Ethics, Law & Academic Standards: Expert Guide

The Legal Architecture of Plagiarism: Copyright, Attribution, and Civil Liability

Most practitioners conflate plagiarism with copyright infringement, treating them as interchangeable concepts. They are not. Plagiarism is fundamentally an ethical violation — the act of presenting someone else's intellectual work as your own — while copyright infringement is a statutory offense governed by intellectual property law. The distinction carries enormous practical weight: you can plagiarize work that sits entirely in the public domain (copying Shakespeare verbatim without attribution violates academic norms but not copyright), and you can technically infringe copyright without committing plagiarism (using a licensed photograph without proper attribution). the legal framework every writer navigates sits at the intersection of these two domains, and understanding where they diverge is the foundation of any serious compliance strategy.

Under the U.S. Copyright Act of 1976, original works fixed in a tangible medium receive automatic protection from the moment of creation — no registration required. Statutory damages for willful infringement can reach $150,000 per work, a figure that has resulted in multi-million dollar verdicts in cases involving academic publishers and digital content platforms. The EU's Directive on Copyright in the Digital Single Market (2019) extended comparable protections with additional platform liability provisions. What both frameworks share is a focus on reproduction, distribution, and derivative works — the mechanical acts of copying — rather than the attribution failures that define plagiarism in academic and professional settings.

Civil Liability vs. Institutional Sanctions

When plagiarism rises to the level of copyright infringement, civil liability attaches. The plaintiff must demonstrate ownership of a valid copyright, access by the defendant, and substantial similarity between the original and the allegedly infringing work. Courts apply both a "fragmented literal similarity" test (direct verbatim copying) and a more nuanced "comprehensive nonliteral similarity" analysis for structural or thematic reproduction. The 2013 Malibu Media v. John Does line of cases and the high-profile Led Zeppelin v. Skidmore litigation illustrate how courts struggle with the boundary between inspiration and infringement. understanding when plagiarism crosses into criminal territory requires examining whether fraudulent intent or commercial deception is present — elements rarely pursued criminally but increasingly relevant in contract fraud scenarios.

Institutional sanctions operate on an entirely separate track. Universities, professional associations, and publishers impose their own penalties — retraction, degree revocation, termination, permanent blacklisting — without any requirement to meet legal evidentiary standards. The Boldt case at MIT and the documented retractions at major journals like JAMA demonstrate that institutional consequences often outpace and outlast any legal proceeding. These bodies apply internal standards of proof closer to "reasonable likelihood" than "beyond reasonable doubt," which means the practical risk calculus for academics and professionals must account for both tracks simultaneously.

Attribution Obligations and Formal Undertakings

Attribution requirements increasingly appear in contractual form. Publishing agreements, grant applications, and institutional submission portals now routinely include explicit anti-plagiarism clauses with indemnification provisions. the binding nature of plagiarism declarations signed during thesis submissions or journal submissions creates a documented paper trail that transforms an ethical failure into potential breach of contract. Practitioners should treat every signed undertaking as a legal instrument:

• Review the specific definition of plagiarism used in the agreement — institutional definitions vary considerably from legal ones
• Confirm whether self-plagiarism (reuse of your own prior work without disclosure) is explicitly covered
• Identify indemnification scope — some agreements shift legal costs to the author upon any allegation, regardless of outcome
• Retain copies of all source documentation used in cited works, with timestamps where possible

The architecture of plagiarism law is not monolithic — it is a layered system of copyright statute, institutional policy, and contractual obligation that professionals must navigate simultaneously. Treating any single layer as the complete picture is where most serious compliance failures originate.

Criminal vs. Civil Consequences: When Academic Dishonesty Becomes a Prosecutable Offense

Most students who submit plagiarized work expect disciplinary action from their institution — a failed grade, academic probation, or expulsion. What they rarely anticipate is a summons. Yet the legal landscape surrounding academic dishonesty is far more complex than campus policy handbooks suggest, and the distinction between civil and criminal liability is where many misconceptions breed serious, real-world consequences.

The Civil Track: Copyright Infringement and Financial Exposure

The majority of plagiarism cases that escalate beyond institutional walls enter the civil court system, primarily under copyright law. When a student reproduces substantial portions of a protected work — a textbook chapter, a published article, a proprietary dataset — without authorization, the copyright holder has legal standing to sue. Civil damages under the U.S. Copyright Act range from $750 to $150,000 per infringed work in cases where willfulness can be demonstrated. A single dissertation containing passages from five separately copyrighted sources could theoretically expose a student to six-figure liability before attorney fees are even considered.

