How do you tell what family what language belongs to ?

The best topics from Languages & Linguistics, kept on a permanent basis.
User avatar
Torco
Smeric
Smeric
Posts: 2372
Joined: Thu Aug 30, 2007 10:45 pm
Location: Santiago de Chile

How do you tell what family what language belongs to ?

Post by Torco »

There's plenty of cases where this isn't too problematic, but imagine an island colonized by late romans which already had been colonized -and forgotten- by chinese. You get something of a romlang with extensive chinese influences [or a sinolang with latin influences] is it an IE lang, or Sino-tibetan?

Or, say, creoles... what family do they fit in?

my bottom-line question is in the title: what's the criteria to say that a 'lang belongs in a particular family?

User avatar
vampireshark
Avisaru
Avisaru
Posts: 738
Joined: Thu Dec 04, 2008 7:02 pm
Location: Luxembourg
Contact:

Post by vampireshark »

As far as I gather, it's mostly the underlying grammatical structure that dictates what categories languages falls under. Maltese, for example, has an extensive amount of Italian borrowings, but its grammar is still heavily Semitic and, as such, 'tis a Semitic language. Likewise, Romanian is heavy in Slavic words, but it's Romance due to the grammar.

Creoles are separate beasts, so I'm not too certain about them.
What do you see in the night?

In search of victims subjects to appear on banknotes. Inquire within.

User avatar
Mbwa
Lebom
Lebom
Posts: 142
Joined: Sat Aug 18, 2007 1:48 pm

Post by Mbwa »

you only speak of odd cases here. Is the first example supposed to be a mixed language? Because areal influence doesn't really matter- it's what langauge it descended from. A romlang that takes a bunch of loans from Chinese doesn't suddenly gain sound correspondences with Proto-Sinitic.

And I think creoles are isolates, though I'm not sure. Actually, I have a feeling someone will prove me wrong on both of these.

User avatar
Radius Solis
Smeric
Smeric
Posts: 1248
Joined: Tue Mar 30, 2004 5:40 pm
Location: Si'ahl
Contact:

Post by Radius Solis »

A language is "related" to whichever other languages it shows extensive regular sound correspondences and systematic grammatical correspondences with. That is all that matters. :) For instance, English is still a Germanic language no matter how much Latinate vocabulary it has borrowed, and always will be.


Creoles are a special case. To extend the "language family" metaphor a bit, creoles are like adopted children: not genetically descended from anyone else in the family, but still a member by any common-sense measure.

True mixed languages, which are very rare, are an even more special case. They essentially have multiple parents, so in one sense it's not unreasonable to say they're related to both language families, but in another sense it's hard to call them related to anything, because they are essentially conlangs. (They are usually created quite intentionally by communities of a mixed ethnic identity, as a reflection of that mixture in language.)

zompist
Boardlord
Boardlord
Posts: 3368
Joined: Thu Sep 12, 2002 8:26 pm
Location: In the den
Contact:

Post by zompist »

Mbwa wrote:A romlang that takes a bunch of loans from Chinese doesn't suddenly gain sound correspondences with Proto-Sinitic.
Yes it does, if the borrowings are massive enough. Japanese, Korean, and Vietnamese are all used to help reconstruct Middle Chinese. English could be used to reconstruct Old French if there were a need for it.

(There are layers of borrowings to disentangle, of course.)

But your main point is correct— a language's core affiliation doesn't change. This is rarely very difficult to tell; the core vocabulary is less affected even by massive borrowings— there's no doubt that English is Germanic and that Japanese is not Sinitic.

Creoles can be set aside as a special case, or assigned to the main acrolectal language— e.g. toss French Creole in with French. Perhaps the most troublesome case is Michif, which takes its nouns from French and its verbs from Cree.

