Visual helpers:

• DBFiddle: found here
  
• Example plpgsql function in a DBFiddle: found here
  
• Dots GraphViz: found here
  
• e.g. Lua sandbox example: found here
  

Struggling to think of how to do this efficiently but my goal is to try to recursively reduce a set of aggregated child_ids to one or more parent_ids, alongside any remainder child_ids.
In the example I attached:

• the set has been reduced from [6, 7, 8] to [5], but the goal would be to reduce it further;
  
• in this case, the result should really be [3] since [5]'s ancestor of [3] is satisfied (and so forth);
  
• e.g. for the case of [6, 7, 9, 12] where [12] is an island/root node, the result would be [3, 12] since [9] would be reduced to [8] and the resultant [6, 7, 8] is contained by [5] etc
  

Any thoughts would be appreciated

​
https://preview.redd.it/m5hd8gtyczma1.png?width=962&format=png&auto=webp&s=9a00b0c471ee2990fbc01b40dd82afb679c05a41
I mean why do we even need letter K! We already have C! Now a lot of people think, C is the culprit/offender who stole from K but false! It's the reverse. C was always hard in Latin including Old English and all was going well for C until K was included for some incomprehensible reason and copied C. We would be better of if only C was used for hard sound, K was non-existent and Q was used for Ch like in Chinese. The fact that Thorn doesn't exist now but K does is a travesty. K is also an ugly letter and a rude text message. K does C but worse because, K is even silent in the kn and the ck digraph for no reason and its quite rare. I propose to remove K, use C for hard sound and Q for Ch. Getting rid of K would also get rid of that obnoxious KKK hate group.
​
Please join this subreddit:
https://www.reddit.com/LetterKsucks/

þere's already Þ replacing the TH digraph. but wait, þere's also a lot of digraphs in english, like sh, ch, ph. what's about þem?
​
1) CH
if we'll ignore any digraphs rules, it will sound like a K. so, þis digraph makes no sense and we should replace it wiþ a letter (i think Ч) (only in cases when it actually sounds like a russian Ч (like in church, chair)).
​
2) GH
sounds like a G at þe beginning of þe words and F at þe end. why? idk lol. this digraph sucks so let's just replace it wiþ a letter. but which one? make it a G at þe beginning and an F at the end?
​
3) PH
well, i þink þis digraph's fine. if you'll try to hold the P sound for a while, you'll get an F. so let's keep it.
​
4) SH
sounds like a russian Ш. wiþ CH, þis digraph also makes no sense and i want to replace it wiþ þis russ.. wait, you use greek letter letter sigma to describe þis sound? got it.
​
5) TH
Þ
​
6) WH
it's just a W. let's remove the H.
​
7) KN
silent K? þat's dumb. now it's actually KN wiþ pronounceable K.
​
8) NG
who made þe Ŋ for þis?
​
9) QU
even Q is a useless letter cuz it actually sounds like a K (if we'll look at þe words with Q not followed by U (as example, two asian countries iraq and qatar)).
​
u know the silent E? i hate it. how about replacing it with anoþer letter, þat will be always silent unlike oþer letters. this is now your чoice, pick up any letter you want þat will replace the silent E. here's my suggestion: Ә

(I'm not a professional or anything (so I'm definitely gonna say something weird or doesn't really make sense), I'm going by personal experience, so take all this with a grain of salt.)
Glass Animals has quickly become one of my top three favourite music groups of all time, honestly, and there's a lot of reasons why I would say that, but I'm gonna be real for a moment and try to talk about this in a way that I can understand it. The psychological impact of GA is one of the most intriguing experiences I've ever had, and I've stated in other posts that they help me with anxiety and calming down.
Now that I've said that, I wanted to look deeper into the reason/reasons why. (Again, I'm not a professional, so take it all with a grain of salt).
The Use of Tempo
I'm gonna start this off by looking at Glass Animals' musical design. When it comes to listening to music, the human body responds to it in different ways depending on who you talk to. The same can be said for relaxation. You can relax however the hell you want, whether it's by jamming out to heavy metal or by listening to something classical, however, in my own way, I experience relaxation when listening to slow music. The tempo in which the music is played has an impact on the human body, and GA is no exception. A lot of GA's music is set at slower speeds, and slow music often results in the heart rate slowing down and giving a relaxing feeling.Songs like Your Love, Gooey, Black Mambo, etc. are all set at slower speeds and give the idea of relaxation.
(Theoretical) Emphasis
Now this one is purely theoretical as I don't know if it was intentional or not by GA themselves, but one of the most fascinating things about their music is the use of the autonomous sensory meridian response, or ASMR, and the emphasis on vowels, consonants, and digraphs. What do I mean by that? Well, vowels such as A, E, I, O, U are used a LOT in GA's songs, and Bayley's emphasis on consonants such as "t" and "s" and digraphs such as "ah", "ck", and "sh". In ASMR, the latter especially is used very often to elicit said response (I've always phrased it as the words "digging into the ear". )If you listen to Take A Slice, here are some of the lyrics with more emphasis on each thing in bold.
I'm filthy and I lOve i...T, STudebaker all gold, got a SHotgun in my poCKE...T [...] Suckin' on a slim Vogue, dark fingernail poliSH
What the ellipsis on the T's mean is the delay before the sound occurs.
The bolded letters and digraphs are emphasised in the song and it might just be my ears, but it feels as though the sounds were 'digging' into the listener's head. Another thing about that song in particular is the vocal distortion to give the auditory appearance of growling, which might also feed into that thing.
Tone
One of the most fascinating things I've heard about GA is their use of tone, in particular Dave Bayley's vocal range. I'll say it right now that the tone he uses for singing is fantastic, but it also plays a part in the previous two paragraphs. How does that work? Well, let's take a look at Gooey and Hazey.
Gooey's music has reverb, and this is given right from the start of the song. If you pick apart a song and add reverb to it, it gives the feeling that it's drawn out into a cave or some sort of chamber. This feeds into the first paragraph where, because it sounds like it's in a cave, the tempo has slowed down (because you have to give it time for the sound to fade). As for the emphasis, let's take a look at the lyrics, and I'm gonna do the same as I did with Take A Slice by bolding the letters and digraphs. One thing that makes Gooey unique in that the first letter sounds like an R is being heard (at least to me).
(r) I come clo...SE, let me SHow you everything I knOW, the jungle SLang, SPinning around my head and I STare.
The alliteration of "s" words helps with identifying the tone that Bayley uses, as it sounds almost like a whistle or a whisper (or both). Usually with these sorts of tones, the speaker is lulling the listener to sleep (in combination with the slow tempo).
Hazey is a little different, as it's slightly more fast-paced than usual and features loud instruments such as brass horns, a badass bass, and finger snaps. There's also what appears to be samples of cat purring in the song. Bayley's vocal range shines here with him going into much higher tones, yet the near-whispering vocals still remain. Again, we're going to do the same as we did with the others and analyse the lyrics.
You know you're So jUICED, you said you kicked the bOOZe [...]
The length in which the "ooo" sounds are drawn out especially help with the tone, as well as the whistle/whisper with the "C" in "juiced" and the "Z" in "booze".
You guys get the point. Now I know that not all of their songs are all slow and relaxing; in my opinion Space Ghost Coast to Coast sounds more grunge-like and "rough", with the use of growling undertones at the chorus; to me it sounds like an auditory snarl, while Take A Slice is loud and boisterous with a lot more "growling" and the use of brass and whatever the hell that last instrument is during the instrumental (wanting to say a distorted guitar, but I have no idea).
Glass Animals' music has an intriguing and fantastic impact on my head, one that's not only one to hear (hehe, pun) from a listener's perspective, but also looking deeper into it makes me think and want to analyse more of their music; it's just that fucking good. Dear lord, I didn't expect myself to type this whole essay, but I did it anyway. if there's anything that I missed, what you guys wanna add to that, your own experiences, if you think I'm spouting nonsense or bullshit, that's fine too. (By the way, if there are any professionals out there that can better word/more accurately phrase what I was trying to describe, let me know and huge thanks.)

