Translation, Localisation, and Adaptation Rights Toolkit: Grants, Vendors, Moral Rights, and Markets

By Casey Scott McKay · 2026-06-12

A translation is a derivative work, which means it needs permission, and it is also an original work, which means the translator owns it unless the contract says otherwise. Almost every problem in this area comes from one of those two facts being overlooked. This toolkit collects the analysis. It works through the scope of a translation grant, why language and territory are different variables that grants routinely conflate, and the reversion and approval terms that decide whether a rights holder keeps control. It then addresses the vendor layer: who owns a translation, a translation memory, a terminology database, and a dubbed performance, and what machine translation does to all four. It closes on moral rights, which are far stronger in most translation markets than in the United States, and on brand localisation and transliteration.

IP and Technology > Copyright | Toolkit | Published 12 June 2026 - Updated 10 July 2026 | Casey Scott McKay - marksy.us

Summary. A translation is a derivative work, which means it needs permission, and it is also an original work, which means the translator owns it unless the contract says otherwise. Almost every problem in this area comes from one of those two facts being overlooked. This toolkit works through the scope of a translation grant, why language and territory are different variables that grants routinely conflate, and the reversion and approval terms that decide control. It addresses the vendor layer — who owns a translation, a translation memory, a terminology database, and a dubbed performance, and what machine translation does to all four — and closes on moral rights and brand localisation.

Keywords: translation rights toolkit · derivative works · localisation vendors · subtitling and dubbing · moral rights abroad · translator ownership · territory and language grants · transliteration · brand localisation · machine translation · translation memory ownership · voice performance rights · terminology databases · game localisation · publishing rights

Start Here

Two propositions govern this entire field and they pull in opposite directions.

A translation is a derivative work. Under 17 U.S.C. § 101 a translation is expressly a derivative work, and 17 U.S.C. § 106(2) reserves to the copyright owner the exclusive right to prepare derivative works. So translating without permission is infringement, whatever the quality or the motive.

And a translation is an original work. 17 U.S.C. § 103 extends copyright to derivative works to the extent of the material contributed, which means the translator owns copyright in the translation — subject to lawful use of the underlying material — unless there is an assignment or a work made for hire arrangement.

The consequence is that a translation has two owners unless the paperwork says otherwise, and the second one is the vendor a publisher paid a fee to and never contracted with properly.

Then the field adds a third complication. Most translation markets are outside the United States, and most of those jurisdictions give translators moral rights of attribution and integrity that cannot be waived — so a contractual arrangement that works domestically may be unenforceable in the market where the translation will actually be used.

And a fourth. Machine translation and post-editing have changed the economics and unsettled the ownership analysis, because the output of a machine may have no author at all.

This toolkit works through grants, vendors, moral rights, brand localisation, and the operational programme, in that order.

The Grant: Language, Territory, and Scope

Language and territory are different variables, and grants conflate them constantly.

A grant of "Spanish language rights" does not say whether it covers Spain, Latin America, or Spanish-language publication in the United States — which is a large market and one the grantor may have intended to keep.

A grant of "Latin American rights" does not say whether it covers Portuguese for Brazil.

So specify both, as a matrix: which languages, in which territories, and whether the grant is exclusive in each combination.

Specify the format and medium. Print, ebook, audio, serialisation, adaptation, dramatisation, and — increasingly the disputed one — whether the grant extends to formats that did not exist when it was made.

Specify the term and the reversion. A perpetual grant of a translation right is a grant of that language market forever, and reversion on failure to publish within a period, or on the translation going out of print, is standard and frequently omitted.

Specify approval. Over the translator, over the translation, over the title, over the cover, and over any abridgement or adaptation. An author who cannot read the target language still needs an approval mechanism, usually exercised through a trusted reader.

Specify the credit for the author and for the translator, in a defined form and position.

And specify what happens to the translation on reversion. Does the rights holder acquire it, licence it, or lose access to it? This is the single most consequential term in a translation grant and it is almost never addressed — with the result that a publisher whose rights have reverted still owns the only translation into that language, and the author cannot licence the market to anyone else without commissioning a new one.

