Saterland Frisian (Sfrs. Seeltersk) is the only living remnant of Old East Frisian. It is an endangered language, with an estimated number of 2250 speakers (Stellmacher Reference Stellmacher1998: 27) and is spoken in the municipality of the Saterland (Sfrs. Seelterlound), which is located in the federal state of Lower Saxony in northwestern Germany.
The municipality of the Saterland consists of the four communities Ramsloh (Sfrs. Romelse), Strücklingen (Sfrs. Strukelje), Scharrel (Sfrs. Skäddel), and Sedelsberg (Sfrs. Sedelsbíerig). Saterland Frisian comprises three local dialects spoken in Ramsloh, in Strücklingen, and in Scharrel and Sedelsberg (Sjölin Reference Sjölin1969, Fort Reference Fort2015); there is no standard variety. The dialect of Ramsloh is the most conservative of the three dialects and has maintained several features of Old East Frisian which are lacking in the dialects of Strücklingen and Scharrel. Most, if not all, native speakers of Saterland Frisian born before the 1970s are trilingual, with a hybrid mixture of Münsterland and Emsland Low German as their second language and Northern Standard German as their third language (Fort Reference Fort and Munske2004). Most younger speakers of Saterland Frisian are bilingual with Northern Standard German as their second language. General descriptions of Saterland Frisian are available from Siebs (Reference Siebs1889, Reference Siebs and Pauls1901), Sjölin (Reference Sjölin1969), Kramer (Reference Kramer1982), and Fort (Reference Fort2015) (for an overview of older contributions see Fort Reference Fort1980: 16ff.). Comprehensive dictionaries are Kramer (Reference Kramer1961) and Fort (Reference Fort2015).
The present description is based on the speech of three middle-class trilingual speakers in their 70s and 80s, two female speakers from Ramsloh and Scharrel, and one male speaker from Strücklingen. The description of the vowels is enhanced by acoustical analyses of 13 male middle-class trilingual speakers from Ramsloh aged between 52 and 70 years (see Schoormann, Heeringa & Peters Reference Schoormann, Heeringa and Peters2017). The recording of the narrative as well as the recordings of the examples in the wordlists and in the running text are taken from the female speaker from Ramsloh. The orthographic conventions used throughout the text are adopted from Fort (Reference Fort2015).
Consonants
Note: Symbols in brackets indicate allophonic variants.
Younger speakers tend to devoice voiced (lenis) obstruents in syllable-final position and before voiceless obstruents. Some older speakers realize /ɡ/ as a velar fricative, which is voiced in all positions except before voiceless consonants and in word-final position. [ʔ] is restricted to the beginning of stressed syllables without an onset consonant and often is missing from connected speech.
Postvocalic /n/ is velarized before tautomorphemic /ɡ/ and /k/ (compare monomorphemic [ˈsɪŋl] Singel ‘fencing’, [ˈslɛŋkə] Slänke ‘rut’, and [slɔːŋk] sloank ‘slim’ with bimorphemic [ˈmɛnkən] Mänken ‘man dim’).
In onset position /r/ is an alveolar trill. Younger speakers tend to replace the alveolar trill with the voiced uvular fricative [ʁ]. Prevocalic /r/ may be preceded by /p/, /b/, /t/, /d/, /k/, /ɡ/, /f/, and /v/. After a vowel in word-final position and before a consonant it is [ɐ] ([bɔː
] boar ‘cash adj’, [bɔːst] Boarst ‘fissure’).
The labiodental fricative /v/ tends to become devoiced in word-final position and before voiceless consonants. It may be preceded by /d/, /t/, /k/, and /s/ in the onset ([dvoː] ‘to do’, [tvoː] ‘two’, [kvoːt] ‘bad’, [ˈsvoːdə] ‘crust’). The voiced bilabial approximant [
] is an allophonic variant of /v/, which occurs in intervocalic position after [u] (see section ‘Diphthongs’ below).
The alveolar fricative /s/ becomes voiced word-medially before /j/, after /l/ and /r/, and in intervocalic position if the preceding vowel is long. Before /p/, /m/, /n/, /l/, /t/, and /v/, /s/ tends to become palatalized in younger speakers, resulting in [sʲ] or [ʃ] (e.g. [ˈsʲpitə] ‘spades’, [ˈʃmiːdə] ‘smithy’, [ˈʃnidə] ‘to cut’, [ˈʃlitə] ‘to wear out’, [ˈʃteːtə] ‘to push’, [ˈʃvɪtə] ‘to sweat’). The phoneme sequence /sj/ likewise tends to become reduced to [sʲ] or [ʃ] (e.g. /sjoːdə/ [ˈsʲoːdə] ‘to boil’, /sjʊnɡə/ [ˈʃʊŋə] ‘to sing’). Word-initially, /sk/ is realized as [sk], [sx], [ʃk], [ʃx], or [ʃ] (see Tröster Reference Tröster1997: 17). Word-finally, [sk] is in free variation with [ʃk].
