Written by Dyami Millarson

Molkwerum Frisian, the native language of Molkwerum (a coastal village with 200+ inhabitants at the time of writing that is located in Southwest Frisia), went extinct in the 19th century. Molkwerum Frisian has traditionally enjoyed an almost mythical status among the Frisians of the Netherlands. Molkwerum Frisian, along with the closely related Hindeloopen Frisian, was famed among Frisians as being “authentic Frisian” owing to its linguistic conservatism. Molkwerum Frisian is, in fact, an isomorphic language, because it did not change much throughout the ages; while Molkwerum Frisian is largely equal in form to what it was before, Molkwerum Frisian, like Hindeloopen Frisian, stayed true to its medieval roots, which meant that both Molkwerum Frisian and Hindeloopen Frisian were in the 19th century a living continuation of medieval Frisian. Calling it “authentic Frisian” as people did in the past has some merit, as it expresses an appreciation of antiquity; Frisians in the past were already very keen on the antiquity of Frisian.

While the assertions of past Frisians about Molkwerum Frisian may seem mythical, there is a kernel of truth to it. Indeed, Molkwerum Frisian is ultimately descended from Old Frisian, but the language itself is not Old Frisian. However, it is old Frisian in the sense that the Molkwerum Frisian language is very archaic. So, to the defense of the Frisians of yore believed in the antiquity of Molkwerum Frisian, there is ambiguity in Frisian and Dutch between “Old Frisian” (as a scientific name of a language) and “old Frisian” (an archaic form of the Frisian language), because just like when one says “Old Frisian” and “old Frisian” in English, there is no distinction between the pair as they are homophones, which are clearly distinguished only in writing (as the former has “Old” with an upper case letter due to being part of a name and the latter has “old” with a lower case letter as it is merely descriptive adjective which affirms the antiquity of a certain kind of Frisian).

In other words, the assertions of Frisians in the past can be taken to be true when they are interpreted as meaning “old Frisian” and “ultimately descended from Old Frisian”. It only requires a slight tweak to be able to see the truth in the assertions from people living in the past. Furthermore, we should definitely be sympathetic to them since they were expression an admiration of both Molkwerum Frisian and Hindeloopen Frisian; they were showering Molkwerum Frisian and Hindeloopen Frisian with praise and they even went on pilgrimages to Molkwerum and Hindeloopen to hear the sound of what was spoken there, this is some praiseworthy dedication to Frisian antiquity! If people had dedicated the same kind of interest to Molkwerum Frisian in the present, then it would have been revived much earlier already. One has to be genuinely zealous, otherewise the aim of bringing Molkwerum Frisian back from the dead cannot be achieved. While the Frisians of yore treated Molkwerum Frisian and Hindeloopen Frisian as sacred, I will treat the revival of Molkwerum Frisian as a sacred mission that must be carried out to completion.

Phonology

Hereby I present my preliminary or tentative phonology of Molkwerum Frisian after studying in painstaking detail all letters and syllables of the texts and words that have been handed down to us. Please bear in mind that I used the spelling of T. van der Kooy for reconstructing the pronunciation. Molkwerum Frisian has never before been written in this spelling, but it is very useful to do so. While the pronunciation is reconstructed and the words have never been attested in the phonetic spelling of T. van der Kooy, I have marked all words with an asterisk. It has never been attempted before to present an overview of the full phonological inventory of Molkwerum Frisian. However, this is absolutely necessary from the perspective of the revival of the language. The phonology is the place to start for any potentially successful attempt at the revival of an extinct language.

