Queer Storytelling

Part I. Queer in Film

International flights these days usually have seatback entertainment systems that allow you to watch movies to pass the time. Over the course of two trans-Atlantic flights last month, I watched eight movies, all of them for the first time. They were, in order: Star Wars: The Last Jedi; The Greatest Showman; The Shape of Water; Horrible Bosses 2; Love, Simon; Lady Bird; The Disaster Artist; and Moonrise Kingdom. I enjoyed them all!

I want to talk about two of these movies in particular, though: Love, Simon and Lady Bird. They’re both teen coming-of-age films; the former has been raved about by a lot of my friends on social media, and the latter was nominated for an Oscar.

People have reporting crying after watching Love, Simon, and most of my gay friends are thrilled by the representation we got on the big screen. That’s really valid, of course: it’s 2018, and Hollywood still has such a small selection of teen comedies with a queer protagonist. Most are indie films. Many have intensely dramatic or traumatic storylines, so they’re not really comedies. Very few have any leading roles that portray queer people of color. From any vantage point, this movie was sorely needed.

lovesimonLove, Simon really does a good job of portraying a “normal” teen life; however, I use scare quotes because obviously everything is a little too polished, everyone is a little too beautiful, every scene a little too efficient. I’m a fan of the film and would definitely watch it again, but it’s clear that it follows a kind of formula for popular films that does not necessarily adhere to reality.

On the other hand, we have Lady Bird, another coming-of-age film about a teenager suffocating in Sacramento, where she feels like there is no culture and no escape from her family’s poverty. The protagonist isn’t queer, but as I watched, I felt really drawn to her story and her circumstance, because it felt eerily similar to my own adolescence. I grew up in a “boring” northern California city, and my upbringing was also heavily marked by organized religion (although unlike Lady Bird, I was 100% willingly involved in church life and academically motivated, not forced to go through the motions at a strict Catholic school).

I think that what made Lady Bird so relatable was the way it told its story through scenes and situations rather than a strictly sequential narrative. I find that I tend to enjoy films that do this more: think Wes Anderson or Hayao Miyazaki films that are heavy in characterization but light on plot, or films like Amélie, Moonlight, Boyhood, and even Forrest Gump. I’ll never find myself in Simon’s specific and weird blackmailed-Cyrano high school love polygon, but Lady Bird’s life circumstances will feel instantly familiar to a whole generation of Californian not-quite-city kids.

ladybirdAnd that’s to say nothing of one very short but emotionally packed scene in which Lady Bird’s ex-boyfriend, whom she had discovered kissing another boy earlier, breaks down in tears while asking her not to tell anyone about his secret. Honestly, watching that almost brought me to tears, out of empathy and an acute personal understanding of the character’s situation. Nothing in the entirety of Love, Simon — not even Simon’s parents’ feel-good heart-to-hearts with their son — moved me the way that a good Christian boy’s realization of how close his perfect life was to crumbling away did, in less than ten seconds, to boot.


There’s a lot to be said for realism on screen. Sometimes you don’t need to go to the lengths of manufacturing a complex narrative to capture your audience; real life is dramatic enough. What’s clearest on my mind as I write this is the first episode of the second season of Queer Eye, the Netflix reboot of the popular “gay guys give makeovers” reality show.

Part II. Queer on TV

Queer Eye is all about spreading the love and doing so in multiple dimensions. The Fab Five, our proud gay protagonists, are of course the poster boys for acceptance of LGBTQ people in American society, but they also spend every episode talking about everything that’s beautiful about the bodies and personalities of people that generally come across as being losers. And they are always overflowing with love for every person they meet, from old white Trump supporters to a devout Christian family. You can see in every word and interaction a determination to be simultaneously the gayest versions of themselves and the friendliest, most compassionate, and least hateable people you could ever meet.

Yes, it’s a reality show and it — by ironic definition — has to stretch reality a little bit, but the story behind this season premiere was astoundingly raw and beautiful. In the episode, the Fab Five are assigned to help a zealous Christian woman in the small, conservative, and predominantly African-American town of Gay, Georgia (you can’t make this up) by cleaning up her house, helping construct her church’s community center, and gently buffering her reconciliation with her gay son.

You might think that the Fab Five would feel a little uncomfortable coming to a small Christian community. In fact, one of them, Bobby, has a really rough history with his own Christian community, one that he had loved and served all throughout childhood, until he came out of the closet and was swiftly kicked out. But Karamo grew up Christian and says that he has only felt love and acceptance from his community despite his sexuality, and Tan is a proud, gay Muslim from the UK. Within the group, there were a lot of different perspectives and backgrounds concerning religion and spirituality, and that was already a great starting point.

The real star of the episode, though, was the woman they helped to make over, Tammye. From the very beginning, she was full of love and acceptance for the Fab Five. She welcomed them into her home and into her church, and even though Bobby didn’t want to set foot inside of the church due to his past trauma, she smothered him with just as much affection1. Tammye explained to them that her gay son had recently moved back home and was feeling apprehensive about rejoining his childhood church because of the way he had been treated. In a truly moving scene, she told the story of how when her son first came out to her, she had not accepted him because of her conservative religious views. But after a few years, she had a change of heart and realized that to love her son no matter what was really what her belief in Jesus called her to do. So she sat down with her son and offered an apology for the way she had made her love conditional, and she asked for his forgiveness.

I don’t think I can describe very well using words how powerful this was to watch. On a personal level, I think it shook me because I thought about my own parents and realized that I have never had an open and honest conversation like that with them. I wouldn’t count on them ever apologizing for how they handled my coming out (I didn’t get kicked out of the house or anything, but it was very uncomfortable nonetheless), and I don’t feel like I actually need one, because I understand their views very well. But to see an example of what this kind of parent-child reconciliation could look like was poignant in a way that, again, I think no fictional narrative could come close to replicating.

But as most allies should know, it can’t just end with the apology. Near the end of the episode, Tammye gave words of exhortation and gratitude to each of the Fab Five in a heartfelt scene. It struck me as unusual that she would invoke words and a style that I only associate with religious blessings, and then I wondered why I thought that was unusual: if Christians really believe in the power of a spoken word blessing, why do we only reserve it for other Christians and not for every last person we meet?

