How exactly is a language classified into language families?

Discussion in 'Etymology, History of languages, and Linguistics (EHL)' started by DaveWen, Oct 31, 2013.

  1. DaveWen Member

    Hi all, hope this doesn't come across as a silly question, but----what is the objective way in which a language is put into a family alongside other languages? In the case of Middle English, for example, how much more Frenchy did it have to be so that it couldn't be considered "English" anymore, and in what aspect of the language?
    Another case that occurs to me are languages spoken in Southern China. In Northern Hainan, for instance, there's a language called "Lin-gao", or "Ong Be" as it's known in the west. It is classified as a Tai-Kadai language. Over its lifetime, it's been heavily influenced by local Chinese languages like Yue and Min, and most recently Mandarin, in vocabulary, grammar and phonology. Conversely, many southern Chinese languages have received a fair amount of influence from local non-Chinese languages. What are the deciding factors that put these languages into their respective language families?

    Thanks in advance.
  2. killerbee256 Senior Member

    American English
    Loan words do not change the family that a language belongs too. Romanian has a large amount of Slavic loan works but it's still a romance language.
  3. DaveWen Member

    Thanks for the reply. I understand the principle, but to say such and such words are loanwords you need to have already put the language in a language family, don't you? Using the examples of the original post of mine, we all know that English is Germanic, Ong Be is Tai-Kadai, and Cantonese is Chinese----what exactly are the deciding factors that put these languages into their respective families? Is it the grammar, for instance?
  4. berndf Moderator

    German (Germany)
    Basically, saying that languages belong to that-and-that group means that that these languages evolved from a common ancestor language or dialect continuum. Thus, claiming that some languages belong to a certain family is not a mere definition but an actual historical hypothesis.
  5. CapnPrep Senior Member

    The classical comparative method is based on systematic correspondences in the phonetic form of native vocabulary. In other words, you have to identify loanwords and exclude them, and no, this is not always easy. The case of English is almost ridiculously simple because we have an extensive historical record written in alphabetic script, both for English itself and for the neighboring languages. It is more challenging, sometimes impossible, to a classify a language if it has no living descendants, no living relatives, massive borrowing, a non-phonetic script (or no script and no historical record), etc. etc. but the methodological principles remain the same.
  6. Gavril Senior Member

    English, USA
    According to the most widespread theory (inasmuch as I understand it), the words and inflectional system of a language have no necessary connection to that language's genetic affiliation.

    If a language is a system for communication within a speech community, then the specific words, inflectional patterns etc. that a community uses for communication are not essential components of their language -- instead, what matters is that the language is theirs, i.e., their means of communicating with one another.

    In theory, every word, affix and construction in a given language could be replaced over time, but as long as this replacement happened without a breach in communication between (or within) generations, the language would still be the "same" language, and would therefore have the same genetic affiliation.

    Thus, English is considered Germanic because it is assumed that modern-day English continues an unbroken chain of communication dating back to Proto-Germanic, but not to French or Proto-Romance.

    I'm still not clear on how linguists define "unbroken chain" here, but insofar as this concept can be defined, it seems possible to use it to identify genetic lineages.
    Last edited: Nov 1, 2013
  7. rbrunner Senior Member

    German - Switzerland
    Well, this sounds good and useful, and looks quite harmless. But I think making this rule part of the rule set that defines when/whether languages are "same" can have some surprising consequences:

    Imagine two peoples living next to each other, speaking 2 clearly different languages A and B at first. They like it other, they have much contact, they visit each other all the time, the intermarry, and as a result the A people happily borrow words and grammatical features from language B. The B people do likewise and pick up more and more features from the A language.

    This continues until the A and the B languages developed and gravitated so much towards each other that essentially they become a single language with only little dialectal variation over the combined space occupied by the A and B peoples.

    Now, would this finally common language be one with two ancestors, belonging to two genetic lineages at the same time? No breach of communication, only gradual change, given rule applies then?

    And if yes, who could affirm that in the case of English there really never was a sizable group of Anglo-French aristocrats starting with their French dialect and over time adopting more and more English words and grammar features until they also arrived at the same English enriched with French things that the peasants spoke? Making the aristocrats the A people of my theoretical example, and the peasants the B people, and English a language that is Germanic and Romance at the same time?
  8. DaveWen Member

    Then, I suppose, the resulting language should be called a creole?
  9. Gavril Senior Member

    English, USA
    Just out of curiosity, what do you think should be the rule set?

    What you describe sounds like what some linguists would call a mixed language, i.e., a mixture of two historical lines of communication.

    I think some linguists would say that the French-speaking aristocrats didn't gradually modify their French to make it more and more like English. I.e., they didn't slowly chip away at their French vocabulary and grammar, but instead they "abruptly" adopted the language spoken by their subjects, and thereby severed the line of communicative continuity with other French speakers.

    However, the concept of "abruptness" in historical linguistics is not fully clear to me -- just like the complementary idea of "unbroken" continuity -- so personally, I don't know how I would answer your question.
  10. Ben Jamin Senior Member

    Can you think of an example of such a perfectly evenly mixed language? As far as I know, one language is usually stronger than the other, and the admixture usually goes one way (at least at the same time). 950 years ago English began to loan heavily from French, now it goes the other way round, but it is not happening at the same time.
    Japanese has borrowed tens of thousands English words, but there are not even 100 words that made their way the opposite direction.
  11. rbrunner Senior Member

    German - Switzerland
    I don't know either. I think your rule that basically says "gradual change cannot destroy a language" makes a lot of sense, but I am not sure whether all its consequences are desirable; that's why I tried an "exercise in logic" to see where it may lead.

    But after all, languages and their developments are messy, and mathematical precision when classifying languages is probably not possible. Maybe a few clear rules and some "strange" cases of classifications resulting from them are better than very complicate rules or rules that are so abstract that most people do not understand them.

    Funny that you mention that - there was an earlier thread here which showed that there are quite conflicting views about creoles...
  12. rbrunner Senior Member

    German - Switzerland
    No, but I don't know whether my still considerable linguistic ignorance is to blame, or whether such cases indeed do not occur.

    And I think that line of reasoning can be dangerous: "You know, these rules logically allow for border cases that would be problematic, but that's no problem, those don't occur anyway."

    To exagerate a little, that's like a mathematician saying that his or her proof is ok because surely nobody intends to divide by zero :)

Share This Page