Permissible Copyright for Re-typesets

[pml]

Hi emeraldimp,

I think you (philosophically) rather miss the point. The score in question had a clearly articulated personal copyright notice with the words "all rights reserved" attached on the first page of music. Now in practice I am not so fussy to enforce all of the rights that could possibly be reserved; I am perfectly happy to have the score redistributed, for it to be used as a study score, and so on. The exception which I wished to exercise was to limit the making of derivative versions, on account of my editorial contributions to the piece.

Lately there was a notice on the work's CPDL page drawing attention to this copyright status, which has been erased as a result of the recent crash. The vocal score currently has the incorrect status, which I had altered back in about March or April to bear the usual CPDL copyright formula. As for the full score's appearance here, I suspect a number of scores were migrated from CPDL after the server was restored.

In general, many of the works I have edited on CPDL have minimal editorial content, and I wholeheartedly agree with them being licensed under a GPL-like license like CPDL's, or under the creative commons unported licenses CC-by or CC-by-sa. I'm not a lawyer however, so for works that for whatever reason I decided should not be completely open, I found it easier just to employ a personal copyright and the well-known formula "All rights reserved."

My edition of the Fauré Requiem is a case where the full score is relatively free to use except for the right to create derivatives. So if I decide I would like to apply a CC-by-nd licence to the edition, then why should IMSLP cavil at this choice? The score is still freely available to be browsed, redistributed, downloaded, copied, printed and so on, just so long as this is done in full; it merely isn't a whatever you please, "free for all". It doesn't prevent anyone from rolling their own version from the p.d. 1900 full score either.

I just mention this, given that some typesets have been put up here without the knowledge of the copyright holders: you really ought to be more careful that the copyright status of the work when it gets listed here at IMSLP is an accurate reflection of that. I suspect the submitter of my work was unaware of the particular conditions I wished to enforce, but by the same token he didn't attempt to inform himself and probably assumed inclusion at the CPDL inferred a GNU-like, CPDL licence.

Lastly without making a careful comparison of new typesets with existing known versions, it would be difficult to always guarantee a typeset has minimal editorial interventions, and not all music editors are disinclined to be litigious, as I think most of you would know from the Lionel Sawkins vs Hyperion Records court case in the UK. The dividing line between literally mechanical transcription and making a new performing edition of a work is considerably more porous as a result.

Regards, Philip

[imslp]

Actually, the CC-BY and CC-BY-SA restriction for new typesets is mostly for legal unambiguity, as they have very close terms with just plain public domain (whereas "public domain" is not allowed as a license in the EU).

To be honest, I never thought about ND while designing the copyrights for typesets, mostly because I honestly really don't see the point of it (sorry ^_^;; ). Even if you don't want people to take your edition and go commercial with it, CC-BY-SA should be more than enough... could you point to a specific reasonable use of the edition the legality of which would differ between CC-BY-SA and CC-BY-ND?

P.S. By the way, feel free to correct the license (or request a removal) of your files that have been submitted to IMSLP. The file will be removed if it infringes copyright.

[pml]

There isn't much difference between the two licences, but pertinently under §3: License Grant, you will find these clauses in the CC by-sa full text:


b. to create and Reproduce Adaptations provided that any such Adaptation, including any translation in any medium, takes reasonable steps to clearly label, demarcate or otherwise identify that changes were made to the original Work. For example, a translation could be marked "The original work was translated from English to Spanish," or a modification could indicate "The original work has been modified."

d. to Distribute and Publicly Perform Adaptations.


A specific, reasonable use that would be allowed under by-sa is to re-orchestrate a score or derive performance parts from it, and as in the instance of the Fauré I wished to make the full score - but not the parts - generally available, hence the most compatible licence is by-nd.

Also, some editors might wish to attach particular conditions to the use of their editions, such as non-commerciality via the by-nc license. Again, would there be any real philosophical difficulty in allowing these conditions to apply to typesets in cases where there is substantial editorial input?

Regards, Philip

[emeraldimp]

I don't think I did miss the point, actually. You claim your score had a non-CC license on it originally. If it was uploaded after the switch to CC only, then it was improperly uploaded and should've been removed (and you should have requested its removal if it was not). This is a legal point, not a philosophical one.

Philosophically, I disagree with your assumption that a creator of a work has any moral right to any control over the work they create. (I'll grant attribution, but that's it.)

