Morphman

Introduction

I won't say Morphman is essential to your studies, but if you already use Anki you might consider Morphman.

It seems scary at first maybe, but I can help you set it up in 5 minutes if you follow this guide. Here's a recap of all the cool things you can do with it.

  • At any point during your studies, you can know exactly how many words you know.

  • You can know exactly how much percentage of a title you understand, helping you finding something with the perfect difficulty for you.

  • You can set the exact percentage of comprehension you want to aim for your next study, either a show, or even the next episode specifically.

  • You can know how many words you need to study to reach that percentage of comprehension, which words they are, and Morphman will prioritize them automatically.

  • You can filter those words by a master frequency list, making sure you still only focus on the most frequent words in japanese, and not just words used in only one show.

  • You can use a master frequency list from all the shows on Netflix, or you can use one based on anime only or video game only. There are two versions: how many times the word appears in all the titles and in how many titles the show appears. You can create a master frequency based on the exact titles you want.

  • The study plan will make sure to order the words you're studying by the episode they appear in.

Get the necessary files

Go to Anki. Tools / Add-ons / Get Add-ons / Type or paste the following numbers. That will install the add-on automatically.

  • Morphman: 900801631

  • Mecab: 13462835

Close Anki, and reopen it, the add-ons are now installed. Reopen the anki add-on window. Select "MeCab ..." and click on View Files

Mecab is used to help Morphman analyze the sentences with better accuracy, there's a Mecab update for it you can get here.

Drag and drop all the files in update archive to your Mecab folder (named 13462835), replace all the files when asked to.

  • Download all the files from my readability List here. We'll go over them on the next page, but at least you have everything ready.

  • Download any deck that you want to study. It can be anime, manga, video games, your own made deck... Anything I made is listed here.

Again, close Anki and reopen it for Mecab to be updated.

What Morphman needs

Morphman can build a database of words you know, but to do this you need to tell him where they are.

Morphman doesn't analyze decks, it analyzes fields. So you need to know what field to look at, because analyzing a field where you have the english text won't do anything. You want to analyze the field where you have the japanese sentence or word that you've learned. You just need to know in which note type that field is.

If you're unsure about note style, field, cards, and all that, go back to the Anki tutorial section.

Note: Morphman is for words, not Kanji. You can't use it on a Kanji deck, simply because single kanji are not words.

For Morphman to work, you need to tell him which cards to analyze. For this, you need to know the note type of the card. But Morphman doesn't analyze the whole card, only a field so you also need to know which field of that note type to analyze. Basically, where you have the sentence or word in japanese.

You can analyze more than one note type, which we'll do in this tutorial, but each time you just need to need to know 2 things:

  • Your note type

  • The name of the field you want to analyze

If you wish to only analyze your cards, that's all you need to know.

If you wish to modify your cards, so they can be reorganized in that +1 order, you need to have a field where Morphman will put the word you don't know.

Morphman preferences depends on the name of YOUR NOTE STYLE, and YOUR FIELDS, so don't necesseraly copy my setting exactly.

Know what Morphman needs

Which note type ?

Go to the browser mode (B). You can see the list of your note types on the left, and if you click on it, you'll see which notes are in there.

Now, let's say that I just finished the Core Anime Deck, and I want Morphman to tell me how many words I know. In this case the note type I'll use is "Japanese Core Anime Vocabulary".

You can also go directly to manage note type (Ctrl+Shift+N), and to check the name of the note type you're using.

Which field to analyze ?

In this case, my sentence is in the field called "Expression".

That's for this note type only. For other note types, the field you want to analyze could be named Front, Word, Morph...

Just click on the name of the field and copy it, so you know there won't be spelling mistake when you have to write it.

Configure Morphman

Morphman's Settings

Press Ctrl+O to open the settings window.

Note Filter

  • Note Type: Pick your note type from the list. In my case, it's "Japanese Core Anime Vocabulary".

  • If you want Morphman to only apply on some cards from that note type, tag those cards, and put the name of that tag in that column.

  • Put the name of the field you want to analyze. No typos, that's why I said you could copy from the browser. Paste or write it here. Mine is "Expression".

  • Pick "Japanese Mecab ..." as your Morphemizer.

  • Check the Read box, if you only want Morphman to collect data from the cards, to build a database of your knowledge. That's what we want to use Morphman for.