The Vendor Layer

Localisation runs on vendors, and the ownership questions multiply with each asset the vendor creates.

The translation itself. Owned by the translator absent an assignment. Work made for hire under 17 U.S.C. § 101 covers a translation as one of the enumerated categories for commissioned works — but only with a signed written agreement stating that it is a work made for hire, so the instrument is required, not merely helpful.

Take an assignment as well as a work made for hire designation, because the designation may fail for reasons of form or foreign law and the assignment is the fallback.

The translation memory. A database of source and target segments built across projects, and the single most valuable asset in a long-running localisation programme. Vendors routinely claim it, and a client that changes vendors without owning its memory pays to translate the same sentences again.

The terminology database and style guide. Same analysis, smaller value, equally often unaddressed.

The dubbed performance. A separate performance with its own rights: the voice actor's performance, the sound recording, and — in many jurisdictions — neighbouring rights and residual obligations under collective agreements.

Subtitles and captions. Text, owned by whoever created them, and frequently produced by a subcontractor of a subcontractor.

Layout and typesetting. Design work with its own thin copyright, and a source of dispute when files are moved between vendors.

Quality assurance and linguistic review output, which is commentary rather than translation, and which nobody addresses at all.

Contract for all seven, with a single assignment clause covering everything created for the client, plus an express obligation to deliver the working files and the memory in a standard interchange format on termination.

Machine Translation and Post-Editing

The economics of the field have been rebuilt around machine output, and the legal analysis has not caught up.

Machine output may have no author. Copyright requires human authorship, and a raw machine translation produced without human creative contribution is unlikely to attract copyright — which means nobody owns it and anybody may copy it.

Post-editing changes that, to the extent of the human contribution, and the resulting work is protectable in the edited portions. Which makes the level of post-editing a legal question as well as a quality one, and one that pricing tiers describe more precisely than most contracts do.

The training question is separate and more consequential. Whether a vendor may use a client's source content, its translations, or its memory to train or improve models is a term that appears in many vendor agreements by default and that a client with confidential or pre-release content cannot accept.

Read that clause in every localisation agreement, and read it again in the terms of any translation tool used directly by the client's own staff.

Confidentiality is the practical exposure. Pre-release product documentation, unannounced marketing, financial materials, and legal documents routinely pass through translation vendors, and a vendor whose tooling sends segments to a third-party service has disclosed them.

Require a stated processing architecture: which engines, hosted where, with what retention, and whether any segment leaves the vendor's environment.

Liability for machine error should be allocated, because an incorrect translation of a safety warning, a dosage, or a contractual term is a different kind of defect from a stylistic one.

And disclose where disclosure matters. Machine-translated content presented as human-authored can be a consumer protection issue in some contexts and a professional obligation issue in others.

Moral Rights, Which Are Stronger Elsewhere

The United States has a narrow moral rights regime — 17 U.S.C. § 106A covers only works of visual art — and most translation markets do not.

In many civil law jurisdictions the author has inalienable rights of attribution and integrity, which survive assignment, cannot be waived in advance, and are enforceable by the author personally and by heirs.

So three parties may hold moral rights in a localised work: the original author, in the underlying work; the translator, in the translation; and, for audiovisual work, the performers and sometimes the director.

Integrity claims in translation are real. Abridgement, bowdlerisation, censorship for a market, and substantial adaptation for cultural fit have all been the subject of author complaints, and a licence permitting adaptation does not necessarily answer a moral rights objection.

Attribution is the more common issue and the cheaper to solve. Translators expect and are frequently entitled to credit, and omitting it is both an infringement in some jurisdictions and a professional norm violation everywhere.

Contract accordingly. Even where a waiver is unavailable, an agreement recording what adaptations are contemplated, obtaining consent to them specifically, and fixing the credit reduces the exposure substantially.

Local counsel should confirm the position in any market where a substantial adaptation is planned, because the answer varies more than practitioners expect between jurisdictions that all describe themselves as protecting moral rights.

And in audiovisual localisation, check the collective agreements, because performer residuals and re-use obligations frequently attach to a dubbed performance and are enforced by unions rather than by individuals.

