Lit Review
Discourse is the overarching, dominant, cultural or institutional “narrative,” and it circulates hegemonically or hierarchically (“vertically”) more or less independent of context (Nafría and María). Dialogue is the process of exchanging “messages” (I assume for any purpose) and it occurs collaboratively (“horizontally”) and is shaped by context (Nafría and María). But I think the relationship between the two is still reciprocal. If I think about my analysis of the poems and literacy circles using these concepts, I might look at how the poems and the dialogue among people “merely” reflect the overarching discourse, but also how those dialogues might disrupt or resist that discourse. This also helps me think of Freire’s work as using dialogue to critique and resist the discourse (“By stimulating ‘perception of the previous perception’ and ‘knowledge of the previous knowledge,’ decoding stimulates the appearance of a new perception and the development of new knowledge. During the decoding process, the participants…externalize a series of sentiments and opinions about themselves, the world, and others…)
Research Questions & Methodology
For my final project I want to analyze digital dialogues (written) about colorism to understand how they reflect and/or resist the dominant discourse on colorism.
In order to accomplish this I will analyze the following sets of texts:
- blog posts
- comments on blog posts
- tweets
- Digital/digitized journal articles, theses, books, dissertations, etc.
using some or all of the following (and possibly other) tools:
- Voyant
- Mallet
- sentiment viz or Python NLTK.
- Stanford NER
to perform the following (and possibly other) tasks:
- Word frequencies
- topic modeling
- Named Entity Recognition
- Sentiment Analysis
In the scope of this semester, my goal for this project is simply to mine the data and see what happens. Therefore, my final product will be an analytical report (generously illustrated with various visualizations, of course).
Rationale
The rationale for such a report is that it will be useful in ‘perceiving perceptions’ and ‘knowing the common knowledge’. It can be a tool for “conscientização” (Freire) or “consciousness raising” in the context of critical dialogue.
Additionally, a clear understanding of the broad, dominant discourse on this topic facilitates the identification and understanding of contraflow when, where, and how it occurs.
This might possibly be the largest “lit review” on colorism ever attempted.
Limitations
One significant piece of the discourse that I’m not sure how best to track (and so probably will not track for this final project) is all of the non-written dialogue. YouTube videos, podcasts, Vimeo videos, songs, speeches, memes, photos, illustrations, etc.
Super interesting, Sarah. I would love to see your findings from this! I am eager to see how the text mining might yield some insight into *why* the term colorism has not been more universally accepted and used. Your post also led me to think about citation politics and how that (at least for academic texts) leads to certain texts becoming more dominant while obscuring others, and non-academic venues (tweets, eg) circumvent the academic structure of replicating hierarchies.
I’m interested in language as used in two specific mediums of communication—academic journal articles and tweets. Furthermore, I’ve narrowed my attention to articles and tweets that explicitly use the term “colorism,” first coined by Alice Walker in 1983 yet still not universally accepted as an official term the way other words are like sexism or racism. Using digital tools to analyze word frequencies, n-grams, and named entity extractions, I will seek preliminary answers to these questions: When people explicitly discuss colorism in academic journals or on Twitter, what are the most common subtopics and key terms or phrases? When people explicitly discuss colorism in academic journals or on Twitter, what entities appear central to these discussions? The initial answers to these questions may be compared within and between the two datasets so that I may then ask the question: How is colorism addressed similarly or differently across time and medium? My goal for generating this report is to get a better sense of the discourse and dialogue on colorism in these prominent mediums. With that understanding, it might be possible to identify or imagine productive dialogical (see Freire) supplements or alternatives for addressing colorism.
This is great! I really love the idea of looking at social media and dialogue to compare to more dominant discourse on colorism.
Are you going to combine the different social media samples together or run them through from separate sources and see how they differ? What sources are you comparing it to? So what sources on colorism are you reading it next to?
The reason I was asking about different social media sites is I wonder if results would be different on different media sites. Maybe you could even do both? Combine all results and see what threads you find and possibly look at them separately and see if there are any distinction?
I think Python would work well for your project, as Python allows you to see different data sets, particularly being able to see it in contexts more easily (or I think more easily) than Voyant.
Thanks for your comments. I will probably do both, combine all social media but also analyze them separately. With Twitter, it’s so easy to archive all the tweets on colorism because people have already designed easy to use apps for that. I’m using Twitter Archiver which is a Google Sheets add on and it gives me lots of good metadata on the tweets too. I haven’t found a similar tool for any of the other social media sites, so I’m not sure how robust my data set will be from the others. In terms of what texts I’m comparing/contrasting to or using as my “test group,” I’m ultimately interested in how the poems that people submit to my contest compare to other dialogues and narratives etc. about colorism. I agree that Python is probably the better tool, but I’m still intimidated by it! Maybe this’ll be the push I need to get good at it.
This project will, no doubt, produce some fascinating insights. I’m curious to know if you think it would useful to organize your research in chronological chunks. When I first read your post, I thought, “It would be interesting to see how conversations re: colorism have (or have not) changed in the past two or three decades, especially with the advent of blogging.” However, I see how it could also be fruitful to analyze a recent blog in parallel with, say, a group of journal articles from the 1980s, especially if both are concerned with a particular sub-topic related to colorism. Are you planning to classify your data by year of publication before (or maybe after) you analyze it?
Hi Gayle,
Thanks for the chronological idea! I actually hadn’t considered it because I was already assuming overarching continuity. But you are right. A chronological organization might reveal some interesting patterns.