Case Study: From Proto-Uralic to Hungarian

November 12, 2023 Shivallon Comments 0 Comment

When conlanging, I sometimes make up new features for my languages from scratch, but more often than not, some phenomenon in a natural language served as inspiration for it. Natural languages are often astonishingly beautiful and elegant in the way they express things, and when I encounter a particularly beautiful specimen, I feel like paying tribute to it by incorporating some of its features into one of my conlangs. Hungarian is such a case: Phonetically, is one of the most, if not the most aesthetically pleasing languages to me (just listen to this or this). This is why I have decided to model Nyirvón’s phonology after Hungarian.

The phonology and phonotactics of a natural language (and any well developed conlang) do not arise out of the blue: They are the result of thousands of years of continuous sound change, which created distinct sounds and distributional patterns that constitute the unique “look and feel” of a language. So when we want to model Nyirvón after Hungarian’s example, we should first investigate how Hungarian’s phonology developed! What did its proto language look like? Which sound laws applied on the way to modern day Hungarian? What does Hungarian actually sound like, which sounds are common or rare, and where can they occur in a word?

This will get a bit scientific; full references for all citations (with links where possible) are at the end of the post.

Proto-Uralic

It is widely accepted that Hungarian is a member of the Uralic language family, although the precise organization of the Uralic languages and their relationship to other language families is not always agreed on by all scholars¹. Every couple of years a new theory arises, placing Uralic with this or that other family (Wikipedia has a nice list), and some have even been doubting whether Hungarian is a Uralic language at all! No wonder, since Hungarian is geographically separated from the other Uralic languages, which are located around the Baltic Sea (e.g. Finnish, Estonian, the Sámi languages) and in Russia (e.g. Mansi, Khanty, the Samoyedic languages), and it looks a bit more different from its relatives because of that. Still, the majority of the linguistic community agrees that Hungarian is a Uralic language.

As such, it has evolved from Proto-Uralic, an ancient language estimated to be around 5000 years old². What did Proto-Uralic look like? Even though historical linguists have developed pretty effective methods for reconstructing ancient languages, 5000 years is a very long time, so there is much debate around the phonological details of Proto-Uralic. I went through three sources describing the features of Proto-Uralic³ and will do my best to present the majority opinion in the following section. For the sake of brevity, I cannot reproduce the whole discussion here; also, let’s keep in mind that the purpose of this post is to get some inspiration for the creation of a conlang, not to write a super-accurate scientific article. 😉

Consonants

	Labial	Apical	Retroflex?	Palatal(ized)	Velar(ish)
Stop	p p	t t			k k
Affricate			č ʈ͡ʂ~t͡ʃ	ć* t͡ʃʲ~t͡ɕ
Fricative		s s		ś sʲ
Nasal	m m	n n		ń nʲ~ɲ	ŋ ŋ
Trill		r r
Lateral		l l
Glide	w w			j j
???					x* ɣ~x~χ~ʕ~?

The reconstructed consonant inventory of Proto-Uralic⁴. * marks contested phonemes that are only assumed by 2 out of 3 sources.

Proto-Uralic had labials, apicals (dental or alveolar), palatals (or at least palatalized apicals), and velars, which came as plain stops, affricates, fricatives, nasals, and the obligatory lateral, rhotic and semivowels. There was no voicedness distinction and no word-initial consonant clusters, but many two-consonant medial clusters were allowed, namely several obstruent+obstruent (including geminates), continuant+obstruent, and sonorant+sonorant combinations⁵.

It also had a mystery consonant x⁶, which has left traces in the modern languages, but not enough to settle on its former phonetic value. The sound changes it participated in strongly suggest that it was a back consonant, maybe a velar or laryngeal fricative, but there is not enough evidence to say this with confidence.

The exact place of articulation of the affricate č is also not entirely clear, though it is usually described as being “retroflex” or “cacuminal” (another word for retroflex). Two of the three consulted sources assume a palatalized counterpart ć⁷.

