Lesson 03: Writing Conventions¶
Physics writing has a dialect. It is not literary prose, and it is not the informal style of a lab notebook. It is precise, first-person, active, and built around a logical skeleton that the reader can skim. This lesson covers the conventions that govern every sentence and paragraph in a HEP paper.
Voice and person¶
Always use first-person plural "we" throughout the paper. Never use passive voice for key results.
| ✅ Write this | ❌ Not this |
|---|---|
| "We show that S4714 is sensitive to..." | "It is shown that S4714 is sensitive to..." |
| "We compute the total event rate." | "The total event rate is computed." |
| "We find σχp ≲ 10⁻³³ cm²." | "A bound of σχp ≲ 10⁻³³ cm² is obtained." |
Lead the first sentence of each paragraph with the subject and an active verb: "We compute...", "We find...", "We demonstrate...", "We propose...", "Figure X shows...", "This approach..."
State conclusions at the right confidence level. Use "we demonstrate" or "we show" for established results. Use "we find" for numerical outputs. Use "in principle" and "could" for genuine possibilities that have not been calculated.
Never start a sentence with:
- "It is important to note that..."
- "It should be noted that..."
- "One can see that..."
- "It is worth mentioning that..."
These phrases add words without adding information. Cut them and start with the actual claim.
Paragraph structure¶
Each paragraph has three parts: an opening sentence that states the main point, middle sentences that support it, and a closing sentence that concludes or bridges forward.
Length: 3–6 sentences. Hard limit: 8 lines in two-column format.
The opening sentence is a claim, not a setup. A reader skimming only the first sentence of every paragraph should understand the logical skeleton of the paper.
| ✅ Write this | ❌ Not this |
|---|---|
| "Atmospheric neutrinos constitute a critical signal." | "In this section, we discuss the atmospheric neutrino flux." |
| "The interaction length constraint follows immediately from the AGN size." | "We now turn to the interaction length." |
| "The UHECR survival requirement sets a direct bound on the DM–nucleon cross section." | "Having established the physical setup, we now consider the constraint." |
| "Figure~\ref{fig:main} shows that the bound scales linearly with DM mass." | "We now present the main result of this paper." |
No orphan sentences. Every sentence must connect causally or logically to the adjacent one. Test: could you swap any two sentences without breaking meaning? If yes, the paragraph lacks cohesion.
Transitions are load-bearing. Use them precisely:
| Purpose | Words to use |
|---|---|
| Contrast | "however", "on the other hand", "while", "by contrast" |
| Elaboration | "specifically", "in particular", "for example" |
| Consequence | "therefore", "as a result", "this implies" |
| Addition | "additionally", "furthermore" (use sparingly) |
| Sequential | "first...then", "subsequently", "finally" |
Avoid "Firstly,... Secondly,... Thirdly,...". Use causal or contrast flow instead.
The opening sentence must not contain a citation. Citations belong in supporting sentences, not in claims.
| ✅ Write this | ❌ Not this |
|---|---|
| "The UHECR spectrum cuts off sharply above 50 EeV. This behaviour was first measured by [ref]." | "The UHECR spectrum [ref] cuts off sharply above 50 EeV." |
| "Solar neutrino detection at DUNE can proceed via neutral-current interactions on argon. This channel was proposed in Refs. [ref1, ref2]." | "Solar neutrino detection [ref1, ref2] can proceed via neutral-current interactions on argon." |
Standard paragraph forms¶
Three patterns cover most paragraphs in a HEP paper.
Claim → Evidence → Consequence (most common in body sections)
Opening sentence states the main point directly. Middle sentences provide the supporting calculation, observation, or prior result. The closing sentence states the consequence or bridges forward to the next paragraph.
Example (from arXiv:2605.16721, adapted):
"Atmospheric neutrinos constitute a critical signal for measuring the CP-violating phase. A precise understanding of the low-energy atmospheric neutrino flux is crucial for multiple physics goals: probing neutrino oscillations, mass ordering, and flavor ratios, and it acts as a background in the search for the DSNB, proton decay, and indirect dark matter searches. We demonstrate in Section III that the subdominant flux uncertainty is the limiting systematic at next-generation detectors."
