Initial clusters reducing to single consonants

[LoneWolf]

I have only heard of a few languages where historically speaking virtually all of the original initial consonant clusters were reduced to single consonant onsets. In fact the only two examples I know are Korean and Lhasa Tibetan. However in the case of Lhasa Tibetan I have little to no idea what were the actual historical developments that lead to such a massive simplification of initial clusters. Does anyone know what were the sound changes that lead to this reduction of initial clusters? Also are there other languages where this has happened and what sound changes were involved in the reduction process?

The reason I'm asking is because I'm making a Proto-Conlang which allows initial consonant clusters. On the other hand I want most of the daughter languages to have simplified these initial clusters to single consonant onsets in all roots. So I'm looking for inspiration on how this might happen in a natural language.

[Zhen Lin]

Some of the other Sino-Tibetan languages are believed to have underwent the same initial cluster simplification, including e.g. Chinese.

[Nortaneous]

One of these days I'll write a bot that automentions Tsakonian in any post about cluster simplification.

Although for onset clusters, the thing about Tibetan is that most, maybe all, initial clusters were actually preinitial + single-consonant onset (although in some cases, two preinitials were possible), so the "cluster simplification" was mostly just preinitials getting nuked.

[Grimalkin]

Yep, Sino-Tibetan is a great language family for tracking all the different ways in which consonant clusters simplify. In many modern ST languages, preinitial consonants have disappeared, but not before affecting the tone of the following vowel. If I remember correctly, pre-initial s- generally put the vowel in a high tone, and I think an initial glottal stop tended to produce a low or falling tone before dropping out.

Another common feature of consonant cluster simplification is coalescence, when two segments merge into a single segment that has features of both the original sounds. /kw/ > /p/ is a very common sound change. /dw/ > /b/ is attested from Latin (not a Sino-Tibetan language obv, but I wouldn't be surprised if it was a common sound change). Initial /s/ + nasal clusters might coalesce into voiceless nasals, which tend to be unstable, so they probably wouldn't last long.

Also, epenthetic vowels are your friend. A hypothetical Proto-lang root *stak might become satak in one daughter lang, estak in another, etc.

Hope this all helps, this post was a bit rushed

[spats]

Spanish has lost a significant number of its initial clusters compared to Latin, either through simplification or epenthic vowels.

[Torco]

epenthetic vowels and assimilation sound like would be common for initial clusters.

[alice]

You could also look at loanwords in Finnish, such as ranta from the source of English "strand".

[Magb]

You could also look at loanwords in Finnish, such as ranta from the source of English "strand".

Finnish never allowed initial clusters to begin with though.

[Grimalkin]

Nonetheless it's a good example of how initial consonant clusters can be simplified. I'd hazard a guess that the most sonorant consonant in a cluster tends to be the most resistant to dropping. To give another example, in Welsh, initial st simplified to s. Compare the words for 'star' in the three Brythonic languages - steredenn in Breton, steren in Cornish, but seren in Welsh.

On second thoughts, I'm probably wrong about the sonority thing. In many Sino-Tibetan languages, medial liquids have dropped out of complex onsets, leaving obstruents intact. Hmmm

[MisterBernie]

On second thoughts, I'm probably wrong about the sonority thing. In many Sino-Tibetan languages, medial liquids have dropped out of complex onsets, leaving obstruents intact. Hmmm

In Finnish loaning, as far as I know, it used to be that simply all but the last consonant in a cluster got dropped. For example, skola -> koulu, stund -> tunti.

[Qwynegold]

I have only heard of a few languages where historically speaking virtually all of the original initial consonant clusters were reduced to single consonant onsets. In fact the only two examples I know are Korean and Lhasa Tibetan.

So what happened to Korean?

[Ser]

If I remember correctly, pre-initial s- generally put the vowel in a high tone, and I think an initial glottal stop tended to produce a low or falling tone before dropping out.