For ad werþe zen nýe Mannen, bez mann hæbbe allhjarted.
[ɸɔɾ ɑ ˈɰɛrːs̪ə ʃɲ̩ ˈnœʏ̯ːjə ˈmɑnːn̩ bəʃ ˈmɑnː ˈʃæbːə ˌɑlːˈʃɑrːtə]
for to become-inf the.m.sg new-def.m.sg Manne-the.m.sg be.fut.sg man.sg have-inf all-heart-def.n.sg
"To become God, you have to walk in everyone's shoes."
- Erish proverb

Erish (ærsk), an a posteriori West Germanic artlang, isn't the first constructed language I've worked on, but it is the first one I can say has come to a point where it is presentable. The concept is that, in the conworld, Erish arises from Proto-West Germanic nearby North Germanic languages as they arise from Proto-Norse, and is still in a sort of sprachbund with them. Intelligibility, particularly in speech, is hampered by Erish's own innovations, especially phonologically.
Here, I would like to provide a summary of the closest thing to a standard Erish pronunciation, as well as an account of the orthography, as its depth tells a bit about the changes that Erish has undergone. With each, I'll give a snippet about the goals I had going into them, as well as feedback questions I myself have - Erish is and will always be a work-in-progress. I am greatly indebted to a variety of resources, so I will provide several of them at the end of this post and the others that may follow it, as well as a concluding gloss.

Phonology

Most Erish speakers simply use their own dialects when speaking, up to and including the King or Queen. The pronunciation taught to foreigners, as well as the one used in national broadcasting, is that of Hamnstead, which was the city where radio broadcasting first developed in Erishland, and which is still a center of national media. The Hamnstead dialect is a Western dialect close enough to Southern dialects that its phonology is sort of a mixture of the two groups, plus its own quirks.
Goals
Personally, this phonology is my attempt at creating one reminiscent of the older stages of Germanic languages, but which feels plausibly modern and plausible in a place where North Germanic contact and influence continues into present. A bit of a summary and highlights of what that means:

• The vowels, especially as phonemes, are not too dissimilar from contemporary and historic relatives, as Germanic languages were and are known for their many vowels.
  • Hamnstead Erish doesn't have the /ɵ,ʉː/ of Southern Erish dialects, but the realization of /eː,øː,oː/ is similar to the Norwegian and Old Norse diphthongs. They even sort of correspond, but with the asterisk that they also correspond to Norwegian /iː,yː,ʉː/ and /eː,øː,oː/.
• The consonants may seem more akin to Spanish than Swedish, though in my view, it's a blend of the latter and Gothic. I do give props to the interpretation of Spanish /ɾ/ being ungeminated / for Erish /'s allophony, though.
  • Word-initial /ɕ,j,ɧ/ in Swedish corresponds to /t͡ʃ,ʝ,ʃ/ in Erish; /ʝ/ is similar phonetically to Old English /j/. However, one word with Erish initial /ʝ/ also corresponds to Swedish /h/; initial /ʃ/ also corresponds to many Swedish /h/'s, and even a few /d/'s.
• Many of the apparent archaisms are actually re-innovations. Why cling to an old way of pronunciation when a change closer to present day can plausibly re-introduce something similar?
  • Case in point: Is the [β] allophone of /b/ you lenited decades ago hard to distinguish from the /v/ you and your neighboring languages have had for centuries? Just merge /v/ with /b/!
• The only notable phonological archaisms of Hamnstead Erish, to my knowledge, are that there is still a short /æ/ from i-umlauted /a/ (something uncommon even among Erish dialects), and that Proto-Germanic *h is still pronounced as /x/ where it hasn't merged with other phonemes.
  • There's /ɣ/ as well, but Dutch and Low German also preserve it. It's also a bit misleading, since /ɣ/ is actually /ɰ/. The /ɣ/ transcription is used for consistency with what otherwise varies between /ɣ/, /ɰ/, and /w/ between dialects.