  • Check the Modify Box if you want Morphman to reorganize your cards in +1 order. I don't check it here, because The Core Anime deck is already in +1 order, Morphman will mess around with the cards and we don't want that for this deck. If you're using the tango decks, you probably don't to change the card order either. We'll do it with the deck for A silent Voice below.

  • Click apply

Now, the only thing you have to do is recalc (Ctrl+M) your database. You'll see the number of your known words change at the top of Anki.

Note that the deck or mecab may change between the moment I write this tutorial and the time you do it, so the numbers may be different.

Morphman counts words in two ways:

  • K: Number of Known words (Ex ね and ねぇ count as 1 word)

  • V: Number of Known words - with iterations (Ex ね and ねぇ count as 2 words)

Now, that I know how many words I know, I want to keep studying with the Deck for A Silent Voice.

Morphman can analyze more than one field at a time, so we'll do just that. The process is the same, you need to know the field and the note type to analyze.

When you, go back to Morphman's option (Ctrl+O) and click on Clone to add another line.

The deck for A Silent Voice is in chronological order. To change that order, this time, you need to check the modify box. This way, you'll see sentences with only one unkown word, they'are called i+1.

We'll see in the newt page of the tutorial how to be more precise about that order with the use of frequency lists.

Let's check the rest of the options because we need to if we to modify the order of cards.

Extra Fields

I just said that Morphman can reorder any deck in a +1 manner, but to do that you need to have a field where that +1 word, the word you don't know is going to go.

This is required for every note type where you checked the modify box.

In addition to that field, you can also create more field where morphman will automatically write some data. It's optional.

So you need to tell Morphman in which field so the information is going to go by simply writing it in the boxes.

  • Focus morph (*). This where the unkown word of your sentence goes. By default, it's focus, I named my field "Morph", so that's what I'm putting here.

  • Index is basically Morphman's order, the higher the score, the sooner you'll see the card, but you have the due column for that anyway.

  • The following four columns tell you how many unknow words there are in your note, but also which ones they are. The unmatures are the words you know but don't know very well yet.

  • Unknow frequency can be useful. It tells you how many times in your collection the "Morph" word appears. When I say in your collection, it only applies to the decks you analyze with morphman. That number has nothing to do with outside frequency files. It applies to you present database in Anki.

  • Focus morph POS, I found useful when making my core2k deck. It tells you the grammatical nature of your "Morph". Same morph can have differents ones, this specifiy which one you're studying. Problem is that it's written in japanese.

If you want to add a field to your note type, you can go to Manage Note Types (Ctrl+Shift+N).

Pick the note type you want, click on Fields..., and then Add.

Name your field. You put the same name in the Extra Fields tab in Morphman's options.

Tags

This not necessary but really useful to know which cards to use, so may want to check the box to add those tags.

  • Vocab note, is what you want, only 1 new word in the sentence. I called it 1T.

  • Fresh and comprehension mean that you've seen all the words in that sentence. So the 1T note became 0T, meaning 0 unknown word. Comprehension means that one new word has matured, fresh means it hasn't yet, you saw it for the first time recently.

  • Not ready is for sentences that have more that 1 unkown word in the sentence. So 2T or more if you will.

  • Already known is very useful if you studied with Anki before using Morphman. You can use those tag to tell Morphman you already know those cards without reviewing the cards first. The tag by itself is not enough, make sure the note types and fields of the cards you tag are analyzed by morphman in the Note Filter.

  • You prioritize your words if you have a priority database. It's not really useful, Frequency works better.

  • Too Short and Too Long is important, you can use those tags to know if the sentence of your note is ... too long or too short. See below.

  • Frequency is also useful, it will tell you if the morph on your 1T card is included in your frequency file. See next page for frequency.

General

Let's get pratical and win some time with our studies. The example below is if I don't learn anything before using the "A Silent Voice" deck.

For this screeshot, I use the advanced browser add-on to see more columns in the browser and created the UnknownFreq field in the note type (see Extra Fields above).

Couple of observations:

  • Cards are organized by Due Date here. Without using a frequency list, Morphman will give to you in priority the words that appear the most time in your own deck collection. In this case, that means 何 appear 93 times in "A Silent Voice", な appears 75 times and so on.

  • I don't have 0T cards as I don't know anything. So 1T cards are first. More than 1T cards are at the bottom as you're not ready to learn them.

  • You can see that the sentences are just one word. (Expression column). Some sentences are longer but they are names put in brackets.

