Learning Na'vi Vocabulary: The Easy Way

Started by MuneTompaTaem, January 07, 2012, 11:48:01 AM

Previous topic - Next topic

0 Members and 1 Guest are viewing this topic.

Taronyu Leleioae

#80
I know it's likely impossible to split it up.  But having just a verbs only course could be helpful and less intimidating.  Then you can associate supporting words (adv, adj) to the verbs as you begin to use them in practice.

Palulukan Maktoyu

well I've been thinking about that.
yes your right, it's impossible to split parts up out of an existing course. However, it is possible it have a course ignore words that you have already learned in another
Fkol syaw oeru Palulukan Maktoyu Ta'lengean

Twitter: https://twitter.com/navi_wotd

Herwìna

#82
I've been thinking of doing common words. In their base form. No adpositions or infixes. The problem is my Na'vi is so crap that translating even a 100-word list is really annoying because I'll have to do it with the dictionary 90% of the time. And it wouldn't even necessarily be a 100 words, because I don't know how many common words cover approximately one third of daily language in English. (That's every third word!) In Finnish it's 100. Then there could be another course that would cover approximately fifty percent of daily language - in Finnish that's 400 words. But I don't know the numbers in English, and cannot seem to find the info effin anywhere. >:(

And then there's the trouble of translating. *crawls into some nook and sulks*
Siyevop nga nìzawnong ayukmì, vaykrr oengeyä mefya'o ultxaräpun fìtsap nìmun.

Oe zawng
nga zawng
nìwotx awnga zawng
fte oeti zeykivawng

Ngal yamom fì'ut srak?!

Palulukan Maktoyu

Quote from: Herwìna on March 14, 2013, 06:10:52 PM
I've been thinking of doing common words. In their base form. No adpositions or infixes. The problem is my Na'vi is so crap that translating even a 100-word list is really annoying because I'll have to do it with the dictionary 90% of the time. And it wouldn't even necessarily be a 100 words, because I don't know how many common words cover approximately one third of daily language (that's every third word!) In Finnish it's 100. Then there could be another course that would cover approximately fifty percent of daily language - in Finnish that's 400 words. But I don't know the numbers in English, and cannot seem to find the info effin anywhere. >:(

And then there's the trouble of translating. *crawls into some nook and sulks*
well there is a good number section in the dictionary, AND from out of the threads is a LOVELY PDF for Na'vi numbers!
http://www.learnnavi.org/docs/The_Na%27vi_Number_System_v1.05.pdf
Fkol syaw oeru Palulukan Maktoyu Ta'lengean

Twitter: https://twitter.com/navi_wotd

Herwìna

I mean. In the Finnish language, the 100 most common words cover approximately 33% of the language people use daily. The 400 most common words cover approximately 50% of daily language. And I cannot find these numbers for the English language anywhere. I could of course do the common words courses based on Finnish, I'm p sure Finnish and English share a lot of common concepts, both being western countries and all, but I just feel slightly silly being so Finnocentric when the Internet is so English.
Siyevop nga nìzawnong ayukmì, vaykrr oengeyä mefya'o ultxaräpun fìtsap nìmun.

Oe zawng
nga zawng
nìwotx awnga zawng
fte oeti zeykivawng

Ngal yamom fì'ut srak?!

Palulukan Maktoyu

hey no biggie. just curious what's wrong wit your pc?
Fkol syaw oeru Palulukan Maktoyu Ta'lengean

Twitter: https://twitter.com/navi_wotd

Herwìna

#86
We think the charger is broken. I'll test it the next time I visit a friend of mine who has a compatible charger. Hopefully it is indeed the charger and nothing more srs :P

Anyway, the idea I had with the common words courses was to base them on the however many common words of English that cover that approximately 33% of daily language (excluding the words that in Na'vi are adpositions). And these words would be the first words one learns in Na'vi; words one uses a lot.

And then the same deal with words that cover approximately half of everyday language. However if I cannot find out how many of the most common words one needs in order to cover that 33,3% or 50% of daily language in English, then I guess I'll have to resort to the Finnish common words lists.
Siyevop nga nìzawnong ayukmì, vaykrr oengeyä mefya'o ultxaräpun fìtsap nìmun.

Oe zawng
nga zawng
nìwotx awnga zawng
fte oeti zeykivawng

Ngal yamom fì'ut srak?!

Ezy Ryder

Well... 50% is not much. You need 80-90% to understand a passage, and I've heard that You need 95% to be able to learn the rest out of context (I think it was Krashen who studied that, he called it the "comprehensible input"). For example, I've tested how many characters I understood in some Chinese Wikipedia articles, the percentage was often ~60%, but I only could figure out what were the articles about, not what did they say.
Since there aren't that many (correctly) written materials in Na'vi, and I have plenty of free time, I guess I could (try to) get a few longer passages (like Eywal Ngati Vewng, and such) and calculate how many occurrences certain words make. What do You think?
PS.: Obviously, there's no guarantee that I'd finish such a project.

