In search of the Old Hungarian vowels
Well, this is unusual. I am writing an article on read.cash about linguistics. Quite the novelty. Still, I wanted to write this article for a long time, and figured - why not put it out here?
Anyway, this article will assume that you are already somewhat familiar with the International Phonetic Alphabet. If you are not, I recommend you to watch these two videos before further reading this article:
Okay, now that all the linguistically illiterate people are out, let's get to the juicy parts: the issues with reconstructing Old Hungarian, more specifically Old Hungarian pronounciation.
Modern Hungarian phonology
Before getting into Old Hungarian phonology, it is important to first give readers a bit of a brief introduction to Modern Hungarian phonology.
It is believed that Modern Hungarian has 14 vowels /ɒ aː o oː u uː ɛ eː i iː ø øː y yː/ respectively written <a á o ó u ú e é i í ö ő ü ű> and and 25 consonants /m p b f v n t d t͡s d͡z s z r l t͡ʃ d͡ʒ ʃ ʒ ɲ c ɟ j k ɡ h/ respectively written <m p b f v n t d c dz sz z r l cs dzs s zs ny ty gy j k g h>, with the following caveats:
There is also a marginal vowel /ɒː/ that appears only in the word <Augusztus> /ɒːgustuʃ/
Some dialects still maintain a distinction between the open /ɛ/ and the close /e/, which has been lost in Standard Hungarian and most dialects. In some folk songs or poems where this distinction is relevant, the short close /e/ can be written as <ë>. Southern dialects - namely, the dialect of Szeged - has instead merged /e/ into /ø/, hence the local pronunciation of the city's name being /søgɛd/ rather than the Standard /sɛgɛd/, indicating that historically it must have been /segɛd/.
/ɒ/ isn't actually an open back rounded vowel [ɒ], but rather a near-open near-back weakly rounded vowel, which is difficult to mark in the IPA: [ɑ̝̜̈], [ɐ̠̜], [ʌ̞̜̈], etc. For the sake of simplicity, I will use /ɒ/.
/ɛ/ is phonetically more open [æ].
/o ø/ is phonetically more mid [o̞ ø̞], while their long counterparts /oː øː/ are close-mid [oː øː].
While the long /iː yː uː/ are truly closed [iː yː uː], the short /i y u/ are somewhat opened and centralized [ɪ ʏ ʊ], though not to the same degree as in English and other Germanic languages. Native Hungarian-speakers still perceive the main difference to be vowel length, rather than vowel quality.
The front rounded vowels /ø øː y yː/ are not truly front, but rather near-front, somewhat centralized [ø̞̈ ø̈ː ʏ̈ ÿː]
Historically, Hungarian also had an (alveolo-)palatal lateral approximant /ʎ/, written <ly>. In Standard Hungarian and the majority of dialects, it merged into /j/, but is still written <ly>.
/n t d t͡s d͡z s z r/ are actually denti-alveolar [n̪ t̪ d̪ t̪͡s̪ d̪͡z̪ s̪ z̪ r̪], while /l/ is just alveolar [l]
/ɲ c ɟ/ are actually alveolo-palatal [n̠ʲ t̠ʲ d̠ʲ]. /c ɟ/ in particular vary between stops [t̠ʲ d̠ʲ] and non-sibilant affricates [t̠ʲ͡ç d̠ʲ͡ʝ]. In the affricate variants, the stop components are alveolo-palatal, while the fricative components are dorso-palatal, non-sibilant.
/j/ has the fricative allophones [ç ʝ] when following a stop consonant, especially before a word-boundary, with /kɒpj/ and /dobj/ respectively being [kɒpç] and [dobʝ].
/h/ has four allophones:
/n/ assimilates to the following consonant, with /np nb/ being [mp mb], /nf nv/ being [ɱf ɱv], /nc nɟ/ being [ɲc ɲɟ], /nk ng/ being [ŋk ŋg], etc.
/d͡z/ is always geminated [d͡zː] when between two vowels
/d͡ʒ/ exists only in loanwords, and is always geminated [d͡ʒː] when between two vowels, except in a few recent loanwords from English. Historically, it was treated like a consonant cluster rather than a proper affricate, just like in Russian, but now it has been established well enough as an independent consonantal phoneme.