Those are the effects of post-codas -s and -ʔ actually: -s produced a falling (or perhaps low) tone (去聲 qù shēng 'departing/leaving tone'), and ʔ a rising (or perhaps high) tone (上聲 shǎng shēng 'rising/top tone'). Old Chinese 數 *skrjo(k)s 'number' > Middle Chinese ʂju^去 > Mandarin shù, Cantonese [sou˧]; OC 數 *skrjokʔ 'to count, to calculate' > MC ʂju^上 > Mandarin shǔ, Cantonese [sou˧˥]. (These two examples come from William Baxter's reconstructions in his Handbook of Old Chinese Phonology, p.228)

To make the terms clear, Baxter reconstructs the structure of Old Chinese syllables like this: (pre-initial) - initial - (medial) - vowel - (coda) - (post-coda), ((C)C(G))V(C(C)).

Baxter presents three Old Chinese pre-initials: *ɦ-, *N- (some nasal consonant) and *s-. *ɦ- disappeared while voicing the next consonant, *N- only appeared before plosives and disappeared while nasalizing them, consonants following *s- generally disappeared leaving *s- alone except if they were coronal plosives (then they became affricates and *s- disappeared).

He also reconstructs the medials *-l- and *-r-: *-l- appeared after velar plosives (and possibly labials too) and disappeared while making them dentals, then there's also OC *hl- > MC tʰ; *-r- generally disappeared while colouring an unrounded vowel (by fronting it, sometimes lowering it) and/or making a preceding dental consonant a retroflex, or a preceding labial plosive the dental l-.

The problem with discussing initial clusters in Old Chinese is that they are full of controversies. Baxter's reconstruction while well known isn't widely accepted at all, on the basis that internal reconstruction can only take one so far. I should note that the entire thing isn't only his, but is based on various hypotheses of features that Old Chinese possibly had by looking at patterns and searching for traces. The problem is presenting them as a full reconstruction, it's quite likely that many other features didn't leave traces...

Old Chinese is "attested" in the sense that the texts are there, the problem is that figuring out its phonology is largely like reconstructing a proto- language.

[Khvaragh]

Baxter presents three Old Chinese pre-initials: *ɦ-, *N- (some nasal consonant) and *s-. *ɦ- disappeared while voicing the next consonant, *N- only appeared before plosives and disappeared while nasalizing them, consonants following *s- generally disappeared leaving *s- alone except if they were coronal plosives (then they became affricates and *s- disappeared).

Are *N- and *ɦ- by any chance cognate to Tibetan 'a-chung?

[Ser]

Are *N- and *ɦ- by any chance cognate to Tibetan 'a-chung?

*ɦ- is said to be.

[Grimalkin]

Ah, interesting. Thanks

Anyone know how good/reliable James Matisoff's handbook of Proto-Tibeto-Burman is? I has a copy and it seems quite good. It's got some information on Old Chinese too.

[Richard W]

I'm surprised no-one mentioned Thai. There is a dialect that simplifies khw- > f- and kw- > f-, thus eliminating all the clusters that are still written in the standard language. The changes, if you start from the standard language, is otherwise the pretty boring loss of liquids:

kl-, kr- > k-
khl-, khr- > kh-
pl-, pr- > p-
phl-, phr- > ph-
kw-, khw- > f-

There's a possibility of tr- > t-, but first one should establish that the dialect ever had initial /tr/.

[linguoboy]

I have only heard of a few languages where historically speaking virtually all of the original initial consonant clusters were reduced to single consonant onsets. In fact the only two examples I know are Korean and Lhasa Tibetan.

So what happened to Korean?

Clusters became tense consonants. The expected development of our hypothetical *stak would be *sttak > *ttak [where doubling represents tenseness, according to the conventions of Korean orthography; whether the tense consonants actually derive from geminates is a moot point]. Cf. MK stol "daughter" > ModK ttal (i.e. [t͈al]).