The civil pathway also extends to contract law. Ghost-writing services and contract cheating platforms create enforceable paper trails that courts have used against both buyers and sellers. Universities in several jurisdictions have successfully pursued civil claims against commercial essay mills, and students who submitted purchased work have faced civil breach-of-academic-contract suits — a legal mechanism that is increasingly formalized in UK and Australian higher education frameworks. Understanding why copying someone else's intellectual work carries genuine legal weight is the starting point for grasping the full exposure students face.

When Cases Cross Into Criminal Territory

Criminal prosecution requires a higher threshold, but it is far from theoretical. Three primary statutes transform plagiarism from a disciplinary matter into a prosecutable offense:

• Wire fraud and mail fraud — used when falsified academic credentials obtained through dishonest work are submitted to employers, licensing boards, or government agencies across state lines
• Identity theft statutes — applicable when a student submits work under another person's name, or when academic records are falsified using stolen credentials
• Federal financial aid fraud — triggered when a student maintains enrollment — and thereby receives Pell Grants or federally subsidized loans — through fraudulent academic performance

The 2019 case of a nursing student in Ohio illustrates the criminal threshold precisely: she was criminally charged not for copying a paper, but because her falsified transcript was submitted to a state licensing body, constituting fraud. Actual incarceration as a consequence of academic dishonesty is documented and prosecutable when the misconduct intersects with these statutory triggers. Sentences in documented cases have ranged from probation to 24-month federal prison terms.

The practical threshold that separates institutional discipline from legal exposure centers on intent, financial benefit, and third-party harm. Submitting a plagiarized essay rarely meets all three criteria. Selling falsified medical degrees, fraudulently collecting financial aid, or misrepresenting credentials to a hiring body routinely does. Students and academics should understand which specific behaviors trigger statutory penalties versus academic sanctions — the overlap is narrower than feared, but the consequences within that overlap are severe enough to warrant serious attention.

Benefits and Drawbacks of Understanding Academic Integrity and Compliance

Institutional Enforcement Mechanisms: Academic Penalties, Disciplinary Procedures, and Long-Term Reputational Damage

Universities and academic institutions have developed increasingly sophisticated enforcement frameworks over the past two decades, moving far beyond the simple "fail the assignment" response that characterized earlier approaches to academic misconduct. Today's institutional mechanisms operate on multiple levels simultaneously — course-level sanctions, departmental reviews, and institution-wide disciplinary boards — each with distinct authority and escalating consequences. Understanding how these systems interact is essential for anyone navigating academic integrity policies, whether as a student, faculty member, or institutional administrator.

The Graduated Penalty Structure

Most research universities in the United States, United Kingdom, and Australia follow a tiered sanctioning model that calibrates penalties to the severity and context of the offense. A first-time instance of inadequate citation in a freshman seminar will typically result in a grade penalty on the assignment — often zero — alongside mandatory completion of an academic integrity module. By contrast, fabricated data in a doctoral dissertation or a master's thesis triggers an entirely different process involving formal hearings, faculty committees, and potential degree revocation. Harvard, for instance, has used its Administrative Board to pursue cases involving degree revocation years after graduation, demonstrating that institutional authority does not expire at commencement. Before accepting any sanction, students should understand the full range of penalties that institutions are legally and procedurally empowered to impose, including those that extend well beyond academic settings.

The procedural pathway itself carries significant implications. Formal academic misconduct hearings are quasi-judicial processes: students may present evidence, call witnesses, and in many jurisdictions appeal decisions to external review bodies. The record generated by these proceedings — transcripts of hearings, written findings, imposed sanctions — often becomes part of a permanent institutional file. This documentation can surface during background checks, bar admission processes, or medical licensing reviews, creating professional consequences that outlast the academic ones by decades.

