Iversen's Guide to Learning Languages (version 3b)

All about language programs, courses, websites and other learning resources
User avatar
Black Belt - 4th Dan
Posts: 4870
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 15359

Re: Iversen's Guide to Learning Languages (version 3b)

Postby Iversen » Sun Jan 31, 2016 1:52 pm

4.4. Copying by hand (scriptorium)

One of the techniques proposed by Alexander Arguelles is the so-called "scriptorium", which he has demonstrated in a video and described as follows on HTLAL:

I have made a short video to demonstrate the proper form for transcribing languages by hand as I do in my “scriptorium” exercise. In order to do this properly, you should:

1. Read a sentence aloud.
2. Say each word aloud again as you write it.
3. Read the sentence aloud as you have written it.

The whole purpose of this exercise is to force yourself to slow down and pay attention to detail. This is the stage at which you should check all unknowns in grammars or dictionaries, although that would have been too tedious to show in the video.

I once commented as follows on this technique and on his shadowing technique in a HTLAL thread called "Language learning series video reviews":

As far as I can see the common factor in both your shadowing and your scriptorium methods is that you transform something that normally is done in a purely passive way into something active, namely listening and reading. In your case it is done in a very active way, cfr. your recommendation of speaking loudly and walk at a brisk pace. When I do similar things I do them silently and sitting in my chair, which suits me better - the basic idea is the same, but there is a distinct difference in style, which probably can be blamed on temperamental factors.

However there is another, more fundamental difference: apparently you have moved from a more conventional way of using dictionaries and grammars to a position where you actually do produce such books yourself, but essentially you assume that the absorption through shadowing and 'lectoriuming' is the main ingredient of language learning. The main difference between your methods and the so called 'natural learning' is that you don't need somebody to speak to because you use preexisting materials, and that you prefer heavily structured learning instead of relying on the hazardous nature of immersive experiences or even chance encounters with suitable natives.

I am convinced that both two methods work, but I have found that shadowing doesn't work for me - I stop listening if I speak - and while I actually do a lot of copying by hand, I do it invariably without vocalization.

Text copying may sound as the final sign of mental madness, but it actually has a purpose: to slow you down and make you notice every detail in the text you have chosen - and which of course has to be shorter than the texts you can rush through when you just read for fun. I always write by hand because I feel it gives me better contact with the words, but younger people may feel more at home with a keyboard. The source can be mono- or bilingual, but because the goal is to catch every meaning of every word and every grammatical quirk I prefer using bilingual texts in my weaker languages - otherwise I would loose too much time on looking things up. I used folded A4 sheets to make them more handy, and I always leave a margin to the right for new words with a preliminary translation. These words (or in some cases short expressions) are supposed to be entered into wordlists later, and there I can check the dubious cases. The result looks like this:


The right column is almost as tailormade to enter a wordlist (cfr. chapter 2.6), but you should be aware that it is more heterogene that the words on a page in a dictionary. In some cases I have just copied the suggestions of a translation, in others I have quoted my own guesses based on the context. I prefer writing the standard form as used in dictionaries (i.e. nominative singular for substantives, the infintive for verbs in most languages - but sometimes the 1.person singular present tense) - but sometimes I make errors. At the end of the day the only cure would be to look all words up in the dictionary, but .. well, it would actually become the end of the day, and it would spoil my 'drive'.

However one good habit with this exercise is to look the unknown words up before you actually start plowing through a sentence or paragraph. Then you will feel that the text is much easier, and you won't loose the thread because you have to look words up all the time. Another tip: If a passage has been particularly bothersome then run through it twice. First time you will be solving a rebus, but at the second passage the text will almost certainly have become comprehensible.

As I said above you ought to make sure that all translations in the column with new words are correct and ready to go into a wordlist, and this may have to wait until you are making that list. But let's face it: checking the words in a dictionary while transferring them to a wordlist can be a chore because the words are spread all over the alphabet (online 'pop up' dictionaries might solve this problem). So once I am advanced enough also to do dictionary based wordlists I permit myself the luxury of skipping the most dubious words among those I collected from the texts. It is better to learn five words from a dictionary and skip one from a text because I would had to look it up first.

There is one reason more: in those cases where you run through the text twice you will automatically be doing a repetition of the unknown words from round one. If you can't understand each and evry word the second time you know you have a problem, and you can do something to memorize the problematic words. So to summarize: when I first started out doing those text copies- cum-study I was adamant about running all unknown words through a wordlist, but I have become more relaxed about it because I know that I have the alternative of doing dictionary based wordlists. Instead I have become more interested in getting a 'flow' in my work with the texts. But this exercise is still an intensive one because the goal is to understand everything and remember as much vocabulary as possible from each text.

With weak languages or with particularly complicated constructions in better known ones I have sometimes added hyperliteral translations, either between the lines or after a paragraph. The text below is taken from the bilingual guidebok "Your key to Korça / Korça çelësi turistik", and it is one of my first sheets with an Albanian text:


My own translation was made on the basis of the English version in the book, but it is (hopefully) closer to the original Albanian text. And it is partly in Danish because I'm Danish, but as you can see there are also single words and passages in English because both my dictionary and the translation in the book are in English. This illustrates clearly that the goal isn't to make a polished translation. The important thing is to fixate your understanding of the text and getting the sentence patterns of of the target language under your skin.

You may also notice that I wrote the Albanian text again (without translation). If you have struggled with the text the first time then it should feel like a breeze to write it once more, and this time you should actually be able to understand it. But of course this only is worth doing when you are new to a language or the text is unusually hard to deal with. I have a rough rule of thumb: if the column with the new words fills up faster than the one with the text then it wouldn't be a bad idea to repeat the text.

But there is also a variant of the repetition layout which works best with fairly strong languages, namely retranslation. This means that you read a sentence in the source and at first just write a literal or hyperliteral translation on the paper, not the original sentence. You are supposed to be able to supply that from your memory or by reconstructing it, but at least in my concept you must end up with exactly the same sentence as in the source. The result will be something like the Ilya Frank format from the preceding chapter.

I tried something like this long ago, but dropped the exercise because the 'reconstructed' sentences rarely became identical with those in the source - and then I felt I just as well could do ordinary translations from a base language into my target language. I took retranslation up again when I heard a video by Luca Lampariello, where he spoke about his own use of retranslation. But in the video he ironically advocates exactly the thing that originally had made me drop the technique, namely doing the retranslation long after you see the source. For me the short interval is important because it has the effect that the original sentence so to say feels like your own creation - and in that way it not only supports your efforts to activate the target language, but it also boosts your selfconfidence.

PS: I have been asked whether I keep my text copies forever. No, I don't. I stash them, but when the piles get too high I throw them out. It is highly unlikely that I return to my own copies, and if I return to the original texts later it is more rewarding to look at them with fresh eyes than it is to waste time on my old errors. I have the same attitude to my wordlists except if they are part of a 'word list campaign', where I might want to do statistics and maybe even belated repetitions. It's much better to start a new wordlist from scratch than it is to wallow in your old scribblings.
You do not have the required permissions to view the files attached to this post.
3 x

User avatar
Black Belt - 4th Dan
Posts: 4870
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 15359

Re: Iversen's Guide to Learning Languages (version 3b)

Postby Iversen » Sun Jan 31, 2016 2:06 pm

5. Fifth part - How to deal with speech and writing

5.1. Shadowing and chorusing

It would be absurd to discuss listening without taking ProfArguelles' Shadowing method into consideration. Among the first references to it in the HTLAL forum is the following one from 2005.

Let us define:

Echoing = listen, then repeat
Shadowing = listen and repeat simultaneously

I believe that echoing can be an invaluable technique when you are working one on one with a phonetician who can correct your pronunciation and intonation. However, when you are working on your own, you really have no way of determining the accuracy of your rendition. When you shadow through earphones, however, you should at least instinctively and immediately perceive when you are "flat," and with constant repetition of the same material, you will begin to automatically compensate by modifying your pronunciation.

Three years (10 March 2008) later he made a video, and in the introduction to this video he wrote these lines:

In order to shadow most effectively, it is important to observe three points:

1. Walk outdoors as swiftly as possible.
2. Maintain perfectly upright posture.
3. Articulate thoroughly in a loud, clear voice.


In the video, you see me initially shadowing without looking at a book, then while looking at one. You will want to shadow without a book when you are in the very initial stages of language study, focusing on phonetics only (= “blind shadowing”), before you study any individual lesson, and then again finally after you have worked through your lessons thoroughly.

In one of his videos on accent formation he explicitly states that shadowing gives much better results than echoing. Nevertheless echoing is the technique that is used not only in classrooms all over the planet, but also in the kind of language labs I remember from my study.

