Daddy's Big Girl Tackles Latin

Don't you dare call her little.  She will set you straight, stomping her foot if necessary.

This wasn't my idea either, at least not for her, and not at the time. Yesterday was a lazy day at home for the Rietz clan.  We are fairly

Christmas Eve

industrious slobs, though; after a couple of bike rides my middle guy was mastering the GoPro camera given him by relatives, my oldest was managing a virtual soccer team, and my daughter was lying on the couch next to me, weaving with her new rainbow loom.  I was on my laptop debugging code.

Although I don't push it too hard, my kids are no strangers to what I do.  My older kids have both played in Linguachet, and several years ago my middle guy (who was 8) actually helped inspire the vocab and forms tabs by the questions he asked me as he worked.  My oldest, who is quite good at graphic design, created my current business cards.  When my daughter asked me today if she could play Linguachet, it didn't surprise me.

She is seven, though.  She's a very good speller, but - like all second graders - she's still learning.  She really just picked up cursive in the last couple of weeks.  She's curious and quick as a whip, but she changes her mind just as quickly.  I certainly didn't picture her composing full sentences in Latin.  I handed her an iPad with a smile, figuring this little experiment would last five minutes at most.

What happened next floored me.  It's exactly what was supposed to happen.  She started typing, looking up almost every word at first.  She's still a hunt-and-peck typist; it still takes her a couple of seconds to find letters like "P".  She didn't whine. She asked questions. She didn't quit.  She fixed all her mistakes.  She celebrated.  She did the first unit.

I figured surely it was time to lighten up...or at least vary the approach before she burned out!  We went outside and threw a "pila" to each other. As we threw, I called the play-by-play in simple, repetitive Latin.  She got cold. She came back in.  She asked to do more Linguachet.

She redid the first unit.  She started the second.

I pulled out Oerberg's Lingua Latina and a globe.  She translated that first page, her confidence surging.  After that she reached for my copy of Cattus Pettasatus. Obviously that wasn't so easy AT ALL, but not being able to do that didn't seem to discourage her at all either.  She went back to Linguachet.

At dinner, I told her that words are fossils.  She immediately started splitting up words she knew.

I took a walk around the front yard with my oldest.  I was worried she'd get frustrated and quit, but I also think kids need space to learn to manage their frustrations. I fully expected to return and find her watching a cartoon, and I would have been fine with that. 

I returned to find her typing Latin. 

Minecraft and Disney get turned down for what!?

At bedtime we often read a Bible story together.  On a whim, I pulled out the Latin Vulgate, reading in Latin from Luke 3, a story she had heard in church - in English - a couple of nights before.  She looked over my shoulder and picked out Latin words she had seen and used throughout the afternoon: "et", "est", "sunt".  Others, like "angeli" she easily picked out from context.

I attach no ongoing expectations here. Today is a new day.  However, as I was first typing that sentence, my daughter woke up walked into my room.  She didn't say good morning.  She asked to play more Linguachet. ;-)

Keeping the Love of Latin Alive...During Exam Prep.?

Exams can mark a nerve-wracking time for all of us.  Younger students (I teach middle school.) need

Blechhhy, Wikimedia

to be walked through their studying more than older ones do.  For years I simply gave a nonstop regimen of quizzes the two weeks before an exam.  They did help measurably, but no one could describe the process as engaging, and it certainly did little to foster the love of Latin!  Sure we sang a bit too, but we had a lot to cover, and one aspires to more than chocolate-covered broccoli, right?

Then I stumbled upon the idea of review via composition.  At first, I was thinking of efficiency, not passion.  I could hit all of the grammar (and much of the vocabulary) in ten to twenty carefully-written Latin sentences.  This did seem to be more authentically challenging, and it actually did work well for my top students.  

My struggling students, though, could get overwhelmed as they self-corrected that action-packed page.  It was too much feedback for them to take in all at once, and there was nothing they could immediately do with that onslaught of information... except hope to do better next time!

