Looking for "cultural feminism" with corpus linguistics

talk given at the Berkshire Conference of Women Historians
May 23, 2014 slides

Alice Echols' trajectory of radical feminism's decline into cultural feminism is one of the most influential interpretations of the women's liberation movement. She describes cultural feminism as apolitical, celebratory, female, alternative and separatist. Cultural feminism as an ideology has been characterized as racist and essentialist (Alcoff 1988).

Following Katie King I find Echols’ thesis to be “a particular form of U.S. socialist feminism expensively writing a taxonomic history of the women's movements, pedagogically powerful but historically mystifying” particularly as I’ve spent two decades writing about the women around Chrysalis Magazine, which she considers "a major outlet for cultural feminism (Echols 1983). However her thesis has been widely adopted by scholars outside of history.

figure 1

figure 2

I first attempted to use metadata from Google books to characterize the spread of Echol's thesis (figure 1 and 2) which made for some pretty data visualizations, but the metadata was so unreliable.   I then turned to JSTOR,  searching for articles from JSTOR containing "cultural feminism" and with one reference to Daring to be Bad, but encountered more metadata issues with how subjects and disciplines are assigned in their database.   I finally hand assigned disciplines to journals and created a very simple line graph to show the relative adoption and diachronic change (Figure 3)

figure 3

My question today is can we use digital tools to disrupt this comfortable historiography that imposes a kind of clarity on a very messy social movement that spawned major intellectual contributions?

figure 4

I worked with three periodicals, Feminist Studies and Feminist Review (digitzed from JSTOR) as well as Chrysalis digitized with the help of graduate assistances Whitney Esson and Maggie Byrd to explore features of feminist discourse (figure 4). This is a subset of a larger set of periodicals I work with. Feminist Studies founded in 1972 in New York (moved to University of Maryland in Fall of 1977 ), encouraged “analytic responses to feminist issues.   Feminist Review began in London in 1979 as ‘a vehicle to unite research and theory with political practice, and contribute to the development of both.’ Chrysalis started in 1977 in Los Angeles as “a magazine for women’s culture.”  Feminist Review is most strongly associated with socialist feminism, Chrysalis with cultural feminism and Feminist Studies with US academic feminism  I focus on the period of 1978 to 1982 when a series of shifts took place in the women's movement around difference and identity and a changed relationship emerged between activism and academia within feminism. 

Using Antconc, a concordancing software, I ran different kinds of analyses on the corpus, to find patterns in the texts (figure 5) (read more about Corpus Lingusitics here).

figure 5

Starting with comparison of the texts to find words more or less frequent in comparison to the other texts, I zoomed in on this difference around women and woman. I found that potentially fascinating ( is this difference an indicator of an individualist v a communal discourse of essentialist v constructionist?).  However I can't just look at women, first of all because there are far too many occurrences, but more importantly because I need to get beyond just counting and searching.  I need to look at women in context, which is why I also looked at clusters, words that co-occur.   

figure 6

women from appears in almost a third of the Feminist Review files.  From is tricky though as it may be used to  to indicate source or origin): ”women from both the third world” may be used to express discrimination “prevent women from being as economically efficient” or it may be used to differentiate “clearly distinguished working women from bourgeois women.” 

With that in mind I start sorting by various words to left and right to see what sorts of patterns emerge (note if I had this corpus POS tagged this would be much easier but alas I do not yet) Of the 58 lines containing women from pretty easy to identify all but 7 into these three categories. The use that predominates is the second sense of “from” to discriminate.  The verbs preceding women from  are emancipate, exclude, free, liberate, prevent, prohibit, separate. In this case, of course, the lines are talking about how to STOP.   A smaller set designates specific women “women from the republic of ireland” “women from merchant families” While interesting, none of this gets me any closer to something called cultural feminism.  

figure 7

To do that I need to compare how women from appears in a second text.  Although “women from” is not as frequent a cluster for Feminist Studies  I compared usage there.  Conducting a similar analysis of patterns with preceding verbs and clauses that follow women from it would appear Feminist Studies is more about discrimination against women than Feminist Review However as the lines shown here reveal there is an interesting co-occurrence that does not appear in Feminist Studies protect isolate separate. I needed to refine my analysis.

figure 8

I zoomed back in to look more at the co-occurrence of words with “women from.” Feminist Review has a chunk of content that refers to empowering women  “liberate women from the oppression” “free women from childcare” and “emancipate women from tradition.” Feminist Studies is not only smaller in use of these terms, containing no uses of free or liberate, but also contains a different usage all together protect. These co-occurences highlight labor “protect working women from the onerous conditions” as well as “domestic violence.”  Chrysalis which is the largest corpus has the fewest instances and they display only very weak alignment with the uses we see in the other two corpora  (figure 8). 

Plotting these usages along the axes Echols uses to define feminism in a US context is impossible. She argues that “liberal feminists” worked “to eliminate sex discrimination in the workplace” (139) but also at the 1969 Congress to Unite Women, both women’s liberationists and NOW feminists were interested in addressing “discrimination in education and employment” (343) and that the infamous SNCC position paper pointed to instances of “sexual discrimination” (30). Discrimination as a marker clearly cuts across all lines.