In addition to these caveats, it is also important to note, that Hungarian employs heavy amount of consonant assimilation:
Stop, affricate and fricative consonants regressively assimilate in regards to voicing, e.g. <Buddha> /budhɒ/ being actually pronounced [bʊthɒ], with the /d/ assimilating to the /h/ by devoicing. The only exception to this is /v/, which does not cause the voicing of preceding consonants, phonologically behaving more like an approximant than a fricative.
/t.s/ and /c.s/ assimilate to [t͡sː]
/t.ʃ/ and /c.ʃ/ assimilate to [t͡ʃː]
/d.z/ and /ɟ.z/ assimilate to [d͡zː]
/d.ʒ/ and /ɟ.ʒ/ assimilate to [d͡ʒː]
/ʃ.s/ assimilates to [sː]
/s.ʃ/ assimilates to [ʃː]
/ʒ.z/ assimilates to [zː]
/z.ʒ/ assimilates to [ʒː]
/n.j/ and /n.ɲ/ assimilate to [ɲː]
/t.j/ and /t.c/ assimilate to [cː]
/d.j/ and /d.ɟ/ assimilate to [ɟː]
/l.j/ historically assimilated to [ʎː], which is now pronounced as [jː] in Standard Hungarian.
For most speakers, /l.r/ assimilates to [rː]. Not for me personally though.
To an extent, Hungarian historically employed vowel assimilation as well. The language still has a feature called "vowel harmony" - also present in other Uralic languages (such as Finnish) and various Turkic languages. It effectively means, that vowels are logically subdivided into two groups - front and back - and that if a a word's last syllable contains a front vowel, all the conjugations must use front vowels too; if a word's last syllable contains a back vowel, all the conjugations must use back vowels too.
A good example to vowel harmony would be the following two nouns and their plural forms:
gyerek (child) -> gyerekek (children)
bot (stick) -> botok (sticks)
At one point, this vowel harmony was even more strict, demanding words to consist solely of front vowels or back vowels without allowing words to mix and match.
It is also important to note, that in many words, /i iː/ act as back vowels, despite being front vowels. This is because historically, they had back variants in an earlier form of Hungarian /ɯ ɯː/, but those merged with the front variants. When did this merger happen? We don't know. It could have happened any time between 1500 BC and 1500 AD. Whenever it happened, words that historically contained /ɯ ɯː/ are still conjugated using the back vowelled affixes, despite the fact that they are pronounced with /i iː/. To make matters worse for language-learners, we make no distinction in writing either, so if you want to learn Hungarian, you will just have to learn the words that historically contained /ɯ ɯː/ as exceptions, treating them like irregular nouns or something.
Let's go back in time
By convention, we split the periods of the Hungarian language into:
Proto-Hungarian: before 896 AD
Old Hungarian: from 896 AD to 1526 AD
Middle Hungarian: from 1526 to 1772
Modern Hungarian: from 1772
While in this article, I will be focusing on Old Hungarian - more specifically, Medieval Hungarian, from 1001 to 1500 - I will also address the other periods as well. I already wrote a good deal about Modern Hungarian phonology in this article, so, without ado, let's go back, as early as we can.
Hungarian is a Uralic language. While it was heavily influenced by Turkic, Slavic, Germanic and Latin languages, taking lots of loanwords from them, at its heart, Hungarian remains a Uralic language, more specifically a Ugric one. No, fellow Hungarian Nationalists, this doesn't exclude the possibility of affinity with Turkic peoples, you can put down the pitchforks (especially if you are like me, and also believe in the Uralo-Altaic language family).
The Ugric family consists of Hungarian and the Ob-Ugric languages (Mansi, Khanty). It is believed that proto-Hungarian split off from the Ob-Ugric languages around 1000 BC. The Ugric languages split off from Proto-Uralic around 2000 BC.
From Proto-Uralic to (proto-)Hungarian
Proto-Uralic had vowels 8 vowels <ä a e o i ü ï u> /æ ɑ e o i y ɯ u/ and a debatable amount of consonants, whose precise pronunciations are still up to debate.
