Learning with Texts for Classical Languages

Last March, U. Chicago’s Alex Lee wrote a great piece for the Dickinson College Commentary blog on the value of flashcards in learning Latin, with particular focus on the electronic flashcard app Mnemosyne. My own experience confirms Lee’s conclusions about electronic flashcards—I attribute a good part of my success on comprehensive Latin and Greek exams to a similar program called Anki, which by drilling 10 minutes a day for several months helped me systematically reinforce the hundreds upon hundreds of words that I needed to pass these exams. (The count for the Greek exam which covered Homer to Lucian was actually a bit over 3,000 words!)

That said, I was immediately struck by the introductory paragraph of the piece which begins:

“In an ideal world all vocabulary would be learned contextually, but when trying to learn Latin in a limited amount of time, we usually need flashcards.”

I was struck by this because in recent months I had begun to work pretty intensively with a new program designed to do exact that: learn vocabulary in context. The program is called Learning with Texts (LWT), developed by J. Pierre and inspired by among other things the language acquisition principles of Stephen Krashen and Steve Kaufmann’s LingQ language-learning website. As noted on the project page, LWT is “100% free, open source, and in the public domain.”

The LWT reading interface

The LWT reading interface

LWT is, in one respect, a reading assistant. It is an interface which keeps your reading material, your dictionaries and your vocabulary notes in one place. In another respect, LWT is a learning assistant, that is, it takes all of the accumulated knowledge built from annotating your texts and provides a second interface for reviewing this material systematically. The seamless combination of the reading and review interfaces makes LWT nothing less than a progressive memory of your language learning, built from the ground up and aware of what you have seen before, what you have reviewed before, and where you need to concentrate your learning efforts. It replicates many of the traditional methods for acquiring languages but in a highly centralized and efficient manner. In addition, it gives instant feedback and can track your progress with basic statistics, providing a game-like learning experience. Lastly, it is highly customizable, allowing learners to use the methods and resources that work best for them.

The LWT page is filled with information about the project, including its own installation instructions, setup guides, links to numerous other tutorials, and so on. The page includes tutorials for several languages, but Latin and Greek are not among them. (I did find this nice video on YouTube from Lingua Ex Deo on using LWT with the YLE Nuntii Latini.) This tutorial is meant to get you up and running as fast as possible with using LWT to read Latin and Greek texts. This tutorial uses the Perseus Word Study Tool for looking up vocabulary, though the program can be customized to use your preferred resources.

Getting Started

Unfortunately, the biggest hurdle for using LWT comes at the outset, namely the installation. The LWT project page does have detailed installation instructions for Windows and Mac. Both the these options require you to run a local webserver such as MAMP or XAMPP. You should only proceed in this direction if you are comfortable with running this sort of software on your computer. (The comments in this post rightly raise necessary precautions and point out possible risks.)

Luckily, there is another, simpler option, especially for those of you who simply want to try out LWT and see if it’s something you want to pursue further. Benny Lewis at Fluent in 3 Months hosts an online version of LWT—something he decided to do specifically because of the program’s “somewhat intimidating installation process”. He explains his site’s LWT features in this blog post. Using Lewis’s setup, I was able to get up and running with LWT in around 10 minutes.

Register for a free account at http://www.fluentin3months.com/wp-login.php?action=register by entering a username and password. You will be emailed a password. With this username and password, go to http://lwtfi3m.co/ to get started with LWT. NB: In the future, this will be the page to visit whenever you want to use LWT.

If you find LWT useful and plan to continue using it regularly and with any significant volume of reading material, I recommend that you invest the time and effort in installing LWT on your own computer. It runs much faster and more reliably than the lwtfi3m.co version (as noted on the project page).

Language Settings for Latin and Greek

Before you can get started reading and reviewing texts in Latin and Greek, you need to add them to your list of languages and then configure LWT to work with language-specific settings and available online resources. Here are the instructions for adding languages and configuring the settings.

