Iversen's second multiconfused log thread

Continue or start your personal language log here, including logs for challenge participants
User avatar
Iversen
Black Belt - 4th Dan
Posts: 4768
Joined: Sun Jul 19, 2015 7:36 pm
Location: Denmark
Languages: Monolingual travels in Danish, English, German, Dutch, Swedish, French, Portuguese, Spanish, Catalan, Italian, Romanian and (part time) Esperanto
Ahem, not yet: Norwegian, Afrikaans, Platt, Scots, Russian, Serbian, Bulgarian, Albanian, Greek, Latin, Irish, Indonesian and a few more...
Language Log: viewtopic.php?f=15&t=1027
x 14962

Re: Iversen's second multiconfused log thread

Postby Iversen » Fri Aug 27, 2021 1:15 am

Today I have added 2x2 languages to my vocabulary asssessment (or 'wordcounting') campaign, Slovak and Greek. Yesterday I did two tests for Polish, and there was quite a large percentage difference - 45 % known with the small Francophone Oxford dictionary, but only 31 % with the big Fat Germanophone Pons. But Pons contains at least 35.000 headwords so even 31% amounts to an estimated vocabulary of 12.300 head words - passive of course. Is that assessment realistic? OK, the test is based on my own personal evaluation of each word, but it is not totally arbitrary.

To put a a word in the 'known' category the optimal situation will of course be that I remember having seen (and understood) it somewhere - either in a text or a wordlist (speech is not an option). But with derivations I may use a more lenient criterion, namely: is it totally obvious that the derived word looks like this way, given that the main word looks like that? If it isn't obvious. for instance because there is another ending or an unexpected spelling then I would just call it 'guessable' and place it in the middle cathegory, where I also place words which I remember, but whose meaning isn't totally clear. Or compounds where I know the parts, but whose the meaning is different from what you might have expected.

So today I added Slovak and Modern Greek, and both with two dictionaries because of my experience with Polish. Slovak gave 53 % with a micro Langenscheidt, but only 40 % with a somewhat larger Lingua (which I bought in Bratislava in 2019) - so here the big dictionary gave the highest score, but for a good reason: by a freakish stroke of luck I escaped the pages filled with verbal prefixes! So according to this result I should know 18.300 headwords - but alas, I don't. Even 40 % is rather high, but that's what I got with a more even mix of page contents (but notice how the dominant colour switches from green in the first column to red in the second one, just because I have hit upon territory dominated by the prefix "od").

And Greek gave 42% with a Greek-Danish Pataki (author: R.Hesse) and 52 % with an old Langenscheidt from before the orthography reform that removed the aspirations and two accents. Both dictionaries have between 35.000 and 40.000 headwords, so maybe I actually do know 15.000 Greek words now, thanks to my wordlists. It's a fairly high number, but since I now can read the Rhodes thing with just a handful of unknown (and unguessable) words per each page it can't be totally wrong.

And now it is goodnight time, and I have placed the Russian guide to Delfi (Дельфы) at my bedside (plus a minuscule Berlitz dictionary) - the Greek guide to Rhodes is almost finished, and I have put it in a pocket on my jacket so that I can read the rest in the bus. And tomorrow I'll try to come up with some techniques to memorize verbs in particular. I remember that I once made a table with all the derivations of some verbal root, including those dictated by the opposition between perfective and imperfective) - something like that but in a setup where more wordroots are included. If words with prefixes are a problem, then the reason must be that the prefixes take away some of the 'visibility' of the roots. And then I automatically think about my Indonesian dictionary where the root of any compound is indicated, and where a number of derivations are listed for the bare root words (the same for Malaysian and Pilipino by the way, but I don't study them - yet).

SLovak-Wordcount-26-08-2021.jpg
You do not have the required permissions to view the files attached to this post.
6 x

User avatar
Iversen
Black Belt - 4th Dan
Posts: 4768
Joined: Sun Jul 19, 2015 7:36 pm
Location: Denmark
Languages: Monolingual travels in Danish, English, German, Dutch, Swedish, French, Portuguese, Spanish, Catalan, Italian, Romanian and (part time) Esperanto
Ahem, not yet: Norwegian, Afrikaans, Platt, Scots, Russian, Serbian, Bulgarian, Albanian, Greek, Latin, Irish, Indonesian and a few more...
Language Log: viewtopic.php?f=15&t=1027
x 14962

Re: Iversen's second multiconfused log thread

Postby Iversen » Fri Aug 27, 2021 10:26 pm

And now a few more wordcounts. For Esperanto I got 55 % with the laudema senjoro Bick's Esperanto-Danish dictionary that contains an estimated 28600 words. 55% would correspond to around 15700 headwords - and I don't find that unrealistic, given that for each 'rootword' there is a healthy family of derived words so the number of wordfamilies would of course be considerably lower. In other words, you get a large vocabulary with a limited investment of time and effort. Other languages also have derivation, but in a less systematic fashion and with more irregularities so there is more stuff to learn there. The dictionary I used listed a lot of plant names and other obscure words, typically based on Latin, and I didn't know all of these (not even in Danish), whereas I knew the vast majority of the 'normal' words.