It is taught in high schools, that when proto-Hungarian split off from proto-Uralic, /mp nt ŋk/ became [b d g], intervocalic /p t k/ were lenited [β z ɣ], word-initial /p/ was sprintalized to [ɸ], word-initial /k/ was sprintalized to [x] only before back vowels, word-final vowels were eventually lost, etc. Out of these consonants, /ɣ/ was eventually lost altogether, /ɸ β/ became labiodental [f v] (I'll get back to the voiced one later) and /x/ debuccalized to [h] in Modern Hungarian.
To be more specific, during the transition from proto-Uralic to proto-Hungarian:
the proto-Uralic /ð ðʲ/ became [l ʎ], though the latter sometimes became [ɟ]. This convinced some linguists that Proto-Uralic /ð ðʲ/ were actually [ɮ ɮʲ].
Proto-Uralic word-initial /s/ became silent, /ʃ/ was usually depalatalized to /s/, but seems to have remained /ʃ/ in some situations
/mp nt ŋk ɲʃ/ became [b d g ɟ]
/ŋ/ consistently became [g]
word-initial /p/ was sprintalized to [ɸ]
word-initial /k/ was sprintalized to [x] only before back vowels
The original proto-Uralic gemination was lost, but then later gemination was re-gained
Most word-final vowels were lost.
During this period, Hungarian came into contact with Permic, Iranic and Turkic languages, and took a heavy amount of loanwords from them. This led to the introduction of words with word-initial voiced stops /b d g/, the reintroduction of word-initial /p/, the reintroduction of /k/ before back vowels and introduction of /x/ before front vowels, etc.
Besides the aforementioned changes to the consonantal system from proto-Uralic, the vowel system, I assume, was probably nearly identical to that of proto-Uralic, except for the introduction of vowel length, introduction of the vowel /ø/ and potential early loss of /ɯ/. Generally, in root words that were originally multi-syllabic in proto-Uralic but became monosyllabic in proto-Hungarian via the loss of word-final vowels, the remaining vowel was typically lengthened, though there are some exceptions.
Examples: PU "kala" -> MH "hal" (fish), PU "kalaw" -> MH "háló" (net), PU "pälä" -> MH "fél" (half), PU "juka" -> MH "jó" (good, river), PU "jäŋi" -> MH "jég" (ice)
From Proto-Hungarian to Old Hungarian
According to official, mainstream history, in 896, the Hungarians settled in the Pannonian Basin, bringing their language with them. While Hungarian did have a writing system - Rovásírás - written sources in it are rare and few between. In 1001, Latin became the official language of the Kingdom of Hungary, and remained so until 1844. Hungarian was seldom written, but when it was, the writing system was very rudementary, missing a lot of distinctions.
What is /v/ in Modern Hungarian - possibly [w] or [β] in Old Hungarian - was inconsistently written as either <v> or <w>
What is /h/ in Modern Hungarian - possibly [x] in Old Hungarian - was inconsistently written as either <h> or <ch>
/k/ was inconsistently written as <k>, <c> and <ch>
both /g/ and /ɟ/ were written as <g>
/n t l/ weren't distinguished from their palatal variants /ɲ c ʎ/, typically being written <n t l>
<s> stood for /t͡ʃ/, /ʃ/ and /ʒ/
<z> stood for /t͡s/, /s/ and /z/, though /t͡s/ was sometimes also written <c>
No real distinction was made in vowel length, though what are /øː oː/ in Modern Hungarian were often written as <ew ou>, potentially indicating diphthongal pronounciations [eu̯ ou̯].
No distinctions made between /u/, /uː/, /y/ and /yː/, all written as <u>, also used sometimes for /o/ and /ø/
No distinctions made between /i/ and /iː/ - or potentially /ɯ ɯː/, if they still existed at this point -, both written as <i>, sometimes <y>.
