Paper Draft Feedback
Here are some general tips after reading through ~1/2 of the paper drafts:
Cite more densely and use more diverse citations.
Higher âcitation densityâ means more inline citations per paragraph. âMore diverseâ means that multiple core arguments should not, in general, rest on a single citation.
Inline citations in BibTeX format are expected for the final submission.
Avoid phrases that indicate vague relationships between concepts.
Examples include:
- âX highlights Yâ (not always bad in the drafts Iâve read)
- âX brings into question Yâ (pretty consistently bad)
- âX raises ethical concernsâ/âX raises ethical questionsâ (pretty consistently bad, and rather uninformative. It is expected that the topic of your paper will raise ethical concerns.)
Be more precise in describing relationships.
Talking precisely about the relationships between concepts will improve your technical writing in general. This precision can be improved by expanding your repertoire of words and phrases for describing relationships.
This is not just a stylistic comment â I am not encouraging variety for the sake of variety. Different conjunctions communicate different relationships.
Most essays make use of common conjunctions like âandâ, âbutâ, and âhoweverâ. I think âhoweverâ, especially, has been overused. Consider alternatives like âat the same timeâ, âin contrastâ, âparadoxicallyâ,âunexpectedlyâ, âcounterintuitivelyâ,âsurprisinglyâ,âdespite thisâ, or âneverthelessâ.
"X. Paradoxically, Y."
means that X and Y are both observed, but appear to be in logical conflict with each other.
"X. Unexpectedly, Y."
means that X and Y are both observed, and one would not expect Y knowing X, but X and Y may not conflict logically.
"X. At the same time, Y."
This is even more relaxed. There may be some tension between X and Y, or Y may be tangentially related to X.
Avoid exagerration, overuse of adjectives, and unnecessarily strong claims.
When you notice yourself using superlatives (âthe mostâ, âthe bestâ, âŚ) or making universal claims (âevery userâ, âall aspects of user dataâ), ask yourself whether you really need to make such a strong claim. In many cases, weaker claims will have the same effect on your overall argument.
Stronger claims increase your exposure to reader skepticism! Relax your minor claims to the minimum strength necessary to argue your thesis.
(Note that this is not the same advice as ârelax your thesisâ. In many cases I would prefer that you sharpen your thesis!).
An analogy: Think of yourself like a secret agent trying to hit a target (reader belief change) inside enemy territory (reader skepticism).
- I am NOT suggesting that you argue for something non-controversial. This would equate to aborting the mission entirely.
- I am asking you to try to change the readerâs mind using relatively non-controversial subarguments and minor claims (i.e. hit the target without making a lot of noise on the way there).
Every time you exaggerate a minor claim, the reader hears alarm bells. They react to this by becoming MORE vigilant and skeptical, making the rest of the mission harder.
Another analogy: For the mathematicians in the class, simpler proofs are often preferred to more complicated ones. Proofs are often considered beautiful when a suprising or useful result can be derived from a minimal set of axioms, in relatively few derivation steps. Minimality, simplicity, parsimony, and elegance are words used to refer to this aesthetic value.
"This ensures that critical situations, anywhere and at any time, are identified and addressed..."
Is is necessary that situations can be addressed at any time?
"Every action a user takes feeds into the system the algorithm uses."
Is it critical to your argument that EVERY argument feeds into the system? Would your overall argument be the same if it were only âthe majorityâ or âsomeâ of user actions?
"This grave danger raises serious concerns about the role of big technology..."
The reader may not agree that the danger is grave or that the concern is serious. If reader agreement on the severity is not important, donât invite the reader to disagree with you on this nonessential point.
âSignpostâ more frequently
Signposting refers to writing practices that help the reader keep track of the structure of your essay.
Re-outline and re-organize
You may or may not have written an outline before writing this draft. If you did, you probably notice that the draft has diverged from the structure you originally planned to use. This is a good sign! It means that the writing process has helped you think (one of the main benefits of writing outside of school). At the same time, your paper may have lost some cohesiveness in the process. Top-down reorganization can help fix this.
Re-outline and re-organize means:
- Lock up your old outline where you canât see it
- Read the draft that you wrote
- Write an outline based on the draft that you wrote
- Assess the new outline for logical flow
- Reorganize the outline for improved logical flow
- Reorganize the draft based on the outline
If you donât like outlining, you can use another high-level organizational tool like a mindmap or flowchart.
Make good use of your fresh eyes
âFresh eyesâ are the eyes that you read with after not having looked at your draft for a while. Fresh eyes make many problems in your writing obvious.
If you write for four hours spread evenly over 4 days, you get fresh eyes 4 times.
If you write in one long block of 4 hours, you get fresh eyes once.
Like any technique, there are tradeoffs. A long stretch of writing time can be useful for getting to clarity on a tricky, complicated part of your project.
Watch this video
Key takeaways:
- In any ambitious intellectual project, writing is a thinking tool, not just a communication tool.
- In your first draft, you will write in a way that is helpful for your own thinking.
- The structure that aids your own thinking will not be the structure that aids your readerâs understanding.
- In school, your writing was read by teachers who were paid to care. Your teacherâs objective was to grade you for what you knew, on a topic area that they already knew.
- Outside of school, your writing will be read (or not) by people who have limited attention and are not paid to care. Your readerâs objectives are to figure out if your writing has any merit, learn from it (if it has merit), and stop wasting time (if it doesnât).
- Value comes from changing the beliefs of a community of readers. Writers can signal value by establishing tension â getting the reader to feel that they (and the rest of the community) have missed something.
- More words != more clarity. If you get feedback that your writing is not clear, or that your reader does not understand, you will react instinctively by adding words to explain yourself further. If your 500 words have been unclear, 1000 words will be twice as many words at the same clarity level. Improving clarity depends on a change in strategy, not more words with the same strategy.
Feedback Key
| code | description |
|---|---|
| 1 | Too broad/vague. Be more precise. |
| 2 | Repetitive |
| 3 | Non-obvious/not enough justification/non-sequitur |
| C | needs citation |
| 4 | unnecessarily strong claim/description |
| strikethrough | delete this |
| 5 | too verbose |
| 6 | what is the implication of this? Unclear significance, take it further. |
| 7 | good content |
| 8 | good style |
| 9 | thesis non-obvious |
| 10 | Find a stronger source or diversify sources |
| 11 | What about the inverse*? Consider counterfactuals. |
| 12 | Why this lens? Justify your some-from-many choice. |
| 13 | Use a different word/ Rephrase this |
| 14 | Obvious/tautological |
| 15 | Define this |
| 16 | Missing basis/denominator |
| 17 | Needs stronger thesis |
| 18 | Needs reorganization for logical flow |
| 19 | Needs signposting |
| 20 | Address counterarguments |
| 21 | Too specific for this location |
| 22 | Too general for this location |
| 23 | Discuss sources |
| 24 | Unclear relevance |
| P | insert paragraph break here |