Brand Localisation and Transliteration

A brand entering a new language market makes three decisions, and each has trademark consequences.

Keep the existing mark, which preserves global consistency and may be unpronounceable, meaningless, or unfortunate in the target language.

Transliterate, producing a phonetic equivalent in the target script, which requires clearance as a separate mark and which the market may adopt in a form the brand did not choose.

Or adopt a new local mark, chosen for meaning, which requires full clearance and creates a second brand to maintain.

Clear whichever is chosen under the local equivalent of 15 U.S.C. § 1052(d), and clear it before the launch is announced, because a first-to-file jurisdiction will otherwise see the name registered by a distributor or a broker.

Register the transliteration even if the brand does not use it. Markets generate their own local versions of foreign brand names, and a brand that has not registered the version consumers actually use has no rights in it.

Screen for meaning and connotation in the target language and in adjacent dialects, which is a linguistic exercise rather than a legal one and which prevents the failures that become industry anecdotes.

Watch the descriptive trap. A word that is arbitrary in English may be descriptive or generic in the target language, and unregistrable there.

Coordinate the domains, handles, and marketplace listings in local scripts, which are frequently registered by opportunists ahead of a launch.

And control the localisation of the brand's own materials, because a translation vendor rendering a tagline literally can create a claim, a descriptive statement, or a defamation exposure that the source text did not contain.

Sector Variations

Publishing. Translation rights are sold territory by territory and language by language, typically by an agent, with an advance and royalty structure and a reversion on failure to publish. The recurring failures are the unaddressed reversion of the translation itself and the omission of translator credit.

Audiovisual. Dubbing and subtitling for each market, with performer rights, collective agreement obligations, and — for streaming distribution — accessibility requirements that make captioning a compliance matter as well as a localisation one.

Games. The most complex, because the text is embedded in software, the localisation includes voice, art, and cultural adaptation, and the vendor frequently touches source code. Ownership of the localised assets, the memory, and any engine modifications should all be addressed, and the confidentiality exposure before launch is acute.

Software and documentation. High volume, high reuse, and therefore the sector where translation memory ownership matters most. Also the sector where machine translation and post-editing are most established.

Marketing. Transcreation rather than translation, producing a substantially new work whose authorship sits with the agency, and where the resulting copy needs the same claims and regulatory review as the source did.

Legal and regulatory. Certified and sworn translation requirements in some jurisdictions, liability for error, and a confidentiality profile that generally rules out consumer translation tooling entirely.

Life sciences. Regulatory submissions and labelling in local languages, where translation error is a safety and compliance issue and where the validation process is part of the quality system.

And user-generated and community translation raises its own ownership question: volunteers own their contributions absent a licence, and a platform relying on community localisation needs contributor terms that most do not have.

Building the Programme

Map the rights you hold and the rights you grant. A matrix of languages by territories by formats, with terms and reversion dates, is the foundational document and most publishers and content owners do not have one.

Fix the grant template: language and territory as separate axes, format specified, term and reversion defined, approval mechanism named, credit fixed, and the fate of the translation on reversion stated expressly.

Fix the vendor template: assignment plus work made for hire designation, ownership of memory and terminology, delivery of working files in interchange format on termination, confidentiality with a stated processing architecture, a training prohibition where required, and liability allocation for error.

Own the memory. Take delivery of it periodically rather than at termination, because a vendor in dispute is a vendor who delivers slowly.

Build the moral rights position market by market for the territories that matter, with local confirmation where substantial adaptation is planned.

Clear the brand localisation — existing mark, transliteration, or local mark — before announcement, and register the transliteration regardless.

Route localised marketing copy through the same claims review as the source, because a literal translation of a permissible claim may be an impermissible one.

And review annually: the rights matrix against actual publication, the vendor arrangements against the assets held, and the trademark position in each market against what consumers actually call the product.

A Worked Example

A publisher licenses Spanish rights to a Latin American house, which commissions a translator on a flat fee with no written agreement, and the book sells well. Five years later the licence reverts.

The publisher does not own the translation. The translator does, because there was no assignment and no work made for hire instrument, and 17 U.S.C. § 103 gives the translator copyright in what they contributed.

