By the way, the Linguistics room is really the nicest room on Stackexchange. Everybody here is always super civil. Even I get the niceness virus when I am here.
2
English is great and very nice too, of course, though some (benign) heated arguments sometimes take place. But this is like an oasis of peace and civilisation, ahhhh enjoys the atmosphere.
French has that rule that whenever a masculine entity is part of a group, the whole NP will default to masculine as far as agreement goes. My native language, German, also defaults gender to masculine, whose forms are usually the least marked. However, I wonder how common defaulting mixed-gender ...
@Alenanno Soo you're designing a car in 3D?? Cool!
The most important things is that I start with a decent vocabulary in all the Romance languages because of Latin.
That effect should happen for you guys too.
Except that the evolution from Latin to its daughter languages is a bit easier to intuit than from sister to sister (I only need to trace the path from Latin to Portuguese, for example), while you need to trace from Portuguese to Latin to Italian, hypothetically.
By that I don't mean that you actually follow those developments, but rather that the connection between Latin and Portuguese is shorter than between Italian and Portuguese, and so the phonological changes are perhaps a bit easier to intuit.
And there is grammar too: a lot of the grammar in Italian is easy to follow based on Latin, and also in Portuguese.
I wish I knew Proto-Germanic...
On the other hand, your Portuguese is much better than my Latin will ever be, so Italian is probably considerably easier for you than for me.
Some time ago I wanted to find some regularities in the phonetic adaptation of Russian loanwords in Dolgan (a tiny Turkic language in north-eastern Siberia), and somewhat to my own surprise ended up using quantitative methods, or in fact coming up with one of my own.
There's this Journal of Quantitative Linguistics. In the end they turned my paper down with some strange (to me, at least) arguments, but a year later they invited me to a Festschrift, and damn, I have no idea what or how to write about
@KamilS I am a bit ambivalent towards Chomsky; partly because I don't know that much about his work, partly because I know he has done some useful stuff.
It's an Arabic name. It comes from k-m-l 'perfect' (hence eventually English camel as well :) The name went to Ottoman, and from there to Polish, and perhaps other languages as well. In Indian, it might be directly from Arabic or perhaps via Persian, I don't know.
They spend much effort in testing and discovering laws in languages. Not the phonetic laws we all know and love, but statistic laws of the Zipf kind (the most frequent words are also the shortest and most polysemic) &c.
I think it's not very well tested yet, but I know that they counted the number of occurrences of certain defective verbs in old German texts, and the fit was quite good.
Let's assume for a moment the law is properly tested and indeed describes how features are born and die in languages.
So my observation is this:
if I counted the number of borrowings from English in Polish and on this basis, calculated the appropriate coefficients, I would obtain a curve the end of which would be in the future and would apparently show when the influence of E. on P. will start fading and eventually die out.
Now, this kind of influence is usually explained by the general prestige of a language, which in turn is obviously one of the effects of a country's political and military domination.
So: would I have thus calculated when the US falls?
horizontal is time and vertical is frequency in texts
So you can see that if the law is correct, a new feature (e.g. an irregular verb becoming regular) begins slowly, then becomes quite rapidly more and more popular, and than slows down until the old form dies out completely
If it were true, it should also describe the number of loanwords from a certain language.
In a statement such as "I like this more than that.", what is the term (is there a term?) used to refer to the this and/or that?
In the above example, the this/that could easily be replaced with a noun. However, the idea could also be expanded to something like "I thought that I would be more t...
Let me get this straight first. The curve only describes the process of adaptation of one average word, doesn't it? So it should be the average of what happens to all foreign words that are used for the first time in a sentence in the host language, measured from the time the word is first so used?
Forgive me if I misunderstand: I recount this on memory.
And another thing: the curve only describes the proportion between one foreign word and one native word, right?
Or does it describe the proportion of speaker who use the word?
Well, as far as I understand, we can't really say that it describes anything as yet because the empirical grounding is not strong enough. Let's just assume for the sake of this conversation that it is.
It's proposed as a curve which describes change in language:
nevermind if it's one form being replaced by another, the proportion of analytic : flectional, frequency of loanwords or anything. Change in language in the very general sense.
This particular curve in fig. 2 is limited to one form, but it's the general shape that counts, not whether it's a little wider or taller.