From Esperanto til Indonesian, both languages that are heavily dependent on derivation mechanisms og which therefore have a lot of their words organized into word families. But I don't count families, I count headwords, and I was actually worried that my Indonesian vocabulary had fizzled away because I haven't used the language much. OK I have done wordlists and studied some texts, but not used it extensively at all. But the result is OK: 39% known words from my Tuttle dictionary. My total estimate for this dictionary clocked in at 14.000 headwords, even though I skipped abbreviations for institutions and things like that plus all compounds that weren't printed in bold types (my total estimates in 2013 and 2104 were much lower -heaven knows why). Given this number og words in the dictionary the estimate for known words has a minimum value of some 5500 headwords, which sounds realistic considering my mediocre, but above zero reading skills. A larger dictionary would of course give a higher estimate, which is why you shouldn't trust absolute figures, but look at the percentages. NB: I wrote yesterday in this message that the total number of words was 9400 words - nonsense! I had written a wrong number of pages. And then I had to revise the conclusion - it actually does seem that I have learn quite a few words since 2014. :P

INDO: Saya mengunjungi Indonesia pada tahun 2002 dan mengunjungi Sulawesi, Bali dan Jawa. Di Sulawesi saya melihat penyembelihan sapi untuk pemakaman, di Bali saya melihat antara lain taman burung dan beberapa candi dan gunung berapi, dan di Jawa mengunjungi Borobodur, Yogyakarta dan Surabaya. Saya belum mempelajari bahasanya saat itu. Lain kali saya mengunjungi negara itu, saya hampir pasti tidak bisa berbicara dengan penduduk setempat, tetapi mudah-mudahan membaca koran. Dan "koran" tidak berarti kitab suci Islam - itu dieja Al-Qur'an, Qur'an atau sederhana Quran.

F2233b04_palmleaf-writings-muzium-Yogyakarta.jpg
You do not have the required permissions to view the files attached to this post.
4 x

User avatar
Iversen
Black Belt - 4th Dan
Posts: 4768
Joined: Sun Jul 19, 2015 7:36 pm
Location: Denmark
Languages: Monolingual travels in Danish, English, German, Dutch, Swedish, French, Portuguese, Spanish, Catalan, Italian, Romanian and (part time) Esperanto
Ahem, not yet: Norwegian, Afrikaans, Platt, Scots, Russian, Serbian, Bulgarian, Albanian, Greek, Latin, Irish, Indonesian and a few more...
Language Log: viewtopic.php?f=15&t=1027
x 14962

Re: Iversen's second multiconfused log thread

Postby Iversen » Sat Aug 28, 2021 8:02 pm

And yeah, we are still in the midst of the wordcounting season.

From Indonesian I proceeded to Latin, using a New College dictionary which I picked up in a bookstore in Manila. It has an estimated 17.000 headwords when you exclude the combos and examples in the articles (which I didn't do in my first counts in 2013 and 2014), and with an estimated 58% known words I get an estimated vocabulary of 9.700 words - and the best thing is that I know how to inflect them. But it is status quo compared to 2014.

And from Latin to Romanian, where I have used my trusty Teora with around 44000 words (which only has one defect: it doesn't mark which verbs use an infix - probably because it was meant to be used by native Romanians). And now it is definitely calm sea and prosperous voyage (the name of an ouverture by Mendelssohn-B): 68% known and an estimated vocabulary of 30500 words! One factor behind this result is that we suddenly got a Romanian teacher at the Romance Institute where I studied French during the 70s. In the beginning we were 5 'pupils' (including two or three teachers), but the last two years I got solo classes two hours every week - 100% in Romanian. But now it is five years since I last visited Romania and three or four years since I last had an opportunity to speak it, so maybe I could have spent my time better on starting out learning some Slavic language - never mind which one.