Nor does the licensee, for the same reason — so the reverted rights are worth less than they appear, because the only Spanish text in existence belongs to a third party.

The United States Spanish-language market was never addressed. The grant said "Spanish language," and the parties assumed different things about whether that included Spanish-language publication in the United States.

The translator was not credited on the second edition, which in the translator's own jurisdiction is a moral rights violation and which in every jurisdiction is a professional grievance that will be aired publicly.

And the audio edition was produced from the translation without a separate grant, because the original licence said nothing about format.

Five problems, one licence, and the drafting that would have prevented all five is a language-territory-format matrix, a reversion clause addressing the translation, a translator agreement with an assignment, and a credit provision — perhaps two hundred words in total.

Scale and Cadence

An author or small publisher needs the language-territory-format matrix, a reversion clause covering the translation, and a translator agreement with an assignment and a credit provision.

A content owner localising into a handful of markets needs the vendor template with memory ownership, a confidentiality architecture, and brand clearance in each market before announcement.

A large localisation programme needs all of that plus periodic memory delivery, a machine translation policy stating what content may pass through which tooling, a moral rights position per market, and a claims review for localised marketing.

A platform relying on community translation needs contributor terms granting a licence, and it needs them before the community exists rather than afterwards.

Review annually, and on five triggers: a new market, a new vendor, a new format, a reversion date, and any adoption of new translation tooling by any team.

Adaptation Beyond Language

Localisation frequently goes further than translation, and the further it goes the more it engages the derivative work right under 17 U.S.C. § 106(2).

Cultural adaptation — changing names, settings, references, humour, and examples — produces a work substantially different from the original and requires a grant that says so. A translation licence is not an adaptation licence, and the two are routinely conflated.

Abridgement and expansion for market conventions, particularly in children's publishing and educational material, similarly require express permission.

Format shifts — a book into an audio production, a text into an interactive experience, a film into a series — are separate derivative works needing separate grants, and a grant of translation rights covers none of them.

Regulatory adaptation in labelling, instructions, and safety material is compelled by local law and is still an adaptation, which means the grant should permit it expressly rather than leaving the licensee to infringe in order to comply.

Censorship for a market is the most sensitive case, engaging the author's moral rights directly, and a licensee asked by a market authority to remove content should return to the rights holder rather than decide alone.

Sequels, spin-offs, and characters are outside every translation grant and belong in a separate negotiation.

And where the adaptation is substantial enough, the resulting work has its own author with their own rights — which is the same problem as the translator's copyright, at a larger scale.

So define the permitted adaptation in the grant, list what requires approval, and record the approvals when they are given, because a decade later nobody will remember what was agreed by email.

Accessibility and Compelled Localisation

Some localisation is required rather than commercial, and it engages a different set of rules.

Captioning and subtitling for accessibility are obligations in several distribution contexts, and the resulting captions are works with their own authorship even when produced under compulsion.

Audio description is a further derivative work, scripted and performed, with the same ownership analysis and the same tendency to be commissioned without paperwork.

Accessible format conversion for people with print disabilities is permitted by 17 U.S.C. § 121 for authorised entities, and the cross-border framework established by the Marrakesh arrangements permits exchange of accessible copies between countries.

That exception does not extend to commercial localisation, and businesses occasionally assume it covers more than it does.

Plain language and simplified versions are adaptations, and where they are produced for regulatory or public interest reasons the grant should still say so.

Terminology mandated by regulators — approved drug names, standardised safety phrases, official translations of legal terms — constrains the translator's choices and should be supplied rather than left to the vendor to research.

And official or sworn translation requirements in some jurisdictions require a qualified translator whose certification carries personal liability, which changes both the cost and the contractual allocation.

Enforcement and Infringing Translations

Unauthorised translations are common and the enforcement analysis is unusual in two ways.

The claim is infringement of the derivative work right, and it does not require that any words be copied literally — a translation copies expression by definition, which is why translation is enumerated in the statutory definition of a derivative work.