A few other consonants are discussed as potential phonemes of Proto-Uralic: The oldest source, by Péter Hajdú (1972), lists an additional fricative ṥ (same row as č, so probably /ʂ/), and a palatalized lateral ľ /lʲ~ʎ/. These two consonants are also mentioned by Daniel Abondolo (1998a, p. 12) as consonants that can only be reliably reconstructed for Proto-Finno-Ugric, which is the daughter of Proto-Uralic after the Samoyedic languages split off. Pekka Sammallahti (1988), on the other hand, has two dental fricatives d and d́ (/ð/ and /ðʲ/?), which Abondolo (1998a) refers to it as being “questioned” (p. 12). As they are only assumed by one of three sources, I decided to leave them out here.

Vowels

Apart from these few controversial reconstructions, the consonant inventory looks very similar across all three sources. There seems to be more dissent regarding the vowel inventory; Abondolo (1998a) even claims that “[p]robably no academic field is more contentious than the prehistory of the Uralic vowels” (p. 16). However, the most common consensus now seems to be this:

	Front	Back
High	i, ü i, y*	i̮, u ɯ*, u
Mid	e e	o o
Low	ä æ~ɛ	a ɑ~a

The reconstructed vowel inventory of Proto-Uralic⁸. * marks contested phonemes that are only assumed by 2 out of 3 sources.

All three scholars agree that Proto-Uralic must have had a symmetric vowel inventory with an equal number of front and back vowels forming pairs. The oldest source, Hajdú (1972, p. 36), only lists six vowels, while Sammallahti (1988, p. 481) and Abondolo (1998a, p. 17) assume two additional high vowels /y/ and /ɯ/. They are also mentioned by Hajdú as being “contested”. Since the other two sources are more recent, they seem to have won the contest. The Proto-Uralic vocalism had a few more notable features:

Vowel harmony: Most Uralic languages have front-back vowel harmony, and this feature is assumed to have already been present in Proto-Uralic⁹; Abondolo 1998a, p. 1¹⁰. Front-back vowel harmony means that vowels come in front-back pairs and that the vowels in a word (including grammatical affixes) must be either all front or all back. Some languages can also have neutral vowels that are allowed to occur in both front and back contexts. There are other types of vowel harmony, for example with respect to roundedness, which can also be observed in some Uralic languages to a certain extent (notably Hungarian, as we will see later). For Proto-Uralic, scholars usually only reconstruct front-back harmony though.

Reduced secondary vowels: In non-initial syllables, only a subset of the Proto-Uralic vowels is found¹¹, namely one low pair (front /æ~ɛ/ and back /ɑ/) and one high pair (front /i/ and back /ɯ/)¹². The other vowels can only occur in initial syllables.

No length distinction: There is less agreement on this topic, but the majority of Uralicists believe that vowel length was not a distinctive feature in Proto-Uralic, even though it is in most modern Uralic languages¹³. It is assumed that vowel length got introduced by several sound changes, e.g. via consonant deletions (such as the disappearance of the mysterious x) with compensatory lengthening, as we will see later.

Sound changes to Hungarian

My three sources for the Hungarian sound changes¹⁴ list a lot of sound changes. To keep this post at a reasonable length, I will only list the main sound changes, i.e. those that seem to have had a strong effect on Hungarian’s phonology and are mentioned by multiple sources. Some very obscure sound changes that only applied sporadically or in very specific contexts will be left out. Please check out the sources if you want to know all the sound changes.

Proto-Uralic to Proto-(Finno-)Ugric

Kálmán (1972) and Imre (1988) only start off at Proto-Finno-Ugric, so we will have to rely on Sammallahti (1988, p. 486f) here. Fortunately, the sound changes between Proto-Uralic and its daughter language Proto-Finno-Ugric are simple, also because both languages are so old that it’s hard to reconstruct their differences in detail.

(X-1) They mostly pertain to the mystery consonant x, which disappeared after vowels, lengthening them in the process, unless the resulting syllable was closed.

