The Underrated Role of Understanding Language

Rules for policy writing, like the training and resources offered by the University of North Carolina at Chapel Hill’s (UNC-Chapel Hill) Policy Office, are helpful tools to improve the overall accessibility and utility of policies. However, as is true in almost all fields, rules have their limits. It never makes sense to apply the same rules to every piece of policy writing. Policy writers need to consider how language furthers their policy goals and institutional values. In order to understand when the “rules” are useful and when they should be ignored, policy writers and editors need to be familiar with why guidelines are given. Fortunately, this is exactly the type of question linguists study: how does human cognition interact with language, and how can that information be used? From this perspective, policy administrators can better examine and justify writing and editing decisions. This is illustrated by looking at a few examples of common policy-editing rules.

Rule: Remove Barrier Language

One of the most obvious changes in updated policy language is the removal of marginalizing or otherwise exclusionary words. This includes gendered terminology, non-preferred labels, or unnecessarily limited categories (e.g., outdated country names, normative descriptors). Generally, though not always, this rule of using inclusive language is conceptually understood – why be exclusive when you could be inclusive? Still, it can seem trivial for organizations to devote resources to combing through old policies, looking for violations of inclusivity rules and making tiny changes. The field of sociolinguistics provides a lot of evidence that this investment is actually not trivial at all.

For example, the use of gendered terminology triggers mental concepts of gender categories, making gendered stereotypes more accessible in the mind. This unconscious process has very real consequences on behavior. When masculine forms are used as “neutral” (e.g., “mankind”), it promotes stereotypes that male is the default, expected category – making those who do not identify as male feel less suited to the environment. A 2021 study of adults in Israel demonstrated that addressing women with masculine (neutral usage for Hebrew) pronouns in online math testing resulted in poorer performance, whereas feminine testing language reduced the gender achievement gap by one-third. The converse held for men, who performed worse when addressed in the feminine. Furthermore, both genders exhibited more effort (measured in time) when taking a test with language corresponding to their gender identity (Kricheli-Katz & Regev, 2021). The use of gendered language influenced the perception of the “prototypical test-taker,” making those of a gender not addressed directly in the test’s language feel alienated from the field of mathematics. The simple act of changing pronouns to be properly inclusive significantly improved test-takers’ attitudes and achievement.

In Sweden, a gender-neutral pronoun was officially incorporated into their language in 2015. This faced backlash, being criticized as a performative action of “political correctness” with little tangible impact (Tavits & Pérez, 2019). Yet experiments here again reveal that gender-neutral pronoun use weakens people’s bias favoring men, and that this reduced salience of masculinity promotes more equal attitudes towards women and members of the LGBTQ+ community. This was displayed in more positive attitudes toward female politicians and less hostility towards LGBTQ+ individuals, and more support for policies that benefit both groups (Tavitz & Pérez, 2019).

These results should be hugely important in the world of higher education policy and administration. The purported goal of education is to promote opportunity without discrimination. By this standard, it is problematic to use language in policies that makes certain groups or individuals feel alienated because this negatively impacts their academic performance and undermines their sense of belonging in the institutional setting. As such, removing barrier language is not about “following the rules” just because they exist, but about recognizing the very real impacts that language has on behavior and ensuring that the attitudes of an institution are represented correctly in policy. As language continuously evolves and preferred, maximally inclusive language changes, a review that is sensitive to the realities of how policy language impacts people is an essential tool.

Rule: Avoid Negative Statements

Looking at more technical elements of policy review guidelines, let’s consider the long-promoted practice of avoiding negative statements. Or, to state the rule more simply: no negative statements. Interestingly, this rule is clearer when stated in a way that violates the rule itself. So why is it such a common recommendation for clear writing?

Traditionally, proponents of avoiding negative statements in policy cite processing difficulties and assert that telling people what to do is more helpful than telling them what not to do. It’s not that these ideas are “wrong.” However, linguistic evidence reveals a more complicated picture than any rule could account for.

In some regards, the “no negative statements” rule has obvious applicability. If a policy intends to have employees submit paperwork to the Human Resources department, saying “submit paperwork to the Human Resources department” is more informative and useful than saying “do not submit paperwork to the Finance department.” A rule to avoid negative statements helps ensure actionable policy statements. Some statements, however, have equally informative positive and negative versions (when they refer to a binary). Still, negative statements have been found to be more cognitively demanding than positive statements (Agmon et al., 2022). This phenomenon is demonstrated in simple experiments measuring reaction time in verification tasks of statements like “the square is blue” and negated statements like “the square is not blue.” The delay of task completion for negative sentences can sometimes be attributed to processing cost (for example, some linguistic theorists posit that double-processing is necessary for negation: first processing a situation to then be able to process its negation). Negation also has a verification cost, which is an additional effort to determine the truth value of a negative sentence (Agmon et al., 2022). Another concern is that negation often increases structural complexity by requiring the addition of auxiliary verbs (e.g., in a sentence like “The student reads,” negation requires the addition of the auxiliary verb “do,” in the form “The student does not read”). Difficulties can also arise from a pragmatic perspective, since readers find negation to be strange if the specific context does not invoke it. In other words, if there is no expectation of some positive statement, it is hard for readers to determine the relevance of its negation (Nordmeyer & Frank, 2014). As such, policies that include negative statements carry a contextual burden that may be lessened by avoiding negative statements.

