The proprietary publishers have established an elaborate co-dependency relationship with academics. Academics depend on journal editorships and citations for promotion. Editors get many perks and prestige as a result of being an editor, but the selection of who becomes the editor is up to the publisher. Reviewers get pre-publication access to results. Yes, the reviewers are supposed to hold the information in confidence, but does pre-publication access help them in thinking about which directions to take in their own work? Absolutely. An extensive web of co-dependence has evolved between the proprietary publishers and the academic community.
Academics generally do not receive royalties from journal articles, but they do from book publications. Who publishes those books? The same publishers that publish the proprietary journals. Who selects which authors will be invited to publish books? The publishers.
Elite institutions and large university systems negotiate discounted and preferred subscription agreements giving their researchers free access to a wide range of journals, which in turn makes it more attractive for academic "stars" to go to those institutions. The faculty at those schools benefit from these favorable access agreements. Are we surprised that University of California faculty voted against open access?
It is also not just speech and language research. The majority of work in fields like cancer research is also published in paywalled journals. Cancer patients may not be able to wait a year before articles appear in open access archives.
The vast majority of academic work is supported by public funding, and charitable foundations support most of what is not government supported. High time to require open access. The academics are not going to do it themselves.
Never trust an operating system.