I think mnemonics are great for getting yourself introduced to a word. I find that a mnemonic is a crutch I use until I've seen the word a few times. After say 10 real word exposures to the word...The mnemonic is long gone. I've been memorizing about 30-40 new words a day. Some require a mnemonic to help them stick. I've notice after a week or so the mnemonic is gone and the word sticks. This is at least the way it works for me....YOU maybe be different.

Mental tricks one brings to bear to make encoding elaborate, deep, and meaningful are the known standard to produce optimum acquisition of most information, including vocabulary. Personal resonance, associations, wordplay, imagery, whatever works for you. Rote repetition has its place, but it's superficial. The former is superior to the latter. You want the short-term to be more difficult in this way for the superiority of long-term retention and retrieval. The mnemonics fall by the wayside as you progress.

Also, everyone has a natural affinity for all natural languages. Japanese isn't unique in this respect, and its salient language features are functionally well-fitted to any human learner. That's how they evolved. Personal experience with an L1 that's similar to an L2 can be beneficial, but it can also be harmful, and these effects are only a labile part of language acquisition, and not remotely the most important part.

Related: http://www.sciencemag.org/content/330/6002/335.full

Kuma01 Wrote:I'll rephrase, it isn't the only way, but the most efficient way when you're at a beginner or intermediate level, especially for a language like Japanese that nobody has a real natural affinity for.

Thank you. I'm willing to agree with you. But Japanese is in no way different, to other languages, when it comes to learning of words, though. Words are sound, meaning, However, it still comes down to the way you are going to learn individual words, that makes all the difference. Is it just a list of words, without context, that you are going to learn? Nothing against a list of words to memorize, i use them too, to some extent. So i do very well know, how easy it is to memorize 190+ words, in a period of 2 to three days.

But how can you tell, if those words you are learning from a list, are worth learning in the first place? I ask this question from the viewpoint of a beginner, who has no prior knowledge of Japanese, what would your answer to this be? How many words would i need to know?

Quote:You only start picking up words via osmosis when you already understand a fair amount of the target language.

This is not true. I pick up vocabulary all the time, by osmosis, and in context of my textbooks, as well as from native media. This way i was able to acquire an estimated amount of close to 2000 words in this one year i have been learning so far. Osmosis, the gradual acquisition of words, as well as learning words in general, doesn't happen just by consuming. If you don't do anything for it, or with it, the outcome will be poor, at best. For learning words this means, you will have to look up definitions, use it actively, write it down, add it to the SRS to review it, use it in speech, translate it, and so forth. I prefer context-based vocabulary acquisition. Because this is where you learn most words.

Quote:But if you want to build vocab fast when you start out then rote memorization is the best way to accomplish that. Mnemonics etc. only complicate the acquisition and increase the lag when recalling the word.

I can fully agree with you on that. More words can give any learner a head-start. If mnemonic is good or bad, is dependent on the learner, and how he is able to process and retrieve information. I never used any mnemonics for learning vocabulary.

I think if you're learning Japanese words purely by rote you're truly missing out on the awesomeness of kanji, and it's such a waste, especially after doing RTK. The mnemonic opportunities are so rich and simple, it's hard for me to imagine learning words through mindless repetition. I said rote's good and fine once you know a lot of words, but really even then there's a certain level of automatic mnemonic trickery (mostly stemming from the interaction of modalities/orthography/context).

Like I said before I love a mnemonic for a word that will just not stick. My son is 8 years old and has Japanese lessons 3 times a week. For some reason he could not get the word 指（ゆび）to stick. It was really frustrating him. So one day I just said "you be number 1" (I made the number one sign with my fingers) I don't think Ive ever seen him looked so relieved about remembering a word. He has long since moved on past that mnemonic but I think it helped him get that word conquered.

Yeah I think using really explicit mnemonics like that are most useful when you're stuck on stuff, or feel the need to really distinguish it. I think the ideal is developing your metacognitive awareness to the point where you're able to subtly and quickly use word features, meanings, and your knowledge and overall strategy to generate and integrate new lexical items into your growing mental corpus, chunking and internalizing them as you proceed with the spaced retrieval system and non-SRS extemporaneous redundancies.

nest0r Wrote:I think if you're learning Japanese words purely by rote you're truly missing out on the awesomeness of kanji, and it's such a waste, especially after doing RTK. The mnemonic opportunities are so rich and simple, it's hard for me to imagine learning words through mindless repetition. I said rote's good and fine once you know a lot of words, but really even then there's a certain level of automatic mnemonic trickery (mostly stemming from the interaction of modalities/orthography/context).