Palulukan Maktoyu

Quote from: Herwìna on March 14, 2013, 06:29:17 PM
I mean. In the Finnish language, the 100 most common words cover approximately 33% of the language people use daily. The 400 most common words cover approximately 50% of daily language. And I cannot find these numbers for the English language anywhere. I could of course do the common words courses based on Finnish, I'm p sure Finnish and English share a lot of common concepts, both being western countries and all, but I just feel slightly silly being so Finnocentric when the Internet is so English.

Maybe if someone were to write a script or see if Toruk Makto ( admin ) can have the the board do this, it could lead to the most common words used. thusly give people a guide where to start?

http://forum.learnnavi.org/vocab-phrases/conversational-vocab-learn-these-first/

Surely there's a way for the board itself to take a copy of every post in nìNa'vi nì'aw then run it all through a script that counts the instances of words that are in the dictionary.
What you will find in there is that people will also sometimes make their own "Borrow words" when a word doesn't exist. most of the time if the rest can be interpreted then you can figure out what the unofficial word/s were meaning. but for the KNOWN one's surely there's a way for someone to automate that and have it run say once a month =) *just thoughts* *wishing i knew coding better now DOH*
Fkol syaw oeru Palulukan Maktoyu Ta'lengean

Twitter: https://twitter.com/navi_wotd

Herwìna

#89
Quote from: Ezy Ryder on March 15, 2013, 05:01:59 AM
Well... 50% is not much. You need 80-90% to understand a passage, and I've heard that You need 95% to be able to learn the rest out of context (I think it was Krashen who studied that, he called it the "comprehensible input").

In Finnish:
Quote100 most common words cover 35.1%
200 ... 42.9%
300 ... 48.0%
400 ... 51.7%
500 ... 54.9%
1000 ... 64.8%
2000 ... 74.0%
3000 ... 78.6%
5000 ... 83.5%
10000 ...89.4%

I think 50% is pretty good if it means that you only need to learn 400 words at first. It's a lot more realistic a goal for starters than to go straight for 5 000 words. 50% means approximately every second word after all! That would mean that until you have learned those 5 000-10 000 words, you would only have to check approximately every second word from the dictionary :P

Nevfur the less, I think it would be awesome if someone could calculate the occurrences of different words from actual Na'vi itself! From the actual users of the language! Oh, yeaahh~ totally needs to happen.

Quote from: Palulukan Maktoyu Ta'lengean on March 15, 2013, 05:15:07 AM
http://forum.learnnavi.org/vocab-phrases/conversational-vocab-learn-these-first/

SRANE. This is exactly the type of thing I would need!


Siyevop nga nìzawnong ayukmì, vaykrr oengeyä mefya'o ultxaräpun fìtsap nìmun.

Oe zawng
nga zawng
nìwotx awnga zawng
fte oeti zeykivawng

Ngal yamom fì'ut srak?!

Ezy Ryder

#90
I'm not saying that You shouldn't learn in the frequency order, I think it's a great idea but imagine the following situation: You want to read a book, let's say there are 700 words on an average page. Could You imagine looking a word in a dictionary 350 times just to get through one page? It wouldn't be fun, it'd be a chore (trust me, I've tried :) ).
About the percentages You've given, do You know how did they define a word, and what did they do the research on? If You know the word "tam", You can pretty much guess words such as "letam" or "teykiyevatsam", even though they might be written quite differently. Also, notice that the language You use with Your friends is quite different from the language used in courts, or quantum physics papers. Also, depending what Your goals are, if You want just to be able to speak, sometimes You can find a way around knowing some words. For example, before I've learned the word "fwew", as in "Oel fwew 'uot", I was just saying "Oel new rivun 'uot" or "Oel fmi rivun 'uot".
As for automatisation of the process... Even in Na'vi there might sometimes occur a problem such as, whether "kilvan" means "a river", or "would've aimed to". Na'vi is also pretty inflectional, so You'd have to write a script that would check each possible conjugation of verbs, each declension of nouns and so on... I don't know how hard would it be.
PS.: Some people don't write correctly in the NìNa'vi nì'aw subforum... I don't know if it's that reliable.

Tsyalatun te Eyktan Txuratu'itan

Quote from: Taronyu Leleioae on March 14, 2013, 05:47:24 PM
I know it's likely impossible to split it up.  But having just a verbs only course could be helpful and less intimidating.  Then you can associate supporting words (adv, adj) to the verbs as you begin to use them in practice.
It's not impossible to split - it's a matter of filtering the words from the dictionary according to the criteria of the course you want to create (which is how I generated the list of adpositions.)