Then, the episode closed with Tammye getting up on the pulpit at her church for the opening ceremony of the community center and delivering a forceful appeal for gay acceptance. She admonished her church for not loving unconditionally (what would Jesus do?), and she told them all about how proud she was of her gay son, while he was in the audience. I found myself in tears again.

This is what active reconciliation looks like. The straight, conservative church can’t just stop protesting against LGBTQ rights; that’s ceasing to harm but it does not heal. They can’t just say sorry for the way they have promoted discrimination for decades; that’s a first step, but it doesn’t require any effort or real sacrifice. To truly reconcile the LGBTQ community to the church, the church has to go out on a limb and fight for equality and rights hand-in-hand with LGBTQ activists. What Tammye did at her church was just a small example of what could be done nationwide — what must be done — in the name of Jesus.

Part III. Queer in Real Life

I guess that what this is all boiling down to is that the best queer stories you can find are probably going to be the ones that come from real life. By “best”, I mean the most influential, whether that’s because they are more heartwarming or heart-rending… it’s probably just because they are just more authentic. So by all means, please continue to support queer narratives in entertainment, and follow the works of queer writers and storytellers. But for every rom-com or reality TV show you consume, remember that there are LGBTQ people all around you who have compelling stories, and if they are willing to share them, it would be your privilege to listen.


Tan, Jonathan, Bobby, Karamo, and Antoni from Netflix’s Queer Eye

– – –

1I will note that despite Tammye’s best intentions, I don’t think it was fair for her to make Bobby revisit a site of trauma and expect him to get over years of pain in one week. But it also isn’t clear to me whether Bobby’s aversion to churches was really that deep-seated, or just a bit of “stunting” for some increased dramatic tension. This is reality television, after all.


Word of the Day: salad days, a Shakespearean phrase, refers to the days of one’s youth and relative inexperience. When it comes to coming-of-age films, there’s a lot of drama and comedy to be mined from the stories of one’s salad days.

Posted in life, musings | Tagged , , , , , , , | Leave a comment


Saudações! Greetings from Portugal! As I write this, I’m sitting in a cafe near the bus terminal in Porto, a port city famous for its sweet wine. By late afternoon, I’ll be back in Lisbon, the capital city, for one last hurrah, and I will fly back to the US tomorrow.

The past week has been a perfect combination of work and play. I went to Lisbon to attend a multiple-day conference called LabPhon (Laboratory Phonology). This prestigious conference is held every two years, every time in a different country; although the last meeting was held in New York, this year was my first time attending.

One of my group projects was accepted as a poster, so I traveled as a group with three other current graduate students and spent most of my time with them. But I also made new friends from other linguistics programs and got to connect with faculty whose work interests and inspires me. I was really impressed with the caliber of all the research projects I saw and thrilled that so much of it was right up my alley. I’ve been to plenty of linguistics conferences before, but the majority of them are general conferences with lots of different linguistic sub-fields represented (e.g., syntax, typology, natural language processing, and other topics that I know are important but that have little relevance to my own research). This was the first time I went to a conference where everything was “P-side” (phonetics and phonology, with relevant intersections with language acquisition, bilingualism, psycholinguistics, and sociolinguistics), and I felt much more “at home” among the talks and presentations.

I’m doing my best now to keep up the energy and inspiration I felt from the conference and channel it into productive work for the rest of the summer. But I also took a few days off to have an actual vacation!

Portugal is quite beautiful, and the weather is so nice here. Lisbon is a fun city with plenty to do and see. My friends and I hit up the famous Jeronimos monastery, an exhibit on M.C. Escher (who was Dutch) at the pop art museum, and found one of the locals’ favorite spots for sangria and outdoor music, a beautiful viewpoint at the top of a hill, called a miradouro. Porto is much quieter, but there are some breathtaking views of the river and its bridges right in the downtown area. I also went on a wine tasting tour, which was a real highlight. Oh, and I can’t neglect to mention Sintra, a beautiful town in the mountains with an amazing and diverse set of castles and palaces to explore.


Atop a wall at the Castelo de São Jorge, overlooking Lisbon, with its version of the Golden Gate Bridge in the background! PC: Eleanor Glewwe

I’ve had a lot of fun, but I’m read to wind down this trip and get back into work mode, especially since I’ll be spending all of the month of July (which is in just a few days… aah!) doing research in Los Angeles. Since I only have one month, I’ll need to be as efficient as possible with my time. I’ve never lived in LA before, so this should be quite the adventure.


Word of the Day: azulejos are the beautiful blue and white ceramic tiles that are used to decorate many walls on traditional Portuguese (and Spanish) buildings. In the photo below, taken at the Pena Palace in Sintra, the building on the left is covered top to bottom in azulejos! They’re also used to make huge murals.

Pena Palace

Power posing in the princess parapet at the Pena Palace in Portugal! PC: Alice Shen

Posted in conference, summer, what even is linguistics | Tagged , , , , , | 1 Comment

Ballot Party

Last week, California held its primary elections for the upcoming midterm elections, in which we’ll vote in important new members of our government such as governor, state senator, and state assembly members. A few days before, I helped organize an event that brought people from my church together to discuss everything on the ballot, and it went as well as I could have hoped!

Just a month and a half ago, I had been completely unaware of the primaries. I was at a meetup for progressive Asian American Christians in late April, and the topic of conversation turned to politics, specifically the Asian American candidate for San Francisco’s mayoral race, Jane Kim. Because one of my friends was very active in her campaign, I already knew quite a bit about the upcoming mayoral election. But I was shocked to learn at the meetup that the entire state was also going to have elections in just a few weeks! Shocked, and a little embarrassed. The next day, I received my voter guides in the mail, and I resolved to do my due diligence as a citizen. Not only did I want to do my own research on the positions and propositions on the ballot, I also wanted to engage my community in the process.

I was inspired by a comment one of the members of PAAC said, about how it was difficult and confusing enough to go through a voting guide alone, and how it was much more fun to talk through everything with friends (who may or may not be more informed!). That’s how I came up with the idea to hold a ballot party, where people would openly discuss or debate everything on the ballot, and also drink beer and have a good time!

