A guide to small consonant inventories

[----]

Tolomako has either 10 or 11 consonants: /m n p t k (s) ts β ɣ r l/. Wikipedia lists its phonemic inventory without /s/ but it also has a sample written with a single <s> in it. However, it appears before an <i> so I dunno if it's allophonic or what

[Nortaneous]

One of the linguists who did fieldwork on it says it has /s/.

[gach]

Does anyone know of attainable references for Australian languages analysed as having only a voices stop series?

The picture I've got is that a good starting point for the analysis of the stop systems of Australian languages lacking a MOA contrast for their stops is that these are underlyingly obstruents just defined for their place of articulation. Depending from the language the stops will then be allophonically unvoiced, voiced or fricativised to varying extents.

Wei Lo (2010) (ling.yale.edu/sites/default/files/files/alumni%20senior%20essays/Justin%20Wei%20Lo.pdf) gives the following basis for an alternative analysis of languages lacking a stop series contrast when we assume the language to have an underlying stop mode of articulation:

In languages with most stops phonetically voiced, as in Yidiny, we would likely draw the the conclusion that in such a language the allophones [p, b, β] would all present /b/ instead of /p/. Each language in Australia would be able to choose their normative articulation in this way.

Dixon's Australian Languages also gives the following short and general statement for correlations between the stop POA and voicing:

There is in some languages a strong tendency for the apico-postalveolar stop to have a voiced articulation, while the other stops are generally voiceless. See, for example, Hercus (1994: 37) on Arabana/Wangkangurru: 'intervocalic ploisives are unvoiced except for the retroflex rd, which is always voiced'. For Yinjtjiparnrti Wordick (1982: 11) reports that all consonants except for rt (and sometimes rr) 'are pronounced as if they were geminate (doubled) in medial position between vowels'.

Not all stops can be expected to follow the same analyses.

Wikipedia lists aɾiakɾe for k-less men's speech, so either that's an error or there are still voiceless stops there. It may be that some /tʃ/ are retained too. Anyone have a PDF of The Amazonian Languages?

I don't have a pdf but do have a paper copy. Here's the main paragraph describing the men's and women's speeches in Karajá (emphasis mine):

Another phonological peculiarity of Karajá is the systematic differentiation of the phonological shape of words between men's and women's speech. Men's speech regularly lacks the velar stops present in the speech of women, as well as the instances of the voiceless alveo-palatal affricate that are historically derived from velar and stops palatalized by a preceding i. As a consequence of the dropping of the velar and alveo-palatal consonants, several vowel contractions may result, thus making the shape of the words and sentences uttered by men more distinct from those pronounced by the women. Even borrowings from Portuguese are subject to the dropping of the velar stop.

Here "regularly" has to be understood as "typically but not always". The table related to this discussion includes not only men's aɾiakɾe for women's kaɾitʃakɾe but also men's aɾihoɔɗekɾe for women's kaɾihokɔɗekɾe. There's no question that also the men's speech has /k/ as a phoneme, it's just much rarer than in women's speech.

[Pogostick Man]

Does anyone know of attainable references for Australian languages analysed as having only a voices stop series?

I don't know if this is what you're looking for, but this is claimed for Gamilaraay at Wikipedia.

[gach]

I don't know if this is what you're looking for, but this is claimed for Gamilaraay at Wikipedia.

To me that just has the basic condition of quick descriptions of phoneme inventories for the Australian languages; there's no indication whether the stops should be analysed with the underlying feature [+voiced] or if the voiced stops are just a graphical convention.

It's certainly true that a more or less constant voicing is a common Australian unmarked background feature. What I would want to see are detailed analyses arguing whether this should be transferred to the stop phonemes or left at the phonetic level.

You can find a similar conceptual problem in the phonemic analysis of Abau (also included in the list of this thread for nine consonants). The language is listed as having /r/ but not /t/. However, the phoneme system of the language is really much less remarkable than this since the /r/ comes from lenition of earlier /t/ and still has both [t] and [r] as common allophones. /r/ is just picked to represent the underlying phoneme since [r] appears in more environments than [t]. Really, a more intuitive description of the phoneme would be something like /t ~ r/, leaving the mode of articulation purely onto the level of allophony.

[gach]

Since I talked about Skou elsewhere, let's add Barupu here as it's missing from the list: /p t k b r m n j w/, 9 consonant phonemes in total.

It has /b/ as its only voiced stop but this isn't really remarkable at all as /r/ groups as the voiced variant of /t/.

Another one from the same family is the Waromo dialect of Vanimo. It appears to have /p t ʔ b d m n ɲ s ɺ j w/ (12 in total), though the SIL sheet is very vague about the dialectal distribution of the phonemes.

[Pogostick Man]

So I've redone the frequency analysis. Note that the duplicate analyses were not separated. These categories are possibly not strictly speaking correct with how the phonemes were analyzed (at least w.r.t. the fricatives and resonants). Also, beware of underspecified rhotics.