Vowels
Hamnstead Erish has a rather bland vowel inventory for an Erish dialect. About the only notable feature, phonemically speaking, is that there is still a short /æ/ distinct from /ɛ/, though that's typical of Western dialects. Phonetically, though, the story's a bit more complicated - Hamnstead Erish is amongst the few dialects that can be argued to, in some limited way, preserve most of the original Old Erish diphthongs, and has re-innovated a very limited form of allophonic u-umlaut.
​

	Front unrounded	Front rounded	Back
Close	ɪ • iː	ʏ • yː	ʊ • uː
Mid	ɛ • eː	œ • øː	ɔ • oː
Open	æ • æː		ɑ • ɑː

• The short vowels are phonetic monophthongs
  • The close vowels are near-close [ɪ,ʏ,ʊ]
  • The mid-front vowels, especially /ɛ/, are mid-front [ɛ̝,œ̝]; /œ/ may also be open-mid [œ̝]
    • /ɛ/ in unstressed syllables is generally [ə], though broadcasters tend towards using an [ɛ̠]
  • The mid-back vowel is either open-mid [ɔ] or, less often, mid [ɔ̝]
    • In unstressed syllables, it may be realized as a retracted, raised [ɞ̟˔], but this is far less common than the [ə] realization of /ɛ/. This may have to do with unstressed /ɔ/ always being morphologically associated with some marked feature, namely the feminine gender, neuter plural, and plural subject of the past tense.
  • The open front vowel may be near-open [æ] or open [a]
  • The open back vowel in regular syllables may vary between completely unrounded open back [ɑ], or a very weakly rounded [ɑ̜]
    • /ɑ/ is fully rounded to [ɒ] if a following syllable contains /ɔ,ʊ/, or the allophone [ɒ]
• The long vowels /iː,uː/ are phonetic monophthongs [iː,uː]
• /ɑː/ is phonetically a monophthong, but may be raised [ɑ̝ː], and follows the allophonic rounding pattern of its short counterpart
• All other long vowels are realized as diphthongs
  • The mid-vowels /eː,øː,oː/ are realized as closing diphthongs [ɛɪ̯ː,œʏ̯ː,ɔʊ̯ː], or [eɪ̯ː,øʏ̯ː,oʊ̯ː]
  • /yː,æː/ are realized as backing diphthongs [yʉ̯ː,æɐ̯ː]

Consonants
Hamnstead Erish, like most Erish dialects, has a consonant inventory that is phonemically similar to the Nordic languages, but the allophony of these consonants is less so. Voiced stops regularly lenit to approximants that devoice and fricate word-finally; this leads to the notorious "Erish hiss". Notable aspects of Hamnstead's phonology are the merger of /ʂ/ and /ɕ/ into /ʃ/, a change which is common but still absent in Southern dialects, and that /ɣ/ is a velar approximant, instead of the labiovelar common to Western dialects.
​

	Vclss. labial	Vcd. labial	Vclss. coronal	Vcd. coronal	Vclss. palatal	Vcd. palatal	Vclss. velar	Vcd. velar
Nasal		m		n		ɲ		ŋ
Stops	p	b	t	d	t͡ʃ	(d͡ʒː)	k	(gː)
Cntnts.	f	(β)	s	(ð)	ʃ	ʝ	x	ɣ
Laterals				l		ʎ
Trill				r

• The nasals are generally realized as [m,n,ɲ̟,ŋ], with the palatal and nasal being inherently geminate. Before a coronal obstruent, they are typically realized as [ɱ,n̪,ɲ̟,ŋ].
  • After a syllable initial voiceless obstruent, /n,ɲ/ are voiceless [n̥,ɲ̟̊].
  • Unstressed, word-final /ɛm,ɛn/ are commonly realized as syllabic [m̩,n̩].
    • Unstressed sequences of /ʃɛn/ and sometimes /t͡ʃɛn/ may be realized as [ʃɲ̩,t͡ʃɲ̩]
• Similar to many other Germanic languages, the voiceless stops /p,t,t͡ʃ,k/ are realized as aspirated [pʰ,t̪ʰ,t͡ʃʰ,kʰ] in stressed onsets if they are not preceded by a sibilant.
• The voiced stops /b,d/ are plosive [b,d̪] if they are morpheme initial, geminated, or post-nasal. Elsewhere, they lenit to the approximants [β̞, ð̞], which devoice and spirantize word-finally to /f,s/.
  • The process of /d/ to /s/, in tandem with /ʝ/ to /ʃ/, is a well-known feature of Erish phonology. Uneducated Erish speakers using other languages may apply it, making the "Erish hiss".
• The voiced continuants /ʝ,ɣ/ are approximants [j,ɰ]; when geminated, they harden to stops [d͡ʒː,gː]. Similar to the voiced stops, the approximants devoice and spirantize to /ʃ,x/ word-finally.
• In emphatic speech, or speech mimicking Eastern dialects, the phonemic and allophonic voiced continuants may be pronounced as [β,ð,ʝ,ɣ], though never with the same frication as the voiceless continuants.
• The voiceless continuants /f,s,ʃ,x/ are consistently realized with strong constriction as [ɸ,s̪,ʃ,x].
  • Some dialects have free variation in the realization of /f/ from [ɸ] to [f], and a few consistently realize it as the labiodental /f/.
  • After a voiced stop, /f/ is realized as [v], which usually has less frication than the voiceless allophone, but is not an approximant like the phonemic voiced continuants.
• The coronal lateral /l/ is normally realized as [l], but when in contact with /t,d,s/, it is laminal denti-alveolar [l̪].
  • After a voiceless consonant, it is voiceless [l̥~l̪̊].
• The palatal lateral /ʎ/ is a traditional phoneme of Erish, but younger generations outside of the South have begun to merge it with /ʝ/. It is realized as an alveolo-palatal [ʎ̟].
  • After a voiceless consonant, it is voiceless [ʎ̥˖].
• The trill / is realized as a full trill morpheme-initially or when geminated; in other contexts, it is an alveolar tap [ɾ].
  • After a voiceless consonant, it is voiceless [ɾ̥].
  • Unstressed, word-final /ɛ is commonly realized as [ɐ].
• As described throughout, the coronal obstruents /t,d,s/ are dental [t̪,d̪,s̪]. Other dialects may use an alveolar realization [t,d,s].
  • Before /k/, /s/ is usually alveolar [s].
• The palatal and velar obstruents /t͡ʃ,ʃ,ʝ,k,ɣ,x/ are all allophonically rounded to [Ḱʷ,Kʷ] before rounded vowels.