The voiced alveolar fricative /z/ is rare and tends to become devoiced in word-final position. Some speakers do not use the voiced alveolar fricative in any position. The fricative /x/ is restricted to word-medial and word-final position. It is usually a velar fricative but in some speakers palatal variants can be observed after front vowels and uvular variants after low back vowels, as in Northern Standard German (see Kohler Reference Kohler1990). The glottal fricative /h/ is restricted to morpheme-initial prevocalic position.
The palatal approximant /j/ is restricted to prevocalic position, where it may be preceded intra-morphemically by /b/, /t/, /d/, /k/, /ɡ/, /f/, /s/, /m/, /n/, and /l/ (e.g. [ˈbjoːdə] ‘to offer’, [tjoː] ‘to pull’, [ˈdjuːvəl] ‘devil’, [ˈkjoːzə] ‘to choose’, [ˈɡjʊxtə] ‘direction’, [fjuː] ‘fire’, [mjʊks] ‘dung’, [ˈnjuːɡən] ‘nine’, [ˈljuːdə] ‘people’; for /sj/ see the previous comments on /s/). Prevocalic /l/ may be preceded by /p/, /b/, /k/, /ɡ/, /f/, and /s/ (e.g. [ˈplɔːtə] ‘disk’, [blo
t] ‘blood’, [kloːt] ‘dress’, [ɡlik] ‘equal’, [flʊks] ‘instantly’, [slɔt] ‘lock’).
Saterland Frisian allows complex onsets with up to three consonants /klj/, /flj/, /spr/, /spl/, /str/, /stj/, /skr/, /skj/, /skl/, /slj/, e.g. [kljɔn] ‘clew’, [ˈfljoːtə] ‘to flow’, [ˈʃprɔːkə] ‘language’, [ˈʃpliːtə] ‘to split’, [ˈʃtrɔːljə] ‘to shine’, [ˈstjɔŋkə] ‘to stink’, [ˈskriə] ‘to write’, [ˈskjoːtə] ‘to shoot’, [ˈsklɔːvə] ‘slave’, and [ʃljʊxt] ‘bad’. The consonant clusters /trj/ and /krj/ are often realized as [tj] and [kj], respectively (e.g. [ˈtjɔːxtɐ] ‘funnel’, [ˈkjoːpə] ‘to creep’). Inflected verb forms may contain codas with up to four consonants after a short vowel and after /ɔː/ (/mpst/, /ndst/, /nkst/, /lpst/, /ldst/, /lkst/, e.g. [vrɪmpst] ‘writhe 2sg.prs’, [bɪn
st] ‘bind 2sg.prs’, [stjɔːŋkst] ‘stink 2sg.pst’, [hɛlpst] ‘help 2sg.prs’, [jɛlst] ‘hold for 2sg.prs’, [mɛlkst] ‘milk 2sg.prs’).
Vowels
Monophthongs
Saterland Frisian has 10 short and 10 long monophthongs. In addition, the vowel system of Saterland Frisian includes [ə], which is restricted to unstressed syllables, such as the weak form ze ‘they’ included in the examples below.
The close vowels can be divided into three classes: short lax vowels (/ɪ ʏ ʊ/), short tense vowels (/i y u/), and long tense vowels (/iː yː uː/). The mid vowels comprise short lax vowels (/ɛ œ ɔ/), long lax vowels (/ɛː œː ɔː/), and long tense vowels (/eː øː oː/). The open vowels /a/ and /aː/ differ mainly in duration. Scharrel speakers replace /aː/ with /a/ before tautosyllabic alveolar plosives.
The close front rounded vowel /yː/ is largely restricted to loans from Low German. The close tense vowels /i y u/ are shorter than /iː yː uː/ but usually a bit longer and more peripheral than the lax vowels /ɪ ʏ ʊ/ (see Heeringa, Peters & Schoormann Reference Heeringa, Peters, Schoormann, Campbell, Gibbon and Hirst2014). There is some disagreement about the distribution of long and short tense vowels (see Fort Reference Fort1971) and the distinction is on retreat. Whereas there are still a few older speakers who have preserved the opposition between short and long tense vowels, most younger and middle-aged speakers replace the short tense with long tense vowels, which results in binary oppositions between short lax /ɪ ʏ ʊ/ and long tense /iː yː uː/, as in Northern Standard German (see Tröster Reference Tröster1996, Reference Tröster1997; Schoormann et al. Reference Schoormann, Heeringa and Peters2017).