Next: Italian with 60% from an old Italian-Danish Gyldendal with around 38200 words - i.e. an estimated vocabolario of at least 22.700. Am I happy? No! How can I know as many words in Romanian as in Italian when I don't have a Romanian TV channel and rarely find anything to read for pleasure in Romanian (on paper or on screen)? By the way: I do have a newer and thicker Gyldendal dictionary, but according to my estimates there are not more words in it than in the old one from 1980. So to get a second opinion I took my old monolingual Garzanti with almost a thousand pages and opened it on page 428 - result 21 known words and 3 dubbiosi in the first column (88%, zero unknown), but back to the realities with 67% after 5 columns. However that's more like what I had expected from my Italian - it shouldn't be lower than my poor mistreated Romanian!

ItalienskDramatik.JPG
You do not have the required permissions to view the files attached to this post.
4 x

User avatar
Iversen
Black Belt - 4th Dan
Posts: 4768
Joined: Sun Jul 19, 2015 7:36 pm
Location: Denmark
Languages: Monolingual travels in Danish, English, German, Dutch, Swedish, French, Portuguese, Spanish, Catalan, Italian, Romanian and (part time) Esperanto
Ahem, not yet: Norwegian, Afrikaans, Platt, Scots, Russian, Serbian, Bulgarian, Albanian, Greek, Latin, Irish, Indonesian and a few more...
Language Log: viewtopic.php?f=15&t=1027
x 14962

Re: Iversen's second multiconfused log thread

Postby Iversen » Sun Aug 29, 2021 2:22 am

And now to the Iberoromance languages, first Catalan. I have used the Catalan-English book in a set from Enciclopèdia, and it has somewhere around 40.000 headwords (latest estimate: 42.000), and with 62% I should in theory know at least 26.200 words. Actually my percentage was higher in 2014 (66% - and 71% with a dictionary from Larousse), but on the pages I culled this time there were more words, so my estimated vocabulary has never been bigger. Statistics is a weird creature...

For Castellano I used a big fat dictionary from Gyldendal with an estimated 44.300 headwords, and my 59% would then correspond to a vocabulary of at least 26.000 - just a few thousand words more than in 2014 so basically status quo. But rather than adding more individual words I should focus on idiomatic expressions - and on regional differences.

Spansk.JPG

And finally Portuguese, which is special in one way: I spent two weeks in Brazil in 2019, and inspired by that I did two wordcounts in 2020, based on one small dictionary (73% known) and one bigger one (67%, corresponding to a modest 33.000 known headwords). This time I used the bigger one of the two, a monolingual monster from Porto Editora Flumense with 1058 pages which I bought in Lisboa as far back as 1981, just after I had left the university - but then I stopped all language learning for 25 years. I only started really to study the language in 2006, a few weeks before a trip to Cabo Verde, but found it surprisingly easy to learn, probably because I already knew some Catalan and some Castellano. The current inventory gave 58%, slightly lower than last time, but with more words on the pages I used - and apparently more difficult words. So the estimate of known words is almost the same both times - around 33.000 headwords from the 55.000 or so in the book. But as usual you should look at the percentages rather than the absolute figures. If I made statistics based on my Bratli Spanish-Danish dictionary (with around 200.000 headwords) then you would see the absolute numbers for Spanish soar.

I look forward to doing French and the Germanic languages tomorrow - and maybe also Irish and Albanian to see some low numbers.

F5913b02_a-cidade-de-Natal-no-Brazil.jpg
You do not have the required permissions to view the files attached to this post.
3 x

User avatar
Iversen
Black Belt - 4th Dan
Posts: 4768
Joined: Sun Jul 19, 2015 7:36 pm
Location: Denmark
Languages: Monolingual travels in Danish, English, German, Dutch, Swedish, French, Portuguese, Spanish, Catalan, Italian, Romanian and (part time) Esperanto
Ahem, not yet: Norwegian, Afrikaans, Platt, Scots, Russian, Serbian, Bulgarian, Albanian, Greek, Latin, Irish, Indonesian and a few more...
Language Log: viewtopic.php?f=15&t=1027
x 14962

Re: Iversen's second multiconfused log thread

Postby Iversen » Mon Aug 30, 2021 2:09 am

FR: J'aurait été assez surpris si mon inventaire de mots-clé français n'aurait pas dépassé les 30.000 .. [EN] but it did - 72 % from a Fransk-Dansk Gyldendal with an estimated 51.700 words, which suggests a lower level for my French vocabulary at around 37.000 words. The words I didn't know came mainly from two sphères: biological names (flowers and small worms etc.) ... and names for maladies, which the French insist on calling something in Ancient Greek. And I have not studied that part of the Greek vocabulaire. If I had hit upon an area with a lot of institutional names I might also have had a problem, but I didn't.