At the same time, there are cases where a negation is processed faster than an affirmative, which is particularly true of “real-world” language use compared to artificial research contexts (Orenes, 2021). Eye-tracking studies have demonstrated that participants quickly fixate on a negative phrase corresponding to a statement they heard without first fixating on the non-negative version. This is evidence for a more immediate integration of negation, contrary to the double-processing theory (Orenes, 2021). As it turns out, both slower and faster processing of negative versus affirmative statements can be true. In comprehending language, people are fast to represent an explicit negation but slower to process an affirmative alternate (Orenes, 2021). The policy implication here is that avoiding negatives is counterproductive if the goal of a policy statement is centrally to avoid some behavior. People are quite adept at integrating meaning for explicit negation. However, issues arise when understanding the explicit negative is insufficient for the given context, and readers have to mentally represent some affirmative that is not explicitly stated. Effective policy writers exhibit flexibility in applying the “no negation rule,” in accordance with principles of cognition.

Rule: Eliminate Passive Voice

There is a similar nuance to consider for eliminating passive voice. This rule has been taught in schools for ages. Passive voice is said to be the enemy of clarity. A passive sentence, such as “The book was read,” provides no insight into the sentence’s subject. This leaves the sentence vulnerable to multiple interpretations (one could assume, as we might in our office, that the book was read “by zombies”). Linguists grate against the idea of banning passive voice, however, because the simple fact that native speakers opt for passive voice when they speak and write indicates that it is a valuable practice. Passive voice is socially useful (e.g., to avoid placing blame), but it is also useful for topicalizing some part of a sentence to focus attention. For example, in the sentence, “Millions of people read the book,” the focal point is the subject (people). If I want the attention to be on the book, passive voice facilitates its topicalization: “The book was read by millions of people.”

In the world of policy, using passive voice to obscure a subject is resoundingly problematic. Policies are meant to be instructive, so it needs to be explicitly known who is meant to be doing what (rarely is it meant to be zombies). In this regard, the “eliminate passive voice” rule makes sense. Another good reason to eliminate passive voice in policy is that passive sentences increase the processing load for the reader. Passive sentences cause a delay in resolution; readers have to search for the subject and then connect it to the rest of the statement. Generally, the sentence “Sam ate an apple” is easier to process than “An apple was eaten by Sam.”

Still, there may be cases in policy where avoiding passive voice introduces more problems than it solves. Sometimes, avoiding passive voice becomes cumbersome, like when there are several potential actors, or if some object is the actual focal point of a statement. This is where it is essential to stay focused on the policy’s goals. A policy produced by student services might discuss a form that students need to submit, and then provide a timeline for request approval: “The form will be reviewed within 10 business days.” If the policy is designed to be student-facing, who reviews the form is not particularly important. If it is simple to indicate the subject (e.g., “The Vice Chancellor for Student Affairs will review the form within 10 business days”), that is probably best practice to promote accountability. However, the form may not have a single, definable reviewer. A student reading the policy will be more burdened by an explanation of all the potential reviewers than by the processing load of not being able to find a subject for the sentence. Only by understanding why “eliminate passive voice” is a rule can a policy writer be equipped to compare the difficulty introduced by passive voice to the difficulty introduced by avoiding it.

This sensitivity to how artificial rules for writing can contradict natural human language use is why a linguistic background is highly valuable in a policy environment. Policy review guidelines, like UNC-Chapel Hill’s, are a useful starting point, but policy writers will inevitably encounter situations where the rules become cumbersome. In these cases, understanding the cognitive basis for writing recommendations is essential to making the best policy language decisions.

References

Agmon, G., Loewenstein, Y., & Grodzinsky, Y. (2022). Negative sentences exhibit a sustained effect in delayed verification tasks. In Journal of Experimental Psychology: Learning, Memory, and Cognition (Vol. 48, pp. 122–141). American Psychological Association. https://doi.org/10.1037/xlm0001059

Committee on the Status of Women in Linguistics. (2003). Guidelines for inclusive language. Linguistic Society of America. https://www.lsadc.org/guidelines_for_inclusive_language

Kricheli-Katz, T., & Regev, T. (2021). The effect of language on performance: do gendered languages fail women in maths? Npj Science of Learning, 6(1). https://doi.org/10.1038/s41539-021-00087-7

Nordmeyer, A.E., & Frank, M.C. (2014). A pragmatic account of the processing of negative sentences. Cognitive Science, 36. https://langcog.stanford.edu/papers/NF-cogsci2014.pdf

Orenes, I. (2021). "Looking at" Negation: Faster Processing for Symbolic Rather Than Iconic Representations. Journal of psycholinguistic research, 50(6), 1417–1436. https://doi.org/10.1007/s10936-021-09797-w

Tavits, M., & Pérez, E. O. (2019). Language influences mass opinion toward gender and LGBT equality. Proceedings of the National Academy of Sciences, 116(34), 16781–16786. https://doi.org/10.1073/pnas.1908156116