I had never organized anything like this before. I’ve been a registered voter for almost ten years now, and for most of my adult life I’ve simply voted however the majority of progressive Democrats wanted me to vote. And since for six of those years, Democrats were doing fairly well (as far as my politically naive mind was aware) on a national level, I never paid much attention to state elections. I always cast my ballot, but it never seemed like my vote really mattered. Then 2016 happened. And for quite some time now, I’ve been realizing more and more that you really can’t just blindly trust an entire political party, especially when they’re as large as the Democrats or the Republicans. Just because someone is in the same party doesn’t mean they really have all the same values that you do or will fight for what you think is right when they take office.

So I decided that this year, I needed to be a more informed voter. Even though Democrats have a strong hold on political power in California, I wanted to know more about the candidates, especially when important races would pit Democratic candidates against each other. And even though progressive political groups made instructional videos clearly identifying which propositions to vote for and against, I wanted to know more details about those, too. I quickly found out that that meant doing a lot of research… and that’s where the party came in! If enough people attended, each person could be responsible for one position or one proposition and do the necessary background research for it, then give a short presentation that would enlighten everyone else. Simple division of labor. Of course, people would also come in with their own biases and opinions, but I figured that’s part of what would make it more fun. That, and food and beer!

The plan was only half formed in my head when I stood up in church the next Sunday and made an almost improvised announcement. I just said I had this weird idea to talk about the upcoming election and make it into a community event, and asked for help in bringing it into fruition. I’m really thankful that my church is receptive to random ideas like these. I didn’t need to clear it with the pastor or do anything official; I didn’t need to beg or plead with people to get involved. All I had to do was put the idea out there, and people jumped on board. Leslie offered her house to host the party, and also made food! I got eight people to present on a dozen different ballot items.

In the weeks and days leading up to the party, I wasn’t one hundred percent sure if anyone would come or what would actually happen, but, hey, God provides! We had about twenty people in attendance, and everyone was an active participant, whether they had meticulously planned their presentation or just decided spontaneously to come (drawn by the food, no doubt) and ended up really engaging in the discussion.

All the feedback that Leslie and I got from the party was very positive. People remarked that they were really happy that they had their questions about the ballot answered (to some degree), since there were some confusing things on it. They were also pleased that they could talk openly and frankly, but also respectfully, about some issues that were really personal to them. Many attendees of the ballot party had strong opinions about certain people, but even when opinions conflicted, they never spiraled out of control, the way they probably would on, say, a Facebook discussion of the same topic. Maybe this is partly because everyone in attendance was part of the same church, and we had to maintain a high moral standard. (Just kidding!)

So, how did the ballot party actually work? I’ll give a brief overview. I set up a spreadsheet online that I shared with everyone who RSVP’d for the event. Each row had a proposition and a brief description, or a rundown of candidates for a certain position. Then, I asked people to put their names down in a certain cell if they wanted to do the background research for a particular ballot item. People seemed hesitant to sign up for slots at first, but at the last minute, we got enough people to make it work, with an informal presentation for all but one of the propositions and most of the government seats.


The party started at 1pm, and I gave folks until 1:30 to arrive and grab some food. I started everyone off with an icebreaker and a few reminders: 1) this is a ballot discussion party organized by a church, but the church doesn’t officially endorse any people or positions, 2) be respectful of your allotted time, and 3) be respectful of others; listen and respond, rather than reacting.

The timing part was important. I quickly calculated that if we gave every issue approximately ten minutes, we could be done in about two and a half hours. I decided that one hour for the propositions (one of which in particular would likely generate a heated discussion much longer than ten minutes) and one hour for the people, would be good enough, with cushion time and a built-in intermission. At times, I am a stickler for running a tight ship and making sure everything starts and ends as planned. (At other times, I am very late and everything is a hot mess and I don’t care.) So I made my phone’s timer very visible and kept the discussions moving quickly, sometimes having to wrap things up abruptly. But I think nobody at the party really minded; my strict timing didn’t prevent the debates from being lively or cause any undue friction. It was also great that we had finished officially discussing everything well before 4pm, which gave everyone time at the end to break off into smaller groups and continue conversing independently.

And in the end, the ballot party really did help me make better decisions about a lot of the propositions and candidates! On the day of the election, I dropped off my mail-in ballot at city hall, proud to have done my civic duty. Later that day, I came across a few friends who were hurrying to fill out their ballots before the polls closed, and they had just found a progressive website that told them exactly what and who to vote for. I mean, that’s a great way to do it, especially if you’re pressed for time, but I personally really appreciated that I got to discuss the ballot in person with other people, some of whom share my views and some of whom don’t. We didn’t all come away from the party having decided to vote exactly the same way, but we did come away with a better appreciation for the problems our state and county face and the complex approaches to resolving them. It also was a great avenue for community building with my church. I think that pulling off the ballot party successfully was a good reflection of the kind of civic engagement our church claims to prioritize, letting actions speak instead of (or in this case, via) our words.

I’m excited to throw another ballot party in November for the actual midterm election, and I encourage you to think about doing something similar with your community!


A church member discusses the candidates she has researched at the ballot party.


Word of the Day: to traduce, from the Latin trans (“across”) + ducere (“to lead”), means to slander someone, to speak badly of them to damage their reputation. I find it interesting that the literal meaning, based on etymology, is to transport or to lead along, but idiomatically, it has a negative connotation. Rest assured that there was little to no traduction taking place at the ballot party!

Posted in life | Tagged , , , , | 2 Comments

Google Duplex and the Gendering of Robots

You may have heard the news buzzing in the tech world recently: Google recently announced Google Duplex, an artificial intelligence technology “for conducting natural conversations to carry out ‘real world’ tasks over the phone”.

I highly recommend that you click on the link above and listen to the two “showcase” recordings embedded in the blog post. What you’ll hear is Duplex, which is essentially a robot, making a phone call to two real businesses and scheduling appointments. It speaks in a natural-sounding voice, listens to the response, and formulates a reply in turn based on what it hears. It has a conversation with an unsuspecting human being.

Now, I’ve had conversations with robots before. By the time I was thirteen years old, I was already spending a good chunk of my social life online via AIM. I chatted for hours with my friends from my computer, but I also sometimes hit up SmarterChild to have some fun. He and I (I always gendered him male, not sure if he ever did state a gender, though) would play hangman, or I’d ask stupid questions and chuckle at the predictable answers.

