Minnan is a language group and not a single language

[xng]

The common misconception is that Minnan is a single language is incorrect with Teochiu and Hokkien being dialects of this language incorrectly stated. To be dialects of a language, dialects should be more than 90% mutually intelligible eg. American English and UK English.

Technically, Minnan should be called a group of mutually unintelligible languages (with around 50% intelligibility) derived from the same ancestral language. Just as we have West Germanic languages which includes German and English, southern Min languages should include several distinct languages.

1. Quan-Zhang branch (colloquially called Hokkien)

The dialects of this language are Quanzhou, Taiwanese and Zhangzhou .

2. Chaoshan branch (colloquially called Teochiu)

3. Qiongwen branch (colloquially called Hainanese)

4. Puxian branch (colloquially called Hing Hua)


Qiongwen and Puxian are only around 50% mutually intelligible with Hokkien branch but in linguistics, languages that descend from the same ancestral language are grouped together in the same language family.

You might be shocked to hear that Chaoshan branch actually descended from old Putian language before they migrated to Guangdong province. Putian dialect was very close to Quanzhou dialect before splitting from Quanzhou during the Song dynasty.

Hainanese seems to be a secondary migration from Chaoshan people.

After the split, Putian dialect underwent influence from Fuzhou language. While Hainan Min underwent influence from local Li (Tai-Kadai) language.

[xng]

Just to give a hypothetical hierarchical comparison to Indo-European language family

1.Sino-Tibetan / Indo-European
2.Sinitic / European

3.Min / Germanic languages
4.Minnan / West Germanic languages
4.Minbei / North Germanic languages
4.Mindong / East Germanic languages

3.PingYue / Romance languages
4.Ping hua / Western Romance languages
4.Yue / Eastern Romance languages

[Abun]

I've pretty much given up on the issue of language vs. dialect because it has become more and more clear to me that the two are impossible to clearly define in scientific terms.

Yes, the most commonly heard definition is that if two local vernaculars are not mutually intellegible, they are different languages, otherwise they are dialects of the same language. Which sounds clear at first but in practice it's not that easy.

The first problem is that language/dialect borders are blurred but form dialect continua. To give an example from West-Germanic languages: Monolingual speakers from Berlin and Amsterdam would not be able to understand the majority of what the other is saying, so one would classify Amsterdam Dutch and Berliner German as two different (albeit related) languages, right? But what if we compare speakers A from Amsterdam, B from Enschede (near the German border), C from Osnabrück (near the Dutch border), D from Hannover, E from Magdeburg and G from Berlin (each speaking their traditional language of course, not the standard languages)? We will find that each of these speakers will be able to communicate with their neighbours (and often at least one over as well). In fact, I suspect the most radical difference between two neighboring speakers would not be at the German-Dutch border, but between D from Hannover and E from Magdeburg because that's roughly where the (equally blurry) border between Low and Middle German is crossed and Low German and Dutch are closer to each other than Low German to Middle German.
In any case, we would face the paradoxical situation that neighbouring pairs of speakers would be classified as speaking the same language, only different dialects, the two speakers at the extremes would be classified as speaking different languages.
The same phenomenon can be found in China, not only within the big families such as Min, Yue, Wu ect., but also on the borders between them.

Another, smaller problem are loaning and sprachbund phenomena (i.e. phenomena which cause originally more distant languages to become closer to each other). A monolingual English speaker for example would be able to understand a substantial bit of the words from, say, Italian because of the large number of Latin and French loanwords in English. But even if he can gather as much as the rough content of what is said, one would not consider English and Italian as dialects of the same language.

Therefore I have come to the conclusion that the use of "language" and "dialect" as two distinct concepts is a purely political one. If somebody proclaims a certain linguistic entity (for want of a better word other than "language" or "dialect") a "dialect", they usually have a political agenda to stress unity with some other group. Conversely, the concept of a seperate "language" is usually used in a context where people have the political agenda to stress distinction from some other group. And as usual it's those in power who impose their political views as the standard. As a saying popularized by sociolinguist Max Weinreich goes: "A language is a dialect with an army and navy."

[xng]

I've pretty much given up on the issue of language vs. dialect because it has become more and more clear to me that the two are impossible to clearly define in scientific terms.

Yes, you have some valid points in your arguments. There are problems with dialect continuum and loan words but these can be overcome as follows.

1. Dialect continuum

Don't think in an overly complex manner, if two monolingual people can communicate with each other 90% of the time, they are considered dialects of each other, otherwise they are different languages. The different dialects are compared with the standard dialect of the language. I doubt two dialects of border languages can have more than 90% intelligibility with each other. That would imply that the standard dialects of the two languages would also have 90% intelligibility, then it shouldn't be called languages but dialects. We compare based on the standard dialect of each language to determine which language it falls under.

2.Loan words

Loan words do not factor into classification of a language, there are loan words in Japanese from Chinese but Japanese is not classified as Sino-Tibetan due to different basic cognates and grammar. Furthermore, loan words won't make a language to be 90% mutually intelligible with the loan language because only a certain percentage make up the loan words for the total vocabulary . Eg. Even though Puxian language adopted some loan words from Fuzhou, most of its cognate vocabulary and grammar still follows its root ancestral language Minnan.