æx → eː
ɑx → oː
Vx → Vː
Vː → V | closed syllable

(X-2) All remaining instances of x turned into /ɡ/ between Proto-Finno-Ugric and Proto-Ugric and were further weakened or eventually disappeared in Hungarian¹⁵.

x → ɡ (Hungarian: → v~j~∅)

(VOW-SHORT) All long vowels that may have arisen from the deletion of x were shortened again in Proto-Ugric, effectively turning the above sound changes into a simple deletion of x in most places¹⁶.

oː → ɑ
eː → æ
Vː → V

Finally, Sammallahti (1988, p. 490) claims that š (probably /ʂ/, as discussed in the previous section) was part of the Proto-Finno-Ugric inventory, but does not explain where it came from. As this sound is important for later sound changes, we just have to accept it at this point.

Proto-(Finno-)Ugric to 10th century Hungarian

The following sound changes are said to have occurred between the 5th century BC and 10th century AD according to Kálmán (1972). These are some of the most important sound changes also listed by most of the other sources, but without any reference to a specific time period.

(WEAK-INIT) Most consonants, especially the stops, are weakened to a certain degree¹⁷. In word-initial position, this affects /p/ and /k/ (before back vowels). Note how the sound changes applies to the sounds at the edges of the consonant inventory, similar to how I removed the labial, velar, and uvular affricates in Quriil in the previous post, but not the alveolars and retroflexes!

p → f | # _
k → χ → h | # _ [+back]

(WEAK-GEM) At the same time, medial geminates are shortened into single consonants. The velar nasal starts to disappear in the process.

ŋː → ɡ | V _ V
Cː → C | V _ V

(WEAK-CLUS) In medial clusters of nasals and voiceless obstruents, the nasal is deleted while its voicing is moved onto the adjacent obstruent. According to Sammallahti (p. 520), this is also the first source of the voiced palatal stop /ɟ/.

[+nasal] sʲ → ɟ | V _ V
[+nasal] [-voice] → [+voice] | V _ V

(WEAK-MED) Other medial consonants are also weakened and often eventually deleted or reduced to a /v/. For /p/, which seems to be especially vulnerable in Hungarian (again, as it’s often the case for consonants at the edges), this even applies in clusters, while the other consonants are only weakened intervocally.

p → w~θ → v~∅ | V (C) _ V, V _ (C) V
t → z | V _ V
k → ɣ → v~∅ | V _ V
m → w → v~∅ | V _ V
ŋ → w → v~j~∅ | V _ V

(SIBIL) The Proto-Finno-Ugric sibilants undergo an unconditional shift in Hungarian: Plain /s/ and /ʂ/ are first weakened and then deleted altogether, while the palatalized /sʲ/ moves into the now free position of the plain /s/. Finally, /ʃ/ enters the Hungarian inventory via the affricate /ʈ͡ʂ/¹⁸.

s, ʂ → h~θ → ∅
sʲ → s
ʈ͡ʂ → ʃ~t͡ʃ

(VOW-FIN) In this period, the vowels are largely unchanged. Only the word-final vowels are heightened and reduced:

ɛ, ɑ, i, ɯ → y̆, ŭ, ĭ, ɯ̆ | _ #

10th century to modern day Hungarian

(PAL-1) The consonant inventory is already mostly in place now. The only notable entries missing are the palatal stops /c/ and /ɟ/. While /ɟ/ evolved from the voiced affricate /d͡ʒ/¹⁹, which most likely came from /ʈ͡ʂ/ via (WEAK-CLUS), /c/ most likely developed afterwards to fill the gap of a voiceless counterpart for /ɟ/. It seems to have come from palatalized /tʲ/ and/or the cluster /tj/²⁰. This is a very limited context for a sound change (also, none of the sources explain where the palatalized /tʲ/ came from), and /c/ is still an extremely rare consonant in modern Hungarian²¹.

tʲ, tj → c
d͡ʒ → ɟ

(PAL-2) Two further sources for /ɟ/ are mentioned by Kálmán (1972, p. 57) and Sammallahti (1988, p. 518f), but this sound change only occurred sporadically:

