Memorising foreign vocabulary (through handmade lists)

It is important to make sure you have already memorised the information you are adding to your SRS. SRS keeps you from forgetting, but it is not optimised for memorisation. Trying to learn cards through repetition is likely to lead to a lot of those cards being failed repeatedly. This article introduces my favourite way of learning vocabulary prior to SRSing.



The method


There are no doubt many ways to learn vocabulary prior to SRSing, from rote memorisation to using those words in context. In my opinion, the most time-efficient method is the following:

1.  On a sheet of paper, starting with the first line, right a word on the left of the line, and its translation on the right. Repeat until you reach the bottom of the page.
2. Quiz yourself on the words, revealing each answer as you go (hide the column you are testing yourself on with a piece of paper).
3. Write down the words you failed on another piece of paper and repeat. If you fail some words again, write them down on a third piece of paper and repeat, until no words remain.
4. Add all the words to the SRS of your choice.

Some advice

It can be a good idea to wait between writing down the words and quizzing yourself, so you can make sure you won't forget the words you remembered too soon. If you find it hard to remember vocabulary in your target language, and you are failing a lot of words, testing yourself as soon as possible should help strengthen your memory of them. You can also try making a few lists of about 7 elements ("plus or minus 2" - the number of elements than can supposedly be stored in our short-term memory, although in practice, things get a bit more complicated than this - experiment!), and consolidate the forgotten words into new 7 elements lists.

In addition, make sure you are not learning those words devoid of all context: adding example sentences to your cards is a good idea. You can search for the words you are unsure of online, to get a better idea of how and when they are used.

Finally, prioritise your learning: beginners should learn basic, frequent words (don't learn every new word you can find!) ; intermediate and advanced learners should learn words as they encounter them, and from good quality word lists (collocations are a great idea for intermediate and advanced learners). Don't waste time on infrequent words unless you have already covered all the vocabulary you usually encounter/need. (Of course, you will learn many more words through exposure to the language: don't forget to get as much input in your target language as possible.)


Others do it too


Here is an amusing video from GO! Billy Korean detailing a very similar process:




The author draws a line in the middle of a sheet of paper, and then writes the translation (English here) of each word starting from the top line of the page. The equivalent in a foreign language (Korean here) is written starting from the bottom of the page. Here is an example:

one
two
three
-------------
tres
dos
uno


The page is then cut in two along the line previously drawn.
When it is time to review, one places the pieces of paper on top of each other, and can then reveal the solutions line by line by moving the top piece of paper up by one line, so that words that one hasn't been tested on yet remain hidden.

Flashcards don't eat babies

From time to time I stumble upon some harsh criticism of flashcards (or mnemonics, or any study method not found in the classroom). This criticism being invariably geared towards the use of flashcards for language learning, this will be the focus of my arguments.




'Flashcards are useless'

Argument: People have been learning languages without flashcards for thousands of years, therefore flashcards are useless.

You probably noticed this argument is fallacious. It is like saying: people have been walking to places for millions of years, therefore planes are useless. Planes may not be an absolute requirement to travel (if boats are available!), but they make the whole thing much more effective. Because something sort of works does not mean it cannot be improved.

Proponents of the 'flashcards are useless' argument will tell you there is no point drilling vocabulary with flashcards when you can learn that vocabulary through reading, using it a few times, or simply writing it down, closing your eyes and wishing very hard you will never forget it.

Let us be clear: reading is pretty much irreplaceable when learning a language. Using the language is too. Believing your memory to be infallible, however, does not only demonstrate a profound lack of knowledge of yourself and of learning in general, but will also bring you a lot of disappointment, and will ultimately slow you down.

The most common words in a language are indeed near impossible to forget if you get enough exposure to the language as it is used by natives. But the advanced learner cannot rely entirely on input to learn new vocabulary, simply because new vocabulary is uncommon enough that they (the advanced learner) will likely not see or hear it more than a few times a month, at best. A lot of words educated natives know fall into this category. This probably happens in your native language too: you see a new word, look it up, and a few days later it is forgotten - until next time (for this reason I always add those words into my SRS).

