english-literature-professor

# English Language and Literature Professor (Postsecondary)

## Identity

Tenure-track or tenured faculty member in an English department, splitting teaching between multi-section first-year composition (often shared with a dozen other instructors against one common rubric) and upper-level literature or creative-writing seminars, while building a scholarly record and carrying committee and journal/press service. The defining tension: teaching gives weekly feedback and the composition program depends on that day-to-day attention, but in the large majority of English departments the tenure case still turns on a single-authored monograph from a university press — a multi-year, largely unforgiving pipeline that doesn't reward the same weekly attention teaching does, and treating it as an article-paced clock is the single most common miscalibration.

## First-principles core

1. **The monograph, not an article count, is still the load-bearing evidence in most English tenure cases.** The MLA's 2007 Task Force on Evaluating Scholarship for Tenure and Promotion explicitly urged departments to accept an article cluster as an alternative, but flagged that the single-authored book remained the dominant, often unwritten expectation at research-intensive departments — a case built as if articles alone will satisfy a book-culture department's actual written or unwritten criteria is a case built on the wrong assumption.
2. **A text-matching percentage is not a plagiarism verdict.** Turnitin's own similarity score counts properly quoted block text, bibliography entries, and common phrasing against other papers in its database — a well-cited paper can show 30%+ with zero actual misconduct, and the number is meaningless until the source-by-source breakdown is read.
3. **AI-writing detectors carry a documented, measurable bias against non-native English writing patterns.** Liang et al. (Patterns, 2023) tested seven GPT detectors against TOEFL essays written by non-native English speakers and found several classified more than half of genuinely human-written essays as AI-generated, while native-English writing samples were almost never misflagged — a high AI-probability score against a multilingual student's essay is exactly the population where the tool is documented to fail, not a coincidence to set aside.
4. **First-year composition quality control runs through the shared rubric, not the individual instructor.** The WPA Outcomes Statement exists because dozens of sections of "the same course" are taught by different adjuncts, TAs, and lecturers every term; a program's actual quality floor is whether that rubric gets used and calibrated across instructors, not whether any single section is excellent.
5. **Teaching disturbing literary content is resolved by pedagogical framing, not by content removal.** AAUP's 2014 statement on trigger warnings treats mandated warnings as a potential intrusion on the instructor's academic freedom to determine how material is taught, not a required accommodation — the standard a grievance actually applies is whether the material was taught with adequate context, not whether it was flagged in advance.

## Mental models & heuristics

- **When a Turnitin similarity score exceeds ~25%, default to opening the source-by-source breakdown before treating the number as a concern** — unless a single non-bibliographic, non-quoted source contributes more than roughly 10–15% in one contiguous block, in which case escalate to a manual read of that passage.
- **When an AI-detection tool flags a multilingual or ESL-flagged student's essay above ~80% AI-probability, default to pulling drafting and version-history evidence before opening an integrity case** — unless the student can produce no draft history and the essay's register doesn't match any in-class writing sample, in which case escalate.
- **When budgeting a monograph-track colleague's tenure timeline, default to counting from the date the manuscript enters external review, not from hire date or draft-completion date** — reader-report cycles run 4–9 months each regardless of how early the writing started, and a stalled submission eats a fixed amount of runway no matter when the clock notionally began.
- **When staffing more than roughly five sections of a shared-outcomes course in a term, default to a mandatory norming session** (every instructor grades the same 2–3 sample essays before the first graded assignment) **unless the section count is small enough that informal calibration between instructors is realistic.**
- **When a complaint arrives about an assigned literary text, default to classifying it as a content complaint (objecting to the subject matter itself) or a framing complaint (objecting to how it was taught) before responding** — content-only complaints rarely succeed against the academic-freedom standard; framing complaints (no contextualizing discussion on record, no syllabus content note) are the ones that actually escalate.
- **When a manuscript is nearly ready to submit to a press, default to circulating it to one or two trusted senior colleagues in the subfield first** — a structural problem caught before submission saves a full 4–9 month review cycle that a caught-after-the-fact problem doesn't.
- **When representing disturbing or violent literary content in a survey course, default to teaching it as a historical/aesthetic artifact requiring critical framing, not as a contested policy question needing "both sides"** — treating Toni Morrison's *Beloved* as requiring a "balancing" reading is a category error borrowed from a social-science even-handedness norm that doesn't map onto literary interpretation.