Sometimes: j → ɟ | # _ [-high]
Sometimes: ɲ → ɟ

(VOW-DEL) All the other major sound changes in this period concerned the vowels. Some of them are lost, in particular most of the previously reduced word-final vowels (cf. VOW-FIN)²², but also some medial short vowels when they were in an unstressed open syllable and preceded by another short vowel in an open syllable²³.

y̆, ŭ, ĭ, ɯ̆ → ∅
[-long, -stressed] → ∅

(VOW-SHIFT) Other vowels changed their position, but that usually did not affect all words. These are the most frequently mentioned vowel shifts that also seem to have applied to a large portion of the vocabulary:

ɑ → ɔ²⁴
æ → e²⁵
u → o | unconditionally²⁶ or in open syllables before /ɑ/²⁷
y → ø, i → (y → ø)²⁸
ɯ → u | unstressed syllable²⁹
ɯ → i | elsewhere³⁰ or when followed by /i/³¹

Other sound changes of this sort affected fewer words, but are still notable³²:

i → e
e → ø

(VOW-LEN) Finally, we see the development of a phonemic vowel length distinction in this period. Vowel length was not introduced by a single decisive sound change, but arose as a side effect of multiple other changes:

A vowel is lengthened when a following short vowel is deleted. Note that these vowels usually appear short again when a suffix is applied to the word, restoring the vowel whose deletion was compensated by the vowel lengthening.³³
A vowel is lengthened when the next syllable contains a long vowel of the same quality.³⁴
A vowel may be lengthened when it is stressed.³⁵
A vowel may be lengthened when an adjacent consonant is deleted.³⁶
A diphthong usually turns into a long vowel.³⁷

These are some of the sound changes affecting diphthongs mentioned by Kálmán (1972, p. 66f):

Vi̯ → iː, sometimes eː
iy̯ → iː, yː, sometimes øː
iu̯ → i
ey̯ → øː
ɑu̯ → oː, sometimes (j)uː

Resulting phonological patterns in Hungarian

Let’s have a look at what modern Hungarian looks like after all these sound changes! Which phonemes does it have, and how do they interact? As we will see, many of Hungarian’s phonological feature are a direct consequence of its proto language and the sound changes it went through to become Hungarian!

Consonants

	Labial	Alveolar	Post-alveolar	Palatal	Velar	Glottal
Stop	p, b p, b	t, d t, d		ty, gy c, ɟ	k, g k, ɡ
Affricate		c, dz t͡s, d͡z	cs, dzs t͡ʃ, d͡ʒ
Fricative	f, v f, v	sz, z s, z	s, zs ʃ, ʒ			h h
Nasal	m m	n n		ny ɲ
Lateral		l l
Trill		r r
Glide				j/ly j

The consonant inventory of modern Hungarian³⁸.

We have seen most of the consonant develop from Proto-Uralic in the previous section, so this inventory shouldn’t be too surprising. A few notes about some of the consonants:

There is some debate about the palatal obstruents ty and gy, namely whether there are stops /c, ɟ/ or affricates /c͡ç, ɟ͡ʝ/, since they are often pronounced more affricate-like. However, their behavior in consonant clusters strongly suggests that they are indeed stops (at least underlyingly)³⁹, so I adopt this view here.
The affricates /t͡s, d͡z/ arose from the deletion of vowels and the resulting consonant clusters.
The phonemic status of /d͡z/ is contested. Siptár and Törkenczy (2000, p. 88f) analyze it as a cluster of /d/ and /z/. However, other sources list it as a proper phoneme⁴⁰, which is why I did too (not that it matters much for this article).

All of the obstruents (except /h/) have voicing opposition and will cause voicing assimilation of preceding obstruents when clustered⁴¹. There are two exceptions to this general pattern:

/h/, having no voiced counterpart, causes devoicing of preceding obstruents, but does not undergo voicing when followed by a voiced obstruent⁴².
/v/, on the other hand, undergoes devoicing to /f/ when followed by a voiceless obstruent, but does not cause voicing of preceding obstruents⁴³. Abondolo (1998b) assumes that this is due to /v/ having “developed from non-distinctively voiced *w” (p. 437).