Latin

  • Click My Languages on the LWT Home screen.
  • Click the green plus (+) sign or the text New Language… to add Latin to your available languages.
  • lwt_latin_settings

    LWT Latin settings

  • Here are the settings we will want to change to get LWT to work with Latin texts.
    • Language: Latin
    • Dictionary 1 URI: http://www.perseus.tufts.edu/hopper/morph?la=la&l=###
      This allows us to use the Perseus Word Study Tool as our main Latin dictionary. If you look at the URL query string, you will see two values:

      • la for language, which is made equal to la for Latin. We will use gr below when we set up the dictionary for Greek.
      • l for lemma, which is set to ###, i.e. the wildcard for LWT. So, when you click on a Latin word in LWT, e.g. verbum, it will query http://www.perseus.tufts.edu/hopper/morph?la=la&l=verbum for the definition.
    • Dictionary 2 URI: You can set this to any online dictionary which uses URL query strings. For example, *http://latin-dictionary.net/search/latin/###.
      Note that you need to include the opening asterisk in Dictionary 2 URI (as well as in GoogleTranslateURI below) so that the page will open in its own window.
    • GoogleTranslate URI: *http://translate.google.com/?ie=UTF-8&sl=la&tl=en&text=###
      Latin is supported by Google Translate. As with the Perseus URL, it uses a URL query and has the following values:

      • sl for source language, which is set to la for Latin
      • tl for target language, which is set to en for English
      • text, which is set to whatever text you want LWT to translate.
    • The remaining default settings can be used for using Latin in LWT.

Greek

  • Just as with the settings for Latin, click My Languages on the LWT Home screen.
  • Click the green plus (+) sign or the text New Language… to add Ancient Greek to your available languages.
  • lwt_greek_setting

    LWT Greek settings

  • Here are the settings we will want to change to get LWT to work with Greek texts.
    • Language: Ancient Greek
    • Dictionary 1 URI: http://www.perseus.tufts.edu/hopper/morph?la=gr&l=###
      See the Latin settings for an explanation of the URL query.
    • Dictionary 2 URI: You can set this to any online Greek dictionary which uses URL query strings. Remember to prefix the URL with an asterisk so that it will open in its own window.
    • Google Translate does not support ancient Greek at the moment. You can leave this blank.
    • RegExp Split Sentences: .!?:;·. LWT uses the characters in this list to determine the context for vocabulary words, that is these characters delimit the surrounding sentence or clause. This list adds the Greek semicolon (·, or the single midline dot) to the default settings.
    • RegExp Word Characters: \x{0370}-\x{03FF}\x{1F00}-\x{1FFF}.
      This tells LWT to use the section of the Unicode character set that includes the polytonic Greek characters.
    • The remaining default settings can be used for using Greek in LWT.

Reading Texts in LWT

You are now ready to add a text and use LWT’s reading interface. Follow the steps below to experiment with the reading interface.

Adding a text

  • Select Latin from Language: list on the the LWT home screen.
  • Click My Texts.
    There are multiple options available on the My Latin Texts page and I encourage you to play around with all of the them. For the purposes of this tutorial, we will concentrate on Read & Edit in the second column of the bottom section of the page.
    LWT's My Texts screen

    LWT’s My Texts screen

  • Click the green plus (+) sign or the text New Text….
  • On the New Text interface, enter the following information for the example text, Ovid’s Amores 1.1:
    • Title: Ovid, Amores 1.1
    • Text: Arma gravi numero violentaque bella parabam
      edere, materia conveniente modis.
      par erat inferior versus; risisse Cupido
      dicitur atque unum surripuisse pedem.
      etc.
    • Tags: Ovid, Amores, Latin, poetry, elegy. Tagging texts is optional but you may find it helpful later in creating custom vocabular sets for testing. (NB: You need to retype these, not cut and paste, for them to appear as separate tags.)
    • Audio-URI: http://bit.ly/1btqVZJ. This is the audio from Turpin’s DCC edition of Ovid’s Amores as read by Meghan Reedy. Audio is also optional and I will not cover it further in this introductory tutorial. Still, it is an amazing feature to have available and is worth being aware of.
  • Click Save and Open to proceed to the reading interface.

Reading a text

LWT uses a color-coded system to measure your familiarity with the vocabulary of a text. When you open the reading interface for the first time, naturally all of the words will be marked as unknown. These words are highlighted in blue. Your text of Amores 1 should something like this.

A new text with no learning history.

A new text with no learning history.

As you read you can mark evaluate how well you know the words you encounter. You rate words on a scale from, least familiar to most familar, i.e. 1-5 (red, red-orange, orange, yellow, green). As mentioned above, blue indicates unknown (unseen, really). A word with no highlighing indicates well-known. After reading Amores 1, your text may look something like this.

