text-refinement

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:

1. Clarity - Every sentence should be immediately understandable
2. Conciseness - Remove unnecessary words without losing meaning
3. Flow - Sentences and paragraphs connect smoothly
4. Coherence - Ideas build logically
5. Precision - Word choice is exact and appropriate

---

Refinement Process

Step 1: Structural Review

Before word-level edits, check:

• Does each paragraph have a clear topic sentence?
• Is information organized logically?
• Do paragraphs connect to each other?
• Is the overall argument/narrative clear?

Step 2: Sentence-Level Editing

For each sentence, ask:

• Is this necessary?
• Is it as short as it can be?
• Is the meaning clear on first read?
• Does it flow from the previous sentence?

Step 3: Word-Level Polishing

For key passages:

• Are words precise?
• Are there unnecessary qualifiers?
• Is jargon appropriate for the audience?
• Are there better word choices?

---

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:

1. Connect to the previous paragraph
2. Present one main idea
3. 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

• Every sentence has a clear subject and action
• Technical terms are defined
• Ambiguous pronouns are resolved

Conciseness

• Unnecessary words removed
• Redundancies eliminated
• Nominalizations converted to verbs

Flow

• Old-to-new information pattern used
• Transitions connect paragraphs
• Section narrative is clear

Coherence

• Parallel structure maintained
• Consistent terminology
• Logical progression of ideas

Precision

• Specific words replace vague ones
• Technical terms used correctly
• Numbers and data are accurate

Tone

• Appropriate formality level
• Hedging calibrated correctly
• Colloquialisms removed

---

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"