Clusters of Stops

A forum for discussing linguistics or just languages in general.
Post Reply
Axthieb
rupestrian
rupestrian
Posts: 10
Joined: 08 Dec 2020 22:25
Location: Austria

Clusters of Stops

Post by Axthieb »

I've wondered whether consonant clusters consisting of adjacent stops have some sort of universal linguistics behind them.
I am certain that there are restrictions in a given language, and i could almost spot a pattern, but to do that on my own I'd need to be knowledgeable about the lexicons of quite a few languages.
My quick google search resulted in nothing usable.

For the purpose of this matter I consider affricates to be sequences, meaning in an adjacent position after another stop, they form a stop cluster like /pts/.

German has has the coda cluster /pt/ in words like "Haupt" where it is a part of the word, and words like "schraubt, schnaubt, staubt, stirbt", which are verbs with a suffixed -t. The suffix is used in 3rd person sg. present tense and 2nd person pl. present tense.
There is also the cluster /kt/ which seems to appear mainly in verbs as a suffix. eg: "sagt, plagt, laugt, zeugt" There is also "Jagd, Magd" where the cluster is a part of the core word. I could also think of special examples, namely "Markt" and "hackt" where /k/ is likely aspirated in sequence, maybe even as [kʰtʰ].
I know of no such clusters in the onset of German syllables.

In English, some Latin borrowings end in /kt/ likely [kʰtʰ] eg: "fact, act, subject" and the suffix of the simple past -ed, probably [kʰt] "checked, hacked, locked, tricked"

I've come across words in different languages as well:

Mongolian: Coda /kt~gd/ "хагд, жигд, богд", possibly [ɣtʰ] I hear in the pronunciation in the audio file on bolor-toli.com of богд
Ancient Greek: Onset /pt/ "πτάσσω, πτῆσις, πτυχίς", Onset /kt/ "κτῆνος, κτίστης, κτίζω"
Georgian: Onset /ɡd/ "გდება, გდია" (these appear to be the only words available in the lexicon)
Russian: Onset /ɡd/ [ɡʲdʲ] "где" /kt/ "кто"

[Edit: I removed two unnecessary paragraphs.]

Edit: I just remembered that I need to add that coda and word ending are of course not the same, neither are onset and the beginning of a word. This is just a way to be certain that the cluster does belong in a syllable coda or onset.

I'd be happy with any further Information about formation, occurences in other languages, or known patterns accross languages.
Last edited by Axthieb on 06 Jan 2021 19:15, edited 4 times in total.
User avatar
Creyeditor
MVP
MVP
Posts: 4148
Joined: 14 Aug 2012 19:32

Re: Clusters of Stops

Post by Creyeditor »

IIRC, the generalization in German is that a coda stop cluster can only end in an alveolar stop and coda stops cannot have the same PoA. I heard an explanation based on markedness. Alveolar stops are least marked, therefore they are allowed in that position. I think a similar generalization might hold in other languages.
Creyeditor
"Thoughts are free."
Produce, Analyze, Manipulate
1 :deu: 2 :eng: 3 :idn: 4 :fra: 4 :esp:
:con: Ook & Omlűt & Nautli languages & Sperenjas
[<3] Papuan languages, Morphophonology, Lexical Semantics [<3]
User avatar
cedh
MVP
MVP
Posts: 368
Joined: 07 Sep 2011 22:25
Location: Tübingen, Germany
Contact:

Re: Clusters of Stops

Post by cedh »

Georgian famously has "harmonic clusters", which consist of a labial or coronal obstruent followed by a dorsal obstruent of the same laryngeal specification (e.g. voiced /bɡ, dɡ/, voiceless /pk, tk/, ejective /p’k’, t’k’/), and which behave like unitary segments in a number of morphophonological processes. These harmonic clusters also include clusters with fricatives though (e.g. /zɡ, sk/ and /bɣ, px, p’χ’/.

