)

Ik ben be dat Wiu wesen un hebe et ir to quäden, un ju quad, ju will et uk ihr Dochter tau quäde. (47215 Ramsloh)

Ik benn bi dett Wiw wesen un häbet hir toqueden, un ju quad, ju wilt ok hire Dochter toquede. (47216 Hollen)

The neuter article forms in West Frisian and Saterlandic can be con-firmed by other sources. According to Tiersma (1985: 48) and the Wurdboek fan de Fryske taal 25 (2011: 281), the lexical gender of wiif is neuter. However, as can be seen from many of the examples in the article of the Wurdboek, agreeing anaphoric forms are usually feminine.1314This is illustrated by the following example, which contains both a (clitic) neuter article (’t) and a feminine personal pronoun (hjar), showing thus that we are dealing with a lexical hybrid:

6. Aste ’t wiif de pong jouste, jou hjar den de broek ek mar ‘If you give the:n wife the purse, give her:f the trousers as well.’ (Wurdboek fan de Fryske taal 25 (2011): 283)

As to Saterlandic, the lexeme Wieuw (Fort’s 1980 spelling) or Wiu (Kra mer’s 1982 spelling) occurs with the neuter article dät and is a lexical neuter according to Fort (1980: 184) and Kramer (1982: 12). However, forms other than the article are usually feminine, as attested by the following example displaying a possessive construction with the feminine, not neuter pos-sessive pronoun, but with the neuter form of the article. This is a clear indication that it is a lexical hybrid:

7. dät Wiu hiere Kloodere ‘the woman’s clothes (literally: ‘the:n woman her:f clothes’)’ (Kramer 1982: 15)

13 Similarly, Tiersma (1985: 48) states that the personal pronouns referring to famke ‘girl’, another lexical hybrid, are usually feminine, but he does not explicitly mention wiif in this respect.


In the East Frisian dialect of Wangerooge and in North Frisian, feminine article forms prevail. This is shown by the following examples of Frisian translations of Wenker sentence 9, which illustrate that feminine article forms1415occur in Wangerooge and in the entire dialectally diverse North Frisian area:

8. Ik sin bi ju Wüf’ wissien un hebb herrit twiddien, un ju tweid, ju weilt uck herri Vaun twidder. (48241 Wangerooge)

Ick ben be jö Wöff wesen un hee dat her saät, un jö säe, jö weel dat herr dogter säe. (46638 Altendeich)

Ik ben bai jü Wüf wesen, end hæv dat her sait, end jü sæ, jü wêl dat ok her Doghter sêde. (46573 Langeness)

Ick benn bei äh Wüff wen en hevt ha sejd, en jü säh, jü wielt ak ta ha Dochter sierde. (46699 Horsbüll)

Jck ban bei jö Wüw wän em häw et ha säid, enn jo säi, jö wäujl ock ha Dochter seede. (46702 Deezbüll)

Jk bänn bei ä Wöff wä’en, enn hev’t har säid, enn jö sä, jö wielt ok har Dogter siäde. (46892 Klanxbüll)

Ik sen bi di Wüf wessen, en jü seid, jü wildt uk nog di Fa’men si. (46887 Keitum)

Ik san bi jü Wüf wēsen an hā’t hör said, an jü sād, jü wul ät uk hör Dāāgter sai. (46748 Utersum)

Ik san bi a Wüf wesan, and ha-t her sad, and jü sad, jü wul dat ok her Fåman sai. (46572 Nebel)

Ek bèn bĕrre Wüöff wē iñ èn hò hö eret sōit èn dje ßōit, dje wull he örrem Fōmel et ŏck sōi. (47862 Helgoland)

The Wangerooge data can be confirmed by Siebs (1923: 249), who also provides a (younger) translation of Wenker sentence 9, which contains bī jū wǖf. 1516In the case of North Frisian the feminine article can be cor-roborated by many other sources. For instance, for the Mooring mainland dialect wüf, the modern form of Old Frisian wīf, is listed as a “regular” feminine used with the feminine article jü (not neuter dåt) according to

14 Or, to be more exact, non-neuter forms, since feminine and masculine forms have collapsed into a single common form in some instances (this concerns especially the shorter definite a/e article in many North Frisian locations).

15 It seems possible that the feminine article is relatively young in the case of Wangerooge. According to Ehrentraut (1849: 405; transcription adapted) wǖf is a neuter. In accordance with that, in one of Ehrentraut’s texts we find dait wǖf (Ehrentraut 1854: 41; transcription adapted), with the neuter article dait, not feminine djû (cf. Ehrentraut 1849: 18).

