Working with a Linguistic Treasure Trove Found in Greek Dialect Features in the Speech of Greek Canadians

In 2020, Dr. Panayiotis (Panos) Pappas was awarded a Social Sciences and Humanities Council (SSHRC) Insight Grant for his project, Dialect features in the Greek of Greek Canadians. Professor and Department Chair, Dr. Pappas’s research mainly focuses on language change in Greek. We recently spoke with him to learn more about his area of research in this first of a series of Linguistics Research Spotlights.

Working in collaboration with linguists from the Modern Greek Dialects Laboratory in Greece (University of Patras), Dr. Pappas and the research team are exploring foundational issues in quantitative sociolinguistics by examining key dialectal features of Modern Greek in the speech of Greek Canadians. The research team is compiling a linguistically annotated machine-readable corpus of Modern Greek spoken by first generation Greek Canadians who arrived in Canada between 1945 and 1975. These interviews were recorded and transcribed during a previous interdisciplinary project, Immigration and Language Change in Canada: Greeks and Greek Canadians (Immigrec), which was completed in December 2018 and on which Pappas was an Associate Investigator.  

Recordings provide a “time capsule” of dialectal areas of Greece

Between 1945 and 1970 more than 107,000 Greek citizens immigrated to Canada, making this country one of the primary destinations of Greek immigration. Pappas explained, “Before this wave of immigration started, there were only about 12,000 Greeks in Canada. From a country that had maybe 2 million people at the time, that was a very significant increase in the community here.” The participants in the Immigrec project represent almost all dialectal areas of Greece, both sexes, as well as a wide range of educational backgrounds. “The speakers came to Canada from all different areas of Greece, and interestingly, also from areas of the Greek speaking world that were not Greece, for example—refugees from Cypress, Alexandria Egypt, Istanbul, and South Africa,” he elaborated.  

“The data are enormous in quantity. We have over 430 interviews (many of which are over an hour), yielding about 300 hours of recordings,” Pappas explained. Most of the speakers are over 70 years old and many of the dialects represented in the sample have suffered attrition in Greece itself due to the spread of Standard Modern Greek. The recordings complement the efforts of the Modern Greek Dialects Laboratory at the University of Patras to preserve dialects of Modern Greek which are quickly vanishing. “The people who came to Canada were never forced to learn a standard that could eat away at their dialect,” said Pappas. “My Dad, for example, from the 1970s on, had to convert towards Athenian and learn that dialect. But his friends who left Greece and came to Canada in the mid-1960s, never had to convert in the same way as my father. Dialect speakers in Greece were under a lot more pressure to abandon their dialect than dialect speakers in Canada. The recordings are a bit of a time capsule.”

Research objectives

The first phase of the project (now nearly complete) involves the creation of a linguistically annotated machine-readable corpus of Greek spoken in Canada using the interviews outlined above. In the analysis phase (currently underway), researchers will quantitatively study four key dialectal features: unstressed vowel raising; palatalization of lateral and nasal consonants; the form of imperfect past tense stems for a subset of Modern Greek verbs; and use of the unstressed prefix for the formation of past tenses. Once the analysis is complete, Dr. Pappas and his collaborators will work together to examine the empirical and theoretical ramifications of the findings.

Highlights and vision

Dr. Pappas was invited to speak about his research at Ohio State University’s 2019 Kenneth E. Naylor Memorial Lecture which brings leading scholars in South Slavic Linguistics to OSU to present a public lecture and to speak to students. “That was a big honour, because it was my alma mater,” said Pappas. He was also invited to speak at La Trobe University in Melbourne, Australia in the same year, which has spawned a parallel project: “My colleagues there are now working on getting a similar project about the Greek of Greek-Australians off the ground.”  

Involving students and training the next generation of researchers is a key part of Dr. Pappas’s work. “I’m really excited about the work we’ve been doing with Laurens Bosman,” Pappas shared. Currently completing his Bachelor of Computational Linguistics at SFU, Bosman worked as a Research Assistant with Panos and created a software pipeline using the Montreal Forced Aligner to automatically align the interview transcripts to the audio recordings. Bosman explained: “I trained a language model using the Greek data that we had access to and used this model to fine-tune the forced alignment process to Greek. By making use of this technology we were able to speed up the process of aligning phonemes to the audio. We were able to align a few hundred interviews by running the pipeline overnight on SFU’s Discourse Processing Lab server.”   

“If we continue on this trajectory of success it will be a significant contribution to Greek Linguistics,” Pappas added. While the Montreal Forced Aligner had been well-used with many of the world’s languages, it hadn’t been used with Greek—until now. “If it does work the way I’m hoping it will work it will be a major breakthrough,” said Pappas.

Dr. Pappas is aiming for increased global collaboration as well. “The private foundation that sponsored the Immigrec project, the Stavros Niarchos Foundation, is quite impressed with the work we’ve done. We’re discussing how to use what we’ve organized so far to become a template for other Greek programs around the world. My hope is that eventually we will develop a picture of what Greek diaspora studies can be based on. That will give the project greater longevity.”