Training for ethical linguistics: a model for building “responsible conduct of research” into department culture

1 Introduction: ethics for a diverse discipline

The contemporary discipline of linguistics is intellectually diverse. Linguists study an enormous range of topics, and they do so from diverse theoretical and methodological vantage points. They work alone and in teams. They produce data ranging from audio and video files to EEG data to systematically elicited grammaticality judgments. They build and query databases. They study speech sounds, vocabulary, grammar, semantics, language variation, and processes of change. They study specific languages and relate the patterns they find to the human cognitive capacity for language. Their methods range from computational modeling to running regression analyses to notebook-and-pen documentary fieldwork in rural villages. Linguists make recordings, transcribe and analyze speech, conduct physiological and psycholinguistic experiments in controlled settings, develop software, and collaborate with communities to support language reclamation and educational initiatives. They study language in spoken, signed, written, and whistled forms. They scrutinize eye gaze, gesture, joking, metaphor, and song. They take interest in the utterances of children, adults, elders, neurodiverse, and queer individuals, and mono-/bi-/multilinguals. They study the language of people from all backgrounds, situating the patterns they document at individual and societal levels. Because using language is part of virtually everything people do, there is hardly a human activity in which there would be nothing of interest for linguists to study. And every one of linguists’ topics, methods, data types, and research situations carries along with it a distinct configuration of ethical considerations and potential pitfalls.

Furthermore, like people trained in other disciplines, linguists do many things besides conduct research: they lead multifaceted professional lives. Those in academic positions teach in settings from community colleges to online classes to university seminars. In that capacity they structure their classrooms and issue student grades. They teach at different levels, advising undergraduates on their programs of study and giving career advice to graduate students. They sit on committees, evaluate their colleagues, run meetings, and write reports. They supervise staff and teaching assistants. They advocate for their programs within and beyond their institutions and participate in hiring decisions. They write funding proposals, journal articles, and books; they review and edit proposals and manuscripts written by others; and they manage grants. They accept and extend lecture invitations. They attend conferences where they share their work and give feedback to colleagues. They respond to calls from journalists seeking comment on matters of current interest. The responsibilities of linguists who work in professional settings outside academia are so varied they are even more difficult to summarize.

What can it possibly mean to provide “training for ethical linguistics” given the enormous diversity of what linguists do? The rest of this article will attempt to provide a meaningful answer. But before doing this it will help to more carefully consider the two terms “linguistics” and “training” that the question presupposes.

2 Ethics for linguistics?

It might seem self-evident that the domain of an ethics training program for linguists should be the research-related aspects of linguistics. Indeed, the very existence of the present publication suggests that there is a substantive common denominator that all linguists share that brings into play an associated array of ethical norms and concerns (though the paragraphs above are meant to raise doubts about that prospect). This question of how the domain should be construed for purposes of ethical guidance and education was one that vexed the Linguistic Society of America (LSA) when, for the first time in the society’s history, it chose to take responsibility for addressing ethics as part of its mission. In 2006, an informal advocacy group coming out of that year’s meeting of the Society’s Committee on Endangered Languages and their Preservation (CELP) proposed to the LSA Executive Committee that the Society take a leadership role in regard to ethics. (The matter arose in CELP at that time because of the new attention that was being paid to linguistic fieldwork in endangered language linguistics.) The proposal, which I helped compose, was that the LSA strive to cultivate a reflexive discourse within the discipline about the kinds of ethical issues that arise in linguistic research. The executive committee responded by establishing a standing ethics committee, and the LSA adopted its first Ethics Statement a few years later, in 2009.^[1]