A text with learning history.

A text with learning history.

To evaluate a given word:

  • Click on the first word, “arma”. Notice that the Dictionary 1 opens in the lower-righthand corner and that your personal notecard for this word opens in the upper-righthand corner. Let’s mark this term in the Word pop-up box as I know this term well. The word will accordingly lose its highlighting, indicating that this word demands no further attention.
    Use this pop-up to rate vocabulary.

    Use this pop-up to rate vocabulary.

  • Click on the second word, “gravi”. Let’s add a Translation to our notecard in the upper-righthand corner—“heavy”. Click Save or press Enter to continue reading. New terms are automatically given a score of 1, although you can manually change this. We can include as much or as little information as will be helpful to us. Remember that this is your personal vocabulary learning system. If you want to include “heavy, weight, ponderous; sing. masc. abl.”, go ahead. I recommend keeping it simple at first, as you will learn over time how much detail you need.
    lwt_note_frame

    Notecard for storing definitions, tags, etc.

  • Note that when you hover over the highlighted words, the information from your notecard appears in the pop-up.
    lwt_hover_text_2
  • Continue reading this way until you have finished your text. Note the To Do counter in the upper-lefthand corner. This keeps track of how many unknown (blue) words remain in your text. You’d be amazed what kind of positive reinforcement this countdown can make—your reading progress now has a measurable goal and each word you learn gets you that much closer to zero.
    lwt_countdown

    The blue To Do counter lets you know how many unknown words are left in a text. Note also the audio interface in this frame.

  • Click on the last word “etc.” We can also tell LWT to ignore certain words. Click Ignore this term in the pop-up window.
  • As you read, you may come across a word where the dictionary gives you some clues, but your understanding remains cloudy. You can use Google Translate to get some contextual clues about the meaning of a word. Click on “pedem” and then Lookup Sentence in the pop-up. Here you can see Google Translate’s attempt at the fourth line—not exactly poetry, but perhaps useful in sorting out which meaning of “pes” might go best with “surripuisse” here.
    lwt_google_translate

    Google Translate can be used (for Latin, at least) to get suggested readings for entire sentences.

  • One last thing to take note of on the reading interface. At the top of the upper-righthand window, there are four icons next to the dropdown menu labelled [Menu]. The third icon, which looks like a pad and pencil, is the icon for Edit text. You can click this if you discover something you would like to fix, add, delete, and so on in your text.

Testing Texts in LWT

LWT’s reading interface does a wonderful job of keeping your vocabulary tools and notes in one place. In this respect, the program acts more or less like a slicker version of a traditional flashcard system, automating the recognition, look up and recording of new vocabulary. (Ok, there’s the color coding, the audio integration and loads of other features too!).

But it is in the review interface where LWT really stands out from traditional systems. As I mentioned at the beginning of the tutorial, LWT can take the vocabulary you want to learn and make contextual flashcards. Simply put, this would be so labor intensive to do using traditional methods on any serious scale—remember the 3,000 Greek flashcards!—that it would be virtually impossible to do.

The LWT Test interface

The LWT Test interface

The good news is that LWT recalls the sentence where you learned a word and can quiz you using this context. This means that you no longer need to rely on a simple one-word flashcard to drill, say, “surripuisse” into your long-term memory. You can now quiz yourself on this word in the context of its original line: “dicitur atque unum surripuisse pedem.” And LWT creates these contextual flashcards automatically for every word you learn.

Now is the time to test our vocabulary from Amores 1. (Actually, LWT will only test your vocabulary after 24 hours. This is practical in terms of actual language learning, but frustrating for the purposes of this tutorial. This means that you will need to resume the tutorial at this point tomorrow. If you try to test before waiting 24 hours, the program will simply tell you that there are no words to review.)

  • From the My Texts page, find your text of Ovid Amores 1 and click the icon for Test, the icon that looks like a blue thought bubble with a white question mark.
  • There are several test options available but for the purposes of this tutorial we will look at the first “..[L2]..”, what I have been referring to as contextual flashcards. Click on the button for “..[L2]..”, located next to the word Choose:.
    The [..L2..] button generates contextual flashcards.