Monophthongs	Phonetic transcription	Examples
a	/a/	*waskje (to wash), *jarre (waste of livestock usually found at a farm: liquid manure, dung water)
aa	/ɑː/	*aame (grandmother)
ââ	/aː/	*ââld (old), *fâân (woman), *lâând (land)
e, i (in -ing, -ig)	/ə/	*de (the), *’e (the), *ketting (necklace), *wòddig (worthy)
e (è)	/ɛ/	*het (what), *flesk (flesh, meat), *bjer (bear), *settel (kettle)
èè	/ɛː/	*rèèg (back), *grèète (great, large), *stèèwe (a long object: stave, rod, bar), *sèè (a source of water, water well)
ee	/eː/	*sjeek (sick), *leep (a kind of bird: lapwing), **sees (cheese)
i	/ɪ/	*jimme (you, 2nd pers. plur. informal) *fisk (fish), *lippe (lip), *sjippe (soap), *libbe (to live), *linnen (linnen)
éé, ee (before r and l)	/ɪː/	*déé (dead, death), *kéémer (chamber, room), *feer (father), **meer (more), *neel (nail)
ie	/i/	*jie (you, 2nd pers. formal) *wiend (wind), *brienge (bring), *ien (in), *bie (by, at)
ii	/iː/	*iis (ice), *tiisdei (Tuesday)
ò	/ɔ/	*bòd (bed), *sòrke (church), *sòlm (self), *enkòrm (each other), *Mòlkòrrer (pertaining to Molkwerum), *bòn (child)
òò	/ɔː/	*dòònsje (dance)
ó	/o/	*tó (to, used with infinites), *róst (???), *dóbber (float of a fishing rod), *jók (yoke)
óó	/oː/	*góóld (gold), *nóós (nose), *hóólt (wood)
ö	/ɜ/	*böst (breast)
oe	/u/	*doe (thou, 2nd pers. sing. informal), *loes (louse), *moes (mouse), *foeke (???), *toenger (thunder), *moer (wall)
ú	/œ/	*stút (tail)
úú	/øː/	*húúp (hip)
uu	/y/	*wuute (to know), *kruus (cross), *huuspanne (rooftile), *kuut (a part of the body: calf)

Diphthongs	Phonetic transcription	Examples
ai	/ai/	*saing (song)
aaₑ	/ɑːə/	*aaₑte (grandfather), *haaₑnd (hand)
aauw	/ɑːu/	*wienkbraauw (eyebrow), *gaauw (soon), *klaauwen (claws)
ééₐ	/ɪːɔ/	*fééₐr (father), bééₐs (a kind of fish: bass)
ééi	/ɪːi/	*wééije (to blow)
ei	/ɛi/	*weide (???), *reit (reed)
eeuw	/eːu/	**leeuwe (to believe)
iₐ	/ɪɔ/	*miₐlt (an internal organ: milt, spleen)
ieₑ	/iə/	*stieₑn (stone), **mieₑne (to think)
òi	/ɔi/	*dòiter (daughter), *tòit (thought, pret. tense)
òòi	/ɔːi/	**mòòi (beautiful)
òₐ	/ɔɔ/	*bòₐn (child),**hòₐt (heart), *hòₐdstéé (hearth), *kòₐse (candle), *gòₐs (grass)
òòₐ	/ɔːɔ/	*tòòₐst (thirst)
òòₑ	/ɔːə/	*mòòₑlken (milk)
óóₐ	/oːɔ ~ oːə/	*kóóₐts (fever), *fóóₐne (a kind of fish)
öi	/ɜi/	*möie (midwife), *flöite (flute), *löiwaigen (scrubbing brush)
oeₑ	/uə/	*tjoeₑne (to perform magic)
ou	/ɔu ~ au/	*oujette (to pour out), *houkje (to clean the ditches, usually of a farm)
uo	/uo/	*ljuocht- (???)

Consonants	Phonetic transcription	Examples
b	/b/	*boer (farmer), *boerin (female farmer), *bemmen (trees), *blómmen (flowers)
ch	/x/	*sechste (sixth), *waachse (to wax, grow)
d	/d/	*den (then), *dòònsje (to dance), *nóósdoeₑk (handkerchief)
f	/f/	*fienger (finger), *foegel (fowl, bird), *flamme (flame)
g (initial)	/g/	*walgje (to be disgusted)
g (intervocal)	/ɣ ~ g/	*swééger (brother-in-law), *brèège (bridge), *foegel (fowl, bird)
g (final)	/x/	*éég (eye)
h	/h/	*hemd (shirt), *hoenger (hunger), *hââ (head)
j	/j/	*jarre (waste of livestock usually found at a farm: liquid manure, dung water)
j (after a consonant)	/ʲ/	*fiskje (to fish), *waskje (to wash)
k	/k/	*klaiwer (clover), *kessen (cushion, pillow), *koê (cow), kéél (calf), *skóólder (shoulder), *skiere (pair of scissors), *kat (cat)
l	/l/	*léén (reward, wage), *lééwer (liver), *lekken (bed sheet)
l (final)	/ɫ/	*hóólt (wood), *góóld (gold), *ââld (old), **wrââld (world)
m	/m/	*mem (mother), *maikje (to make)
n	/n/	*nei (to, preposition), *neel (nail), *nekke (neck)
ng	/ŋ/	*tange (a kind of tool: forceps, a pair of tongs), *joenge (boy), **toenge (tongue)
ng (separated by a syllable)	/ŋ.g/	*iengewand (internal organ of the body)
nk	/ŋk/	*bank (bench, sofa), *tienke (to think), *drienke (to drink)
p	/p/	*spieₑ (to vomit)
r	/r/	*riene (to rain)
s	/s/	*slieₑpe (to sleep), *strieₑ (straw)
sj	/ʃ/	*sjippe (soap), *sjipsóp (soap suds)
t	/t/	*tòòₐst (thirst), *toen (garden), *tómme (thumb), *trieₑ (thread)
tj	/c/	*tjoeₑne (to perform magic)
w	/v/	*wònsdei (Wednesday)
w (after a consonant)	/ʷ/	*swelle (a kind of bird: swallow)
z	/z/	*wèèze (to be), *deizenstók (a bar for stewing meat above the hearth)

Phonological remarks:

• The phonology of Molkwerum Frisian is closely related to that of Hindelopen Frisian.