Reputational and Professional Spillover

The long-term reputational damage from a documented integrity violation is frequently underestimated at the time of the incident. Law schools, medical schools, and graduate programs routinely ask applicants to disclose prior disciplinary actions, and a substantiated plagiarism finding can disqualify an otherwise competitive candidate. The consequences become substantially more serious when misconduct involves published research: journal retractions are indexed by databases like Retraction Watch and remain publicly searchable indefinitely, effectively marking the researcher's name in their professional field. Several high-profile cases — including the retraction of multiple papers by prominent nutrition researcher Brian Wansink — illustrate how a pattern of misconduct, once exposed, can unravel years of otherwise credible work.

Institutions also increasingly require students to submit formal integrity declarations before submitting major work. These undertakings carry real legal and procedural weight — signing one falsely can transform an academic matter into a potential fraud issue, which is why understanding what these declarations actually commit you to is non-negotiable before you put your signature on one. Similarly, while most institutional proceedings remain civil in character, cases involving large-scale contract cheating operations or falsified credentials can cross into criminal territory, and students who assume otherwise are taking a serious risk — the boundary between academic penalty and criminal prosecution for academic dishonesty is narrower than commonly assumed.

• Document everything: Keep copies of all submitted work, communications with instructors, and any detection software reports you receive.
• Engage legal counsel early if a formal hearing is initiated — institutional processes may feel informal but produce legally consequential records.
• Review your institution's appeals procedure before accepting any penalty, even a minor one; accepting a sanction often waives your right to contest it later.

Linguistic Plagiarism and Patchwriting: Identifying Subtle Forms of Intellectual Theft

Most academics and writers understand that copying text verbatim without attribution constitutes plagiarism. What far fewer recognize is that linguistic plagiarism — the appropriation of another writer's distinctive sentence structures, rhetorical patterns, and stylistic choices — represents an equally serious breach of intellectual integrity. A writer can swap out every noun and verb in a source paragraph while preserving its syntactic skeleton and argumentative logic, and the result remains plagiarism in all but the most superficial sense. Detection software increasingly flags these cases, and university honor boards have begun treating them with the same severity as word-for-word copying.

Patchwriting: The Most Pervasive Form You've Never Heard Of

Patchwriting describes the practice of lightly editing source material — replacing a few words with synonyms, rearranging clauses, or breaking one long sentence into two — while retaining the original's intellectual architecture. Rebecca Moore Howard, who coined the term in 1993, estimated that the majority of undergraduate writing contains significant patchwriting, often without any conscious intent to deceive. The problem is structural: students and junior researchers who haven't fully internalized a concept tend to restate it in borrowed language because they lack the cognitive ownership to rephrase it genuinely. Understanding how language itself can be stolen requires examining not just word choice but the deeper architecture of ideas and expression.

Consider a concrete example. If a source reads: "The mitochondrial pathway initiates apoptosis through cytochrome c release, which activates downstream caspases," a patchwritten version might read: "Apoptosis is initiated via the mitochondrial route when cytochrome c is released, subsequently triggering caspase activation downstream." Every technical term remains; only the grammatical scaffolding has shifted. No citation appears. This is not paraphrase — it is intellectual theft with cosmetic alterations.

Legal and Institutional Exposure

The legal dimensions of linguistic plagiarism are less settled than those of verbatim copying, but they are far from negligible. Copyright law protects expression, not ideas — yet courts have found infringement where the structural similarity between two works was sufficiently close, even when individual words differed substantially. Writers operating in commercial contexts should be especially alert: the legal framework governing written work extends beyond simple copy-paste scenarios and can expose publishers, agencies, and individual authors to liability. In academic settings, the consequences are often more immediate, ranging from course failure to expulsion and permanent transcript notation.

What distinguishes legitimate paraphrase from patchwriting comes down to cognitive distance. Genuine paraphrase requires the writer to fully comprehend the source, set it aside, and reconstruct the idea from memory in their own voice. The resulting text should reflect the writer's syntactic habits, not the source's. Institutions increasingly train instructors to recognize the difference, and tools like iThenticate and Copyleaks now analyze structural similarity alongside lexical matches. Given that academic and legal consequences for intellectual theft can follow a writer for decades, developing genuine paraphrase skills is not optional — it is a professional survival competency.