However, there's an amount of manual work necessary to maintain a course, so I'm not looking to create lots of courses, but if people have a few well defined ideas which can be easily filtered from the dictionary data, I don't mind creating such a course.  I have the tools to be able to do this. :)

The other thing to be aware of is what memrise levels are supposed to be.  It's probably worth reading through the course creation guide at http://www.memrise.com/course-creation/ to see how the memrise stuff all hangs together - levels aren't limited to being "words N-N" - their titles can be set, and they can contain as many or as few words as is desired, and they can have words added in a random order from courses database of words.

For instance, inside a "Na'vi verb" course, we could have a level called "10 Most Commonly Used Verbs" followed by "15 Useful Verbs" or something like that - we don't need individual courses for each of these.  We could still have a single course - if we can split up adpositions into categories, we could have a number of adposition levels, a number of verb levels (split up how I described previously), a few nouns, etc.  We could order these levels so that levels 1-4 gets you a basic but workable vocabulary of a few nouns, a few verbs and a few adjectives/adverbs/adpositions.

The point I'm making there is... it's flexible, and there's a lot of different ways to arrange this without having multiple separate courses with their own individual databases that would need individually and independently updating.

Quote from: Ezy Ryder on March 15, 2013, 07:03:43 AM
As for automatisation of the process... Even in Na'vi there might sometimes occur a problem such as, whether "kilvan" means "a river", or "would've aimed to". Na'vi is also pretty inflectional, so You'd have to write a script that would check each possible conjugation of verbs, each declension of nouns and so on... I don't know how hard would it be.
I have such a program, and when I'm struggling, I'll either give it a single word or a full sentence to sort out for me.  It's not perfect - and suffers from the above problem, and because it only produces one word, it sometimes gets it wrong.  However, it's already fuzzily matching against the dictionary, so internally it is finding all possible matches and then trying to sensibly pick one from those.  If you're on the LN.org IRC or Teamspeak's Kelutral channel, you can interact with it by saying: "vrrtep: tt " and following that with your word or phrase.  I'd point out though that it isn't running on a fast machine though.  Here's an example:
<Tsyalatun> vrrtep: tt Oel yerom wutsot.
<vrrtep> Tsyalatun: oel (oe): pn. I, me, case: l | yerom (yom): vtr. eat, infix(es): er | wutsot (wutso): n. dinner, served meal, case: t

Which is "I am eating dinner."

However, this is not something I'm putting a great deal of effort into - I'd much rather spend that time learning Na'vi so I don't need it. :)

Herwìna

#92
Quote from: Ezy Ryder on March 15, 2013, 07:03:43 AM
About the percentages You've given, do You know how did they define a word, and what did they do the research on?

I haven't got the book the percentages are from, I copy pasted them from a blog post, but I'm quite sure they defined the word as the word in its basic form, without suffixes. Just like dictionaries do. Generally the research material for things like this are books, newspapers and other works written in the Finnish Literary Standard Language. Apparently there is a frequency dictionary of Finnish dialects as well though, and therein the most frequent words are quite different. (A lot of fillers.)
Anyway, I'm not saying that one should only learn 50% of a language's words and stop there. I just think that it would be immensely useful to learn the most common words first. First the 100 most common, then the 300 most common after those, and then, say, from 400 to 5 000 or whatever one wishes to do.

What I mean is, if by learning only 100 words (+ grammar) you can understand every third word, and by learning 300 more you can understand every second word, I'd say that's a quite good deal as opposed to learning 100 or 400 random words and in the end not understanding jack s***. With the common words you'd at least know that you will see these words frequently.

Obviously people will learn the most common words soonest one way or another, but I would still like to make a Memrise course for them at some point to help speed learning.
Siyevop nga nìzawnong ayukmì, vaykrr oengeyä mefya'o ultxaräpun fìtsap nìmun.

Oe zawng
nga zawng
nìwotx awnga zawng
fte oeti zeykivawng

Ngal yamom fì'ut srak?!

Blue Elf

Quote from: Tsyalatun te Eyktan Txuratu'itan on March 15, 2013, 09:30:21 AM
I have such a program, and when I'm struggling, I'll either give it a single word or a full sentence to sort out for me.  It's not perfect - and suffers from the above problem, and because it only produces one word, it sometimes gets it wrong.  However, it's already fuzzily matching against the dictionary, so internally it is finding all possible matches and then trying to sensibly pick one from those.  If you're on the LN.org IRC or Teamspeak's Kelutral channel, you can interact with it by saying: "vrrtep: tt " and following that with your word or phrase.  I'd point out though that it isn't running on a fast machine though.  Here's an example:
<Tsyalatun> vrrtep: tt Oel yerom wutsot.
<vrrtep> Tsyalatun: oel (oe): pn. I, me, case: l | yerom (yom): vtr. eat, infix(es): er | wutsot (wutso): n. dinner, served meal, case: t

Which is "I am eating dinner."