  • Hindeloopen Frisian may be used to reconstruct some phonology distinctions such as the three-way distinction between éé, ee and èè and the two-way distinction between óó and òò.
    • Due to all the idiosyncratic sounds, reconstructing Molkwerum Frisian phonology feels like reconstructing that of a proto-language, such as Indo-European or Proto-Germanic.
• I initially reconstructed *aate and *hâând, but based on the well-attested Hindeloopen Frisian, it should be *aaₑte (grandfather) and *haaₑnd (hand) with the diphthong aaₑ. Nevertheless, based on the Molkwerum Frisian source materials, one has to conclude that there is remarkable phonological similarity between the Hindeloopen Frisian and Molkwerum Frisian. Therefore, where a case of phonological uncertainty may arise, I am inclined to follow Hindeloopen Frisian as a model, because its phonology is well-documented and it is the most closely related to Molkwerum Frisian. For instance, one may wonder why it should not be *hâând like *lâând; after all, these are pronounced the same in the other Frisian languages: Schiermonnikoog Frisian has haun and laun, East Terschelling Frisian has hôn and lôn, Shire Frisian has hân and lân, Sagelterland Frisian has hound and lound, Heligolandic Frisian has hun and lun, and Northern Goesharde Frisian has houn and loun (monophthongised to hoon- and loon- in compounds). However, Molkwerum Frisian may have followed the same pattern as Hindeloopen Frisian in this case, because the former is geographically very close to the latter and may have undergone a similar phonological evolution based on linguistic cross-influence (which definitely seems to have occurred between the two). Of course, *hâând is the result one gets if one disregards Hindeloopen Frisian in this case; when one strictly follows evidence from all the other languages, that would not be wrong, but it is doubtful that Molkwerum Frisian would not have been like Hindeloopen Frisian in this case, because Molkwerum Frisian and Hindeloopen Frisian are showing similar phonological patterns in many ways as further detailed below.
  • Conclusion: based on the Molkwerum Frisian source materials, one has to conclude that there is remarkable phonological similarity between the Hindeloopen Frisian and Molkwerum Frisian. Therefore, in the case of uncertainty, I will generally follow the Hindeloopen Frisian example. Hindeloopen Frisian is my model for reconstructing Molkwerum Frisian, because it is the most well-attested language that is the most closely related to Molkwerum Frisian.
    • How “ik” and “is” were pronounced in Molkwerum Frisian remains an open question; were they pronounced “iek” and “ies” like in Hindeloopen Frisian, or were they pronounced “ik” and “is”? Molkwerum Frisian shows “ie” in many places just like Hindeloopen Frisian, but it may deviate from it with these two words. Sagelterland Frisian has iek and Wangerooge Frisian has îk. If Molkwerum Frisian has “iek”, it would be one of the Frisian languages that has this form, but it would certainly not be the only one of the Frisian languages in the Netherlands or Germany. Therefore, it is not unfathomable or unimaginable that it would exhibit this pronunciation for the words “ik” and “is”. I will do my best to resolve this matter and arrive at a satisfactory answer. The use of ie in “iek” and “ies” could be idiosyncratic to Hindeloopen Frisian, whereas the use of aaₑ in *aaₑte (grandfather) and *haaₑnd (hand) might easily be a shared property between Hindeloopen Frisian and Molkwerum Frisian. Namely, the doubt arises from the fact that the ie in “iek” and “ies” seems to have come from analogy with words where ie was pronounced, yet i was expected in other Frisian languages. As for *aaₑte (grandfather), this seems likely to have been pronounced that way; though one may raise more questions about *haaₑnd, but since aaₑ would exist as a rare sound only employed in certain exceptional cases in Molkwerum Frisian, it would certainly not be impossible for *haaₑnd to be realised in the same way as in Hindeloopen Frisian. However, ik and is are ubiquitous words; the question is whether they were spelled that way in Molkwerum Frisian under influence from Shire Frisian, which, according to the evidence, seems to have been the case with other words in Molkwerum Frisian, or were they genuinely pronounced with a lax rather than tense vowel? Frequent words in Molkwerum Frisian seem to have slavishly followed the Shire Frisian spelling of the time, and this makes it hard to discern what is being meant. Either way, it is conceivable that it would be “ies” and “iek” like Hindeloopen Frisian.