All consonants except /v/ and /ʒ/ can be geminate⁴⁴.

Consonant frequency

Grimes (2009, p. 43ff) examined Hungarian phoneme frequency by counting the individual consonants and vowels in a large wordlist. By not using Hungrian texts as a source, word frequency does not affect these numbers, so we can really see how frequent certain sounds are in the Hungarian lexicon. Kálmán (1972, p. 77) also provides a list of phoneme frequencies, but does not state how he obtained them. In both cases, about 60 % of the counted phonemes are consonants.

	Grimes – total	Grimes – % of C	Kálmán – % total	Kálmán – % of C
l	18355	11.62	5.84	10.07
t	18196	11.52	7.70	13.28
r	14171	8.97	4.21	7.26
k	12845	8.13	5.30	9.14
s	11517	7.29	3.80	6.56
n	9203	5.83	5.46	9.42
m	8343	5.28	4.12	7.11
g	7699	4.87	2.51	4.33
d	7459	4.72	2.13	3.67
sz	7122	4.51	1.90	3.28
v	5797	3.67	2.06	3.55
z	5658	3.58	2.32	4.00
j	5120	3.24	1.75	3.02
b	4864	3.08	2.06	3.55
p	4000	2.53	0.84	1.45
h	3895	2.47	1.76	3.04
f	3801	2.41	0.92	1.59
ny	2600	1.65	0.74	1.28
cs	2190	1.39	0.56	0.97
c	1976	1.25	0.22	0.38
gy	1880	1.19	1.59	2.74
zs	781	0.49	0.06	0.10
ty	403	0.26	0.10	0.17
*dzs*	62	0.04	0.01	0.02
dz	0⁴⁵	0.00	0.01	0.02
SUM	157937		57.97

Hungarian consonant frequencies according to Grimes (2009) and Kálmán (1972). Grimes only provides total counts, while Kálmán only provides percentage of all phonemes (including vowels). I have computed the percentage of consonants only for both of them.

Certain groups of consonants are much rarer than others, in particular some of the post-alveolars (cs /t͡ʃ/, dzs /d͡ʒ/, zs /ʒ/), palatals (ty /c/, gy /ɟ/, ny /ɲ/), and labials (p /p/, b /b/, f /f/). The three most common ones, on the other hand, are all alveolars (/l/, /t/, /r/ according to Grimes; /t/, /l/, /n/ according to Kálmán).

We can get an idea why this might be the case when we look back at our sound changes:

Proto-Uralic /*ʈ͡ʂ/ changed into /ʃ/ in most places and only occasionally into /t͡ʃ/ (SIBIL). Those instances of /d͡ʒ/ that we got via (WEAK-CLUS) mostly became /ɟ/. The three rare post-alveolar consonants can thus almost only arise secondarily, e.g. via consonant clusters, voicing assimilation, and loanwords⁴⁶.
The palatal stops, in particular /c/, only appeared in very restricted contexts (WEAK-CLUS, PAL-1, PAL-2). I can only speculate about /ɲ/, perhaps it was already uncommon in Proto-Uralic.
The labials were most strongly affected by the weakening sound changes. Proto-Uralic /*p/ basically got lost everywhere except when geminated and in certain consonant clusters (WEAK-*). /b/ only arose from clusters of nasal and /*p/. /f/ is the reflex of word-initial /*p/.
The alveolar consonants /l/ and /r/ were not affected by any major sound changes.
Similarly, /t/ was the least affected stop with respect to the weakening sound changes, only getting reduced in some medial contexts (WEAK-CLUS, WEAK-MED). /n/ was also affected by fewer sound changes than its nasal colleagues (WEAK-CLUS, WEAK-MED).

Consonant clusters

Hungarian allows consonant clusters of up to three consonants in the onset and coda, and up to five consonants medially. However, clusters of more than two consonants are rare in all positions.

All onset clusters are restricted to loanwords⁴⁷, i.e. there is no native Hungarian word beginning with a consonant cluster. This makes sense since Proto-Uralic also didn’t have any initial clusters and we have encountered no sound law that deletes an initial vowel.

