Joan Corinne Bachenko

Joan Bachenko began her public career in linguistics as a graduate student when she and her colleague, Virginia Teller, gave a talk to a standing room only crowd at the 1974 Linguistic Society of America Summer Institute arguing against the highly disputed tenets of Generative Semantics. The talk was well-received as much, I think, for their courage as for its well-argued content. Joan´s desire for elegant explanation and her drive to build useful language tools were stand-out qualities in her professional life. Her courage and her ability to nurture characterize both her professional and her personal life.

Joan began her research career in 1982 at the Naval Research Lab in Washington, where she worked on a system to analyze military messages including the parsing of the sublanguage effects that arise in these messages. Always with explanation in mind, she hired Don Hindle, who brought his deterministic parser, Fidditch, to the task of sublanguage parsing. Joan hired me to build the dictionary of navalese to support the parser. Joan was delighted by an oddity that we discovered in the messages: verb usage in the messages was consistently different from verb usage in standard English, a fact that explained how the messages could be understood unambiguously even though they were pared down versions of standard English.

Don had also been consulting at AT&T Bell Labs with Mitch Marcus who invited Joan to present her sublanguage work there. This resulted in the offer of a member of technical staff position to Joan, who moved from D.C. to what became her beloved New Jersey, first in Hoboken and later out in the New Jersey horse country in Oxford, NJ.

At Bell Labs, Joan continued her work on deterministic parsing as it applied to speech synthesis, in particular to prosodic phrasing, an area that Mark Liberman encouraged her to explore as an area in need of research. This was Joan´s most productive work, with twelve publications and a patent that converted written Deaf English to spoken standard English for phone messaging. It was adopted by AT&T subsidiaries. She also developed many friendships that she nourished at the labs and sustained for the rest of her life.

In the late 1990´s, family needs took Joan back to her roots in Minnesota, where she and Michael Schonwetter co-founded Linguistic Technologies, Inc. with the Mayo Clinic as its first client and the task being the transcription of doctors´ recorded notes to text. Joan employed transcribers, who typed the text as had always been done, and a research team who worked on automated speech recognition, in particular to convert the disfluencies of the recorded speech into readable English. Joan was also teaching at the University of Minnesota and mentored two Ph.D. candidates, Serguei Pakhomov and Guergana Savova, who became part of the LTI research team.

Joan´s courage was at the fore here. Building a brand and a clientele, and making payroll every week were not skills Joan trained for but her grit turned LTI into a successful company, bought out by Lernout and Hauspie in 2001. At that point, Michael was approached by a legal firm asking whether it would be possible to use natural language processing (NLP) tools to detect deceptive speech. Rising to the challenge, Joan investigated the state of the art in psychology research on deception, saw a niche for the use of NLP and, again with Michael Schonwetter, founded a company, Deception Discovery Technologies, which took on a few legal cases in Minnesota before Joan moved back to New Jersey, bought a farm that she converted into a horse farm, Golden Pastures Equine Retirement, and invested a good sum in Montclair State University´s Linguistics graduate students to continue the experiments in automatic deception detection. Within two years, Joan, the students, and I were engaged in work with the Center for the Advanced Study of Language (CASL) at the University of Maryland analyzing job interview data. The work achieved a 92% accuracy in identifying true and false statements based on the words chosen by the speakers. While the study was classified and not published, it was used to train interviewers in successful methods of obtaining truthful statements from interlocutors.
Most of the deception work at this time used fabricated laboratory data because it is easy to verify. But Joan and I were dedicated to using real world data and found a colleague, Tommaso Fornaciari, who held our view of the data, using Italian court data for his deception studies. With Tommaso we organized two workshops on deception detection at EACL in 2012 and NAACL in 2016, and authored a text for Morgan and Claypool, Automatic Detection of Verbal Deception in 2015. Joan continued to work on deceptive language until the pandemic, when she became ill.
Joan was a wonderful colleague, exciting to work with, able to see new opportunities, ready to take them on and produce useful tools while always looking for elegant explanations for facets of the phenomenon she was studying. She also had a remarkable ability to nurture. She played a major role in the early professional lives of her research assistants and students both at Minnesota and at Montclair State. She played a similar role in my life, employing me at NRL and then at Bell Labs, which set me up for a tenure track position at Montclair State. And it´s not surprising that the horses, cats, and rabbits in her care thrived, as did her partner, Dave Vanderhoof, and the young helpers at the farm, who flourished under her instruction. She is deeply missed by us all.

Remembering Joan...