User avatar
Mbwa
Lebom
Lebom
Posts: 142
Joined: Sat Aug 18, 2007 1:48 pm

Post by Mbwa »

zompist wrote:
Mbwa wrote:A romlang that takes a bunch of loans from Chinese doesn't suddenly gain sound correspondences with Proto-Sinitic.
Yes it does, if the borrowings are massive enough. Japanese, Korean, and Vietnamese are all used to help reconstruct Middle Chinese. English could be used to reconstruct Old French if there were a need for it.
Interesting. I suppose the sound correspondence thing was a bad example. What I was trying to get across is that even if a language borrows intensively from another, it's not going to become part of that family, as you just said.

zompist
Boardlord
Boardlord
Posts: 3368
Joined: Thu Sep 12, 2002 8:26 pm
Location: In the den
Contact:

Post by zompist »

An example: there are regular č / š and dž / ž correspondences between Norman French borrowings into English and French.

User avatar
Morrígan
Avisaru
Avisaru
Posts: 396
Joined: Thu Sep 09, 2004 9:33 am
Location: Wizard Tower

Post by Morrígan »

Zompist wrote:Yes it does, if the borrowings are massive enough.
Strictly, the language would have correspondences with the donor language, but couldn't go an farther than that. So as you say, we could use English to look at Old French, but if we know about Old French, English won't help us with Vulgar Latin or Romanz, or whatever.

This is what makes inherited vocabulary special, I suppose.

User avatar
Torco
Smeric
Smeric
Posts: 2372
Joined: Thu Aug 30, 2007 10:45 pm
Location: Santiago de Chile

Post by Torco »

So what matters is grammatical similitude, instead of shared lexicon?

User avatar
linguoboy
Sanno
Sanno
Posts: 3681
Joined: Tue Sep 17, 2002 9:00 am
Location: Rogers Park/Evanston

Post by linguoboy »

Torco wrote:So what matters is grammatical similitude, instead of shared lexicon?
Actually, the best evidence is shared morphology. Lexical borrowing is dirt common, grammatical borrowing takes place wherever languages are in contact for an extended time, but it's really quite rare for one language to borrow another's inflectional morphemes.

User avatar
Morrígan
Avisaru
Avisaru
Posts: 396
Joined: Thu Sep 09, 2004 9:33 am
Location: Wizard Tower

Post by Morrígan »

Torco wrote:So what matters is grammatical similitude, instead of shared lexicon?
I want to say "no, not quite" but really, it's complicated.

If we are talking about language evolution, than it is a relatively simple matter of tracing inheritance. But looking back, we don't actually know what happened. We can infer that languages A and B descend from language C by comparing their lexicons, the sound changes inherited by the lexicon, and their grammars, with respect to morphology and syntax.

In highly synthetic languages, morphological similarities can be far more useful than sound correspondences. Vajda's connection of Yeniseian to Na-Dene was done largely on the basis of verb morphology; I don't recall seeing anything explicitly about sound correspondences in that paper.

hwhatting
Smeric
Smeric
Posts: 2315
Joined: Fri Sep 13, 2002 2:49 am
Location: Bonn, Germany

Post by hwhatting »

TheGoatMan wrote: In highly synthetic languages, morphological similarities can be far more useful than sound correspondences. Vajda's connection of Yeniseian to Na-Dene was done largely on the basis of verb morphology; I don't recall seeing anything explicitly about sound correspondences in that paper.
Well, he didn't just note similarities in the morphology, but identified individual elements, which implies that there are sound correspondences. Without that identification of individual elements, one could argue that Dene-Yeniseic is just an old areal group with analogical structures that later became disrupted by the inflow of "Altaic" languages, not a language family.
And if we're talking about this paper, he spends about 23 pages on morphology and 33 pages on establishing sound correspondences based on vocabulary comparisons. You're certainly right that the morphological comparisons are the really impressve part and form the main basis for establishing the relationship, but the lexical comparisons help in adding a lot of meat to the bones.