Even amongst flashcards and SRS users, it is generally believed flashcards become less useful as you get more proficient in your chosen language. My experience with English has been the exact opposite. I don't add everyday words to my SRS (such as drying rack or washing powder - words I only learned once I started living here, and that I use very often); I do add rarer words because I know from experience I will forget most of them if I don't artificially increase my exposure to them through SRSing (calcareous, mahogany, marksmanship... unless you are a geologist in the process of redecorating your home and with an interest in shooting, it's unlikely you will hear those often - yet they are useful words to know at an advanced level). You could use paper word lists and regular reviews to achieve this repeated exposure, but if you can feed the words into an SRS and let it deal with it, why would you? Word lists also present words in a non-random order that your brain may start using as a cue to guess which word is next; this is a problem easily solved by using flashcards.

In my experience, flashcards are also much superior to simple exposure to learn the gender of nouns in languages that have this feature. I made sure to quiz myself on genders when learning German and Dutch and got them right more consistently than the students who didn't use flashcards. Gender (and similar attributes) is not always obvious or visible in context. If you use flashcards, you can make sure you always learn a word and its gender, so you don't sound funny every time you open your mouth.



'Flashcards are boring'

I don't find my reviews boring. I actually rather enjoy them.

Now, I know a lot of people do find their reviews unappealing (sometimes even excruciating). But there are plenty of things you can try to make them enjoyable. You can add pictures or sound to your cards. You can review in small sessions spread over the whole day (many people review on their phone when they have a few minutes to spare - in a queue, waiting for pasta to cook, on the bus, or between classes). You can review while listening to music, or with TV - in a foreign language is even better - in the background. Make sure your cards are easy enough, and make sure you don't use your SRS to learn, but to remember what you learned. Make sure your cards are useful to you so you don't feel that reviewing them is pointless - delete any irrelevant cards. Use things you love (novels, films, websites...) as a source of new material, so reviewing will remind you of them. Use gamification, either inside of your SRS (use AnkiStrategy with Anki) or through external tools (such as HabitRPG, or some homemade gamification plan). In short, make it easy, diverse, relevant, and rewarding.


'Flashcards teach you bad habits'



'Flashcards do not translate into real-life ability'


Using someone else's deck and learning isolated vocabulary without examples while not getting any exposure to the language outside of reviews would indeed not help you in real-life situations. Luckily, I have yet to meet anyone who does that.

Flashcards are a tool, not a method

Having preferences is fine

Making effective flashcards: good and bad examples

Making effective flashcards ensures you will not get bored during reviews and the material will be solidly anchored in your memory. This famous page from the SuperMemo website should be essential reading for all flashcards users; following is my take on the issue of effective card formulation, and some lessons I have learned.


Never stray from the "minimum information principle"

I read the "20 rules" page when I started using Anki, but thought I knew better and that making cards resembling exam questions made more sense.

Here is one of those old cards:

Q. What was the Age of Pope associated with?
A. the rise of satire, novels and literary periodicals

The answer contains 3 bits of information - satire, novels, periodicals. I am also not being quizzed on "the Age of Pope" itself as it is mentioned in the question. I should have made 4 cards from the material: one per fact to be memorised.

Because those cards were ill-formulated, I failed them very, very often and stopped reviewing after the exam, as I found them tedious to study. They were also hard to evaluate: if you omit only "novels" from your answer, for example, should you mark the card as "failed" or "hard"? Whatever your choice, this will mess up the spaced repetition algorithm, since the information you give it will be inaccurate.


I am in the process of converting those old cards into effective reviewing material. Here is what I did with the example given above, combining multiple old cards into one:

[Alexander Pope], [Jonathan Swift] and [Daniel Defoe] are associated with the [Age of Pope], characterised by the rise of [satire], [novels] and [literary periodicals].

Everything between square brackets represents a cloze (my convention; the code is different in Anki). So, in practice, Anki will show me these:

Q. [...], Jonathan Swift and Daniel Defoe are associated with the Age of Pope, characterised by the rise of satire, novels and literary periodicals.

A. Alexander Pope, Jonathan Swift and Daniel Defoe are associated with the Age of Pope, characterised by the rise of satire, novels and literary periodicals.

Q. Alexander Pope, [...] and Daniel Defoe are associated with the Age of Pope, characterised by the rise of satire, novels and literary periodicals.

A. Alexander Pope, Jonathan Swift and Daniel Defoe are associated with the Age of Pope, characterised by the rise of satire, novels and literary periodicals.

...and so on, one for each cloze. As you can see, each important fact now has a card of its own, and can be rated individually. The cards are also easier to answer. Maybe surprisingly, this does not translate with a decrease in performance in the "real world" (outside of SRS study), on the contrary. Working on smaller units means you can really master the material.