I'm having difficulty in understanding how one would apply this law in practice. Consider the use French words in Dutch. I think it rose and fell during the Middle Ages, then started to grow faster and faster from ca. 1650 onwards; I'd say growth reached a peak somewhere in the 19th century, then it started to recede very slowly; the decline was accelerated a great deal by each World War, and it is now at quite a low level, let's say around the 1700 or 1600 level.
To which part of this historical development should the formula be applied?
I don't think this would be a problem. I suppose one would need to slice the entire history into several chunks, each of which is described by a different set of coefficients.
Just a thought: perhaps one of the main strengths of this formula is that it was invented for use where one form took the place of another, where the two forms are nearly 100 % synonymous.
(So suppose you were going to measure the total frequency of English words in Polish, you would collect data from the past 100 years, plot them on a graph, and see if the curve fits in anywhere?)
I went to a lecture on stylometry recently. Apparently, by examining the list of most frequent words in a text, and comparing it to preset specimens, it is possible to determine the author of the given text with up to some 99% accuracy.
(That could be manageable (except that the data collection would be a huge amount of work, unless you have a properly prepared corpus ready where someone else has done the job of marking words English v. Polish.).)
The 99% accuracy is only achievable in perfect conditions. But in more standard conditions, you can get about 80% accuracy for e.g. Latin, which I would say is still quite impressive.
Here's where I come back to the answer: when I hear a statement like this, I think what you said: Not at all sure this is true. But the guy did quite a lot of tests, and how can I argue with that?
He had a corpus of say 30 texts by five different authors. He took another text and asked the computer to determine which of the 30 texts it is closest to, based on the list of five or so thousand most frequent words.
It all seems like magic which I'm not really prepared to believe. But I'm not ready to argue with numbers, either.
The 100k words I only gave as an example. I can't really remember how many words it was. I can remember he had an English corpus of around 30 19th c. novels.
And by "random", does that mean you let the computer pick one word from a random place in the database, then another word from another random place, and so on?
If that is so, then consider this. An author's style changes as he writes. He may be in a different mood during the first few chapters, because the story is still fresh for him. He may use different words as the plot develops, when people go to a different location (I am presuming novels, why not). He may write a single chapter in a short time span.
This would mean that the closer two pages are together in the book, the more similar they will be in language. This in turns means that you sample only a limited part of his style if you only take the first few chapters, while you would get a slightly more representative sample if you also sampled later chapters.
This is just the first thing that pops up in my mind: perhaps this isn't really relevant.
If I understood you correctly (of which I am not sure), then it would indeed be better to take random words spread throughout the novel than taking all words in the first few chapters.
Taking a completely random sample frees us from "this chapter just happens to happen in a dungeon which is all dark, wet and gloomy, and that's why he uses these words so often".
I'm not sure. The lectures were primarily aimed at raising interest.
Well, I for one didn't figure that out.
are you familiar with the concept of ZPG, Zero Player Game? A funny thing for those who'd rather not spend too much time playing, and yet would like to play.
If you're interested, I've started playing godvillegame.com a few days ago. If you decide to join, please send me your email at [email protected], and I'll send you an invite. (Not that you need it, but we'll be friends in the game :)
Basically, you create a hero (so it's like RPG – that you know, I presume?), for whom you are god.
But your hero has a mind of his or her own (admittedly, not a terrible lot of it ;), and you can leave him or her for days or weeks unattended and he'll play the game on its own.
You can limit yourself to just checking on your hero from time to time, and becoming more active e.g. on weekends.
I did Erepublik for about a year. A fantastic game but horribly time consuming. I just found this and am happy I'm not spending my whole life in it :) Laters!
> So, you just got a personal heroine! It's good to have a follower who will pray to you and do stupid things in your honor, isn't it? You don't need to control her directly, because she's smart enough to figure things out on her own. All you need to do is to visit her once in a while and show some signs of life, because the heroine needs a god to believe in. Your heroine will write down all the important events of her life in the diary, so you will always know what she is up to.
That doesn't make sense.
The God doesn't need to visit the creature, it's the other way around.
> I think I get it now. I have the great honor of being chosen for an inscrutable purpose by a supreme being. Thank you, my Mistress. I will not let you down!
:D
Bah, the heroine got married in a few minutes and I am single after many years.
In a statement such as "I like this more than that.", what is the term (is there a term?) used to refer to the this and/or that?
In the above example, the this/that could easily be replaced with a noun. However, the idea could also be expanded to something like "I thought that I would be more t...
lɪŋˈgwɪstɪks
lɪŋˈgwɪstɪks