User avatar
WeepingElf
Smeric
Smeric
Posts: 1630
Joined: Wed Mar 08, 2006 5:00 pm
Location: Braunschweig, Germany
Contact:

Post by WeepingElf »

hwhatting wrote:You're certainly right that the morphological comparisons are the really impressve part and form the main basis for establishing the relationship, but the lexical comparisons help in adding a lot of meat to the bones.
Well put. Many long-range comparativists compare only dictionaries and not really languages, and thus end up with meat without bones ;)
...brought to you by the Weeping Elf
Tha cvastam émi cvastam santham amal phelsa. -- Friedrich Schiller
ESTAR-3SG:P human-OBJ only human-OBJ true-OBJ REL-LOC play-3SG:A

User avatar
sangi39
Avisaru
Avisaru
Posts: 402
Joined: Sun Mar 01, 2009 3:34 am
Location: North Yorkshire, UK

Post by sangi39 »

WeepingElf wrote:
hwhatting wrote:You're certainly right that the morphological comparisons are the really impressve part and form the main basis for establishing the relationship, but the lexical comparisons help in adding a lot of meat to the bones.
Well put. Many long-range comparativists compare only dictionaries and not really languages, and thus end up with meat without bones ;)
I remember one guy on a FB group trying to suggest a genetic link between the Eskimo-Aleutan languages and Japanese by comparing words from Nunavut and Japanese alone without even trying to find any sound correspondences. He was basically just one of your run-of-the-mill "hey, look how similat these words are" comparativists.
You can tell the same lie a thousand times,
But it never gets any more true,
So close your eyes once more and once more believe
That they all still believe in you.
Just one time.

David McCann
Sanci
Sanci
Posts: 25
Joined: Thu Mar 16, 2006 12:27 pm
Location: London

Post by David McCann »

Is there necessarily a single rule?

1. Copper Island Aleut has replaced the conjugational suffixes of the verb by Russian ones. So much for the persistence of morphology.

2. Sardinian speech is a continuum from Sardinian influenced by Italian to Italian influenced by Sardinian. It's not always easy to draw a line: in the Middle Ages, Corsican was a language like Sardinian, but now it's an Italian dialect.

3. Pidgins take a vocabulary mainly from a foreign language and a simplified grammar from a local one. We call Tok Pisin an English pidgins, classifying by vocabulary, but that may be just because we're English speakers; why shouldn't it be a Tolai pidgin?

3. Michif uses French for the nouns, Cree for the verbs. I'd say that's a conlang!

User avatar
Jacqui
Lebom
Lebom
Posts: 125
Joined: Fri Apr 23, 2010 5:29 pm

Post by Jacqui »

Root words, I guess.

User avatar
Basilius
Avisaru
Avisaru
Posts: 398
Joined: Thu Nov 16, 2006 5:43 am
Location: Moscow, Russia

Post by Basilius »

WeepingElf wrote:
hwhatting wrote:You're certainly right that the morphological comparisons are the really impressve part and form the main basis for establishing the relationship, but the lexical comparisons help in adding a lot of meat to the bones.
Well put. Many long-range comparativists compare only dictionaries and not really languages, and thus end up with meat without bones ;)
Where there are sound morphological comparisons, they mean a lot indeed.

But lack of morphological similarities means nothing. That is, just nothing.

In particular, it cannot be used as negative evidence (at considerable timedepths at any rate).

Also, if supposedly common morphological material does not display sound correspondences also found in lexical cognates, it probably represents chance resemblances - as easily as lexical look-alikes often do.

Torko's question was, basically, what is needed to demonstrate kinship with a known genetic grouping. For this particular task lexical similarities alone can work very well, if only you know how to cook them (which is in fact a very important reservation).
Basilius

User avatar
Morrígan
Avisaru
Avisaru
Posts: 396
Joined: Thu Sep 09, 2004 9:33 am
Location: Wizard Tower

Post by Morrígan »

I'm rereading Vajda (2008), and he there is a very good point that he makes in the introduction —
Vajda wrote:If Yeniseic is demonstrably relatable to Na-Dene, the evidence should be able to help solve Na-Dene internal problems by providing hitherto unknown external comparative data
Which is the basic point of what makes a theory a theory and not a hypothesis; a theory (or a good one, anyway) should have an explanatory scope which extends beyond what it is originally intended to explain.

