#language classification

where substantial shared material exists, one should default to a relationship and not propose mass loaning / creolization / "mixing" without further arguments

correspondence sets without proper reconstruction can be almost as important for tracing external relationships as fully reconstructed proto-forms

not every language has numerals and thus not every relationship involves them either

This paper [Vovin, Alexander. "Towards a new classification of Tungusic languages." Eurasian Studies Yearbook 65 (1993): 99-113] is an attempt at a lexicostatistical classification of Tungusic. Based on percentages of cognates in 100 word Swadesh lists, Vovin proposes the following classification:

A strikingly unusual feature of this tree is the position of Even as the second language to branch off after Manchu. Usually, Even is placed in the same branch with Evenki, Solon, Oroqen and Negidal. 

Indeed, Even mostly shows lower cognate percentages with other Tungusic varieties than any other language save Manchu:

It would be interesting to check the accuracy of Vovin’s wordlists, but I will not try to do it here. Instead, we can look at some of those items of the Swadesh lists where Even differs from other languages. These are the very items which cause lower cognate percentages.

‘blood‘: Even huneel [actually huŋeel] vs. reflexes of *sääksä in all other languages (Manchu senggi contains a different suffix, but has the same root)

‘bone‘: Even ikeri with a cognate in Negidal vs. reflexes of *gïramsa in all other languages (Manchu giranngi contains a different suffix, but has the same root)

‘die‘: Even køke= vs. reflexes of *bö(d)- in all other languages, including Manchu

‘drink‘: Even kool= vs. reflexes of *umï- in all other languages, including Manchu

‘ear‘: Even korat with an apparent cognate in Udihe vs. reflexes of *sian in all other languages, including Manchu

‘full‘: Even milteree vs. reflexes of *ǯalo- in all other languages save Udihe

‘sleep‘: Even huklee= vs. reflexes of *au- in all other languages, including Manchu

‘sun‘: Even nøølten vs. reflexes of *sigöön in all other languages, including Manchu (Evenki, Solon & Negidal also reflect another root: *dïlačaa)

This is not a pattern we would expect from a second outlier in the family. One could rather expect that Even would share some archaisms with Manchu as against innovations in the common ancestor of all remaining languages. There are no such cases in the Swadesh list. Instead, we see Manchu sharing common words with the rest of the family except Even. Given that Manchu is the first outlier in Vovin’s classification (as well as in most other classifications), words shared by Manchu with other languages are by definition archaisms, while isolated Even words must be innovations.

A small collection of innovations across Cushitic that don't seem to establish any coarser subclassification: broad phonological changes, plus one of the more prominent grammatical divisions — loss vs. preservation of the prefix conjugation, thought to be already old Afrasian inheritance. This already suffices to distinguish every unambiguous group from each other, even though none of the features are unique to some single group! Hopefully this will illustrate the relative heterogeneity of South and especially East Cushitic.

Loss of pharyngeals and ?introduction of labialized consonants indeed continue to Ethio-Semitic, as shown. (Loss of lateral obstruents, again supposedly Afrasian inheritance, however is not quite areally shared between the two: Cushitic has mainly *ɬ > l, Ethio-Semitic *ɬ > s.) To reiterate, it would be also nice to compare further with North Omotic eventually, it already features prominently in any papers on the Ethiopian Sprachbund; but I've not seen thorough enough comparative coverage to assemble anything analogous to this.

The arrangement of the Cushitic subgroups is roughly geographic, in terms of what's adjacent to what. This ends up kind of showing also the large range of Oromo and Somali, not in a way proportional to the real geography though, e.g. HEC + Konsoid + Dullay + Arboroid are actually all cramped in an area smaller than that of just Beja or just Afar, and Agaw is really multiple language islands mainly within the range of Ethio-Semitic, instead of forming a neat cluster between Oromo and Beja. If I had put Ethio-Semitic in here geographically too, all of it would go in the same area where Agaw is now: between Beja, HEC, Oromo and SA.

As per some previous brainstorming on Twitter and initial further reading up: is Polynesian really just one unremarkable subgroup of the Oceanic languages, or should it be maybe considered a distinct entity from the diverse Melanesian / Micronesian groups?

One of the more unusual features in reconstructed Proto-Oceanic as it stands today are the "labialized labial" consonants, *mʷ *mbʷ *pʷ — reflected fairly widely and yielding in daughters variously just that (e.g. Tawala: /mʷ bʷ pʷ/), or velarized labials (e.g. Marshallese: /mˠ pˠ/) or labial-velars (e.g. Lewo: /ŋm kp/), or labialized velars (e.g. Kwaio: /ŋʷ gʷ kʷ xʷ), or also various other less distinct stuff. Where these have common Malayo-Polynesian etymologies (though most don't), they seem to just arise from plain *p/b, *mp/mb, *m without any single conditioning factor. This, among other things, gives many Oceanic languages relatively rich consonant inventories by Austronesian standards.