 I needed to find another pattern to look at to explore these idea further.  The second cluster I identified,  against women seemed very interesting to me to given the conflation of cultural feminism with "victim feminism" or an over emphasis on women as victims of violence against women.  Again I identified the co-occurences in each line (contained word or variant of word shown in graphs.) 

figure 9

For Feminist Review  against women appears in about 25% of the files and again, as in women from, has an association with the idea of discrimination either systemic “historical analysis of legal discrimination against women in Britain and the United States” or against groups as in “discrimination against women artists.” However most of the occurrences are not aligned with any clear meaning. Interestingly a term we might expect “violence against women” pops just 2x in just 1 article in 1981 'I Just Wanted to Kill a Woman.' Why? The Ripper and Male Sexuality by Wendy Hollway 

figure 10

Conversely, Feminist Studies contains against women in fewer files (15%) than FR (25%) even though it is the bigger corpus and against women displays a stronger association with violence as in “rally to end violence against women” and “upsurge of incest and violence against women.” As incest in this last example hints at, Feminist Studies contains an association of sexual with against women as in “sexual offenses and domestic crimes against women” or “violence against women, sexual harassment, and reproductive.”   

figure 11

This association is even stronger in Chrysalis, the largest corpus which contains a similar percentage of against women as Feminist Studies (15%) (figure 11).  “rape is a primary form of violence against women in patriarchy” “male violence against women” “conference on violence against women.” 

How would we place these periodicals then on a spectrum of feminist ideologies?  Echols cites as a key moment in the transition of feminist thought this shift to focus on physical violence done to women by men:   “previously when radical feminist spoke of violence they were often referring to …”the violence of the mind” [and] “psychic damage”… “while radical feminist sometimes acknowledged the role played by physical violence in maintaining male supremacy they tended to emphasize other [institutional] factors” (201).  Chrysalis contains the greatest emphasis on overall violence, while Feminist Studies is larger for a focus on sexual aspects.  Clearly looking closer at uses of rape and violence  might help us to sort this more. 

Because this notion of men perpetrating violence against women appeared in the corpus,  I went back to the trigrams that popped in my initial analysis women and men and men and women (which appear respectively in FS and FR in about 40% files at freq greater than 100) (This is not to say that the patterns don’t appear reverse in the other corpora, just in smaller proportions and with even less strong associations) Were these conjoinings or comparisons? Again I identified the co-occurences, although we don’t see as strong associations here as we did in the prior example women from. 

 In what has now become a familiar trend I see that Feminist Review is more about difference, equality and work, while Feminist Studies contains relationships, family and work. Difference is an extraordinarily complicated term in feminist discourse, so I’m going to bracket it here. Echols asserts that “cultural feminists organized women around the idea of female difference” while “radical feminists mobiliz[ed] women on the basis of their similarity to men” (6). I don’t think anyone ever claimed the women of Feminist Review were cultural feminists as their socialist leaning are pretty well acknowledged so clearly a term like “difference” would require much investigation and one that takes into account change over the course of the movement to account for the introduction of French feminist theory. 

figure 12

Turning then to  Feminist Studies the emphasis on relations and the family might highlight what we call a personal is political approach as in “which made personal relations between women and men a central political issue” or “of the family and of personal relationships between women and men, the political consequences of” Echols identifies “the family” as a pivot point in feminist ideology. While socialist feminists saw the family as an economic unit, radical feminists sometimes argued that a focus on the family as an institution shifted blame away from men. Ellen Willis quoted in Echols as “to say … that the family oppressed women was to evade the fact that our husbands and fathers oppressed us”(148). so called cultural feminists focused on forming their own alternatives to the nuclear family rather than challenging its hegemony or reclaiming it as a site of potential empowerment.  Echols' quotes Robin Morgan’s gloss of Jane Alpert’s Mother Right “each individual woman gaining self-respect and yes power … within her family” (251). This all hints at what Echols’ characterizes as “cultural feminisms’ ” retreat from confronting patriarchy. The problem here is that these trigrams are not as salient for Chrysalis. While they do occur  it is in numbers too small to identify. [men and women present in only about 25% and women and men in 15%] What does seem clear is that men is underutilized in Chrysalis as compared to the other 2 corpora (men overused in FR compared to FS). Does this mean we rank them on an engagement with patriarchy scale of most Feminist Review to least Chrysalis? The problem with that analysis is that patriarchy is over-used in Chrysalis as compared to the other two texts.  Clearly I need to investigate patriarchy further.


That leaves us with how women appear in the three corpora (figure 13).   By this analysis it seems clear that  in Chrysalis there is more affiliation between women and identity modifiers than in the other two journals. This would seem to at the least weaken Echols’ assertion that “cultural feminism” emerged as a way to evade difference among women. 

figure 13

All of this was to test out my ideas and methodology to see if this project warranted expansion, which I believe it does.  I have a larger data set coming of  Frontiers: A Journal of Women Studies, Letras Femeninas , Questions Féministes Signs, and Women’s Review of Books. I hope to create more digitized movement periodicals and especially to add black feminist works to my corpus.

I'd love feedback either here in the comments, via email professmoravec at gmail or via twitter @professmoravec

Works cited
Alice Echols "Cultural Feminism: Feminist Capitalism and the Anti-Pornography Movement". Social Text,  Vol. 7, pp. pp. 34-53.  Alice Echols, Daring to Be Bad: Radical Feminism in America, 1967-1975 (Minneapolis: University of Minnesota Press, 1989). 
Katie King, Theory in Its Feminist Travels: Conversations in U. S. Women's Movements, Indiana University Press, 1994