Doing this in Linguachet did improve things a good bit:  students got specific immediate feedback as they were working. They could learn from their mistakes. Furthermore, that immediate feedback

really motivated them...even when I was simply grading completion!  

For highly challenging, cumulative exercises, though, this motivation occasionally became a bit of a problem: struggling students struggled too hard, and occasionally I would find an exercise 1/2 done to total perfection with a massive amount of time expended.

From talking with my students through this process, I began to realize that - for many of them - making more than 10 attempts per question often amounted to brute-force hacking.  Formative feedback drew them deeply into battle, but - in certain situations - Pyhrric victories were more likely to result than meaningful learning.  Ironically, I had built Linguachet partly to enable differentiation, but seeing my own students' learning data from it was what fully convinced me.  Students' attention and effort are precious, limited resources, and I had to start learning to steward them.  This may even be more true during exams, when many students also need to catch up!

From there, I began to watch Linguachet's results tab data differently.  My goals began to change: from getting everyone through the same work to keeping each student in a state of flow so he or she could progress.  Armed with live data and plenty of ready choices, I began to improvise differentiation a bit.  I began encouraging certain students to use the skip button, not to avoid challenges but to select just a few of them.   I began to experiment with assigning the concept-dense review exercises to my strongest students while targeting concept-specific units for the struggling ones.  I also experimented with self-prescription, having students look at a recent test and choose 3 weak areas to target in Linguachet.  Knowing the homework was targeted to help them individually did put more of the focus on learning for some.

Everybody loves a Cinderella ending.  All my C students became A students.  That didn't happen.  Writing 12 sentences doesn't make every single vocabulary word magically stick either, but it definitely helps with grammar.  My steps were tentative.  I still gave the same exams.

What else happened, though, was far, far cooler in its own way.  More of my students began to get the idea that they could DO Latin.  Some started working ahead in Linguachet.  One first-semester Latin student, in the peak season of exam stress, decided to translate "Somewhere Over the Rainbow" into Latin, frame it, and present it to me for Christmas.  In January, my Latin students crashed their school's talent show by performing "Free Bird"...in Latin.  On at least one occasion, my assigning Linguachet homework was met with a classroom of...cheers?

I still don't believe in a holy grail of education.  My classroom has often been a lab but never a test tube.  I was simultaneously experimenting with oral Latin, embedded stories, comprehensible input, and TPRS, and I think all of these things buoyed students' enthusiasm.  My students were also watching ME risk, improvise, and learn, which probably added to the energy.  Perhaps anything that makes you risk, improvise, and learn from your mistakes is going to catalyze a growth mindset.  I hope that Linguachet will do that for a lot of students...even during exams...perhaps especially during exams!

Two Facts Collide in the Brain

I recently heard a developmental molecular biologist address us teachers about research on the human brain and its possible implications for how we teach our students.  He is an animated speaker whose own brain runs at warp speed, and I learned a number of fascinating things from him.  Here and now, though, I will focus on one of them: 300 milliseconds.

200-300 milliseconds (1/5 to 1/3 of a second), apparently, is the minimum time it takes the executive center of the human

This is your brain on colored dye, with credit to Gutenberg Encyclopedia.

brain (which processes requests one at a time) to change from one simple, conscious task to another.  Again, that's only for simple tasks, whatever those are.  Complex ones can take much longer to begin.  That's not even the time it takes me to think about something.  That's the minimum time it takes me to start consciously thinking about anything.

I suppose we all come to the same facts from varying perspectives.  Perhaps 300 milliseconds doesn't seem like very much time.  After all, it may be faster than the blink of an eye, which clocks in at 300 to 400ms.

Personally, that 300 ms changes the way I see myself as a human being.  I am monumentally slower than I want to think I am.