It's not like I mindlessly try to remember words, obviously I know the general english meaning of the kanij, and when it correlates somewhat with the meaning of the word itself then that serves as an association to strengthen the memory, so in a way I'm not purely doing rote memorization. Also, while I've not done RtK 2 I have noticed patterns in readings after so many hundred words (certain radicals that are very often read with the same on reading etc. I think Heisig calls them pure groups). I've actually stopped learning the words as seperate units but instead I try to link the readings of the kanji in new vocab to words I already know, that way it's much easier to remember the readings. I don't consider that as a mnemonic per se, since you often tend to naturally start associating words and readings after a while.

I guess there's a reason I try to avoid words like rote and mnemonic.

Edit: I see the word ‘strategy’ is actually pretty popular, so mnemonics tend to be classified more expansively as ‘memory strategies’ in taxonomies of learning (e.g. Norbert Schmitt). Edit: Just so we're clear, rote is mechanical (speaking, writing, rehearsal) repetition with only cursory attention to meaning, especially w/ vocabulary 1:1 L1/L2 word pairs in lists). Mnemonics refers to incorporating prior knowledge, using imagery, personal, affective, elaborate, semantic, sound, etc. associative strategies. By definition rote is shallow and mnemonic is deep, the former being considered inferior to the latter for long-term memory.

This looks like a good book: http://books.google.com/books?id=PHqqd785IzYC

At any rate, my point is that being strategic and metacognitively adaptive and using elaborate encoding for spaced retrieval is best. Seems ultimately I'm always defending that difficulty in the short-term leads to long-term superiority, by ensuring short-term difficulty is used in a complementary, streamlined fashion.

Nagareboshi Wrote:There are so many ways to learn vocabulary, you pick words up by reading, listening, watching a movie, or working through a book etc. You can than create a word list,

All forms of rote memorization.

Unless we've intentionally pushed rote memorization into an arbitrary and easily rebutted sub corner.

There's no harm in mnemonics, because they tend to slip away once a word is encountered with enough frequency. It's not like the word continues forever to be retrieved via the mnemonic. They are a temporary crutch. The better question to ask is are they a necessary crutch?

The problem with mnemonics is that they are a waste of time to create for information you want to keep/use in the longterm, because the brain can and does learn fine without them. Let's say you want to memorise a list of 30 words. Depending on the word, it may take up to 2 minutes or more per word, and that's if you're not distracted by other things, so it'll take you maybe an hour to initially learn 30 words by making mnemonics for them. These words then get reinforced via review in srs or elsewhere (where perhaps more time is spent reinforcing the mnemonic) and really nailed down properly via repeated exposure in context. It is the distantly spaced review and exposure in context (and possibly also the act of forgetting) that insures the longterm retention. If I want to remember where I parked my car, or a phone number I'll need later in the day (assuming I don't have paper), I'll use a mnemonic. If I want to remember a word in the longterm, I'll jut be sure to review it as necessary.

My new method for learning Mandarin vocabulary now i'm done with Japanese, is simply to:
1. write 30 words or so into a list (usually takes 20 minutes or so) into a notebook along with their pinyin pronunciation and meaning in English or japanese and say them out loud.
2. date it
3. ignore that list for 2-3 weeks
4. 2-3 weeks later, check which words I still remember.
5. Create a new list for the words I forgot.
6. go back to 2.

3 or 4 reviews spaced over 1-2 months is enough to remember most words in each initial list. Some tougher words take maybe a month longer. I just write as many lists as I have time for in the day, with words taken from frequency lists, reading material etc.

By doing it this way, I don't bother wasting anytime on short term memorisation (via closely spaced reviews as in srs) and am able to get things into the long term memory spending as little time as possible on a per word basis.

There's nothing wrong with rote memorisation as long as it's not just cramming to the short term memory via over-reviewing over a short time (day before the exam etc).

That sounds boring and slow and ineffective. ;p You might be misunderstanding what ‘rote’ is; generally rote means non-distributed repeated rehearsal.