And now to the Germanic languages. I take them with the newest so let's start out with Afrikaans. I have a couple of small dictionaries, but only one midsize Tweetalige Skoolwoordeboek from Pharos which is printed on lousy paper, but apart from that it is quite good. I got 54% known words out of around 36500 headwords, which is quite OK. And Dutch gave 69%, which with a big fat Gyldendal with around 44.000 headwords (and a weight of 1½ kg) yields a vocabulary estimate of something like 30.400 words. I skipped Low German because I don't have a decent dictionary and continued to High German, also here with a big fat Gyldendal as the basis for my enquiries. Result: 74% known words from a dictionary with more than 80.000 headwords. Or a vocabulary of 60.000 words - and now the alarm bells should toll because that's far more than the average native person is supposed to know, according to the experts.

I also began a test based on a similar-sized English dictionary with maybe 50.000 headwords, probably more - and then realized that I knew almost all words in the first three columns I checked, and then it would be absurd to continue. As you may remember my technique is that I open the dictionary at a random point and select a random column, and then I take equidistant pages, same column until I have enough words.

OK, if a red Gyldendal isn't big enough then get the heavy artillery then - Webster's Unabridged in the 1979 edition (4.250 g, 2127 pages and several hundreds of thousands of headwords)! Here I did actually run into some words I didn't know, but over long stretches I knew the majority of the words. And now the alarm bells are banging like hell because it seems that the sky - or rather the size of the dictionary used - is the limit. I can't know more than 100.000 English words - that's preposterous.

The only logical solution could be to count word families instead of single headwords - for instance I got 21 known words starting with "Energie-" from just one column in the German dictionary, and that makes my sums go berserk. But dictionaries aren't subdivided according to families - the nearest thing to this would be my Indonesian Tuttle dictionary, where several derivations are listed for every 'root word', but even there it would not be feasible to switch to family counts. Or maybe continue as usual, but divide the results with a factor that reflected the ratio of headwords to word families for that language (and that dictionary!). But you would have to know that ratio, and you don't. It's almost like assessing your active vocabulary - you know that it is a fraction of your passive vocabulary, but you don't know the ratio, and there is no easy way to measure it.

The internet site Testyourvocab claims in its faq that

"There are many different factors that go into measuring someone's vocabulary size—for example, do you count "quick" and "quickly" as two words or just one? After all, "quickly" is just a simple and predictable derivation of "quick." Depending on how different questions like these are decided, vocabulary tests can give widely different results. We take a conservative approach, and count the number of headwords (not derived words) which you are estimated to know in a standard dictionary. In the end, what really matters is not your absolute number, but rather your score relative to others who take the same test, no matter how the test is put together."

Well, I basically count the things that are printed in bold and starts an article (my definition of headwords), but in some dictionaries there are also bolded words or word combinations inside the articles, and sometimes these could just as well been made into as headwords, whereas they in other cases just are examples. And some headwords are clearly just standard derivations - like "quickly" from "quick". But I do count them, and generally I don't count bolded items inside dictionary articles - but there are exceptions, depending on the style of the dictionary.

The testyourvocab site published in 2013 in its blog this overview (for English of course, not for German or French, let alone Danish):

• Most adult native test-takers range from 20,000–35,000 words
• Average native test-takers of age 8 already know 10,000 words
• Average native test-takers of age 4 already know 5,000 words
• Adult native test-takers learn almost 1 new word a day until middle age
• Adult test-taker vocabulary growth basically stops at middle age
• The most common vocabulary size for foreign test-takers is 4,500 words
• Foreign test-takers tend to reach over 10,000 words by living abroad
• Foreign test-takers learn 2.5 new words a day while living in an English-speaking country


I actually took the test some years ago, and as far as I remember I landed in the upper 30 kilos range - and another online test gave me more than 50.000 words. However in spite of my work with wordlists I can't believe that I know more words than elderly ladies who read novels and do crosswords (probably the most word-savvy segment within the native population). I remember that I once ran through a Danish dictionary, and there I knew almost all the words. Those elderly ladies could probably do the same, and I would venture the guess that a German lady from this population segment also would know most of the words in even her fattest dictionary.