Problem → Technique → Scope (used at section and subsection openings)
Opening sentence poses the question this section answers. Middle sentence names the method or criterion used. Closing sentence states the scope constraint or conservative assumption.
Example (from arXiv:2512.18093):
"Assuming that UHECRs near the GZK cut-off (E_CR ~ 50 EeV) are all accelerated within AGNs, we can set a basic requirement on the interaction length for CR–DM interactions. The interaction length λ_CR,χ needs to be larger than the acceleration distance l_acc of the CR. Conservatively, we can set l_acc to the size of the AGN."
Result → Refinement → Forward pointer (used in multi-subsection derivations)
First sentence states the result just derived. Middle sentence identifies the leading approximation that can be improved. Closing sentence points to the next subsection.
Example (from arXiv:2512.18093):
"We can refine the estimate in Equation 6 by observing that if a particle escapes its acceleration region, it no longer accumulates energy."
Section titles¶
Section titles are short noun phrases, not full sentences, not questions, and not gerunds.
| ✅ Write this | ❌ Not this |
|---|---|
| "Flux Measurement" | "Measuring the Flux" |
| "DM Densities" | "The Dark Matter Density Profile" |
| "Conclusions and Outlook" | "Summary and Conclusions" |
| "Measurement Potential" | "The Measurement Potential of Hyper-K" |
The final section is always "Conclusion" (singular) or "Conclusions and Outlook". Never "Summary and Conclusions".
Order-of-magnitude estimates¶
Every paper should contain at least one clearly signposted order-of-magnitude (OOM) estimate, an argument a reader can reproduce with pencil and paper in under five minutes. Its role is to build intuition for why the answer comes out the way it does, not to replace the full calculation.
Where to place it:
- In the introduction or at the end of Section 1, for the paper's main result.
- At the opening of each Results subsection, as a scaling argument before the full calculation.
- After a surprising result, to explain which physical effect bridges the gap from naive expectation.
Set it off explicitly: "To build intuition, we note that..." or "A simple estimate gives..."
Structure of a well-formed estimate:
- Name the one or two quantities that dominate the answer.
- Write the scaling: a product or ratio of those quantities, using ~ freely.
- Plug in numbers and state the result (one significant figure).
Example (from arXiv:2512.18093):
"Assuming that UHECRs near the GZK cut-off (E_CR ~ 50 EeV) are accelerated inside AGN, a basic constraint on the DM–nucleon cross section follows immediately. The CR mean free path must exceed the AGN size, l_acc ~ 10¹⁶ cm, giving λ_CR,χ = (nχ σχp)⁻¹ ≳ l_acc. With the local DM density ρχ ~ 10⁻²⁴ g cm⁻³ and mχ ~ 1 GeV, this requires σχp ≲ 10⁻³⁰ cm², already three orders of magnitude below existing direct-detection limits."
Always close the estimate paragraph with a sentence comparing it to the full result:
| ✅ Write this | ❌ Not this |
|---|---|
| "The full calculation (Section III) gives σ ≃ 2.3 × 10⁻³³ cm², consistent with this estimate." | (Present only the full result with no estimate to orient the reader.) |
| "The full result is a factor of ~3 larger because of the spike profile, discussed in Section IV." |
Notation for estimates: Use ~ for order-of-magnitude equalities, ≃ for approximate numerical equality (within 10–20%). Do not use ≈ for rough estimates. Reserve it for numerical approximations in equations.
Before and after: a poorly-written versus a well-formed estimate¶
❌ Poor version: no controlling quantities named, no algebra shown, no comparison to the full result:
"The DM–nucleon cross section is constrained to be small. The mean free path of a UHECR must be large enough for it to escape the AGN. This gives a rough bound on the cross section."
This paragraph says nothing a reader could reproduce. It has no numbers, no scaling, and no link to the full result. A reader cannot check it and gains no physical intuition from it.