Further, Georgian onset clusters (which can consist of up to six consonants) mostly have a clear structure that can be summarized roughly as follows, with all slots optional but occurring in a fixed order:
  1. /b p p’ m/
  2. /r/
  3. /d t t’ dz ts ts’ dʒ tʃ tʃ’ z s ʒ ʃ/
  4. /ɡ k k’ ɣ x χ’/
  5. /v/
  6. /r l m n/
(The /r/ in position 2 is apparently "optional", and the /v/ in position 5 has alternatively been analyzed as labialization of the preceding consonant. It also seems that there are a few other clusters in Georgian that do not fit into the above pattern, but those are much less complex and typically consist of only two segments.)

You can find more information in Marika Butskhrikidze's dissertation (2002), especially on pages 101ff.
Salmoneus
MVP
MVP
Posts: 2189
Joined: 19 Sep 2011 19:37

Re: Clusters of Stops

Post by Salmoneus »

I'm not really sure what you're looking for here. Stop clusters are rare (compared to the total number of languages), but they're so widespread it's hard to really generalise about them.

[one weird thing in PIE is the presence of "thorn clusters": an alveolar stop followed by a velar stop. The weird thing is that although PIE had quite a few of them, they only survived in Anatolian and Tocharian: everywhere else, they metathesized, and in many branches assilibated.]


If you consider the famous cluster (/word) /xɬpʼχʷɬtʰɬpʰɬːskʷʰt͡s/ (then he had in his possession a bunchberry plant), you can see that Nuxalk does at least have kʷʰt͡s sequences. You can also see other sequences in the words /st͡sʼqʰt͡sʰtʰx/ (that's my animal fat over there!), /qʼtʰ/ (go to shore), /qʷʰtʰ/ (crooked), /skʷʰpʰ/ (saliva), and so on.
User avatar
Pabappa
greek
greek
Posts: 456
Joined: 18 Nov 2017 02:41
Contact:

Re: Clusters of Stops

Post by Pabappa »

Russian has /kt/ and /kp/ all over the place, just not in single written words ..... there is a preposition /k/~/ko/, where the vowel appears before some onsets and not others. however, simple onset /t/ and /p/ go with the unpadded form, and therefore Russian abounds in onset /kt/ and /kp/. I suspect likewise it also has lots of /gb/ and /gd/.
I'll take the theses, and you can have the thoses.
Khemehekis
runic
runic
Posts: 2585
Joined: 14 Aug 2010 09:36
Location: California über alles

Re: Clusters of Stops

Post by Khemehekis »

Also, English has lots of words with /pt/ at the end: apt, concept, opt, adopt, corrupt, script, tapped, dropped, slipped, gripped, stepped, jumped, burped, grasped, etc.
♂♥♂♀

Squirrels chase koi . . . chase squirrels

My Kankonian-English dictionary: 67,500 words and counting

31,416: The number of the conlanging beast!
Axthieb
rupestrian
rupestrian
Posts: 10
Joined: 08 Dec 2020 22:25
Location: Austria

Re: Clusters of Stops

Post by Axthieb »