235

Jörgensen (1978: 12).1617The same holds for the dialect of the Wiedingharde, in which wüf takes the feminine article jü (not neuter dat) according to Jensen et al. (1994: 380). In the Sölring insular dialect Wüf takes the ar-ticle di (standing for masculine and feminine), not neuter dit according to Schmidt (1969: 74). For 19th century Fering, Ebert (1998: 270) quotes a source which has jü wüf (which does not correspond to 20th century usage though; see note 17). Thus, according to 19th as well as 20th century North Frisian sources, the modern forms of OFr wīf seem to behave as a regular feminine noun.1718

6. Discussion and conclusion For the history of German, there is no doubt that OHG wīb was and NHG Weib still is a lexical hybrid: Although minor changes can be observed, in the oldest as well as in the youngest records articles are always neuter, whereas anaphoric pronouns are mostly, though not exclusively, femi-nine (see Fleischer 2012). This fits the definition of a lexical hybrid. In Frisian, however, things are different: in addition to feminine attribu-tive agreement, which is attested as early as in Old Frisian, there is also evidence that the declensional pattern of wīf was changed from original neuter to feminine. Therefore, it might make sense to consider wīf a “dual gender word” (Boutkan 1996: 52) or to consider it as a noun “with hetero-clitic gender” (Rauch 2007: 366).

16 Note, however, that in Mooring the neuter article can be used if wüf is used in a pejorative sense (see Sjölin, Walker & Wilts 1988: 275). This is not necessarily a rest of the old neuter though, since the neuter in a pejorative sense seems to be productive in Mooring: As Jörgensen (1978: 13) states, both feminines and masculines can be used as neuters if one wants to express something diminishing or pejorative (“wan hum wat latjs un feråchtliks ütkrööge wal”). For our purposes it is important to note that this use of the neuter is not restricted to (original) lexical hybrids displaying a conflict between neuter and feminine, as the example dåt hün (Jörgensen 1978: 13) shows: According to Sjölin, Walker & Wilts (1988: 109) hün ‘dog’ is a masculine.

17 Unless, of course, feminine and neuter have collapsed into one gender, as is the case in the 20th century insular dialects of Föhr and Amrum (see e.g. Wahrig-Burfeind 1989: 193, Hoekstra 1996). Note, however, that according to Ebert (1998) this collapse is far from complete yet; specifically, as indicated by Ebert (1998: 270), wüf can be used both with “neuter” at and “common” a, at wüf meaning ‘woman’, a wüf meaning ‘wife’. We cannot discuss these more modern developments here. For our purposes, it is most important to state that, as the above examples from Utersum and Nebel show, there was a stage in late 19th century Fering-Öömrang in which the feminine (or non-neuter) article forms are attested with wüf.


However, Old East Frisian is not uniform in this respect. As far as inflec-tion is concerned, the Old Frisian data presented above indicate that wīf is still a neuter in the western Ems Old Frisian Hunsingo and Fivelgo manuscripts: there is no evidence for the -e genitive, nor for any other form corresponding to a feminine nominal class exclusively. Of course (and pace Rauch 2007), this is not contradicted by the feminine agreement forms that are triggered by wīf on anaphoric pronouns: This behavior simply reflects the fact that wīf is a lexical hybrid. In the east, the very same holds for the Riustring manuscripts where feminine dative attrib-utes are attested, but no unequivocally feminine declension forms of the noun wīf itself. As a matter of fact, Boutkan’s (1996) classification of OFr wīf as a feminine in the dative singular cannot be based on morphology, which is the same in all genders in the strong dative singular, but exclu-sively on the fact that attributive feminine agreement forms are attested. Only in the Brokmer and Emsingo manuscripts it might be unavoidable to consider wīf a “dual gender word” from a morphological point of view. In the genitive singular, we find the -e form exclusively (B) or in variation with -es (E), thus, a form that fits feminine declensional classes such as the feminine ō-stems, but not neuter a-stems, whereas in the nominative/accusative we rather find the neuter a-stem -Ø ending than the pervasive feminine e-ending. However, even if we base our assessment as to the gender of OFr wīf on agreement it is clear that wīf is not yet a canonical feminine noun, because attributive agreeing forms in the nominative/accusative are still neuter without exception, and this holds even in the most progressive Brokmer texts.The situation is different in the modern North Frisian varieties and in the Wangerooge dialect, however. Here, the incipient increase of feminine at-tributive agreement that we encounter most prominently in eastern Ems Old Frisian and in Weser Old Frisian has come to its completion. In these modern varieties, OFr wīf has arguably become “de-hybridized” and devel-oped into a regular feminine noun. Since there are no declensional class-es left and since all agreeing forms are exclusively feminine, no traces of the original hybrid behavior can be seen (cf. Fleischer, Rieken & Widmer 2015: 8–9). In the Hunsingo and Fivelgo manuscripts from west of the Ems river, however, this development did not take place, and the same holds for some modern varieties (Saterlandic and West Frisian), in which we are still dealing with a lexical hybrid (as would be the case in Dutch and German, respectively). Thus, unless nominal gender is entirely given up, as it is the