Unfortunately the method didn't function for me - I simply stopped listening while I spoke (so I'll never become a good simultaneous interpreter).

Chorusing would normally be done in a class room setting with a teacher. Chorusing would normally be done in a class room setting with a teacher. In a discussion on HTLAL someone mentioned the Swede Olle Kjellin, and he not only participated in person in the discussion, but gave also some references to literature about accent reduction techniques, including the article "The Role of Deliberate Practice in the Acquisition of Expert Performance" by Erickson, Krampe and Tesch-Roemer, which also tells about the changing views through history on the concept of training. Let's first quote some general remarks about training:

There is a relatively widespread conception that if individuals are innately talented, they can easily and rapidly achieve an exceptional level of performance once they have acquired basic skills and knowledge. Biographical material disproves this notion. In their classic study of expertise in chess, Simon and Chase (1973) observed that nobody had attained the level of an international chess master (grandmaster) "with less than about a decade's intense preparation with the game" (p. 402). Simon and Chase estimated that the amount of knowledge a chess master has acquired is comparable in size to the vocabulary of an adult native speaker of English. It takes normal individuals approximately a decade to acquire this vocabulary (...)

The article briefly mentions the role of feedback:

Studies focusing on speed of performance tend to use easy tasks, where highly accurate performance is rapidly attained, and subjects are instructed to increase the speed of performance while maintaining the high level of accuracy. Under these conditions subjects' performance improves monotonically as a function of the amount of practice according to the power law (...) In the absence of adequate feedback, efficient learning is impossible and improvement only minimal even for highly motivated subjects. Hence mere repetition of an activity will not automatically lead to improvement in, especially, accuracy of performance (..).

.. but then also points out a flaw in some of the research:

When laboratory training is extended over longer time periods, studies show that providing a motivated individual with repeated exposure to a task does not ensure that the highest levels of performance will be attained. Assessment of subjects' methods shows that inadequate strategies often account for the lack of improvement. (...)

In the HTLAL thread I wrote the following:

As language learners we can essentially try to absorb an humungous amount of unrelated pieces of grammatical information, or we can try to find patterns. And it seems logical to try the same thing with phonetics. You shouldn't primarily spend time on teaching people how to pronounce specific words, but rather go for series of words that point to some general problem - which for instance could be that all /i/ sounds are too flat, or that some ending shouldn't be pronounced clearly unless you are spelling the word out to a foreigner.

Unfortunately that's difficult to do in a group setting. As a teacher in that situation you have to send out a cloud of informations, which may or may not be relevant for any specific pupil. And as a result of hearing so much irrelevant noise they don't get the advice which would be relevant for them personally. And just letting them hear some excelent native speech can't garantee that they pick up what they themselves make differently.

Teaching languages in a group setting is as dangerous as operating patients by throwing knives and scalpels around the operating theater in the hope that they hit someone in just the right spot. I am not sure pronouncing nothing but my own address in Danish (or for that matter Swedish) could keep me attentive and responsive for a whole week, but I'm fairly sure that 1 hour used intelligently is better than 10 spent without aim and purpose in the company of 9 random and uninterested co-students.

And Kjellin answered as follows:

I can't agree less, having experience of both teaching and operating! :) The target pronunciation is the same for everyone, the tools for attaining it (the speech apparatus) is the same for everyone, the effect of listening to one another's road to success can't be underestimated, etc. This is an effective listening practice, where they learn to detect the "center of gravity", as it were, for when the pronunciation is "correct", and when it deviates too far. Actually, "correct" pronunciation is *never* a point, as in IPA, but more like electron clouds of allophones with statistical centers for each phoneme. Just like any quality control procedure with specified tolerance ranges. During the process, they practice a zillion times to also get that electron cloud feeling motorically through what I call the "audio-motor memory", which will enable them to monitor their own speech and compensate for various external factors just as native speakers can do; so-called compensatory articulation. Which is automatic and subconscious for natives, but rather non-existent for L2 learners having suffered ordinary language classes with minimal pronunciation practice.

In my 4 decades of experience, chorus practice alternating with individual exercises in groups of 8 people or more outperformes one-to-one instruction by several orders of magnitude. Also it puts less strain on the teacher and students, so we can do it all day without fatigue.
Quite magically, 7 or less is too few. It's a clear step function at 8. (Indeed, I was told the same thing by amateur choir singers - when they meet to sing and are fewer than 8, it's less fun.)

Unfortunately that's where I quietly leave through the backdoor. Even if it may be possible to move the pronunciations of a group in the right direction through a massive collective effort it's something I wouldn't want to experience.
1 x

User avatar
Black Belt - 4th Dan
Posts: 4870
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 15359

Re: Iversen's Guide to Learning Languages (version 3b)

Postby Iversen » Sun Jan 31, 2016 2:57 pm

5.2. Listening like a bloodhound

Let me understand this: you can pick out every syllable, but they don't combine into words? Not even if you actually know the words? Well, now I'm going to be a bit provocative: one reason for being unable to understand spoken words from for example TV is that people are obsessed with understanding what is being said, when the basic problem may be that they can't even parse the stream of sounds. Parsing is the process of dividing the stream of sounds into syllables and words on the fly, and to follow a stream of speech you must be able to perform the operation automatically. If you try to understand everything at this stage then each unknown word will block you, and you will loose the next couple of sentences. Therefore 'listening like a bloodhound follows a trail' (or 'active listening') can sometimes be a help, even though it goes right against all the other good advice about listening to films, music etc.

Here you deliberately don't care at all about the meaning, but try to parse the stream of sounds into sentences and words, following the speaker 'like a bloodhound on a trail'. If you have a transcript the whole thing of course becomes easier, but if you have at least some feeling for the sounds of the language you can do it without support. As your vocabulary grows and you become adept at reading the language, you will discover that the meaning of the spoken form suddenly becomes crystal clear even though you haven't tried to understand it. The meanings of those words you actually do know suddenly will pop up in your mind, but before you know enough of them they won't combine automatically into some grand meaning of the whole thing.

When I have listened to something in this way I have sometimes seen a string of grey words running across my inner field of vision, though blurred so that only the shape of the words is seen. At other times I just listen for the borderlines between words, phrases and sentences. In both cases you should focus on changes in intonation, pauses, words you have heard before and common endings and affixes, which means that all those things will seep into your mind. Then some day the meaning appears, and you will have learnt to understand that language without passing through the perilous stage of translation.

This technique will of course demand your utmost concentration, and therefore it is counted as an intensive method, which shouldn't be applied for hours on end.

Now this is my advice to people who can't understand spoken utterances. As I wrote above, you mostly will get exactly the opposite advice if you search for tips on how to learn to listen. For instance at the site "Skills you Need", where number 9 of altogether 10 principles go like this:

Listen for Ideas – Not Just Words
You need to get the whole picture, not just isolated bits and pieces. Maybe one of the most difficult aspects of listening is the ability to link together pieces of information to reveal the ideas of others. With proper concentration, letting go of distractions, and focus this becomes easier.

So what's wrong? Who is the false prophet?

In my opinion the beginner who can't follow a recorded speech or TV program has the choice between two strategies: 1) going for the meaning and only getting the gist, based on isolated clues in the form of single words or the context of the speech, or 2) going for the way isolated sounds are tied together for form words and phrases, but leaving the general meaning for later. The first is relevant if you are teleported to a foreign country and have to survive, but no. 2 is the intensive method that saves your skin if you want to understand everything later.

The other 9 principles of this site are concerned with sensible things like noticing non-verbal cues and stop talking to yourself. As Mark Twain wrote: "If we were supposed to talk more than we listen, we would have two tongues and one ear." Those who start out early contacting speakers of their target language (like Benny 'the Irish Polyglot') may be lucky to meet individuals who speak clearly and in simple sentences to help them. And these learners obviously don't have the option of disregarding the meaning while they parse the messages of their conversation partners.
2 x

User avatar
Black Belt - 4th Dan
Posts: 4870
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 15359

Re: Iversen's Guide to Learning Languages (version 3b)

Postby Iversen » Sun Jan 31, 2016 3:11 pm

5.3. Training phonetic awareness

All languages have a phonetic level and a phonematic level. The phonetic level is the level with the actual sounds emitted by people who speak, while the phonematic level is a simplified version of this, where the concrete sounds are assigned to a small number of building blocks called phonemes. Two sounds from a given language belong to different phonemes if you can find at least one word or word combination with different meaning in that language (a so-called minimal pair) which only differ by having one or the other of these sounds. If you can't find such a pair then even fairly different sounds can be assigned to the same phoneme. Such sounds are then called allophones.