No distinctions made between /ɛ/, /ɛː/, /e/, /eː/, all written as <e>
If the language had separate /a/ and /aː/, no distinction was made between them, both written as <a>
If the language had separate /ɔ/ and /ɔː/, no distinction was made between them, both written as <o>
It is believed that Old Hungarian's consonantal inventory was virtually identical to that of Modern Hungarian, except for the absence of /d͡ʒ/ and presence of /ʎ/. However, their pronounciations were most likely different from Modern Hungarian, namely:
What is /h/ in Modern Hungarian was probably [x] in Old Hungarian, due to the frequent spelling as <ch>
What is /v/ in Modern Hungarian was probably [w] or [β] in Old Hungarian, due to the frequent spelling as <w>
/l/ was dental [l̪] rather than alveolar [l]. It could have been possibly velarized [ł̪] like in neighbouring Slavic languages, but this is purely speculative.
There appears to have been variants of Old Hungarian, where palatal consonants were depalatalized in certain positions, evidenced by the Modern Hungarian doublets "disznó" and "gyisznó", "lány" and "lyány". While Modern Hungarian retains the palatal consonants, in the aforementioned cases, the depalatalized forms became the standard ones, "lány" and "disznó" instead of "lyány" and "gyisznó".
There may have been a highly unstable /ɣ/ early on, inherited from Proto-Hungarian. This consonant eventually vocalized, creating diphthongs first, which eventually monophthongized. I'll get back into this later.
While Modern Hungarian has 14 vowels, Old Hungarian - at least, according to the mainstream - had a whopping 16 vowels /ɔ aː ɛ ɛː o oː ø øː e eː u uː y yː i iː/, though I personally also theorize the existence of /a ɔː/ and possibly /ɯ ɯː/ too, which means that Old Hungarian could have had anywhere between 16 and 20 vowels.
The Old Hungarian vowels /aː ɛ ø øː o i iː y yː u uː/ clearly map to their Modern Hungarian equivalents /aː ɛ ø øː o i iː y yː u uː/. All the other vowels are slightly more complicated.
The Old Hungarian /ɛ/ and /e/ merged as /ɛ/ in Modern Standard Hungarian, though some dialects either retain /e/ or have merged it with /ø/ instead. The latter is in particular what Southern dialects - such as the Szeged dialect - do.
In some cases, the Standard Hungarian form of the word also uses /ø/ instead of /ɛ/, with "sör" (beer) and "csöcs" (boob) coming to mind, both having archaic and dialectal variants "ser" and "csecs". However, it is also entirely possible that the the forms with /ø/ were the original ones, and the archaic variants simply came from a dialect that unrounded them.
The Old Hungarian /ɛː/ and /eː/ merged as /eː/ in Modern Standard Hungarian, though some dialects raised the original /eː/ to /iː/ before the shift to prevent a merger.
If you are a staunch believer of the 16-vowel model, then the Old Hungarian /ɔ/ corresponds to the Modern Hungarian /ɒ/. However, if you are like me and believe that Old Hungarian also had /a/, then the Modern Hungarian /ɒ/ is a result of a merger of /a/ and /ɔ/. I'll make the case for /a/ and /ɔː/ later in this article.
If you are a staunch believer of the 16-vowel model, then the Old Hungarian /oː/ corresponds to the Modern Hungarian /oː/. However, if you are like me and believe that Old Hungarian also had /ɔː/, then the Modern Hungarian /oː/ is a result of a merger of /ɔː/ and /oː/. I'll get back to this later.
We don't know when did the back unrounded /ɯ ɯː/ merge into /i iː/, but it must have happened by the 16th century at latest. As I said, it could have happened at any date between 1500 BC and 1500 AD, we just don't know.
If we are to use the 16-vowel model, then the Old Hungarian /ɔ aː ɛ ɛː o oː ø øː e eː u uː y yː i iː/ map to the Proto-Hungarian /ɑ ɑː æ æː o oː ø øː e eː u uː y yː i~ɯ iː~ɯː/, which in turn correspond to the Proto-Uralic /ɑ~o æ o~u y e~i~ɯ u y i~ɯ/. In Slavic loanwords in particular, the Old Hungarian /ɔ/ corresponds to the Slavic /o/, with words like <pop> /pop/ -> /pɔp/ -> <pap> /pɒp/ (priest).