  • Those cards should be tags are mm_tooShort, but I changed the options to show you this first.

  • Now, my first card should be "A_KoenoKatachi_0570" where I'd learn 何.

  • I pass the card. The next one should be "A_KoenoKatachi_1037". But it's the same morph : 何. So maybe I don't want to lose time to see that card because, I already picked a card I'm gonna use to learn that word.

  • So if I want to move on to learn な and my second card instead to be A_KoenoKatachi_1031, check this option:

Maybe you didn't see the morph today, but you still don't want to learn additional 0T for 何 cards when you already have one designated for that, you can check the first two boxes as well. The difference is wether or not 何 would be consider as Fresh or Mature morph.

  • Include all missing morphemes, will add morphemes to your frequency list that weren't include because of your study plan settings. See next page for that.

  • You can choose to ignore morphemes that are categorized as grammar points.

  • You can ignore what's in brackets in your sentences. So if it's a word you don't know but it's in brackets, Morphman won't count that as a Morph. This option may unlock cards as 1T if names are in bracket like the last 3 sentences or my example above. It use add-ons like Migaku for pitch accent or analyze a field where you have furigana, you may want to check the bracket options.

  • In case names are not in brackets, you can check propor nouns as known. So you won't have to learn names of characters of towns...

  • Remember that if you fail cards too many times, they become leeches ? You can ignore those cards as well.

Click apply, you're done here.

Morphman's config.py

There's one last thing to configure. Go to your add-ons window, select Morphman, and click on View files. That takes you the Morphman add-on directory.

Open the "Morph folder" and open config.py with any notepad.

Comprehension and fresh

So those comprehension and fresh morph we mentionned just above, you can say to morphman which ones they are, and how long it takes for a fresh morph (known) to become a comprehension (mature) morph. You should leave those settings by default anyway, since it's recommended to skip those extra cards.

Pick cards with a good sentence length

We all know that length matters, especially when you have to study japanese sentences.

You can set threshold for what is too short or too long. If the sentence doesn't fit it will get some malus point and not be prioritize, but they won't get suspended or anything. Morphman will just prioritize cards with a good length.

Some like longer sentences to practice their listening skills of long sentences, because sentences in immersion are usually long. I like shorter sentences, it makes it easier to recognize the morph I don't know, and makes it easier to memorize it. It also make reviews faster.

Find a balance that works for you. The numbers hear apply to the number of Morph (words) not characters.

Prioritize words on your frequency list

We'll see about frequency list next page. But if you're using one, it's because you want to learn words in that file in priority.

So make sure the number here is high enough so they are prioritized in your queue list.

Using Morphman doing reviews

Pick the right settings

Maybe you'll change your setttings once in a while, that's okay, I did too, before I found the one that were working for me.

I use Anki just to be reminded of the vocab I know and keep it fresh in my head, but the real studying or practice takes place during immersion. It's by watching an anime, or most likely playing jrpgs that I really see my progess or when I'm reading, and I don't need to check the definition of a word because I saw the card not too long ago. So I make sure I spend as little time as possible doing reviews to make them efficient. Short length, good cards to memorize, skip all the redundant 0T out of the process, disregard proper nouns and words in brackets...

That's just me, as always, figure out what works for you.

Pick the right card

Unless you make your own cards, good chance you'll use subs2srs decks for anime. If that's the case, some cards are not that great because the progress was automatic. The screenshot doesn't help, the audio is missing at the beginning or the end of the sentence...

So when doing reviews, simply press "L". It will show you all the others cards that have the same Morph. Use the preview button to check them out.

If you find a better card that the one Morphman gave you, select it, you can use Edit - Learn now (Ctrl + Alt +N) or Cards - reposition, and ok. I use the latter since you have a shortcut in the header.

Update your database

Morphman doesn't update your database of know words automatically. You need to do it manually. Press Ctrl + M to do so.

Maybe not useful to do it every time you learned a new card, but I do it at the end of my reviews.

So do your new cards for the day, recalc the database. In the header of the main screen you'll see the K and V numbers change.

Why two numbers ? It's because of iterations of the same word. So ね and ねぇ are the same word but written a bit differently. Think of "hey" and "heyyy".

If that was the only know word your numbers would be this; K 1 (ね) and V 2 (ね and ねぇ).

Overall the way to count words is not perfect, so don't stress about it too much. When updating mecab for example, your numbers may be lowered, it doesn't mean you loose knowledge, it just means there's better accuracy.