Houston’s piece represents my first encounter with an article that applies computational text analysis. Computational text analysis cannot replace rich, deep qualitative and ethnographic narratives, but it can abstract large bodies of text that an individual could not observe and produce consistent diachronic descriptions of context. Though perhaps the scope is not very sophisticated compared to other methods, it can be useful when combined with other methods and data.
I reproduce below some key ideas from the article to understand what CTA does and how:
Feminist Computational Criticism: involves a reflection on not only the object of study, e.g. Victorian women’s poetry, but also on the decisions to make selecting the dataset, conducting analysis, and interpreting the results.
Text Analysis: the predominant analytical method in literary studies is close reading. The primacy of the text in “service of a larger theoretical approach.”
Distant reading: “examines the socio-cultural systems of value that produce the very category of literature itself.”
Analysis: use of free open software: Antconc, a corpus analysis tool, and the R programming language and libraries. Sample: 1,284 poems in Stedman’s anthology.
Unsupervised methods of machine learning: “a form of distant reading that can bypass some biases of human judgements. Unsupervised machine learning algorithms harness the computer’s pattern-matching power to reveal semantic patterns in texts, without any predetermined instruction in what to look at or what is significant.”
Topic modeling: “approach to understanding large sets of texts using an unsupervised machine-learning algorithm that implements an iterative, probabilistic assessment of word-occurrences in the documents in the corpus (…) it gathers information about which words co-occur in the same document, and the probability that those same words would co-occur in another document of the corpus.”
What kind of discourses or themes are present in the volume? Topic model “can help us examine semantic patterns in large sets of texts to see what kinds of discourses are present.”
Topic: “words that co-occur in documents at a greater than average probability together make up a topic, a cluster of semantic meaning.”
“The programmer selects the number of topics that the algorithm will locate in the corpus (…) the algorithm was programmed to discover 15 topics.”
Stopword: “stopword list was applied to the corpus before modeling it, excluding common articles, pronouns, and numbers.”
Black feminist scholar, Patricia Hill Collins, popularized intersectionality in the nineties. Carastathis makes a detailed intellectual history of intersectionality pointing when it was introduced for the first time and what conceptual predecessors it has. Intersectionality aims to encompass multiple forms of inequality that are organized via a similar logic, to encompass in a single word the simultaneous experience of multiple oppressions. Carastathis highlights four analytical benefits of intersectionality as a theoretical paradigm and research methodology: simultaneity, complexity, irreducibility, and inclusivity.
Simultaneity: the author refers to simultaneity as the nonfragmentation of a phenomenological experience. Intersectionality as a knowledge project seeks to explain the ways in which multiple dimensions of inequality intersect and co-create one another in social life and in institutions. It points out that race, class, gender, sexuality, ethnicity, nation, ability, age operate not as unitary, mutually exclusive categories, but are always intertwined with one another
Irreducibility: intersectionality moves beyond a mono-categorical focus on inequality. Traditional approaches to study inequality foreground a single dimension of inequality such as race or class and conceptualize these processes as parallel. Intersectionality points out that race, class, gender, age, ability, nation, ethnicity, and similar categories of analysis are best understood in relational terms rather than in isolation from one another. It favors interpretative tools that can show the relational complexity of things, that indivisibility is important – therefore intersectionality does not simply add new variables.
Complexity: intersectionality introduces a greater level of complexity into conceptualizing inequality. Asking questions about, for example, how racial inequality works in the US, we will not understand or capture the complexity of racial inequality if we do not consider how citizenship, ability, sexual orientation, race, and other ways in which human beings are categorized and organized work together to produce inequality.
Epistemic standpoint: Individuals and groups differentially placed within intersecting systems of power have different points of view on their own and others’ experiences with complex social inequalities, typically advancing knowledge projects that reflect their social locations within power relations.
Intersectionality as praxis, social justice: intersectionality is positioned in the academic literature and in activism as a critical social theory. It not only describes how inequality works but makes interventions and thinks about how we can make the world a better place. Intersectional scholarship is explicitly committed to social justice.