• The aauw has a different phonological origin from ou: aauw occurs only in words loaned from Dutch.
• The ö has a different phonological origin from ú: ö comes from an o that is modified by an r that has disappeared and ú comes from an
• The öi has a different phonological origin from ui: ui occurs only in Dutch loanwords with ui.
  • The vowel qualities of öi and ui may be the opposite (perhaps öi should be pronounced /øi/ while ui should be pronounced /ɜi/), this has to be investigated.
    • The Hindeloopen Frisian öi corresponds with ui in Dutch: flöiter is related to fluit.
      • The word meuije has been included in Heeroma’s wordlist (here initially transcribed as *möije) is related to Shire Frisian muoie, muoike; Hindeloopen Frisian möike, -möi (only still used in compounds); Dutch meu, moei. The correspondence with Dutch eu makes it tempting to believe that meuije might have been pronounced with an /øi/ and should therefore be spelled with ui. However, the eui clearly corresponds with Hindeloopen Frisian öi. The distinction between /øi/ and /ɜi/ is not easily noticed by the Dutch ear.
        • Luiwagen and fluite (as spelled in Heeroma’s wordlist) have been initially transcribed here as luiwaagen and fluite, where ui stood for /øi/, contrasted with öi which stood for /ɜi/, while noting that luiwagen and fluite were loaned from Dutch and may thereofre have sounded different from meuije. Nevertheless, I was not satisfied with this explanation and I began comparing the words with their Hindeloopen Frisian equivalents: the Hindeloopen Frisian words for luiwagen and fluite are löiwaigen and flöiter. This means that lui- and fluite should be transcribed as *löi- and *fluite. This means that ui /øi/ should be deleted for the time and all the words that were spelled with ui should be moved to öi; Heeroma’s spelling contrasting ui and eu may nevertheless mean that meuije was actually pronouced with /øi/, although that does seem unlikely. As for luiwagen, the second element of the compound should probably not be transcribed as *-waagen but as *-waigen.
• The eei may perhaps be pronounced ééi, or it may even be identical with ei.
  • The word is waaije in Hindeloopen Frisian. The equivalent of this Hindeloopen Frisian aa in Molkwerum Frisian is éé, therefore the pronunciation is wééije with /ɪːi/, not weeije with /eːi/. This diphthong is, however, very rare in Molkwerum Frisian. It seems, at first, somewhat strange for this reason, but Heeroma’s wordlist and the comparison with Hindeloopen Frisian surely suggest that the long diphthong ééi existed in Molkwerum Frisian.
• The quality and length of the dipthong òₑ/òòₑ is uncertain; the fact that bòn (cp. Hindeloopen Frisian bòn) may be short indicates perhaps that it is bòₑn (with short diphthong), not bòòₑn. Furthermore, Molkwerum Frisian (ò)òₑ corresponds with short diphthong èₐ in Hindeloopen Frisian; this may serve as etymological evidence that (ò)òₑ is indeed short and that therefore it should be written òₑ, not òòₑ. Nonetheless, the correspondence of òₑ with èₐ raises questions about whether the Molkwerum Frisian diphthong ends in ₑ /ə/ or ₐ /ɔ/. The fact that the diphthong is written as oä might suggest the latter, because ₐ /ɔ/ is the most a-like sound of the two choices. At the same time, the situation may be compared with óóₐ, which, although originally óóₐ, alternates with óóₑ. Based on this etymological reasoning, perhaps the original one may be òₐ, but it may have been allophonic with òₑ just as the distinction between óóₐ and óóₑ is not phonemic in Hindeloopen Frisian. It should additionally be noted that òₐ/òₑ does not exist in Hindeloopen Frisian because the r is generally not lost before the r in Hindeloopen Frisian except in the case of bòn, which does, however, not exhibit the òₐ/òₑ as the short form bòn, which also exists in Molwerum Frisian, came to replace the long or diphthongised form, which only still existed in Molkwerum Frisian.