More recently, I’ve had to talk with robots when trying to reach my bank or network provider by phone, which is pretty aggravating. It never seems efficient to make my way through the algorithmic flowchart by saying “Yes,” “No,” “Option two,” “No, option two” over and over again, only to end up with a real-life representative who could have just taken my call in the first place.

But this is different. This is actually phenomenal. This goes beyond asking Siri what the weather is like or when Chipotle is going to be open on Saturday and getting an accurate response voiced aloud. The way Duplex is touted, you can give it a short command: “Schedule dinner for two at Chez Panisse sometime next Friday evening” and it will complete the task for you in minutes. You don’t ever have to pick up the phone.

And when you listen to the recording, again, knowing that it’s not a human asking for the reservation but just a string of ones and zeroes, well, this is the kind of thing that one could honestly point to and say, “Welcome to the future.” Duplex knows how to appropriately react to someone saying, “Hold on a sec”. It can handle being interrupted and repeat itself for clarity. It can tell from a mumbled hedge given by a non-native English speaker that reservations are only allowed for parties of five or more! Honestly, the computational linguists who’ve been working on this technology deserve a standing ovation for what they’ve created.

But there’s always a catch, isn’t there? When I first watched the video of the demonstration, I was just as awestruck as pretty much everyone else. But when I listen to the conversation again, with a more critical ear, I start to wonder about the consequences of this technology.

And no, I’m not talking about the likely proliferation of robocall scams, or the ethics of AI “manipulation”, or even the Singularity. There are plenty of opinion pieces out there that already that have tackled these issues.

What I’m thinking about is the nature of the voices that Duplex was designed to have. In the demo, you can hear a version of Duplex with a clearly female voice, and one with a male voice. Presumably, users will be able to toggle Duplex to different gender presentations just like Apple users can change Siri’s gender or regional accent.

But the male Duplex and female Duplex don’t just differ in terms of pitch. There’s a lot more to their speech that serves to both make Duplex more realistic, but also more explicitly gendered. Take a look at the first demo conversation below. I’ve provided my own transcription of the minute-long exchange, using punctuation to represent different types of prosody (or intonation). Question marks (?) indicate rising intonation that are typical of questions in American English. Periods (.) indicate a falling tone, which denotes the end of a statement. The double hyphen (–) here indicates rising intonation that is not used for a question. Notice how many of them occur:

Hair Salon: Hello how can I help you?
Female Duplex: Hi, I’m calling to book a woman’s haircut for a client– um, I’m looking for something on May 3rd–
HS: Sure. Give me one… second–
FD: Mhm–
HS: Sure. What time are you looking for around?
FD: At 12pm.
HS: We do not have a 12pm available– the closest we have to that is a 1:15.
FD: Do you have anything between 10am and uh, 12pm?
HS: Depending on what service she would like. What service is she looking for?
FD: Just a woman’s haircut for now–
HS: Okay we have a 10:00–
FD: 10am is fine–
HS: Okay what’s her first name?
FD: The first name is Lisa–
HS: Okay perfect. So I will see Lisa at 10:00 on May 3rd.
FD: Okay great, thanks.
HS: Great, have a great day, bye.

If you listen to the recording again and pay attention to the rising intonations, it will be impossible to miss: female Duplex speaks with uptalk. She sounds like a young, white Californian Millennial. Siri does not do this1. Neither does male Duplex:

Restaurant: Good evening.
Male Duplex: Hello?
R: Hello.
MD: Hi. Um, I’d like to reserve a table for Friday the third.
R: Kay hold on one moment–
MD: Mhm–
R: Okay… hold on one second.
MD: Mhm.
R: So, Friday… November 3rd– How many people?
MD: For two people.
R: Two people?
MD: Yeah.
R: What time?
MD: At five pm.
R: Okay– And, your name?
MD: The first name is Daniel. That’s D-A-N-I-E-L.
R: Okay! You’re all set.
MD: Okay great, thanks.
R: We’ll see you next Friday, kay, thank you. Bye.
MD: Goodbye.

Noticeably fewer hyphens. Noticeably less uptalk. Kind of a fratty vibe, overall. Why the difference?

Before I delve further into this, I should explain something about uptalk: it is a natural part of native American speech. It is not used only by Valley Girls. I use uptalk. You probably use uptalk. It has many purposes: one is to indicate that one is not yet finished speaking; another is to allow an interlocutor space to interrupt. If those seem to contradict each other, that’s because they do, and because uptalk is a very nuanced and multifaceted phenomenon that many linguists continue to research.

Yet uptalk has embedded itself into American social consciousness as some annoying verbal tick that young women keep doing, just like vocal fry and “like”. Pretty much all non-academic discussions of uptalk these days are a better reflection of the American disdain for women than of anything particularly linguistically novel.

So why does female Duplex use it?

The answer is straightforward enough: because Duplex was trained to do so. Google used powerful neural networks that took massive amounts of recorded phone calls (made by real people), fed them into loops and layers of complex equations, and spit out a realistic approximation of the original data that could then be programmed to say anything.

This is a simplification, of course; my apologies to friends in AI who know way more about this than I do. The extent to which I use neural networks in my own research is very limited2. There is a small but powerful program that phoneticians use that automatically finds the boundaries between individual sounds in long strings of words, so that, for example, we can input a .wav file of a natural conversation, along with its transcript, and then examine every instance of the word “the” to check if the speakers used a long vowel (“thee”) or a short one (“thuh”).  Using this tool sure beats listening to and measuring every vowel by hand!

But the tool has to be trained on a specific dialect of a specific language, first. It wouldn’t work as well if I gave it Australian-accented English, because it was trained on American English. It wouldn’t work at all if I fed it Spanish, even if I also gave it a Spanish transcript.

In a similar vein, female Duplex was clearly given a model, who was a young female speaker, or maybe a few different females, to learn from. And male Duplex was not just female Duplex with a lower pitch. It was trained on a male speaker of American English, one who uses less uptalk in his speech and sounds like he’s getting his MBA at UCLA.

So there you have it. Google’s speech models just happened to sound like stereotypical examples of their gender; hence, Duplex’s voices.