This means that if you transcribe a spoken passage using phonemes you can't be absolute sure how it should sound. There may be some rules of thumb which tell which allophones to use in a given context, but at the end of the day you have to learn the actual pronunciation by listening to native speakers. If you know something about the sound system of a given language you will have an easier time identifying the phonemes, which is necessary because it is the combination of the phonemes that produce meaningful elements like words. The phonetic level may contribute to the meaning at a more global level, but the phonemes are defined in such a way that they suffice to describe all the meaning bearing units of a language.

If you read a standard book about the phonology of a language you will be presented with a lot of details about the way the phonemes are represented as allophones in different environments, and you will learn about the relevant mouth positions and probably also something about dialectal variations. But we don't all study academic books about phonology. They are notoriously dry and hard reading - especially if you don't already have the 'tone' of a certain language ringing in your ears. So instead we just rely on the lists of sounds in textbooks or language guides to help us along, and that is not quite enough.

The main problem is that we have to do a phonematic reduction of the information to get the meaning of the things we here, and in this process we deliberately disregard the details and variations and intonations and everything else that isn't absolutely necessary for the choice between different meanings. And this situation is made worse by the way such lists use sounds of the base language of the book to represent the sounds of the target language - even in cases where the similarity isn't too overwhelming. And because the whole thing has to be written down the idiosyncrasies of the base language make the lists even harder to deal with.

Not to single any particular language out, but English is a particularly problematic base language because it has such a pathetic writing system and so many local variations. For instance you have to write "oo" to indicate a flat /u/ sound, and this makes a transcription of for instance Bahasa Indonesia harder to read than the original (this language has an almost perfect phonetic writing system). For instance a guidebook ("buku petunjuk") is rendered as "booˑkoo peˑtoonˑjook" in Lonely Planet Indonesian from 2006.

So phonetic awareness means going back to the original sounds and listening to them. You need to know which words were intended in order to know which phonemes are intended, but at this level the global meaning of a sentence itself isn't important. If we speak about intonation you do however need to consider larger blocks of speech and you need to know the grammatical structure, but even here the general meaning constitutes a background information rather than your study object.

The most efficient technique I have used for studying minute differences - below the phonematic level - is problematic because it demands concentration and absence of outside disturbances. The idea is that you find a short snippet of speech which you can repeat again and again while you try to write down exactly what you hear. Ideally you should do this in a recognized transcription system like IPA, but then you have to learn that first. I use homebrewed 'alphabets' and keep telling myself that this is OK because I don't have to communicate the results to others - the important thing being the listening process itself.

Of course I have listened to speech before and done some parroting of the usual kind, but seeing how long time it takes me really to hear what is said is shocking because it tells me how much I have missed earlier - even when I have replayed a certain passage several times. It is like walking from platform 9 to 10 at King's Cross, and suddenly you discover that there is something weird in between. Let's take a few concrete examples:

Is príomhoide teagaisc mé
/Ish pri·vɔdjə tægæshχmæ/
Is! principal (of)teaching I
I am a teaching principal

The first line is an original Irish phrase, which I put into the excellent Irish speech synthesizer Abair, using the Cabóigín voice. Below in the second line is my impression of what the result sounded like, written in my own totally idiosyncratic sound writing. Oh yes, I ought to learn the IPA system which is the norm for transcriptions in the scientific world, but it would take me a lot of time to get as comfortable with that as I am with my own homegrown systems - and we are not studying Irish here, so the details aren't important here. The third line is a hyperliteral translation as described in part 2, and the fourth line contains a 'normal' translation into standard English.

Another example: I took a sentence from the homepage of Burgers Zoo in Arnhem, the Netherlands, and then I let four different voices from the Acapela Box read it aloud. The result, which I published on HTLAL in 2012, clearly shows how much variation there is between different speakers - a lesson which language learners and teachers definitely should take seriously:

U kunt tegenwoordig een combi-kaart voor Burgers' Zoo en het Nederlands Openluchtmuseum kopen

/e könt teχəvo·Rdeχ e·n kɔmbikαRRt fo·R BöRhɔrs so· æn ət ne·dɔlansə-öpənleχtmyseåm kåubə /
/y könt teχəvo·RRdeχ e·n kɔmbikαRRt fo·R BöRgɔs so· æn-ət ne·dɔla·Rns opɔlöχtmyseöm ko·pə /
/y könt teχəvo·Rdeχ en kɔmbikαRt fo·R BøRgɔs so· æn (h)ət ne·dɔlansə opɔ(R)löRχtmyzæöm kåbə/
/y könt teχəvɔ·Rdeχ e·n kɔmbikαRRt fo·R BøRχɔRs so· æn ət ne·dɔla(R)ns opɔ(R)löRχtmyzæjöm kåupə/

It takes surprisingly long time to catch the subtleties in even a few sentences, but after you have done it you are better equipped to hear what really is going on in genuine speech. And then you can translate the rough indications in dictionaries etc. to the actual sounds which you have heard with your own ears. And yes, I ought to use some internationally recognized standard system for the transcription task, but I feel more at home with my own homebrewed systems - which to boot differ from language to language.

But few people do this exercise, and instead they rely on an automatic osmosis mechanism which means that the correct pronunciations seep into their system, where they replace erroneous preconceptions based upon standard orthography and even the supposedly phonetical spellings in dictionaries etc. And this doesn't always work out as intended because you hear what you expect to hear and not what really is said. You may be one of the exceptional individuals who can remember and repeat any spoken sequence with 100% precision, but don't count on it.

Some readers may at this point ask themselves how I can trust a speech synthesizer. Well I can't, but much depend on the quality of the synthesizer, and it is practical to have a machine that accepts everything you tell it to say, and who can do it again and again using exactly the same pronunciation

OK we have discussed single sounds, but at the other end of the scale you find the suprasegmental intonation patterns, which almost never are indicated in language learning materials - you are (once again) expected to absorb them from the things you listen to through some kind of unconscious osmosis. Again there is one thing you can do to make this process partly conscious: get some text in both a spoken and a written version and print out the written version with large spaces between the lines. Now mark up two things: stress patterns and tone level (in practice you may have to do one thing at a time).

Write a wavering coloured line for the tone level - and check this against the type of sentence. Is it a question? An order? A triumphant conclusion? Sometimes there a jumps in the tone level which seemingly don't have any grammatical explanation, but they may still be necessary for the authentic 'nativelike' sound.

This is a useful exercise with any language, but especially important for languages heavy use of sentence 'melody' - like Swedish and Norwegian. And one of the funny experiences is of course when something unexpected happens in the pronunciation. Just as an experiment I took a question from a Swedish Youtube video and put it into Acapela, and then I drew the following lines:


Only the first of these lines shows the actual intonation as spoken by a living human being, but the seven others definitely sound like Swedes resp. Norwegians in spite of being compiled from sampled fragments by a machine. And at least one of the Norwegian voices actually shows the archetypical jump upwards at the end of a sentence - which in this case also is a question. However the software must have overlooked this when it produced the flat intonation pattern of the first Norwegian voice, or the speaker speaks a dialect with an inonation pattern that is closer to Swedish than it is to other Norwegian dialects.

These lines only indicate tone level, but the distribution of stress points can of course also be notated in a number of ways (including the use of accent signs in the writing system itself). With the stress patterns the main error is to assume that there are a few localized stressed syllables and then no stress at all around them. But there are many levels of stress, and different languages have different patterns - some have large variations, some have small variations and some have regular intervals like telegraph poles in a landscape, others aren't nearly as regular. We all try to learn these patterns by listening and listening and listening, but my claim is that you can listen more efficiently when you have made clear what you are listening for. And having to notate exactly what you hear is an excellent way of forcing yourself to listen carefully, which carries over into your listening to authentic speech later on. Again I have to say that you may be one of the lucky ones who can absorb sounds without ever studying them in detail, but we aren't all born with this talent.
You do not have the required permissions to view the files attached to this post.
0 x

User avatar
Black Belt - 4th Dan
Posts: 4870
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 15359

Re: Iversen's Guide to Learning Languages (version 3b)

Postby Iversen » Sun Jan 31, 2016 3:33 pm

5.4. To know NOTHING but still understand something

If you read or listen to something in a dialect or language that is closely related to something you already know then you will often be able to understand it. The reason for this is that most words are the same in the two languages, just with some differences that reflect the different sound developments in each language. The historic linguists of the 18th century worked out long lists over the systematic changes of different sounds in different surroundings that led to the modern languages, and by rolling backwards through these changes and comparing different languages they have even been able to propose something Proto-Indoeuropean, which apparently was very different from its modern descendant on several points. But the relevant thing here and now is not how people spoke many thousand years ago, but the fact that the changes have led to parallel sound systems, so that you just have to find out which sound transformations lead from one language or dialect to its neighbour.