Practical countermeasures include:

• The cover-and-reconstruct method: Read the source, close it completely, wait several minutes, then write from memory before comparing
• Structural auditing: After drafting a paraphrase, count how many clause types match the original — more than two consecutive matches signals patchwriting
• Voice-consistency checks: If a paragraph reads noticeably differently from the surrounding text, it likely borrowed more than ideas
• Over-attribution as a safety habit: Citing a source even for a paraphrase costs nothing and eliminates ambiguity entirely

Self-Plagiarism, Recycled Research, and the Ethics of Reusing Your Own Work

Most researchers understand that copying someone else's work is plagiarism. What catches many academics and professionals off guard is the discovery that reusing their own previously published material can be just as ethically problematic — and in some contexts, legally actionable. Self-plagiarism, sometimes called text recycling or duplicate publication, occurs when an author resubmits substantial portions of their own prior work without proper disclosure. The consequences range from manuscript rejection and retraction to permanent damage to one's academic reputation.

The core ethical issue isn't originality in the creative sense — it's transparency and informed consent from publishers, editors, and readers. When a journal publishes what it believes to be new research, its editorial board, peer reviewers, and ultimately its readership are operating on a shared assumption: the findings are being reported for the first time. Submitting a paper that draws heavily from a previous publication without disclosure violates that assumption entirely. A 2017 analysis of retracted papers in biomedical literature found that duplicate publication accounted for roughly 14% of all retractions — a figure that underscores how seriously the academic community takes this issue.

Where the Line Actually Falls

Not all reuse is problematic, which is where the nuance becomes critical. Methodological sections are commonly considered the gray zone: if your experimental setup is identical across two studies, reproducing that description almost verbatim may be practically necessary. Many journals now explicitly permit this with proper citation. What crosses the line is repackaging the same results, conclusions, or data sets as new findings. Publishing the same study in two journals to maximize citation counts — sometimes called salami slicing when data is divided artificially to generate multiple papers — is a clear ethical violation regardless of how the text is modified.

The linguistic dimension matters too. Some researchers attempt to sidestep detection by paraphrasing their own earlier work extensively, believing that changed wording eliminates the ethical concern. It doesn't. Understanding how language itself functions as a vessel for intellectual claims clarifies why: the ethical breach lies in the undisclosed reuse of ideas and findings, not just verbatim sentences. Detection software like iThenticate routinely flags self-plagiarism, and many publishers cross-reference submissions against an author's own prior publications.

Practical Guidelines for Legitimate Reuse

There are scenarios where building on prior work is not only acceptable but expected. Conference papers expanded into journal articles, dissertations reworked into monographs, and systematic reviews that incorporate earlier meta-analyses all represent legitimate scholarly progression — provided disclosure is handled correctly. The standard practice is explicit acknowledgment in the submission letter and, where appropriate, in the published text itself.

• Always disclose prior publication to editors at submission stage, even for partial reuse
• Obtain copyright clearance if your previous work was published under a publisher's copyright — you may not legally own the right to reuse it
• Cite your own previous work in the manuscript just as you would cite any other source
• Check journal-specific policies on text recycling — thresholds vary significantly, with some journals tolerating up to 10% overlap and others rejecting anything above zero

The copyright dimension is frequently overlooked. When you sign a publication agreement, you typically transfer copyright to the publisher — meaning that substantial reuse of your own published text may constitute a copyright infringement under intellectual property law, not merely an ethical lapse. Researchers who have signed broad transfer agreements should request written permission before reproducing more than a paragraph or two of their own previously published prose. The formal commitments many institutions require around originality — addressed in detail when examining what plagiarism undertakings actually obligate signatories to — typically extend to self-plagiarism as well, a fact that surprises many early-career academics encountering such documents for the first time.

Digital Detection Technology: How Plagiarism Checkers Work, Their Accuracy Limits, and Algorithmic Bias

Modern plagiarism detection software operates on two primary mechanisms: string matching and fingerprinting. String matching compares submitted text against indexed sources character by character, while fingerprinting breaks documents into overlapping token sequences — typically 5 to 8 words — and hashes them for rapid comparison. Turnitin, the dominant player in academic markets, maintains an index of over 1.3 billion student papers, 99 million journal articles, and billions of web pages. iThenticate, its counterpart for researchers and publishers, draws from CrossRef and a proprietary journal database exceeding 155 million documents. Understanding these architectures matters because the database scope directly determines what gets caught and what slips through.

