C2Text analysisStylisticLayer 1 (Deterministic)

Verbosity

Detects unnecessary wordiness and padding phrases that inflate text length without adding substance, a common trait of AI-generated academic writing.

Technical description

Identifies filler phrases, redundant modifiers, and padding expressions using curated dictionaries. Measures word-to-content ratio by counting substantive content words versus discourse markers, intensifiers, and filler. Computes padding density (filler_count / sentence_count) and flags sentences exceeding a word count threshold with low information density.

How it works

Layer 1 (deterministic): Matches text against a dictionary of ~150 padding phrases ('it is important to note that', 'in the context of', 'a wide range of'). Counts redundant modifiers and pleonasms. Calculates padding density per paragraph. Flags sentences with more filler words than content words.

Why this matters

Language models tend to produce verbose text because they are trained to generate plausible continuations, leading to repetitive elaboration and padding. Human academic writers, constrained by journal word limits, typically write more concisely. Excessive verbosity is a reliable surface-level signal of machine generation.

Score thresholds

0-1: Concise writing with minimal padding
2-3: Moderate use of filler phrases
4-5: Heavily padded with redundant expressions throughout

Limitations

Some academic writing traditions (especially in social sciences) use more elaborate phrasing naturally. Non-native English speakers may use padding phrases as discourse connectors. Grant proposals often contain formulaic language that may trigger false positives.