Figure 1 shows mean values of monophthongs produced in nonce words of the form /hVt/ recorded from 13 male speakers of Ramsloh Saterland Frisian aged between 50 and 75 years with three repetitions. To obtain a target vowel the speakers were instructed to read a /hVt/ word with the intended vowel immediately after a rhyming Saterland Frisian word (e.g. Hoot after Sfr. Poot ‘paw’ to elicit /oː/; for further details see Schoormann et al. Reference Schoormann, Heeringa and Peters2017). As the productions of /i y u/ and /iː yː uː/ do not differ either by duration or by F1 or F2, we conclude that the speakers have merged the short and long close tense vowels.
Figure 1 Mean formant frequencies at 50% of vowel duration of short (left panel) and long monophthongs (right panel) in /hVt/ words of Ramsloh Saterland Frisian (after Schoormann et al. Reference Schoormann, Heeringa and Peters2017). Ellipses indicate one standard deviation. Formant extraction was carried out using the acoustical analysis software Praat (Boersma & Weenink Reference Boersma and Weenink2015).
Diphthongs
Saterland Frisian has seven phonemic diphthongs, all of which are falling and closing: /o
/, /ɛ
/, /œ
/, /ɔ
/, /a
/, /ɔ
/, and /a
/ (see Bussmann Reference Bussmann2004: 82, who, however, omits /o
/). The first part of the diphthongs may be lengthened but there are no durational contrasts between the phonemic diphthongs. The diphthong /ɔ
/ is realised as [ɔː
] except before tautomorphemic voiceless plosive (e.g. [brɔː
t] ‘scald 3sg.pst’ with heteromorphemic /t/ vs. [ˈflɔ
tjə] ‘to pipe’ with tautomorphemic /t/; see Bussmann Reference Bussmann2004: 76f., 80f.). The diphthongs may end with a more centralized vowel. This is especially true for /a
/, /a
/, and /ɔ
/, which may be realized as [a
], [a
], and [ɔ
], respectively (see formant values in Figure 2 below). As the location of the final part of these diphthongs in the F1-F2 space is variable, we assume that they are diphthongs which end with a tense vowel but may be realized with different degrees of undershoot (see Fort Reference Fort2015 and Bussmann Reference Bussmann2004). The second part of /œ
/, /o
/, and /ɔ
/ may be both centralized and rounded, resulting in [o
], [œ
], and [ɔ
].
Figure 2 Mean formant frequencies (in Hz) of Ramsloh diphthongs in /hVt/ words measured at 20%, 50%, and 80% of diphthong duration (beginning, center and end of arrow). Left panel: phonemic diphthongs; right panel: allophonic variants of /iːv/, /iv/, /ɪv/, /eːv/, and /ɛv/.
The diphthong /y/, which Fort (Reference Fort2015) finds in a single word, Sküüi ‘meat juice’, was unknown to our speakers. Fort’s (Reference Fort2015) [uːi], which is attested for [truːjə] ‘to threaten’, is an allophonic variant of /uː/ before /j/.
Saterland Frisian has five additional diphthongs, [ɪ
], [i
], [iː
], [eː
], and [ɛ
], which are allophones of the phoneme sequences /ɪv/, /iv/, /iːv/, /eːv/, and /ɛv/, respectively. When /v/ occurs in intervocalic position, the diphthong is followed by a voiced bilabial approximant [
].
Just as the distinction between /iː/, /i/, and /ɪ/ is under pressure, there is a tendency to merge /iːv/, /iv/, and /ɪv/. Table 1 lists the vowel–/v/ sequences that were found to be merged in at least one of the three local dialects by Schoormann et al. (Reference Schoormann, Heeringa and Peters2017).
Table 1 Mergers of vowel–/v/ sequences.
According to Fort (Reference Fort2015: xv) there is an additional diphthong [ɛː
] in the pronoun sääuwen ‘self’, which is a variant of säärm, and which our speakers no longer used.
Figure 2 shows mean values for the phonemic diphthongs (left panel) and allophonic realizations of vowel–/v/ sequences (right panel) in /hVt/ words, which were elicited by using rhyming dialect words and recorded two times from the 13 Ramsloh speakers (see section ‘Monophthongs’ above).
Word stress and intonation
The distribution of primary word stress in Saterland Frisian can be characterized with reference to the phonological word as defined by Wiese (Reference Wiese2000) for Standard German. Specifically, a phonological word in Saterland Frisian can consist of one of the following: (i) a lexical stem, which may be extended by (a) prefixes with a reduced vowel, (b) the prefix /bi/, or (c) suffixes that begin with a vowel; (ii) a prefix with a full vowel (except /bi/); and (iii) a suffix that has a full vowel and begins with a consonant. Morphological words comprise one or more phonological words. In morphological words comprising a single phonological word the primary word stress falls on the last stressed syllable. If they comprise more than one phonological word, the primary word stress falls on the last stressed syllable of the first phonological word. An illustration of how stress works in Saterland Frisian complex words is given below. Phonological words are in parentheses. Stressed syllables are underlined. Note that the rules do also apply to loans like /zəˈlɔːt/ and /ˌeləˈfant/.
