My predictive text and I have something very important to say:

As a society, we are in a good place to make a good point. The only problem is that the game doesn’t play with the other games and I can’t stop playing it. 

Why does my predictive text sound like a dude bro in his first year of a philosophy major? I truly hope that outside of my predictive text, I don’t sound like someone who is stereotypically high waxing poetic, but never actually saying anything of substance (that would be mortifying).  Is this exemplary of how I write? Is predictive text a reflection of who we are? In some ways, the answer is both yes and no. Predictive text uses machine learning in order to assess which words we tend to use more often and creates a personalized dictionary that scores those words based on the probability that we’ll use them again (NBC News, 2017). Additionally, the predictive text algorithm also uses something called “probabilistic language modelling” which considers the context of what is being written and how certain words tend to go together (NBC News, 2017). Ultimately, the algorithm of predictive text is a combination of machine learning (about how I speak) and language probability (how everyone speaks). This leads me to wonder, how much of the predictive text is me and how much of the predictive text is language modelling?

Interestingly, some predictive text, such as Smart Compose in Gmail is based off of a finite corpus of emails from Enron, (though I’m unclear if all predictive text uses this base of emails for developing its algorithm) (Mars, 2020). As Amanda Levendowski points out, It doesn’t take a lot to imagine how basing predictive text and other AI off of a specific set of emails might bias the technology (Mars, 2020). Is it possible that the probabilistic language modelling of my phone’s predictive text algorithm is based off of the emails of Enron dude bros embroiled in an early aughts corporate scandal?

In some ways using predictive text feels like using a Ouija board with your friends when you’re eleven years old . You may think it’s random or believe that it’s ghosts, but in reality you and your friends exert force onto the planchette of the Ouija to make it spell something coherent (Romano, 2018).  Similarly, predictive text is not random, nor is predictive text a ghost writer, we guide the predictive text by the choices we make and it coughs out a sentence. 

I tried the predictive text a few times with a few different prompts, partly because it was so fun, but also to see if I could get a result that I thought sounded like me. Interestingly, I found the predictive text leaned more toward a positive affect, for example the words ‘good’ and ‘great’ came up quite often (as well as the word ‘birthday’). Every time I put in the prompt “as a society, we are…” I got sentences like “we are going to be able to make it”  or “we are in a good mood” or “we are in a good place.” Interestingly, I don’t think I would ever finish that prompt that way. I’m quite critical of society actually, but the predictive text never gave me the option!

I also found that the predictive text never really led to saying anything quite meaningful, consider this sentence generated by predictive text: I think we should have some more of the things we need to make sure we are doing the next year. It reminds me of a segment of John Oliver’s Last Week Tonight in which John Oliver compares Trump’s actual presidential speech patterns with that of predictive technology (Last Week Tonight, 2017,  3:10).

As an extension to this activity, I also tried using predictive text for something that I wanted to say (rather than using a provided prompt and seeing where the predictive text meanders). In this example I’ve highlighted the words that the predictive text offered that were indeed predictive of what I was intending to write (I did not include the words that the predictive text offered after I had typed a few letters of the word):

You should check out the episode “you’ve got mail” from 99% invisible which is about how basically all AI technology, like predictive text, is based on the released emails from Enron. 

Certainly there is a difference between using predictive text to generate content versus using predictive text to support content. In a study about predictive text the authors found that when using it to caption photographs it led people to write shorter, more predictable captions (Arnold et al., 2020). When I think about predictive text in various contexts, I can see its value in situations that require more formulaic writing or in contexts where brevity is valued. Writing work emails would be an example of this type of context; there is a specific language and way that people conduct themselves in work emails. Anyone that has had to write or receive an email for work would be well acquainted with the phrases “please see the attached document,” or “at your earliest convenience,” and the dreaded “as per my last email.” However, in the end, I would make the case that predictive text works best when we use it to support our ideas, not generate them. 