Where Detection Fails: Technical and Conceptual Blind Spots

Similarity scores are routinely misread as plagiarism scores — they are not the same thing. A legitimate literature review might return a 35% similarity rating due to standard citations and discipline-specific terminology, while a paper with 8% similarity could still contain substantive academic fraud through paraphrasing. Turnitin itself warns users that similarity scores require human review and "should never be used as the sole basis for a plagiarism determination." Yet institutional policies frequently set automatic thresholds — commonly 20% or 25% — that trigger sanctions without deeper analysis. Before submitting documents to any cloud-based detection tool, faculty and students alike should review what actually happens to text after it's uploaded, since many platforms retain submissions permanently in their comparison databases.

Detection algorithms also struggle fundamentally with paraphrase plagiarism, translation plagiarism, and idea theft. A student who rewrites a source sentence by sentence — swapping synonyms and restructuring clauses — can reduce a similarity score from 40% to under 5% while reproducing the original argument entirely. Translation plagiarism, where source material is taken from a foreign-language publication and rendered into English, remains largely invisible to most detection tools, which index primarily English-language content. This is not a minor edge case: a 2019 study published in Scientometrics estimated that translation plagiarism accounts for 8–12% of detected misconduct cases in multilingual academic environments.

Algorithmic Bias and Equity Concerns

Detection tools carry measurable bias that disproportionately flags non-native English speakers. Research published in the International Journal of Educational Technology in Higher Education documented that texts written by ESL students received higher false-positive rates due to reliance on common syntactic structures and limited vocabulary variation — patterns that linguistically overlap with source material even in genuinely original work. This connects directly to broader questions about how linguistic identity intersects with originality standards, where native speaker conventions are baked into what algorithms treat as "suspicious." Institutions using these tools without bias-awareness training risk penalizing students for writing style rather than academic misconduct.

The legal consequences of false determinations can be severe, which makes algorithmic literacy non-negotiable for educators. A student expelled based on a misread similarity report has grounds for appeal — and in some jurisdictions, legal recourse. Anyone navigating institutional proceedings should understand the full spectrum of how academic misconduct findings translate into formal and legal consequences. The practical recommendation for institutions: treat any detection report as the beginning of an investigation, not its conclusion, and document the human review process explicitly in academic integrity policies.

• Cross-language detection gap: Most tools index fewer than 15% of non-English academic sources
• AI-generated text: Turnitin's AI detection feature reports a 1% false positive rate at 20% threshold — still consequential at scale
• Database access asymmetry: Open-access institutions often use free tools (PlagScan, Copyscape) with databases 10–50x smaller than commercial alternatives
• Metadata blind spots: Plagiarism checkers cannot assess whether a cited idea was properly attributed, only whether text strings match

Data Privacy and Security Risks of Submitting Academic Work to Third-Party Platforms

Every time a student uploads a dissertation, thesis, or research paper to an external plagiarism checker or AI detection tool, they are transferring intellectual property to a server they have no control over. This is not a hypothetical risk — it is a structural reality of how most commercial academic integrity platforms operate. Understanding what happens to that data after submission is no longer optional for institutions or individual researchers; it is a matter of legal and professional self-preservation.

What Platforms Actually Do With Your Submitted Work

The business model of many third-party plagiarism detection services depends on building and expanding their reference databases. Turnitin, for example, explicitly retains submitted papers in its repository to enable future comparisons — meaning your original work becomes part of their commercial infrastructure. Smaller or less reputable tools may sell metadata, share document content with advertising partners, or store files on servers in jurisdictions with weaker data protection laws. Before submitting any academic work, students and faculty should carefully review the terms of service, specifically the sections covering data retention policies, ownership of submitted content, and third-party data sharing agreements. A detailed breakdown of what reputable versus risky tools actually do is covered in this analysis of how different plagiarism tools handle your data.

GDPR (EU), FERPA (US), and similar frameworks impose legal obligations on institutions that use these platforms on behalf of students. If a university mandates Turnitin submission without disclosing retention practices to students, it may be violating data protection obligations. In 2021, a UK university faced formal complaints to the Information Commissioner's Office precisely because students had not been adequately informed about how their work was being archived. Institutions need Data Processing Agreements (DPAs) in place with any third-party tool they mandate — this is not optional compliance theater, it is a legal requirement under GDPR Article 28.