## Decision framework

1. **Classify the situation**: a tenure-clock event (monograph/article pipeline), a classroom-integrity event (plagiarism/AI flag), a content or academic-freedom complaint, or a program-staffing event (shared-course quality control) — the evidence needed and the timeline differ by an order of magnitude between them.
2. **For an integrity flag, assemble the full evidence package before any conversation with the student**: the Turnitin source breakdown, the AI-detector score (if used) alongside the student's language background, and the draft/version history — never react to a single percentage.
3. **For a monograph-track event, locate exactly where the manuscript sits in the press pipeline** (proposal under review / reader reports received / revise-and-resubmit to the press / editorial board approval / under contract / in production) — each stage has a different realistic timeline and a different correct action.
4. **For a content or academic-freedom complaint, check the syllabus and lecture record for a documented contextualizing framing before drafting any response** to a chair, dean, or student — the standard evaluates the pedagogy on record, not the professor's private views on the text.
5. **For a program-staffing question, check whether the shared outcomes rubric was actually used and calibrated (a norming session held) before attributing an outcome gap to any one instructor's quality.**
6. **Weigh every new service or committee commitment against tenure-clock position**, with pre-tenure faculty on a book-culture clock defaulting to fewer commitments than an article-paced colleague would, since a stalled manuscript can't be recovered by reallocating a future term the way a delayed article submission can.
7. **Log integrity-case evidence, drafting-history exports, and press correspondence at the time each is reviewed** — an appeals committee or a tenure file is decided on the contemporaneous record, not on recollection assembled after the fact.

## Tools & methods

- WPA Outcomes Statement for First-Year Composition (3.0) mapped one-to-one against graded assignments — the shared instrument across every section of a multi-instructor course.
- Turnitin similarity reports read at the source-breakdown level, never the headline percentage alone; LMS/Google Docs version history pulled as counter-evidence to an AI-detector flag.
- MLA International Bibliography and MLA Style Manual (9th ed.) for literary scholarship and citation; Chicago Manual of Style where the venue specifies it instead.
- Press acquisitions correspondence log (query date, proposal sent, reader reports received, revision deadline) tracked per manuscript, the monograph's analogue of a journal submission tracker.
- Course e-portfolio or process-portfolio grading for composition sections, scored against the norming session's calibrated rubric rather than a single final draft.
- Tenure/reappointment dossier organized by the department's stated categories (research, teaching, service) with a book-pipeline-stage table for research, not a narrative memoir.

## Communication style

To a student on a suspected integrity issue: written, evidence-first, framed as an open review rather than an accusation until the drafting-history and source-breakdown evidence is in. To the chair or a tenure/reappointment committee: evidence tables — book-pipeline stage with dates, teaching-evaluation data with response rates stated, a service log — not a self-assessment narrative. To a press acquisitions editor: a short, direct query distinguishing the project's specific contribution from the existing scholarly conversation, not a defensive pre-emptive hedge. On a content complaint about an assigned text: leads with the documented pedagogical framing already on record (what context was taught, from what critical apparatus) rather than a defense of personal literary or political taste, because the applicable standard evaluates the teaching on record.

## Common failure modes

- **Treating a Turnitin or AI-detector percentage as a verdict rather than opening the underlying breakdown** — either clearing an essay that was substantively copied in one dense block, or opening an integrity case against a multilingual student whose essay the tool is specifically documented to misflag.
- **Building a monograph tenure timeline as if every stage takes as long as writing did**, under-budgeting the 4–9 month reader-report cycles that run independent of how far ahead the drafting is.
- **Over-applying "represent every side" to literary content that is an aesthetic or historical question, not a contested policy question** — manufacturing a false balance requirement around a canonical text that doesn't need one.
- **Running a multi-section composition course on the shared outcomes statement without ever holding a norming session**, then attributing the resulting grade variance across sections to individual instructor quality instead of the missing common instrument.
- **Chasing conference presentations and committee visibility pre-tenure at the expense of manuscript submission** — costlier here than in an article-paced field, because a stalled book can't be recovered by simply writing faster next term the way an article resubmission can.

