Several years ago, I attended a conference at the Berlin Max Planck Institute. This was the time Larry Lessig was in residence in Berlin, so it must have been more or less June 2004. On the occasion, as you may expect, we discussed Creative Commons licenses. There we debated what I since always describe as the input/output issue.

Here are more or less the terms of the question. Suppose we have an artist who wants to release music under an open license, e.g. Creative Commons attribution. Maybe the artist is the writer and composer of the lyrics and of the music; maybe is also the performer; and may be also is doing the digital recording. In this case the artist probably is in total control of all the creative items which he or she sets out to license under CC. However, it may also well be that some of the inputs, be they the lyrics, the music, the performance, the sound recording, come from another person and entity. If this is the case, then the licensor might wish to make sure that she or he is in fact is authorized to license also these items by their respective right holders; and, more to the point, may wish to make sure that the authorization is specifically referred to a release under CC.

In Berlin, we also started to think in terms of making available to the general public information held by public institutions. Indeed, our host was the Max Planck Institute responsible for the History of Science; and it is well known that this is an institution which greatly cares for dissemination of information, data and knowledge among the public. This was probably the first time I started to consider PSI issues.

A lot of the material which was candidate to dissemination was (and is) in the public domain. However, the input/output issue cropped up also in that connection. It was considered that the process itself of digitization may lead to creation of IP protected works or materials. Consider the metadata: the blueprint (or format) for arranging them probably is creative most of the times. Even short descriptions of an item may attract copyright protection. No doubt, once these metadata are systematically arranged and individually retrievable, a data base is obtained; which means that possibly the sui generis right is triggered. So, to make available under an open license a final product incorporating material in the public domain plus metadata, one should wonder whether authority to make available metadata under open terms is available to begin with.

Another time I had to consider this constellation of facts was a major Italian digitization project, undertaken by an Italian public body and intended to be coordinated at the EU level with other Member State projects. Here again the input/output issue cropped up; and was complicated by the fact that the public body was cooperating with other local bodies and nationwide quasi-public institutions.

These situations do have a number of common features. How are the individual items created which go into the final, complex product, consisting of the original works plus any IP protected data which may be generated in the process of digitization? One possibility is that the job is done by employees of the digitizing institution. Also outside consultants may be employed for this purpose. It is also possible that the job is outsourced, that this that an outside contractor digitizes, creates the metadata, writes the program to make the content and the ancillary information available and so forth. Now, if this is the case, the digitizing institution should make sure that it has been granted an authorization for each and all the inputs which go into the final product by each of the persons or entities which created the individual inputs; and more specifically it should make sure that this authorization extends to the making available of the individual item under a free license.

This is not always a foregone conclusion. Indeed, we should consider that dealings in copyright protected works are based on a doctrine which continental Europeans describe as Zweckübertragung and common law lawyers describe as “divisibility”. This doctrine implies that copyright is made out of a bundle of exclusive rights; that each of these rights is independent of the others; and that the transfer, be it by assignment or license, of one right does not imply that also other rights are transferred. A transfer in gross of all the sticks of the bundle would not have effect; the parties must specify which rights they intend to transfer. If there is not enough by way of specificity, this means that the parties intended to transfer only the rights which were indispensible for the purpose the parties envisaged at the time of the contract.

This principle applies to copyright; but we know that also software is protected by copyright and moreover we may expect that several of the rules to be adopted in connection with data base rights are likely to be developed by case law on the basis of an analogy with copyright principles.

Now, application of this doctrine in the pre-digital environment makes a lot of sense. After all, most of the time, the creator is the weaker party; and it does make sense that he or she should not part with rights unless he or she has specifically accepted to transfer them (and has obtained consideration for this).

However, application of the doctrine may create uncertainty if applied to digitization projects and more generally to PSI. What happens if the employee of the digitizing institution has not signed away his or her rights (on the inputs) providing that he or she authorizes that these are incorporated into a larger product and that this product is disseminated via an open license? What happens if the outside contractor has not signed away its rights with sufficient specificity to comply with the doctrine of Zweckübertragung? Of course it may be argued that the scope of the authorization for the inputs may be inferred from the fact that the creator of the input did know what the contribution was intended for.

This is a solution to the input/output problem which may make sense in many, if not all cases. However, it is also possible to see the issue as one in which ideally a default rule might be created by legislation. It would be sufficient that legislatures spelled out that, any time a employee or an outside contractor works for a public body and the purpose of this public body directly or indirectly includes the dissemination of knowledge, data or information, then any transfer or license from the employee or outside contractor to the public body is presumed to include authorization to make available the input downstream, individually or in combination with other works or materials, under an open license. This is an area which is within the jurisdiction of each Member State, as it usually is the case in connection with rules on ownership. So this is the level in which a solution may be proposed.

This proposal was presented to the Italian government several years ago (in 2006); but it never was taken up. Maybe the time has come to think about it again. In the meantime, it may make sense to look at the contracts with all the parties involved in digitization projects or the creation of PSI data sets, to make sure that the output may be released openly to the outside world without the nasty surprise that some of the inputs which go into it are not clearly available on open terms.