Fort (Reference Fort2015) lists a number of compounds which apparently are loans from Northern Standard German (Nsg.) but have a different stress pattern than their German cognates, such as Sfrs. ˌäärmˈzoalig vs. Nsg. ˈarmˌselig ‘miserable’, Sfrs. lichtˈfäidig vs. Nsg. ˈleichtˌfertig ‘careless’, and Sfrs. ˌstjúurˈfräi vs. Nsg. ˈsteuerˌfrei ‘tax-free’. The deviating stress patterns of the Saterland Frisian words may result from treating the German loans as non-compounds consisting of a single phonological word whose last stressed syllable attracts the primary word stress.
Saterland Frisian shares most if not all intonation contours with Northern Standard German (see Fort Reference Fort and Wätjen1995: 527, Peters Reference Peters2008; for Northern Standard German see Grice, Baumann & Benzmüller Reference Grice, Baumann, Benzmüller and Sun-Ah2005 and Peters Reference Peters2014). In today’s language, there is no clear indication of a word accent distinction between gestossener Ton (‘push tone’) and geschliffener Ton (‘dragging tone’), as assumed by Siebs (Reference Siebs1889) for the Saterland Frisian of his time.
Transcription of recorded passage
The passage was recorded from the female speaker from Ramsloh who distinguishes between short and long tense close vowels (see Heeringa et al. Reference Heeringa, Peters, Schoormann, Campbell, Gibbon and Hirst2014). The symbols ǀ and ‖ are used to mark minor and major intonational phrases.
Phonemic transcription
di ˈnɔdəˌviːnd ʊn ju ˈsʊnə ‖ ˈɪnsən ˈstreːdən sɪk di ˈnɔdəˌviːnd ǀ ʊn ju ˈsʊnə ǀ vɛl fɔn do beː ǀ vɛl di ˈstɛːrkərə vaːs ‖ as aːn ˈvɔndərsˌmɔn ǀ di ɪn aːn ˈvoːrmən ˈmɔːntəl ˈpakəd vaːs ǀ dɛn va lɔːŋs koːm ‖ jo ˈvudən ˈeːnɪç ǀ dɛt di fɔr dɛn ˈstɛːrkərə ˈjɛldə skul ǀ di dɛn ˈvɔndərsˌmɔn tvɪŋt ǀ sɪn ˈmɔːntəl ˈuttɔˌlukən ‖ di ˈnɔdəˌviːnd ˈblɔːzədə mɛt ˈalə ˈkrɔːzjə ǀ ˈɔːvər ˈumso moːr hi ˈblɔːzədə ǀ ˈumso ˈfɛːstər ˈpakədə sɪk di ˈvɔndərsˌmɔn ǀ ɪn sɪn ˈmɔːntəl in ‖ ˈɛnt
lk ˈrɔː.tə di ˈnɔdəˌviːnd dɛn kamp ap ‖ nu ˈmɔːkədə ju ˈsʊnə mɛd ˈhirə ˈfryntəlkə ˈstrɔːlən ǀ ju lʊft voːrm ‖ ʊn al ˈɛtər mɪn ˈɔːɡənˌblɪkə ǀ loːk di ˈvɔndərsˌmɔn ǀ sɪn ˈmɔːntəl ut ‖ do ˈmɔːstə di ˈnɔdəˌviːnd ˈtɔˌreːkə ǀ dɛt ju ˈsʊnə fɔn hɪm beː ǀ di ˈstɛːrkərə vaːs ‖
Allophonic transcription
Orthographic version
Die Noudewíend un ju Sunne
Insen streden sik die Noudewíend un ju Sunne, wäl fon do bee wäil die Stäärkere waas, as ʼn Wondersmon, die in ʼn woormen Moantel pakked waas, dän Wai loangs koom. Jo wuden enig, dät die foar dän Stäärkere jäilde skuul, die dän Wondersmon twingt, sin Moantel uuttouluuken. Die Noudewíend bloazede mäd alle Kroazje, oaver jee moor hie bloazede, uumso fääster pakkede sik die Wondersmon in sin Moantel ien. Äntelk roate die Noudewíend dän Kamp ap. Nu moakede ju Sunne mäd hiere früüntelke Stroale ju Luft woorm, un al ätter min Oogenblikke look die Wondersmon sin Moantel uut. Do moaste die Noudewíend toureke, dät ju Sunne fon him bee die Stäärkere waas.
Acknowledgements
The research reported in this paper has been funded by the German Research Foundation (DFG), grant number PE 793/2-1. I thank Marron C. Fort, Wilbert J. Heeringa, and Heike E. Schoorman for their support and all consultants for giving me their time and best efforts. Furthermore, I am grateful for helpful suggestions from the editor and two anonymous reviewers.