The Case for Old Hungarian /w/, /a/ and /ɔː/, and maybe /ɣ/?
The Modern Hungarian reflex of the Proto-Hungarian /w/ and /β/ is [v]. However, what is now /v/ in Modern Hungarian was inconsistently written as either <v> or <w> in Old Hungarian, with the latter particularly being preferred in the more German-influenced orthograpies - and we know for a fact, that the German <w> represented a labio-velar approximant [w] rather than a labiodental fricative [v] well into the 17th century.
So why does it matter, if at least some - but possibly all - of Modern Hungarian /v/ was actually /w/ in Old Hungarian? It matters a lot, because it makes it so much easier to explain why so many Slavic loanwords with /av/ in the original Slavic source have /oː/ in Modern Hungarian, with notable examples being Ladislav /ładisłav/ -> László /laːsloː/, Stanislav /stanisłav/ -> Sztaniszló /stɒnisloː/, židov /ʒidov/ -> zsidó /ʒidoː/, etc.
If we postulate Old Hungarian had /w/ instead of (or in addition to) /v/ or /β/, then it naturally follows - based on the above-presented evidence - that the Slavic /stanisłav/ became either /stanislaw/ (my model) or /stɔnislɔw/ (strict 16-vowel model) in Old Hungarian, which eventually became /stɒnisloː/.
However, the problems with the mainstream model don't end there. I find it hard to believe, that Slavic, Germanic, Latin and Turkic /a/ would become /ɔ/ in Old Hungarian - or that the proto-Uralic /ɑ/ would be reflected as /ɔ/ - so I postulate the existence of a short /a/ in Old Hungarian, which I believe merged with /ɔ/ to create the Modern Hungarian /ɒ/ in the 17th century.
I also postulate the existence of a long /ɔː/ in at least Late Old Hungarian (1300-1500), which was must have been a result of a monophthongization of an earlier /aw/. This /ɔː/ merged with /oː/ in Middle/Modern Hungarian. The monophthongization of /aw/ to /ɔː/ was actually a fairly common vowel shift in Late Medieval European languages (namely Middle English, Middle French and Southern German dialects), so even the time frame seems to match. If we want to make even more analogy between Middle French/English and Late Old Hungarian, the monophthongization of /ew/ or /ɛw/ to /øː/ and /iw/ to /yː/ could also count.
Under my 18-vowel model, out of the 18 Old Hungarian vowels /a aː ɛ ɛː ɔ ɔː o oː ø øː e eː u uː y yː u uː/, /a aː ɛ ɛː ɔ oː e eː u uː y yː i iː/ clearly correspond to the Proto-Hungarian / Proto-Uralic /ɑ ɑː æ æː o oː e eː u uː y yː i iː/, with /ɔː øː/ being exclusive to Late Old Hungarian and being a result of a monophthongization of earlier /aw ew/, /ø o/ mostly being results of Proto-Uralic /y u/ being lowered, and most instances of /oː/ coming from earlier /ɔw/ or /ow/.
My theory, my 18-vowel model is backed up by the fact, that Old Hungarian words of native Uralic origins used <a> rather than <o> for spelling the Hungarian reflex of the Uralic /ɑ/, as well as the obvious monophthongization of Slavic /av/. The monopthongization theory is also backed up by some forms of words in Modern Hungarian, e.g. tó (lake) - tavak (lakes), ló (horse) - lovak (horses), etc.which seems to indicate the original forms being /tɔw/ and /low/ before the monophthongization.
But wait, there's more!
You see, I kept mentioning proto-Hungarian having a voiced velar fricative /ɣ/ that was supposedly lost. Well... it did not disappear without a trace. A few written sources from the 9th century, written in the Runic alphabet, spell words that have /oː øː yː/ in Modern Hungarian as <aɣ eɣ iɣ> respectively. The most salient example is the Modern Hungarian word "tű" (needle) /tyː/ being spelled "tiɣ", or "menő" /mɛnøː/ being spelled "meneh" /mɛnɛɣ/. What's going on here?