Lets for instance take High and Low German ('Plattdeutsch'). They separated from a common stem several thousand years ago through the so-called Second (or High) Germanic consonant shift, which in High German among other things changed p into ff or pf, depending on the context. There is a long list of similar changes which have been traced back to just this one event (check Wikipedia for a simple overview over the changes). Other sound shifts caused changes in the vowels, such as the one that changed long u into au in High German ("hus" in Low German, "Hause" in High German), and then it is suddenly not surprising that the following sentence from "Geiht das ok 'n beten fixer?" by Ines Barber ...

Suurkruut schmeckt opwarmt erst richtig goot.

... looks like this in standard High German:

Sauerkraut schmeckt aufgewärmt erst richtig gut.

When the systematic side to sound shifts was discovered the linguists hoped that ALL differences between the sound systems could be explained through series of shifts, which could be put in chronological order for each and every dialect or dialect. The world turned out to be more complicated than that, but finding the resulting systematic differences is still the basis for understanding any foreign speak through something that resembles it. This means that you can guess many words in the 'other' language by applying simple phonetic transformation rules.

But notice also the word "richtig" which is the same in both languages. I am fairly sure that the original form - if it existed - in Low German must have been something like *"riktig" (similar to "rigtig" with a hard first 'g' in Danish), because 'ch' in High German is another typical result of the second consonant shift. Such loanwords will of course not adhere to the normal transformation rules. Actually some words are borrowed twice. This can result in two words, one of which has been through a number of sound shifts which the other was spared.

Grammar changes aren't as systematic as phonetic ones, but please notice the missing "ge" in Low German. Unlike both Dutch and High German this prefix isn't used in Low German, - and on this point Low German resembles the Nordic languages. The reasons for this are not relevant for the ordinary language learner, but they show that a simple tree structure can't be expected.

If you have learnt (or discovered) a number of simple conversion rules, some facts about grammar and some of the more conspicuous false friends, then you are ready to read and understand Low German, and that is probably more than you can expect of most people from the Northern part of Germany, especially among the younger generation.

Now let's consider a slightly more distant relative, - Albanian, which is Indoeuropean, but occupies it own private branch on the Indoeuropean 'Stammbaum'. Here the conversion rules aren't simple at all, so you have to use other methods. I visited the country in 2009, and there I bought some bilingual guidebooks to Albanian towns. Below I quote a passage with a hyperliteral translation into English from a bilingual guide to the town Korça ("Your key to Korça çelësi turistik") - the English translation in the book follows below:

Pozite gjeografike
Position geographical

Rrethi i Korçës shrtihet në pësën Juglindore të Shquipërisë, në Krahinën Malore Qëndrore.
District of Korça* lies in [Eastern South]** of Albania, in region-mountain(eous) Central

Përfshin tre qytete: Korça, qendër e rrethit, Bilishti e Maliqi dhe 27 komuna.
(it) contains three towns: Korça, center of district, Bilishti + Maliqi and 27 communes.

Qyteti i Korçës është një nga qytet kryesore të Shqipërisë.
TownThe of Korça is on of (from) town(S) mostimportant of Albania.

Geographical position. The Korça District is situated in Southeastern Albania, in the Central Mountainous region. It encompasses three cities and towns: Korça, Bilisht and Maliq, as well as 27 communes. The City of Korça is one of the main cities in Albania.

* the name of the town is Korçë, the -a is a postclitic definite article, which is used here because that's what you do to feminine town names in Albanian, - and ë is a schwa sound, mostly silent in the final position of a word
*** deliberately erroneous translation by me - see below

You can probably already now see the parallels between the two versions, helped by the geographical names which are easily recognizable. If you want to find texts which almost can be understood alone from their loanwords then don't look for common babble or serious literature, - check out scientific or historical or geographical magazines, because it is here you have the highest proportion of loanwords, and with an unknown language you need those to grasp the structure of the foreign sentences. And often this will be enough to let you guess at the general meaning of the foreign texts, - in particular if you have got some context (such as you have when you are looking at a thing in a museum and there is a label telling you what it is called in Albanian).

With a bilingual text you can almost start building your own dictionary: "Pozite" = 'position' (noun), "gjeografike" = 'geographical', "rrethi" = 'district' etc. etc. I have assumed that "juglindore" meant South (because it ressembles "jug" in Serbian, which means 'South') and "pësën" would then have to be "Eastern", - but the capital letter in just one of the two words was confusing. I therefore took a peek in a dictionary I bought in Kosovo on an earlier trip and the truth was revealed: "Juglindore" alone means "Southeastern" and "pësën" means 'five' (which doesn't appear in the English translation). So I had nearly misunderstood the word "pësën" - but the dictionary saved me.

This example suggests that Albanian sentences aren't totally incomprehensible if you have some loanwords and proper names to start with. You can guess the meaning of other words ("i" = 'and', "në" = 'in' etc.), and by simple logical thinking and some guesswork you can soon extend your knowledge of Albanian, and you can even begin to dream about reading extremely hardcore stuff, such as books for youngsters and silly gossip magazines.

And even if you don't have a translation you can still use the same kind of clues: international words, loanwords, other comprehensible words, parallel passages and some simple logic. It goes even better if you know a few things about the language, like for instance that Albanian has postclitic articles and that the whole country is called 'Shqipëri' in Albanian. Some persons are better at this than others, but we all use such clues to understand neighbouring languages and dialects.

However I also tried the same technique on the bilingual in-flight magazine of Maleev, and I soon had to realize that Hungarian isn't nearly as close to English as Albanian apparently is, - but after looking up most of the words I'm still fairly sure that you could learn much about Hungarian by comparing the Hungarian and the English versions of a simple airline magazine. The burning question is: would I be able to learn Hungarian simply by reading magazines or listening to TV programs with subtitles? My guess is that it would take forever, and I would end up speaking Finnish like a Spanish donkey. But as a preparation for learning to understand spoken Hungarian (or Albanian) it might function, and you could probably also pick up some of the language melody and individual sounds with your right brain hemisphere while trying to find recognizable words with the left one.

Let me end this chapter with a slightly different topic, namely how to know something, but understanding a bit more than expected. And to illustrate that I would like to quote Krashen's "Principles and Practice in Second Language Acquisition", p. 20ff:

We acquire, in other words, only when we understand language that contains structure
that is "a little beyond" where we are now. How is this possible? How can we understand
language that contains structures that we have not yet acquired? The answer to this apparent
paradox is that we use more than our linguistic competence to help us understand. We also
use context, our knowledge of the world, our extra-linguistic information to help us understand
language directed at us.

The input hypothesis runs counter to our usual pedagogical approach in second and
foreign language teaching. As Hatch (1978a) has pointed out, our assumption has been that
we first learn structures, then practice using them in communication, and this is how fluency
develops. The input hypothesis says the opposite. It says we acquire by "going for meaning"
first, and as a result, we acquire structure! (For discussion of first language acquisition, see
MacNamara, 1972.)

We may thus state parts (1) and (2) of the input hypothesis as follows:
(1) The input hypothesis relates to acquisition, not learning.
(2) We acquire by understanding language that contains structure a it
beyond our current level of competence (i + 1). This is done with the help of
context or extra-linguistic information.

A third part of the input hypothesis says that input must contain i + 1 to be useful for
language acquisition, but it need not contain only i + 1. It says that if the acquirer understands
the input, and there is enough of it, i + 1 will automatically be provided. In other words, if
communication is successful, i + 1 is provided. As we will discuss later, this implies that the
best input should not even attempt to deliberately aim at i + 1. We are all familiar with syllabi
that try to deliberately cover i + 1. There is a "structure of the day", and usually both teacher
and student feel that the aim of the lesson is to teach or practice a specific grammatical item
or structure. Once this structure is "mastered", the syllabus proceeds to the next one.
This part of the input hypothesis implies that such a deliberate attempt to provide i + 1 is not necessary(...)

Thus, part (3) of the input hypothesis is:
(3) When communication is successful, when the input is understood and there
is enough of it, i + 1 will be provided automatically.
The final part of the input hypothesis states that speaking fluency cannot be taught
directly. Rather, it "emerges" over time, on its own. The best way, and perhaps the only way,
to teach speaking, according to this view, is simply to provide comprehensible input.

In most cases people refer to Krashen's ideas using the formula "N + 1", and the usual interpretation of this formula is that you should understand everything except a couple of words here and there in a N+1 source, but as the quote above shows he is more optimistic - apparently he expects some mysterious n+1 structure to be accessible even in n+2, n+3 or higher. And he actually taunts those who try to produce N+1 materials for learners - so there the whole business of graded learners went pouff. In fact he believed so much in this mechanism that he downplayed the usefulness of formel language learning tools like Anki, grammars etc.