✅ Well-formed version (from arXiv:2512.18093):
"To build intuition, we note that the UHECR mean free path must exceed the AGN acceleration region, $l_\mathrm{acc} \sim 10^{16}$~cm. The interaction length is $\lambda_{\mathrm{CR},\chi} = (n_\chi \sigma_{\chi p})^{-1}$, where $n_\chi = \rho_\chi / m_\chi \sim 10^{-3}$~cm$^{-3}$ for a local DM density $\rho_\chi \sim 10^{-24}$~g~cm$^{-3}$ and $m_\chi \sim 1$~GeV. The requirement $\lambda_{\mathrm{CR},\chi} \gtrsim l_\mathrm{acc}$ then gives $\sigma_{\chi p} \lesssim 10^{-30}$~cm², already three orders of magnitude below current direct-detection limits. The full numerical result in Section~\ref{sec:results} tightens this to $\sigma_{\chi p} \lesssim 10^{-33}$~cm² once the spike profile is included."
Three steps visible: (1) controlling quantity named ($l_\mathrm{acc}$), (2) scaling written out with numbers, (3) estimate compared to the full result. A reader can reproduce this in two minutes.
Figures in text¶
Always introduce a figure before or at the start of the paragraph that interprets it.
| ✅ Write this | ❌ Not this |
|---|---|
| "Figure 3 shows the predicted event rate as a function of energy." | "As can be seen in Figure 3, the event rate..." |
| "The results are presented in Figure 3." | "The event rate (Figure 3) is..." (when Figure 3 hasn't been referenced yet) |
Always spell out "Figure", never "Fig.". See Lesson 04 for the LaTeX conventions.
Lists¶
In the main text, use in-line enumeration rather than bullet points: "We compute (i) the total energy exchange, (ii) the orbital modulation, and (iii) the sensitivity to the spike profile."
Bulleted lists (itemize) belong in appendices or for genuinely unordered parallel items such as a list of assumptions.
Numbered lists (enumerate) are appropriate only when sequence or hierarchy matters, and should contain 3–5 items at most.
If you find yourself using bullet points in a Results section, convert them to prose.
Common phrases: use and avoid¶
The left column is the standard phrasing in this group's papers. The right column is the version that will be flagged in review, sometimes by a co-author, sometimes by a referee, and sometimes by an automated checker.
| ✅ Use | ❌ Avoid | Why |
|---|---|---|
| "We show that..." | "It is shown that..." | Passive hides agency; "we" is shorter and clearer |
| "Figure~\ref{fig:X} shows..." | "As can be seen in Figure~\ref{fig:X}..." | The figure does the showing; the reader does not need to be told they can see it |
| "We have computed..." (conclusion) | "We have tried to compute..." | "Tried" implies failure; state what was done |
| "In this work, we..." | "In this paper, the authors..." | You are one of the authors; use "we" |
| "as we discuss below" | "as will be discussed later" | Forward reference should be specific and active |
| "We are grateful for helpful discussions with..." | "The authors would like to thank..." | Direct and in first person; "would like to" is hedged |
| "we find" (numerical output) | "it turns out that" | "It turns out" is conversational and implies surprise; "we find" is the standard result verb |
| "we adopt" (model or parameter choice) | "we assume without loss of generality" | WLOG is almost never literally true in a phenomenology paper |
| "in principle" | "theoretically speaking" | "In principle" has a precise meaning (allowed by the laws of physics but not yet achieved); use it accurately |
| "We compute the sensitivity to..." | "We attempt to constrain..." | State what was done, not what was attempted |
| "This bound is three orders of magnitude below..." | "This is a significant improvement over..." | Quantify the improvement; "significant" is vague |
Further reading¶
George Gopen and Judith Swan's "The Science of Scientific Writing" (American Scientist, 1990) is a two-page paper that explains, from a linguistic perspective, why readers struggle with scientific prose and what structural choices cause those difficulties. It is not a style guide for physics specifically, but the underlying theory of why paragraph structure matters is worth reading alongside this lesson.
What to read next¶
Lesson 04 covers the LaTeX conventions that implement these writing rules in practice: figure environments, equation formatting, cross-references, and citations.