But between those two moments was a significant process delay while the relevant constituencies contemplated and debated whether an LSA ethics statement should address what I jotted down in my notes during a phone meeting at the time as “normal ethical conduct not specific to the field”. On the one hand, if people in all fields need guidance on matters such as peer review, mentoring, data management, and so on, then one could argue that it is unnecessary to include them in a discipline-specific ethics statement. On the other hand, if the LSA was the main professional society to which linguists belonged, then that is where they would be looking for guidance on matters of ethics in all areas, not just those that pertained to their research. In the end the latter position prevailed, and the Ethics Statement that was issued included text addressing linguists’ responsibilities to students, colleagues, and the public. These dimensions were elaborated further in the 2019 revision, which reflects the changing times and takes the perspective that being an ethical linguist involves whole persons in all their capacities (LSA Ethics Statement, Section 2):

It is the responsibility of linguists, individually and collectively, to ensure that we uphold the ideals and principles of equality and anti-discrimination in our professional interactions – in our language and in our behavior, on a personal level, in our interactions with others, in our institutional roles, and in our academic commitments […]

Furthermore, these responsibilities […] apply to all settings where linguists conduct their work – such as regular workplace settings, both live and digital, field settings in which many linguists train and work, and other settings associated with professionally-sponsored programs and activities (e.g. meetings, publications, honors and recognition, governance programs, and all appointed, elected, and volunteer positions).

The question the LSA Executive Committee pondered is important because before one can even begin to think about what “training for ethical linguistics” might look like, it is critical to clarify the boundary between what properly belongs inside the work we do “in linguistics”, as opposed to what can be safely ignored as beyond its scope.

Most instructive in regard to that boundary are two major ethics crises that troubled the LSA (and the field of linguistics more widely) in recent years. In 2017, following accusations that University of Rochester language scientist Florian Jaeger had a history of harassing women,^[2] the LSA was urged to set up a task force on misconduct that resulted in the Society adopting a code of conduct for all LSA meetings, institutes, and other Society-sponsored events.^[3] In 2020, as part of the public outpouring of fury over the murder of George Floyd at the hands of the Minneapolis police, the LSA was again called upon to act in the name of ethics; an open letter to the LSA with over 600 signatures demanded that the Society remove Dr. Steven Pinker from its roster of distinguished fellows and media experts. The letter presented instances of social media posts going back over nearly a decade in which Pinker seemed to be using his professional authority to downplay racist violence in defense of an unjust status quo. In response, the Society convened two more task forces, one to revisit its process for inviting (and un-inviting) Fellows, and another to establish explicit policies and procedures for endorsing public experts (see Kastner et al. 2022 for a timeline and summary of what transpired).

These cases of controversy are interesting because the ethical breaches they raise – sexual harassment, speaking publicly in a way that seems to minimize the seriousness of societal racism – have nothing to do with linguistics as such. Nor are they related in a direct way to the conduct of research. Rather, they are about the ethics of wider professional life for those committed to “doing linguistics”.^[4] Given the prominence and importance of these challenging cases, it seems that the LSA was on the right track when it constructed its Ethics Statement to apply to the more expansive understanding of “linguists” as a professional identity category, rather than “linguistics” as a topic of research. I endorse this understanding, and in the remainder of this article it is what I will assume.

3 Ethics training?

In the United States and Canada, linguistics students often receive their first introduction to ethics education in the form of university training focused on the protection of human subjects, the individuals whose speech, bodies, and actions provide data for answering researchers’ questions. For researchers to work ethically with people in this capacity means (at minimum) safeguarding their welfare: seeing to it that the harms incurred by participants are minimized, that their consent to participate is properly informed, and that they understand and limit the circulation of any data that identifies them. In the United States, human subjects protections are overseen by Institutional Review Boards (IRBs); in Canada this responsibility is held by Research Ethics Boards (REBs). IRBs and REBs are administrative bodies meant to ensure compliance with national ethics regulations.^[5] These in turn operationalize more general ethical principles like respect for persons (respecting individuals’ right to make choices for themselves), beneficence (striving to do good and avoid harm), and justice (fairly distributing exposure to risks and access to benefits of research).^[6]

Although high-level ethical principles such as these are worthy aspirations, it is doubtless imperfect to manage adherence to them through a legalistic system focused on auditable actions like data storage practices and creating records of informed consent. Many have argued that the humans subjects regulatory framework is a flawed means of ensuring the ethical quality of work in the humanities and social sciences (see, e.g., Brydon-Miller and Greenwood 2006; Lederman 2006; Schrag 2011; van den Hoonard and Hamilton 2016).