Pabappa wrote: 06 Jan 2021 15:55 Russian has /kt/ and /kp/ all over the place, just not in single written words
Well admittedly my research on Russian was too quick, and needless to say I don't know Russian. Thank you for sharing this.
Khemehekis wrote: 06 Jan 2021 16:13 Also, English has lots of words with /pt/ at the end: apt, concept, opt, adopt, corrupt, script, tapped, dropped, slipped, gripped, stepped, jumped, burped, grasped, etc.
I did have the feeling that I missed something.
Salmoneus wrote: 06 Jan 2021 15:49 [one weird thing in PIE is the presence of "thorn clusters": an alveolar stop followed by a velar stop. The weird thing is that although PIE had quite a few of them, they only survived in Anatolian and Tocharian: everywhere else, they metathesized, and in many branches assilibated.]
I wonder why they're called thorn clusters.
I think this points towards stop clusters being unstable.
Salmoneus wrote: 06 Jan 2021 15:49 If you consider the famous cluster (/word) /xɬpʼχʷɬtʰɬpʰɬːskʷʰt͡s/ (then he had in his possession a bunchberry plant), you can see that Nuxalk does at least have kʷʰt͡s sequences. You can also see other sequences in the words /st͡sʼqʰt͡sʰtʰx/ (that's my animal fat over there!), /qʼtʰ/ (go to shore), /qʷʰtʰ/ (crooked), /skʷʰpʰ/ (saliva), and so on.
I'm impressed there's even t͡sʰtʰ which I wouldn't have even thought was possible. It seems that this language readily clusters consonants. This includes stops.
cedh wrote: 06 Jan 2021 14:18 Georgian famously has "harmonic clusters", which consist of a labial or coronal obstruent followed by a dorsal obstruent of the same laryngeal specification (e.g. voiced /bɡ, dɡ/, voiceless /pk, tk/, ejective /p’k’, t’k’/), and which behave like unitary segments in a number of morphophonological processes.
This means that I missed a lot. There may be some more words after all.
Salmoneus wrote: 06 Jan 2021 15:49 I'm not really sure what you're looking for here. Stop clusters are rare (compared to the total number of languages), but they're so widespread it's hard to really generalise about them.
First I guess what I already got is what I was looking for. I'd gotten the impression stop clusters were rare, about as rare as ejectives, or maybe a tad more common.
I'd thought it was possible to figure out whether for example a velar stop would be more likely to be followed by a coronal stop than a bi/labial stop.

I was also under the impression that some practical articulatory mechanics might govern or influence any part of this phenomenon, if I may call it that.

Consider the following quote from the Wikipedia article on implosives.
The vast majority of implosive consonants are voiced, so the glottis is only partially closed. Because the airflow required for voicing reduces the vacuum being created in the mouth, implosives are easiest to make with a large oral cavity.[citation needed]

Bilabial [ɓ] is thus the easiest implosive to pronounce and so it is the most common one around the world. Velar [ɠ], on the other hand, is quite rare (uvular [ʛ] is even rarer).[citation needed] That is the opposite pattern to the ejective consonant in which it is the velar articulation that is most common and the bilabial articulation that is rare.
https://en.m.wikipedia.org/wiki/Implosive_consonant
I noticed just now the looming reminders demanding citations.
This is in part of where I got the idea from.

Ultimately I just wanted to know more about stop clusters. I still wonder if there's more to know

A thank you for everyone who has posted thus far.
User avatar
Sequor
sinic
sinic
Posts: 308
Joined: 30 Jun 2012 06:13

Re: Clusters of Stops

Post by Sequor »

Axthieb wrote: 06 Jan 2021 19:11I think this points towards stop clusters being unstable.
Sometimes they can last a long time though. English and Spanish still have the [st] of PIE *h₂stḗr 'star' in star and estrella, and that'd be what, maybe 5000 years? And similarly, the cluster [nt] of PIE *h₁énteros 'inside' is still there in Spanish entrar 'to enter', Armenian ընդերք /ənˈdɛɾkʰ/ 'entrails', modern Greek έντερο /ˈendero/ 'intestine' (colloquially also άντερο /ˈandero/).
I'm impressed there's even t͡sʰtʰ which I wouldn't have even thought was possible. It seems that this language readily clusters consonants. This includes stops.
13th century Old Spanish (when Old Spanish starts being written in large amounts, and in spellings close to pronunciations) had /tst/ and /dzd/, and not just had them around but favoured them, as /ts dz/ were the preferred outcome for the coda formed after the "second syncope" that created clusters from dzVC, dVC and ðVC, with the syncope deleting the vowel. The sound change in fact was still undergoing in that century.

