brycewang-stanford Academic Tone Guide
A skill for adjusting the tone, register, and voice of academic writing to match disciplinary expectations and target venue conventions. Covers the spectrum from informal drafts to publication-ready prose, including hedging language, impersonal constructions, discipline-specific stylistic norms, and common tone pitfalls that undermine credibility.
Understanding Academic Register
The Register Spectrum
Academic writing occupies a specific position on the formality spectrum. The appropriate register depends on the genre (journal article, conference paper, thesis, grant proposal, blog post) and the discipline (humanities tend toward more elaborate prose; STEM fields favor concise, direct statements).
Register Levels in Academic Writing:
Level 1 - Informal (lab notebooks, internal emails):
"We tried the new method and it worked way better."
Level 2 - Semi-formal (conference talks, blog posts):
"We tested the proposed method and observed significant improvement."
Level 3 - Formal (journal articles, theses):
"The proposed method was evaluated against the baseline, yielding
a statistically significant improvement in accuracy (p < 0.01)."
Level 4 - Highly formal (legal briefs, policy documents):
"The aforementioned methodology was subjected to rigorous
evaluation, the results of which demonstrate a statistically
significant enhancement in predictive accuracy."
Most academic papers should aim for Level 3. Level 4 often sounds stilted and impedes readability. Level 2 is acceptable for workshop papers or invited commentaries but too casual for top-tier journals.
Discipline-Specific Norms
STEM Fields:
- Prefer active voice for methods: "We collected samples..."
- Use passive voice for established processes: "Samples were centrifuged..."
- Short sentences, minimal adjectives
- Precision over elegance
- Numbered equations and defined variables
Social Sciences:
- Mix of active and passive voice
- Moderate hedging: "The results suggest..." rather than "The results prove..."
- Theory-heavy introductions with clear operational definitions
- APA style typically required
Humanities:
- Complex sentence structures are acceptable
- Argumentative, first-person voice common in many fields
- Extensive engagement with prior scholarship expected
- Discipline-specific jargon must be deployed precisely
- Chicago or MLA style common
Hedging and Boosting
Hedging Language
Hedging is the use of cautious language to qualify claims. It signals intellectual honesty and awareness of limitations. Over-hedging weakens arguments; under-hedging invites criticism.
Hedging Devices:
Modal verbs (strength order):
might < could < may < can < should < would < will < must
"This approach might explain..." (weak claim)
"This approach may explain..." (moderate claim)
"This approach can explain..." (confident claim)
Hedging phrases:
"It appears that..."
"The evidence suggests..."
"One possible interpretation is..."
"To some extent..."
"Under certain conditions..."
Hedging quantifiers:
"Most participants..." (not "all")
"In many cases..." (not "always")
"A substantial proportion..." (not "everyone")
When to Hedge and When Not To
HEDGE when:
- Reporting your own results (findings suggest, data indicate)
- Making causal claims from correlational data
- Generalizing beyond your sample
- Interpreting ambiguous results
- Discussing implications and future directions
DO NOT HEDGE when:
- Stating established facts ("DNA is a double helix")
- Describing your own methodology ("We recruited 200 participants")
- Reporting objective measurements ("The temperature was 37 degrees C")
- Citing published findings ("Smith (2020) demonstrated that...")
Common Tone Problems and Fixes
Informal Language
Problem: Colloquialisms and contractions
Before: "We didn't find any significant results, which was kind of
surprising given that lots of previous studies had found
strong effects."
After: "No statistically significant effects were observed, a finding
that contrasts with the robust effects reported in prior
studies (Smith et al., 2019; Jones, 2020)."
Rules:
- No contractions (didn't -> did not, it's -> it is)
- No colloquialisms (kind of, a lot, pretty much)
- No rhetorical questions in formal papers
- Avoid first-person opinion statements without evidence
- Replace vague quantifiers with specific numbers
Overconfident Claims
Problem: Making claims stronger than the evidence supports
Before: "Our results prove that social media causes depression in
teenagers."
After: "Our results indicate a significant positive association
between social media usage and depressive symptoms among
adolescent participants, consistent with the hypothesis
that excessive social media engagement may contribute to
negative mental health outcomes."
Note: "Prove" is almost never appropriate in empirical research.
Use "demonstrate," "indicate," "suggest," or "support."
Emotional Language
Problem: Using emotionally charged words that bias the reader
Before: "The appalling lack of research on this devastating condition
is shocking and must be remedied immediately."
After: "Despite the substantial disease burden, the existing literature
on this condition remains limited. Further research is needed
to address critical knowledge gaps."
Authorial Voice and Person
First Person Usage
The question of whether to use "I" or "we" in academic writing varies by discipline and journal.
Active first person (increasingly accepted in STEM):
"We designed the experiment to test..."
"We propose a novel framework for..."
"In this paper, we argue that..."
Passive alternative:
"The experiment was designed to test..."
"A novel framework is proposed for..."
"This paper argues that..."
Guidance:
- Check your target journal's style guide
- "We" is standard in multi-author papers, even in passive-heavy fields
- Single-author "I" is common in humanities, less so in STEM
- Avoid "the authors" to refer to yourself (awkward in most contexts)
- Never use "you" to address the reader in formal papers
Practical Revision Checklist
Tone Audit Checklist:
1. Search for contractions -> expand all
2. Search for colloquialisms -> replace with formal equivalents
3. Check hedging on all claims -> calibrate to evidence strength
4. Verify consistent voice (active/passive) -> match journal norms
5. Remove emotional/judgmental adjectives -> use neutral descriptors
6. Check for rhetorical questions -> convert to declarative statements
7. Ensure technical terms are defined on first use
8. Verify that "significant" always means "statistically significant"
9. Remove filler phrases ("it is important to note that") -> state directly
10. Read abstract aloud -> does it sound like published papers in the venue?
This checklist can be applied during the final revision pass, after substantive content editing is complete. Tone adjustments should be among the last changes made, as earlier revisions frequently introduce new informal language that needs to be caught.