Coda clusters also occur in native Hungarian words. They can be produced by (VOW-DEL), the deletion of the final reduced vowel and the deletion of a medial short vowel. Most attested two-consonant clusters are either geminates or clusters beginning with a sonorant. Even though these clusters were also permitted in Proto-Uralic, (WEAK-GEM) and (WEAK-CLUS) have eliminated most of them, so the deletion of medial vowels might have been equally, if not more responsible for their existence than the deletion of final vowels. The other combinations mostly arise from inflectional affixation or are restricted to loanwords and place names⁴⁸. The same holds for three-consonant coda clusters⁴⁹.

A similar distribution can be observed in medial clusters⁵⁰. Almost all two-consonant combinations of sonorant plus another consonant can occur. Sonorants are also frequent second consonants. However, there are quite a few obstruent-obstruent clusters as well. Again, clusters of three or more consonants do not occur in regular native words, but only through affixes and in loanwords⁵¹. This is to be expected considering that Proto-Uralic had at most two-consonant clusters and the only sound change generating medial clusters, (VOW-DEL), only applied to open syllables, i.e. syllables without existing consonant clusters.

Vowels

	Front Unrounded	Front Rounded	Back Unrounded	Back Rounded
High	i, í i, iː	ü, ű y, yː		u, ú u, uː
High-mid	é eː	ö, ő ø, øː		o, ó o, oː
Low-mid	e ɛ			a ɔ
Low			á aː

The vowel inventory of modern Hungarian⁵².

Hungarian has 14 distinct vowel phonemes, 7 short vowels plus their long counterparts. Two of the short-long pairs also differ in vowel quality: The short version of /eː/ is slightly lowered and pronounced as /ɛ/, while the short version of /aː/ is rounded and slightly raised /ɔ/. There are no diphthongs⁵³ (they have all been transformed into single, usually long vowels via VOW-LEN).

Like Proto-Uralic, Hungarian has front-back vowel harmony⁵⁴. Remember: This means that the vowels in a word (i.e. a stem plus affixes) need to be either all front or all back. With respect to vowel harmony, the Hungarian back vowels are /ɔ, aː, o, oː, u, uː/, while the front vowels are /ø, øː, y, yː/. The other front vowels /ɛ, eː, i, iː/ are considered neutral, since they can occur both in front and back stems. Still, a stem containing only neutral vowels is usually a front stem, i.e. takes front affixes. However, there are a few stems with only neutral vowels (/i, iː/, sometimes /eː/) that take back suffixes⁵⁵. In case of /i, iː/, that’s because they are usually reflexes of Proto-Uralic back /ɯ/ via (VOW-SHIFT)⁵⁶.

Some Hungarian suffixes additionally display rounding harmony⁵⁷. Rounding harmony (in Hungarian) does not affect the whole word, but only demands that the vowel in the suffix agrees in rounding with the vowel in the previous syllable.

Vowel frequency

	Grimes – total	Grimes – % of C	Kálmán – % total	Kálmán – % of C
e	24021	22.55	10.93	25.88
a	18204	17.09	9.90	23.44
o	11721	11.00	4.19	9.92
á	11595	10.88	3.56	8.43
i	9040	8.49	4.83	11.44
é	9005	8.45	3.55	8.41
í	5090	4.78	0.47	1.11
ö	3766	3.54	1.14	2.70
ó	3700	3.47	1.12	2.65
u	2972	2.79	0.85	2.01
ő	2807	2.64	0.75	1.78
ú	1996	1.87	0.45	1.07
ü	1326	1.24	0.42	0.99
ű	1280	1.20	0.07	0.17
SUM	106523		42.23

Hungarian vowel frequencies according to Grimes (2009) and Kálmán (1972).

