| name | text-refinement |
| description | Improve draft text for clarity, flow, coherence, and academic quality. Use when polishing sections, tightening prose, or enhancing readability while preserving meaning. |
| metadata | {"author":"core-team","version":"1.0.0"} |
Text Refinement
Your Task
When refining draft text, improve it for:
- Clarity - Every sentence should be immediately understandable
- Conciseness - Remove unnecessary words without losing meaning
- Flow - Sentences and paragraphs connect smoothly
- Coherence - Ideas build logically
- Precision - Word choice is exact and appropriate
Refinement Process
Step 1: Structural Review
Before word-level edits, check:
Step 2: Sentence-Level Editing
For each sentence, ask:
Step 3: Word-Level Polishing
For key passages:
Clarity Techniques
Cut Empty Phrases
| Wordy | Clear |
|---|
| in order to | to |
| due to the fact that | because |
| at this point in time | now |
| in the event that | if |
| for the purpose of | to, for |
| with regard to | about |
| in terms of | [usually remove] |
| it is important to note that | [remove, just state it] |
| the fact that | that (or remove) |
| it should be noted that | [remove] |
Eliminate Redundancy
| Redundant | Better |
|---|
| completely finished | finished |
| advance planning | planning |
| basic fundamentals | fundamentals |
| end result | result |
| past history | history |
| future plans | plans |
| consensus of opinion | consensus |
| unexpected surprise | surprise |
Prefer Active Voice
| Passive | Active |
|---|
| The data were analyzed by the team | The team analyzed the data |
| It was found that | We found that / Analysis revealed |
| The method was developed | We developed the method |
When passive is appropriate:
- Focus is on action, not actor: "The samples were collected at dawn"
- Actor is unknown: "The manuscript was written in the 12th century"
- Discipline conventions require it
Reduce Nominalizations
| Nominalized | Verbal |
|---|
| make a decision | decide |
| give consideration to | consider |
| conduct an investigation | investigate |
| make an adjustment | adjust |
| provide an explanation | explain |
Conciseness Techniques
Tighten Sentences
Before: "It is worth noting that the algorithm has the capability to process large amounts of data in a very efficient manner."
After: "The algorithm efficiently processes large datasets."
Combine Choppy Sentences
Before: "The study examined color perception. The participants were students. There were 50 participants. They came from three universities."
After: "The study examined color perception among 50 students from three universities."
Cut Filler Words
Common fillers to remove:
- really, very, quite, rather
- actually, basically, essentially
- clearly, obviously, certainly (unless necessary for hedging)
- simply, just, literally
Flow Techniques
Sentence-Level Flow
Use the old-to-new pattern: each sentence should begin with familiar information and end with new information.
Poor flow:
A new algorithm was developed. Efficiency was the main goal. Processing speed increased significantly.
Good flow:
We developed a new algorithm with efficiency as the primary goal. This focus on efficiency led to significant improvements in processing speed.
Paragraph-Level Flow
Each paragraph should:
- Connect to the previous paragraph
- Present one main idea
- Transition to the next paragraph
Transition Techniques:
- Repeat a key term from the previous paragraph
- Use transition words (However, Therefore, Furthermore)
- Ask a question that the paragraph answers
- Use demonstrative pronouns (This finding, These results)
Section-Level Flow
Sections should:
- Have clear internal logic
- Build on previous sections
- Preview what comes next
- Connect to the paper's main argument
Coherence Techniques
Maintain Consistent Focus
Within a paragraph, keep the same grammatical subject when possible:
Inconsistent:
The algorithm processes data efficiently. Large datasets can be handled. Minimal memory is required.
Consistent:
The algorithm processes data efficiently. It handles large datasets while requiring minimal memory.
Use Parallel Structure
Non-parallel:
The method is fast, reliable, and doesn't cost much.
Parallel:
The method is fast, reliable, and affordable.
Signal Relationships Explicitly
| Relationship | Signals |
|---|
| Addition | Furthermore, Moreover, Additionally |
| Contrast | However, Nevertheless, In contrast |
| Cause/Effect | Therefore, Consequently, As a result |
| Example | For instance, Specifically, To illustrate |
| Comparison | Similarly, Likewise, In the same way |
Precision Techniques
Choose Specific Words
| Vague | Specific |
|---|
| good | effective, efficient, precise |
| bad | inefficient, inaccurate, flawed |
| big | substantial, significant, extensive |
| thing | factor, component, element |
| stuff | materials, data, content |
| a lot | numerous, substantial, considerable |
Distinguish Similar Terms
Ensure you use the right word:
- affect (verb) vs. effect (noun/verb)
- compose vs. comprise
- fewer (countable) vs. less (uncountable)
- that (restrictive) vs. which (non-restrictive)
- who (subject) vs. whom (object)
Maintain Technical Precision
- Use terms consistently throughout
- Define terms on first use
- Don't use synonyms for technical terms (it causes confusion)
Academic Tone Calibration
Avoid Colloquialisms
| Colloquial | Academic |
|---|
| a lot of | numerous, substantial |
| get | obtain, acquire, receive |
| stuff | materials, factors, elements |
| okay | acceptable, satisfactory |
| pretty much | largely, substantially |
| kind of | somewhat, to some extent |
Calibrate Hedging
Over-hedged (too cautious):
It might be possible that the results could perhaps suggest that there may be a relationship.
Under-hedged (too confident):
The results prove that there is definitely a relationship.
Appropriately hedged:
The results suggest a relationship between the variables.
Match Formality to Context
| Context | Formality Level |
|---|
| Abstract | High |
| Introduction | High |
| Methods | High |
| Results | High |
| Discussion | Medium-High |
| Acknowledgements | Medium |
Before/After Examples
Example 1: Cutting Wordiness
Before:
In the context of the current study, we endeavored to investigate the manner in which the algorithm processes data inputs with respect to efficiency considerations.
After:
We investigated how efficiently the algorithm processes input data.
Example 2: Improving Flow
Before:
The method was tested. Good results were obtained. Some limitations exist. Future work will address these limitations.
After:
Testing yielded positive results, though some limitations emerged. Future work will address these issues.
Example 3: Enhancing Precision
Before:
The thing worked pretty well in most situations.
After:
The algorithm performed effectively across 85% of test scenarios.
Example 4: Fixing Passive Overuse
Before:
It was determined by the research team that the hypothesis was supported by the data that was collected.
After:
The research team found that the collected data supported the hypothesis.
Refinement Checklist
Clarity
Conciseness
Flow
Coherence
Precision
Tone
Quick Reference
| Problem | Fix |
|---|
| Too wordy | Cut empty phrases, combine sentences |
| Choppy flow | Add transitions, use old-to-new pattern |
| Vague | Replace with specific terms |
| Overly passive | Convert to active voice |
| Inconsistent | Maintain same subject/terminology |
| Over-hedged | Reduce qualifiers, be more direct |
| Under-hedged | Add "suggests," "may," "indicates" |