Let me share a different perspective.  I develop computer code that is shooting Latin sentences from students in places as remote as New Zealand or Tel Aviv to a high-powered server in Dallas, doing all kinds of conditional logic and data gymnastics, and sending them back again.  I've now invested years of my life in that action-packed round trip, which involves a relay of code in several different computer languages.

Today, from my desk in Tennessee, that entire process takes about 93 milliseconds.  It will go to Dallas, do its thing, and shoot everything back here before I can even start to think about opening Facebook.

My friend and mentor Scott Salisbury, whose company develops such solutions for corporate clients,

photo credit: Hustvedt

assures me that this figure is basically all server time.  After all, signals traverse fiber-optic cable at the speed of light, which would take me from here to Dallas in less than 4 ms!

There are times when speed matters, too.  300ms seems like a very long time for my brain to switch from engaging in a cell phone conversation to considering whether... perhaps... I should slam on the brakes.  And to go back from driving to the phone conversation?  Another 300ms.  And each time my brain switches, the car I used to drive is carrying my body another 31 feet while my mind is somewhere else. Yes, my eyes have never left the road.  My hands may still be on the wheel.  The lights are on, but I'm just not quite home at the moment.

Iraqi children - photo credit: Christiaan Briggs

We all still want to believe that we really can truly, meaningfully multitask.  We believe we experience it every day.  We insist that we can prove it.  We've raised multiple kids, for crying out loud!  What about a guitarist riffing his scales at 120 beats per minute while watching football on TV?  Heck, that guy even looks bored!  Apparently many things get chunked together and/or delegated to muscle memory.  That's what practice does.  But the chunks of conscious thought - however elaborate - still pass through the executive center of the brain, one at a time.

Even more importantly, how much additional "server time" does my brain spend actually listening to the other person instead of driving?  How much time is spent processing emotions?  How much time does it spend composing a response? After all, if - as we were repeatedly assured - all executive processing is sequential, 300ms is only the transition time, not the total time off task.  It's just a drop in the bucket of things I am not noticing.

Even with the best of intentions, I've seen students run into a similar problem: they think they can complete another task on their devices and still be listening in class.  They really believe they are listening too, but when I talk to them, they aren't "there."

300ms humbles me. Like the drunk guy who wants the keys, I may not be in the best position to judge objectively my own limits, because there is a lot of life going on while I am neurologically unable to pay attention.  But I still think I can!  So, if you call me while I'm driving, my phone probably will ring...inside my glove box, where I cannot reach it.

How do I introduce my students to Linguachet?

I teach with a woman I admire a great deal.  She is a master at growing deep, lasting bonds with her

students.  Not only does this make her effective while teaching them, but she has a continual stream of former students stopping in for years after she has taught them, particularly at this time of year.

Her example is teaching me that all teaching happens through relationship.  If this is true, it's not just what what we do in a classroom that matters; how we do it matters even more.  Maybe sometimes, in our eagerness to bolt on a magical new teaching tool or method, we forget this relational context that colors everything we do.  And if we do that, then we may be puzzled and frustrated at the results we get.

With this relational context in mind, I'd like to offer three things you can tell your students about Linguachet. 

1. The point of Linguachet is to help you teach yourself, something even more important than Latin per se.  As your teacher, I'm very interested in helping you do that.
2. Nobody's perfect...but we can all get better. Learning from our own mistakes can help us a lot.  It's not where we start that matters.  It's where we finish.
3. Linguachet will never be perfect either.  It can make mistakes.  But if it helps us learn from our mistakes, then it will USEFUL to us.

An amazing former boss had the following sign on her door: "People may doubt what you say, but they will always believe what you do."   So, once I say these things, students will be watching to see if I mean them. 

Used purely as an electronic taskmaster, Linguachet would become a source of drudgery.  Used simply as a way to test and assign grades, it would become a source of anxiety.  Presented and used as a helper, though, it can become a source of gratitude.  

With our task environment in mind, I'd like to offer a few tips for using Linguachet with students.