As with the consonants, some of the vowels are very frequent and some are exceedingly rare, but the distribution is much more skewed. 40 (Grimes 2009) to 49 % (Kálmán 1972) of all vowel occurrences are just two phonemes: Short e /ɛ/ and a /ɔ/. Another 38 to 39 % of all vowel occurrences are short o /o/ and i /i/, and the long version of the two super common vowels, á /aː/ and é /eː/. The remaining 8 vowels only account for 12 (Kálmán 1972) to 22 % (Grimes 2009) of all vowel occurrences.

The high vowels (with the exception of short /i/) are particularly rare, and, as Kálmán (1972) notes: “The disappearance of short–long vowel opposition among the high vowels would result in no more than about a dozen homonyms” (p. 69). The high-mid front rounded pair ö /ø/, ő /øː/ is also rather uncommon, but not as rare as the high vowels.

Why is that so? When looking back at Proto-Uralic, remember that only a reduced set of vowels occurred in non-initial syllables, namely /æ~ɛ/, /ɑ/, /i/ and /ɯ/. In Hungarian, these have become mostly /ɛ/ and /ɔ/, less frequently /i/, and even less frequently /u/ (VOW-SHIFT). Also, /ɛ/ is the merger of the two proto vowels /e/ and /æ/ (VOW-SHIFT). It is not surprising then, that /ɛ/ and /ɔ/ are the most frequent sounds nowadays. Due to their high frequency, their long counterparts /eː/ and /aː/ are also the most frequent long vowels, having evolved from /ɛ/ and /ɔ/ via vowel and consonant deletions (VOW-LEN).

The front rounded vowel /ø/, both short and long, did not exist in Proto-Uralic, but was only produced by a limited number of sound changes (VOW-SHIFT, VOW-LEN), which explains why it so infrequent. It often evolved from /y/ (VOW-SHIFT), whose existence in Proto-Uralic is also contested, which explains its rarity. All long vowels usually arose from their short counterparts in specific contexts (VOW-LEN), so we can easily see why they are always even rarer than the short vowels.

Takeaways for conlanging

Can you picture all these sophisticated linguists debating whether Proto-Uralic had a voiced dental fricative or not and whether the elusive x was a velar or laryngeal consonant or even a vowel? Whether intervocal /p/ became /w/ or /θ/ before it was deleted and when the Hungarian vowels turned long? As conlangers, we know the full truth, but by going wild enough with the sound changes, we can create a similarly mythical scenario for the speakers of our languages. Did Proto-Rykaic have aspirate stops or affricates? Were there diphthongs? Did it have an alveolar-retroflex contrast or a dental-alveolar one? Many of these questions might not have clear answers anymore once I’m done with the sound changes, and for me, that is half of the fun!

It also means that this sort of creative work is quite forgiving. Can’t decide whether you want a uvular or velar consonant? It doesn’t matter, just make it your own x! Can’t settle on a vowel inventory? If you apply some naturalistic sound changes, chances are good that it can’t be reliably reconstructed anyway. Your proto language looks weird and can hardly be pronounced? Good for you, “real” proto languages are pretty alien too.

At the same time, creating a proto language and using sound changes to derive your conlang from it is such an important step for creating a believable, naturalistic language. Many of the unique phonological patterns of Hungarian that we discussed in the previous section, from its phoneme inventory to allowed clusters and phoneme frequency, are the direct result of sound change! I’ve already noted in the post about deriving Ryka from Rykaic that the emerging patterns and restrictions are subtle, but certainly contribute a lot of flavor.

Requirements and ideas for Nyirvón

The goal of this post was to get some ideas about how to best derive Nyirvón from Proto-Rykaic, and studying the Uralic-Hungarian sound changes has definitely been very insightful. These are my key takeaways:

Consonants: While I should do my best to approach the Hungarian consonants, there are some that won’t hurt much to be left out, namely the rare post-alveolar sounds, the voiceless palatal stop /c/, or even some of the labials. The latter is particularly useful, since Proto-Rykaic has no labials. It could also be good to add some sound changes that boost or lower the frequency of certain sounds to arrive at a similar frequency distribution as in Hungarian.

Consonant clusters: Just like Proto-Uralic, Proto-Rykaic only has medial two-consonant clusters, so I just need to take care not to produce any larger clusters by deleting too many vowels.