237

case in modern English,1819hybrid lexical behavior can be retained in sys-tems that display only a twofold opposition between neuter and common nominal gender such as in very many Germanic varieties on the southern North Sea littoral. Interestingly, the clear geographic patterns with respect to the agreement behavior of Ems and Weser Old Frisian are reflected to some extent by the modern Frisian varieties. Thus, dialectal differences with respect to agreement seem to be diachronically persistent in this case. The geographic variation with respect to the declensional pattern and the attributive agreement forms of OFr wīf can be considered additional evidence for the assumption that within the various Old East Frisian legal manuscripts we are dealing with diatopic differences rather than with diachronic developments. In our case, interestingly, the differences ob-served in the oldest attestations are mirrored to some extent in the mod-ern dialects, some of which have gone consistently further in a develop-ment that was already attested in some of the oldest sources of Frisian. It remains to be investigated at what date exactly Frisian wīf gave up the last traces of its original hybrid agreement behavior in the eastern varieties.

> Quoted textsB1: http://kulturerbe.niedersachsen.de/viewer/objekt/isil_DE-1811-

HA_STAOL_Best_24_1_Ab_Nr_3/1/LOG_0000B2: Buma, Wybren Jan & Wilhelm Ebel 1965. Das Brokmer Recht.

Göttingen: Vandenhoeck & Ruprecht.E1–E4: Buma, Wybren Jan & Wilhelm Ebel 1967. Das Emsiger Recht.

Göttingen: Vandenhoeck & Ruprecht.F: Buma, Wybren Jan & Wilhelm Ebel 1972. Das Fivelgoer Recht.

Göttingen: Vandenhoeck & Ruprecht.H1+H2: Hoekstra, J. 1950. De eerste en de tweede Hunsinger Codex. ’s Graven-

hage: Nijhoff. H2: Buma, Wybren Jan & Wilhelm Ebel 1969. Das Hunsingoer Recht.

Göttingen: Vandenhoeck & Ruprecht. J: Buma, Wybren Jan & Wilhelm Ebel 1977. Westerlauwerssches Recht

I: Jus Municipale Frisonum. Erster Teil. Göttingen: Vandenhoeck & Ruprecht.

R1+R2: Buma, Wybren Jan & Wilhelm Ebel 1963. Das Rüstringer Recht. Göttingen et al.: Vandenhoeck & Musterschmidt.

18 The development of both the morphology and agreement behavior of OE wīf could provide an interesting comparison for the question how and when exactly the “de-hybridization” took place in the history of English.


> ReferencesBoutkan, Dirk F. H. (1996): A concise grammar of the Old Frisian dialect of the

first Riustring manuscript. Odense: Odense University Press.Boutkan, Dirk F.H. (2001): Morphology of Old Frisian. Handbuch des

Friesischen/Handbook of Frisian Studies, Horst Haider Munske (ed.), 620–626. Tübingen: Niemeyer.

Bremmer, Rolf H. Jr. (2009): An introduction to Old Frisian. Amsterdam: Benjamins.

Buma, Wybren Jan (1996): Vollständiges Wörterbuch zum westerlauwersschen Jus Minicipale Frisonum. Ljouwert: Fryske Akademy.

Corbett, Greville G. (1979): The Agreement Hierarchy. Journal of Linguistics 15: 203–224.

Corbett, Greville G. (2006): Agreement. Cambridge: Cambridge University Press.

Ebert, Karen H. (1998): Genussynkretismus im Nordseeraum: die Resistenz des Fering. Sprache in Raum und Zeit 2, Winfried Boeder et al. (eds.), 269–281. Tübingen: Narr.

Ehrentraut, H[einrich] G[eorg] (1849): Mittheilungen aus der Sprache der Wangeroger. Friesisches Archiv 1, H[einrich] G[eorg] Ehrentraut (ed.), 3–109, 338–416.

Ehrentraut, H[einrich] G[eorg] (1854): Mittheilungen aus der Sprache der Wangeroger. Friesisches Archiv 2, H[einrich] G[eorg] Ehrentraut (ed.), 1–84.