Phonotactics
Valid onset consonants:

• Zero consonant
• All consonant phonemes other than /ɲ/ and /ŋ/
• Obstruent + nasal: sm, sn, ʃɲ, kn, xn
• Obstruent + continuant: tf, df, sf
• Obstruent + lateral: pl, bl, fl, sl, ʃʎ, kl, ɣl, xl
• Obstruent + trill: pr, br, fr, tr, dr, sr, kr, ɣr, xr
• Continuant + stop: sp, st, sk
• Continuant + stop + lateral: spl
• Continuant + stop + trill: spr, str, skr

Valid coda consonants:

• Zero consonant
• Consonant: m, n, p, t, t͡ʃ, k, f, s, ʃ, x, l, r
• All consonant phonemes as geminates
• Nasal + nasal: mn, ɲn, ŋn
• Nasal + obstruent: mp, nt, nd, ns, ɲt͡ʃ, ɲʃ, ŋk
• Obstruent + obstruent: sp, pt, ft, st, ʃt, kt, xt, ps, fs, ts, ks, xs, sk
• Lateral + nasal: lm
• Lateral + obstruent: lp, lt, ld, ʎt͡ʃ, lk, lf, ls, ʎʃ, lx
• Trill + nasal: rm, rn
• Trill + obstruent: rp, rf, rt, rd, rs, rt͡ʃ, rʃ, rk, rx
• Trill + lateral: rl

Valid medial consonants:

• All consonant onsets and codas above, excluding zero consonants
  • If hiatus occurs, /j/ is inserted if the first vowel is front; if back, /ɣ/ is inserted.
• Nasal + obstruent: mb
• Lateral + obstruent: lb, lɣ
• Trill + nasal: rŋ
• Trill + obstruent: rb, rʝ, rɣ

All stressed syllables are inherently heavy - if there is no long vowel, the first consonant to follow is geminated. Neither long vowels nor geminates may occur in unstressed syllables.
Prosody
Like most other Germanic languages, the most common syllable to be stressed is the first (and often only) of a given word. Loanwords can follow different patterns, but a rule of thumb is that the syllable before the last consonant of a root is the one to be stressed.
Feedback Questions
Questions I personally have are:

• For those familiar with the phonologies of older Germanic languages as well as Proto-Germanic, does the consonant system seem like a good "modern version" of that type of phonology?
• Is it a good idea to leave all long vowels other than /ɑː,iː,uː/ as phonetic diphthongs? Would it take a short time before the diphthong allophones become phonemes of their own, or would they remain stable?
• I'm fairly comfortable with /ʝ/ devoicing and spirantizing to /ʃ/ word-finally because there are direct attestations of such final-obstruent devoicing, but is it plausible for /d/ to similarly go to /s/?
• Does anybody have a resource that consistently gives phonemic transcriptions of Swedish, Norwegian, and Danish words, especially regarding pitch-accent/stod?

Orthography & History

Erish is written with the standard 26 letters of the Latin alphabet and the additional letters þ, æ, and œ. Unfortunately, pronunciation cannot be succinctly described for Hamnstead Erish because the orthography is fairly etymological, essentially reflecting the pronunciation of Old Erish with a few simplifications and updates. It is, though, fairly regular, and the hope of Jugar Raskson, the father of modern Erish orthography, was that the written language would let as many people as possible derive their own pronunciation from the spelling.
Goals
Erish orthography is essentially fighting two battles: the battle to modernize spelling and keep the rules as regular as possible, and the battle to keep important cultural and religious texts from the Old Erish period as intelligible as possible. I've looked at Icelandic and Faroese for inspiration, but also incorporated principles from languages with less orthographic depth.

• As will be seen, Erish orthography is more than willing to keep native spellings a millennium old. It is less concerned about loanwords, which may be mangled to keep regularity.
• With no small amount of resistance from traditionalists, many old graphemes have either been dropped or restricted in usage. Joches "horses", for example, used to be spelled johves.
• The use of the letter k for /k,t͡ʃ/ is actually fairly recent; the traditional letter in Erish was c, and this is still seen in the basic long forms of k, g, and h being ck, cg, and ch.
  • Similarly, the use of w is a modernism; v was used for both itself and w, and this is still seen in the spellings of Cv clusters.
• Erish tries as much as it can to avoid diacritics and special characters, so it makes use of every letter of the Latin alphabet as a regular part of its orthography.
  • The accented vowels are there because digraphed vowels would suggest vowels in hiatus or perhaps even long vowels.
  • The letters æ and œ are used along with þ more or less out of in-universe tradition.

Vowels
The following table presents the pronunciation of vowel graphemes in Erish. Note that the "jV" digraphs are only pronounced in this way if the j is word-initial or can soften a preceding consonant:
​

Letter(s)	a	e, í, eì, aì	i	o, á, ú, aù, où	u, ó	y	æ, já	œ, jó, jú, ý, oì, eù
Short	/ɑ/	/ɛ/	/ɪ/	/ɔ/	/ʊ/	/ʏ/	/æ/	/œ/
Long	/ɑː/	/eː/	/iː/	/oː/	/uː/	/yː/	/æː/	/øː/

• The long vowels are only used if the vowel is stressed and followed by no more than a single consonant; when there is a long vowel before multiple consonants, an apostrophe is inserted after the first consonant, as in gód't "good (n.)" /ˈɣuːst/. In all other cases, the short vowels are used.
  • Several common grammatical words are pronounced in common speech with a short vowel, such as jis "ye" [jɪs], mostly because they are unstressed.
• The grave accented letters ì and ù are used to represent the second element of potential diphthongs in Erish dialects that arose from loanwords. In Hamnstead Erish, they are simply (phonemically) monophthongs.
• Because it is impossible for vowels to stand in hiatus, any sequence where two vowels stand in hiatus, such as fríen "free" /ˈfreːʝɛn/, is treated as though a g stood between the two vowels.
• The accented letters correspond to Old Erish /aː,iː,oː,uː,yː/, which shifted during the Middle Erish period. In the Hamnstead dialect, /aː,oː/ shifted to /ɔ~oː,ʊ~uː/.
  • /iː,uː,yː/ merged with the Old Erish diphthongs /ɛi̯,ɔu̯,œy̯/, and then later monophthongized to /ɛ~eː,ɔ~oː,œ~øː/
  • The demonstrative pronouns zó "this, that (f.)", zád "this, that (n.)", zós "these, those (f.)", zóm "these, those (dat.)" are written with acute accents, but their sound value is that of the unaccented letter. This is done to distinguish them from the definite article.
• The Old Erish glides /jaː,joː,juː/ became /æː,øː,yː/, so long as the /j/ didn't palatalize the preceding consonant and wasn't word-initial; this /yː/ would go through the process described above,

