Battle of Ideas

Copyright should last much less long

A steelman map of the public fight over copyright term — arguments for and against, with sources and assumptions.

AI-generated · paired steelman agents · independently red-teamed · Pass-1 source spot-checks only · framing-fidelity not independently verified · single model family

How long copyright lasts — not whether copyright should exist. Berne's minimum vs life-plus-seventy and term extensions. Software, music, and books are in; patents are a sibling, not this claim.

AGAINST 5

no further strong arguments at this depth

FOR 5

no further strong arguments at this depth

People also ask

Questions people actually type. The two columns above are the cases — open a card for sources, assumptions, and counters.

Why does copyright last so long?

Is copyright duration too long?

Copyright term of protection

Copyright term extension act

Copyright time limit

Copyright lasts too long

Ordering within each column: strongest first — validation tier, then source quality, then representativeness.

AGAINST · Copyright should last much less long
Logically validP1

Berne's life-plus-fifty is the international bargain — 'much less long' that undercuts it is a treaty walk-out, not a calibration

Berne Convention Article 7(1): the term of protection 'shall be the life of the author and fifty years after his death.' That is the minimum 181 parties have accepted so that a book sold in Nairobi is not instantly pirate-legal in Paris. WIPO's summary is the same sentence. Countries may grant more; the EU did, in the 1993 Term Directive (now 2006/116/EC), and the US matched in 1998 so that American catalogues would not be the short-term dumping ground under the rule of the shorter term. AGAINST can grant Lessig everything about Mickey and still say: the claim's strongest economic models (Pollock's ~15 years) sit below the Berne floor. Implementing them as domestic law on foreign works either violates the treaty or uses the rule of the shorter term to strip foreign authors, which is a trade weapon, not a public-domain gift. 'Much less long' is ambiguous in ordinary speech. If it means 'cut the extra twenty,' you are arguing with the CTEA and the EU overlay — a real fight, and a smaller one. If it means Pollock, you are arguing with Berne. FOR's page is the second argument wearing the first's clothes. Patents are the sibling that already has a short, disclosure-based term; they are not a Berne copyright, and pretending they are is how you smuggle a 20-year cap into a life-plus system the rest of the world still runs.

Key assumptions

  • Staying in Berne (life+50 minimum) is a constraint the claim must respect, not a lobby talking point untestable
  • FOR's 'much less long' is in practice the Pollock/Founders band, not merely repeal of the extra twenty years partial

Red team — the strongest counters

Berne is a floor countries already overshoot — it is not a veto on cutting the extra twenty

If the claim is 'much less long' than life+70, life+50 is available inside the treaty. AGAINST's 'walk-out' sentence only lands if FOR means Pollock-15. Ordinary language of the claim ('much less long') can be the CTEA/EU overlay. Using Berne as a shield for life+70 is how the overlay becomes invisible. Patents show domestic law can be shorter than maximalist practice without leaving the WTO overnight — copyright's treaty is tighter, but the extra twenty is still a choice.

Treaties have been revised; 'the club' is not geology

Berne's term was not always life+50, and TRIPS locked a version of it. That is path dependence plus exporter politics, not a welfare proof. A coalition to lower the floor is exactly the claim at world scale. Pointing at 181 parties is prevalence, not efficacy — the phones-in-school error applied to WIPO.

Sources

Confidence, decomposed

Logical validity●●●●●
Premise support●●●●○
Representativeness●●●●●
Source quality●●●●●

Provenance

Generated by a paired steelman agent (single model family) · red-teamed by an independent adversarial agent · sources Pass-1 spot-checked (existence and rough fit) — framing-fidelity not independently verified. Judged on merit: per the founding rule of this project, AI authorship is disclosed at site level and arguments stand or fall on their content.
AGAINST · Copyright should last much less long
Logically validP1

A shorter-term island next to life-plus-seventy trading partners is a gift of the catalogue

Copyright is territorial. A US or Kenyan cut to a short term, with Berne's rule of the shorter term in force in many markets, means foreign users can copy your works sooner while your users still cannot copy theirs. Eldred's majority took 'harmonise with the EU' as a legitimate congressional purpose for going up. The inverse is the trade fact AGAINST is pointing at: going down alone is not a public-domain gift to your public. It is a one-way opening. Film, music, and software exporters care about this more than poets, which is why the politics are ugly, but the arithmetic does not care. A coordinated Berne amendment to a lower floor is a different claim — a treaty project. A domestic 'much less long' is free-riding by the longer-term countries on your newly opened stock. Software is the exception that proves the coordination problem: it already leaks, forks, and dies young, so a short term's extra leak is smaller. Novels, films, and recorded music are the goods that travel. AGAINST does not need to love the CTEA. It needs to say that duration is a club good, and clubs that cut first donate the backlist.