See also: http://books.google.com/books?id=C2uZNzg...22&f=false - I'll add some quotes down below since some pages get cut off in the preview.

It takes me a few seconds, tops, to construct a mnemonic for a word; generally I forget it by the time the next review rolls around; if it didn't do its job the first time (e.g. if I fail or have a hard time), I'll spend a few seconds again. Perhaps it's not even a few seconds, it's become so automatic. Often these days I don't even have to do that much, I think, so I should be speaking mostly in the past tense when first establishing a base of a few thousand words, and the usual line of progressively needing this stuff less and less or using it better and better follows. I have so much elaborative encoding, multisensory type information, especially if combined with the spaced retrieval practice of Anki and the way I've designed the Question sides of cards, and so much prior knowledge of the language, context, etc., there's not much need...

Studies on why spaced retrieval (stuff I've linked to recently on desirable difficulty and the like Edit: http://forum.koohii.com/showthread.php?p...#pid133017 and/or http://www.sciencedirect.com/science/art...1310002081) works are based around it using the same principles that explain why mnemonic memory strategies are proven so much more superior to rote rehearsal for vocabulary and other forms of learning and long-term retention. (Especially that one on the mediator hypothesis, which takes things a bit further in relating spaced retrieval to literal mnemonic devices).

Edit: Just found this—only thing I don't like is that mnemonic word learning often gets studied as the keyword mnemonic for some reason, when it's just a fraction of the whole, of memory strategies as discussed in books such as that one on vocabulary strategies mentioned before (quoted at the bottom of this comment): Comparing and combining retrieval practice and the keyword mnemonic for foreign vocabulary learning

Abstract: ‘Experiment 1 compared the effectiveness of retrieval practice, the keyword mnemonic and rote rehearsal for learning foreign language vocabulary. Both mnemonic methods produced similar recall and were superior to rote rehearsal. In Experiment 2, participants learned German vocabulary using keywords, retrieval practice or their own method. Retrieval practice and keyword-based recall were similar and superior to self-directed study. In Experiment 3, participants studied using keywords, retrieval practice, a combination or an elaboration strategy. Criterion testing occurred immediately and after a week. For receptive learning, retrieval practice and keywords were equally beneficial but for productive learning, retrieval practice was more effective. Combining strategies produced mixed results with significant benefits only for receptive learning in the delayed test.’

I think that last bit points to finding a combination that doesn't involve really explicit, concrete mnemonic devices for complex targets such as words or promotes use of devices such as RTK with its incorporation of weak cues for singular target information such as kanji. Edit: Interestingly, in the Discussion they note that the combined strategies gave the best production improvement/results vs. reception in the initial tests, and that this might be maintained through spacing for delayed tests. Oh, and even for more heavy-handed mnemonic devices such as the keyword technique for words, it's 20 seconds at the most per item.

I'm not against rehearsal, obviously I'm a fan (especially subvocalization/speaking audio, and handwriting), but only as a part of a overarching deep, semantic, multisensory, elaborate, metacognitive, spaced methodology. Schmitt's taxonomy from way back in 1997, which gets referenced a lot from what I see, has a similar breakdown. Mnemonic strategies have a category, but so do rote strategies and metacognitive strategies, all interacting and looked at from learner-generated and teacher perspectives. That seems to be the norm. Brute forcing stuff just seems so backwards and silly to me, and far more effort than its worth.

Quotes: From a couple previously linked books:

Mnemonics: “Most memory strategies (traditionally known as mnemonics) involve relating the word to be retained with some previously learned knowledge, using some form of imagery, or grouping.” (Schmitt's taxonomy)

“Thornbury (2002) has compiled a list of principles that facilitate the transfer of the learning material into the long-term memory. These include multiple encounters
with a lexical item, preferably at spaced intervals, retrieval and use of
lexical items, cognitive depth (cf. Schneider et al., 2002), affective depth,
personalisation, imaging, use of mnemonics (4) and conscious attention
that is necessary to remember a lexical item.”

