In the early nineties our little research group had a couple of servers, one of which was called “mort” after the Terry Pratchett character. TP was very popular around that time in the group. When we replaced the machine, I suggested we call it “umfi” which — I claimed — means “death” in Xhosa. Today I was looking up another Xhosa word in Fischer (one of the two classic Xhosa dictionaries; the other is McLaren). I happened to notice that they translate “death” as “ukufa”. I was a little shocked because my Xhosa is very bad and I might easily have made a mistake all those years ago. What did we really name the machine?
Fortunately, Xhosa is quite a subtle language (but, like Finnish, it has a very logical internal structure). “Ukufa” is an abstract noun, a little like an infinitive, which means “to die” or “the act of dying”. All class 15 words (prefix “uku-”; 15 in the Meinhoff classification; I don’t know what McLaren’s classification is called, but it’s class 8 in his dictionary) are similar; they denote “the act of …”. So, in Xhosa the verb “to die” is “-fa”. It can also mean to be very ill. As with many Xhosa nouns, there are a large number of derived words you get by throwing “-fa” into different classes: “ilifa” is an inheritance, “imfa” is mortality, “isifo” is illness, and “umfi” is the dead or deceased. (The final vowel needs to be adjusted in some cases, as you see.) Yippee!
Xhosa verbs can also be transformed with a kind of causal-final construction: “fela” (= to die for something, or to die at some specific place). This new verb can in turn be transformed with a passive construction: “felwa” (= to be bereaved (Xhosa has stative verbs) or to lose someone through death). Verbs can also be transformed with a causative construction: “fa” -> “fisa” (= to make someone die), or “fela” -> “felisa” (= to make someone die for something). From these words come “umfisi” (= a murderer) and “isifiso” (= assassination). The list is endless and includes words like “umfana” (= a young man or servant) and “imfambele” (= a friendly orphan, or a cow with blind teats), and (I think) “ifaca” (= a weak, emaciated, starving person).
I have really fallen in love all over with Xhosa. It is such a rich language. I won’t have time this year to study it, but perhaps next year I’ll pick it up again. I’ve done about one third of the Xhosa second year course at the university, but that was long ago. It used to be excellent, and I’m sure it still is. I hear nothing but good about our university’s language courses.
> Fortunately, Xhosa is quite a subtle language
What does this mean?
> but, like Finnish, it has a very logical internal structure
What does this mean?
> I hear nothing but good about our university’s language courses.
Apparently they don’t teach basic linguistic facts:
“There is no evidence that any language is any more ‘logical’ than any other”
Linguists do not agree on everything, but see:
http://www.llas.ac.uk/resources/gpg/135
where the quotation was taken from.
Wow, you are so strict! But, admittedly, I used a lot of imprecise terms and I can appreciate that linguists do not like to see their science disrespected.
By subtle language, I meant that it has a large vocabulary and is able to make fine distinctions, perhaps finer that some people expect. Perhaps all languages are equally capable of this to some degree and in different ways, but in my experience most people are unaware of their own languages’ richness of, for example, modal expression.
Xhosa is “logical” in the informal sense that it is very regular. Well, in a mathematical sense all languages are perfectly “regular” (not in the automata-theoretic sense, of course, but information-theoretically), and differ only in the size of the description of the grammar. Perhaps one could even measure this using some form of Kolmogorov complexity.
As for our university’s language courses, I don’t think they teach linguistics. That would be a linguistic course.
Thanks for the link. The article is really interesting.
> Wow, you are so strict!
Am I?
> But, admittedly, I used a lot of imprecise terms and I can appreciate that linguists do not like to see their science disrespected.
I don’t know what to say. Why is it that when a linguist points to linguistic facts being wrong this is seen as an emotional issue? Not something to be calmly discussed… Don’t get me wrong but we get these kinds of answers all the time when we are simply pointing things out.
> As for our university’s language courses, I don’t think they teach linguistics. That would be a linguistic course.
That was just a joke. I myself have been studying languages alongside linguistics and know that in particular language classes the teachers present wildly unlinguistic factoids. Sad but true.