But in Polynesian there's again no systemic sign of these. Supposedly they just merge back into the labials! (I gather same also in a variety of other Oceanic languages, but the literature I have checked doesn't bother explicitly listing all known cases.) Some cases still also with *m > *mʷ > Proto-Polynesian *ŋ, but my contact linguist senses suggest that this might be loans from some *ŋʷ or *ŋm type language.

So why should we think the labialized labials ever existed in Polynesian at all? It seems to me that the rise of this kind of a highly distinctive new phonological category is unlikely to just immediately disappear again in a single step. The later Polynesian languages do show a trend towards inventory simplification, but very step-by-step. Intermediate stages towards something highly minimal like Hawai'ian can be indeed still attested in the other languages. This is not the case with the loss of the *Pʷ series as far as I can tell. And it's also not the case that Polynesian would be known to group deep into a tree of Oceanic languages, it's usually put as basically direct daughter of a primary branch (Central Pacific) together with Rotuman and Fijian. Neither of which seems to have generally kept the *Pʷ series either. (Fijian has instead unrelated labialized velars, seemingly from POc. plain velars again without much of an explanation. Ross 2011 proposes also a distinct *kʷ already into POc…)

So perhaps there instead existed a "Proto-Near-Oceanic" or the like, which was the real node innovating the *Pʷ series, and which later spread partly further across the original spread of Oceanic — leaving Polynesian now kind of insulated in the east (mostly: there are still the Polynesian outlier languages which have come back west into Micronesia and Melanesia). Maybe there could be even a series of nodes adding the labiolabials in gradually. Some languages like Western Fijian are claimed to not distinguish e.g. *p = *pʷ, but do distinguish e.g. *mʷ | *m (as ŋʷ | m). I don't know the archeology of the original spread of Oceanic in a ton of detail, but it would make sense that skilled seafarers might have formed their first substantial Far Oceanian settlements (i.e. not just trading outposts or the like) on newly found islands entirely like, say, Vanuatu, versus require at least some time to properly claim ground in places like New Britain and Bougainville that already had long-established "Papuan" populations.

(incidentally why do we call these "Papuan", when many of the families have no presence on the island of New Guinea itself? shouldn't something like "Paleo-Oceanic" be more appropriate for this gaggle of non-Austronesian languages?)

All available grammatical information should be systematically examined, but vocabulary leads most swiftly to the correct hypotheses as a general rule. The effectiveness of mass comparison of basic vocabulary, for all its apparent simplicity, is illustrated in Table 3 by only a few forms from all the contemporary languages of Europe. Note that, even by the time the second word has been examined, the correct hypothesis emerges.

Not impressed by this assertion. Sure the comparison of just '1' and '2' will easily suggest some language groups like Germanic, Baltic, Slavic, Finnic, and leave Basque so far isolated. But Hungarian egy (also mistranscribed here as /ed/ and not /ɛɟ/ or even /edʲ/) will not be readily identified as (and probably isn't) cognate to Finnic *üksi. One might rather end up thinking it is a reduced reflex of Slavic #jadin (actually *edinъ but we will not exactly be doing any real reconstruction based on just this data). Két also shows no highly compelling similarity to *kaksi; maybe would do so more if G. had bothered to inform that the inflected stem is kahte-. So just by comparing the first two numerals' citation forms, the most sound conclusion might be that Hungarian is either a divergent Slavic language, or an isolate that has borrowed '1' from Slavic!

Do we really retrieve correct sorting of the other IE languages from just these either, if we pay no heed to G. having already sorted the data according to what we should think? I see no consistent way to divide Celtic from Romance, rather e.g. Irish and French seem to come close to each other. And nothing whatsoever even suggests Balto-Slavic. Instead, I think, we'd end up with a basic division Slavic / Rest by different words for '1', which then suggests that the cluster *tv- in Germanic is archaic and that groups with only *d- are a single unit; and Greek, of course, subgroups with Baltic by virtue of suffixal -as in '1' and evidence for a close front element in '2'. Maybe it would remain there even after considering the rest of thsi data, thanks to e.g. #afci 'ear', #(m-)aki 'eye', and the i-suffix in #dand-i 'tooth'!

The standard view I've usually seen is that *s > initial /h/, mediofinal /θ/. Apparently this comes with quite a bit of positional variation though, including also several island dialects (light blue, khaki, grey) that have a preceding assimilation *s-ʃ > /ʃ-ʃ/ — seemingly proving that *s > θ is late areal drift. This also claims the "h" reflexes to vary between pharyngeal, uvular, velar, and even epiglottal (I wonder if this is somehow over-transcribed, when plain [h] makes no appearence).

An interesting further feature is the small eastern area with *ps > ? *ph > /ɸː/ and *ks > ? *kh > /xː/ (supposed to be two dialect areas really but I can't find the darker purplish brown anywhere on the map). This has a close typological parallel in Vasjugan Selkup where *s > /h/ similarly goes with *ps > ? *ph > /f/.