So the most logical conclusion is that absolute vocabulary sizes are bullshit, and you should focus on the percentages. If you know half of whatever you are counting then it would probably also be half of what you didn't count - and half of what you might have counted if you had used a bigger dictionary. I broke off my studies of the monstrous Webster (and avoided using my equally monstrous Spanish Bratli), but if I sampled an insane lot of pages from those then I might discover whether the use of monster dictionaries really lead to lower percentages, but that'll be an exercise for a later date - not now. I might also want to do counts for things like Swedish and Icelandic and Ancient French and Irish and Albanian and maybe a selection of languages I haven't studied yet, but again ... that'll have to wait, I'm not going to do those exercises now.

And this concludes the inventory series.

Kunst172.JPG
You do not have the required permissions to view the files attached to this post.
7 x

User avatar
Iversen
Black Belt - 4th Dan
Posts: 4768
Joined: Sun Jul 19, 2015 7:36 pm
Location: Denmark
Languages: Monolingual travels in Danish, English, German, Dutch, Swedish, French, Portuguese, Spanish, Catalan, Italian, Romanian and (part time) Esperanto
Ahem, not yet: Norwegian, Afrikaans, Platt, Scots, Russian, Serbian, Bulgarian, Albanian, Greek, Latin, Irish, Indonesian and a few more...
Language Log: viewtopic.php?f=15&t=1027
x 14962

Re: Iversen's second multiconfused log thread

Postby Iversen » Mon Aug 30, 2021 10:02 pm

Iversen wrote:And this concludes the inventory series.


Ahem - no, it didn't. After I had finished my last rant I continued to ponder the abbyssal opposition between my conception of headwords (basically article starters printed in bold types in a dictionary, with some exceptions and some addenda) and that of the Testyourvocab guys (basically the same, but reduced to heads of language families). And then it struck me that I actually have the materials to discuss the problem in the notes I have made to calculate my vocabulary sizes in different languages.

First I finished the count based on the Webster monster, and the results were not quite so aberrant as I first had expected: 52% known words, 26% dubious cases and 33% definitely unknown words. That would give a vocabulary size of at least 66000 headwords from a total of around 127.000 headwords with their own articles (far less than I had expected based on the promises, but that's what I got). The next step would then be to reduce the lists to just one word from each family, and this is what happens:

I started out with 83 known, 23 dubious and 53 not known headwords (according to my understanding of the word 'head word'). If I do the clean-up the really brutal way (for instance keeping "oblique", but eliminating "obliquity") then there are 24 known ones left, 3 guessable ("Rudesheimer" about wine from Rüdesheim", "Rabelaisian" and "Rudolphine tables" about astrological tables ordered by the notoriously superstitious emperor Rudolph II) plus only 8 totally unknown ones. That actually makes the matters more mysterious, with 73% known out of a mere 32.700 supercleansed wordfamilies, corresponding to roughly 24.000 known families (from a dictionary of more than four kilos - ridiculous!).

But why should "proboscis monkey" go down the drain just because the word "proboscis" is a recognized pater familias? And of course the wise Testyourvocab guys didn't demand that - they quoted the case "quick"/"quickly", so they probably just intended to eliminate typical affixes that are used for common wordclass conversions - though I wonder whether they would accept "obliquity" as an independent word or just a derivatum from "oblique". OK, using my common sense I reincluded the cases where I see unforeseen transformations based on a certain root. And this opens for a flood of new headwords, including "obliquity", but not "obliqueness", and "rabbit fever" (from "rabbit"), but not "rabbiter" for a hunter that shoots rabbits, nor "rabbitry" - both noted as guessable in my original assessment. But I accept "paddler" (from "paddle") for an oldfashioned robber because I definitely couldn't have guessed that this word existed. Likewise "paddle ball" and "paddle bean". There is a great deal of arbitrariness in this process, but I ended up with 39 known, 6 borderline cases and a full 35 unknown items (half of my original total sum), or as percentages: 49%, 8% and 44%. The estimate for total number of families in the book has gone up to 64.000, and the funny thing is that even though the majority of the readmitted items are unknown, my estimated known vocabulary has gone up to 31000 items because the total number of recognized items has gone up.

And then you can ask whether this is more in line with the recommandations of the Testyourvocabians (and behind them the guru in this field, Paul Nation) - it is still far more than they would have expected from a foreigner who never has lived in an Anglophone country, but now I'm using their own definitions. By the way Denmark is not seen as an Anglophone country, but we are sliding in that direction (even though the Netherlands apparently have overtaken us in that respect), and personally I read and hear and think as much in English as in Danish - especially since I stopped working 5 years ago.