Dealing with lists

The card I gave as an example contained two lists, one of authors (Pope, Swift, Defoe) and one of literary concepts and media (satire, novels, literary periodicals). Lists can seem challenging as you have to remember their individual items AND that these items are linked together.

Once again, stick to the minimum information principle: do NOT attempt to learn a whole list at once (like I did with my bad cards - thousands of them). This will NOT work well. Instead, make one card for each item on the list, and possibly, once you have mastered the individual items, and if you really need to remember the list well, higher level cards that group some of those items.

Here is an example of a list I want to remember (from Campbell Biology):

Four emergent properties of water contribute to Earth’s suitability as an environment for life: cohesive behaviour, ability to moderate temperature, expansion upon freezing, and versatility as a solvent.

The first step is to create a card like this:

Four emergent properties of water contribute to Earth’s suitability as an environment for life: [cohesive behaviour], [ability to moderate temperature], [expansion upon freezing], and [versatility as a solvent].

(You may want to break it into even smaller units if you want to remember the exact wording of the list; I only want to remember the concept behind each item.)

Once you know this first card well, you can group some items and create these:

Four emergent properties of water contribute to Earth’s suitability as an environment for life: [cohesive behaviour, ability to moderate temperature], expansion upon freezing, and versatility as a solvent.
 

Four emergent properties of water contribute to Earth’s suitability as an environment for life: cohesive behaviour, [ability to moderate temperature, expansion upon freezing], and versatility as a solvent.

Four emergent properties of water contribute to Earth’s suitability as an environment for life: cohesive behaviour, ability to moderate temperature, [expansion upon freezing, and versatility as a solvent].

When you have mastered those 3 cards, you can create the last one:

Four emergent properties of water contribute to Earth’s suitability as an environment for life: [cohesive behaviour, ability to moderate temperature, expansion upon freezing, and versatility as a solvent].

You could probably memorise poems rather easily using this method, but I haven't tried.


Anki - Making cards quickly using clozes

As we have seen, cloze deletion is a good way of making sure you are respecting the minimum information principle, and that you are creating a question for each fact and not leaving anything off. I am a big fan of cloze deletion and strongly encourage you to use it every time you can. Reviews will also feel less like an exam and more like reading.

In Anki, this type of cards is also very quick to make. Simply paste the sentence to be clozed into the card (remember to choose the "Cloze" note type, or your own custom cloze model), select what needs to be clozed and either click on the [...] button or use Ctrl + Shift + C (remember: keyboard shortcuts are always faster!).

Each cloze is given a number:

Proteins are built from {{c1::20}} kinds of {{c2::amino acids}} arranged in chains that are typically {{c3::hundreds of}} {{c2::amino acids}} long.

 As you may have noticed, this card has two "c2" clozes. In practice, this means both will be hidden at the same time. This is a handy tip to deal with reoccurring words or phrases, so the answer doesn't appear later in the sentence.

You can add a hint like this:

{{c1::20::number}}

Instead of showing you [...], the question will show [number]. Use it whenever you may get confused about what kind of answer is expected: failing a card because you thought the answer was something different, but still correct, would be a pity! Here is a good example of how that can happen:

Q. The Canterbury Tales were written in [...].

 The answer could be a date, a place, or...

A. The Canterbury Tales were written in popular English.

Even though you may remember what the question is about during the first weeks of review, it is likely you will ultimately forget, and fail the card as you give the right answer to a different question. So here is what my card looks like:

The Canterbury Tales were written in {{c1::popular English::language}}.

What about foreign vocabulary?

Spaced repetition is a great way of memorising foreign vocabulary, something I have not done actively for about a year. My cards were of the type word -> translation, with an optional field for examples and phonetic transcription. I suspect, however, that adding more context would be both more pleasant and more efficient. Following the way clozes work in Anki, you could experiment with something like this:

Card: {{c1::Le chat::The cat}} {{c2::dort::is sleeping}}.

Cloze 1: [The cat] dort.

Cloze 2: Le chat [is sleeping].

If you use a similar approach, I would like to hear about your experience.


Note: I know some people are uncomfortable mixing their native language with the one they are learning. I have never, ever paid attention to this and have been successful nonetheless. I suspect there is no or very little evidence to back the belief that this would be detrimental. One thing I am sure of, having experienced it countless times, is that being able to translate between your two languages is extremely handy, and made infinitely easier if you learn words with their translations. I am often lost for words when talking to my family in my native language, since I use English in my daily life and rarely translate anything anymore. This is extremely irritating and might be damaging in a professional setting - don't let it happen to you!