Also problematically, the prominence given to human subjects training in formal ethics education means that ethics is often encountered by novices as a matter of regulatory compliance, i.e., a monitored following of pre-established rules. This has some unfortunate repercussions. There is a danger that compliance-related concerns “eclipse the contemplation and discussion necessary for responsible disciplinary practice” (Dobrin and Lederman 2012), as anxieties about gaining research approval understandably dominate students’ experience. Associating ethics with rules separate from the researcher and the research situation – an approach sometimes referred to as deontological ethics (Alexander 2021), procedural ethics (Guillemin and Gillam 2004), or in the literature on applied linguistics, macro-ethics (Kubanyiova 2008) – means that formal institutional approval can come to be taken counterproductively as a proxy for ethical research itself. Furthermore, in order to ease the administrative burden of compliance tracking, many North American institutions in the U.S. provide training in the form of online modules, which are too easy to complete and then forget.^[7]

Compliance-oriented ethics training makes sense if the point is making sure researchers know and agree to follow a set of rules. Of course, it also makes sense from the point of view of institutions eager to protect themselves from potential lawsuits. But it is far from adequate for preparing members of a community of practice to exercise ethical judgment in all aspects of their professional activities throughout their careers. It is inadequate, in other words, for the cultivation of a generalized ethical sensitivity, practical wisdom (Luckman and Gunsalus 2023), or as one European research ethics handbook puts it, ethics as “a state of mind” (Pauwels 2007: 20). For that, an entirely different approach is called for, one that is more akin to socialization into ways of being – habits of reflection, discourse, and behavior – than it is to following rules. A recent exploration of how researchers understand the development of professional integrity suggests a process not of training but of “sensitization” to acting with integrity, something that arises out of past (especially early) experience and exposure to role models both positive and negative (Satalkar and Shaw 2019: 6). The purpose of systematic ethics education would be to reinforce and further develop this ethical sensitivity.

4 Responsible conduct of research: a broader framework for ethics education

Focusing training on human subjects regulations is problematic not only because of its emphasis on rule following. It is also too limited in scope, as it draws learners’ attention to only a narrowly defined fraction of the ethics-related challenges they are likely to face in their capacity as professionals: as discussed above, recent public ethics controversies in linguistics have centered on the relation between researchers and one another and the relation between the profession of linguistics and society at large, rather than on the way research itself was being carried out.

But there is another set of regulations that I believe can be built upon to play a useful role in structuring ethics education because it is capacious in content and less restrictive in its compliance requirements. In the late 2000s, a group of U.S. federal agencies began requiring that institutions plan to “provide appropriate training and oversight in the responsible and ethical conduct of research” to students and postdoctoral scholars involved in sponsored research.^[8] The term generally applied to these regulations is “Responsible Conduct of Research” or “RCR”. The regulations require that institutions seeking funding from the relevant federal agencies – which include the National Institutes of Health (NIH), the U.S. Department of Agriculture’s National Institute of Food and Agriculture (USDA NIFA), and the National Science Foundation (NSF) – attest that their researchers have been provided with training in RCR. Although I will be focusing specifically on the U.S. regulatory framework in this article, documents with similar coverage have been issued in the time period since then by international bodies such as the World Conferences on Research Integrity (WCRI) and the All European Academies (ALLEA).^[9]

RCR is concerned very broadly with ensuring that researchers will be ethical actors in the social world. In addition to human subjects protections, the domain of RCR covers conflicts of interest; mentor-mentee relationships; collaborative arrangements of different kinds; the peer review process; data management, sharing, and rights; misconduct; authorship and publication; and societal impacts of research. The full listing of RCR subject matter from the most recently issued guidance is presented in (1).