My recommendation is still to look at the percentages. The percentage for known items was 52% with all headwords included, but after a thorough Testyourvocabization process it slided down to 49% - mainly because there are a bewildering lot of animals and plants that are named after the rabbit and a bewildering lot of maladies that are pneumatic in some way. And who knows all of them by heart?

Webster.jpg

Epilogue: I couldn't resist the temptation to take the test again, and it must be at least 5 or 6 years since I did it last time so it should be permissible by any means. The somewhat disappointing result can be seen below - conclusion: I still have a lot to learn (but so do the native Anglophones!)...

Testyourvocab_2021.jpg
You do not have the required permissions to view the files attached to this post.
5 x

User avatar
Iversen
Black Belt - 4th Dan
Posts: 4768
Joined: Sun Jul 19, 2015 7:36 pm
Location: Denmark
Languages: Monolingual travels in Danish, English, German, Dutch, Swedish, French, Portuguese, Spanish, Catalan, Italian, Romanian and (part time) Esperanto
Ahem, not yet: Norwegian, Afrikaans, Platt, Scots, Russian, Serbian, Bulgarian, Albanian, Greek, Latin, Irish, Indonesian and a few more...
Language Log: viewtopic.php?f=15&t=1027
x 14962

Re: Iversen's second multiconfused log thread

Postby Iversen » Fri Sep 03, 2021 8:40 pm

I have visited my mother again this week, and I brought along my Rhodes guide plus a minuscule Micro Langenscheidt to help me over the worst potholes in the road. And then what happened: Wednesday at. 15:00, right after a Danish TV program, all channels disappeared. I called the cable company to hear the prospects, and they said it was a problem that had hit a sizeable area, and the repair might be done in an hour, or it could last until 22 (where the repair workers apparently go home). The signal wasn't back before the following morning, so we listened to cassette tapes instead, and I read my guide to to Rhodes .. AND FINISHED IT!!!!! I have been writing about that goddam book for at least two weeks, and the problem is that when I use a book as goodnight reading then I read maybe one page and fall asleep. But Wednesday evening I got through forty pages in Modern Greek with spurious dictionary aid. I read the last few pages in bed, but still wasn't sleepy, and then I continued with the dictionary, which was small enough to keep in one hand.

And the day after I made a wordlist with German words - mostly to get some genders mended, but also some whose meaning was slightly different from what I remembered. And of course I also watched some TV in English, German, Swedish and Norwegian in those periods where the cable company deigned to provide us with the entertainment that we expected. But I did more garden work including painting a raised garden bed (plus some cooking, sewing and other menial tasks) than I did language studies - which is quite normal during family visits. And then today I visited a zoo with my sister. But now I'm at home and can study to my heart's delight, except that I intend to glue a few hundred photos into an album, paint a painting, listen to a string of composers whose names start with B and .. well, you get the idea. I may be retired, but I'm just as busy as when I still worked.

GR: Χρησιμοποίησα το μικρό ελληνικό λεξικό για άσκηση σκέψης σε ξένη γλώσσα. Σκέφτεστε μια σκέψη, συναντάτε μια τρύπα ή ασάφεια στο λεξιλόγιό σας και στη συνέχεια την αναζητάτε στο λεξικό και σκέφτεστε περαιτέρω. Είναι μια άσκηση που απαιτεί ένα πολύ μικρό (και ελαφρύ) λεξικό και συγκέντρωση - όχι κάτι που γίνεται ενώ παρακολουθείτε τηλεόραση - αλλά είναι ένας καλός τρόπος για να εκπαιδεύσετε τη σκέψη στη γλώσσα -στόχο. Παρεμπιπτόντως, το λεξικό ήταν πάρα πολύ μικρό: έλειπαν τουλάχιστον οι μισές από τις προβληματικές μου λέξεις, αλλά μπορώ να συμπεράνω από αυτό ότι το λεξιλόγιό μου καλύπτει τα κεντρικά μέρη της γλώσσας - όταν μπορώ να ανταγωνιστώ ένα λεξικό με 15.000 ελληνικές λέξεις, δεν μπορεί να είναι εντελώς άθλιος.

Kunst102.JPG

Apart from that: I have had a couple of dreams in English. One was situated in a town called Manchester, but it didn't look like the English city. In the other I didn't get a place name (or I have forgotten it), but in the part I remember I was going to visit a museum. First I walked through a shopping arcade (where I was tempted to enter a couple of bookshops that looked more like some I have seen in Madurai in India than anything British), and then I walked through an underpass that was directly nicked from the connection between Estação do Oriente in Lisboa and the neighboring shopping center, and when I came out the other side there was another shopping complex - but at last I found an tourist information desk, and beside it I found the stairs up to the museum on the second floor - but then I woke up. Since I don't delve into dream interpretation much, I don't have a large background stock of other people's dreams, but my impression is that there is a lot of talking going on plus meetings with other human beings. In the dream I just mentioned I saw very few people and didn't speak to anybody during my perambulations, but I read some booktitles in the bookstore, and I read some texts at the tourist information and greeted the lady behind the desk - but that's all. No conversations at all, just running around looking at things. I'm an incarnate tourist even in my dreams...