But proving it requires the target language. A rights holder who cannot read the infringing text needs an expert, and the comparison is between the translation and the original rather than between two texts in the same language.

Registration governs the remedies. 17 U.S.C. § 411 conditions suit on registration, Fourth Estate Public Benefit Corp. v. Wall-Street.com, LLC, 586 U.S. 296 (2019), requires that the Office have acted, and 17 U.S.C. § 412 governs statutory damages and fees — which for a foreign-language edition of a registered work is usually available on the underlying registration.

Foreign works are protected without formality in most cases through treaty relations reflected in 17 U.S.C. § 104, and the registration precondition applies to United States works.

Platform takedown handles most of it. Notice and takedown under 17 U.S.C. § 512 reaches unauthorised translations posted online, and marketplace processes reach printed editions.

Fan translation is the recurring practical problem in games, manga, and serialised fiction. It is infringement; it is also frequently the market's own promotional infrastructure; and the enforcement decision is commercial rather than legal.

Where a rights holder wants to convert rather than suppress, a community licence with quality and credit terms is available and is used more often than practitioners expect.

And derivative work status cuts both ways. An unauthorised translator has no enforceable copyright in an infringing translation, so a rights holder who later wants to use it is negotiating from strength rather than from need.

Practical Drafting Notes

Define language by standard code rather than by name. "Spanish" is ambiguous; a locale code is not, and it distinguishes regional variants that matter commercially.

Define territory by country list, not by region, and address digital distribution expressly because a digital edition sold in one territory is available everywhere unless it is geo-restricted.

Address the digital availability question directly. Territorial exclusivity in a world of global platforms requires either geo-restriction obligations or an acceptance that exclusivity is nominal.

State the format list exhaustively and add a clause allocating formats that do not yet exist, in whichever direction the parties intend.

Fix reversion triggers: failure to publish within a period, going out of print, falling below a sales threshold, insolvency, and change of control.

State the translation ownership on reversion — assignment, licence, or nothing — and price it.

Fix the credit for author and translator, in a defined form and position, on all editions and in all marketing.

Require delivery of files. The final text, the working files, and the memory, in a stated format, within a stated period.

Include an audit right proportionate to the economics, and a clear royalty definition including the treatment of deep discounts, bundles, and subscription distribution.

And include a moral rights acknowledgement listing the adaptations contemplated and consented to, which is worth more in most markets than a waiver whose validity is doubtful.

A Ninety-Day Programme

Days one to fifteen. Build the rights matrix: every work, every language, every territory, every format, every licensee, with terms and reversion dates. Most content owners discover during this exercise that they cannot say what they have granted.

Days fifteen to thirty. Identify the translations the organisation uses and cannot prove it owns. Missing translator assignments are the commonest finding, and retrospective assignments are usually obtainable for a modest fee while the translator is contactable.

Days thirty to forty-five. Audit the vendor agreements for memory ownership, terminology, working file delivery, confidentiality architecture, and any clause permitting use of client content for model training.

Days forty-five to sixty. Take delivery of the memories and terminology from every current vendor, in interchange format, and establish a periodic delivery obligation rather than a termination one.

Days sixty to seventy-five. Fix the grant and vendor templates, and check the brand localisation position — existing mark, transliteration, or local mark — in every market where the organisation sells.

Days seventy-five to ninety. Write the machine translation policy, route localised marketing through the claims review, and set the annual review with its five triggers.

The output is one matrix, two templates, and a set of delivered assets — and the matrix alone is usually enough to change how the organisation licences.

A Closing Note

Translation is the part of copyright practice where the doctrine is simple and the paperwork is uniformly bad.

Everyone knows that a translation is a derivative work. Rather fewer act on the second half of the proposition, which is that a translation is also an original work with its own author — and so the field is full of publishers, platforms, and content owners using translations they do not own, produced by people they never contracted with, in markets whose moral rights regimes they have not checked.

The commercial consequences arrive late. A licence reverts and the text does not come with it. A vendor changes and the memory does not. A translator is uncredited and complains publicly. A market authority requires a change and nobody knows who may authorise it.