Key assumptions

  • Other Berne members will not match a unilateral short term on a relevant deadline partial
  • The rule of the shorter term and ordinary territorial copying would actually open the domestic catalogue abroad testable

Red team — the strongest counters

The rule of the shorter term already lets you be the short country without donating everything

Article 7(8) means your short term follows your works abroad unless the other country opts out. That is a reason exporters hate unilateral cuts. It is also a reason a country that values its public domain more than its catalogue exports can choose to be short. 'Gift of the catalogue' assumes the national objective is export revenue. FOR's objective is the unused middle at home. Those can trade off; they are not a theorem that you must match the longest partner.

Harmonising up is how maximalists always win

If every extra twenty years in the EU forces a US match, and vice versa, the ratchet has no down switch. Eldred's harmonisation rationale is the machine FOR is attacking. AGAINST using it as a constraint is conceding the ratchet. A worldwide claim has to be allowed to argue for coordinated down, not only 'never go first.'

Sources

  • Berne Article 7(8) — rule of the shorter term Term in the country where protection is claimed shall not exceed the country-of-origin term unless that country provides otherwise. Pass-1: WIPO Article 7 text exists. This is the legal machine behind 'unilateral short term donates the catalogue.' P1 checked
  • Eldred v. Ashcroft — harmonisation rationale Majority treats matching EU life+70 as a permissible purpose of the CTEA. Pass-1: 537 U.S. 186 exists. AGAINST uses the inverse: mismatch in the other direction is also a trade fact. P1 checked

Confidence, decomposed

Logical validity●●●●●
Premise support●●●●○
Representativeness●●●●○
Source quality●●●●●

Provenance

Generated by a paired steelman agent (single model family) · red-teamed by an independent adversarial agent · sources Pass-1 spot-checked (existence and rough fit) — framing-fidelity not independently verified. Judged on merit: per the founding rule of this project, AI authorship is disclosed at site level and arguments stand or fall on their content.
AGAINST · Copyright should last much less long
Logically validP1

Orphan-works procedure is the targeted fix — a short term is a blunderbuss that also hits the living catalogue

The US Copyright Office's 2006 and 2015 orphan-works reports recommended limited liability after a reasonably diligent search — a use-it-or-risk-it rule for the unfindable, not a term cut for everyone. EU Directive 2012/28/EU built a version of that for cultural-heritage institutions. Those instruments exist because the problem is identification, not the existence of a term. Heald's disappeared mid-century books are in part an orphan and out-of-print problem, which a diligent-search plus a 'use it or lose the injunction' rule would unlock without opening the still-managed backlist. AGAINST can grant the freezer is real. The remedy that matches the diagnosis is a search standard and a damages cap, plus maybe a registration or renewal formality for the extra decades (the Landes–Posner renewable-term idea). Cutting the unitary term to fifteen years to solve orphans is how you also cut the author who is 60 and still in print. FOR's cleanest duration argument is the unused middle. The cleanest instrument for the unused middle is not this claim. It is a formality or an orphan exception. Two columns that only offer 'life-plus-seventy vs much less' hide that third tool.

Key assumptions

  • A diligent-search / limited-remedy regime would actually free most orphans rather than die in the same lobbying that killed the US bills partial
  • Formalities or renewal for extra decades are politically and treaty-compatible enough to be a real alternative partial

Red team — the strongest counters

The targeted tool has failed in the US for twenty years — that is data

2006 and 2015 Office reports, dead bills, and a narrow EU heritage exception that does not cover ordinary creators are the record. 'Just pass diligent search' is the third camp as a hope. FOR's duration cut does not need a new bureaucracy of search standards that rights-holders can always call inadequate. Political failure of the surgical instrument is a reason to consider the blunt one, not only a reason to try surgery again.

Formalities conflict with Berne's no-formality rule for the core term

Berne Article 5(2) forbids formalities as a condition of protection. Renewable extra decades might be structured to survive that. They also might not, which is why the US dropped formalities when it joined Berne. AGAINST's 'register the extra decades' is a clever design that the treaty and the politics have not actually delivered. Until they do, it is a sketch, not an alternative on the table.