(4): “Mnemonic devices are techniques based on cognitive processes
which are used to enhance retention of material one would otherwise
forget. The mnemonics can be classified into verbal (reduction, elabo-
rated coding, semantic elaboration, rhyme and rhythm), visual (imagery,
the Loci Method, method of spatial page organisation) and mixed
mnemonics (the Peg Method, the Keyword Method, association mne-
monics, rituals, process mnemonics)... ” (Takač)

“... the teaching and use of rote strategies has been heavily criticized because rote strategies are usually less efficient and effective than higher level strategies that are meaning based.

At first glance, rote repetition may seem efficient and effective. A process that re-
quires minimal attention and no comprehension is efficient in the sense that it re-
quires little involvement by working memory. And rote repetition can be deemed
effective because it certainly preserves information for the short term. Nevertheless, if
long-term retention is the goal, rote memory strategies fall short. The amount of
learning resulting from merely repeating an item is typically small and dependent on
the type of material.The assumption that simply holding verbal information in
short-term memory for a longer period of time leads to more learning and retention
appears to be fallacious (Baddeley, 1990; Shiffrin, 1999). It seems that rote rehearsal
of information does not necessarily enhance transfer to long-term storage (Estes,
1999). When it comes to learning, simple rehearsal-based strategies are sometimes
inferior to no strategy at all (Scruggs & Mastropieri, 1990).

The actual amount and durability of information transferred to long-term mem-
ory is thought to depend on whether the information is just being passively rehearsed
or is being more actively and consciously manipulated. There is little evidence to
suggest that simply holding information in short-term passive storage facilitates
learning. Rather, the degree of learning depends on the manner in which the infor-
mation is processed in working memory (Baddeley, 2006). Similar to the levels-of-
processing model advocated by Craig and Lockhart (1972), deeper processing of the
information seems to lead to better long-term retention than superficial processing,
such as rote repetition. In essence, rote strategies are inferior to relational strategies
that attach meaning to the information to be learned. Another drawback to rote strat-
egies is that they do little to increase the overall processing efficiency and capacity of
working memory (Parente & Herrmann, 1996). Despite the disadvantages, for some
individuals, especially those with low cognitive abilities and constricted working
memory capacity, training in rote strategies may be very beneficial and necessary be-
fore higher level strategies can be taught.

Relational Strategies

In contrast to rote strategies, extended higher level working memory processing does
increase the probability of long-term storage. Higher level processes, classified as rela-
tional strategies, make the information more meaningful, thereby increasing the
probability of retaining information over the long term. When related schemas are
available, relational processing can occur automatically; when a developed schema is
absent, effortful relational strategies can create relationships where none existed pre-
viously. Nearly all mnemonics, especially those involving imagery, can be considered
relational strategies. The learning principle underlying the effectiveness of relational
strategies is that meaningfulness strongly influences learning; attaching meaning to
information makes it more memorable.” (Dehn)

kazelee Wrote:

Nagareboshi Wrote:There are so many ways to learn vocabulary, you pick words up by reading, listening, watching a movie, or working through a book etc. You can than create a word list,

All forms of rote memorization.

How so? Let me give you an example. I am currently working with AIAIJ. Of course, there is a vocab list in every chapter, that i memorize. Also there is a list of kanji the book expects me to be able to read and to write. In one of the chapters there are words like　日付変更線 , or 知識豊富, 時差ボケ but the difference is, that i haven't had to memorize them from a list.

Those words and the words you can build with them [日付変更線], International Date Line, [知識豊富], knowledgeable, [知識] knowledge, [豊富] bountiful, plenty, [時差ボケ] jet-lag, [時差] time difference, would mean little to me without any context. And i guess i would have a hard time remembering them without it. But these words stick. Because there is context everywhere, in every section of a given chapter. And, thanks to the dialogues and grammar examples, and Audacity, i have this context in Anki as well. This is why i remember these words quite easily, without having to see them 10 times in Anki.

So this can hardly be seen as a memorizing process that involves routine or repetition, without being able to comprehend what has been memorized.

Quote:Unless we've intentionally pushed rote memorization into an arbitrary and easily rebutted sub corner.

If i had to learn for a test, and i know that i need to know a certain amount of words to pass it, i would just rote memorize them. But this way there is no guarantee, that i will remember any of those words, even after constant exposure to them, if they don't mean anything to me. So, rote memorization has it's merits, but if there are other ways, i'd prefer them over rote anytime.