  • To argue that the length of òₑ is short, however, goes against the evidence from the other Frisian languages: Shire Frisian bern (= bên) and East Terschelling Frisian bên. The West Terschelling Frisian hort, which is the only equivalent of the Molkwerum Frisian e > o sound shift outside Hindeloopen Frisian, is pronounced short: hòt. However, the evidence from the rhymed pairs in the poetical work of G. Knop seems to suggest that the o in West Terschelling Frisian is long (see my previous article). Based on Shire Frisian and East Terschelling Frisian, it seems plausible that it ought to be written òòₐ/òòₑ.
    • The word mòòₑlken does occur in Hindeloopen Frisian, and therefore the long diphthong òòₑ is extant in Hindeloopen Frisian, while the short diphthong òₑ is not. Therefore, it seems reasonable to believe that òòₑ would also have occurred in the Molkwerum Frisian word for child.
      • I have checked the entire dictionary of T. van der Kooy to see whether óóₑ, óóₐ, òòₑ, òòₐ, óₑ, óₐ, òₑ, òₐ are extant in Hindeloopen Frisian. The combinations óóₑ, óóₐ, òòₑ, òòₐ are extant: tòòₐst, nóóₐd, óóₐs, móóₐd, dòòₐst, bòòₐdskip, fòòₐlik, oeelebòòadoen, behooₐlik, skriebóóₐd, sòòₐtjen, sóóₐt. I did know that óóₐ exists in Hindeloopen Frisian because words like óóₐs (otherwise, differently), sóóₐt (kind, species, type) and nóóₐd (north) are quite frequent, but the fact that òòₐ exists actually had escaped my conscious attention; the examples with these sounds are actually quite few in Hindeloopen Frisian, therefore rendering the sounds quite rare if it were not for the frequent use of a few words containing these sounds. All examples with ₐ are the result of the loss of a historical r. Furthermore, òòₐ is particularly relevant for our research: the Hindeloopen Frisian tòòₐst is an important example, because it confirms that the equivalent Molkwerum Frisian word, which occurs in Heeroma’s wordlist with the spelling taôst, could have been pronounced with òòₐ as well. I was not surprised that I could not find examples óₑ and òₑ in Hindeloopen Frisian as far as I checked (it is possible that I did not check this thoroughly enough). However, I was somewhat surprised to learn that óₐ and òₐ existed in 19th-century Hindeloopen Frisian. Let us take the example with òₐ as this is crucial for our phonological research: bòₐttedwaaₑn which later became bòttedwaaₑn before becoming an obsolete word. This example demonstrates the evolution from òₐ to ò, and this is exactly what one may presume to have happened with the Hindeloopen Frisian word bòn, whicht ought to have been bòₐn at one point. Since both òₐ and òòₐ did occur in Hindeloopen Frisian, both most certainly could have occurred side by side in Molkwerum Frisian as well.
        • An etymological argument could be made as well for having both òₐ and òòₐ in Molkwerum Frisian: *tòòₐst comes from *tòòrst with a long stem-vowel while *bòₐn comes from *born with a short stem-vowel. So, both the òₐ and òòₐ would have occurred in Molkwerum Frisian because these two sounds have a different etymological origin; they are the result of a different vowel lengths. The etymological argument is definitely compelling, especially since Hindeloopen Frisian has demonstrated that these two sounds could technically exist side by side and this is no mere etymological fiction that is created by excessively following the linguistic reconstruction methods. I find the òòₐ interesting because it requires around 3 morae to pronounce it, this reminds me of the sound ô which is reconstructed for Proto-Germanic and is supposed to have been overlong (i.e., not 2 morae long but 3 morae long; a short sound is 1 mora). When I looked at the evidence from Molkwerum Frisian, I was skeptical at first that a vowel length difference could produce òòₐ and òₐ (while I had not been aware of such a sound pair in other Frisian languages) and that is why I initially grouped them together, but the evidence from Hindeloopen Frisian persuaded me to follow the evidence to its logical conclusion. Nonetheless, it should be noted that the pair òòₐ and òₐ is too rare for the sounds to be distinguished as 2 different phonemes; the two could be easily interchanged since the distinction is non-phonemic, but this interchanging would likely have been prevented by the fact that vowel length is generally phonemic in Molkwerum Frisian and Hindeloopen Frisian. Therefore, not enough examples did exist to specifically distinguish òòₐ and òₐ, there were enough other examples to distinguish vowel length in general.