Sources

Confidence, decomposed

Logical validity●●●●●
Premise support●●●○○
Representativeness●●●●○
Source quality●●●●●

Provenance

Generated by a paired steelman agent (single model family) · red-teamed by an independent adversarial agent · sources Pass-1 spot-checked (existence and rough fit) — framing-fidelity not independently verified. Judged on merit: per the founding rule of this project, AI authorship is disclosed at site level and arguments stand or fall on their content.
AGAINST · Copyright should last much less long
Logically validP1

The author's remaining life is not a rounding error in the bargain

The moral and incentive core of Berne is not 'fifty was a number we picked from a hat.' It is: the author should not see their work go free during their working life, and their dependents should have a defined after. Eighteenth-century 14-year terms assumed short commercial lives and short remaining lives. A novelist who debuts at 30 and lives to 90 would, under Pollock, have been in the public domain since 45 — competing with free editions at the peak of a reputation the book itself built. That is a different authorship deal than 'we pay you a salary.' Copyright is how independent authorship exists without a patron or a wage. AGAINST can grant that life-plus-seventy is a lobby overlay and still defend life-plus-something as the human-scale unit. Software employees already got a wage; their extra decades are the weakest part of the unitary term. Novelists, composers, and illustrators did not. A reform that splits software off is the sibling conversation. A reform that uses software's depreciation to cut the novelist's life-plus is the category error running in the other direction. Orphan-works procedure (diligent search, limited remedy) is how you free the unfindable middle without declaring the living author a public utility.

Key assumptions

  • An author competing with free copies of their own work during life is an unacceptable deal, even if the first 15 years captured most NPV untestable
  • Dependents' post-mortem years are part of the incentive, not a pure transfer partial

Red team — the strongest counters

An author can sell books while alive under a 28-year term; they did

US 1909-Act terms and the Founders' 14+14 did not make 19th-century authorship impossible. Life-plus is a 20th-century thickening of the bargain, coinciding with longer lives and thicker lobbies. 'Competing with free copies at 45' is a real disutility; it is not a showing that the books would not have been written. Salaries, patrons, and serialisation already exist as complements.

'Two generations of descendants' is a recital, not a finding

The EU directive's moral arithmetic (author plus two generations) is a political sentence. It is how you get life+70 from life+50, not how you know grandchildren must collect. Dependents can be covered by a shorter post-mortem (25, 30) without 70. Using the recital as if it were evidence is AGAINST doing what FOR is accused of with Mickey: a slogan doing legislative work.

Sources

  • Berne Convention Article 7 and WIPO summary Life + 50 as the general minimum; designed around the author's life, not a fixed 15-year clock. Pass-1: WIPO text exists. P1 checked
  • Directive 2006/116/EC recitals EU term directive recitals: Berne's life+50 was intended to protect the author and two generations of descendants; the EU added 20 for harmonisation. Pass-1: EUR-Lex exists. The 'two generations' sentence is the official moral arithmetic AGAINST is using, not a finding that two generations is optimal. P1 checked

Confidence, decomposed

Logical validity●●●●●
Premise support●●●○○
Representativeness●●●●●
Source quality●●●●○

Provenance

Generated by a paired steelman agent (single model family) · red-teamed by an independent adversarial agent · sources Pass-1 spot-checked (existence and rough fit) — framing-fidelity not independently verified. Judged on merit: per the founding rule of this project, AI authorship is disclosed at site level and arguments stand or fall on their content.
AGAINST · Copyright should last much less long
Logically validP1

The works that still earn are not a rounding error — they are why a long tail exists

Pollock's decaying-demand curve is an average. AGAINST's objection is selection: the socially and commercially important remainder is exactly the fat tail the average treats as noise. Beatles catalogues, Disney's remaining in-copyright features, Tolkien, hymnals, textbooks that stay in print, photographic archives, and enterprise software that still runs a hospital are not 'Mickey as rent.' They are works whose incentive and whose reinvestment (remastering, translation, restoration, security patches) continue for decades. Landes and Posner (and the later 'indefinitely renewable copyright' line) take the tail seriously: a one-size term that kills the tail to free the median unused work is a policy that taxes hits to subsidise unused inventory. Heald's Amazon hole is real for the median mid-century book. It is not a finding that Steamboat Willie's cousins should have been free in 1955. Music collecting societies' distributions, however skewed, are the empirical remainder: some compositions earn in year sixty. A short term would move that earning into whoever can industrialise public-domain hits fastest — not into a pastoral commons of amateur orchestras. FOR can still want the unused middle in the public domain. The instrument 'much less long for everything' is how you hit the tail you can see to free the middle you cannot.