1. Topics that fall within the scope of RCR instruction according to 2022 NIH guidance.^[10]

   1. conflict of interest – personal, professional, and financial – and conflict of commitment, in allocating time, effort, or other research resources

   2. policies regarding human subjects, live vertebrate animal subjects in research, and safe laboratory practices

   3. mentor/mentee responsibilities and relationships

   4. safe research environments (e.g., those that promote inclusion and are free of sexual, racial, ethnic, disability and other forms of discriminatory harassment)

   5. collaborative research, including collaborations with industry and investigators and institutions in other countries

   6. peer review, including the responsibility for maintaining confidentiality and security in peer review

   7. data acquisition and analysis; laboratory tools (e.g., tools for analyzing data and creating or working with digital images); recordkeeping practices, including methods such as electronic laboratory notebooks

   8. secure and ethical data use; data confidentiality, management, sharing, and ownership

   9. research misconduct and policies for handling misconduct

   10. responsible authorship and publication

   11. the scientist as a responsible member of society, contemporary ethical issues in biomedical research, and the environmental and societal impacts of scientific research

The RCR framework is valuable not because it is an alternative regulatory system that improves upon the human subjects protections system (though it is, and it does), but because it takes an expansive view of what should be covered in ethics education. For example, the issues raised by the ethical controversies discussed in Section 2 above would be addressed under RCR topic categories (c) mentor/mentee relations, (d) workplace harassment, and (k) scientists’ societal responsibilities. The kinds of questioning taking place as I write this about the role of faculty and students in rejecting their university employers’ entanglements with other institutions that are argued to be causing harm (i.e., calls for divestment) would likewise be discussed under the category of the scientist as a responsible member of society (k).

While the U.S. RCR policy imposes no particular requirements for formal RCR instruction, the associated guidance documents do provide recommendations for best practices.^[11] The primary educational format should involve substantial face-to-face discussions. These can be supplemented by video-based formats so long as they are “utilized in a way that fosters discussion, active learning, engagement, and interaction among the participants” (NOT-OD-22-055). Online courses are explicitly rejected as an adequate method of RCR education (NOT-OD-10-019). The participants in formal ethics education should include not only students or “trainees”; mentors should also contribute, whether as presenters, discussion leaders, or course organizers. (As many RCR experts have observed, the involvement of faculty members or other senior professionals signals that RCR is a core community value; see, e.g., Anderson 2016; Bird 2012.) In order to support the consolidation of learning, RCR education activities should be distributed over an extended period of time like a semester or year, as opposed to a single intensive time-compressed event, and they should involve at least eight contact hours. Structured ethical reflection on RCR topics should ideally occur at every stage of education and career, from undergraduate research assistants all the way up to the faculty level.

In an insightful programmatic article that I urge anyone interested in ethics education to read in its entirety, Michael Kalichman argues that “the primary goal of RCR education [should be] to foster a research culture in which conversations about responsible conduct of research are expected and acceptable” (2014: 69). Focusing on building an ethically engaged community implies a very different educational approach than the compliance-oriented “training” discussed earlier. While there are certainly rules to be learned (not only about human subjects but also about animal care and use, data management, and conflicts of interest), the ethical issues that arise in the course of a linguist’s professional life will rarely be straightforward breaches of regulations and policies; they will almost always call for navigation of ambiguity, uncertainty, competing pressures, and divergent points of view. Ethical sensitivity and alertness can keep linguists from stumbling into “questionable research practices” (QRPs), ill-advised but not always premeditated choices (such as over-budgeting on grant proposals, cherry-picking data, retrofitting hypotheses to suit the results, excessive self-citation, etc.) that threaten the ethical integrity of the research process at every stage, even if such practices do not rise to the status of outright misconduct (Plonsky et al. 2024).^[12]