Contextual learning's a different beast and is often opposed to rote learning in the literature; one salient aspect I think relevant to the idea of word acquisition is the tendency of natural sentences to give rise to thematically related word groupings. Thematic clustering of lists seems to be seen as facilitating learning of new words, while semantic clustering causes interference. I believe resolve incorporates some knowledge of this in Anki (by way of sibling cards?). Supermemo probably also mentions it. It's something we've noticed and mentioned here a few times I think (in the context of KO2001?).

At any rate, my newest interest in learning words in semantically related groupings stems from what I've been reading about how we store lexical items in the brain, and also I believe that with a vertical approach as in hypernym-hyponym in WordNet rather than horizontal (synonyms, sets of similar hyponyms under a hypernym), interference will be replaced by facilitation (well, that's what I'm noticing so far); there's also metacognitive awareness and variability induced by various consolidation methods and representation of information, etc. to keep things distinct. There's some other stuff regarding word sketches and case frames and whatnot being developed for Japanese terms (e.g. Japanese FrameNet).

But I digress.

nest0r's citation Wrote:Abstract: ‘Experiment 1 compared the effectiveness of retrieval practice, the keyword mnemonic and rote rehearsal for learning foreign language vocabulary. Both mnemonic methods produced similar recall and were superior to rote rehearsal. In Experiment 2, participants learned German vocabulary using keywords, retrieval practice or their own method. Retrieval practice and keyword-based recall were similar and superior to self-directed study. In Experiment 3, participants studied using keywords, retrieval practice, a combination or an elaboration strategy. Criterion testing occurred immediately and after a week. For receptive learning, retrieval practice and keywords were equally beneficial but for productive learning, retrieval practice was more effective. Combining strategies produced mixed results with significant benefits only for receptive learning in the delayed test.’

The problem with this study and likely a lot of studies about language learning and memory, is that it is far too short term. Testing immediately and one week after is still well in the realm of short term memory and therefore not really useful for people like us interested in acquiring vocab for the long-term and in as little work as possible.

nest0r Wrote:Brute forcing stuff just seems so backwards and silly to me, and far more effort than its worth.

It depends how you define brute forcing. In computing, brute forcing refers to solving problems by comparing every possible solution against some solution criterion (Relying on the speed of modern computers) rather that using a clever algorithm. So I don't really see how the term "brute forcing" applies here but anyways, mnemonic heavy methods sound a lot more like "brute forcing" to me in the sense that you spend unnecessary mental time and energy remembering additional information that is only useful in keeping the word in the short-term memory, something I've come to realise realise is completely unnecessary if the goal is long-term retention. Mnemonics tend to slip away anyway once something is in the longterm memory, unless more time is wasted periodically reinforcing it. So mnemonics are useful if you want to remember something an hour later or a week later but beyond that you're much better off just relying on effortless longterm memorisation. Mnemonic memorisation is useful for parlour tricks or for remembering useless information in exchange for time and mental effort. So the way I see it, you can put in the additional time (via mnemonics/sentences etc) and srs to insure that you still remember something 1 hour, 1 day, 1 week, 1 month later, but after a few months, none of that extra work really makes any difference to the end result

nest0r Wrote:The assumption that simply holding verbal information in
short-term memory for a longer period of time leads to more learning and retention
appears to be fallacious (Baddeley, 1990; Shiffrin, 1999). It seems that rote rehearsal
of information does not necessarily enhance transfer to long-term storage (Estes,
1999). When it comes to learning, simple rehearsal-based strategies are sometimes
inferior to no strategy at all (Scruggs & Mastropieri, 1990).

This would all seem to back up my ideas...This is why I ignore the short-term intervals and wait at least 2-3 weeks before checking what I remember. Spending a lot of time bouncing the word through your working memory, thinking of images and so on sounds a lot more like rote-learning to me, by your own definition "non-distributed repeated rehearsal".

The method I described in my previous post cuts out all the fat and is just a way of insuring minimal longterm repeated exposure to the things I want to learn, what the words sounds like, what they look like on paper and what they mean.

Boy.pockets Wrote:

nadiatims Wrote:

Boy.pocket's friend Wrote:I have my code, and I have my iOS device, but I can't put my code on the device unless I go through the Apple machine.

That's Apple's way of ensuring quality. You're can just use the iOS simulator for testing though...