Consonants
To keep Erish consonants simple, and as allophony has already been covered, their transcription here is phonemic. In educational materials for Erish, consonants are generally divided into four groups, plain, strong, weak, and the letter g, which are grouped based on whether and how they can soften.
Basic rules about consonants include:

• Consonants are long if they are not word-initial, and they are doubled, their grapheme is composed of more than one letter, or they precede another consonant without an intervening apostrophe.
  • The one exception is qu, which is treated like a short consonant for both consonant and vowel length.
• Several consonants can also become fricatives word-finally if they are short.
• Before another d, s, t, frication may also occur, with the resulting consonant, in line with pronunciation rules, being long or short.

Plain consonants
Plain consonants are so-called because they cannot soften under any circumstances.
​

Letter(s)	Hard	Fricative
b	/b/ binde "to bind" /ˈbɪnːdɛ/	/f/ lab "lab(oratory)" /ˈlɑːf/
f	/f/ faþer "father" /ˈfɑːsɛ
m	/m/ móte "must" /ˈmuːtɛ/
ng	/ŋ/ wing "wing" /ˈɣɪŋː/
p	/p/ pá "on" /ˈpoː/
r	/ rotte "rat" /ˈrɔtːɛ/
v	/b/ virus "virus" /ˈbiːrʊs/	/f/ livte "lived" /ˈlɪfːtɛ/

• V may also represent /f/ after a consonant, such as in tves "two" /ˈtfeːs/, as least in Hamnstead Erish. Erish dialects generally vary between this v representing /f/, /w/ (the original Old Erish value), or /x/.
  • V otherwise reflects non-initial instances of Old Erish /f/, which was realized as [v], and later became its own phoneme. When /b/ began to be lenited to [β] around the 1700s, this facilitated the merger of /v/ with /b/ (as well as the shift of /f/ to [ɸ]).

Strong consonants
Strong consonants are "strong" enough it takes a j or z to soften them. Because strong consonants are long when softened and non-initial, they lack soft fricative forms. They include the coronal consonants which got palatalized to retroflex consonants, though z was retroflex for most of Erish history. The retroflex spellings tend not to occur in modern loanwords unless Henskland adopts a retroflex form; it's the only Land left that hasn't merged the retroflex consonants with the palatals.
​

Letter(s)	Hard	Soft	Fricative
d	/d/ dœr "door" /ˈdøː	/ʝ/ bedje "to ask" /ˈbɛʝːɛ/	/s/ sæd "seed" /ˈsæːs/
l	/l/ láte "to let" /ˈloːtɛ/	/ʎ/ ljúht "light" /ˈʎɔxːt/
n	/n/ naht "night" /ˈnɑxːt/	/ɲ/ hænje "to hang" /ˈʃæɲːɛ/
s	/s/ synge "to sing" /ˈsʏŋːɛ/	/ʃ/ sjelv "self" /ˈʃɛlːf/
st	/st/ stóren "big" /ˈstuːrɛn/	/ʃ/ stjarne "star" /ˈʃɑrːnɛ/
t	/t/ sten "stone" /ˈsteːn/	/t͡ʃ/ sitje "to sit" /ˈsɪt͡ʃːɛ/
z	/ʃ/ meze "more" /ˈmeːʃɛ/
þ	/s/ þing "thing" /ˈsɪŋː/	/ʃ/ þjúv "thief" /ˈʃoːf/

• In the definite article -ed as well as core grammatical words like ged "it" /ˈʝeː/, zad "the; that (n.)" /ˈʃɑː/, and gvad "what" /ˈɣɑː/, d is silent.
• D does not fricate after l or r, as in hald "hold (imp.)" /ˈxɑlːd/ and gard "farm" /ˈɣɑrːs/, but does after t, as in wlet'de "searched" /ˈɣleːstɛ/
• The use of j or z for softening strong consonants is largely predictable.
  • J is only used at the beginning of words, whilst z is normally used word-medially.
  • Word-medial palatalizing j is an indication that the verb - and it is always a verb - has an irregular conjugation that involves hardening (depalatalization).
    • J, however, is always used after l or n; lz and nz indicate /ʎʃ,ɲʃ/
• Lj and nj can only soften initially if they are word-initial, or part of the onsets hlj, hnj, slj, or snj.
• Before k, n is pronounced /ŋ/, as in tank "tank (container)" /ˈtaŋːk/
  • Before c, sk, tz, or z, it is pronounced as /ɲ/ as in lunc "lunch" /ˈlʊɲːt͡ʃ/
• S softens before lj and nj to /ʃ/, such as in snjó "snow" /ˈʃɲuː/
• T fricates before d, tt, or apostrophized t, as in hlot'de "allotted" /ˈxloːstɛ/
• Z reflects Old Erish /ʂ/, which was the main reflex of Proto-Germanic *z in Old Erish (some earlier instances did get devoiced to /s/ instead).
  • /z/ never occurred word-initially in Proto-Germanic, but it did in Old Erish, such as in zat "the; that (n.)" (whence modern Erish zad). This developed from some instances where a weakly stressed /θ/ got voiced to [ð], which soon assibilated to /z/, which later devoiced to /ʂ/.
• Þ reflects Old Erish /s/, the main reflex of Proto-Germanic *þ in Old Erish. Around the late 900s, the original dental fricative /θ/ assibilated to /s/, but remained contrastive with /s̺/, the reflex of Proto-Germanic *s. Although this contrast was lost by the 1400s, it is still reflected in Erish orthography.