The topical coverage of any RCR education curriculum will necessarily be incomplete; after all, research trajectories and careers often take unexpected turns, and even when they don’t, it is impossible to anticipate the kinds of situations that any one of us will ultimately confront. There is a reason we so often marvel to ourselves that “life is stranger than fiction”! Kalichman makes the point that the real entity subject to positive transformation through successful RCR education is in any case not the individual participants or “trainees”, but rather the wider community to which they belong. That community includes everyone affected by research along a spectrum of roles from students to society at large, and it will include areas of professional life that extend beyond the narrow practice of research: How often, and in what kinds of settings, should mentors meet with their mentees? Which factors are appropriate to consider in determining the order of authors on a publication deriving from a joint project? For how long is it justifiable to embargo one’s dissertation data? How can workplace diversity be created, supported, and sustained? An advisory report on RCR education recently issued by the Association for Practicing and Professional Ethics (APPE) advances a perspective in very much this same vein (APPE 2024).

5 Building a culture of ethical engagement at home

The APPE report just mentioned offers detailed strategies for furthering the interests of four major stakeholders in the research ethics ecosystem: funders, research institutions, accreditation bodies, and professional associations. One of the report’s key recommendations for research institutions is that they “provide and encourage participation in appropriate, tailored, and engaging RCR training for all who are engaged in research” (APPE 2024: 6). The tailoring they seem to have in mind involves differentiation according to disciplines. They call for “abandon[ing] the ‘silver bullet’ myth […] that one type of intervention serves all fields equally” and instead encourage “the development of discipline-specific topics and education experiences”, welcoming the assistance of professional societies in doing this (APPE 2024: 6). Specialists in ethics education are broadly in agreement that RCR programming should vary by discipline and address issues “identified by experts within the fields themselves” (Bulger and Heitman 2007: 877; see also Field et al. 2024). But while I completely agree that any generalized RCR education program is likely to be insufficiently tailored, whether for linguistics or for any other field, a discipline-specific ethics education program does not improve on this as much as it might seem at first blush. After all, there are simply too many ways of “being a linguist” (or a sociologist, an environmental scientist, or whatever) to be confident that even a discipline-specific program will target the relevant issues for everyone who participates.

Nevertheless, to the extent that linguists are socialized into their professional identities in their home graduate departments/schools/programs, a degree of discipline-specificity will follow naturally if that is where ethics education programming is developed and sited. There are certainly other settings that contribute to junior researchers’ socialization into their fields of study, including individual courses, research labs, professional meetings, social media sites, and listservs. But for purposes of identity formation and professionalization, the departments where we receive our graduate education are comparable to our families of origin: even in the absence of active teaching, our home departments provide one of the earliest professional settings into which we are actively incorporated; it is in them that we learn how to interact with others in our capacity as emerging linguistic researchers and practitioners. We learn which of our seniors and peers to admire, what kind of questions are legitimate to pose, what professional success and failure look like, and how conflicts are handled (and hopefully how they are resolved). Departments constitute a formative social space where the values of a disciplinary culture are acquired.

6 Conclusions

Linguists are special, but they are not that special. Like researchers and practitioners in all fields, as students go through their graduate education in their home departments they become socialized to take certain kinds of problems seriously and to approach them and reason about them in certain ways. As part of this socialization, they acquire habits of speaking and writing, arguing, teaching, organizing themselves, promoting their work, dealing with conflict, and so on, that are never intentionally “taught”, yet somehow end up being learned. One of the habits they can learn in this way is to bring an ethical alertness to any of the myriad things they will end up doing and witnessing in their lives as linguists, including things that are not directly tied to the conduct of research. There is no better way to cultivate this habit than to give linguistics students “the tools, resources, and motivation to engage in regular, open conversations” with their seniors and peers about the ethical issues that touch them (Kalichman 2023: 206; Kalichman et al. 2022). The RCR workshop offers a model for building ethical reflection into the social fabric of the departments where our professional identities begin to form.