Yes I agree. It is a good thing for quality and I can test using the iOS simulator and test it well. But I want to put my app on my iPod Touch... and I can't. :'(

You can, just jailbreak.

I think that short interval repetition helps by improving clarity and comprehensibility of input.

For example, when a Japanese audio clip is unclear to me, I find that listening to it repeatedly helps to clarify it and process it at natural speed. Without spending that time on building clarity, it is incomprehensible input, which is slower for the brain to learn from automagically. IMO that's a solid example of the value of short term repetition.

@vosmiura

Indeed.

Also, don't worry about examples, we've entered crazytown again, so you can pretty much make up whatever you want at this point and no one can argue with you. ;p

nest0r, what do you mean? 肯定か、否定か?
And btw, way to go derail the Goi thread when we have no idea what kind of study method it uses.

vosmiura Wrote:nest0r, what do you mean? 肯定か、否定か?
And btw, way to go derail the Goi thread when we have no idea what kind of study method it uses.

I know, but I don't want people who don't understand how memory works or the process of learning vocabulary to insult possible premises for the program and make general dismissals before we know anything about it.

Are you asking what my ‘indeed’ meant? It was neutral, but like I said, I think rehearsal has its place, which will entail some short-term repetition. I think microspacing beyond first setting out to study a new item is a waste but if it helps some I can't denigrate it, especially as long as it doesn't interfere with the spacing effect which needs a substantive delay (e.g. 24 hours) between initial encoding and the initial retrieval. I also want to ensure that folks don't attempt to replace quality with quantity, as the former is what's important for encoding and retaining information.

If you're asking what the rest of my comment meant, I was referring to selective, contradictory quoting, not understanding concepts like rote rehearsal and long and short-term memory, and just general nonsense in someone's previous comment. I made the mistake of forgetting who I was replying to and wasted some time writing out a response containing factual information backed by research.

I'll assume you were actually replying to me with your dismissal " we've entered crazytown again"...

nest0r if you're going to dismiss what I'm saying as crazy talk how about actually explaining why in your own words rather than just throwing about poorly understood research paper abstracts and pointing back at other longwinded forum posts from other threads. Can't you explain anything with any brevity?

vix86 Wrote:With Android owning a larger market share than iOS now, I'm have a hard time understanding why people still only make exclusive Iphone/Itouch apps.

Same here, can't care since I only use android.

Coding for iOS is admittedly a personal choice. We are a small team and we cannot code for both. Having said that, I have coded for android in the past and it's a nightmare compared to coding for iOS. Also nadiatims is entirely right, iOS users are better customers than their android counterparts. I don't mean to start a flame war, as I said we had to choose one and we chose iOS. We are iOS users ourselves so it was a natural decision.

This thread has turned into a fantastic resource on various vocabulary acquisition strategies. I am glad that our application has sparked this debate. It seems that a latent interest in "smarter" methods of repetition has been uncovered. This too makes me excited because that is at the core of our application. I too have been using Anki and other spaced repetition software and have found it to be excellent. However, when designing Goi we specifically target vocabulary and only vocabulary so that we could implement several unique features.

I won't give too much away, however, several of you have touched upon things that are at the foundation of our application, which is great because this means that people other than ourselves have been thinking along the same lines. I really hope that you like what we have done. Goi was created out of our own desire for a better tool and I think you will like it. We are looking forward to the beta testing period and will be seriously considering all of your feedback. This thread has made me realize, again, what a passionate group of learners make up this forum and I'm glad that many of you are dedicating your time to help turn Goi into an amazing way to learn vocabulary.

I'll attempt to respond to several questions, again without divulging too much. Firstly, while the ability to add/share mnemonics is built in it is entirely optional and unobtrusive. Many users will use it frequently, others less so. After reading the debates in this thread on the value of mnemonics I am glad we have done it the way we have.

Relational contextual memorization... I'll leave it at that. Sorry to be a tease, but the beta testers will find out soon enough!

Google search for "relational contextual memorization" only points to this page.

You should coin that, snozle . RCM?

Seems there's an upcoming publication of the results of a study (Diagnosing criterion level effects on memory: What aspects of memory are enhanced by repeated retrieval?) on how retrieval practice is not only superior for target memory, but cue memory and associative memory.