References

Arnold, K., Chauncey, K., & Gajos, K. (2020). Predictive text encourages predictable writing. Paper presented at the 128-138. https://doi.org/10.1145/3377325.3377523

Last Week Tonight [LastWeekTonight]. (2017, November 13). The Trump presidency: Last Week Tonight with John Oliver (HBO) [Video]. YouTube. https://www.youtube.com/watch?v=1ZAPwfrtAFY

NBC News. (2017, November 8). Predicitve texting: How your phones keyboard figures out what you might type next [Video]. YouTube. https://www.youtube.com/watch?v=OfzMkERVFu8

Roman, M. (2020, November 9). You’ve got Enron mail (episode 421). [Audio podcast episode]. In 99% Invisible. Radiotopia. https://99percentinvisible.org/episode/youve-got-enron-mail/

Romano, A. (2018, September 6). How Ouija boards work. (Hint: It’s not ghosts.). Vox. https://www.vox.com/2016/10/29/13301590/how-ouija-boards-work-debunked-ideomotor-effect

Golden Record Curation Network generated by Palladio for ETEC 540

To be honest, I don’t have a strong understanding of algorithms and the terminology that I use in this reflection may make that quite obvious, but please bear with me. I do understand algorithms to be a set of instructions that help solve problems, but I wonder what problem is being solved with the Palladio program. When I look at the whole class graph/network it all seems pretty straight forward, the program has shown the connectivity of the class and their respective song choices for task 8. I could draw something like this by hand fairly easily by drawing a node for each song and each person, and then connecting them with a line. There wouldn’t be a need, necessarily, for any math in order for me to do this. Similarly, if I play around with the data and force groups of people, say for example forcing a grouping between Meipsy, Nathan, and Allison, I can see their connectivity and understand exactly how Palladio generated that graph.

Forced network grouping of Meipsy, Nathan, and Allsion

However, I wonder what set of instructions Palladio is using to group the members of each class. I would suspect it’s optimizing the in-group connectivity, but I’m not sure. Why does it put students into groups of 3 or 4? Why not two groups of 8 and one group of 7? Or why not groups of 4 or 5? Is this a default setting, is it part of the optimization, or was this predetermined by our instructor? 

The groupings generated by Palladio grouped James, Nathan, and myself together. As you can see in the image below, we have a high degree of connectivity with each other (with James and I being slightly more connected in the grouping than Nathan).

Palladio generated network grouping of Nathan, James, and Deirdre

Nathan has a good write up in his reflection about why he suspects we had such similar song choices. I don’t necessarily want to duplicate what he’s written, but I do agree that part of the reason is because the three of us created similar constraints for our song choices. Other similarities we share (those that I could glean from the info hidden in our blog posts) include the fact that we’re all high school teachers in math and science, and I’m assuming, generationally we’re all millennials. Though there are other people in our class who fall into those demographics (e.g. Ying)  and their song choices were not similar enough to be grouped with us. I suspect the constraints we determined for our song choices is a better reflection of grouping than anything about us demographically. But then again, maybe our demographics affected the constraints we decided on. Interestingly, Nathan and I actually know of each other having worked in the same remote region of Canada. Is there a shared personality trait that would lead us to both work in the Arctic and teach math and science and come up with the same choices of songs? I wonder how many other small world connections there might be in the other groupings. 

Had Nathan, James, or myself just chosen what we personally liked, would we have still ended up in the same group? 

Curiously, Palladio excluded Katrina from our group despite the fact that she shared 7 song choices with myself, 7 songs with James, and 6 with Nathan. I’ve included images of Katrina grouped according to Palladio and then grouped with my grouping.

Katrina’s original network grouping generated by Palladio

Katrina grouped with Nathan, James, and Deirdre

Interestingly, when visiting Katrina’s blogpost, she had similar constraints for herself when it came to song choices. Why then, did Palladio’s algorithm parse the data the way it did?

Another area of exploration is looking at who I shared the least amount of song choices with. In this case, Ben and myself only shared 2 song choices. I can’t speculate on why, given that I don’t have access to Ben’s reflections, but if my hypothesis is right, I would suspect that with respect to song choices, Ben had constraints that differed from mine.

Network depicting the connections between myself, Ben, and our song choices

One last thing I am curious about is if there are any two people in our class whose song choices excluded each other. I wasn’t able to determine an easy way to manipulate the information in Palladio in order to see that without manually examining each pairing of students.

Essentially, the issue at hand is whether or not Palladio’s groupings can accurately reflect anything more than just the network of connections of song choices, and the strength of the choices. Perhaps it would be helpful to have some insights into the biases or instructions used by the Palladio algorithm. Otherwise, I’m not convinced that a network like this can tell us anything about “why.” Why did our class overwhelmingly choose Johnny B. Goode and so few of us chose Men’s House Song?

This is the first time I’ve been exposed to Nasa’s Golden Record and I was struck with the thoughts that simultaneously it was so wholesome and so arrogant. The Voyager Golden Record seems wholesome in its quest to share the music of Earth and to find common ground among intelligent life, for example consider its inscription “to the makers of music – all worlds,  all times” (Taylor, 2019). However, it is also arrogant in its quest to represent humanity in a single LP, as well as arrogant to think any life in space would even care. It is true that in the podcast interview with Dallas Taylor (2019) that Timothy Ferris is at least a little bit self aware when he mentions that most great music is going to be excluded. I was curious who made up the committee chaired by Carl Sagan in determining the playlist of the Golden Record and by all accounts they are a team of white people (Ferris, 2017). It seems fairly arrogant, narrow minded, and colonialist that a team only made up of white people would be able to accurately and justly curate a playlist that represents humanity. I wonder about the cultural permissions of some of the songs that were included. Just because the Golden Record Committee secured song rights doesn’t necessarily mean they went through the process of ethically getting the correct permissions from the knowledge keepers of the songs. Ironically, in this task, I am just another white person curating the list down even smaller, I am no more qualified to say what should go on an intergalactic best of album. With that in mind, I did try and stay true the original goals of the record which is to be inclusive as possible and to make a good record (Taylor, 2019). 

With respect to inclusivity, I tried to pick songs that 

• represented a variety of sounds and instruments
• represented disparate geographic areas
• balanced out a range of male and female vocals 

With respect to making a good record, well that’s subjective. When I was choosing between songs in similar geographic regions, I picked the song I liked best. Additionally, I tended to favour songs with vocals rather than instrumental alone. I justify this choice since the album is human centric, why not reflect the actual sounds humans make. 

Here is my list of ten songs from the Golden Record with brief rationale:

1. Alima Song – This song from the Democratic Republic of Congo is beautiful to listen to and is the song that I chose from Africa. I favoured it because it also met the other criteria that I set for this task which was having a track list with more vocals and a track list that adds parity when it came to representation of gender. Although, this is a song that makes me question whether the culture who owns this song was consulted about its use. 
2. Jonny B. Goode by Chuck Berry was the first song I knew I was going to pick because it has the most personal relevance to me. My partner and I have been having an ongoing debate about collective cultural decay and whether or not we think the Beatles will be relevant in 50-100 years from now. From our research, including an email to beloved Canadian music writer and broadcaster Alan Cross, it seems that if any rock and roller is to be remembered it will be Chuck Berry, and specifically the song Johnny B. Goode. According to another music writer, Chuck Klosterman, the only memory of rock and roll at all will be of Chuck Berry. Seems like if we want to send any message to the cosmos about rock and roll, Chuck Berry is a good messenger.
3. Sokaku-Reibu (Depicting the cranes in their nest) – I like the symbolism of this song with respect to the Voyager.  According to this piece the music uses a technique called Koro-koro which is meant to imitate the flapping of wings of cranes and the song itself represents the raising of young cranes who eventually leave the nest. It seems as though this is meant as a metaphor for the two Voyager spacecrafts, from the birthing of the idea, to building them, to actually launching them into space (hence leaving the nest). From what I understand, cranes are also a symbol of a long and happy life, which seems fitting for the Voyager spacecrafts. I also chose this song for inclusion because of the unique instrumentation and because this song is from East Asia. 
4. Tchrakulo – Unlike some of the other regions of Earth (e.g. Europe), the Middle East was not entirely represented. The songs on the original track listing from Georgia and Azerbaijan come the closest to being from the Middle East. I ended up choosing Tchrakulo because it features humans singing. 
5. Melancholy Blues by Louis Armstrong and His Hot Seven was chosen because it met the criteria for including songs that had unique instruments on it, and this was the only track of the original set of tracks that featured brass instruments (with maybe the exception of the Stravinsky track). 
6. Symphony number five in C minor –  At least in the western hemisphere, Beethoven’s fifth symphony, first movement is still enduring and recognizable to this day, in other words it still has cultural weight. In fact, this song inspired a disco song in the 70s and was also used in Disney’s fantasia.  In addition, I chose Beethoven because I like the fact that Chuck Berry happens to have a song called Roll Over Beethoven which is a song about Rock and Roll replacing classical music. I like the juxtaposition of Chuck Berry and Beethoven on this track list. Also in choosing Beethoven I’ve chosen to eliminate Mozart as they were contemporaries and I think one classical artist is enough. Lastly this serves as my song choice from Europe. 
7. Navajo Night Chant –  Like the Alima song, this song makes me wonder if anyone from the Navajo Nation was consulted in its inclusion on the record. From a cursory search on the internet, it seems that this is quite a sacred song. I wonder if perhaps this song is only meant for those who are of Navajo heritage. Nevertheless, I chose this song as my North American pick. 
8. Solomon islands pan pipes – Of the three songs from the South Pacific I liked this one best. It also has the unique sounds of the panpipes, thus meeting the criteria for diverse instruments. Additionally, I did not choose the Australian Morning Star/Mokoi song because according to this article from the Atlantic  it might not mean what the original creators of the Golden Record thought it meant. In fact it might have a darker message than what was intended when it was sent to space.
9. Wedding song – It’s such a tragedy that the young women singing the song has never been identified or credited for the recording. It’s at least a beautiful sounding song and I am curious about the translation, I suspect it’s devastating. Additionally, I chose this as my song from South America and because it features female vocals. 
10. Bhairavi: Jaat Kahan Ho – This article translates the lyrics “Jaat Kahan Ho akeli, gori” as “where are you going alone, girl” and this seems metaphorically fitting for a spaceship alone in space. This song fits a lot of the criteria I set for myself when narrowing down the original track list and yet I hesitated to include this song because of something about the recording artist that I read here by Vikram Sampath: “Being an orthodox musician, Kesarbai was suspicious of the recording medium. She considered it a compromise on the art itself given the limited time available on a record. Her stand was in contrast to the other women musicians who preceded her and easily took to recording. She would often say that her music was not meant for someone sitting in a tea stall and listening to it casually while having a chat.”  Given her stance on recorded music, I wonder if Kesarbai Kerkar would have approved it being launched into space. 

Like I briefly mentioned, my partner and I have been having an ongoing debate about music’s relevancy and the collective memory. We wonder who will be culturally relevant in the future. If I made a Golden Record of music of the aughts, would anyone care in 50 years? What about 100 years? This task of curating the present for the future really elucidates Abby Smith Rumsey’s commentary on the difficulty of determining what data is valuable and needs to be preserved for the future (Brown University, 2017). In her talk at Brown University (2017), at the 38:20 minute mark Smith Rumsey states “what has long term value? We actually don’t know the value of anything until way in the future because its actual meaning is determined by events and contexts we don’t know about.” Ultimately, I do think the concept of the Golden Record is very cool, even if it ends up being kind of meaningless.

References

Brown University (2017, July 11). Abbey Smith Rumsey: Digital memory: What can we afford to lose [Video]. YouTube. https://www.youtube.com/watch?v=FBrahqg9ZMc.

Ferris, T. (2017, August 20). How the voyager golden record was made. The New Yorker. https://www.newyorker.com/tech/annals-of-technology/voyager-golden-record-40th-anniversary-timothy-ferris

Taylor, D. (Host). (2019, April 22). Voyager Golden Record (No. 65) [Audio podcast episode]. In Twenty Thousdand Hertz. Defacto Sounds. https://www.20k.org/episodes/voyagergoldenrecord