F6010b06 _ Madurai bookstore.jpg
You do not have the required permissions to view the files attached to this post.
6 x

User avatar
Beosweyne
Yellow Belt
Posts: 80
Joined: Fri Sep 18, 2020 4:59 am
Languages: En (L1), Fr (A2), Af (A1), Ar (beg.), Zu (beg.)
Ancient:
La (A2), Gr (A2), Eg (beg.)
Language Log: https://forum.language-learners.org/vie ... 15&t=16342
x 342

Re: Iversen's second multiconfused log thread

Postby Beosweyne » Sat Sep 04, 2021 2:16 am

Iversen wrote:Oh no, I knew that Google Translate has problems with Latin, but this is ridiculous. It is hard to point to anything which the poor beast hasn't misunderstood. Unless I have forgotten the last bit of whatever I once knew about Latin the passage should be translated as follows:

Me: Addendum: today (Friday) I woke up from a dream partly in the Latin language. I have scrutinized a Latin document where certain passages (places) seemed dubious to me, and then another person behind me indicated with his finger that "2." definitely wouldn't be comprehensible to the Romans - you would need (to use) 'second'. And once awake again, I strived to recall my dream in Latin. I haven't worked on Latin texts recently, and so it is a wonder to experience this language in my dreams.


It's uncanny how just days after you wrote this, the team at Google finally installed a massively improved Latin translator! For example I input your final paragraph and got this fine translation: Addendum: hodie (Veneris) e somno excitatus sum partim in lingua Latina. Documentum Latinum examinavi ubi dubia quaedam loca mihi videbantur, et deinde alius homo digito post me indicatus est "2." certe Romanis comprehendi non posset - opus esset (utar) 'secundum'. Et iterum evigilans, somnium meum Latine recolere studui. In textibus Latinis nuper non elaboravi, ideoque mirum est hanc linguam in somnis experiri.

Latin-to-English translation also seems to work great now.
3 x

User avatar
Iversen
Black Belt - 4th Dan
Posts: 4768
Joined: Sun Jul 19, 2015 7:36 pm
Location: Denmark
Languages: Monolingual travels in Danish, English, German, Dutch, Swedish, French, Portuguese, Spanish, Catalan, Italian, Romanian and (part time) Esperanto
Ahem, not yet: Norwegian, Afrikaans, Platt, Scots, Russian, Serbian, Bulgarian, Albanian, Greek, Latin, Irish, Indonesian and a few more...
Language Log: viewtopic.php?f=15&t=1027
x 14962

Re: Iversen's second multiconfused log thread

Postby Iversen » Sun Sep 05, 2021 11:35 am

I'll definitely try it out - the translation above looks promising. It can't have been my feeble voice in the far back alone that finally pushed the sleepy giant to mend its Latin translator, but it was obvious that something had to be done.
1 x

User avatar
Iversen
Black Belt - 4th Dan
Posts: 4768
Joined: Sun Jul 19, 2015 7:36 pm
Location: Denmark
Languages: Monolingual travels in Danish, English, German, Dutch, Swedish, French, Portuguese, Spanish, Catalan, Italian, Romanian and (part time) Esperanto
Ahem, not yet: Norwegian, Afrikaans, Platt, Scots, Russian, Serbian, Bulgarian, Albanian, Greek, Latin, Irish, Indonesian and a few more...
Language Log: viewtopic.php?f=15&t=1027
x 14962

Re: Iversen's second multiconfused log thread

Postby Iversen » Wed Sep 08, 2021 6:27 pm

Two major things have happened today: I have started a new painting (and that will diminish my study time for a couple of weeks), and I have done some work on my Irish for the first time in many months, using an article from Wikipedia about somebody named Cú Chulainn from Ulster. But I soon realized that my grammar skills had eroded too much, and then I searched for the notes I had prepared for making green sheets - but I didn't find them. I did however find some printouts from the splendid "Gramadach na Gaeilge" and I have some grammar books on my shelves, so all is not lost. But Irish is a weird language that doesn't easily fit into my green sheets routines. For instance you have to separate the things that happen at the beginning of nouns from the typical endings for the five declensions and three or four or five cases - otherwise you will end up with a bewildering array of combinations that simply can't be fitted into a one sheet format.