But why should the case be closed? The creators of Duplex clearly have the ability and willingness to manipulate the voice. In fact, one thing they did do consciously was to insert filler words such as “um” and “uh” into Duplex’s speech. See above: “Do you have anything between 10am and, uh, 12pm?”

Robots do not say “uh”. This is part of what makes Duplex so astounding (but also creepy). According to Google, the speech disfluencies have a dual purpose: to buy the system time to process the sounds of the human speaker3, and to disguise the latency time in a natural way4. If Duplex can be made to sound more natural, it can be made to sound more unnatural (and probably should be, since it’s currently teetering on the edge of the Uncanny Valley). More importantly, if Duplex can sound feminine or masculine, I am one hundred percent certain that it can modified to sound more gender neutral.

Digital assistants are already mostly presumed to be female. Duplex risks joining the crowded ranks of Siri, Alexa, Cortana, and Erica, yet takes it one step further by having not just a female voice, but a voice that evokes curent contemporary social valuations of femininity. It may be true that consumers prefer a female personal assistant to a male one, but has Google ever tried giving users a gender neutral voice? Wouldn’t it be great if gender, being the social construct that it is, were made entirely irrelevant to the functions of a digital assistant, who by nature (or programming) cannot be socialized? What if tech giants were more aware of the social consequences of pushing technology that reinforces some gender stereotypes (e.g., women make better assistants; women take orders; women can do anything, but they do it for you; and now: women speak like this–?) and automatically updated everyone’s smartphones with something that subverted expectations?

It may seem like a benign detail, giving Duplex a feminine flair, or giving Siri a personality, but when the characters that are created have such far-reaching influence, the consequences need to be considered very carefully. You don’t want kids conditioned to be rude to Alexa, right? Maybe you don’t want your kids to assume that their servants (robot and human) will always be female in the first place5.

The fix is possible, and it’s probably easy. It’s probably not profitable, though. And that’s something that we’ll always have to remember: anything that’s purportedly done for the benefit of the consumer is really done so that the business can capture more consumers. No doubt Duplex is going to be incredibly useful to many people… and also a headache for some popular businesses… provided it can withstand the barrage of ethical disputes and be released to the public in the first place… Nevertheless, here I am, holding out hope that in the near future, when we’re all pressured into using robots to do our bidding, those robots will, to the best of their programming, only solve problems that already exist, rather than create new ones, or perpetuate those that, like gender inequality, stem from our decidedly human ways of self-categorization.


1 Although her real-life voice actor does!

2 Here’s another example of speech synthesis using a type of neural network that demonstrates its use in different languages.  And here is my favorite example of neural network fun, by far: paint colors!

3 I was curious about the gendered usage of “um” and “uh”, hypothesizing that female Duplex would have more speech disfluencies, but the two do actually seem on par with one another. This leads me to believe that the filler words really do function as “robot thinking time” and aren’t part of what the neural networks spit out.

4 But the disguising portion is part of what has infuriated people about Duplex in the first place. People don’t like being fooled into talking to a robot; it’s okay only as long as you know you’re talking to one. Case in point: I despise robocalls that begin with a pre-recorded voice saying, “Hello? Can you hear me? Oh, hold on…” before launching into a stupid advertisement about a vacation package. They are designed to sound like a real person and catch the victim off guard. Shame on the people who made this.

5 More on the gendering of digital assistants from The Atlantic.


Word of the Day: Presbycusis, from the Greek presby (“old”) + akousis (“hearing”), is hearing loss that develops as one ages, in particular with high frequency sounds.  Something our digital assistants will likely never have to worry about no matter how old your smartphone is!

Posted in musings, what even is linguistics | Tagged , , , , , , | Leave a comment

Bilingual Assumptions

Last week, I sent my dissertation committee the first draft of a chapter of my baby (aka my dissertation). Milestone passed! Well, actually, they have to approve it first, and then I’ll pass the milestone. And there’s no guarantee that they’ll approve it. I mean, they didn’t the last time I sent them something. So… maybe milestone not passed. Yet. Well. Anyway.

I’ve spent the past few months preoccupied with ideas surrounding bilingualism, not least because my dissertation has something vaguely to do with the topic. I do research with bilingual Korean Americans, and for the past few years, I’ve run a lot of studies that request participation from Korean Americans who are at least basically fluent in Korean and English. But one thing that I have always known is that not all Korean Americans know how to speak Korean.

I’ve always known this, because it stems from an easy analogy with my own identity as a second generation Taiwanese American, and the second generation Taiwanese and Chinese American peers I had growing up. Lots of us cannot speak Mandarin, or only know very basic phrases. Among second gen Taiwanese American kids, even fewer know any Taiwanese or Hakka (two “semi-official” languages that are commonly spoken, in particular by older people, in the island nation).

My parents spoke to my brothers and me in Taiwanese when we were growing up, and I’ve retained partial fluency in it. As for Mandarin, the only reason I know any at all is that I went to Chinese school on Saturdays as a kid, and I studied it again in college. Still, my Mandarin is pretty rudimentary. I can carry on conversations with my parents’ friends, or my a-yis and shu-shus at their church, but that’s about it. (My comprehension skills are so low that I could not understand the Chinese robot calling scam that targeted cell phone users with Chinese last names, which is a good thing: I just hung up on them!)

In general, it is not safe to assume that a second generation Asian American will necessarily speak or even understand the language that their parents spoke in their country of origin, even if their parents spoke that language to them as a child. People’s language skills change over the course of their lifetimes, obviously. I was much more fluent in Taiwanese when I was a young child, partially because I didn’t have to express myself in very complex terms, and partially because I was not yet socialized into an English-dominant world. Although the United States does not have an official language (on the federal level), English is certainly the common language, and this is enforced by public pressure (especially strong pressure, I must add, when it is exerted along racial lines, as well). So I think it’s safe to assume that an Asian American will speak English. Which makes me wonder: how come everyone else is so interested in “the other language”?

