Error Spotting – Advanced Study Notes (English Language)
Audience: University‐level students (MIT‑style) in English Linguistics or Applied Linguistics
Goal: Master the identification and correction of grammatical, syntactic, lexical, and stylistic errors in written English, using formal linguistic theory and empirical evidence.
1. Introduction
Error spotting is the systematic identification of linguistic mistakes in a text. It is a key skill in editing, teaching, and natural language processing (NLP). The discipline blends:
| Domain | Focus | Example |
|---|---|---|
| Grammar | Parts of speech, agreement, tense, voice | “He went to the store” vs. “He go to the store” |
| Syntax | Constituent structure, dependency relations | “The dog that chased the cat was loud” |
| Lexical Choice | Collocation, register, cognates | “The bibliography of the book” (should be bibliography of the author) |
| Stylistics | Cohesion, coherence, rhetorical devices | Repetition, parallelism, parallel structure |
2. First Principles
-
Universal Grammar (UG)
- Chomsky’s hypothesis: All human languages share a set of abstract syntactic principles.
- Error spotting relies on detecting violations of UG constraints (e.g., Subjacency, Island Constraints).
-
Minimalist Program
- Merge and Move operations build syntactic trees.
- Errors often involve failed Merge (missing elements) or illegal Move (exceeding locality constraints).
-
Lexical Functional Grammar (LFG)
- Separates constituent structure (c‑structure) and functional structure (f‑structure).
- Helps spot mismatches between syntactic roles and functions (e.g., subject–verb agreement).
-
Construction Grammar
- Treats idioms and fixed patterns as first‑class syntactic objects.
- Errors arise when a construction is misapplied (e.g., a good day vs. a good day).
-
Corpus Linguistics
- Empirical frequencies guide “typical” vs. “atypical” usage.
- Error spotting becomes data‑driven: compare against large corpora.
3. Systematic Error Categories
| Category | Typical Errors | Linguistic Insight | Detection Technique |
|---|---|---|---|
| Morphological | Wrong inflection, missing plural markers | Agreement constraints (Subject–Verb) | Spell‑check + POS tags |
| Syntactic | Misplaced modifiers, dangling participles, broken parallelism | Constituent order, movement restrictions | Dependency parsing, tree‑based validation |
| Lexical | Wrong word choice, false friends, collocation violations | Lexical semantics, polysemy | Word sense disambiguation, collocation databases |
| Pragmatic | Tone mismatch, inappropriate register | Contextual appropriateness | Discourse analysis, pragmatic markers |
| Stylistic | Redundancy, vague pronoun reference | Cohesion & coherence | Coreference resolution, readability metrics |
4. Advanced Theoretical Tools
4.1. Feature Geometry
- Agree: Features such as [Person], [Number], [Gender] must be matched.
- Error: She go to school (number mismatch).
- Correction: She goes to school.
4.2. Constraint Satisfaction
- Minimality: Closest agreement target.
- Island Constraints: Prevent extraction from certain domains.
- Error: *Whose picture did you see the picture of? (extraction from a relative clause).
4.3. Semantic Role Labeling (SRL)
- Detect mismatches between verb arguments and their roles.
- Example: She gave the book to me vs. She gave the book to me (correct); She gave the book to me (ambiguous if me is an object or a location).
4.4. Discourse Representation Theory (DRT)
- Handles anaphoric relations and discourse coherence.
- Error: John said he would arrive. He left early. (unresolved he).
5. Edge Cases & Common Pitfalls
| Edge Case | Explanation | Typical Mistake | Correction |
|---|---|---|---|
| Zero‑Subjunctive | In some dialects, verbs lack explicit subjects. | Went to the store (missing he/she). | Add subject or use context‑appropriate ellipsis. |
| Mixed Tense | Switching tenses inconsistently within a clause. | She runs and will run. | Align tenses: She runs and runs. |
| Non‑Finite Clause Coordination | Coordinating two infinitives without a shared auxiliary. | To eat and to sleep (unnecessary repetition). | To eat and sleep |
| Dual‑Headed Conjunctions | Conjunctions like both…and require agreement on both conjuncts. | Both the students and the teacher is | Both the students and the teacher are |
| Back‑shift Errors | Wrong tense shift in reported speech. | He said I am tired. (should be I was tired). | Correct back‑shift: He said I was tired. |
6. Practical Detection Pipeline
- Tokenization & POS Tagging
- Use spaCy or NLTK to annotate parts of speech.
- Dependency Parsing
- Verify grammatical relations (subject, object, modifier).
- Feature Matching
- Check number, person, gender agreement.
- Lexical Validation
- Cross‑reference with WordNet or Wiktionary.
- Pragmatic Check
- Coreference resolution (AllenNLP) to detect ambiguous pronouns.
- Stylistic Analysis
- Readability scores (Flesch‑Kincaid), redundancy detection.
7. Challenging Problem (MIT‑Level)
Problem Statement
Given the following paragraph, identify all linguistic errors and provide corrected versions. Explain the theoretical reason for each error and why the proposed correction satisfies the constraint.
Paragraph
The committee of the university, which include the ten professors and the dean, has decided that the new research grant will be distributed among the faculty. However, it has not yet been approved by the