Personally I doubt very much that this will function, but using sources which are marginally harder than you can cope with here and now serves 1) to strengthen your ability to use the things you already know AND 2) to point out a manageable number of things you should make an effort to learn. With N+1 the amount of new things to learn is limited, with N+100 it would be ridiculuously high. You could still study such sources, but it would be like solving riddles. Wouldn't riddle solving work too? Well, maybe - in the long run, but the result might be the ability to solve linguistic riddles rather than the ability to use the language freely.

The formula N+1 does NOT mean that you are restricted to ridiculously simple word combinations like "el perro blanco". Depending on the relations between the new language and those you already know and the amount of non-linguistic context, you may actually be able to start out with things at the level of Game of Thrones or the Divina Commedia of Dante Aleghieri. But in my interpretation it becomes more and more necessary to use dictionaries, grammars and translations to make head and tail of too difficult sources. And the one who already knows 2000 words in a language will be better positioned to make intelligent guesses about no. 2001 than the one who only knows 200.
0 x

User avatar
Black Belt - 4th Dan
Posts: 4870
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 15359

Re: Iversen's Guide to Learning Languages (version 3b)

Postby Iversen » Sun Jan 31, 2016 4:39 pm

5.5. Don't let a weird spelling lead you astray

Partly quoted from the thread "Early Modern English" at HTLAL - which may explain the choice of reference texts.

The Anglosaxon poem Beowulf can't be read without some study of Anglo-Saxon even when using a bilingual edition. But the Pearl poet? I didn't know this author so I had to look him up, and luckily Wikiquote had the Pearl poem with a translation by Brian Stone (http://en.wikiquote.org/wiki/Pearl_Poet). And frankly, I can't see why this stuff in Middle English should be a problem for advanced learners and even less for native speakers. Sometimes you have to think creatively to understand a word, but with the help of the translation you can make decent guesses. Let's have a look at the 1. verse:

Allas! I leste hyr in on erbere;
Þurз gresse to grounde hit fro me yot.
I dewyne, fordolked of luf-daungere
Of þat pryuy perle wythouten spot.

Alas! In a garden I lost it, let
It go to the ground on a grassy plot.
Bereft of love, I am racked by regret
For Pearl, my own Pearl without a spot.

The letter Þ (still found in Icelandic) is the same as "th" (which again is the real culprit behind the "y" of "Ye olde shoppe" and similar aberrations). If you didn't know it you should be able to guess it when you reach "þat" ('that'). "leste" is one wowel away from "lost", but this is just another dialect from another time - so look out for other cases where "o" is rendered as "e" by the pearl poet. "erbere": think 'herbal'. "Þurз" may not be the same word as 'turf' (as in 'home turf'), but pretend that it is, and "gresse" looks distinctly like "grassy" (again with an 'e' in place of something else, - now you should almost be able to imagine how this dialect sounded! - something like /wek wek wek/).

I won't go through the whole thing, but with a translation anybody who can cope with Shakespeare should also be able to understand this poem, and once you have conquered one poem like this one you are ready to tackle other writings from the 14. century, with or without a translation. There will probably be a number of unknown words in them, and once in a while you may even suspect that something more is going on than a simple translation can render. Am I the only one to see a saucy reference in the last stanza?

"Moteleз may so meke and mylde",
Þen sayde I to þat lufly flor,
"Bryng me to þat bygly bylde
And let me se þy blysful bor."
Þat schene sayde: "Þat God wyl schylde;
Þou may not enter wythinne hys tor.

"Moteless maiden so meek and mild,"
Then said I to that fairest flower,
Bring me to that bountiful pile
And let me see your blissful bower."
"God will forbid it," the bright one said.
"You shall not enter his holy place."

Or in modern language: Poet: May I *** you? Fair maid's answer: Nope!

Well, maybe my timid ventures into the world of Anglo-Saxon, my mediocre knowledge of Old Norse plus rudiments of Icelandic, French and Latin plus ample, but not recent reading of works by people like Chaucer and Shakespeare has given me a background which makes the Pearl seem almost understandable. But I have to accept the statements from two competent HTLAL members: the Pearl may be just one step beyond comprehensible even for smart native speakers.

But what is there to do about this? Short of following courses in Early Middle English there may be one trick you could try. I have sometimes read texts in dialects or languages related to something I already knew, and then it helped to try to see through the spelling and just listen to the words in your mind. As I said this poem has often an e (probably open) instead of a and o, which I irreverently rendered as wek-wek-wek. Imagine that you close your nostrils and try to sound slightly tired, detached and ironical - then I think the accent will be there. And then listen to the poem while you read the letters. Then the spelling "gresse" won't lead you astray - you'll hear 'grassy'. The unknown words will of course pose a problem, but there the translation should help you out - although this is a typical literary translation - with all the vagaries this entails - and not a hyperliteral one. But even with this caveat you should have less problems understanding the poem than when you just looked at the aberrant spelling and said: "no way, this is not my kind of English". And now let's leave English, past and present.

Major languages mostly have a standardized spelling, but you can't count on that for dialects or marginalized languages. Sometimes there is a spelling system which have become an unofficial standard. For Low German it is the one defined by Sass (which also is the one used at the low Germann Wikipedia), and I have a small dictionary in that orthography. But even within that dictionary there are often three for more variants, that illustrate the pronunciations of different (sub)dialects:


To add to the confusion there are also other standards, and one of these is used in the "Neues hochdeutsch-plattdeutsches Wörterbuch", where the translation of "befestigen" is given as "fastmåken, anmåken". To boot there as dialects or languages close to Platt, but yet not really the same thing - as the Plautdietsch, or Mennonite Low German, which according to Wikipedia originally was "a Low Prussian variety of East Low German, with Dutch influence, that developed in the 16th and 17th centuries in the Vistula delta area of Royal Prussia (...)". Later it was carried by immigrants to the Americas, where there also are colonies of emigrants speaking other kinds of German. Now how do you survive studying something that is so heterogene? The key here is to learn enough to give you a reasonably stable 'inner voice'. It is not strictly important that it is correct in the sense that at least one native speaker speaks like that, but it must be reasonably stable, and it will probably based on the variants you hear most often . And then you try to apply that voice even to texts in other orthographic systems than the one you are most familiar with. If a passage written in a special orthography then sounds like something you have heard before, chances are that it is.

On the other hand: if you are familiar with the pronunciation of a certain dialect, and you know someone who speaks it as a native speaker, then try to image that you hear that person read the text aloud. You may be able to understand the sounds you hear in your head when they 'come' from some known person.
You do not have the required permissions to view the files attached to this post.
0 x

User avatar
Black Belt - 4th Dan
Posts: 4870
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 15359

Re: Iversen's Guide to Learning Languages (version 3b)

Postby Iversen » Sun Jan 31, 2016 4:44 pm

5.6. Automatizing your target language

One goal of your language learning is to make the cumbersome construction (or decoding) of utterances automatic. In a HTLAL thread from 2010 member Cainntear wrote:

As an absolute beginner in language (French) I had to think things out -- I relied on conscious knowledge to construct the sentence. After a while, I started to notice that while I was thinking about how to form the correct sentence, there was a fully formed sentence appearing in parallel. It seemed like that voice had been there for a while, but I was just thinking too hard to notice it..

This is a very interesting statement, and one I can recognize from myself - but in a weak language I would have single words and sticks-and-boxes fleating around in my mind before the complete sentence pops up. I don't feel that there is a sharp border between those fragments and the final fullyfledged sentence - they belong to the same process that produced it.

I have written about intensive and extensive strategies and activities again and again, mostly - but not exclusively - in the realm of passive learning. And I have stated that intensive activities are most important in the beginning, whereas the scales tip more and more in the favor of extensive activities the further you progress in your learning process. And the reason that this happens is mainly that more and more of the knowledge you get through intensive activities is covered by automatic and even subconscious mechanisms.

One example of this is that you don't have to run through a conjugation table to find an ending and construct a certain form of a word - you have seen this form so often that this particular form is stored as an individual item. Other examples are the influence of one sound on surrounding sounds, the choice of preposition after a noun or verb and the choice of case after a preposition. But even though these individualized reflexes are the end goal they don't have to be absorbed as single items, which would last forever. Tables and list of rules are there to show you where all those scattered pieces of information belong.

An interesting development has been that the two storage mechanisms - as indivisible items and as computational objects - have been studied at the neurophysiological level, for instance in a study " How the brain processes complex words: an event-related potential study
of German verb inflections" by Penke et al. where complex words were studied by presenting German native speakers with deliberately erroneous forms. Quote:

We conclude that two complementary mechanisms co-exist in how our brain processes morphologically complex words: i. accessing full-form entries stored in memory; and ii. a computational system that decomposes complex words into stems and affixes. This distinction is reflected in different ERP responses to i and ii.