Every one of those is prevented by a page of drafting at the point when the relationship is formed, and none of them is fixable afterwards without the cooperation of somebody who no longer needs to cooperate.

Which makes the rights matrix the highest-value document in this area. It is tedious to build, it is never funded on its own merits, and it is the only thing that lets an organisation answer the question it will eventually be asked: what have we granted, to whom, in what language, for how long — and who owns the translation.

What Clients Actually Ask

"We paid for the translation — don't we own it?" No. Payment is not assignment, and 17 U.S.C. § 204 requires a signed writing. A translation is one of the enumerated categories eligible for work made for hire treatment, but only with the instrument, so get both the designation and the assignment.

"Can we use machine translation on this?" Ask what the tooling does with the text and whether the content is confidential or pre-release. Then ask how much post-editing there will be, because that determines whether the output is protectable.

"The vendor says the memory is theirs." It probably is, under their standard terms. Negotiate it, and take delivery periodically rather than at termination.

"Our licence reverted — do we get the Spanish text?" Only if the grant said so. This is the term that decides whether a reverted market is worth anything.

"Do we need to credit the translator?" In most translation markets, yes as a matter of law and everywhere as a matter of professional norm. It costs a line and its omission is the grievance most likely to become public.

"Can we adapt it for the local market?" Translation is not adaptation. Get the adaptation right expressly, and record the specific changes consented to.

"What is our biggest exposure?" A decade of translations produced by freelancers on purchase orders, with no assignments, underpinning a catalogue the business believes it owns.

Working With Translators

Translators are professionals with strong norms, and a practitioner who understands them gets better outcomes than one who treats them as vendors.

Credit is not negotiable in practice. Professional bodies have campaigned on it, publishers have conceded it, and omitting it generates public criticism disproportionate to its cost.

Royalty participation is standard in literary translation in many markets and unusual in commercial localisation, and conflating the two produces friction.

Rates and turnaround are the commercial terms, and a rate that assumes machine post-editing should say so, because the professional distinction between translation and post-editing is real and pricing that blurs it is resented.

Confidentiality is well understood and readily accepted, and translators are frequently better at it than the businesses that hire them.

Rights buy-outs are accepted where they are explicit and priced, and resisted where they are buried in a purchase order.

Attribution of machine output to a human translator is a professional integrity issue and should never be requested.

And the collective and freelance context matters. Many translators work through agencies that impose their own terms downstream, and a client with an assignment from the agency may still have a gap if the agency's own contract with the translator is silent.

Ask for the chain, once, at the start of the relationship. It is a reasonable request, it is rarely refused, and it is the difference between owning a catalogue and believing you do.

Diligence in a Transaction

A buyer of a content catalogue asks four questions about localisation, and most sellers answer three of them badly.

Which languages and territories are granted, to whom, for how long? The rights matrix answers this in a page, and its absence adds weeks to the process.

Do you own the translations? Assignments from every translator, or work made for hire instruments, or an honest statement that certain translations are used under implied licence — which is a discount rather than a deal-breaker if disclosed.

What comes with the vendor relationships? Memories, terminology, and working files, with evidence that they have been delivered rather than promised.

And what are the reversion dates? A catalogue whose principal foreign licences revert within eighteen months is a different asset from one whose licences run for a decade.

Two further questions arrive in regulated and audiovisual sectors: whether performer residual obligations attach to dubbed material, and whether accessibility obligations are being met in every distribution channel.

The seller who prepares these answers controls the process. The seller who does not spends the diligence period reconstructing a decade of purchase orders while the buyer's counsel drafts an indemnity — which is the ordinary outcome and the avoidable one.

A Note on Proportion

A single work licensed into two languages needs a grant with a matrix, a reversion clause, and a translator assignment. That is three paragraphs and it covers the position completely.

A publisher with a list needs the rights matrix and the two templates, maintained by whoever handles subsidiary rights, which is a role that already exists.

A software or games business localising continuously needs the vendor discipline: memory ownership, periodic delivery, confidentiality architecture, and a machine translation policy — which is a procurement exercise more than a legal one.

And a business entering a market for the first time needs the brand localisation decision made deliberately, before announcement, with the transliteration registered whether or not it is used.

