Aidan Pine

My name is Aidan Pine. I'm a settler of European ancestry born in Victoria, B.C. and I developed this site in order to bring together the various themes of my research as well as the language revitalization and open-source software projects that I am involved in.

I fell in love with language revitalization work as a way of combining my interests in language, linguistics, anti-colonial/decolonial methodologies, and computer science.

I am a full-time researcher at the Canadian National Research Council, where I was most recently technical lead for the Speech Generation for Indigenous Language Education project (2022-2025). I am also the founder and lead developer of Mother Tongues, and former technology consultant for the First Peoples' Cultural Council, and the University of British Columbia.

Photo of Aidan Pine

Research Themes

The sections below help connect my publications to four core themes that run throughout most of my research. You can also view a thematically organized timeline of my publications or a filterable list of all of my publications below.

Language Revitalization

I have loved language ever since I was a child. But, it wasn't until I was an adult that I learned that the linguistic reality of the place I grew up in had been hidden from me.

I learned that this attempted erasure was the result of myriad colonial policies aimed at replacing Indigenous languages with English or French and removing Indigenous people from their lands.

I strongly believe that being an effective supporter of language revitalization efforts requires understanding the very harmful processes of language de-vitalization that led to this situation in the first place.

Being able to direct my passion and interest for language into something that supports language revitalization and reclamation efforts has been an endlessly fulfilling pursuit.

Speech Technology

Many communities I worked with had a seemingly simple request: to be able to hear the language in the text-based tools we had been developing together. However, recording the tens of thousands or millions of words that would be in the average dictionary or verb conjugator would be a near impossible task, so the research question became: "What is the minimum number of recordings needed to build a TTS system good enough for an educational setting?".

That question brought me to the University of Edinburgh for my M.Sc., led to a Best Special Theme Paper at ACL based on my dissertation work, and eventually to a three-year grant collaboratively building community-owned TTS systems for SENĆOŦEN, nêhiyawêwin, and Kanien'kéha.

Digital Lexicography

As a First Nations' Languages and Linguistics major during my undergraduate degree, I worked to support the Gitksan Dictionary Project where I worked with elders and speakers of the language to document words and phrases for an online dictionary. The final step of turning the carefully documented dictionary data into something accessible to learners was expensive and challenging. Many web development companies at the time developed products that didn't properly render symbols in Indigenous languages, and the search algorithms they used were designed around English.

This became the focus of my undergraduate honours thesis, and later led to the Mother Tongues dictionaries; a free, open source tool which allows for customizable approximate search and has been used for dozens of Indigenous languages world-wide.

Foundational Language Technology

Speech synthesis and other cutting edge technologies are cool and all, but if the language in question doesn't have reliable access to a Unicode input system/keyboard, then what use is text-to-speech? There's a surprising amount of foundational language technology that is required to build other language technologies. Many NLP researchers working on languages like English take Unicode, keyboards, phonemizers, and forced alignment for granted, but these present real challenges.

One thread of my research is building tools that are either directly requested by communities (like text-audio aligners or orthography converters), or prerequisites to other desired digital technologies (like grapheme-to-phoneme engines).

Research Publication Timeline

The timeline below shows my academic publishing record according to the above themes. While Language Revitalization work has remained constant throughout my career, other themes follow a different pattern. For example, I only start working on Speech Technology after completing my M.Sc. in Speech & Language Processing at the University of Edinburgh in 2021.

Publications (25)

Google Scholar
Presentations (55)

Contact

If you're a language revitalization practitioner or a researcher interested by similar things, I'd love to hear from you! If you're a student looking for summer internship possibilities at the National Research Council and have experience in related research areas, please reach out to my NRC email.

[email protected]

GitHub: @roedoejet