Specific Risks Students and Researchers Face

The risks are not uniform across all users. Graduate researchers working on unpublished findings, patent-pending innovations, or commercially sensitive data face the highest exposure. Submitting such work to a third-party platform before publication can constitute prior disclosure, potentially invalidating patent applications or giving competitors early access to research directions. For students, the risks are different but equally concrete:

• Permanent archiving of work without explicit consent or the right to deletion
• Cross-institutional matching that can flag legitimately reused personal work as self-plagiarism
• Breach of institutional confidentiality when collaborative or sponsored research is uploaded externally
• Jurisdictional gaps when data is stored on servers outside the student's or institution's home country

Many students sign plagiarism undertakings without reading them carefully, not realizing that some of these documents implicitly authorize third-party submission. Understanding the full legal weight of a plagiarism declaration — including what you are consenting to — is essential before signing anything. The consequences of academic misconduct findings triggered by platform mismatches or faulty detection can be severe, and as the legal consequences of plagiarism rulings demonstrate, these are not merely academic penalties.

The practical safeguard is straightforward: use only platforms that offer one-time submission without retention, verify that any mandatory institutional tool has a signed DPA available for review, and never submit unpublished research containing novel findings or proprietary data to any external service without explicit legal clearance from your institution's research office.

Research Integrity Frameworks: International Standards, Institutional Policies, and the Principal Investigator's Legal Obligations

Research integrity operates within a layered architecture of obligations that Principal Investigators (PIs) frequently underestimate until they face a formal inquiry. At the international level, the Singapore Statement on Research Integrity (2010) and the European Code of Conduct for Research Integrity (ALLEA, revised 2023) set the normative baseline. These documents carry no binding legal force in themselves, but they are routinely incorporated by reference into national funding agency regulations — meaning a violation of their principles can directly trigger grant termination, repayment demands, or debarment from future funding. The U.S. Office of Research Integrity (ORI), for instance, processed 32 formal findings of misconduct in fiscal year 2022 alone, with fabrication and falsification accounting for the majority of cases.

Institutional Policies: Where Abstract Principles Become Enforceable Rules

Every research institution receiving federal or public funding is contractually required to maintain a written Research Integrity Policy and to designate a Research Integrity Officer (RIO). These institutional policies transform international norms into enforceable employment and contractual obligations. A PI who allows a graduate student to submit a manuscript containing undisclosed borrowed text can face personal liability under institutional misconduct procedures, not merely reputational damage. Understanding the scope of what a plagiarism undertaking actually commits a researcher to is therefore not a formality — it is a legally significant act with direct career consequences.

Institutional policies typically mandate that PIs take explicit responsibility for the integrity of all work produced within their laboratory or research group. This supervisory liability extends to co-authored publications, grant applications, and even conference abstracts. The obligation includes verifying that collaborators — particularly early-career researchers — understand applicable standards before submission.

The PI's Legal Obligations: A Practical Breakdown

Beyond institutional policy, PIs operate within a web of statutory and contractual requirements:

• Grant agreement compliance: Most funding instruments (NIH, ERC, DFG) require PIs to certify that all submitted materials are original and accurately attributed. False certification can constitute grant fraud.
• Data management mandates: Retention of raw data for a minimum of 10 years is standard across European and North American frameworks; destruction or manipulation constitutes a separate misconduct category.
• Authorship accountability: The ICMJE criteria define authorship as a legal and ethical commitment; gift authorship and ghost authorship are explicitly prohibited and increasingly prosecuted.
• Disclosure obligations: Conflicts of interest, funding sources, and prior publication of related data must be disclosed; omission is treated as an integrity violation regardless of intent.

The legal dimension of text-based misconduct deserves particular attention. The legal framework governing unattributed copying extends well beyond academic policy into copyright law, and researchers who reproduce substantial portions of prior work — including their own — without proper disclosure face potential civil liability. In jurisdictions where plagiarism intersects with fraud statutes, the consequences can include criminal charges, particularly when the plagiarized content supported a funded grant application.

A dimension that remains chronically underaddressed in PI training is the integrity of linguistic and structural choices themselves — paraphrasing that preserves the substantive intellectual contribution of a source without attribution constitutes misconduct under virtually every operative framework, even when verbatim copying is absent. PIs should conduct mandatory pre-submission integrity checks using validated detection tools, document those checks, and maintain records as part of their standard research governance practice. The record itself becomes evidence of due diligence if a complaint is ever filed.