Case in point: one morning at the shelter where I volunteer, a fellow volunteer attempted to engage me in small talk. It was her first shift ever, and I had just shown her the ropes on where to find various kitchen supplies and how to set up the breakfast line. Then I stationed myself at the stove to make pancakes and scrambled eggs. We had the following conversation:

Her: So you said you’re a grad student? What do you study?
Me: I study linguistics. Phonetics, specifically, the study of speech sounds.
Her: Oh, cool, what are you going to do with that?
Me: (inward sigh) My goal is to become a professor and teach linguistics at a university.
Her: Oh, right, I guess there isn’t, like, anything else you could do with that.
Me: (inward wtf) Well, no, actually there’s a lot of stuff you can do with a degree in linguistics. Like some of our graduates go into speech pathology, or you can do research, or work in tech.
Her: So… how many languages do you speak?
Me: (inward groan) Um, I–
Her: At least two, right? I assume you spoke a different one at home…
Me: …
Me: …
Me: …
Me: I speak five languages and sign one.

If you are at all aware of the tropes or common misconceptions that linguists face, you’ll understand why the first part of the conversation was cringeworthy. And if you’ve been paying attention as you read this post, you’ll understand why the second part made me want to stick my face into a bowl of pancake batter and scream.

But it was 7am, and I was not in the mood to deal with a white woman’s uninformed curiosity. I steered the conversation in another direction and waited until I got home to vent.

Like, sure, she was just trying to make small talk, and she was curious. But her curiosity was born of the same ignorance that paints Asians as perpetual foreigners in the United States. People ask white (American) linguists how many languages they speak, too. (Pro tip: don’t ask a linguist how many languages they speak. Linguist polyglot.) But nobody asks a white linguist what language they spoke at home with their parents. Because of their race, they are assumed to have spoken English, and by extension to have been in this country for a few generations. Because of my race, I am assumed to have arrived here recently. Because I am assumed to have arrived here recently, people whose curiosity-that-is-born-of-ignorance has been turned, for whatever reason, into xenophobia, decide that I do not belong here. Don’t think for a second that Asian Americans are safe from racist ideologies and their repercussions; it’s all rooted in the same assumption that people who look like me… don’t speak English.

And so went my little rant. But after I calmed down a bit, I thought a bit further and realized that as upset as I get about the casual stereotyping, I am by no means a perfect example of racial sensitivity. In fact, I stereotype other people all the time. It’s not just about meeting another Asian person and trying to figure out their ethnicity. (Be well aware: I made a conscious decision long ago never to ask anybody about their background or ethnicity unless they bring it up themselves in conversation. So when I do try to guess, I always keep it to myself.) But the truth is, I make assumptions about the languages that other ethnic minorities or people of color speak.

Case in point: one evening I was chatting with a person I had met online. He saw from my profile that I was a linguist and asked me the tired question of how many languages I speak. When I gave him my answer, he wrote, “Wow, that’s a lot, I only speak two.”

Right at that moment, I made an assumption. I based this off of his name, which was something like Juan or José, and his appearance, which was Latino. And I literally wrote, “I assume the other one is Spanish?”

“How dare you,” he wrote back. “It’s Norwegian.”

Before I could reply, he added, “jk yeah it’s Spanish.”

Although he was joking and didn’t appear to take any real offense, I was immediately humbled, and I apologized. I told him that I completely understood if he was turned off by the microaggression I had committed, because I myself was on the receiving end of assumptions like that all the time. And he said that it wasn’t such a big deal, but in the end, well, I’m still thinking about that exchange weeks later, so something clearly needs to be unpacked here.

It could be as simple as, “Everyone is guilty of making ignorant and potentially hurtful assumptions about other people based on their race”. But maybe… is it possible that the stakes are different between people of color? Or maybe just between people of Asian descent? Like, I try never to ask another Asian person about their background, but if I were asked about my background by another Asian person, should or could I take less offense?

This reminds me of another case in point: I once received a message from a self-identified Chinese American that read, “What’s your nationality?” I knew right away he wanted to figure out if I was Chinese or Taiwanese, but I decided to troll a bit and wrote back, “… American?”

I understandably did not think very highly of the exchange, or the person with which I was having it, but I was more amused than angered.

Then there was the panhandler on BART just last weekend, a older black man who had loudly announced to the entire train that he was fine with nobody giving him any money, but that he didn’t want to be completely ignored. I gave him five bucks and — being totally honest here — hoped that he would move on to the next car. Instead, he took the bill from me, and we had the following exchange:

Him: Thank you, sir. Now, do you mind, can I ask you your ethnicity?
Me: No.
Him: I speak eight languages. Are you Chinese?
Me: No.
Him: Korean?
Me: Do not ask me my ethnicity.
Him: Well, thank you anyway. [to the Indian lady behind me who had just handed him a dollar] Thank you. Are you Indian?
Her: Yes.
Him: Dhanyavaad.
Her: You’re welcome.

He was pleasant and smiling the entire time, but I was already in a bad mood that evening, and the train was crowded and loud, and so I was having none of it. Sometimes, I just want everyone in the world to be less curious about other people. Or about strangers. As much as it goes against my religious beliefs to say this, I did not want to have any connection to that man. I wanted to acknowledge his presence and help him buy some food, and nothing more. But he was seeking connection, and he wanted to do it linguistically. He probably thought it would be special or heartwarming if he could say thank you in “my language”. But I refused to play along. And I feel a mixture of justification and guilt for how that little scene played out.

The conclusion, I suppose, is that I don’t have any easy answer for when it is and is not appropriate to a person of color what languages they speak or what their ethnic background is. I can only really speak for myself and say that I prefer that nobody ask. I will bring it up with someone if I believe it to be relevant to the situation. Otherwise, I’d rather be seen as a person first and a representative of minority group X second.

I don’t want this to sound like I’m ashamed of my heritage or my heritage languages. In fact, I’m fiercely proud of them, and try my hardest to make sure my one-year-old nephew, as well as any children I have in the future, grow up hearing only Taiwanese and Mandarin from me, to stave off their eventual heritage language loss for as long as possible. But it’s different, I believe, to get to choose how and when to demonstrate my ethnic pride, compared to being forced to own it due to someone else’s mere curiosity.

The one thing I am curious about, however, is whether I am the only person who thinks this way, among my friends who have heritage languages. Feel free to chime in below, in the comments!