You see, /ɣ/ did not exactly disappear without a trace. After a consonant, it was vocalized to /j/, while after a vowel, it was vocalized to /w/. Hence, /tiɣ/ -> /tiw/ -> /tyː/, /mɛnɛɣ/ -> /mɛnɛw/ -> /mɛnøː/, /ʃaɣ/ -> /ʃaw/ -> /ʃɔː/ -> /ʃoː/.
A perfect example is the Hungarian word for Pecheneg, "besenyő", which clearly comes from the Turkic "Beçenek": /bet͡ʃenek~beʃenek/ -> /beʃeɲeɣ/ -> /beʃeɲew/ -> /bɛʃɛɲøː/
The Case for a schwa-like /e/
A friend of mine - to whom I shall only refer as "Tangent" - postulated that because the Old Hungarian <ë> /e/ is reflected as [ɛ] in Standard Hungarian and most dialects, [e] in some dialects and [ø̞] in Southern dialects, it must have had a vowel quality that was somewhere between all three, with his bet being a central mid vowel [ə].
The Case for a Late Loss of /ɯ/
"Feheruuaru rea meneh hodu utu rea" - /fɛhɛːrvaːru reaː mɛnɛɣ hɔdu utu reaː/
So goes the sentence according to the mainstream.... Or, in modern Hungarian: "Fehérvárra menő hadi útra" - /fɛheːrvarːɒ mɛnøː hɒdi uːtrɒ/.
However, I slightly disagree with the mainstream, especially due to the supposed shift of /u/ to /i/. So, my reconstruction of the pronounciation would be /fɛxɛːrwaːrɯ rea mɛnɛɣ xɔdɯ utɯ rea/.
My theory is based on the fact, that /ɯ/ eventually merged into /i/, hence the /xɔdɯ/-/hɒdi/ correspondence seeming far more plausible than /xɔdu/-/hɒdi/, I also went with /utɯ/ instead of /utu/, because it comes from the proto-Uralic "uktï" /uktɯ/. I went with short /a/ instead of long /aː/ in /rea/, because it ended up becoming /rɒ~rɛ/ in Modern Hungarian, and it seems far more plausible if both vowels were short to begin with. In fact, it isn't even that implausible, that the affix had already taken on its modern form by then, and it was just the scribes not bothering with the vowel harmony, instead writing out both vowels in a sequence.
/ɯ/ being misheard - and thus misspelled -as /u/ isn't that implausible.
As I previously said, the loss of /ɯ/ could have happened at any time between 1000 BC to 1500 AD... but if my theory is true, it actually happened some time between 1050 AD and 1500 AD. It's interesting to note, that a lot of Turkic /ɯ/ is actually reflected as /ø/ in Modern Hungarian.
The Case for a Creolized Old Hungarian
My friend Nika theorizes that there must have been at least two variants of Old Hungarian: one spoken by the actual Magyar migrants (possessing a vowel system closer to proto-Uralic, just with vowel length distinction and the extra Ö), the other spoken by Magyarized Slavs.
True Magyars would have used [ɑ ɑː æ æː o oː ø øː e eː u uː y yː i iː β h] in "pure Hungarian", while Magyarized Slavs would have used [ɔ aː ɛ ɛː ɔ oː ø øː e eː u uː y yː i iː w x] in their Slavo-Hungarian creole, and Middle Hungarian would have been the result of some sort of decreolization process, the collapse, the merger of the two varieties into one. One possible evidence to this is the fact that in Old Hungarian, Slavic loanwords preserved forms closer to their originals - in terms of phonotactics too - while in Modern Hungarian, they often have epenthetic vowels inserted. A good example is the Modern Hungarian "barát", which was "brat" in Old Hungarian and in the various Slavic langauges. She also made this suggestion based on the observation that Old Hungarian was "more Slavic" than Modern Hungarian, due to the presence of /x/ and /w/.
While the hypothesis that the early Hungarians came in with just the men and married Slavic women is obviously untrue (we wouldn't be speaking Hungarian now, if it was true, since the mother tongue is inherited from the mother), it is true that there was a lot of intermixing and intermarriage between Magyars and Slavs during the Middle Ages, nearly always leading to the Magyarization of Slavs. That, and a continuous stream of immigrants - not just of Slavic stock, but also Germanic, and to a lesser extent Latin/Frankish - getting assimilated, Magyarized.