Weak consonants
Weak consonants are "weak" enough that vowels can soften them in addition to j; in instances where a soft pronunciation is used, it is either because a hard letter is used, or because there is a v "shielding" them. They include the reflexes of the Old Erish velar consonants /x,k,sk/, as well as the labiovelars /xʷ,kʷ,skʷ/. The basic principle governing their softening is that they do so before certain vowels initially, and after other vowels elsewhere. If neither of these conditions are met, they use a hard pronunciation. It is to be noted that, barring one exception, acute accented letters have the same effects as their unaccented counterparts, and so they are not treated seperately here (grave accented vowels never soften a consonant).
​

Letter(s)	Conditions or example	Softens initially	Softens elsewhere	Hard
h (short), ch (long)	Conditions	to /ʃ/ before e, i, y, æ, œ	to /ʃ/ after e, i, y, æ, œ	/x/ in all other conditions
	Examples	himmel "sky" /ˈʃɪmːɛl/	reht "justice" /ˈrɛʃːt/	hús "house" /ˈxoːs/
k (short), ck (long)	Conditions	to /t͡ʃ/ before e, i, y, æ, œ	to /t͡ʃ/ after i, y	/k/ in all other conditions
	Examples	kyng "king" /ˈt͡ʃʏŋː/	rík "realm" /ˈreːt͡ʃ/	bók "book" /ˈbuːk/
sk (short and long)	Conditions	to /ʃ/ before e, i, y, æ, œ	to /ʃ/ after a (not á), e, i, y, æ, œ, or a consonant	/sk/ in all other conditions
	Examples	skíne "to shine" /ˈʃeːnɛ/	fisk "fish" /ˈfɪʃː/	busk "bush" /ˈbʊsːk/

• The use of h to represent /ʃ/ is more or less confined to native Erish vocabulary. Sk is far more common in loanwords.
• Similarly to s, h may also soften before lj and nj, as in hnjóse "to sneeze" /ˈʃɲuːse/
• Soft non-initial k is confined to native Erish vocabulary, as Eastern dialects failed to palatalize it in those conditions. Instead, c and zk are used in loanwords.
• The behavior of sk palatalizing after a but not á is because sk palatalized non-initially unless a back vowel preceded (or, technically, followed). Though in Old Erish they were phonemically /a,aː/, /a/ was not a back vowel (most likely being [ä]), but /aː/ was, being either [ɒː] or [ɔː].

Erish used to be far more inconsistent in representing when exceptions to these conditions occurred, but modern spelling is extremely regular in this regard. Soft consonants in instances where a hard consonant is expected are represented in different ways depending upon if they are initial, non-initial and short, or non-initial and long.
​

Desired soft consonant	Location and length	Grapheme(s)	Example
h /ʃ/	Initial	hj	hjarte "heart" /ˈʃɑrːtɛ/
	Non-initial and short	N/A
	Non-initial and long	hj (singular instance)	hlæhje "to laugh" /ˈxlæʃːɛ/
k /t͡ʃ/	Initial	kj	kjúe "to chew" /ˈt͡ʃoːɣɛ/
	Non-initial and short	c	koc "coach" /ˈkoːt͡ʃ/
	Non-initial and long	zk, rarely kj	þækje "to think" /ˈsæt͡ʃːɛ/
sk /ʃ/	Initial	skj	skjá "cloud" /ˈʃoː/
	Non-initial and short	N/A
	Non-initial and long	sc	broscyre "brochure" /ˈbrɔʃːʏ

• The use of non-initial hj, kj is, similar to the strong consonants, a signal that the verb's conjugation is irregular and involves hardening.
• Hj is an etymological grapheme that mostly occurs in native Erish vocabulary. Hlæhje "to laugh", is the only example in Erish of a non-initial hj.

Hard consonants are represented with similar treatments, with the grapheme depending upon whether the instance is word-initial, non-initial and short, or non-initial and long:
​

Desired hard consonant	Location and length	Grapheme(s)	Example
h /x/	Initial	hv	hvint "hint" /ˈxɪnːt/
	Non-initial and short	x	exo "echo" /ˈeːxɔ/
	Non-initial and long	hh, rarely hv	sehve "to see" /ˈsɛxːɛ/
k /k/	Initial	qu	quinne "woman" /ˈkɪnːɛ/
	Non-initial and short	qu	kliqu "clique" /ˈkliːk/
	Non-initial and long	kk, rarely kv	republikk "republic" /rɛpʊbˈlɪkː/
sk /sk/	Initial	squ	squeìt "(ice/roller) skate" /ˈskeːt/
	Non-initial and short	N/A
	Non-initial and long	squ	fresque "fresco" /ˈɸrɛsːkɛ/

• Hv is irregularly used before a in native Erish vocabulary, such as in hval "whale" /ˈxɑːl/
• Medial hv, kv have similar uses to the consonant plus j digraphs, indicating that the verb's conjugation is irregular and may result in softening.
• The hv, kv, qu spellings are rooted in how Old Erish, in contrast to the rest of the West Germanic languages, never resolved labiovelars into /Kw/ sequences. Instead, these were preserved into Old Erish, and failed to palatalize. When they were lost as phonemes, they simply delabialized from /Kʷ/ to /K/, helping phonemicize the palatal allophones.