I'm not sure I feel great about Colarusso's Pontic Hypothesis*, but the fact is that he is able to link roots in West Caucasian and PIE, and not just on the basis of words, but in terms of groups of words formed from a single semantic root. Thus, the claim that these groups are related is not just based on similarity between words or even sound correspondences, but that problematic PIE roots can be explained as derivations of Pontic roots. This is exactly the kind of thing the Octaviano consistently failed to do, it seems to me.

* This is partly because Pontic and NWC are so synthetic that I can't tell if he is just making things up; I trust that he isn't, but I'm not familiar enough with NWC to tell

User avatar
linguoboy
Sanno
Sanno
Posts: 3681
Joined: Tue Sep 17, 2002 9:00 am
Location: Rogers Park/Evanston

Post by linguoboy »

David McCann wrote:Is there necessarily a single rule?

1. Copper Island Aleut has replaced the conjugational suffixes of the verb by Russian ones. So much for the persistence of morphology.

2. Sardinian speech is a continuum from Sardinian influenced by Italian to Italian influenced by Sardinian. It's not always easy to draw a line: in the Middle Ages, Corsican was a language like Sardinian, but now it's an Italian dialect.

3. Pidgins take a vocabulary mainly from a foreign language and a simplified grammar from a local one. We call Tok Pisin an English pidgins, classifying by vocabulary, but that may be just because we're English speakers; why shouldn't it be a Tolai pidgin?

4. Michif uses French for the nouns, Cree for the verbs. I'd say that's a conlang!
Copper Island Aleut and Michif are conlangs. They are mixed languages, of which Rad said upthread:
Radius Solis wrote:True mixed languages, which are very rare, are an even more special case. They essentially have multiple parents, so in one sense it's not unreasonable to say they're related to both language families, but in another sense it's hard to call them related to anything, because they are essentially conlangs.
Also, I'm not sure what your point is about Corsican and Sardinian. These are still both classified as Romance languages. And you're not exactly the first person to note the shortcomings of the Stammbaum model.

User avatar
Radagast
Lebom
Lebom
Posts: 97
Joined: Sat May 21, 2005 9:46 am
Location: Denmark
Contact:

Post by Radagast »

Mixed languages are of course not conlangs. They are completely natural languages. They are natural languages because all languages are mixed to one degree or the other. Being a mixed language is the natural state of a language. Some language are just called "mixed" because the circumstances of their geneses so mixed that they are confusing for the historical method and the basic assumptions of the tree model. That doesn't make them conlangs - nobody sat down and decided to create them, they emerged in a community of speakers completely naturally.
[i]D'abord on ne parla qu'en poésie ; on ne s'avisa de raisonner que long-temps après.[/i] J. J. Rousseau, Sur l'origine des langues. 1783

User avatar
vohpenonomae
N'guny
N'guny
Posts: 91
Joined: Sat Nov 02, 2002 4:23 am

Post by vohpenonomae »

Radagast wrote:Mixed languages are of course not conlangs. They are completely natural languages. They are natural languages because all languages are mixed to one degree or the other. Being a mixed language is the natural state of a language. Some language are just called "mixed" because the circumstances of their geneses so mixed that they are confusing for the historical method and the basic assumptions of the tree model. That doesn't make them conlangs - nobody sat down and decided to create them, they emerged in a community of speakers completely naturally.
This brings up an interesting question--is there any difference between a naturalistic conlang and a natural language, aside from the (obvious) reality of the speaking population and the number of creators? I find it hard to formulate any principle that would distinguish the two.
"On that island lies the flesh and bone of the Great Charging Bear, for as long as the grass grows and water runs," he said. "Where his spirit dwells, no one can say."

User avatar
Terra
Avisaru
Avisaru
Posts: 571
Joined: Tue May 24, 2005 10:01 am

Post by Terra »

I'm surprised that nobody's mentioned modern Israeli Hebrew yet. On one hand it's got clear semitic influences with the tri-consonant root system, but at the same time it was adopted by mostly various european jews, who undoubtedly subconsciously molded it according to their first language.

Edit: I can't spell.
Last edited by Terra on Fri May 21, 2010 4:41 pm, edited 1 time in total.

