As I expected, the Classification of Chinese Language post is coming under some very serious criticism. I’m deleting all of the comments from the individual making them, but I will deal with them in this post.
First of all, it is alleged, by this person and others, that mutual intelligibility is a very fuzzy concept and is unscientific at core. This is not the case at all. I am not aware of any kind of fighting going on in Linguistics, other than complaints by a few cranks, about the issue of mutual intelligibility in general or between any lects. The concept is quite clear – speakers either understand other speakers or they don’t. It’s that simple.
How do linguists determine whether speakers of Lect A can understand Lect B? Real easy. They do something “completely unscientific.” They sit down and ask speakers of Lect A: “Can you understand speakers of Lect B?” Really, that’s how it’s done. It’s called Sociolinguistics, and these determinations are usually made by personal interview. There’s usually no need for fancy tests with phonology, lexicon, tones, or syntax. It’s pretty much a yes or no answer.
Obviously, many times you get something other than a yes or no answer. If they understand Lect B 20%, they say, “We can barely understand those people at all!” If they understand 1/2 of what they hear, or 70%, or 85%, or 90%, that’s a little harder, but people are not morons. Often they will attempt to tell you what percentage of Lect B they can figure out. They will say, “We can understand 50/65/70/85/90% of Lect B.” Really, that’s how it’s done.
Sometimes some speakers say they can understand more than others. There are speakers of Portuguese and Spanish who say they understand the other language perfectly. Are they in the majority? Surely not. In those cases, you just go with the majority. On the other hand, if one person says, “I can only understand 70% of Lect B,” and everyone else in the village says, “We can understand Lect B perfectly,” you go with the majority again and dismiss the first person as anomalous.
If you look at Ethnologue‘s Mexico page, you can see that Ethnologue‘s sources have gone all over rural Mexico determining the variable intelligibility of different lects of Mexican Indian speech. This is in an area where just about every village speaks a different language or dialect. Under Tijaltepec Mixtec, we find the following:
Speakers have 89% intelligibility of San Miguel el Grande and Yosoyúa, 82% of San Mateo Peñasco, 81% of Sinicahua and 66% of Teita.
So how did Ethnologue arrive at those figures? Who knows. Are these figures controversial in the field of Linguistics, that is, do we fight about endlessly in our journals and books? Not at all. Are those figures above right? Probably. Are they wrong? Anything’s possible. Do we fight about it? Other than a few nuts on the Internet, no way.
I assume Ethnologue is just sitting people down in front of tape recordings of the other lects and see how much they understand. I assume there are scientific ways of calculating how much they get and how much they don’t.
What is Ethnologue? It’s the main publication of SIL. Who is SIL? Why, they are the very folks who are giving out ISO codes, the codes that determine whether something is a language or not – languages get ISO codes, dialects do not. Was giving this job to SIL controversial? Other than a few cranks here and there, not really.
There are currently some great big dustups in Language Policy over a few lects. Valencian speakers are fighting like mad to get their lect recognized as a language apart from Catalan. Ethnologue seems to be turning them down so far, because the two lects can understand each other.
Moldavian was granted an ISO code, then Ethnologue snatched it away and decided that Moldavian was a dialect of Romanian.
Are there intense passions aroused in these fights? Sure. Is mutual intelligibility one of the things that they fight about? Not generally, because the figures are out there, and everyone but for a few cranks agrees that the figures are scientific.
There are also indices of lexical similarity and there are various ways of doing this too. Spanish and Portuguese have 89% similarity, Italian and French have 87%, English and Frisian have 64%, English and French have 27%, on and on. Does anyone argue about these figures and say that they are wrong? Other than a few cranks, not really.
Finally, the very concept of a pilot study has been challenged by folks who feel that everything they write down has to be correct. I’ve also been accused of changing my research too much as a result of criticism.
This is wrong on various counts.
Pilot study is the way to go in science. Furthermore, it’s very humble. The first study really is a pilot study. Pilot study just means the first one. Is it making any claims about much of anything? Hardly. So what’s the purpose? To stimulate further thinking, theorizing, criticism and studies. That’s it. It’s the first guy to the lake, tossing out his line and reporting what bites. That’s all it is.
Research, and science, should be anything but certain. If you read academic journals, particularly the more scientific ones, at the end, they usually list about 3-5 reasons why their conclusions might be wrong. That’s as humble as you can get. Then they go through and say why they think all those reasons are probably wrong. Conclusions, when presented, are often couched in tentative language.
Does this mean scientists are a bunch of insecure wussies? Not even.
That’s simply how one does science.
Science minus huge doses of uncertainty and humility might be interesting, but it ain’t science. What is it? I don’t know. Religion, opinion, polemics, politics, advertising, fiction? Any or all of those.