However, this is not something I'm putting a great deal of effort into - I'd much rather spend that time learning Na'vi so I don't need it. :)
The same sentence analysis is included in vrrtepcli by Tirea Aean (version 2.0 for Linux only - I played with this some time ago; don't know if there's any progress...)
QuoteWhat I mean is, if by learning only 100 words (+ grammar) you can understand every third word, and by learning 300 more you can understand every second word, I'd say that's a quite good deal as opposed to learning 100 or 400 random words and in the end not understanding jack s***. With the common words you'd at least know that you will see these words frequently.
Well, I read somewhere, that you need to know just about a few hundreds of words to be able to use language (of course daily usage only, not reading books or science etc.)
IMO it is true - I'm not sure how many Na'vi words I know (less than 50% I guess), but still I mostly understand main idea. For simple usage you really do not need too much.
Oe lu skxawng skxakep. Slä oe nerume mi.
"Oe tasyätxaw ulte koren za'u oehu" (Limonádový Joe)


Tsyalatun te Eyktan Txuratu'itan

Quote from: Blue Elf on March 15, 2013, 01:44:34 PM
The same sentence analysis is included in vrrtepcli by Tirea Aean (version 2.0 for Linux only - I played with this some time ago; don't know if there's any progress...)
Yes, I have that too. :) I know Tirea improved it recently with some of my ideas for matching recently, but I don't know whether anything further has happened.  However, as this is off topic for this thread (let's move this discussion to the vrrtepcli thread!) I won't say too much more about it here.

Taronyu Leleioae

For Tsyalatun since the PM system is blocking sending PM's to you (from me)?

On the adposition memrise, wondering if the database will be updated and/or can be tweaked.  I'm thinking certain words might have double entries or the earlier memrise system couldn't identify certain spellings.  ì vs i.  I've tested it and know that mi vs mì is allowed as is mikam instead of mìkam.  However it does know that ilä is incorrect.  (All the ä words seem to be ok.)

Obviously there are some missing adpositions in the list.  sko, nuä, kam, kay, lisre, and talun.

Tìtstewan

Quote from: Taronyu Leleioae on March 20, 2013, 01:45:14 PM
For Tsyalatun since the PM system is blocking sending PM's to you (from me)?
Quote from: Toruk MaktoThe latest security patch for SMF seems to have broken the sql query for PM "display name" checks. The workaround is to click on the "Find Members" link next to the "To:" field and do a wildcard search for the member (for example, to find "tsu'tey", do a search for tsu*tey and then select the member from the results). This will place the member's login ID into the To: field which should work.
;)

-| Na'vi Vocab + Audio | Na'viteri as one HTML file | FAQ | Useful Links for Beginners |-
-| Kem si fu kem rä'ä si, ke lu tìfmi. |-

Taronyu Leleioae

Tried this method prior to posting here but still came up with the same name and blocked the PM...  It just doesn't like me...  :(

Tsyalatun te Eyktan Txuratu'itan

Quote from: Taronyu Leleioae on March 20, 2013, 01:45:14 PM
For Tsyalatun since the PM system is blocking sending PM's to you (from me)?

Please see the reply from Tìtstewan - I'm not aware of any forum settings blocking anything.  An alternative way to contact me is via email, which should be showing a learnnavi.org address.

Quote from: Taronyu Leleioae on March 20, 2013, 01:45:14 PM
On the adposition memrise, wondering if the database will be updated and/or can be tweaked.  I'm thinking certain words might have double entries or the earlier memrise system couldn't identify certain spellings.  ì vs i.  I've tested it and know that mi vs mì is allowed as is mikam instead of mìkam.  However it does know that ilä is incorrect.  (All the ä words seem to be ok.)

Obviously there are some missing adpositions in the list.  sko, nuä, kam, kay, lisre, and talun.
The adposition course is looked after by Herwìna, and I've already passed the list to her, and volunteered to provide a file that she can use to bulk-load those six.  I hope that it can be updated soon, but I believe Herwìna has issues with her computer at the moment.  (See older posts in this thread.)

Taronyu Leleioae

Quote from: Tsyalatun te Eyktan Txuratu'itan on March 20, 2013, 02:00:32 PM
...An alternative way to contact me is via email, which should be showing a learnnavi.org address.

FYI that I had tried the published work around prior to posting, but no joy.  And your profile has your email "hidden".  Hence why I went to "Plan C" and posted here.