1. Personalize assignments:  Now that it's easier to do, you can either differentiate or offer students choices wherever it makes sense.
2. Slash busywork:  You can tell students what you want them to learn, encouraging them to skip unneeded units or questions where appropriate.
3. Roam the room and coach:  Since students are now looking up basics
   on their own, you may be able to focus your attention toward a smaller number of students who need it.  If you're carrying your iPad, refresh your results tab and see who's logging lots of attempts.
4. Explore results:  If you didn't have to grade an assignment, why not use a fraction of the time saved to click through the results and explore how each student is learning?  Over time, what I noticed started to change my beliefs about teaching.  
5. Experiment with assessment:  Often I gave simple completion grades, only to find the students driving themselves toward 100% accuracy anyway. Sometimes I assigned several units at once, but told students to choose no more than 50% of the questions in each.  I also experimented with letting students choose between taking a quiz and doing Linguachet.  Students' insights helped me understand what I was seeing and decide how to respond.

If you think about it, these are all best practices anyway; they're just a bit easier to do now!  Maybe we can experience renaissance in the art of teaching as we refocus ourselves on the things that ONLY teachers can do.

Finally, passionate teachers never fully agree.  Maybe you'll even disagree with at least one thing that I've said above and give me some pushback in the comments!

Reinventing Latin Homework: A Gateway to Differentiation?

Today another teacher using Linguachet with her classes asked me how I make assignments with it.  I spent a couple of years experimenting with this myself, and I hope to hear other teachers' ideas very soon! 

I must warn you that seeing all this behavioral data from my own students has begun to change my thinking about how I teach. Formative feedback can be a powerful motivator, but - as we know - there is no single holy grail in education.  To use it most effectively, I've been slowly learning how to watch the process data, then differentiate assignments to keep every student moving forward in flow.  

When the task is reachable for them, I find that students will knock themselves out trying to get to 100% on every question, even if I grade for completion.  When a unit is short and not overly challenging, assigning entire units works well; weaker students will just take a bit longer (and improve more) to reach the same point, but everybody will still get there and that's no problem.  The struggling students often feel encouraged because they saw their struggle pay off.  As a teacher, I feel good because I've emphasized learning as a process, reinforced a growth mindset, etc, etc..

By the same token, however, that learning process is different for every student, as we can now see like never before.  I keep an eye on the average-number-of-tries-per-question column; that shows you how challenging each

More challenging exercises expose significant differences in learning process.

student found the exercise.  When the numbers are green or yellow, they're usually learning from their own mistakes, but over 10 attempts at a question (red numbers) can signal guessing and frustration.  Those students may be spinning their wheels.  I click that yellow part of the row for details to see which questions are troubling them, then click in again to see which words in a question are giving them trouble.  That helps me know how to help them.  

Since the results tab defaults to showing each student's latest unit, you can easily have everyone doing different things without killing yourself, especially during exam review. Sometimes I give ad hominem unit prescriptions ("You need to work on your noun-adjective agreement. Do unit 62.").  More often I have students self-prescribe ("Practice your weakest concept from our last test.").  Sometimes I tell students ("Practice the imperfect.  Do units 29-33, but only pick half the questions in each unit.  Skip the ones you think are too easy.").  This is why Linguachet computes accuracy totally  independently of completion; it allows you to emphasize quality over quantity.  When students see that I am willing to break the curriculum for their learning, they tend to respect the assignment more, and I spend less time policing completion.

The more challenging an exercise is, the more differentiation is needed to help keep everyone in flow.  This is especially true in the review units.  You'll probably find that your strongest students will do challenging sentences at least FOUR TIMES faster than your weakest ones.  For this reason, I now think it's best to assign challenging units (like the reviews) by total time ("Spend 20 quality minutes on this tonight.") rather than percent covered.    Since students can return to a unit again and again, it's difficult to compute total time spent on a unit. However, Linguachet will estimate the amount of time spent on each question (not including first attempt) when the data appear to fall in a credible range.  (If you really must verify something with absolute certainty, you can click three layers deep in the data, turn on the datetime column, and see the timestamp (GMT) for every single student response.  I believe the teacher features video on the site shows you how to do things like this.)

