Coming up in two and a half weeks, McGill will be hosting a computational field (linguistics) workshop.

On May 27th, we’ll have two talks and a wine & cheese reception: iLanguage Lab will be presenting about LingSync, and Alexis Palmer, a postdoc working for SEASIDE, will be giving a talk entitled “Computational Linguistics for Low-Resource Languages.”

May 28th, we’ll have two workshops and two talks: iLanguage will host a session entitled “Plugging in to LingSync” and Alexis Palmer will coordinate one focusing on what exactly one should do with field linguistic data. Our talks will be from Erin Olson (presenting on the Prosodylab’s Forced Aligner and its use in segmental analysis) and Robert Henderson (“Reclaiming SIL Bibles for Linguistic Research”).

More details available in our newly-minted workshop section!

The L’nui’sultinej Conference is the place to be for Mi’gmaq speakers and learners; as such, we are proud to announce that we are planning on going to Antigonish to talk about our projects and our classes!

As the organizers say on their site:

“L’nui’sultinej na tesipunqek mawita’mk wjit Mi’kmaq Kisna ta’n pasik wen ta’n ketu’ kinu’tmasit l’nui’sin kwlaman kisi klo’tesnu ksitunnu iapjiw. Ula mawita’mk wjit msit wen aqq ta’n pasik wen wlta’sualaten.

L’nui’sultinej is an annual conference which brings together Mi’kmaq people from across the Mi’kmaq Nation to look at education issues related to language preservation, enhancement and revitalization. It is a conference for parents youth, elders, community members, educators, administrators, counselors and researchers. We welcome all Mi’kmaq into the circle whatever their level of language proficiency.”

Mali-Beth, Carol, and I (Elise) will be attending the conference to talk about how linguists and speakers and learners can work together, inside and outside a language classroom. We hope to see you there!

Since we’re getting a lot of work done with the LingSync application* from the programming side of things, we decided it was high time we start using it the way it was intended to be used; as a database making it easier for us to share data and collaborate.

As we’ve started putting in data, we’ve been laying down some conventions for us to follow in the rest of the database. This blog seemed like a good place to discuss the conventions we’ve established, and an even better place to debate new conventions for areas we haven’t fully fleshed out yet (verbs…).

All of this information and more is stored also on the specific wiki page for LingSync glossing: http://wiki.migmaq.org/index.php?title=LingSync_Glosses

Our general guiding principles are as follow:

• Gloss everything–no defaults!
  • This is mainly to make search easier and more intuitive. If we had, for instance, “animate” as the understood default person and only glossed inanimate morphology as such, it would be very difficult to get a datalist of all animate words. Having no default glossing means that all our glosses will be very explicit and therefore easy to search.
  • This will be painful at the start, but once we have enough data in there, LingSync will autogloss and make our lives much easier! Hang in there.
• When in doubt, don’t parse it out!
  • Only separate morphemes if you and a collaborator are completely, 100% sure that they are separable. Make sure you pass your theory by someone else’s eyes first, too!
  • Feel free to use dots frequently in your glosses. It is safer, generally speaking, to group morphemes (and later split them up) than it is to be over-enthusiastic about splitting them up (and later having to go back and re-group).
• In general, be faithful to the surface/pronounced form when drawing morpheme boundaries. (ie match the morpheme line to the utterance line as closely as possible)
  • Please use the Notes section to leave comments about phonology if you think there is a predictable process going on!
  • (One exception is the palatalization of ‘t’ at morpheme boundaries. Throughout LingSync we will assume that t -> j / _-i, so it is safe to have the utterance and morpheme lines different here.)

As far as specifics go, there is more information on the wiki page itself (WordPress hyperlinks seem broken, here’s the address again http://wiki.migmaq.org/index.php?title=LingSync_Glosses ).

Please use the comments to discuss…

• Verbs! Since we are glossing with the maximal amount of information, including tense and mood (ie. present indicative), where should we put this information? So far we’ve been sticking it onto the end of the root using dots (ie. tli’ma-tis = tell.TA.PRES.IND-1SG>2SG). Any other suggestions for the placement of tense/aspect/mood?
• Verbs part 2! How should we identify the difference between various evidentialities? And what about tense, aspect, mood? Right now we’re marking present indicative, imperative, future, evidential past, inferentialiOpen discussion in comments below!

*(The LingSync extension can be found here: https://chrome.google.com/webstore/detail/lingsync/ocmdknddgpmjngkhcbcofoogkommjfoj )

It’s been pretty quiet on the blog for the last little while, but that’s not because we’re working less on the project! We’ve been working a whole lot on a big aspect of our Mi’gmaq Research Partnership: specifically, trying to get money.

This past month, we’ve been working on our application for a grant. It’s allowed us to get a good look at the different strengths and goals of this project, and we’ve begun to get very excited about the coming work! Turns out, we’re doing good work, and there are a whole lot of interesting areas for our project to grow and explore.