On the other there are also studies that point to a stronger role for whole-form retrieval ("Neurophysiological evidence for whole form retrieval of complex derived words: a mismatch negativity study" by Hanna & Pulvermüller).

We found that congruent derived words elicited a stronger MMN than incongruent derived words, about 150 milliseconds af ter perception of the critical morpheme. This pattern of results is consistent with whole-form storage of morphologically complex derived words as lexical units, or mini-constructions. (..) In addition, neurophysiological results reflected the frequency of derived forms, thus providing further converging evidence for whole form storage and against a combinatorial mechanism.

But remember: these studies have tested native speakers. They don't say that newbee language learners can or should learn all word forms as individual indivisible units - including the complex ones. It should however be seen as a sign of progress if you can stop forming complex words on the fly and instead retrieve them as fullyfledged entities. For once the learning process may profit from another process type than the one you expect to use at the end of the day.

The special situation for learners is described in a number of papers, for instance " Second language learners’ processing of inflected words:" by Hahne & Müller:

Two different subsystems of German inflection were studied, participial inflection and noun plurals. For participial forms, L2 learners were found to widely generalize the -t suffixation rule in a nonce-word elicitation task, and in the ERP experiment they showed an anterior negativity followed by a P600 – both results resembling previous findings from native speakers of German on the same materials. For plural formation, the L2 learners displayed different preference patterns for regular and irregular forms in an off-line plural judgment task. Regular and irregular plural forms also differed clearly with regard to their brain responses.

In other words: this is a fairly complex question, but the brains of learners of German do show some consistent differences in the way the brain reacts to erroneous forms. I think we'll have to leave the question there, but there will definitely be more research in brain patterns in the coming years. Right now the field seems to be in an earlier stage where the reports still conflict.

In the beginning of a language study there is basically only one active extensive activity available for you: parroting. You can hear an expression (in a situation where you can guess the meaning), and then you repeat it. Actually you don't even need you left brain with its language centers for this - the right brain can store this kind of unreflected fragments. The intensive language production starts already when you make the first minimal change - you have to know something about the mechanism of the phrase to make that change. And unless you continue along the parrot trail most or your active language production will consist of constructed utterances for a very long time.

But lo and behold, at a certain point you may experience that some little gnomon in the back of your mind starts making sentences on his/her own, and then you can start your extensive language production career. Your language production has become partly automatized. The problem is that this is more likely to happen if you do a lot of language production the hard way first. For those of us who rarely speak in foreign languages at home this can partly be solved through silent thinking and (for the more energetic ones) by speaking to yourself - I remember an excellent advice about pretending that you speak into a mobile phone, if you are to shy to walk around speaking to nobody in public. However I personally mostly stick to silent thinking.

Nevertheless I have found one trick that might be worth exploiting for others: I listen to something on TV or on the internet in a language I know fairly well, and then I make a simultaneous translation to a related weak language on the fly. Of course it will be totally rubbish in the beginning, but just having to hammer through something like a translation at the speed of a native speaker will mean that I don't have time to construct anything - and then the little lurking gnomon has the chance to appear without any interference from my internal schoolmaster.

And with time the little fellah may even be able to speak in somewhat sounding like the real thing - but only because the systematic part of me constantly feeds him with information about the language in question.

And no, I don't feel like a victim of multiple personality disorder.
0 x

User avatar
Black Belt - 4th Dan
Posts: 4870
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 15359

Re: Iversen's Guide to Learning Languages (version 3b)

Postby Iversen » Sun Jan 31, 2016 5:01 pm

5.7. Dissecting languages for fun

There are areas in the world where one language is so dominating that learning other languages isn't necessary, and if you nevertheless do so the choice will typically be English or another 'world language'. There are other places where it may be necessary to learn one language to speak to your neighbour and another to buy groceries at the market. And the authorities will communicate with you in third or fourth language. In such a place the problem is not to explain why you should learn more than one language, but you may still be in a situation where you have to explain why you choose to learn a language which isn't necessary for your bare survival. In Denmark English has become so intrusive that you can't avoid it, but other languages are under pressure. Due to new rules dictated by the government one of our universities - Sydjysk Universitetscenter - has just announced that it will drop all courses and research in Chinese, Arabic and Spanish, leaving only English and German. Other universities try to keep a wider range of language alive, but too many decision makers believe that English is all you need.

When I studied French in the 70s at the university in Århus we had an institute for Romance languages, and 125 students started to learn French in the same year as me. Many seeped away, but we still had lots of students and a gamut of teachers who offered very different courses. Some twenty years later or so later I saw that the new intake per year was down to 25 students, and while I write this the French studies have been lumped together in a far corner of "Arts" with other small and insignificant languages. Chinese and Arabic have been incorporated into area studies, where you study the culture as much as a language. I actually don't know whether you can pass an exam there exclusively by using English articles and reports, but the languages will inevitably suffer in such a setting. And the downwards spiral continues.

So my conclusion is that language learning for fun may become the rescue for anything apart from English and a handful of other languages with sufficient commercial clout.

I could in principle survive with just Danish and English, and the main effect would be that my travels would be somewhat less interesting. But I already have to survive on English if I visit the far East or parts of Africa so that wouldn't be a novel experience. My main motivation for studying a few languages more than necessary is .. knowing those languages. And adding new ones is like visiting new countries or towns or reading about advances in science.

There may other benefits to multilingualism. It has been claimed that multilingualism delays Alzheimer, but some studies has raised doubts about this (Like the 'nun experiment', which puts the blame in something called the ApoE4 allele, i.e. on genetic factors). Nevertheless it seems logical that you brain benefits from training, and learning new languages is certainly a way to put it to work.

Let's look at the process from a language learning perspective. In the chapter about associations (2.4) I mentioned innerlinguistic associations as one of the most efficient ways to memorize new words. You can compare the meaning of a word with the neurons of a brain cell: if it gets a stimulus from any of its neurons, it gets excited. So the more connections the better. A fluent speaker has met each word many times and attached new memories to it. And the goal of the association making in memorization is to form a mesh of 'memory hooks' which will make it more likely that you can recall that word later. A simple picture may do the trick, but a linguistic association leads to other words, and from there the process can continue.

It is often stated that learning the first foreign language is hard, but after that it becomes easier with every new anguage. When you think of the benefits of learning related languages it is evident that it is an advantage to get a lot of words for free, either identical or just with minor changes. This is not limited to sister languages like Spanish and Portuguese or Danish and Norwegian, but thanks to loan words it also holds true for more distant languages. That there is such a thing as a shared vocabulary in related languages is obvious in cases like Spanish "posada" / Portuguese "pousada". Why? Because there are lots and lots of words which have "o" in Spanish and "ou" in Portuguese. Why? Because these two language have a common ancestor in the form of Vulgar Latin, and historical linguists have constructed socalled 'sound laws' (or in less commiting terms: 'sound shifts') to explain the differences. It is evident that seeing a number of examples of a certain shift can help you to see parallels with other word pairs. It is less obvious whether reading about the changes in scientific lingo will help you to grasp them. Unless you have a solid education in phonology you'll learn more from looking at the concrete examples.

The second source of shared vocabulary is loanwords, where a word is taken over 'as is' or with minor changes due to the sound system and spelling rules of the recipient. Once you have noticed that "post" became "pos" in Bahasa Indonesia you are obviously prepared to look for more examples. There is a very comprehensive analysis of the loanwords in Dutch in the article "Loanwords in Dutch" by N.van der Seis in the "Loanwords in the World's Languages - A Comparative Handbook", and at page 332 she has combined topic with source language and got some somewhat surprising results. It is not a surprise that 13,8% of all words in her corpus within the field "Modern World" are English loanwords. The surprise is that this is the only topic where English is a major player (commercials weren't defined as a separate topic). The overall percentages are as follows:

French 6,8%, Latin 6,1%, High & Low German 2,7%, English 1,5%, Other Romance languages 1,1%, other 0,3%, total loanwords 19,1%, rest 80,9%.

There is one danger to be considered in this context, namely that you learn the 'easy' loan words and understandable heritage words, but neglect the incomprehensible heritage words. I have seen this tendency with myself and my Romanian, and the obvious solution for me was to do wordlists with this last category alone. For those who only learn words from context it must be harder to address the problem in specific terms.

The multiple language discount is also relevant for grammatical features. Let's examine a few concrete examples.

Let's take the Russian word "принадлежать" (pri-nad-lezhatj, 'belong to'):

"при-" as a verbal prefix normally indicates direction towards of position near, "над" as a prefix normally indicates that something happens on top of something, and the homonymous preposition means the same thing or movement towards the space above something. Finally "лежать" means "lie" or "lie down", so the whole things basically means "lie on something (or somebody)". Now the funny thing about Russian is that owning something rarely is expressed with a specific verb (even though there is a verb with a suitable meaning: иметь). Instead the Russian use a combination of "y" and the dative, or in other words: in Russian you don't own thing - they just are near you. And "принадлежать" fits nicely into this logic. A logic you wouldn't have known about without plucking the word to pieces.