Latin placitum [ˈplakɪtũː ˈplakɪtʊm] > *[ˈpladzedo] > Old Spanish plazdo [ˈpladzdo] (soon > plazo [ˈpladzo])
Latin *port-āticum > *[poɾˈtadego] > OSp portadgo [poɾˈtadgo] > portazgo [poɾˈtadzgo]
Latin jūdicāre [juːdɪˈkaːrɛ] > *[dʒuðeˈgaɾe] > OSp judgar [ʒuðˈgaɾ] > juzgar [ʒudzˈgaɾ]

You may find it interesting that from Latin to Old Spanish, clusters of three consonants formed in the second syncope where the second and third were stops tended to simplify by 1) devoicing the third consonant if the second is not voiced and 2) deleting that second consonant. So you get:

Latin masticāre [mastɪˈkaːrɛ] > *[masteˈgaɾe] > st'g > OSp mascar [masˈkaɾ]
 (g devoiced to k because preceding t)
Latin panticem [ˈpantɪkẽː ˈpantɪkɛm] > *[ˈpantedze] > nt'dz > OSp pança [ˈpantsa] (with -a)
 (dz devoiced to ts because preceding t)
Latin vindicāre [wɪndɪˈkaːrɛ] > *[vendeˈgaɾe] > nd'g > OSp vengar [venˈgaɾ]
 (g still g because preceding d)
hīc sunt linguificēs. hēr bēoþ tungemakeras.
Axthieb
rupestrian
rupestrian
Posts: 10
Joined: 08 Dec 2020 22:25
Location: Austria

Re: Clusters of Stops

Post by Axthieb »

Sequor wrote: 07 Jan 2021 18:16 Sometimes they can last a long time though. English and Spanish still have the [st] of PIE *h₂stḗr 'star' in star and estrella, and that'd be what, maybe 5000 years? And similarly, the cluster [nt] of PIE *h₁énteros 'inside' is still there in Spanish entrar 'to enter', Armenian ընդերք /ənˈdɛɾkʰ/ 'entrails', modern Greek έντερο /ˈendero/ 'intestine' (colloquially also άντερο /ˈandero/).
It's always interesting to read about PIE relationships that still have something in common.
I don't see any adjacent /t/ /p/ /k/ or other stops, that cluster in the onset or coda of a given syllable.
Sequor wrote: 07 Jan 2021 18:16 Latin placitum [ˈplakɪtũː ˈplakɪtʊm] > *[ˈpladzedo] > Old Spanish plazdo [ˈpladzdo] (soon > plazo [ˈpladzo])
Latin *port-āticum > *[poɾˈtadego] > OSp portadgo [poɾˈtadgo] > portazgo [poɾˈtadzgo]
Latin jūdicāre [juːdɪˈkaːrɛ] > *[dʒuðeˈgaɾe] > OSp judgar [ʒuðˈgaɾ] > juzgar [ʒudzˈgaɾ]
I think the noteworthy /tst/ clusters are not in the same syllable here. [ˈpladzdo] seems like [ˈpladz.do], which is essential, because that seems very pronounceable to me.

The word "act" [ækt] for example contains a stop cluster in the coda of this single syllable word. Just changing the word to "acting" makes the consonant cluster split to [ˈæk.tɪŋ].
User avatar
eldin raigmore
korean
korean
Posts: 5738
Joined: 14 Aug 2010 19:38
Location: SouthEast Michigan

Re: Clusters of Stops

Post by eldin raigmore »

Is it true that a cluster of coda stops will have the further-back PoA before the further-front PoA unless there’s a following vowel?
And a cluster of onset stops will have the further-front PoA before the further-back PoA unless there’s a preceding vowel?