A really cool in-class differentiation strategy here is to train your strongest students (who finish sooner) to coach others by circulating the room and asking process questions ("What's the subject?  It that a 'she' or a 'they'?  Does that change the ending of your verb?") just as you might do.  My mantra to student coaches is, "Don't give them the fish; teach them to fish." This helps your struggling students get scaffolding that keeps them in flow while engaging the stronger ones in rehearsing these same skills in a more advanced way.

As the data expose just how differently our students learn, we'll ultimately want to create review units differentiated by difficulty; one student can handle all the grammar concepts packed into five sentences, while another needs them spread out over twenty so he stays "hooked" and can keep moving forward without shutting down. This is one of the many reasons I'm finishing a custom-content app for teachers.

I hope this helps!  Thanks for asking great questions, and please share whatever you learn!

Pride and Prejudice

Although it contains nerdy language, this post is really about listening...and how I need to learn to do it better.

About a week ago, a Latin-teaching FB friend emailed me to say he couldn't make an account in Linguachet, that the web address wasn't working.

In righteous consternation, I rushed to the site, tested yet again what I'd tested so many times before, and found it working.  Then I looked at the web address he sent me, which appeared to match nothing I'd ever created on the site (a good chunk of which I can recite from memory).  Where could THAT be coming from?

Sometimes what I think I know keeps me from seeing what I don't.

I tested everything I could think of.  But a few other people came on and created accounts without apparent issue or comment.  Like many of you, I am busy starting a new school year with my own students...so - like everything else in life that seems to be working fine - that comment eventually went to the back burner. 

Well, thankfully, I got a similar email today from someone else.  And it had the same bizarre web address. And I finally realized the source of the problem. It wasn't among the 6,700 (pre-compressed) files I have labored among for years. It wasn't in the complex server configuration that has taken me months of toil to learn and covers a large whiteboard on my wall.  It was just a plain-Jane web link on my own FAQs page, which was missing a single quotation mark.

And, being a helpful and tech-savvy soul himself, the first guy had more or less suggested that sort of issue on the front end.

Now, here I was, hyper-focused on making sure my app itself was running perfectly...gearing myself up to do more reeeeally complicated things...and missing the forest for the trees for that very reason.

Sometimes what I think I know keeps me from learning.

I wish I could say I didn't recognize this problem in myself.  But at least I can take comfort in implicating OTHER people too, right?  When I read it years ago, Pride and Prejudice seemed to be ALL about Lizzy's inability to see Darcy...mostly because she was too busy looking at herself.  Then I picture a flustered Atticus Finch at the end of To Kill a Mockingbird, who has so steeled himself to turn his own son over to justice that he is momentarily blinded to a critical fact: his son isn't guilty at all.  With all the zeal and integrity he can muster, the poor man is just barking up the wrong tree.

I'm so glad that Harper Lee saw fit to include Atticus' limitations.  I'm so glad my Bible is chock full of other men who just don't get it.  I am so grateful for the patience - and the pushback - of those of you charitable enough to call yourselves my friends, because sometimes I need a Heck Tate to rattle my porch.  In a world of casual likes in which many of us just move on when we don't, honest criticism may be the office of a true friend. And if you know any of us passionate souls who throw ourselves into causes, crusades, or businesses, be aware that we can lose perspective because of our intense investment; your honest critique could prove priceless.  Even if we end up rejecting your particular advice, we still need to develop the ability to see as others see if we're going to be able to serve anybody.  In this regard, I am grateful for the providential lessons of friends and failures, repeating themselves as needed, so that I can begin to see and hear what I don't! 