          • An additional argument for òₐ is the fact that òₐ corresponds with èₐ in Hindeloopen Frisian (in the cases of gèₐs, hèₐt, kèₐs and hèₐd) or ò (in the case of bòn), while òòₐ does not correspond with èₐ but actually corresponds with òòₐ (in the case of tòòₐst). Therefore, the Molkwerum Frisian words should not be spelled *bòòₐn,**hòòₐt, *hòòₐdstéé, *kòòₐse, *gòòₐs but *bòₐn,**hòₐt, *hòₐdstéé, *kòₐse, *gòₐs.
            • Conclusion: it all boils down to the fact that òₐ and òòₐ have different etymological origins and therefore they exist as separate phonological entities.
              • The rare Hindeloopen Frisian with óóₑ are: póóₑt, sóóₑlje and kóóₑts (which should etymologically be have óóₐ). The last example shows that óóₐ and óóₑ might in practice be interchangeable. The distinction would certainly not be phonemic. This is an acceptable reason not to include óóₑ and to allow for the possiblity that óóₐ might have been pronounced as óóₑ. I may include óóₑ as a separate sound at a later date, however.
• It should not be *tiₐm (intestine) but *tirm, because Hindeloopen Frisian preserves the r before the m in this case as well: tirm (intestine). Therefore, *tirm is not an example of iₐ.
  • The iₐ is very rare in Hindeloopen Frisian and only occurs in a few words, such as: jiₐld (money), fiₐld (field), kiₐld (cold weather), tjiₐn (against), tiₐcht (close; closed). However, the iiₐ is much more common than the iₐ in Hindeloopen Frisian. The iₐ seems to have changed to ò in Molkwerum Frisian: *jòld for *jiₐld. However, I do not find it very plausible that Molkwerum Frisian would have simplified this diphthong to a single vowel: it might also be the case that the Molkwerum Frisian texts misrepresent this sound since it is known in other Frisian languages. One could easily mishear iₐ as ò (after all, it may sound like this to the non-native ear), or uncertainty about this sound could lead one to spell it as o. Either way, it means that Molkwerum Frisian had the word jiₐld at some stage as well.
    • The fact that Heeroma’s wordlist spells miïlt suggests that *miₐlt existed in Molkwerum Frisian (see my previous article). This is a clear example demonstrating the existence of iₐ. Furthermore, the word miₐlt exists in Hindeloopen Frisian as well. This means that *jiₐld, *fiₐld, etc. probably existed in Molkwerum Frisian as well; it is unlikely that miₐlt would have been an isolated case.
• All the Molkwerum Frisian diphthongs are falling diphthongs with the sole exception of uo, which is a rising diphthong.
• Maitsje and lietj have been attested, but using a k instead of t (e.g. maikje) is more conservative in this phonetic environment. The Molkwerum Frisian intervocal tj/tsj is derived from kj: *lietje < *liekje (cp. Hindeloopen Frisian likje), *maitsje < *maikje (cp. Hindeloopen Frisian meikje).
  • Curiously, laakje (to laugh) instead of laitsje has been attested even in the late material that otherwise shows the new innovation -t(s)j- instead of the conservative -kj-. Laakje is definitely native to Molkwerum Frisian, whereas maitsje and lietj are not.
• Hindeloopen Frisian has aig instead of éég. Could it be possible that Molkwerum Frisian once had *aig < *aag as well? Éég seems a Frisian loanword.
• The clusters /ts/ and /tsj/ only occur in Shire Frisian loanwords which had come to replace the native vocabulary in late Molkwerum Frisian: *tsiere and *tsjiis for *tjiere and *sees.
• Unvoiced plosives are unaspirated in Molkwerum Frisian.
• Plosives can be palatalised or labialised in Molkwerum Frisian. For instance, *swelle and *waskje are pronounced /sʷɛ.lə/ and /vas.kʲə/. The labialised consonants in Molkwerum Frisian were probably under-rounded (i.e., less rounded).
• The intervocal g may be pronounced as a fricative or plosive. The most conservative pronunciation would be with a fricative, but there is a tendency in Hindeloopen Frisian towards pronouncing it as a plosive according the logic that an intervocal g starts a syllable.
• The attested form is *klaawer, but the more conservative form would be *klaiwer. The ai and òi occur in Molkwerum Frisian just like in West Frisian Hollandic and Hindeloopen Frisian.