This example only involved one language (Russian), but while I learnt the first words in Polish and later in Serbian I regularly used associations involving Russian words, and my personal experience is that without constructing the mesh of associations right from the start I simply don't learn words efficiently.

Let's widen the scope and look for syntactic examples with more than one language involved. Here the goal isn't to memorize single words, but to understand constructions, which with a bit of luck might help you to internalize them.

Let's start with a simple argument. In English you say "another thing", and in Italian you have an analog construction: "un'altra cosa" - but not so in Spanish: here you drop the article "otra cosa". To learn this just from reading or listening is definitely possible, but if you are an English speaker with some knowledge of Italian who now wants to add Spanish you will be severely tempted to add the article in Spanish. Or even worse: drop it in Italian. My suggestion would be not to rely on subconscious mechanisms, but to make the difference visible and conscious - and maybe even memorize an example which almost can function as its name - like for instance "un'altro - otro".

A slightly more complicated example: ergativity. It is normally seen as something weird and exotic, but in essence it just means that the object of a transitive verb is in the same case as the subject of an intransitive - and then the subject of the transitive verb has to be in something else. Just to illustrate this I'll quote a typical description from the internet (http://www.sjsu.edu/faculty/watkins/ergative.htm)

The ergative case is hard to understand because it refers to something not found in English. Generally it has to do with a language having a special indication that the subject in a sentence which has a transitive verb and direct object. If the subject is the agent of the action involved for the transitive verb then there is special marker attached to the noun which is the subject. This puts the noun into the ergative case. For sentences with intransitive verbs the subject is treated grammatically the same as direct objects are treated in sentences with transitive verbs. Few languages have the ergative case feature. The most notable are Basque, Sumerian and Greenlandic.

English is not an ergative language and furthermore it does not distinguish between the use of a noun for a subject and for an object. However, personal pronouns in English are generally so distinguished. We say, He kissed her and She kissed him. In English we say He puckered up, but if English were an ergative language we would say Him puckered up. There is a certain logic to this construction in that the phrase Puckered him up, sounds right although unusual. (Somewhat like Star Wars Yoda's syntax.)

But even Wiktionary.org has noticed that this isn't quite correct: according to Wikipedia English has no less that 322 ergative verbs. Well, English doesn't use cases much anyway, but it has an accusative form of some pronouns, and besides it uses word order to distinguish things:

They bent me to their will.
The road bends to the right

This may not be fullblown ergativity, but it can help a learner to understand the logic behind it.

I'll give one final example which I noticed during the flight home from Budapest after the Polyglot conference in 2013. I studied the grammar section of Kauderwelsch Band 31, Ungarisch Wort für Wort, and there I noticed a funny pattern:

Hungarian verbs have two paradigms for the present tense, the ordinary one and a definite present which only is used when the direct object is definite (which of course implies that the verb must be transitive). Their endings are quite different, and they are both illustrated with one verb that represents a 'dark' (or 'deep') vowel harmony and another which shows the 'light' (or 'high') vowel harmony. For my purpose it is enough to show one of them, namely the verb "tudni" (to know". First the ordinary present, then the definite one (in the usual order 1,2,3 singularis + 1,2,3 pluralis):

Ordinary: én tudok, to tudsz, ö/ön tud, mi tudunk, ti tudtok, ön/önök tudnak
Definite: én tudom, to tudod, ö/ön tudja, mi tudjuk, ti tudjátok, ön/önök tudják

(ön and önök are 'polite' pronouns)

In a totally different context there is a chapter about something called "Besitzverhältnis" (ownership), which isn't an ordinary genitive. If I say "the man's trousers" then it is the man which has a pair of trousers, not the inverse. That's how a genitive functions. If you say "táskája" in Hungarian (with táská = handbag) then the owner is a third person entity and the handbag is the owned object. So the flexive element is possessive, but not genitival.

Now take the 'dark' (or 'deep') endings through the usual persons, applied to "táska" by me (the book only gives the endings):

táskám, táskád, táskája, táskánk, táskátok, táskájuk

I may have used a wrong vowel in some places, but the system is clear: here you actually see a substantive with the same endings through number and person as a verbal form, namely those of the definite present. And I'm sure this isn't a coincidence.

Maybe it is worth pointing out that Hungarian like Russian and Irish doesn't have a simple verb for 'to have' (although they do have verbs that mean 'to own') - instead they all use a formulation like "to me is..." instead of "I have". And they all three have a strong tendency to leave out the forms of 'to be' in the present. Kauderwelsch illustrates the Hungarian construction with this example, where you of course immediately recognize the ending "ja":

Péternek - autója van
Peter+dative Auto+sein ist (approx. "forPeter carHis is")
Peter has a car

What now about the common present and its endings you might ask? The system there is not quite as regular, but almost, and it involves the dative, not the "Besitzverhältnis". In Hungarian the dative of substantives is marked by the endings -nak or -nek (low resp. high). So the interrogative pronoun has the dative "kinek" (nominative "ki", accusative "kit"). And the forms of the personal pronoun are built on "nek". NB: there are separate polite forms so there will be 2 x 4 elements below (1,2,3 and this polite thing - and please notice that there aren't separate genders in the paradigms - the Finno-Ugrian languages don't operate with morphological gender):

Nominative: én, te, ö, ön / mi, ti, ök, önök
Accusative: engem, téged, öt, önt / minket, titeket, öket, önöket
Dative: nekem, neked, neki, önnek / nekünk, nektek, nekik, önöknek

As you can see the positive forms in the accusative and dative are treated as ordinary 'high' substantives, but it is equally clear that there is a remarkable similarity between the endings of the other dative forms of the personal pronoun and the endings of the ordinary present as presented above (apart from 3 p.sing "-i"). And the accusative? Well, the ending of nouns in the accusative is 'something with a t', and you actually see this 't' in all forms except 1. & 2. person singular (where the endings are more like those of the definite present and the dative).

I have not yet learned Hungarian, but as this example shows it may be exotic, but not devoid of logic. In my opinion the best way of internalizing an utterly different grammatical system is first to try to see whether there is some kind of governing logic - and then decide whether there are a small number of things that might be worth learning by heart. And then you can start learning the details in an interplay between grammar and the real world.

There are a number of practical problems with the acquisition and maintenance of multiple languages, like finding time and getting exposure enough, but as you can see there are also benefits - in particular between related languages, where you get a lot of vocabulary and sentence patterns almost for free, just with the obligation to be alert to differences and false friends.

One irritating problem is that people may ask you how many languages you speak, and if you then say a number they will expect that you speak them all fluently. But in practice it will be a ladder with a number of fluent languages at the top and those you just know theoretically or are able to read at the bottom. The Russian author Spivak wrote in his book "Как стать полиглотом" ('How to become a polyglot') about his 'seven languages rule' - implying that seven languages (plus/minus one) was the maximal number a human being could keep fluent and ready for use at all times. The rest would then be slightly rusty so that it would need some brushing-up to get them up to speed again. My experience with this is that this can happen to languages you don't use much, but your passive skills are much less affected by a few weeks or months of inactivity. And if you could establish a situation where you used ten or eleven languages on a daily basis it might be feasible to keep that number of languages kicking and alive - but few of us have that opportunity.
0 x

User avatar
Black Belt - 4th Dan
Posts: 4870
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 15359

Re: Iversen's Guide to Learning Languages (version 3b)

Postby Iversen » Sun Jan 31, 2016 5:05 pm

5.8. Total and less than total immersion

In my opinion all language learning ultimately boils down to the notion of comprehensible input. For a total 100% novice nothing is comprehensible, so either you 'cheat' by offering a few words with translations in the beginning or you indicate the meaning by non-linguistic cues. In my opinion there is nothing gained by avoiding the translations.

After that there comes a stage where you can learn new words or expressions if they come in a context where their meaning is clear. For a 95% novice this means that you only can use special prepared texts, graded very carefully, or contexts where there are sufficiently clear non-linguistic clues ("this is a..." pointing to an object). Even at this stage I find that that it is unnecessary masochism to avoid two-way dictionaries and other external sources of information. And you can't trust the guesses of a novice - even if that novice is yourself.

For a person who already is fairly advanced it is less harmful to rely on monolinguistic settings because he/she already has some idea about which guesses are credible, and such a person is capable of processing much larger quantities of input. Nonetheless I still trust a good dictionary or grammar more than my own judgment even at this stage.