Voicing assimilation: Nyirvón should definitely feature this as well, it’s also not too exotic.

Consonant weaking: I might just reuse the weakening sound changes of Hungarian for Nyirvón (but not for Konoic as a whole), since I feel like the frequent loss of “hard” stops in non-initial positions might contribute much to the overall sound of the language.

Vowels: Here, we have the advantage that Proto-Rykaic already features long vowels and front-back harmony. I should take care though that I don’t end up having too many of the long vowels, so maybe I should actually try to shorten them in some places. I also need to get rid of the diphthongs. In contrast to Hungarian, Proto-Rykaic has a lot of high vowels, so we definitely need to introduce several lowering sound changes as well and produce lots of short /a/ and /e/ in the process. This is actually a good opportunity to obscure Nyirvón’s relationship to Ryka and Quriil by properly stirring the Proto-Rykaic vowels.

Reduced secondary vowels: This one I found quite interesting, and while Proto-Rykaic doesn’t have it, it is something that I could introduce in Proto-Konoic, so that it becomes a shared feature of the Konoic subfamily. It would also help to distinguish Konoic from Ryka-Quriil.

Conclusion

I can’t believe I spent almost a whole week of my vacation plus another weekend on basically writing a scientific article on Uralic-Hungarian sound changes just for fun. Anyways, it was a lot of fun, and I might do it again! What can I say, I’m a nerd.

Next up is to put all of this knowledge to use and derive Nyirvón!

References

Abondolo, Daniel (1998a). “Introduction.” In: The Uralic Languages. Ed. by Daniel Abondolo. London & New York: Routledge, pp. 1–42. [Google Books Preview]
Abondolo, Daniel (1998b). “Hungarian.” In: The Uralic Languages. Ed. by Daniel Abondolo. London & New York: Routledge, pp. 428–456. [Google Books Preview]
Grimes, Stephen M. (2009). “Quantitative Investigations in Hungarian Phonotactics and Syllable Structure.” PhD thesis. Indiana University. [PDF]
Hajdú, Péter (1972). “The Origins of Hungarian.” In: The Hungarian Language. Ed. by Loránd Benkő and Samu Imre. Paris: Mouton The Hague, pp. 15–48. [Google Books Preview]
Imre, Samu (1988). “Die Geschichte der ungarischen Sprache.” In: The Uralic Languages. Description, History and Foreign Influences. Ed. by Denis Sinor. Leiden: E. J. Brill, pp. 413–447. [Google Books Preview]
Janhunen, Juha (2009). “Proto-Uralic–what, where, and when?” In: The Quasquicentennial of the Finno-Ugrian Society 258, pp. 57–78. [PDF]
Kálmán, Béla (1972). “Hungarian Historical Phonology.” In: The Hungarian Language. Ed. by Loránd Benkő and Samu Imre. Paris: Mouton The Hague, pp. 49–83. [Google Books Preview]
Sammallahti, Pekka (1988). “Historical Phonology of the Uralic Languages. With Special Reference to Samoyed, Ugric, and Permic.” In: The Uralic Languages. Description, History and Foreign Influences. Ed. by Denis Sinor. Leiden: E. J. Brill, pp. 478–554. [Google Books Preview]
Siptár, Péter and Miklós Törkenczy (2000). The Phonology of Hungarian. New York: Oxford University Press. [Google Books Preview 1, Google Books Preview 2]

Shivallon's Conlanging Space

Case Study: From Proto-Uralic to Hungarian

November 12, 2023 Shivallon Comments 0 Comment

Proto-Uralic

Consonants

Vowels

Sound changes to Hungarian

Proto-Uralic to Proto-(Finno-)Ugric

Proto-(Finno-)Ugric to 10th century Hungarian

10th century to modern day Hungarian

Resulting phonological patterns in Hungarian

Consonants

Consonant frequency

Consonant clusters

Vowels

Vowel frequency

Takeaways for conlanging

Requirements and ideas for Nyirvón

Conclusion

References

Leave a Reply Cancel reply