Everything else in this toolkit is refinement, applicable in proportion to the size of the localisation programme and the value of the catalogue it produces.

But the three paragraphs are not optional at any scale, because the failure they prevent — using a translation nobody assigned, in a market nobody defined, under a grant that never said what happens at the end — is the same failure whether the catalogue is one title or ten thousand.

The scale changes the cost of the remedy, not the availability of the prevention.

Which is why the three paragraphs belong in the first template a client adopts, not in the remediation exercise they run five years later.

A Suggested Reading Path

Start with the doctrine in Saying It in Another Language.

Then the practice in Localising Content for Foreign Markets.

Then the audit in the translation and localisation checklist.

For ownership fundamentals, Who Owns the Work? and Transfers, Licenses, and Termination Rights.

For moral rights, The Artist's Other Rights and Advising on VARA and Moral Rights.

For international filing and priority, Filing on a Foreign Basis and the Global Brand Enforcement Toolkit.

For the brand naming decision, the Trademark Clearance and Brand Selection Toolkit.

For the AI and tooling layer, the AI Content and IP Toolkit and the AI Procurement and Governance Toolkit.

And for audio and performance rights, Everything You Hear and the Podcasting and Audio Rights Toolkit.

Primary Authorities

| Authority | Proposition | |---|---| | 17 U.S.C. § 101 | Derivative work; work made for hire categories | | 17 U.S.C. § 103 | Copyright in derivative works | | 17 U.S.C. § 106 | Exclusive rights including derivative works | | 17 U.S.C. § 106A | Attribution and integrity in the United States | | 17 U.S.C. § 201 | Initial ownership | | 17 U.S.C. § 204 | Transfers require a signed writing | | 17 U.S.C. § 203 | Termination of transfers | | 17 U.S.C. § 302 | Duration | | 17 U.S.C. § 104 | National origin and treaty protection | | 17 U.S.C. § 411 | Registration precondition to suit | | 17 U.S.C. § 412 | Statutory damages and fees | | 15 U.S.C. § 1052(d) | Likelihood of confusion in local clearance | | 15 U.S.C. § 1126 | Foreign filing basis and priority | | Feist Publications v. Rural Telephone Service | Originality and human authorship | | Cmty. for Creative Non-Violence v. Reid | Employee versus contractor | | Stewart v. Abend | Derivative works and underlying rights | | Andy Warhol Found. for the Visual Arts v. Goldsmith | Purpose and character | | Campbell v. Acuff-Rose Music | Transformative use | | Fourth Estate Pub. Benefit Corp. v. Wall-Street.com | Registration before suit | | Berne Convention moral rights | Attribution and integrity abroad | | Translation memory ownership in vendor contracts | The asset vendors claim by default | | Machine translation and copyright authorship | Human contribution and protectability | | Dubbing and performer neighbouring rights | Performance rights in localisation |

Forms and Templates

The License Agreement Template supplies the structure for a translation and territory grant, and the provisions that decide its value are the language-territory-format matrix, the term and reversion, the approval mechanism, the credit obligations, and — the term this toolkit returns to repeatedly — what happens to the translation itself when the grant ends. The Assignment Agreement Template covers translators, editors, voice performers, and localisation vendors, and should combine a work made for hire designation with an assignment so that a failure of the first does not leave the client with nothing. The Portfolio Inventory Template adapts into the rights matrix that this field requires: works, languages, territories, formats, licensees, terms, reversion dates, and translation ownership. Beyond those, maintain three records: a vendor asset register covering memories, terminology, and working files with the date each was last delivered; a machine translation policy stating what content may pass through which tooling; and a market-by-market note on moral rights and brand localisation.

Related Toolkits and Checklists

The Museums, Libraries, and Cultural Heritage IP Toolkit covers institutional translation and access programmes. The AI Content and IP Toolkit covers machine output and training exposure. The Trademark Clearance and Brand Selection Toolkit covers the local naming decision. The Global Brand Enforcement Toolkit covers protection of the localised brand, and the Podcasting and Audio Rights Toolkit covers voice performance and audio rights that dubbing engages.