What happens to manuscripts when tech-focused companies publish scholarly journals according to their for-profit logic? Unsatisfactory quality. I would like to share my recent experience of the production process as an example, and provide a warning to prospective authors.
For me, the publisher was Springer Nature, which produces thousands of journals. Actual production of accepted manuscripts is out-sourced to Straive, a private equity backed company that integrates AI into its editing and publishing services; business process improvements are coupled with AI tools embedded along the publishing workflow. I don’t have any particular insight regarding the implementation of this technology, but as an author I have interacted with it. After the submission and review process, the accepted manuscript is taken over by the compulsion of productivity, the speed and efficiency with which the business can make a commodity.
If you have an accepted manuscript with this publisher, you will be provided access to their online proofing module, and tasked with quickly reviewing, responding to queries, and making corrections. Your manuscript will have already gone through a copyediting process, which, for me at least, remains opaque. In this eProofing interface you will be presented with a version of the manuscript that may include changes to grammar, punctuation, and style. The changes will be highlighted, but won’t indicate what human or non-human operator has created them. So, while the ostensible task for the author (now proofreader) is to only review for new errors that have appeared, you will also have to review these new anonymous edits.
Authors are not necessarily qualified to proofread this final version, but they are the only ones who care enough to, so they get the unpaid task. My first word of warning is to not assume the document you receive is adequately copyedited, it may not be. In my case the changes were not overly intrusive, I did accept some as improvements; but there were errors, unnecessary comma additions, unexplained deviations from guidelines, and inconsistencies.
Besides the html eProofing interface, you will also be provided with a link to a PDF of the preliminary typeset document. A second piece of advice: you must review this PDF in addtion to the html version, because some formatting appears only there. According to a presentation made by a representative of Straive a few years ago, only about 15% of authors (proofreaders) open this PDF.* Since you may be the last entity to review your manuscript before publication, you had better take a close look. In my case, I did look at the PDF, but as I was focused on the work to be done within the html interface, I missed a few typesetting errors generated by the publishing software. One of these errors was particularly troubling, and represents a technology induced worsening of print-based literacy.
There is a convention of good writing to not end a paragraph with a block quote: generally quotations are inserted within paragraphs, integral to the flow of the paragraph. However, in the print version of my article, the first line of text following every block quote was mistakenly indented, as though it was the beginning of a new paragraph. The Straive publishing software ecosystem does not know the convention. Some mistake was made in converting the manuscript into the typesetting markup language. There must be a suitable command for not indenting text after a block quote, but it was not included. Experienced copy editors and proofreaders, with time to do their jobs, would have seen that the result was incorrect. A comparison with my accepted manuscript would also have sufficed.
I am not the only author whose work has been affected this way. I have since observed that it is a feature of other journal articles. Many, I want to say. Books as well. For readers, it is easy to miss. I am only now obsessively aware of it. These systematic indentations are unacceptable errors, each one is the corruption of a constructed paragraph. The breadth and persistence of the error in scholarly publications is the naturalization of what technology creates on our behalf. Print culture was a set of tools at the disposal of human communication; here, human communication is at the disposal of a set of tools. We can worry about the growing and intrusive use of AI, but AI is just one error-prone machine among others maintained by companies like Springer and Straive, pumping out publications.
This moment may pass. Maybe a systems administrator will hear my complaint. The layout rule of typesetting software can be corrected. The service provider can teach its automaton better behaviour, or be replaced. Not for the sake of quality; the commercialization of scholarly publishing will continue. It will foster the necessity of maintaining, not the substance, but the staging of scholarship. So, improvements will be made, errors will be rectified, quality will be assured. But the current poor results will remain in published versions. The knowledge of this particular error will be transitional, mechanisms will become more opaque and normalized. To maintain revenue from the universities, human-like outputs, scholarly-like outputs, will need to be generated.
Springer Nature and Straive exist to make money. In for-profit journal production, what had been work performed by authors, editors, typesetters, and proofreaders is reduced to segments of a workflow, all subject to replacement by adequacy. AI embedded in publishing processes, other automation tools, and proofreading by authors, follow the same reasoning that leads publishers to outsource in the first place: reducing costs. You, the author, first freely provide the essential input of the value chain. You, the proofreader, then freely contribute further labour to the value chain’s proofing module. You may have worthwhile contributions to make in your field, accumulating an ephemeral “cultural capital”, you may be struggling with professional precarity; but in the global publishing economy you are an atomized and untrained provider of unpaid labour, checking off a workflow box.
* “How the AiKira Publishing Platform from Straive is Setting the Standard”
https://www.youtube.com/watch?v=ZGTwgZc6yAo
Originally published on August 7 as a LinkedIn article:
https://www.linkedin.com/pulse/for-profit-scholarly-publishers-degrading-paragraph-dan-farrell-ni2tc
What is a Petition? 