mentecuerpo wrote:(...)Could you please explain a little bit more about obtaining words from your text readings, not from the dictionary as a source? Do you put short phrases of the text on three columns? I don't think there is enough space when to write the sentences in three columns, so I am wondering how you accomplish this.
No, it's not as complicated as that. I fold a sheet of paper (because it is easier to manipulate then), and then I reserve something like 3 to 4 cm from the right edge for new words by drawing a line. New words go into that column, and I copy the text by hand to the rest of the space. If I fill up the margin before I have filled up the text area then the text is difficult, and I do something more with it - like copying difficult passages or interesting constructions once more, or in rare cases I write a translation myself. When I use a bilingual text I only check the dictionary at this point if I'm doubt the correctness of the translation (which unfortunately happens quite often, even with man-made translations)
I only put very short expressions in the margin since there isn't space for long ones there - sometimes I just underline the long ones with a wawy line in another colour, sometimes I write a hyperliteral translation between parentheses after the passage or I write the relevant passage in another color - there is simply no fixed rule for what I do. And I don't use the three-column wordlist format for memorization of long expressions - sometimes I do nothing (which isn't the smartest thing to do), sometimes I transfer them to a list which I then look through when I'm done with the text in question, but there are also no fixed rules here.
When I then transfer the new words to a wordlist I check the dictionary everywhere where there might be some doubts as to the correctness of the translation or where the grammatical class or gender or whatever isn't totally clear. And then I use the wordlist method just as I would do with words from a dictionary. However as I get better at a certain language the number of unknown words is supposed to fall so with time I get more and more of my new words from a dictionary.
I only stop copying text when I'm so far with a language that I hardly find any new words and I understand how each and every sentence is constructed. For the moment that would be the case for maybe seven or eight languages (Spivak's famous seven!). My main problem at that stage is that I rarely take notes when I'm studying extensively (like watching TV or reading stuff on the internet), so from that point on I'm just hoping that something will stick - like those who claim to learn language just from watching films and reading books.
An example (based on a text in Bulgarian about trilobites):
As you can see the number of new words is nicely aligned with the amount of text I have copied, so this text is at just the right level for me at my currrent level. I might be able to read the text and more or less get the main points without copying the whole thing, but then I would skip over all the murky points without even noticing them. When I copy text I have time to think about each sentence and make sure that I understand everything in it, including the grammar.
As you also can see I did include the expression "за разлика от" in the right column (although I forgot one з ). The translation into Danish is "til forskel fra" (literally 'to difference from' in English), and I didn't need a dictionary to figure that out. But that's about the maximal length of expressions that go into the right columns and - consequently - into my wordlists.
I checked a few words before writing the translation, like "пипало" for 'antenna' ('følehorn' in Danish, literally "feeling horn') and "хрылé" for 'gills' (because I wanted to know whether it was used in the singular) - and then I wrote "??" at the word "археолоцият(ите)" because it was translated as "archeologists" by Google translate, and I'm fairly sure that no fossil of any archeologist has ever been found in strata from the Cambrian so it must be something different - or it could just be a spelling error.