Travis B.
Sumerul
Sumerul
Posts: 3570
Joined: Mon Jun 20, 2005 12:47 pm
Location: Milwaukee, US

Post by Travis B. »

I would probably say that modern Israeli Hebrew is a Semitic language, albeit one with quite significant Indo-European substratum influence. As to whether it is a conlang or not, well, that is an exercise for the reader.
Last edited by Travis B. on Fri May 21, 2010 4:49 pm, edited 2 times in total.
Dibotahamdn duthma jallni agaynni ra hgitn lakrhmi.
Amuhawr jalla vowa vta hlakrhi hdm duthmi xaja.
Irdro. Irdro. Irdro. Irdro. Irdro. Irdro. Irdro.

User avatar
vohpenonomae
N'guny
N'guny
Posts: 91
Joined: Sat Nov 02, 2002 4:23 am

Post by vohpenonomae »

hwhatting wrote:
TheGoatMan wrote: In highly synthetic languages, morphological similarities can be far more useful than sound correspondences. Vajda's connection of Yeniseian to Na-Dene was done largely on the basis of verb morphology; I don't recall seeing anything explicitly about sound correspondences in that paper.
Well, he didn't just note similarities in the morphology, but identified individual elements, which implies that there are sound correspondences. Without that identification of individual elements, one could argue that Dene-Yeniseic is just an old areal group with analogical structures that later became disrupted by the inflow of "Altaic" languages, not a language family.
And if we're talking about this paper, he spends about 23 pages on morphology and 33 pages on establishing sound correspondences based on vocabulary comparisons. You're certainly right that the morphological comparisons are the really impressve part and form the main basis for establishing the relationship, but the lexical comparisons help in adding a lot of meat to the bones.
You don't necessarily need to establish a large body of similar elements to establish genetic relationship. Sapir connected Wiyot and Yurok to Algonquian in 1913[1] based largely on morphological similarity and very few similar elements; only the basic prefixes look obviously connected between the three Algic branches. But their morphologies are so consistently alike that evoking areal influence is a real stretch. Areal influence rarely, if ever, results in a large number of consistent correspondences in morphological structure; it tends to be more slapdash.

Cheyenne was established as Algonquian in a similar way, long before its phonetic history was worked out. Again, the morphology, grammatical distinctions, etc. showed that it couldn't be anything but Algonquian.

Of course, both of these hypotheses were controversial for years (the Algonquian-Ritwan controversy persisted until 1958), because sound correspondences are the gold standard for showing genetic relationship, then as now. But, as all this shows, you can know that something is the case without necessarily knowing how it is. Comparative morphology is still in early days relative to comparative phonology.

[1] A controversy remains today as to whether Sapir proved the relationship in 1913 (and with subsequent publications) or whether it was proven by later 1950s fieldwork in Wiyot and Yurok. Some pieces of Sapir's evidence were faulty, due to relying on bad records; but the critiques of his 1913 paper were equally unsound. However, it was these critiques--most notably those by Truman Michelson--that held sway until the 1950s. Michelson was the pre-eminent Algonquian scholar of his time, and his opinions exerted great influence on his and subsequent generations. I'm of the view that Sapir proved the relationship in 1913 and most linguists simply toed the Michelson line until that line became untenable--as happens so often in the history of science and the humanities.
Last edited by vohpenonomae on Fri May 21, 2010 7:14 pm, edited 1 time in total.
"On that island lies the flesh and bone of the Great Charging Bear, for as long as the grass grows and water runs," he said. "Where his spirit dwells, no one can say."

TomHChappell
Avisaru
Avisaru
Posts: 807
Joined: Wed Dec 28, 2005 2:58 pm

Re: How do you tell what family what language belongs to ?

Post by TomHChappell »

Torco wrote:my bottom-line question is in the title: what's the criteria to say that a 'lang belongs in a particular family?
This isn't a perfect rule, but in Western Europe (and less reliably also in Eastern Europe) languages in the same family tend to have the same last name.










:wink: :mrgreen:

Post Reply