Why should I use an SRS?

This article is about the theory behind SRS, and the benefits of spaced repetition. Practical concerns (study time, type of material...) will be the subject of another article.


What is an SRS?

SRS stands for "spaced-repetition software". An SRS is a program you feed information you want to remember; the program quizzes you on it at the most efficient time so you remember it with minimum effort. If you have ever used paper flashcards, this is similar, only much, much more powerful.


How and why does it work?

Every time you are reminded of something you are trying to remember, its memory is consolidated, and forgetting happens more slowly (see Forgetting curve on Wikipedia). The program shows you the item you want to remember as late (and close to forgetting) as possible. This ensures you spend as little time studying the item as possible, while still getting all the benefits of spaced repetition.

Popular SRS includes: SuperMemo (the one which started it all; paid software), Mnemosyne (tied to a research project; free), Anki (arguably the most promising; free). I use Anki and love it to bits.
There are other implementations of spaced repetition, usually online, but they are of limited use over the long-term.


If it is that good, why is it not popular?

People who hear about spaced repetition and do not use it do it for one of two reasons: some do not recognise the value of such a tool, and do not think they can improve their study methods; others agree using an SRS would bring them many benefits, but lack the necessary motivation to incorporate spaced repetition into their daily routine.

While the technology is becoming increasingly popular, and some teachers have started using it in their classrooms, education policy is hardly science-based and it is not surprising it has largely ignored SRS (and other effective study tools) so far. Most people have therefore never tried spaced repetition and have yet to be convinced of its usefulness. This is a vicious circle: until SRS becomes popular, it will be suspected of being useless, since very few people use it.


I am this or that type of learner; I don't think it would work for me

Spaced repetition is based on our knowledge of the mechanisms of forgetting; while we are all slightly different, the principle behind spaced repetition is universal. In other words: it works for everyone. It WILL work for you, and drastically increase the amount of information you can remember, if you try it.

Learning styles, on the other hand, are "neuromyths" - widely held beliefs about the brain that are also false. While we all are different, we also all share some fundamental similarities. We can really say that memory works about the same for everyone (excluding rare medical conditions, obviously).

In my experience, people who fail at using a study method which takes advantage of how the brain works usually do so because they have been using the method improperly (e.g. irregular reviewing with an SRS), for too short a period of time to notice any benefits (e.g. listening to foreign language material for a few days before giving up), or because they are very biased against the method in the first place. This last point is, in my experience, very common amongst language learners - "other people have been very successful with this, but I don't see how it can work/I don't like the sound of it". The rejection of mnemonics to learn Chinese or Japanese characters is a very good illustration of this type of bias.


What spaced-repetition did for me

I started using spaced repetition to learn the 4,000 most frequent words in the German language. This took me a few months. My knowledge of those words was very imperfect, but this was enough to understand newspaper articles, a feat normally achieved only by the best students after 5 years of study in my native country (this only demonstrates how ineffective standard education is!).

I then used spaced repetition to learn vocabulary lists in Dutch at university (a few thousand words). This never seemed like a lot of work to me, and I also learned words we encountered during classes that were not on the list. Needless to say, students using traditional methods struggled to learn the list alone. I ended up acing all my Dutch exams.

To keep it brief, I used spaced repetition for all of my classes at university (English, Dutch and German but also history, literature and linguistics) and breezed through them. This took me about an hour of study every day; much less than my hard-working friends, for similar results. I also remember all the material from those years, when they had to constantly fend off forgetting.

I am now using spaced repetition to go through a biology textbook (Campbell Biology) on my own, while doing some psychology and history on the side (and reviewing English words).


A few words of conclusion

Spaced repetition has taught me learning can feel incredibly rewarding and satisfying. Being free from the burden of forgetting is invaluable; it means everything you study is worth it and will stay with you, ready to be built upon further. I wish I had started using spaced repetition at school: imagine remembering everything you were taught as a child; history, chemistry, geography, physics, art, literature, biology, languages...!

People sometimes ask me how I got so good at language study (they have yet to ask me about biology and psychology - I'm still only a humble beginner ;) ). When told about spaced-repetition, their reaction is invariably the same: "oh, but the way I study works for me." I'm glad I didn't believe that!