Another possibility (found in both S-T and I-E) which I don't think has been mentioned explicitly is vocalisation of liquids, e.g. Latin PLANUS > Italian piano. This can be viewed as intermediate between retention and full palatalisation (cf. PLANUS > Ligurian cian [ʧan]).

[Bristel]

I have only heard of a few languages where historically speaking virtually all of the original initial consonant clusters were reduced to single consonant onsets. In fact the only two examples I know are Korean and Lhasa Tibetan.

So what happened to Korean?

Clusters became tense consonants. The expected development of our hypothetical *stak would be *sttak > *ttak [where doubling represents tenseness, according to the conventions of Korean orthography; whether the tense consonants actually derive from geminates is a moot point]. Cf. MK stol "daughter" > ModK ttal (i.e. [t͈al]).

Another possibility (found in both S-T and I-E) which I don't think has been mentioned explicitly is vocalisation of liquids, e.g. Latin PLANUS > Italian piano. This can be viewed as intermediate between retention and full palatalisation (cf. PLANUS > Ligurian cian [ʧan]).

Is Latin PLANUS > Spanish llano, and what is that occurrence called?

[linguoboy]

Is Latin PLANUS > Spanish llano, and what is that occurrence called?

It's another instance of palatalisation, i.e. /pl/ > /pʎ/ followed by regressive assimilation of the stop. The initial development is the same as in (Standard) Italian, except here /ʎ/ delateralised to /j/. Gallaico-Portuguese shows the same further development as Ligurian (i.e. chão), but I believe there are some minor varieties in northern Iberian which preserve the /pʎ/ stage.

Some people claim these developments show that the palatalisation was a feature of Vulgar Latin, but I think that's refuted by the Catalan evidence: inherited initial /l/ goes to /ʎ/ in all cases, but /pl/ remains intact (i.e. pla).

[Ser]

but I think that's refuted by the Catalan evidence: inherited initial /l/ goes to /ʎ/ in all cases, but /pl/ remains intact (i.e. pla).

And French [pl] in plein.

Unless it could be argued that Catalan and French depalatized the /ʎ/ later?

[linguoboy]

but I think that's refuted by the Catalan evidence: inherited initial /l/ goes to /ʎ/ in all cases, but /pl/ remains intact (i.e. pla).

And French [pl] in plein.

Unless it could be argued that Catalan and French depalatized the /ʎ/ later?

That's why I think Catalan offers a stronger argument than French. Why would /ʎ/ depalatise after initial consonants but not in absolute initial position? Is there really such a salient difference between com a ploms and en cap lloc?

[Qwynegold]

I have only heard of a few languages where historically speaking virtually all of the original initial consonant clusters were reduced to single consonant onsets. In fact the only two examples I know are Korean and Lhasa Tibetan.

So what happened to Korean?

Clusters became tense consonants. The expected development of our hypothetical *stak would be *sttak > *ttak [where doubling represents tenseness, according to the conventions of Korean orthography; whether the tense consonants actually derive from geminates is a moot point]. Cf. MK stol "daughter" > ModK ttal (i.e. [t͈al]).

Another possibility (found in both S-T and I-E) which I don't think has been mentioned explicitly is vocalisation of liquids, e.g. Latin PLANUS > Italian piano. This can be viewed as intermediate between retention and full palatalisation (cf. PLANUS > Ligurian cian [ʧan]).

Oh, that's interesting! Thanks! I thought no one would reply to this.

[Richard W]

The Indic languages also simplified their initial clusters. There's information on the simplification that took place in the development of Pali (which did however retain vy-) from a Sanskrit-like dialect in http://en.wikipedia.org/wiki/Pali#Pali_and_Sanskrit .

Several Tai languages, instead of simplifying initial /kr/ > /k/, change /kr/ to the aspirated consonant /kh/ parallel to their changes of /r/ > /h/. This is attested in the Northern Thai spelling conventions, as well as being fairly clear from the reconstruction of Proto-Tai.