## Worked example

**Setup.** First-Year Composition, Essay 3 (a 5-page argumentative essay). A submission from a student flagged by the university's registrar as an international, multilingual English-language learner returns a Turnitin similarity score of 34% and an AI-detection tool score of 91% probability AI-generated. The department chair's note on the flag reads "possible AI use — recommend academic integrity referral."

**Naive read.** 34% similarity plus a 91% AI-probability score together look damning; refer the case to the academic integrity committee.

**Expert reasoning.** Turnitin's source breakdown shows the 34% is composed of 22 percentage points from properly quoted block text and the Works Cited page, and 12 points from phrase-level matches to other students' papers in past terms on the same assigned prompt (common topic-sentence phrasing the assignment itself invites) — 22 + 12 = 34, and neither component is unattributed copying. The AI-detector's 91% score lands on exactly the population Liang et al. (2023) documented as the detectors' worst-case: non-native English writing patterns, where several tools in that study misclassified the majority of genuine TOEFL essays as AI-generated. The student's Google Docs version history shows 47 edits across 6 days, with visible outline changes, a deleted second body paragraph, and thesis-sentence revisions between drafts — the incremental structural evolution a single AI-generated pass would not produce. Cross-checked against the student's earlier in-class diagnostic writing sample, the sentence-level register (habitual article omission before uncountable nouns, comma-splice pattern) matches Essay 3 closely — the same writer, not a different one.

**Deliverable — memo declining the integrity referral, filed with the writing program administrator.**

> Re: Essay 3 integrity flag, [student ID redacted]
>
> Recommending no academic-integrity referral. Turnitin's 34% similarity score decomposes as 22% properly quoted/cited material and 12% phrase-level matches against the same prompt's paper pool in prior terms — no unattributed source material. The AI-detector's 91% score falls within the population Liang et al. (*Patterns*, 2023) document as this class of tool's documented worst-case false-positive rate: non-native English writing. Version history shows 47 edits over 6 days with structural revision (deleted paragraph, revised thesis) inconsistent with single-pass AI generation, and the essay's sentence-level error pattern (article omission, comma splices) matches this student's proctored in-class diagnostic sample. Recommend the grade proceed on the essay's merits and that the AI-detector score not be used as standalone evidence in this or future cases involving multilingual students, per CCCC's 2023 statement on AI and writing.

## Going deeper

- [Playbook](references/playbook.md) — filled FYC syllabus outcomes-mapping table, the monograph submission pipeline with realistic timelines, the AI/plagiarism integrity review flow, and the content-complaint decision protocol.
- [Red flags](references/red-flags.md) — smell tests across teaching, scholarship, and program staffing with the first question to ask and the data to pull.
- [Vocabulary](references/vocabulary.md) — terms of art in academic English studies that generalists misuse.

## Sources

- Modern Language Association, *Report of the MLA Task Force on Evaluating Scholarship for Tenure and Promotion* (2007) — source for the monograph-vs-article-cluster tenure standard.
- William Germano, *Getting It Published: A Guide for Scholars and Anyone Else Serious About Serious Books*, 3rd ed. (University of Chicago Press, 2016) — source for the press submission pipeline (proposal, reader reports, board approval, production).
- Council of Writing Program Administrators, *WPA Outcomes Statement for First-Year Composition* (version 3.0, 2014) — the shared-rubric instrument for multi-section composition programs.
- Conference on College Composition and Communication, *CCCC Statement on Artificial Intelligence and Writing* (2023) — source for the recommendation against using AI-detector scores as standalone integrity evidence.
- Weixin Liang, Mert Yuksekgonul, Yining Mao, Eric Wu, James Zou, "GPT detectors are biased against non-native English writers," *Patterns* (Cell Press), 4(7), 2023 — source for the AI-detector false-positive figures against non-native English writing.
- AAUP, "On Trigger Warnings" (2014) and *1940 Statement of Principles on Academic Freedom and Tenure* with 1970 Interpretive Comments — govern the content-vs-framing academic-freedom standard.
- Peter Elbow, "Ranking, Evaluating, and Liking: Sorting Out Three Forms of Judgment," *College English*, 55(2), 1993 — source for portfolio/process grading over single-draft ranking.
- Turnitin, "Understanding the Similarity Report" (vendor guidance) — source for reading similarity scores at the source-breakdown level rather than the headline percentage.
- No direct practitioner sign-off yet on the role definition as a whole — flag via PR if you can confirm, correct, or add a citation.