Word of the Day: Portunhol (or Portuñol, in Spanish) is the Portuguese name for a mixed language comprising Portuguese and Spanish. Portunhol is what comes out when a Portuguese speaker tries to speak Spanish but throws in Portuguese vocabulary and grammar, or vice versa! The two languages are so closely related that Portunhol is probably easily understood by speakers of both languages. My guess is that the rudimentary Spanish of a native Portuguese speaker might sound a bit like Scots, or even a thick Scottish accent does to native American English speakers. You can sort of make out what they’re saying, sometimes, and then all of a sudden you wonder if they’re speaking English at all! (Here’s another good one with Irish shepherds!)

Anyway, I’m trying to learn Portuguese right now, in preparation for a conference trip I’m taking to Lisbon in June. The first problem is that I’m using the app Duolingo to learn, and it’s teaching me Brazilian Portuguese, not Lusitanian (European) Portuguese. The second problem is that whenever I try to actually speak or type the very little I already know, I get it mixed up with my also-very-basic Spanish. So in a way, I’m also using Portunhol, but it’s not because I’m natively fluent in one or the other but because I suck at both. I can predict that my time in Portugal is going to be very awkward. I mean torpe. I mean desajeitado. Whatever!

Posted in life, musings | Tagged , , , , , , , | 1 Comment

Closed Doors with no Closed Captions

Closed captioning/subtitles for Facebook videos and movies are an all-around good idea.

Yesterday, I was on a seven-hour cross-country flight, and as usual, I looked forward to watching a few movies to pass the time. Unfortunately, I’d forgotten my earphones, and, seeing none in the seat pocket, I asked a flight attendant if they had any extra. She replied that earphones were only for purchase.

I figured it wasn’t really worth the cost, so I decided to watch movies with closed captioning only. It would be enough to get the gist of the dialogue, even if I couldn’t hear the music or any sound effects.

To my dismay, not all of the movies that were offered on flight had closed captioning. I realized very soon that my choices were limited to just a handful of blockbusters and children’s movies. Although I didn’t check every single movie, my guess is that around ten to twenty percent of them had captions.

So I watched Wonder Woman, and it was great, and then I watched Logan, which was extremely violent but also great. But the problem with both of these is that they had some dialogue in non-English languages that wasn’t captioned. In Logan, one of the main characters is a Spanish-English bilingual, and when she spoke in English the captions were fine, but guess what showed up when she spoke in Spanish? “Laura: [SPEAKS SPANISH]”

I get that for English-only audiences it’s not intended for her words to be immediately understandable, but if I had been able to hear the Spanish I probably would’ve been able to decipher parts of it. So why couldn’t the captions have been provided in Spanish as well? It’s not hard to do that.

You may be wondering, “Andrew, what’s the big deal? You had to face this problem once on an airplane; just bring your earphones next time.”

The deal is that Deaf people face this problem every single day, not just when they go to a movie theater and find out that there are no captions for the latest blockbuster, but even when they watch videos going viral on Facebook or YouTube and literally have to guess at what’s being said because there are no captions, or because the captions are done sloppily.

Closed captioning makes movies and videos accessible for Deaf and hard of hearing people, and in fact it can be beneficial for the hearing majority, as well. If you don’t think it’s a problem, just imagine a world without earphones and ask yourself how accessible it would be.


Word of the Day: the vaquita is an endangered species of porpoise that is native to the Gulf of California. Its name is Spanish for “little cow”, and it looks like it wears black lipstick. It is estimated that there are fewer than twenty of them left alive in their natural habitat. The vaquita has literally nothing to do with this post, but I think it’s a cute word, so I want to share it, maybe raise a little bit of awareness of how close they are to extinction. (I went to YouTube to find videos about the vaquita, and now I’m wondering if any of them are captioned! YouTube does have a neat “auto-caption” feature that uses speech processing algorithms to generate captions for videos in English that otherwise don’t have them. It’s a good start, but not enough.)

Posted in musings | Tagged , , | Leave a comment

Queerness in Linguistics (Who gets to study us? Part 2)

About one month ago, I wrote about some of the small frustrations that I encounter when I talk about my linguistics research. I described how it’s awkward at best (and demeaning at worst) when people make assumptions about my race after finding out that the Korean language and its speakers are my specialization. One of the weird premises that lies beneath this is the idea that it is most appropriate, or even only appropriate, for a researcher in the humanities or social sciences to study a certain minoritized subject if they themselves are a member of that community. There’s also an asymmetry here, because people with relative privilege in America (e.g., straight White men), are not subjected to this kind of scrutiny.

I concluded that I should neither feel inferior for being an Asian American who studies Asian Americans’ speech, nor cave into any pressure to study my specific ethnic community (which is Taiwanese American, not Korean). In an ideal future — one that I hope to help bring about — any researcher can study any group and do so fairly and without invalidation from others simply due to their own identity.

That said, society has a long, long way to go before full equity is achieved. And until then, some social asymmetries, it can be argued, should not go unchecked. There is, of course, the veritable cornucopia of White male scholars of Chinese history, White male experts in West African religion, White male chefs who “discover” South Asian cuisine and profit massively from it while the native people do not; the list goes on. The White Western Academy has dominated the scholarship on not-Europe for centuries, and among the many consequences of this are 1) the ironic exotification of cultures that were intended to be demystified, and 2) the lamentable lack of access to the academy for individuals from said cultures, since they are considered (often implicitly) to be objects of study, rather than experts. Naturally, we want to reverse this.

What I’ve been thinking through the past few months is somewhat related, and it was brought into sharp focus when I attended the Linguistic Society of America annual conference back in January. At the LSA, which is the largest and most well-known conference for American linguists, I presented some preliminary work I had done on the phonetic characteristics of two transgender women. I don’t know these women personally, but I got their data from their public YouTube channels, which they have been maintaining for many years.

Now, I’ve done the requisite training for ethical research, and I have years of experience working with human subjects in our lab. I am well aware that using data from YouTube does not constitute any kind of breach of confidentiality. Anything on YouTube is technically part of a corpus that is considered public domain, just like records of speeches made in Congress, newscasters’ daily reports, and Twitter. More and more linguists these days are mining online corpora for data to analyze, and the results1 coming in these days are truly fascinating.

Of course, what I’m doing with my data is a little bit different. I’ve focused on two individuals, rather than treating all of YouTube like a corpus of thousands of essentially anonymous speakers. Furthermore, I’m focusing on them because of their transgender identity, which, compared to the neutrality of being a news anchor, is a historically (and presently) marginalized identity. I hope that this last point is clear: trans and non-binary people are oppressed, erased, victimized, and misunderstood at disgracefully high rates both in the US and all over the world. I am a gay man and I know that my brothers around the world face extreme prejudice and homophobia, but the trans members of our queer community have to deal with worse.