3. Language group

Different languages in the same group don't have to be mutually intelligible. They only need more than 50% lexical similarity. Eg. German has low mutual intelligibility with English but has a higher level of lexical similarity at 60%. We use the predominant number of basic cognates and similar grammar to determine which language group it falls under. We also consider whether the two groups were once the same people speaking the same language in ancient times before the two groups migrated into different directions.

[Abun]

Don't think in an overly complex manner, if two monolingual people can communicate with each other 90% of the time, they are considered dialects of each other, otherwise they are different languages. The different dialects are compared with the standard dialect of the language. I doubt two dialects of border languages can have more than 90% intelligibility with each other. That would imply that the standard dialects of the two languages would also have 90% intelligibility, then it shouldn't be called languages but dialects. We compare based on the standard dialect of each language to determine which language it falls under.

That's a little clearer (especially the specification of a degree of mutual intellegibility is necessary since that is a continuous scale). However it entails problems as well. The most obvious is the definition of standard dialect. That works if you want to compare Standard German and Standard Dutch, but for Teochew and Hokkien? What are their standard dialects? That of Teochew probably that of Chaozhou proper because that has the most speakers? But of Hokkien? Xiamenese? Taiwanese? These are prestige dialects, yes, but like the number of speakers that's a sociological criterion. Whatever you choose, it will be sort of arbitrary. Maybe you compare Xiamenese and Chaozhou-ese, get an average of 85% mutual intellegibility and classify them as distinct languages. But who's to say that the next study won't get quite different results, just because their criteria for determining a standard dialect were slightly different?

Oh and btw, "dialects" of bordering "languages" can be very much intellegible with each other as well as with their respective standard "dialects" as the Dutch/German example shows. I'm from a traditionally Low German-speaking area, and although I grew up a few hundred kilometers from the border (and never learnt Dutch), it is noticably easier for me to understand Dutch people who come from the eastern parts of the country than from the West (I mean in spoken form of course. In written form Standard Dutch is no big problem, either). And somebody from the northern parts of the Middle German-speaking area (e.g. Berlin) will have the same feeling with Low German.
So if you take Standard German as a point of comparison, Dutch is definitely a different language and so are most variants of Low German. But some variants near the Low-Middle German border will have >90% mutual intellegibility with standard German while at the same time also having >90% MI with other variants that have <90% MI with the standard. Since there is no standard variant of Low German it is thus impossible to give a definitive answer to the question whether it is a distinct language from German. And even if there was a standard variant, depending on which one would be chosen, there would either be an area of overlap where the local languages have >90% MI with both standards or a "no man's land" where both are <90%. In the latter case we would be forced to conclude there is a third language in between, just because someone for some arbitrary reason chose one variant over another as the standard dialect.

Loan words do not factor into classification of a language, there are loan words in Japanese from Chinese but Japanese is not classified as Sino-Tibetan due to different basic cognates and grammar.

Of course, I wouldn't suggest otherwise. What I was hinting at was that there are certain traps one might fall into when ascertaining the degree of mutual intellegibility. Such as equating it with the amount of cognate words in both languages. Measuring the time in which monolingual speakers understand each other is more accurate but it has its practical problems as well: Firstly, unless you ascertain "understanding" it by monitoring the participants brain waves, you will be forced to let the participants assess it for themselves which is rather subjective. Also, the results will vary greatly depending on the topic, because some word fields may for whatever reason have a greater number of easily recognizable cognates than others. And lastly it may prove difficult to even find monolingual speakers. Try finding Low German speakers who don't know any standard German...

[xng]

but for Teochew and Hokkien? What are their standard dialects? That of Teochew probably that of Chaozhou proper because that has the most speakers? But of Hokkien? Xiamenese? Taiwanese? These are prestige dialects, yes, but like the number of speakers that's a sociological criterion. Whatever you choose, it will be sort of arbitrary. Maybe you compare Xiamenese and Chaozhou-ese, get an average of 85% mutual intellegibility and classify them as distinct languages.

The defacto standard dialect of Hokkien/QuanZhang branch is Taiwanese/Xiamen dialect due to the proliferation of Taiwanese songs and movies. (This reason is the same for Cantonese due to HK movies and songs). In fact, there are yearly international Hokkien song competition held and the competitors are from around the world, mostly South East Asia, Taiwan and China. If that's not defacto standard, what is.

Firstly, unless you ascertain "understanding" it by monitoring the participants brain waves, you will be forced to let the participants assess it for themselves which is rather subjective. .

I am not sure which dialect is the prestige dialect for Chaoshan and it certainly doesn't have 85% intelligibility with Xiamen, I read somewhere it's more like 50-60%.

It's not subjective to assess at all, show them videos of common everyday life situation and test them to what extent they can understand the video. Try understanding the video below and please do not read the Chinese characters. See whether you can understand 85% of the time.


https://www.youtube.com/watch?v=0JYsTHh-nxI

[xng]

moved