Last month, I received a phone call that my dear friend Joan Bachenko passed away. I knew she´d been having health issues, but I didn´t realize the dramatic downturn she´d taken. I didn´t get out to see her while I was helping my mom, so I wanted to say just a few things, remembering her friendship.
Joan was a great friend! She was smart, driven, genuine, modest, thoughtful, and loved animals-especially horses! Her colleagues would tell me she was "a genius" at her work! She worked at Bell Laboratories/AT&T Labs before moving to St. Peter, Minnesota to start a new research development company called Linguistic Technologies, Inc. with a couple colleagues. The office building on Third Street that is now occupied by State Farm Insurance was once the place where she, Dave Schulz and others were diligently working to develop speech recognition software, geared toward the medical industry. Those five years....1995-2000 were the years I really got to know Joan. We´d meet for lunch at Ooodles Café, take our break from the workplace....and laugh! Those were such wonderful breaks in the day!
Joan (and her co-owners) sold LTI, to a firm out of Brussels, a company that also purchased "Dragon," another speech recognition software. After that sale, Joan moved back out East closer to old friends and colleagues. She purchased a hobby farm in New Jersey, about an hour outside of New York City. She boarded horses along with her own horse, Flyer. She met Dave, a farrier (horse shoer,) who soon became her partner in life on the hobby farm. I visited Joan in New Jersey and remember how she was NOT a lover of cooking....In fact she did more grazing than eating entire meals. She asked me to assist her with a brunch for some friends, and I still recall getting into the car with a box of pancake mix, and Joan very seriously told me, "Study this!" I looked at her, laughing and replied, "Joan, it´s pancakes! Ya add a little milk or water, maybe an egg....and pour it on the griddle!" Then we both just burst out laughing! Neither one of us was afraid to laugh at ourselves....maybe that´s why we got along so well!
I also remember her phone calls when she was working on a new fraud detection project. She was so excited when the US Department of Defense was interested in the project, but before Joan could make her presentation to the US DoD, she had to pass a lie detector test. She was given three chances to pass the lie detector test, and she was so nervous that she flunked the first two tests! (Now I understand why the tests are NOT submissible in courts.) She was calmer for the third test, and alas, she did pass!
Back in St. Peter, I remember Joan lived in the third-floor apartment that was remodeled by Carlstrom Construction...the old Nicollet Hotel. Joan loved the location as it was on the same block as her office. She really liked her apartment, but after the ´98 tornado hit St. Peter, she would have to move after it damaged her building. Actually, Joan was IN her apartment during the tornado! She, like many others, didn´t think it would really hit the town, and when she realized the tornado was indeed hitting her building, she could not get downstairs! On instinct, she grabbed a plastic laundry basket, put it over her head and stood under a landing that seemed safer than other spots in her apartment. She was not hurt, but she could see the sky from inside of her apartment and was shocked to see the demise of her building when she did get outside!
Another great memory I have of Joan, was when she invited me to help host some of her investors at the Margaret Thatcher Dinner at the Convention Center in Minneapolis. It was a fun event, and I made a fool of myself not recognizing a former Minnesota Viking, yet he and his wife were so very gracious! I was embarrassed, and talked with them a bit before I was seated next to a surgeon who would not let me get a word in edgewise the rest of the evening. It was a memorable dinner!
Joan, I´ll remember laughing with you over the years! I remember when I made that checkerboard cake to celebrate with my parents after they "opened up" for the Tina and Lena Show fundraiser. (Mom and Dad played piano.) I tried to make that cake in honor of my Aunt Per, and the darn cake was raw in the middle, and took so much time to make and had NO flavor! Everyone was so quiet around the table, until I said, "This is terrible!" Then everyone burst out laughing! I remember when you told my dad you´d have a "Virgin Mary" to drink at a party, then Dad came back from the bar to ask you, "Virgin WHO?" Again, we laughed! Ah, I wish I had been able to get out to see you this past month, but I am so grateful you were able to visit my place in Florida prior to COVID! Thanks for the years of friendship and laughter, Joan! RIP, dear friend!
Lisa

Joan and I met in September, 1971 in my first class on my first day of graduate school at New York University. Together we navigated our way through the Linguistics PhD program, surmounting every hurdle along the way, and became such close friends that our department chair started calling us the Bobbsey twins.

Our careers took different trajectories, but we always stayed in touch. I visited Joan during her years in Minnesota, but our visits became much more frequent after she moved back to New Jersey in 2001 and bought her horse farm, where she lived for 20 years.

Joan came into Manhattan (where I live) regularly, and we would go to the opera, theater and other events together. Best of all, every couple of months we would meet for lunch before going to our hair stylist Liell (a Midwesterner like us) for joint appointments, and the three of us would engage in lively conversation the whole time we were there.

In 2021 Joan was diagnosed with dementia, and she spent her final two years in assisted living and memory care in Hackettstown, NJ. As her Power of Attorney and Health Care Proxy, I managed all of Joan´s affairs during this time, visited her frequently, and did my best to ensure that she got proper care and treatment.

The photo I chose to accompany my remembrance was taken in July, 2009 on her farm. It captures the essence of Joan in my mind - she´s on the horse farm where she lived her dream for 20 years with her devoted partner Dave Vanderhoof, and she´s standing next to a portrait of one of her beloved horses.

Joan - I love you and miss you. As always, Virginia