Key assumptions

  • The long-earning tail is large enough, or important enough, that a unitary short term is the wrong tool for the unused middle partial
  • Continued exclusive rights are what fund remastering, translation, and preservation of that tail — not just a transfer to grandchildren partial

Red team — the strongest counters

The tail can be licensed without a century of exclusion of the unused middle

Renewable terms, registration for extra decades, or a use-it-or-lose-it formality (Landes–Posner adjacent) would keep Beatles-scale catalogues owned and dump the Heald hole into the public domain. AGAINST's 'don't hit the tail' is a reason for a formality, not for unitary life+70. The hits are visible; they can check a box.

Remastering and preservation happen in the public domain too

Project Gutenberg, Internet Archive, and competing editions of Dickens are the existence proof that exclusive rights are not required to keep a work available — Heald's own result. Studios restore in-copyright film because they own it; public-domain film is restored by collectors and archives when the legal risk dies. The 'we need a term to restore' sentence is true of some budgets and false as a general law.

Sources

  • Indefinitely Renewable Copyright Landes & Posner, University of Chicago Law Review 70(2):471–518, 2003. Argue against a short hard cutoff and for a renewable term that lets still-valuable works stay owned. Pass-1: Chicago Unbound / JSTOR exist. P1 checked
  • Eldred v. Ashcroft, 537 U.S. 186 (2003) — majority on existing-term extensions Ginsburg majority: Congress may extend terms, including on existing works, under the Copyright Clause; harmonisation with EU life+70 was a stated purpose. Pass-1: Justia exists. Not a welfare proof — the legal fact that the tail was the point of the Act. P1 checked

Confidence, decomposed

Logical validity●●●●○
Premise support●●●○○
Representativeness●●●●○
Source quality●●●●○

Provenance

Generated by a paired steelman agent (single model family) · red-teamed by an independent adversarial agent · sources Pass-1 spot-checked (existence and rough fit) — framing-fidelity not independently verified. Judged on merit: per the founding rule of this project, AI authorship is disclosed at site level and arguments stand or fall on their content.
FOR · Copyright should last much less long
Empirical — moderateP1

Orphan works are the tax of a term longer than anyone can find the owner

The US Copyright Office's 2006 Report on Orphan Works, reaffirmed in the 2015 Report on Orphan Works and Mass Digitization, is the official description of a market failure the Office cannot legislate away: works still in term whose owner cannot be identified or located, so a user who would pay or who would make a non-commercial use does neither, for fear of a later infringement hammer. The 2008 congressional testimony of the Office put the scale in the only honest unit — 'millions of orphan works' locked out of productive use by authors, publishers, filmmakers, archives, museums, historical societies. Europe built a narrower orphan-works directive (2012/28/EU) for cultural-heritage institutions; the US bills died. FOR's causal claim is not that a diligent-search exception is useless. It is that duration is the machine that manufactures orphans. The longer the term, the more likely the creator is dead, the publisher dissolved, the contract in a box, the heir unfindable. Life-plus-seventy on a 1955 trade book is how you get a work that is neither commercially exploited nor legally usable. A term of a few decades would have released the bulk of that stock into the public domain — the regime that already lets Project Gutenberg reprint 1910 without a detective story. This is worldwide: every Berne country that stacked twenty extra years onto a registration-free, automatic copyright enlarged the freezer. Shorter duration is the cleanest orphan-works policy because it does not depend on a search standard that never quite satisfies rights-holders.

Key assumptions

  • A large share of orphans would be used if they were public domain, not merely theoretically 'locked' partial
  • Shorter duration beats a diligent-search exception as the remedy — the exception is the third camp FOR is not required to occupy partial

Red team — the strongest counters

Orphans are an identification failure; a short term is a blunderbuss

The Copyright Office recommended diligent search plus limited liability, not a 15-year cap. EU 2012/28/EU already built a heritage exception. FOR's 'duration manufactures orphans' is a real mechanism. It is not the only instrument, and it hits living catalogues to free unfindable pamphlets. If the US bills died of lobbying, that is a political failure of the targeted tool, not a proof the blunt tool is better.