But we have to deal with the world as it is, which includes harsh copyright laws. *shrug* It wasn't always this way, and it won't always be. Until then, we have to make do with what we have.

(Call me a hypocrite if you like for my use of CC-BY-SA on the pieces I've retypeset; I'll be changing those when I have the opportunity.)

[pml]

Hi Emeraldimp,

the facts: yes it did have a non-CC license, yes it was improperly uploaded, but rather than request the wholesale and immediate removal of the score (as I could have done), instead I have temporarily changed the copyright notice on the work page to emphasise that I view the copyright terms to be more or less equivalent to CC-by-nd, and drawn this to your attention here.

The ball is in the court of IMSLP to decide whether CC-by-nd is too restrictive a licence for you to be bothered with. If you are inclined to disagree and insist on CC-by or CC-by-sa only, then it might be arguable that all typesets must be removed unless you have clear statements from the rights holders to agree to the works being made available according to the terms of those licenses.

Whether you regard moral rights as a valid concept or not is (unfortunately) irrelevant to the legal argument here because the law indeed acknowledges their existence. I'm not going to call anyone a hypocrite here for taking any particular position under the current state of copyright laws as being in their opinion the best compromise.

Regards, Philip

[Leonard Vertighel]

I haven't thought much about ND yet, but I believe that there is a very pragmatic reason for avoiding NC for typesets. It seems to me that the reason why someone might want to use NC is because he wants his share if someone else makes money from his work (and not because he doesn't want anybody to ever make any money at all). But with the passing of time it can get very difficult if not impossible to track down the typesetter (or possibly his heirs), having possibly no other information than a nickname in a wiki. I see a serious risk of locking material into NC-only usage forever.

I believe that this is quite different with new compositions, since for those as far as I understand the copyright collectives should take care of the licensing issues.

I think that similar arguments apply to ND as well.

[imslp]

I would be somewhat inclined to agree with Leonard.

I don't really have extremely strong feelings against or for ND; it is not as restrictive as the NC licenses, so I'm not strongly against it, though I'm not strongly for it either. Maybe this is partly due to the fact that I really still can't see any major reason why someone would prefer ND over NC in normal retypesets (I'll take this case as an exception). Though I agree with Leonard that the piece would be locked in ND perhaps forever, I suppose in this case it is slightly better since we know your real name.

My inclination is to just manually modify the license to ND, because I don't think the usage of ND for typesets would be anything more than a very rare exception. But this is just for this particular typeset; if there is some other typeset that also wants to go under a ND license, we will have to reopen this discussion, and decide on whether to allow ND, or to remove both files.

But I'm still open to arguments from both sides

[Carolus]

I think part of the confusion here lies in what constitutes an "edition" vs. a "retypeset" (or re-engraving) vs. "reconstruction." PML, I think you sell your efforts a bit short here, since what you've apparently done with Faure's Requiem is really more along the lines of a reconstruction than an edition (as the term is most commonly used).

As I understand it, the only version of the work that exists in complete form is the final one from 1900, which may actually be the work (at least partially) of Faure's student Jean Roger-Ducasse. The earlier versions (1887 and 1893) apparently exist only in fragmentary form, which required you (or anyone attempting to prepare the earlier versions for actual use) to reconstruct them from the existing fragments plus sections of the public domain score from 1900. If I were to prepare a reconstruction of the earlier versions, the chances are that mine would differ in several aspects from yours, simply because we're two different individuals. My own view is that your reconstruction would be legitimate subject of a full copyright (your lifetime plus 70 years in the US) - much like that of a new arrangement or a new orchestration. In light of this (please correct me if my impression is wrong), I would say your desire for no derivatives is quite understandable.

This is a different case than the edition of Eine Kleine Nachtmusik issued by Baerenreiter in the 1950s as part of the NMA, which is essentially a lightly edited re-engraving of the one that Breitkopf issued as part of the Gesammtausgabe eighty years before. It is unlikely that one could actually hear any difference if the same orchestra were to play from one edition vs. the other. In contrast to a reconstruction, stating there should be no derivatives for this type of edition is tantamount to stating there should be no derivatives of Mozart.

This is all a rather complicated issue. The best line of reasoning I've been able to thus far come up with is by breaking down things iinto the following categories:

1. Retypesets. Mutopia offers some examples of this - where a contributor, working in a program like Lilypond, literally re-enters the music - detail for detail - from a single public domain source like the Joseffy edition of the Chopin Mazurkas. There is no original contribution involved here apart from the typographical layout itself. Even in those countries (Germany and the UK) where such typographical contributions are subject to copyright protection, it's for a very limited term (like 25 years). Such items would not be the subject of copyright protection in many countries - including the USA.