Fleischer, Jürg (2012): Grammatische und semantische Kongruenz in der Geschichte des Deutschen: eine diachrone Studie zu den Kongruenzformen von ahd. wīb, nhd. Weib. Beiträge zur Geschichte der deutschen Sprache und Literatur 134: 163–203.

Fleischer, Jürg, Elisabeth Rieken & Paul Widmer (2015): Introduction: the diachrony of agreement. Agreement from a diachronic perspective, Jürg Fleischer, Elisabeth Rieken & Paul Widmer (eds.), 1–25. Berlin: de Gruyter.

Fort, Marron C. (1980): Saterfriesisches Wörterbuch mit einer grammatischen Übersicht. Unter Mitarbeit von Hermann Dumstorf. Hamburg: Buske.

Grimm, Jacob. ([1837] 1898): Deutsche Grammatik 4. Herausgegeben von Gustav Röthe. Gütersloh [Reprint 1967 Hildesheim: Olms.]

Haan, Germen J. de (2001): Syntax of Old Frisian. Handbuch des Friesischen/Handbook of Frisian Studies, Horst Haider Munske (ed.), 626–636. Tübingen: Niemeyer.

Hoekstra, Jarich (1996): Transitive pronouns and gender syncretism in Fering-Öömrang (North Frisian). NOWELE 27: 45–66.

239

Jensen, Peter et al. (1994): Freesk uurdebuk: Wörterbuch der Wiedingharder Mundart. Auf der Grundlage eines Manuskripts von Peter Jensen zusam men gestellt und bearbeitet von Adeline Petersen, Bo Sjölin, Alastair G. H. Walker, Ommo Wilts. Herausgegeben von der Nordfriesischen Wörterbuchstelle der Christian-Albrechts-Universität Kiel. Neumünster: Wachholtz.

Jörgensen, V. Tams (1978): Kort spräkeliir foon dåt mooringer frasch. 4. aplååge. Bräist: Noordfriisk Instituut.

Kramer, Pyt (1982): Kute Seelter Sproakleere: Kurze Grammatik des Sater-friesischen. Rhauderfehn: Ostendorp.

Meertens, P. J. (1936): Niederländische Mundartforschung. Zeitschrift für Mundartforschung 12: 125–127.

Rauch, Irmengard (2007): Gender semiotics, Anglo-Frisian wīf, and Old Frisian noun gender. Advances in Old Frisian philology, Rolf H. Brem-mer Jr., Stephen Laker & Oebele Vries (eds.), 357–366. Amsterdam: Rodopi.

Richthofen, Karl (1840): Friesische Rechtsquellen. Berlin: Nicolai.Schmidt, Hermann (1969): Wörterbuch der Sylterfriesischen Sprache (Söl’ring

Uurterbok). Keitum: Söl’ring Foriining.Siebs, Theodor (1923): Vom aussterbenden Friesisch der Insel Wangeroog.

Zeitschrift für deutsche Mundarten 18: 237–253. Sjölin, Bo, Alastair G.H. Walker & Ommo Wilts (1988): Frasch uurdebök:

Wörterbuch der Mooringer Mundart. Herausgegeben von der Nord-friesischen Wörterbuchstelle der Christian-Albrechts-Universität Kiel. Neumünster: Wachholtz.

Steller, Walther (1928): Abriss der altfriesischen Grammatik. Halle: Niemeyer.Tamminga, Douwe Annes (1963): Op ’e taelhelling: losse trochsneden fan Frysk

taellibben. Boalsert: Osinga. Tiersma, Pieter Meijes (1985): Frisian reference grammar. Dordrecht: Foris. Wahrig-Burfeind, Renate (1989): Nominales und pronominales Genus im

südlichen Nordseegebiet. München: Tuduv.Wenker, Georg ([1889–1897] 2013): Schriften zum Sprachatlas des Deutschen

Reichs: Gesamtausgabe. Band 1: Handschriften: Allgemeine Texte, Karten-kommentare 1889–1897. Herausgegeben und bearbeitet von Alfred Lameli unter Mitarbeit von Johanna Heil und Constanze Wellendorf. Hildesheim et. al.: Georg Olms.

Wurdboek fan de Fryske taal/Woordenboek der Friese taal 25: v–z. Ljouwert/Leeuwarden 2011: Fryske Akademy.

When lexical hybrids become feminine: the declension and agreement behavior of OFr wīf ‘woman; wife’ in Old Frisian and modern Frisian varieties

Documents