'Millions of orphans' is a scale word without a use-count

The Office's 'millions' is an order-of-magnitude from comments, not a census of works that would actually be digitised if free. Heald's availability gap is better evidence of unused stock, and even that is Amazon-in-print, not 'filmmakers are waiting.' Demand for the orphaned middle may be thin — which is why nobody renewed.

Sources

Confidence, decomposed

Logical validity●●●●●
Premise support●●●●○
Representativeness●●●●●
Source quality●●●●●

Provenance

Generated by a paired steelman agent (single model family) · red-teamed by an independent adversarial agent · sources Pass-1 spot-checked (existence and rough fit) — framing-fidelity not independently verified. Judged on merit: per the founding rule of this project, AI authorship is disclosed at site level and arguments stand or fall on their content.
FOR · Copyright should last much less long
Empirical — moderateP1

Copyrighted mid-century books vanish from print; public-domain books of the same age do not

Paul Heald (Journal of Empirical Legal Studies, 2014, 'How Copyright Keeps Works Disappeared') sampled more than 2,000 new books for sale on Amazon, grouped by decade of original publication. Titles from before the US 1923 public-domain line were in print in large numbers; titles from the mid-20th century — still in copyright — were a trough. Public-domain books averaged something on the order of four times more editions per title than copyrighted books in his comparison. A companion look at bestsellers found 94% of 1913–22 public-domain bestsellers available as e-books by 2014 versus 27% of 1923–32 bestsellers still in copyright. That is the opposite of the 'we need a long term so publishers keep the backlist alive' story. The backlist that is free is the backlist that is alive. The backlist that is owned is the backlist a rights-holder can sit on, fail to digitise, or price as a collector's item. FOR can grant that some living catalogues (Beatles, Disney, Stephen King) are commercially worked. The Heald hole is about the typical work, which is what a term of life-plus-seventy mostly governs. Shorter copyright would move the trough into the public domain while the paper copies still exist and the audience for a 1960 novel is not yet a specialist. This is not a US-only Amazon artefact: CREATe's later Nordic extension of the availability work finds the same public-domain vs in-copyright gap. Duration that outlasts commercial life is how culture disappears with a © on the corpse.

Key assumptions

  • Amazon-in-print counts track cultural availability rather than just US retail mix partial
  • Rights-holders sitting on works is the mechanism, not a lack of demand for 1930s–60s titles partial

Red team — the strongest counters

Heald's hole can be demand, not only sitting-on-rights

Mid-century books that are out of print may be out of print because nobody wants them at a cost-covering price, while 19th-century public-domain 'classics' are a selected, taught, already-famous stock. Comparing Dickens-in-the-commons to a 1955 forgotten novel is not an apples-to-apples availability experiment. Four times more editions of public-domain titles can be a classics effect.

Some rights-holders do keep backlists alive, and a short term would free-ride on that investment

University presses, estate-managed authors, and studios that restore film are the tail Heald's median misses. A short term transfers the still-valuable remainder to whoever reprints fastest, which can be a race to cheap scans rather than a garden of careful editions. Public-domain availability is real; public-domain quality is a different variable.

Sources

Confidence, decomposed

Logical validity●●●●●
Premise support●●●●○
Representativeness●●●●○
Source quality●●●●●

Provenance

Generated by a paired steelman agent (single model family) · red-teamed by an independent adversarial agent · sources Pass-1 spot-checked (existence and rough fit) — framing-fidelity not independently verified. Judged on merit: per the founding rule of this project, AI authorship is disclosed at site level and arguments stand or fall on their content.
FOR · Copyright should last much less long
Empirical — moderateP1

The models of optimal term land in years, not in life-plus-seventy

Rufus Pollock (Review of Economic Research on Copyright Issues / 'Forever Minus a Day?', 2009) writes down the standard trade-off: a longer term raises the incentive to create at a decreasing rate (because of discounting) and raises the social cost of excluding users and follow-on creators. Calibrating to observed decay of demand for recordings and books, he gets a point estimate around 15 years, with a 99% interval that still tops out around 38 years — below the Berne minimum of life-plus-fifty, let alone life-plus-seventy. That is a model, not a trial. FOR does not need the point estimate to be exact. It needs the shape: almost all of the present value of a typical work is collected in the first years and decades, so the last forty years of a life-plus-seventy term are nearly all cost. Heald's later graphics of how quickly a publisher recoups present value make the same point without Pollock's specific 15. The US Founders' 14-year term (renewable once) was in this band. So is the patent term. Software, which is obsolete inside a decade, is the extreme of the same curve: life-plus-seventy on a 2015 app is a joke about grandchildren. Novels live longer; they still do not need a term that outlasts the author's grandchildren's mortgages. 'Much less long' on this page means: decades, not life-plus-seventy. It does not mean zero, which would be a different claim.