• LingSync application/Chrome extension. This is a really exciting application (at least, exciting for linguists) which will let us store, organise, and selectively share our language data (with encryption on all of it for the sake of our speakers’ privacy). Right now, we’re working with iLanguage on building on the existing code. On our to-do list is…
  • Making a Conversation! LingSync expects people to enter data as single lines, unconnected to each other. What we’d like to do is add another level, which lets users add conversations, dialogues, and discourse so that the separate lines of data are linked to each other just like speakers connect to each other in real life.
• Mobile language-learning app. We’ve got (very rough) Android and Chrome extension prototypes for this already, and have a lot of ideas about where it’s going to go from here. The parts that we want to focus on the most are…
  •  its online/offline abilities, since a lot of language-learning apps work best online and have very limited offline capabilities
  • its tool for also letting students build their own lessons! Most language lessons are “read-only”, and students are expected to consume what is put on their app and be satisfied with that. But what we want to do is let students create material with speakers in their lives, and tailor their app to their own needs.
• Master-Apprentice program. This is my favourite part of this project right now. It’s a wonderful opportunity for learners of Mi’gmaq to get some support and guidance, and partner one-on-one with fluent speakers. Again, get in touch with Vicky or look at an older blog post for more details about it!

Hello, folks! Reporting from Listuguj here, where there has been a lot of work done towards setting up CAN-8 for real live users. We’ve been focussing on

• making connections with (potential) users of the program
• establishing a work-flow that works for people involved in the project (more on this below), and
• recording and uploading content

We’ll start off talking about that first point, making connections. We’re hoping to create a CAN-8 that is geared towards the people using it. Part of this is listening the input of anybody who has an opinion to give. Vicky Metallic has been getting the word out to people outside the Education Complex that the CAN-8 project is happening, and that another tool for Mi’gmaq-learning will be added to the considerable resources that the community already has at its fingertips. We’ve all been talking to people, and in doing so, we’ve also been taking the advice of the people who have seen the program. We’ll use their input to select scenarios for dialogue-making, so that our content is relevant to the lives of the people who are using it.

Next up, the question of “how do the dialogues get made, anyway”? The fine details of this process are still being hammered out, but we have a general outline that works well for us! We write scripts in part-English, part-Mi’gmaq, then Joe Wilmot and his trusty team work with us to tweak the scripts so they flow like real Mi’gmaq. Once the scripts are speaker-worthy, Joe et al record them. The resulting sound files and Mi’gmaq scripts go into CAN-8, so far by my hands, but hopefully this will become a Montreal job soon. And later, we will write up helpful meta-linguistic notes which also go into CAN-8 for the users’ enjoyment.

There are two kinds of dialogue: implicit teaching ones, and daily life ones. These two categories do overlap quite a bit sometimes, but they aren’t one and the same!

For the implicit teaching dialogues:

• Conor and I work on establishing corners of the grammar that might be tricky for learners to pick up directly from speakers–there is a lot that people can learn from the speakers around them, but we’d like to give people some scaffolding to base that knowledge on, and make speaker-knowledge more accessible to more people.
• Once these corners are picked, we then use a theme to construct a little scene. One starts with a boy telling a girl, “I love you!” and she replies, “You love me, but I don’t love you.”–they talk to each other, and explain the predicament to others, and (spoilers!) he eventually wins her over. The aim of this dialogue is to demonstrate TA morphology in a (hopefully) fun and funny way!
• Lastly, we give an overt, meta-linguistic description of the phenomenon in question. We also encourage the CAN-8 learner to go back to the dialogue they just heard with these meta-linguistic notes in mind, and will provide a CAN-8 space to experiment with the forms on their own.

The daily life dialogues are a little different. We don’t have many of them, but we’ll collect more as the summer goes on!

• Two (or maybe three) speakers record a mostly-improvised chat about topics that are likely to come up. One chat that we have already uploaded features a phone call between two speakers, where one invites the other to go across the river for some shopping.
• We then go over the recording with the speakers in question, and transcribe the exchange. This transcription gives CAN-8 users a chance to familiarize themselves with the orthography we’re using, as well as something to hang on to (as this speech tends to be a little faster than the scripted dialogue speech).
• The dialogues are then uploaded! The advantage of these improvised chats is that they are less artificial than the scripts, and will give the CAN-8 user a window into natural speech that they are likely to hear (and eventually produce) with the speakers in their lives.

We’ve also uploaded the rhythmic profiles that we’ve talked about in a previous post, and Roger Metallic is working on Mi’gmaq-izing some of the Akwesasne verb conjugations and integrating them into the Listuguj CAN-8, too. There’s a lot going on, here!

And, as always, please chat with us in comments!

• What do you think of our plan and strategy? What are the things that we’re missing?
• Do you have any suggestions for material to cover?
• Do you have suggestions for new methods to use to show linguistic patterns?
• What have been your favourite bits about self-directed language-learning programs in the past? Should our CAN-8 program have them, too?