G-Consonants
The letter g is not the only consonant grapheme of the last group of Erish consonants, but the set is essentially used to indicate hardness and softness not unlike h, k, and sk. It is grouped by itself because it represents six phonemes and has less consistency in which graphemes are used to represent hardness and softness for those phonemes. In principle, g is just the voiced counter part to h, k, and sk, with the addendum that before nasals it has a nasal pronunciation that can be hard or soft; the fricativization it experiences is expectable because it is voiced.
​

Letter(s)	Conditions or example	Softens initially	Softens elsewhere	Hard
g (short), cg (long)	Regular conditions	to /ʝ/ before e, i, y, æ, or œ, and another vowel	to /ʝ/ after e, i, y, æ, or œ, and another vowel	/ɣ/ in all other regular conditions
	Regular examples	gæst "guest" /ˈʝæsːt/	weges "ways" /ˈɣeːʝɛs/	gá "to go" /ˈɣoː/
	Fricative conditions	N/A	to /ʃ/ between the vowels e, i, y, æ, and œ, and the end of a word or the consonants d, s, or t	to /x/ between all other vowels or l or r, and the end of a word or the consonants d, s, or t
	Fricative examples	N/A	ig "I" /ˈiːʃ/	dag "day" /ˈdɑːx/
	Nasal conditions	N/A	to /ɲ/ between the vowels e, i, y, æ, and œ, and n	to /ŋ/ between all other vowels or l or r, and n
	Nasal examples	N/A	regn "rain" /ˈrɛɲːn/	Ragnar /ˈraŋːnɑ

• cg softens only if there is a preceding i or y, similar to k and ck.

Ideally, g should have a regular distribution of "soft in hard contexts" graphemes and "hard in soft contexts" graphemes, but the reality is that etymology means different graphemes are used in what should be regular contexts. Nonetheless, there aren't so many graphemes in use:
​

Desired consonants	Location and length	Grapheme(s)	Example
Soft g (/ʝ/, /ʃ/, /ɲ/)	Initial	j, gj (less common)	jorþ "earth" /ˈʝɔrːs/, gjos "they (f.)" /ˈʝoːs/
	Non-initial and short	j	garaj "garage" /ɣaˈrɑːʃ/
	Non-initial and long	zg, gj (rare)	brizg "bridge (game)" /ˈbrɪʝː/
Hard g (/ɣ/, /x/, /ŋ/)	Initial	w, gv	gvad "what" /ˈɣɑː/, west "west" /ˈɣɛsːt/
	Non-initial and short	w	intriw "intrigue" /ɪnˈtriːx/
	Non-initial and long	gg	rigg "rig" /ˈrɪɣː/

• Gj is mostly confined to native Erish vocabulary, and is not particularly common initially, and even more so non-initially. As with other Cj graphemes, non-initial gj indicates irregular conjugation.
• J is the main grapheme for representing soft-g in hard contexts, and g for representing regular soft-g, but there exceptions like Jesu "Jesus".
• In native Erish vocabulary, gv is restricted to wh-words such as gvad "what", but it is regularly used in loanwords when the donor language has g, such as gverilja "guerrilla" /ɣɛˈrɪʎːa/.
  • The wh-words in Erish were originally /xʷ/, but underwent the same vocalization as the third-person and demonstrative pronouns, making Old Erish have /ɣʷ/ as a marginal phoneme, until it delabialized during Middle Erish.
• W is the main consonant used to represent initial hard-g in native Erish vocabulary, and is more rare with loanwords. However, it is mandatory in non-initial, short contexts, much like j.
  • W was originally /w/ in Old Erish. It merged with /ɣ/ as a consequence of the fricative leniting to an approximant, effectively making /w/ first merge with /ɣʷ/, and then delabialize /ɣ/.
• Similar to the situation with k, long /ɣː/ failed to palatalize in Eastern dialects, and so zg is used, even though the grapheme is redundant in Hamnstead Erish.

Feedback Questions
Questions I personally have are:

• Does the type of spelling I have decently blend etymological orthographies like Icelandic and more phonemic ones? I recognize this is more an aesthetic question, but I'm interested in opinions.
  • One area I wonder about in particular is whether j and w should be used for representing non-initial /ʝ/ and /ɣ/. In my mind, they make sense since they are representing approximants, but w strikes me as rather... odd.
• Do the phonological developments seem like a good mixture of sharing some of the innovations that occurred in Norwegian and Swedish and Erish following its own path?
• I know a sibilant like [ʐ] (the likely realization of Proto-West Germanic *z, if not Proto-Germanic *z) devoicing to /ʂ/ isn't abnormal. However, given how universal the change of /z/ to / was throughout the surviving Germanic languages, how plausible is it to have Erish do this?

Resources

This last section is dedicated to the resources I think have been most useful in the creation of Erish, and that are valuable to people looking to make a Germanic language, or even a conlang in general. I'm certain some of these sources will be familiar to many members of this sub, but they're there for those to whom they aren't:

• Agee, A Glottometric Subgrouping of the Early Germanic Languages - This thesis provides a pretty good overview of how the Germanic languages developed at their earliest stages after Proto-Germanic. Even if you're not trying to make a Germanic language, I can't imagine that this wouldn't give you at least some ideas.
• Index Diachronica - Index Diachronica is a good site in concept and mostly in implementation. If you want to search for a general to sometimes comprehensive idea of the phonological histories of languages, or how certain phonemes tend to change over time, this is a good place to stop by.
  • I got a history of Proto-Norse to Old West Norse here, although I wound up having to directly go and sift through the """""human-readable""""" source because ID wasn't clear at times (and apparently misinterpreting at points). That's the major problem with ID - its sources or interpretation of those sources may be "sketch" or incomplete, to say the least.
• Jackson Crawford - It's one thing to read internet articles, or even textbooks about old Germanic languages. It's another to have some sort of access on demand to somebody who can actually explain to laypeople an old Germanic language, much less its culture. I cannot recommend his channel enough as a starting and reference point for somebody interested in Germanic conlanging, especially if you're going for a (sort-of-)North Germanic language.
• Ringe, A Linguistic History of English, vols. 1 & 2 - If you are looking for information about old Germanic languages, these are some of the best materials to look through for vocabulary, morphology, and phonology. Erish would not exist as it does without Ringe's reconstruction of Proto-West Germanic in Vol. 2, much less the far more consistent account of the phonological history of Northwest and West Germanic than is available on Wikipedia.
• Simon Roper - Though I can't say Simon has been as much a resource for me as Dr. Crawford - and as Simon himself acknowledges, his field is archaeology, not linguistics - he still provides good material about the phonological history of English, and information about Old English.
• Wikipedia - Wikipedia has articles about all manner of languages and their phonology, grammar, vocabulary, orthography, and so on (much less articles about those in and of themselves). Although obviously I've mainly relied on Germanic languages, I would be remiss to say that one should look exclusively at a particular family. Case in point, /d,ʝ/ devoicing and spirantizing to /s,ʃ/ was a feature inspired by Nahuatl.
• Wiktionary - Wiktionary has multiple features which are extremely useful. Beyond a Proto-Germanic category and Germanic Swadesh list appendix, Wiktionary has an immensely handy feature where translations are often provided through individual senses of English words - anybody wanting to reduce relexing should take note.