Key assumptions

  • Discounted present-value models are the right way to set a cultural term, rather than an author's-life moral right untestable
  • Demand-decay observed in 20th-century books and recordings still describes streaming-era hits partial

Red team — the strongest counters

Pollock's 15 years is a calibration of 20th-century decay, not a law of streaming hits

Discount rates and 'cultural decay' estimated on old recordings and books need not describe a 2020s catalogue that still earns on Spotify in year forty. The 99% interval up to ~38 years is already a different policy from 15. Treating a working-paper optimum as the claim's number is how a model becomes a slogan. Patents are a sibling with a different disclosure bargain; they are not a copyright RCT.

NPV-in-the-first-years ignores option value, reputation, and the author's life as the unit

A novelist may not recoup in year two and still be building a name that makes book five possible. A term that expires in year 15 while the author is 45 is a deal in which the author's later reputation is a public-domain externality they cannot capture. That is AGAINST's life-plus point, and a present-value model that ignores it is answering a different question than Berne asked.

Sources

Confidence, decomposed

Logical validity●●●●●
Premise support●●●○○
Representativeness●●●●○
Source quality●●●●○

Provenance

Generated by a paired steelman agent (single model family) · red-teamed by an independent adversarial agent · sources Pass-1 spot-checked (existence and rough fit) — framing-fidelity not independently verified. Judged on merit: per the founding rule of this project, AI authorship is disclosed at site level and arguments stand or fall on their content.
FOR · Copyright should last much less long
Logically validP1

Term extensions on works already made buy no new creation — they buy Mickey

The incentive theory of copyright is prospective: a term is a bargain to induce a work that would not otherwise be made. Retroactive extension fails that test by construction. The US Copyright Term Extension Act of 1998 (Sonny Bono / 'Mickey Mouse Protection Act') moved the US from life-plus-fifty (the Berne floor) to life-plus-seventy for authors, and to 95 years from publication for corporate works. Steamboat Willie (1928) was nearing the end of its then-term; Disney and other catalogue owners lobbied; Congress added twenty years and applied them to existing works. Eldred v. Ashcroft, 537 U.S. 186 (2003), upheld the Act. Lawrence Lessig's Free Culture is the public FOR voice: the extension was a wealth transfer, not a creativity policy. The first Mickey shorts have since entered the US public domain on the 95-year clock — which is the tell. If a century of exclusive rights was required to get a talking mouse drawn, the mouse would not have been drawn in 1928 under a much shorter term. The EU had already gone to life-plus-seventy in the 1993 Term Directive (now 2006/116/EC), and the US followed to 'harmonise.' Harmonising up is not evidence the extra twenty years induce novels. It is evidence that trading partners with valuable catalogues prefer the extra twenty. FOR's claim is duration, not existence. Berne's life-plus-fifty is already a political floor. Life-plus-seventy and 95-from-publication are the overshoot this argument is for cutting. Patents, which die in twenty years from filing, are the sibling that shows the legal system already knows a short exclusive right can fund expensive invention.

Key assumptions

  • A term that cannot have caused an already-created work is a transfer, not an incentive — including when it is sold as 'harmonisation' untestable
  • Catalogue owners' lobbying is the actual political cause of the extra twenty years, not a new theory of authorship partial

Red team — the strongest counters

Retrospective extension can still be a harmonisation and a bargaining chip, not only Mickey

Eldred's majority treated matching EU life+70 as a legitimate purpose. US catalogues in Europe would have been the short-term dumping ground under the rule of the shorter term. That is a trade argument, not a creativity argument — FOR is right that it cannot have caused Steamboat Willie — but it is a reason a country might extend existing terms without the cartoon being the whole story. 'Mickey' is the slogan of a coalition, not a complete legislative history.