If either of those is not an absolute universal, is it at least a statistical universal?
User avatar
Dormouse559
moderator
moderator
Posts: 2759
Joined: 10 Nov 2012 20:52
Location: California

Re: Clusters of Stops

Post by Dormouse559 »

Axthieb wrote: 08 Jan 2021 22:47I think the noteworthy /tst/ clusters are not in the same syllable here. [ˈpladzdo] seems like [ˈpladz.do], which is essential, because that seems very pronounceable to me.
English does put /tst/ in a single syllable, with verbs that end in /ts/. Take "spritzed" and "blitzed" for example.
Khemehekis
runic
runic
Posts: 2585
Joined: 14 Aug 2010 09:36
Location: California über alles

Re: Clusters of Stops

Post by Khemehekis »

Dormouse559 wrote: 09 Jan 2021 02:51 English does put /tst/ in a single syllable, with verbs that end in /ts/. Take "spritzed" and "blitzed" for example.
"Waltzed" is another good example. Or "schmaltzed", as in "schmaltzed up". Mostly the past tenses and participles of verbs borrowed from German or Yiddish.
♂♥♂♀

Squirrels chase koi . . . chase squirrels

My Kankonian-English dictionary: 67,500 words and counting

31,416: The number of the conlanging beast!
User avatar
Pabappa
greek
greek
Posts: 456
Joined: 18 Nov 2017 02:41
Contact:

Re: Clusters of Stops

Post by Pabappa »

eldin raigmore wrote: 08 Jan 2021 23:48 Is it true that a cluster of coda stops will have the further-back PoA before the further-front PoA unless there’s a following vowel?
And a cluster of onset stops will have the further-front PoA before the further-back PoA unless there’s a preceding vowel?

If either of those is not an absolute universal, is it at least a statistical universal?
i dont think its about frontness. easy counterexamples, such as many languages having /kt- gd-/ but not /tk- dg-/. And likewise having /-pt/ but not /-tp/. If anything i'd say that the tendency is for peripheral consonants (labials and dorsals) to occur first regardless of whether it's a coda or onset cluster.

To some extent you see this with other sounds, .... I'd bet /mn/ is more common than /nm/ cross-linguistically, for example.
I'll take the theses, and you can have the thoses.
User avatar
Creyeditor
MVP
MVP
Posts: 4148
Joined: 14 Aug 2012 19:32

Re: Clusters of Stops

Post by Creyeditor »

Pabappa wrote: 09 Jan 2021 12:36
eldin raigmore wrote: 08 Jan 2021 23:48 Is it true that a cluster of coda stops will have the further-back PoA before the further-front PoA unless there’s a following vowel?
And a cluster of onset stops will have the further-front PoA before the further-back PoA unless there’s a preceding vowel?

If either of those is not an absolute universal, is it at least a statistical universal?
[...] If anything i'd say that the tendency is for peripheral consonants (labials and dorsals) to occur first regardless of whether it's a coda or onset cluster.
See also my post above. Not sure about onset clusters, but I'd expect the mirror image.
Creyeditor
"Thoughts are free."
Produce, Analyze, Manipulate
1 :deu: 2 :eng: 3 :idn: 4 :fra: 4 :esp:
:con: Ook & Omlűt & Nautli languages & Sperenjas
[<3] Papuan languages, Morphophonology, Lexical Semantics [<3]
User avatar
eldin raigmore
korean
korean
Posts: 5738
Joined: 14 Aug 2010 19:38
Location: SouthEast Michigan

Re: Clusters of Stops

Post by eldin raigmore »

Pabappa wrote: 09 Jan 2021 12:36 i dont think its about frontness. easy counterexamples, such as many languages having /kt- gd-/ but not /tk- dg-/. And likewise having /-pt/ but not /-tp/. If anything i'd say that the tendency is for peripheral consonants (labials and dorsals) to occur first regardless of whether it's a coda or onset cluster.
To some extent you see this with other sounds, .... I'd bet /mn/ is more common than /nm/ cross-linguistically, for example.
That does not match my experience; but my experience has only to do with English, and loan-words into English.
Maybe you’re right.
My “sample” is not really cross-linguistic.

Creyeditor wrote: 09 Jan 2021 14:12 See also my post above. Not sure about onset clusters, but I'd expect the mirror image.
Thanks!
Post Reply