Of course living in a place where your target language is actually spoken will be the best possible background for perfecting any language. The better you are, the more profitable such a stay will be (said by someone who has no intention about emigrating). An alternative is to build a sort of 'language bubble' in your own country where you try to eliminate everything from home and include as much from the foreign culture as possible. The archetypal example of this is known under the acronym AJATT - "All Japanese All The Time". But even this can't always be done, and then we have to rely on videos, films, books, Skype calls and maybe the occasional visit to a relevant country. All this is positive, but there are people who are better than me to find occasions to use their foreign languages at home, and besides it is hard to give advice about such extensive activities without making that advice sound totally obvious.

I would however like to point out that the thing that really can slow down your progress is complacency – and this also happens at much lower competence levels. To counteract this tendency the simplest trick is sometimes to tell yourself to stop reading or listening for content, and instead to concentrate on HOW things are formulated. And you should also continue doing some of the things that originally brought you to the advanced level, such as active working with a grammatical issue or doing wordlists or flashcards or whatever. The new thing is that these activities now can be supplemented with studies in the more elusive stylistic and cultural elements of language.

I would like to add a little history from my own school days: when I was 15-17 years old and studied at the local 'gymnasium' (grammar school, high school, lyceum) I also had classes in French. Our teacher was a firm believer in the immersion method, but the class was following the mathematical line - so they were not very motivated to learn languages.

For two and a half year this teacher - one of the most brilliant teachers I have ever met - tried to teach this class French by the total immersion method, and he used all the tricks in the book plus a couple more. But he failed. So just half a year before the final exam he blew us all off our feet by suddenly speaking Danish to us for the first time ever. He told us that quite frankly most of us had learnt close to nothing until that point, and almost all of us were going to fail miserably unless he did something drastic.... so from now on he would do a strictly traditional course. No more modern antics, just old-fashioned black schooling with translations, grammars, dictionary and all the old stuff that he otherwise had tried to avoid.

The very same teacher succeeded in teaching this miserable bunch of half-boiled mathematicians enough French to pull most of us through the final exam just three or four months later. We learnt more in the last short period than we had learnt through 2½ years of entertaining, but fruitless immersion. My belief in pure immersion has never been the same since then.

It is however quite another matter with partial immersion. Again the point is the need to learn from comprehensible input, so the better you are the higher the proportion of native materials can be. And at the bottom of the ladder I cringe at the antics you have to live with in order to avoid using your knowledge of other languages, including your own. The idea of letting newbees guess the idea of drawings or situations and (presumably) learning foreign words and idioms while solving riddles is based on a false premise, namely that the learner doesn't involve his/her own language in the guessing process. Is the brain of a newbee a vacuum with just a few foreign words floating around in the pitchblack emptiness? Of course not, it is full of expressions from your native and other languages, and you can't avoid that these are used as models for your guesswork. So the whole circus of trying to make a purely 'foreign' experience for a newbee is silly. In fact giving the newbee a translation may involve a shorter involvement with his/her native language than letting him/her ponder for a long time in his own language over a problem - and maybe even guessing wrong.

So total immersion is a crock, but partial immersion is a blessing. I personally only experience immersion during during my travels, and I have made my own personal criterion for when a language should figure as one I speak: I should be able to spend a week or so in a country just using the local language. This may not entail hours on end of conversations in the language, but I should be able to have the same amount and kind of conversations in it as I would have had if I had spoken English or French or some other 'world language' to the local people.

If this is too hard there is an 'immersion light' version, where I walk around and speak to people in some common foreign language, but in my mind I make a running translation of all that happens into the local languages - with all the holes and errors I would be hesitant to foist upon some innocent foreigner. I'll come back to this kind of private simultaneous interpretation later, but suffice to say that it isn't as difficult as it sounds.
1 x

User avatar
Black Belt - 4th Dan
Posts: 4870
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 15359

Re: Iversen's Guide to Learning Languages (version 3b)

Postby Iversen » Sun Jan 31, 2016 5:08 pm

5.9. Activating languages

I have a couple of personal guidelines:

I call a language intermediate when I can keep on thinking in the language, but with lots of holes and errors. If I'm traveling with a language on this level I can ask for things in shop or make short comments, but I couldn't manage a extended discussion.

I speak about basic fluency when I can go to a country where the language is spoken and stay there for several days without using other languages, not even when I'm having longer discussions with the local people. I still can't avoid making errors, and I may have a strong accent, but I can have discussions about more or less the same themes as I would have had in for instance English or French or German (though I have to make do with a smaller vocabulary).

I speak about advanced fluency when I am confident (or I'm told by a competent source) that I speak almost correctly, - but not necessarily without an accent. And at this stage I would obviously not let anybody scare me into speaking English or some other tourist language.

But these definitions only concern one skill, namely speaking the languages (because that's my poorest discipline in each and every language). In fact you should indicate a level for both the active skills: thinking, writing, speaking, and the passive ones: reading, understanding speech. You can in principle be able to read even the weirdest poems in a language and still be unable to ask for an ice cream.

There are also some more official evaluation systems, such as the one used by the "The Common European Framework" of the European Union. This system divides learners into six levels:

A Basic User:
. A1 Breakthrough
. A2 Waystage

B Independent User
. B1 Threshold
. B2 Vantage

C Proficient User
. C1 Effective Operational Proficiency
. C2 Mastery

The CEFR describes what a learner is supposed to be able to do in reading, listening, speaking and writing at each level. But as every other system of this kind suffers from the defect that it attaches just one level of skill to a person for a given language, but the reality is that you probably are much better at some activities than at others, - for instance my speaking ability will always be the lowest, while writing abilities are more problematic for others (including native speakers). Passive skills will almost always be better than active skills.

If you have learnt a passive language to a high level you need to have a steady stream of input to keep the language alive. In the case of Latin it means reading Latin on a daily basis. No input, no activity. This reduces your chances of keeping the language alive during a dry spell. Learning a language as an active thing simply makes it more robust because you always can do some thinking in an active language wherever you are.

The activation of passive languages is precisely the thing I described with my Latin as an example. All the grammar and the words I had learnt in the 70s had in fact hibernated, and when I started to relearn my Latin I didn't have to hammer through everything again from scratch, I just needed some repetition rounds, and then I was ready to start thinking and writing in Latin. The only catch was that I wanted to think about things that didn't exist while Latin was still alive, - for instance this forum, my computer, trains, modern town planning and nuclear physics. So I have been busy modernizing my conception of Latin. But everything I learned about Latin as a passive language have come to good use now where my goal is broader. (…)

In retrospect, my first period with Latin was so lopsided because the teaching - as most teaching of Latin - followed an ancient and venerable method called grammar-translation, i.e. an outdated theory where the main goal was to be able to read certain venerable classical authors, but not to be able to use the language. The ironi is that I now have to battle against other theories that dismiss explicit learning of grammar and vocabulary, such as the methods of Krashen and other protagonists of so-called natural learning.

My current position is that you always should try to develop active skills, even in dead or 'undead' languages like Latin, but of course the consequences of having a 'bad' accent are less obvious if you never have to speak to anybody. With a language like Latin this implies that you should try to find dictionaries and homepages and other sources which try to update Latin. My preferred source for reading in NeoLatin is the web-newspaper Ephemeris (ephemeris.alcuinus.net), and I have found wordlists and dictionaries that make suggestions about words that cover contemporary phenomena. Not only will it be more amusing to spend time on Latin when you see it as a living thing, but all aspects of your Latin knowledge will be more robust.

Languages or dialects that you can understand because you know something that resembles them falls in another category because it isn't an immense, but fragile skeleton of grammatical rules and passive vocabulary that keeps them accessible for you, but something which you actually use - namely your own language or another language which you know well. However there is a difference between understanding and actually being able to use such a language or dialect - and the consequences of disregarding this distinction can be quite unbearable.

A good example of this is the relation between the Scandinavian languages. Some outsiders have suggested that they really just are dialects, and the proof is supposed to be that Scandinavians (Danes, Norwegians and Swedes) can understand each other. Well, the reality is that it takes a fair amount of exposure to make that possible, and right now it seems that the intercomprehensibility between the three main Nordic languages is seeping away because English comes in between them. But even a person who can understand practically everything on TV or in magazines in another Nordic language can't just go ahead and speak it. The result will be the persons own language plus some borrowed words and some of the more conspicuous intonations from the other language.

If you want to make a passive language/dialect of this kind active then my advice would be: do the same things as you would do with an unknown language - i.e. supplement your listening and reading with explicit grammar studies, vocabulary acquisition et cetera. The good news is that each phase will last for a much shorter time.

And that's about what I had to say about language learning for the moment.

You do not have the required permissions to view the files attached to this post.
8 x

Return to “Language Programs and Resources”

Who is online

Users browsing this forum: No registered users and 2 guests