The incentive theory is prospective, and creators take the current term as the deal

Authors and studios making works today bargain against life+70 / 95-from-publication. Cutting the term on future works changes that deal; cutting it on existing works is the clean taking. FOR's best shot is the retrospective part. A forward-only shorter term is a different, smaller claim, and Berne still sits under it.

Sources

  • Eldred v. Ashcroft, 537 U.S. 186 (2003) Supreme Court upheld the CTEA's move to life+70 / 95 years from publication, including as applied to existing works. Pass-1: Justia opinion exists. The dissent (Stevens, Breyer) is the judicial FOR: retrospective extension is not 'progress of science.' P1 checked
  • Copyright Term Extension Act (Pub. L. 105–298, 1998) US term to life+70 (authors) and 95 years from publication (works for hire). Colloquial 'Mickey Mouse Protection Act.' Pass-1: statute/CRS/Wikipedia legislative history exist. Disney-lobby story is the standard political history, not a secret memo. P1 checked
  • Free Culture (Lawrence Lessig) Lessig 2004: CTEA as a transfer to existing owners; Eldred litigation. Pass-1: book exists. Public AGAINST-the-extension voice, not a duration RCT. P1 checked

Confidence, decomposed

Logical validity●●●●●
Premise support●●●●○
Representativeness●●●●●
Source quality●●●●●

Provenance

Generated by a paired steelman agent (single model family) · red-teamed by an independent adversarial agent · sources Pass-1 spot-checked (existence and rough fit) — framing-fidelity not independently verified. Judged on merit: per the founding rule of this project, AI authorship is disclosed at site level and arguments stand or fall on their content.
FOR · Copyright should last much less long
Logically validP1

One term for software and for novels is a category error — and we already know how to write a short exclusive right

Copyright is automatic, registration-optional in Berne countries, and unitary: the same life-plus-seventy (or 95-from-publication) covers source code, a pop song, a textbook, and a film. That is a legislative convenience, not a finding that those markets share a depreciation curve. Software's commercial life is short; reuse of old code is exactly the follow-on creation a long exclusive right blocks (forks, interoperability, abandoned tools). Music catalogues have a longer tail. Literary works longer still. FOR's 'much less long' can be a single shorter term — Pollock's decades — or a split term that Berne does not currently require us to refuse. Patents already split: twenty years from filing for inventions, because the point of the bargain was disclosure plus a race, not a dynasty. The sibling is the point. Nobody serious thinks a drug patent should run life-plus-seventy; the deadweight of excluding generic manufacture is obvious. The deadweight of excluding a 1962 textbook, a 1987 font, or a 2004 Java library is the same shape at lower visibility. If the AGAINST column wants a long tail for novels, the honest design is a longer term for works that still earn, not a global life-plus-seventy that turns every GitHub repo and every schoolbook into a trap for orphans. Duration reform is how you stop using the novel as a human shield for the studio catalogue and the software estate.

Key assumptions

  • Different classes of work have different economically useful lives, so a unitary long term is the wrong default partial
  • Patent-style short terms are an existence proof that exclusive rights need not track the author's remaining life plus seventy testable
  • Patents remain a sibling: this argument uses them as a design comparison, not as this claim untestable

Red team — the strongest counters

A split term is the honest design, and it is not this claim

FOR's software-vs-novels point is a reason to unbundle, not a reason to cut the unitary term to software's life. Berne already allows shorter minima for photos and film. A software-specific term is a treaty-and-statute project. Using GitHub's depreciation to take the novelist's life-plus is the category error AGAINST named, running in FOR's voice.

Patents disclose; copyright does not — the sibling is the wrong shape

A 20-year patent is a bargain for a teaching disclosure and a race to invent around. Copyright does not require source, stems, or a plot outline to be filed. Treating the patent term as existence proof that exclusive rights 'need not track the author's life' skips the disclosure half of the patent bargain. Software might belong in a patent-like box. Novels do not.

Sources

Confidence, decomposed

Logical validity●●●●●
Premise support●●●●○
Representativeness●●●●○
Source quality●●●●○

Provenance

Generated by a paired steelman agent (single model family) · red-teamed by an independent adversarial agent · sources Pass-1 spot-checked (existence and rough fit) — framing-fidelity not independently verified. Judged on merit: per the founding rule of this project, AI authorship is disclosed at site level and arguments stand or fall on their content.