Whaaat? Did you say FREE??

Salvete, Magistri et Magistrae!  This will probably not be my most eloquent post.  As the school year begins, I feel like I am returning from another planet.  And this year, my to-do list is...sorry in advance...out of this world.

Do you ever find yourself in odd or novel situations, having to make important decisions without enough information?  If you are a teacher, the answer is probably yes.  After a while, this peculiar brand of insanity becomes your new normal.

I am not a businessman by training, and there are scores of things I do not know.  However, I AM a teacher, and there are many things I believe passionately enough to work and risk for them.  We don't just act on the things we know, for we never know enough.  We act on the things we believe, then learn what we need to know to live them out.

A bit over four years ago, Linguachet was a "wouldn't it be cool if" in my head.  From 16 years of teaching, I had realized writing in Latin put a lot of things together for my students, that it could be engaging and fun, and that it could be very concise and powerful in review.  From 30 years of general nerdiness, I believed that something like this was possible given current technologies.  After discovering that no one had created what my teacher brain wanted, I decided to take the plunge. At the core, it was a simple idea. I figured I'd be done in six months.

Have you ever been thankful for what you DIDN'T know?  I am really glad I DIDN'T know that Linguachet would take me almost 4,000 hours of work, thousands of dollars...and counting!  That's insane.  But when I got my first glimpse of students responding, when I saw how their learning process started to change, I knew this HAD to become reality.

When I started, I just believed that quick, specific feedback and flexibility would be game changers. I figured we might get more of students' best answers instead of just first answers.  I hadn't yet figured out that you could design a tool to help teachers differentiate instruction without killing ourselves...to blur the lines between formative and summative assessment, to promote learning as an ongoing process, to put teachers and students back on the same team, to give us new insights into HOW our students learn that we could then factor back into more effective diffentiation.  As a teacher, I haven't fully achieved all of these things in my own classroom yet either; it scares me a little to think of it all.  But what I have seen substantiates my belief that we truly can do it...and that it can breathe even more new life into the teaching of Latin... and that is what keeps driving me back into code after full days of teaching. 

That last paragraph makes it sound like I've stumbled onto some sort of Holy Grail.  In actuality, I don't believe there is one; I believe that good teaching is an insightfully blended cocktail of divergent - even contradictory - tools and methods.  I believe it must be insightfully blended - with love - by you, the person who knows and cares for your students.  Because I believe that teaching is love, I don't believe that technology can EVER replace a teacher.  What I do hope is that this tool gives you another lens by which you can see - and an extra arm with which you can help - your students.  Maybe it can free you just a little bit more to do the things that ONLY teachers can do.

I know teachers need volume discounts.  To work out the details, I need real-world data on how many students I can serve well at one time.  The software industry has simulators, and I have tested many of them, but the ones good enough to give predictive data in this case cost serious bling...and they know nothing of rotating class schedules, etc!  So, I could go that route - and increase costs, and still be guessing - or I just beef up my server by a lot, take another chance, then crowdsource the problem, and test with real-world traffic as I explain below...

I know nothing about marketing either, but I believe you start by serving.   Unfortunately I have learned that many schools will refuse to look at a tool based on price alone...without even looking at the learning ROI...and their students will miss out.  So, fellow teachers, take this into your classroom for free between now and Sept. 12th.  Make new accounts.  Play with it.  Have your Latin I students make accounts and play with Latin.  Watch some kids race ahead because they can.  Notice what kinds of questions you get.  Give little review prescriptions to your Latin IIs...or let them choose their own areas of weakness.  Then go explore the data and see what you learn about your students.  Consider having your administrators in to see how your kids are learning...and let them see the data too.  For those at ACL this summer, I'm clearly behind schedule but haven't forgotten what you shared with me in those surveys.  Send me honest feedback and ideas at rietz@linguachet.com.  

Per aspera ad astra,

George Rietz