2. Urtext Editions. The chief difference here lies with the number of sources consulted. The autograph (if available), early manuscript copies and first editions are typically consulted. As with the category above, in those few countries that recognize a copyright for such ediitons, it's a limited one only. The problem here lies in the ambiguity of many copyright statutes - which means it would take a court case to determine if any given edition was subject to copyright protection in many countries - the USA in particular. If it actually came down to a courtroom battle, many would not stand due to insufficient originality, but some might. As a general rule, the further back one goes in musical history, the more original contribution is required on an editor's part - which can propel the item into Category 4 when a reconstruction process comes into play (see below).

3. Performing Edtions. These are often literally reprints of public domain scores with editorial markings added for phrasing, articulation and other details which do have effect upon the sound of the piece when played. Sort of the antithesis of an Urtext edition, they are frequently subject to a full copyright term of the editor's life plus xx years, or 95 years from publication in the US if published between 1923 and 1977. Such editions can be particularly confusing to many because that International reprint of an old Gutheil score with Leonard Rose's added articulations and phrasing can't be posted here while a clean re-engraved Urtext of the same piece can be.

4. Reconstructions, Transcriptions, Arrangements, Orchestrations. These are true joint works which involve a considerable original contribution from the secondary author. For example, any orchestration of Mussorgsky's Pictures at an Exhibition is only partly Mussorgsky's work, since he wrote the work for piano only. The Albinoni Adagio is another fine example - it would not exist in the well-known form absent Giazotto's contribution. These are subject to full derivative work status nearly everywhere. Last surviving author plus xx years, or 95 years from publication in the USA if published between 1923 and 1977.

I don't know of any way to automate the decision process. Each work has be to examined in order to make the determination. At some point in the future, when the expansion of this site slows down, it will be possible to sift through these items more deliberately and categorize them in a rational fashion. Right now, I certainly do a lot of "shooting from the hip" when confronted with a fresh Sibelius score. It helps a great deal when posters identify things as new arrangements, editions, or reconstructions.

[imslp]

If what PML did was not only a re-engraving but a reconstruction, it certainly has enough added to warrant any of the licenses permitted on IMSLP.

I think the confusion is partly my fault as I haven't spoken clearly as to my intention in limiting the licenses to choose from. The reason I limited retypesets to CC-BY and CC-BY-SA is for legal clarity. The "retypeset" category is really there for retypesets which either do not have copyright in them, or have borderline copyright. And thus, to clarify the legal status of these retypesets, I limited the license choice to CC-BY or CC-BY-SA, as they are also sufficiently close to actual public domain.

For creations which unambiguously contain copyright, such as the #4 Carolus mentioned above, the correct score type is not "retypeset", but "new composition", which allows the entire gamut of copyrights permitted on IMSLP. And this would be the category that PML's Faure reconstruction falls under Maybe someone could find a better name for this category?

[pml]

Thanks, Carolus, for your clear outlining of the argument here. As much as some of my editions have sat squarely under one of your four categories, other editions have straddled several (I can think of a performing version of one of my editions involving aspects of 3 and 4, which is also available as an urtext score plus reconstruction, so combining 2 and 4).

Your comment about "shooting from the hip" is very apt - the determination in such cases is non-trivial and requires time and knowledge, so that is helpful to have information from the submitter as guidance

Although a reconstruction necessarily involves a secondary author (or authors - Mozart's Requiem is an even more notorious example) I regard my contribution to be so much the minor that I have no qualms for the work to be presented under Fauré's name alone - and concert goers usually have no appreciation if they hear the piece performed that they may be hearing the work of Fauré and inter alia, Jean Roger-Ducasse, John Rutter, Jean-Michel Nectoux, or myself.

To return to the subject. I would point out that there are a large number of scores (at WIMA or CPDL for example) which are typesets or urtext as Carolus describes, which have non-commercial licenses specified: but I see little danger of these being locked into NC forever as Leonard fears, because in at least several jurisdictions the copyright protection is only for a limited term (~25 years in the UK or Germany) or non-existent. So if you are really planning for the long-term, I can't see any harm in implementing NC for some of these works, which can be reverted to a more favourable open licence at some point in the future.

