The Anim Languages of Southern New Guinea
John Cowan
Trond Engen
> https://www.academia.edu/25123762/The_Anim_Languages_of_Southern_New_Guinea
>
> This is a beautiful paper exhibiting a hitherto unrecognized genetic
> relationship between four small language families of southern New
> Guinea.
> It's a very nice, clear, yet completely realistic illustration of how
> the comparative method works.
establish a relationship even from quite short wordlists, since
correspondences in morphological systems are beyond chance resemblance
almost by occurence, and at the same time you have a master key to the
historical phonology of the whole group.
> I strongly recommend it.
And I gladly second that.
--
Med vennlig hilsen
Trond Engen
John Cowan
and of the power of morphology
David Marjanovic
"Franklin’s cautious stance as a classifier contrasts with the expansive drive of Wurmand Voorhoeve, who were engaged in major lumping efforts in the highlands of Papua New Guinea and in the south of Irian Jaya, respectively. If one follows the literature up to 1970, it looks as if the two lumping approaches would eventually clash in the Gulf District. This danger was averted with the invention of the Trans-New Guinea phylum in 1970, which allowed the peaceful unification of both supergroups. Franklin stood in between and successfully blocked any further expansion of the East New Guinea Highlands stock and the Central and South New Guinea stock. The Inland Gulf languages were given a neutral classification as an independent subphylum of the Trans-New Guinea phylum."
The authors do accept, however, that the Anim languages probably belong to Trans-New-Guinea ("a demonstration" being "beyond the scope of this paper", p. 138).
> As our own Marie-Lucie has said several times at the Hat, morphology is overwhelmingly the first line of attack for the comparative method. Only where it is absolutely unavailable should one fall back on merely lexical comparanda. One suffix can literally be worth ten thousand words, if it is a suffix that is applicable to that many bases.
"In this paper, we have made no mention of the fact that the morphology and syntax of Kuni and Ipiko [two Anim languages] differ profoundly from each other. We have not mentioned this because it is irrelevant. It has been known since Greenberg (1963) that the use of typological features in language classification leads to egregious errors. The only valid evidence is cognate signs."
("Signs" apparently means _signifiants_ here.)
Contrast on p. 138:
"Given the wide geographical distribution of the Anim languages and thefact that we required a reflex in one of the two western families as well as in one of the two eastern families, it is unlikely that our Proto-Anim reconstructions include many undetected borrowings. Chance similarity as the reason for the lexical agreements can also be excluded. There are too many exact phonological matches in words expressing concepts that are known to be diachronically stable in New Guinea."
My take is that morphology is wonderful evidence up until the point when it (relatively) suddenly disappears and/or is rebuilt from scratch. As in biology, a "total-evidence approach" should be used; there will be missing data, but they're not some kind of dealbreaker. And so, still on p. 138:
"The data situation in the Anim stock resembles that of many other language groups in New Guinea. Out of 17 languages, there are only two with a dictionary. For the remaining languages, we had to rely on survey word lists. Pessimists who think that historical-comparative work can only be properly embarked upon when there is a grammar and adictionary for most of the languages compared are, in fact, advocating that comparison be deferred indefinitely. We do not share this view. That word lists were only subjected to lexicostatistical analysis in the past was a choice that was not necessitated by the data. The comparative method can very well be applied to word lists. The initial insight into the genetic unity of the Anim languages came from a study of the available word lists. We have profited from the unusually large word lists and the morphological data collected by Drabbe [published 1954, 1955] and hope that his example will be imitated by others."
But anyway: "breast" in Were turns out to be [buːb] < *mbumb(V). Add that to Persian _bad_ and Mbabaram _dog_.
Marie-Lucie Tarpent
Hi fellow hattics,
I don't know the languages in question but I just want to comment on the difference between Morphology and Typology, which seem to be confused in one statement below. It is true that the typology (which refers to general types of language organization)
of a language can change quite drastically over time and therefore is not necessarily relevant as a criterion for language relationship. Examples of typological features which are often shared between two languages which are otherwise unrelated genetically
are: the number of vowels, the presence of unusual consonants, the presence or and number of tones, the indication of nominal features such as case through affixes or through word order, and similarly for the indication of verbal features such as tense, the
presence or absence of articles, of grammatical gender, and many others. Such general resemblances can be due to actual relationship, but also to contacts between neighbouring but unrelated languages. For instance, in South Africa the Zulu language uses
"clicks" as part of its stock of sounds, like the neighbouring Khoisan languages, but it is not related to them. Instead it is a member of the Bantu family which covers a very large area but most languages of which do not have clicks, which are present in
all the Khoisan languages. Historical records kept in tribal memories (as well as other data) show that the Bantu languages (along with cultural hegemony) expanded southward at the expense of the Khoisan ones.
Morphology on the other hand deals not only with the existence of general features but with the formation of words and the actual morphemes, the bits of language used in forming words and (if relevant) in adding the indication of extra features. For instance,
all the Romance languages have a complex verb morphology largely inherited from Latin, their common ancestor. Even though this morphology has become somewhat simplified (thus affecting the typology), what is left is still very obviously Latin in both general
form and actual verb endings, such the infinitive in -re. All these languages also have articles, which Latin did not have, but Latin did have several demonstratives which survive in the shapes of the modern articles, as in French le, la, Spanish el, la,
and others, from Latin ille, illa. The change from the absence of articles in Latin to the presence of articles in its descendants is a typological change, but that does not affect the validity of the relationship (which we would accept even if the history
of the language group was not known) since the morphemes still exist, albeit with a slight change in their function.
This is why I am not very happy with the authors' statement about the great differences in morphology and syntax (the latter indeed very subject to hange) between the languages they are dealing with, and their almost dismissing of morphology in a couple of sentences. Large numbers of resemblant vocabulary, even with rules of sound correspondences, are often due to contact and borrowing rather than genetic relationship. On the other hand, morphemes which are indispensable in the respective languages (such as those indicating verbal tense in most Indo-European languages) are usually present in large number of words, so that their form and meaning are readily identifiable within a single language, and can easily be compared between languages.
An obvious example is the determination of whether English and French are closely related or not. There are obvious resemblances between large numbers of French and English words, usually with close sound correspondences. Relying on lexical items alone (as in danse: dance, or cheminée : chimney) could lead to placing them together in a single genetic group (something which has actually been suggested). But comparison of their verbal morphologies makes this impossible: French verbs borrowed into English have been adapted to English verb structure, losing all their French features, and similarly for the (less common) opposite change through which English verbs borrowed into French have lost their typical English endings and added French verbal morphemes instead. Wider comparison involving Spanish, Italian, Dutch and German (at least) verb morphologies places French firmly in the Latin camp and English in the Germanic camp in spite of the fact that much English vocabulary is similar to that of French or Latin rather than Dutch or German. In this case, we do not have to guess about the source of the similarities, since we know of the historic event after which England was dominated for a few centuries by French-speaking conquerors, a situation in which the dominated population often adopts many words and sometimes other features of the dominant language.
So, I would say that the authors' conclusion should still be considered as tentative rather than definite, until they deal more thoroughly with the morphological aspects.
Sorry to take so long!
Sent: April 15, 2017 12:00:55 PM
To: hat...@googlegroups.com
Subject: Re: [hattics] The Anim Languages of Southern New Guinea
You received this message because you are subscribed to the Google Groups "hattics" group.
To unsubscribe from this group and stop receiving emails from it, send an email to hattics+u...@googlegroups.com.
To post to this group, send email to hat...@googlegroups.com.
Visit this group at https://groups.google.com/group/hattics.
For more options, visit https://groups.google.com/d/optout.