It's funny to think that I once had reached the level where I could construct sentences that might not be totally wrong in at least one dialect or in the official standard, which everyone in Ireland in principle has to learn (and then they hate it). Take for instance the socalled Dative case, which only is used after certain prepositions (with or without an article). Here is a short quote from the article about the articles (not substantives) in the Anglophone (short) Gramadach (there is also a Gramadach in German, if you prefer that):

• E/L: In the standard, there is generally a choice between lenition or eclipsis after most prepositions + article .
e.g.: ar an mbord/ar an bhord = on the table
• E: In Munster and Connacht, one eclipses after most prepositions + article.
e.g.: ar an mbord = on the table
• L: In Ulster, one lenites after most prepositions + article.
e.g.: ar an bhord = on the table
(.................)


Luckily there is a simple practical rule further down:

In the so-called "standard", by most prepositions, lenition or eclipsis are "a choice". After don, den, sa(n), lenition always follows. For words with s- there is the t-prefix only preceding feminine nouns, even if the speaker would otherwise prefer eclipsis. d,t are not lenited, but also not eclipsed (words with s-,d-,t- are also treated as in Connacht).


But all the examples in the English version are in the singular - what about the plural? Well, no direct answer, but there is a big table that shows just how different the dialects are in the singular - but also that the Standard allows you to choose freely what to do, and according to that table there is neither lenition nor eclipsis in the plural in any of the main dialects.

So... lenition after don, den, sa(n),, at least in the singular, t before s with feminine nouns - and otherwise let the few remaining Irish speakers fight among themselves over what mutations you could have applied if you had bothered. And for those of you who still haven't let the Irish language spoil your state of blessed equanimity: lenition roughly means that you write a 'h' after a consonant, and then it sounds totally differently (although 'fh' simply becomes silent), and eclipsis means that you add another letter in front of the original one which then isn't pronunced - but you still write it, and if it for some reason it should be written as a capital letter it is the original one which is capitalized - not the newcomer (as in "ar an gCraobh Rua" - in the Red Branch, i.e. in one of the three royal houses of king Conchúr Mac Neasa). And sometimes you put a t- or -h or -h before a vowel or an s, just to make things more complicated. Nothing is simple in Irish!

The following example from the article I mentioned earlier can illustrate some of these phenomena:

Cú Chulainn.
I luarthlitríocht na hEirann, ba é Cú Chulainn (...) priomhlaoch na nUltach sna laochscéalta Rúraiochta agus an priopmhphearsa sa seanscéal Táin Bó Cuailgne. (...) Ba é an gaiscioch ba cháilúla é ar an gCraobh Rua.


my hyperliteral translation: C.C: In early-literature of-the Ireland, was he C.C. (..) main-hero of-the Ulster in-the main-tales (of) Rúraiochta (whatever that is) and the main-person in-(the) old-tale T.B.C. (...) Was he the warrior most famous he in the Branch Red.


RU: Ранее я упоминал, что намеревался использовать русскоязычный путеводитель о Дельфах в Греции как ночного чтения, и это прекрасно работает - я читаю максимум одну страницу, и потом спию. Вот почему я тоже начал немного читать, когда сыпаюсь утром. В качестве вспомогательного средства я использую свой карманный словарь из Berlitz, но ситуация известна по греческому - большинство слов, которые я хочу проверить, в книге нет. Некоторые из них, однако, потому что на самом деле былы греческие имена, просто написанные кириллическими буквами - но мне часто удается искать, прежде чем я это замечаю. Однако это не относится к таким случаям, как "святилище", поэтому здесь я должен угадать - альтернативой было бы встать и взять словарь большего размера, и честно говоря - не труждаюсь вставать. Кстати, я обнаружил, что у меня есть эта книга тоже по-гречески и по-итальянски, и я посмотрел в итальянской версии несколько раз, но проще принять размытость.

SP: Por cierto, he visto las noticias en español en TVE y parece que a los hoteles españoles no les va tan mal - en un lugar el 100% de las habitaciones estaban ocupadas (he olvidado el cual)- pero al parecer todavía tienes que usar una mascarilla. Dos buenas razones para no viajar aún ...

F3447b03_Delphi.jpg
You do not have the required permissions to view the files attached to this post.
4 x


Return to “Language logs”

Who is online

Users browsing this forum: No registered users and 2 guests