For works where you can't necessarily be sure of how much a work may be a straightforward typeset and how great the contribution of the editor was, I think it would be simply minimising risk for IMSLP to specify more stringent copyright protections initially, until you can justify a more lenient one. In other words, typesets go in under the most stringent CC licence, and the vetting process, which you already must follow to approve pieces, then can scale it back to a licence compatible with the work offered (as Carolus described) or compatible with the terms given by the rights holder.

Regards, Philip

PS Some further thoughts on this. Anyone requiring an NC or ND licence for their work perhaps should be identified somewhere by their real name, on the grounds that an end-user may wish to contact them as the rights holder, to ask permission for a waiver of certain of the conditions of the licence. A quick glance at any of the main CC licences confirms that "Any of the above conditions can be waived if you get permission from the copyright holder."

With reconstructions NC or ND licences for full copyright works (US life of author + 70) present more of a problem if the death date of the last surviving author is unknown, or the heirs of the author cannot be traced; but you have to deal with this anyway in the case of establishing death dates for obscure composers, so I don't really see a problem. Having a work available under a restricted licence is surely better than it not being there at all.

[Carolus]

Perhaps a simple change of name from "New Composition" to "Copyright" would suffice. The "Copyright" designation would encompass categories 3 and 4 above along with all new original works. This would leave the full range of copyright options open not only to composers who wish to post their own works, but also those who might wish to post new transcriptions, arrangements, orchestrations, performing editions, or reconstructions (which include continuo realizations, btw). The "Typeset" designation could then be reserved for works which fall entirely into either category 1 or 2 - qualified for a limited type of protection only (Dover can't wantonly hijack your nice new Sibelius engraving of the Bach WTC, for example, even if it is only a re-engraving of the Bach Gesellschaft score, but someone wanting to use your typeset to make their own orchestration for three krumhorns and marine trumpet is free to do so). Hybrids containing elements from categories 3 or 4 would fall under the "Copyright" designation to avoid any possible appropriation of original creative work without permission.

[emeraldimp]

pml: Aye, we must operate within the bounds of the law. But, within the law we can choose how to operate, and which licenses to allow. This is where a philosophical bent can come in. I favor freer.

I think Carolus' idea is a good one, although I'm not sure that 'copyright' is the right title. I'm concerned that it would be overused and used inappropriately.

[davidp_newton]

You are wrong about the EU. The UK most certainly allows people to release something into the public domain. I cannot speak for the other countries but look at the Copyright Designs and Patents Act 1988 and you will see nowhere which says that copyright cannot be waived completely. In the UK some moral rights cannot be waived, but they are separate from copyright.

[pml]

Bump!

Hi Feldmahler,

will you please add to your list of things to fix:

Re-typesets with significant editorial contributions should be able to utilise ANY of the CC licences when specified using the submission page, for the numerous reasons I have cited above, in respect of my own editions. They are neither “vanilla” re-typesets, nor new compositions, and the existing categorisation under the submission form is frankly, inadequate.

I make no apology for differing from certain other posters in this thread, despite wishing in principle for the music public domain to be as widely defined as possible. Unfortunately, the prevailing legal climate does not allow one to be naïve to the realities of the world!

In the case of the score which I've just uploaded, this is for my own protection as an individual: when you see who the default publisher of the work is, I think you, particularly, will appreciate why.

The work is essentially a retypeset with significant editorial contributions (hint: check out the bottom two staves), which however falls well short of being a “new composition”. In other words, it falls squarely into "category 3", as ably defined by the post that Carolus made last year, further up the thread: it's a Performance Edition (permission granted for Canada, Australia, & the US, but NOT the EU).

The bug in the submission process is that only after informing me of the successful upload, did it also mention that only "CC-by" and "CC-by-sa" are permitted for re-typesets – and referred me to a Wiki page that referred to this forum thread (quelle surprise!): and philosophy to one side, I know the arguments are not clear-cut for a number of hard, legal reasons.

If the restriction is being enforced, then the notification should come before the file is submitted, so the uploader has the option to “back out”. The file was uploaded under that particular licence, so I'm standing by it, until the bug is fixed or the policy clarified. (Naturally, I had to visit the page to manually edit the copyright field to CC-by-nc-nd (Unported) 3.0. Grrrrrrr!!!!!!!)

Best regards, PML

[Carolus]

BTW, it is possible to change the license after the fact once the files are uploaded without the system shaking its finger at you. (You probably already know this - but just in case!)