After the transition to Old Hungarian to Middle Hungarian, we see a slight decline in the Slavic character of Hungarian, and instead, the language taking on more German(ic) overtones - still obviously remaining a Uralic langauge to its core though.
Modern Hungarian to Old Hungarian vowel correspondences under the 20-vowel model:
If a word has /ɒ/ in Modern Hungarian, it must have had either /ɔ/ or /a/ in Old Hungarian. If the etymology of the word is known, it's worth looking at its closest Slavic, Turkic, Germanic or Uralic cognates. If the etymology of the word is unknown, it is impossible to determine if it had /ɔ/ or /a/.
If a word has /aː/ in Modern Hungarian, it clearly had /aː/ in Old Hungarian.
If a word has /o/ in Modern Hungarian, it clearly had /o/ in Old Hungarian.
If a word has /oː/ in Modern Hungarian, it had /ow/, /ɔw/ or /aw/ in Old Hungarian, depending on the etymology.
If it has /uː/ in some dialectal forms, it must have been /oː/ in Late Old Hungarian, thus /ɔw/ or /ow/ in Early Old Hungarian
If it has /ɒu/ or /ou/ in some dialectal forms, it must have been /ɔː/ in Late Old Hungarian, thus /aw/ in Early Old Hungarian
If a word has /u/ in all forms of Modern Hungarian, it clearly had /u/ in Old Hungarian too
If a word has /uː/ in all forms of Modern Hungarian, it clearly had /uː/ in Old Hungarian too
If a word has /ø/ in all forms of Modern Hungarian, it clearly had /ø/ in Old Hungarian.
If a word has /øː/ in all forms of Modern Hungarian, it was /øː/ in Late Old Hungarian, which was originally /ew/ or /ɛw/ in Early Old Hungarian.
If a word has /y/ in all forms of Modern Hungarian, it clearly had /y/ in Old Hungarian too
If a word has /yː/ in all forms of Modern Hungarian, it was either /yː/ or /iw/ in Old Hungarian
If a word has /ɛ/ in all forms of Modern Hungarian, it clearly had /ɛ/ in Old Hungarian too
If a word has /ɛ/ in Standard Hungarian but /ø/ in some dialects - or vice versa - it clearly had /e/ in Old Hungarian
If a word has /eː/ in all forms of Modern Hungarian, it clearly had /ɛː/ in Old Hungarian
If a word has /eː/ in Standard Hungarian but /iː/ in some dialects, it clearly had /eː/ in Old Hungarian
If a word has /i/ in all forms of Modern Hungarian, it clearly had /i/ or /ɯ/ in Old Hungarian, depending on the affixation of the word
If a word has /iː/ in all forms of Modern Hungarian, it clearly had /iː/ or /ɯː/ in Old Hungarian, depending on the affixation of the word
Obviously, this cannot be applied entirely consistently, because Old Hungarian also had some random funky vowel switches - such as the Modern Hungarian "malaszt" /mɒlɒst/ being "milozt" /milɔst/ in Old Hungarian, but otherwise, I feel like it's a good place to start.
And with that, I leave you with the reconstruction of the pronounciation of the first three sentences of the Funeral Sermon and Prayer in my 20-vowel system:
Latiatuc feleym zumtuchel mic vogmuc. yſa pur eſ chomuv uogmuc. Menyi miloſtben terumteve eleve miv iſemucut adamut. eſ odutta vola neki paradiſumut hazoa.
/laːtjaːtok fɛlɛim sømtøkːel mik wɔɟmuk. ɯʃa, por eːʃ xɔmow wɔɟmuk. meɲːi milɔstben tɛrømtɛweː elɛwɛ miw iʃemykøt aːdaːmot. eːʃ ɔdotːaː wɔlaː nɛki paradit͡ʃomot xaːzɔwaː/
I did not just try to transliterate the spelling - I also kept in mind the modern pronounciations of the words, and worked my way from there.