Plosives
/p/ - 153
/pˠ/ - 1
/b/ - 63
/ɓ/ - 1
/t/ - 174
/tʰ/ - 2
/d/ - 54
/ɗ/ - 2
/c/ - 4
/ɟ/ - 6
/k/ - 173
/kʰ/ - 2
/kʷ/ - 16
/g/ - 47
/gʷ/ - 1
/q/ - 2
/ʔ/ - 80
/ʔː/ - 1

Affricates
/tʙ/ - 1
/ts/ - 15
/dz/ - 1
/tʃ/ - 38
/dʒ/ - 11
/gʟ/ - 1

Nasals
/m/ - 145
/mˠ/ - 1
/n/ - 144
/ɲ/ - 12
/ŋ/ - 46

Fricatives
/ɸ/ - 8
/β/ - 18
/βˠ/ - 1
/f/ - 25
/v/ - 15
/θ/ - 4
/ð/ - 3
/s/ - 110
/z/ - 5
/ɮ/ - 1
/ʃ/ - 16
/ʒ/ - 1
/ç/ - 1
/ʝ/ - 3
/x/ - 8
/ɣ/ - 4
/ʁ/ - 1
/h/ - 9

Resonants
/ʍ/ - 1
/w/ - 122
/ʙ/ - 1
/ʋ/ - 2
/l/ - 42
/r/ - 110
/ɺ/ - 1
/ɾ/ - 1
/ɭ/ - 2
/ɽ/ - 1
/ɰ/ - 2
/j/ - 105

[Nortaneous]

Underspecifying the rhotics was a deliberate choice on my part, since I didn't feel like copying and pasting the tap character.

[Pogostick Man]

Would it be more accurate to say that 113 langs had rhotics, or that (113 - x) langs had rhotics and, of these, x langs had more than one?

[Tropylium]

I'd say the former. Maybe spell out explicitly that (113-x) only had one.

[vec]

This thread is such an amazing resource.

[----]

Found a new one!

Nakanai: /p t k b d g m β s h l r/
This comes from SIL, who analyze the glides as being allophones of the high vowels. I'd believe it.

EDIT: And another!
Inanwatan has /p t k ʔ b d g f s m n/.

[Birdlang]

Hawaiian has /p k ? m n h w~v/. There is a t gap from a sound change.

[----]

It's already been counted.

[Birdlang]

It's already been counted.

OK. I didn't check the whole thing

[----]

You didn't even check the first post of the thread?

[Pabappa]

Old Basque is often analysed with 9 consonants : /p t l r k s z n h/ I believe. A counterexample to the /p n/ implies /m/ rule. The modern reflexes of all stops are voiced unless doubled, so one could propose they were voiced in Old Basque as well. (The Trask Basque dictionary does this, and even leaves out /h/, setting it to 8 consonants, but says /h/ existed as a suprasegmental feature.)

[Nortaneous]

Old Basque is often analysed with 9 consonants : /p t l r k s z n h/ I believe. A counterexample to the /p n/ implies /m/ rule. The modern reflexes of all stops are voiced unless doubled, so one could propose they were voiced in Old Basque as well. (The Trask Basque dictionary does this, and even leaves out /h/, setting it to 8 consonants, but says /h/ existed as a suprasegmental feature.)

Didn't Old Basque have /b/ but not /p/? (/p/ but not /p:/, w/e)

[Nortaneous]

Update. Haven't added anywhere near all the eligible languages in PHOIBLE but it's a start.

[Birdlang]

Abau has /m n p k s h ɾ j w/ 8 consonants
Ache has /m p mb t̪ n̪d̪ k ŋɡ t̠ʃ d̠ʒ β ɺ̪/ 10 consonants
Andoque has /ɲ p b t d ɟ k ɸ s h ɾ/ 11 consonants
Bariai has /m n ŋ p b t d k g s r/ 11 consonants
Bandjalang has /m n n̠ ŋ b d d̠ʒ g l r j w/ 12 consonants
Cherokee has /m n d dz g ʔ ʃ h l j w/ 11 consonants
Comanche has /m n p t̪ t̪s̪ k kʷ ʔ s h j w/ 12 consonants and phonemic voiceless vowels.
from PHOIBLE! I copied the word PHOIBLE and I am amazed!!!

[Herr Dunkel]

Most of those are already on the list. Please check the OP at least before you decide to post something

[Nortaneous]

There's only one that isn't on the list, and there's no point digging through PHOIBLE because you can sort by consonants. I'm going to add everything in there eventually.

[Birdlang]

There's only one that isn't on the list, and there's no point digging through PHOIBLE because you can sort by consonants. I'm going to add everything in there eventually.

Thanks Nortaneous for pointing it out. I would love to see what interesting languages come up. Which one is not on the list? Maybe Bandjalang?

[Nortaneous]

bandjalang has been on the list for months, do you not know how to use ctrl+f

[Birdlang]

bandjalang has been on the list for months, do you not know how to use ctrl+f

I am using an IPad.