    The [..L2..] button generates contextual flashcards.

  • A contextual flashcard will be generated from your text. The word being tested will be highlighted in yellow.
  • Click the word. The definition will replace the tested word and information from your notecard about this word will appear in the pop-up.
    You can rate how well you know a word on a scale of 1 to 5 using the LWT test mode.

    You can rate how well you know a word on a scale of 1 to 5 using the LWT test mode.

  • Inside the pop-up, you will have options for Got it!, Oops!, and the 1 through 5 recognition scale. If you know the word, click Got it! and the recognition score will increase by one. If not, click Oops! to reduce the score. Alternatively, you can choose your own score along the recognition scale. Click Wkn if you know the word and do not wish to be tested on it in the future.

Flashcards work best when you can move through them efficiently and at a comfortable pace. Using the mouse or trackpad to click on the scores can unnecessarily slow down this process. Accordingly, I recommend using the following keyboard shortcuts.

  1. Pressing the spacebar will bring up your notes in the upper-righthand box.
  2. Once the notes are onscreen, you can press the up and down cursor keys to change your recognition score by one.
  3. As in the information pop-up, you can choose your own score. To do this with keyboard shortcuts, press the number keys from 1 to 5 or W to mark a word as well-known.
  4. Note that you can only start using the keyboard shortcuts by clicking the highlighted word in the lower-lefthand window. Once you have clicked inside this window at the beginning of your test section, you can continue to use keyboard shortcuts for the remainder of the review session without using the mouse again. This is a little quirk that will take some getting used to.

The lower-lefthand corner keeps track of some useful statistics for your review.

Session timer and card remaining statistics in the LWT Test interface.

Session timer and card remaining statistics in the LWT Test interface.


There is a timer to keep track of your review session. (I have personally found 5 minutes per day to be the ideal review schedule.) On the right side is a running count of the cards on which you are being tested. For example, when you first open the test you might see an equation such as “4 = 1 + 1 + 2”. The first number is the total number. The yellow number is the number of cards remaining. The red number are cards that you missed and the green those you got right. The bar in between the timer and the count is a visual reminder of your progress.

When you are finished with your cards for the day, you will see a screen that reads Nothing more to test here! Tomorrow you’ll find x tests! Note that LWT uses a algorithm to determine which words need to be tested and in which order. The specifics are given here.

Conclusion

I hope that this tutorial was helpful and that Learning with Texts will be as helpful for you as it has been for me. I wish that this program had been available when I was taking my Latin and Greek reading exams. When I think back on how many hours I spent marking up my texts, writing out—and later typing up—flashcards by hand, rereading my texts for contexts, and so on, without exaggeration, I think it would have spent around 50% less time in preparing for the exams. The fact that there is so much classical literature available, between the Perseus Collection, The Latin Library, et al., ready to be cut-and-pasted into the program makes it even more attractive. You are just two keystrokes away from reading something new. Not only that, but you are two keystrokes away from reading something new with a system that knows every word you’ve ever seen and can automatically generate reading prompts based on your own notes. Apologies in advance for the hyperbole, but this is language-learning magic.

[This tutorial is based on a presentation I gave at the Dickinson Summer Latin Workshop at Dickinson College in July 2012. I want to thank Prof. Chris Francese for his support of cutting-edge methods in Latin and Greek pedagogy as well as the participants of the Summer Workshop whose interest in LWT, not to mention their insightful questions and constructive feedback, helped bring the tutorial into its current form.]

Advertisements

6 comments

  1. […] lot of information in this post is totally derived from this post over at DIY Classics. I 100% tell you to go and read that post. My aim is to supplement that post […]

  2. I copied and pasted out Greek text from Perseus but somehow LWT does not recognize it. I had no problem for it to recognize Latin text. Would you know what the issue could be?

  3. OK, I found the problem and solved it by using your set up instructions. Thank you so much!

  4. Giacomo · · Reply

    Fantastic

  5. Raimundo Brilhante de Oliveira Neto · · Reply

    estou tendo dificudade para add audio em inglês.
    Obrigado pela ajuda! Simplesmente perfeito.

  6. […] students on the words which they have looked up in a given session. Patrick Burns responded to this with a blog piece of his own, which is a tutorial on how to use Learning with Texts (LWT) to learn vocabulary in […]

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: