In creating media based on myths, symbols, and idols, we may lose the value and meaning of the original if recommender systems and search algorithms do not find a harmonic balance between meaning and object. The burden lies on the recommender to recognize the difference between the “common” media definition and some other potentially more pure definition or connotation. Whether the former is any more appropriate than the latter is also investigated, along with a generalized form of this phenomenon, comparing words as they exist on common search/recommendation engines, and the same words when they connote other (prior or “purer”) concepts.
This morning, I was reading a few pages from a book, and there was a passage that referenced Prometheus. To remind myself a bit more of his story and the history behind the name, I decided to google1 “Prometheus,” and I found that it directed me (i.e., its very first result took up the whole of my iPhone screen2) to the 2012 film Prometheus, directed by Ridley Scott, starring Michael Fassbender and Charlize Theron. This got me thinking about the theoretical person who has absolutely never heard the name “Prometheus,” and has no idea that it could be anything more than just the title of a science-fiction movie about mercenaries and space travellers. Further, I decided to search for exactly the same word (with the same capitalization, in case that is sensed as well) on DuckDuckGo. Now, the very first result (which instead took up about half of my screen, beginning the list below) is an image and snippet from and a link to the Wikipedia article about the Greek mythological Titan, Prometheus.
Symbols and synecdoches, metaphors and metonymies, adaptations of words used in media to represent some related but different entity than that for which the word was originally intended. It is not a crime, of course, but let us try to be precise about what is happening when we search for (namely, when we google) the meaning of terms.3
Let $M$ be the media (be it social or otherwise), where $M(x)$ is the entity $x$ as it is portrayed/defined in the media, $E(s)$ represent the ease and frequency with which we access some source of information $s$ (e.g., $s = M$ when that source is media), and then let $\Omega(X)$ be our understanding of the world under some condition $X$.
I’m led to believe that there must be a more pure source of information $P \neq M$ that precedes $M$. That is, $M(x)$ is necessarily informed by $P(x)$, and we tend to think because of this fact, the more original intention $P(x)$ is more pure than $M(x)$. In another example, consider the movie The Imitation Game, starring Benedict Cumberbatch and Keira Knightley. In hopes that I do not give away any spoilers (I am vehemently against unsolicited spoiler baring, so if you are at all worried, maybe skip to the next paragraph), this movie contains no significant reference, explicit or otherwise, to the original Imitation Game invented and described in Alan Turing’s seminal work Computing Machinery and Intelligence.4 Instead, there seems to be a sort of allegory being made between the words of “The Imitation Game” (namely, the word imitation) and the actions taken by the UK in response to Germany during the war. In no way, though, are we talking about The Imitation Game described by Turing himself. Here, we have the idea behind the phrase “The Imitation Game”, $m$, the way it is defined/used in media, $M(m)$, and the original coining of the phrase which is being referenced, $P(m)$.5
Consider someone who has never read or heard of Turing’s paper, but has seen this movie, and in so doing makes a connection between $m$ and $M(m)$. When this person hears the words “the imitation game,” especially used in reference to the movie, they will think of some plot point or message being made in (or of) the movie rather than one made by the original author of the paper for which the movie’s title was inspired. I don’t think that it is a farfetched statement to say that $E(M) > E(P)$ in this case, especially (if anything) considering the difficulty with which it can be to read the paper compared to the ease with which we can enjoy the film. And, I believe that since the proliferation of blockbuster movies like this one is so immense, people of this kind (i.e., those who ingest $M(m)$) are becoming more and more likely than people of the alternative kind (i.e., those who ingest $P(m)$). Does this mean that through time, we as consumers of the media are becoming less familiar with the purer forms of human invention like Prometheus or The Imitation Game? Do we inadvertently bury the beauty of purer forms of thought by bastardizing the words used to describe them?6
But then, what is “pure”, or “purer” for that matter? Why does/n’t it matter? Can we design a scenario where $M(x)$ actually manifests as a portrayal of $x$ which appears more pure than $P(x)$? We’ve already shown above (i.e., with the examples of “Prometheus” and “The Imitation Game”) that the reverse is quite possible, and our notion of purity seems to correlate with primacy. But, this might not be strictly necessary. Though I am currently at a loss for a sound example of entity $s=c$ in which $P(c)$ seems less pure than $M(c)$, I leave it to the reader to consider it a possibility, potentially sharing examples with me via email.
It seems to me that purity of meaning is somehow intuitive, subjective, and illogical. If words are only symbols to represent something else in our world, and if meaning is just the connection between words and their corresponding “something(s),” then what’s to say one meaning is purer than another? If I experiment with the word “tree,” considering the fake plastic Christmas tree sitting now in my living room (say, for some theoretical person, all trees they knew were just like this one), and the Oneonta tree currently standing at Rockefeller Center, something inside tells me the latter is an example closer to the purer meaning. Without skirting too far into the ideal/realism argument, I think there is something to be said about our intuitive understanding of purity, and here, how it changes over time.
Maybe this is just a sort of special case of historical bias or deployment bias. But, I’m convinced that it should be treated at least slightly different from either. For example, with either of these kinds of biases, we’re talking about inevitable changes in society (take the evolving intricacies of macroeconomics, for example) or allowing an algorithm to be deployed or used in a way it was not originally intended (e.g., using Facebook to as a news source). This discussion of portrayal seems different though, referencing how we use algorithms to offer up the definition of a thing, or how we portray an idea from the media that was once not in the media.
When searching for a phrase, users (consumers) are interested to know what that phrase is. But then, do they want to know what something is in the context of something more recent and easily accessible, or do they want to know what something is in reference to a “purer” potentially more original context?
a few questions to ponder
Is our understanding of the world at all affected by Google searches? When concepts have been adopted and remade in media, and when these adoptions pervade societal interactions more than their original form, what reason do we have to respect the latter as a primary definition? How does this generalize, and why does it matter?
Interesting that this has become its own self-proclaimed verb. Google, the company, has influenced our world so much that society decided to create the generonym (once neologism) “google” that we use so fondly. It’s worth pointing this out, as the reach of Google is an important aspect of this essay. ↩︎
Very likely the primary mode of googling and looking things up nowadays. ↩︎
I find that “Search Engine Optimization” and “Recommender Systems” have become sufficiently intertwined. In searching or googling, we (users) seek a recommended article (or set of articles) to shed light on what is in the search bar. The observations I make, below, must be a consequence of our brave new dictionary paradigm. ↩︎
As an aside, I find it a bit unfortunate how often this paper is misunderstood. For one, sure, Turing does propose the game as a sort of “test,” but I am of the school of thought which believes he had absolutely no intention of testing whether a machine can think, on the contrary, the whole idea of the paper is to do something completely different. ↩︎
To be fair, I’ll grant that this phrase (or even idea) “The Imitation Game” could actually be even older than Alan Turing himself. But, even if that’s the case, either it is just another unrelated use of the phrase, or we have an older version of the same dichotomy, and the point made above still stands. ↩︎
I will grant that the ideas I’m discussing here are roughly related to a similar phenomenon in social media, and could be extended to describe something parallel to Chomsky’s viewpoints on the evolution of language. Though, to be fair, I’m discussing in particular the position of the recommender system (e.g., Google’s algorithm) as a crucial actor in the preceding question. ↩︎