Conclusion

It's been a long post, so I won't take up so much more space. Seeing as how it's a common enough translation in initial posts, though, I would like to provide the Lord's Prayer in Erish:
Written Erish:

Faþern osren, Hlárden gwen bez í Hjomn,
Be namen zín werþe heligende;
Be ríked zítt kome;
Be wiljo zí skehe pá jorþo zí som í Hjomn;
Be geve til oss í dag ossert daglige brod,
end forláte oss skuldostos osros sá som wid forláte osros skuldos;
End be bringe ick pá oss í fresnos, men frælse pá oss frá yvel.
Zítt bez ríked, end mahten end ero í œighedo.
Amen.

​
Hamnstead Erish pronunciation:

[ˈɸɑːs̪ɐn ˌɔs̪ɾn̩ ˈxl̥ɔrːd̪n̩ ɰn̩ bəʃ ˌɛɪ̯ ˈʃɔmːn]
[ˈbɛɪ̯ː ˈnɑːmn̩ ˌʃɛɪ̯n ˌɰɛɾs̪ə ˈʃɛɪ̯ːlɪjn̪̩d̪ə]
[ˈbɛɪ̯ː ˈrɛɪ̯ːt͡ʃə ˌʃɛt̪ ˈkʷʰoːmə]
[ˈbɛɪ̯ː ˈɰɪʎ̟ːɔ ˌʃɛɪ̯ ˈʃɛɪ̯ːʃə ˌpɔʊ̯ ˈjʷɔrːs̪ə ˌʃɛɪ̯ s̪ɔm ˌɛɪ̯ ˈʃɔmːn]
[ˈbɛɪ̯ː ˈjɛɪ̯ːβ̞ə ˌt̪il ˌɔs̪ ˌɛɪ̯ ˈd̪ɑːx ˌɔs̪ɐt̪ ˈd̪ɑːxˌlijə ˈbɾoːs̪]
[n̪̩d̪ ɸɔɾˈlɔʊ̯t̪ə ˌɔs̪ ˈs̪kʷʊl̪ːd̪ɔs̪t̪ɔs̪ ˌɔs̪ɾɔs̪ ˌs̪ɔʊ̯ s̪ɔm ˌɰi ˈɔs̪ːɾɔs̪ ˈskʷʊl̪ːd̪ɔs̪]
[n̪̩d̪ ˈbɛɪ̯ː ˈbɾɪŋːə ˌɪt͡ʃ pɔ ˌɔs̪ ˌɛɪ̯ ˈɸɾ̥eːs̪ˌnɔʊ̯s̪ mn̩ ˈɸɾ̥æl̪ːs̪ə pɔ ˌɔs̪ ˌɸɾ̥ɔʊ̯ ˈyʉ̯ːβ̞l̩]
[ˌʃɛt̪ bəʃ ˈrɛɪ̯ːt͡ʃə n̪̩d̪ ˈmɑxːt̪n̩ n̪̩d̪ ˈɛɪ̯ːɾɔ ˌɛɪ̯ ˈœʏ̯ːˌjiˌʃːɛɪ̯s̪ɔ]
[ˈɑːmɛ̠n]

​
Gloss:

father-the.m.sg our.incl-m.sg lord-the.m.sg who.m.sg.dir be.fut.sg in Heaven
be.fut.sbjv name-the.m.sg thy.m.sg become-inf sanctify-ptcp.prs.m.sg
be.fut.sbjv realm-the.n.sg thy.n.sg come-inf
be.fut.sbjv will-the.f.sg thy.f.sg happen-inf on earth-the.f.sg thy.f.sg as in Heaven
be.fut.sbjv give-inf to us.incl.dat in day our.incl-n.sg daily-def.n.sg bread
and forgive-inf us.incl.dat guilt-the.f.pl our.incl-f.pl so as we.excl forgive-inf our.incl-f.pl guilt-pl
and be.fut.sbjv bring-inf not acc us.incl.acc in temptation but free-inf acc us.incl.acc from evil
thine be.fut.sg realm-the.n.sg and power-the.m.sg and glory-the.f.sg in eternity-the.f.sg
amen

​
English translation:

Our father, the Lord who will always be in Heaven,
May it always be that thy name sanctifies itself;
May it always be that thy kingdom comes;
May it always be that thy will comes upon your Earth as in Heaven;
May it always be that thou givest us our daily bread,
and forgive our guilts for us, as we forgive our guilts;
And may it always be that thou dost not bring us into temptation but free us from evil.
Thine will always be the kingdom, and the power and the glory in eternity.
Amen.

Notes about the translation:

• In essence, bez in this Prayer is used for "(will always) be...", and be for "(may it always be that)...". Bez and be are the future tense forms of weze ("to be") (the only verb with them), but this tense is often gnomic, indicating a timelessness and fundamental-fact-of-the-universeness.
  • Be is used here as a relic subjunctive with optative mood. In Erish, it is often used in prayers - Christian, Ardist (the native Erish religion), or otherwise - as a sort of "polite asking".
• Hjomn "(Christian) Heaven" in lines 1 and 4 is, similar to many Erish terms relating to Christianity, a loanword from Old Anglic (English) heofon, and is a doublet of Erish himmel ("sky")
• Werþe heligende (lit. "become sanctifying") in line 2 is an analytic mediopassive voice construction, similar in function to the -s suffix of North Germanic languages.
• When the consistently inclusive "we" (Christians and non-Christians) throughout the prayer changes to the exclusive wid in line 6, it indicates that Christians forgive everybody's guilts.
  • Christianity never really took off in Erishland beyond the Allamunnic minority, and "wasn't particularly cared for" until recently.
• Pá "(up)on" is used here as a direct object marker for "us" in line 7. In Erish, pá is used for animate direct objects.