As for other things, your distinctions don’t fare quite far. Let’s take the issue of Xhosa being very regular. How so? I’m guessing morphology here. Morphology is not the whole end-all of language. And besides would you categorize Vietnamese (a language without any morphology) as not logical or thoroughly logical. Why? As for Kolgomorov complexity that won’t work. I can say that without knowing what Kolgomorov complexity is (never got that far with math/stochastics/statistics). There have been numerous attempts to couple language with logic, none have fared well. If you are interested in why, may I suggest you post a query in the sci.lang newsgroup (they have a surprising amount of mathematicians there), as I think I cannot explain the weak points in combining mathematics with linguistics here.
Anyway, all the best!
> And besides would you categorize Vietnamese (a language without any morphology) as not logical or thoroughly logical. Why?
I might have added your native tongue (Afrikaans if I’m not mistaken) with little morphology here:
–
Nominal Morphology
Afrikaans can be considered as the most analytic of all Indo-European languages. In the process of development its declension was subject to radical simplification. The noun fully lost the gender distinction; the plural number is formed by the endings -e and -s. The declension disappeared, the the relations in the sentence are expressed by word order or by means of prepositions. There is a definite article die which is not declined, and an indefinite one ‘n in singular.
Verbal Morphology
The verb is characterized by the complete loss of person and number. The past simple tense was lost, and of all the tenses, the Present, the Perfect, the Future I-II, and the Future in the Past were preserved. The original Germanic strong verbs (English irregular verbs) do not exist, all verbs are weak.
–
You might see that some of the functions (and they are also hard to describe accurately) have moved from morphology to syntax. So morphology in itself is not enough in language categorization.
I was planning to just placate you, since in the final analysis, this topic is not life-and-death. However, your comments has really made me think about this issue.
First, I should point out that nothing that I said in my original post is untrue, and you never claimed that it was. Rather, your objection was, and I admitted, that it is imprecise. For example, my use of the word “logical” is not precise, because what exactly do I mean by it?
But then I noticed that your quotation from the 83-statements-linguists-agree-on uses the same word!
I’m not sure what the linguists behind that article meant, but what I, at least, had in mind the fact that Xhosa is more regular, sticks to its rules more strictly, has fewer exceptions that other languages. And this is really what Kolmogorov complexity is all about. It measures the complexity of the (or a, if you like) description of a language.
I have not studied linguistics, so I have the luxury of disagreeing with what seems to be the consensus among linguists that all languages are equally logical (or at least, that they have not found evidence to the contrary). In the same way non-mathematicians are free to believe that 1 + 1 = 2, even though mathematicians know that this is not always true or sensible.
I am willing to accept that all languages are equally powerful and can express the same concepts, but it seems foolish to me to believe that all languages have exactly the same degree of complexity/regularity. Here, of course, complexity refers to an objective measure, not a subjective opinion. Even if we do not know how it is defined, I believe that however it is defined, there will be some variation among languages. And my hypothesis is that there are some languages which, for most reasonable definitions, will tend to fall on the “easy” side of the scale.
Afrikaans is perhaps one such a language. (One technicality: there are exactly two irregular “strong” verbs: is-was-gewees (to be) and het-het-gehad (to have).) Despite the opinions I expressed earlier about Xhosa and Finnish, I’m not so sure. Xhosa has 15 “genders” or noun classes, and Finnish has at least twelve (??, you might know), though they are not called “genders” (I’m thinking of “kuu” v. “vesi” v. “nainen” etc.).
I suppose I should put my money where my mouth is, or just shut up! Computational linguistics is a big field, and at one point I thought about exploring it, but one needs more linguistic background than I have or can easily acquire. But this kind of “complexity” problem is very enticing. Do linguists in hiding work on this kind of stuff?
Sorry for being such an ass. Anyway, I was serious about posting to sci.lang. You might, well, I don’t know. But it is a revelation (there are SERIOUS mathematicians there)… …I personally know some before-mathematicians-turned-linguists. It is a different discipline, but consider what they ((former-)mathematicians and nowadays professors of Mathematics) have to say.
I MUST add this:
http://xkcd.com/114/
I’m one of them