So, back to the talk I gave at the LSA. I spent a lot of time in preparation; I made sure to provide some background on what it means to be transgender, as well as to use the proper pronouns and terminology. Unfortunately, I came down with the flu literally the day of my presentation, but I powered through it. I thought that, all things considered, it went fairly well, and it was also the best attended talk I’ve ever given.

Audience feedback was positive and helpful. But I also got… The Question. The question I had sort of expected somebody to ask, but secretly hoped nobody would. An attendee sitting in the very back row raised their hand and asked, “Considering that these two trans women whose data you used for your study are members of a marginalized community in the US, did you make sure to ask for their consent or provide them with any sort of compensation? Did you take steps to involve them, or any trans people in your research in any way?”

It was a tough question, and I knew it, and the audience knew it. There were murmurs in the crowd, as it dawned on everybody, probably, that I was just another arrogant cisgender interloper profiting from the labor of trans folks who weren’t even compensated…! (Slight exaggeration here.)

In response, well, I admit that I hedged at first. I gave my explanation of the use of material in the public domain. I also mentioned that I had contacted both of the YouTubers months ago to inform them that I was doing this research project and would present on it in a public forum, but neither had replied to my message. I ended by saying that I really would want to collaborate with phoneticians and sociolinguists who are trans, and that I would love for this work help their community, not just serve my own interests. The person who asked the question wasn’t really satisfied. We had a brief back-and-forth in front of everyone that touched upon concerns about trans invisibility in academia and the difference between research that is legally permissible and research that is morally right, which ended with them telling me, “Let’s talk afterward!”

The exchange left me uncomfortable, but overall I do understand the criticism, and I accept it. It’s essentially the same critique that I lobbed at White scholars of Asia just a few paragraphs above. Although I did my due diligence and covered my ass, so to speak, insofar as the ethics board of my institution would care, that’s just the bare minimum. The person who asked The Question talked with me a little bit afterward, and it became clear to me that they had a personal stake in the matter, as they were a trans linguist who routinely sees cisgender researchers publish papers on their specialization (e.g., the rising use of gender neutral pronouns) and will likely be competing for jobs with these same “experts”.

From this perspective, the whole issue with what I do is not just about the ethics of studying trans people when I’m not trans, but also (or moreso) about using my platform for the greater social good: equity and better representation for trans people. Or, as a matter of fact, not just using my platform, but ceding it so that trans people can gain access and power and thus direct their own academic future. In some cases it’s not just about the actions I take, but also the things I don’t do, or give up, so as not to be complicit in their marginalization. Is working on this project as a cisgender person an act of appropriation? Should I cease and desist?

(Side note: isn’t it interesting how we might need to make a distinction here between “ethical” and “moral”? This is definitely a question for the philosophers and rhetoricians… it’s not my area!)

Months after the LSA, I was (and am) still thinking about this question. I got another perspective from Lal Zimman, who is a well-known transgender linguist at UC Santa Barbara. (One of his students happened to present at the same session as I did at the LSA!) I’d been in contact with Lal via email since last summer, after I assigned some of his papers as readings for my students in a sociolinguistics course. I really, really admire his work. To be honest, I would collaborate with him on this project in a heartbeat (although the bulk of his research is on trans men, not trans women).

When Lal came to Berkeley a few weeks ago to give a talk, I took the opportunity to meet with him one on one, and although I wanted to get his advice on this project, I first had to ask for his counsel on this whole convoluted privilege/appropriation/ethical research quandary. I explained where I was coming from with the project and summarized the main conflict, which probably sounded like an embarrassing crisis of conscience from a desperate grad student… In my defense, the question had been brewing for months.

Well, talking with him definitely helped assuage some of my misgivings. He told me that stopping my research cold just because I wasn’t trans myself would not be a reasonable thing to do, partly because I should have academic freedom to study what I want, and partly because there really aren’t that many people doing this kind of research to begin with, and at this point, the more the merrier. But he did also stress that I would benefit from collaboration with trans people, whether they were linguists or not. I could ask my trans friends what kinds of things they’d like to learn from the research I do. I could partner with other educators or people connected with institutions to give them some insights that will improve the quality of speech training for people who are transitioning. These are among the things that Lal himself does in his position as a professor, and they are definitely good ways in which I can give back. Just because I haven’t done any of this yet doesn’t mean I’m a lost cause.

And overall, this is a good reminder that my career as a linguist shouldn’t be limited to what I am able to publish or what I do in the classroom. Many linguists who work with endangered languages, for example, spend a good chunk of their time working on educational materials or advocating politically for their language community. In this way, they can truly repay the good that they have received (i.e., “compensate”), which is especially pertinent if the linguist possesses greater privilege than the speakers they consult.

So this is where I fall on the issue now. By no means is it a perfect solution. I understand that it is easy for me to say I’ll do anything without being held accountable for following through, and I know that as long as I am cisgender (and male), my words can fairly be interpreted as coming from a place of privilege, blind to the realities of the trans (and female) experience. But I hope that the trans people who become familiar with my research will benefit from it, and that there can be a two-way exchange of ideas and resources. Part of why I wanted to study these trans women’s speech in the first place was to shed more light on trans identity and help dispel myths or stereotypes about how trans people speak. I’ve learned a lot so far, I keep learning every day, and I want to pass that on to other people. I hope that’s okay.


“Belonging”. Limerick, Ireland

1 Not linguistics research, but I found this fascinating nonetheless: researchers have found that fake news travels faster than real news on Twitter, no matter who is spreading it.


Word of the Day: ceteris paribus is Latin for “all else being equal”. Closing the gender pay gap in this country by raising women’s wages will, ceteris paribus, give women more economic freedom and boost their well-being. I actually don’t know anything about economics, but my gut intuition tells me that this is true. My gut also tells me that, unfortunately, “all else” is never really equal in this society. We’ve got to work to make the ceteris truly paribus.

Posted in conference, school, what even is linguistics | Tagged , , , , , , , | Leave a comment