--- title: Are metadata facts? layout: post image: feature: header_orange.png --- This post picks up an argument that I made in [Theses on the Metaphors of Digital-Textual History](https://www.amazon.co.uk/Metaphors-Digital-Textual-History-Stanford-Technologies/dp/1503614883/ref=sr_1_1) about facts and copyright. Namely, that although facts are exempt from copyright, factual status is not necessarily stable. Some thoughts reproduced from the book: The Catch-22 situation in which factuality finds itself is as follows: 1. An academic publisher publishes a fact as constructed by scientific realist principles; 2. This fact, not subject to copyright, is extracted in plain descriptive language and is published, stored, and preserved; 3. The study’s results are disproved, changing the factual status of the proposition; 4. As a result, the fact is no longer a fact and, therefore, can be placed under copyright again; Publishers can end up, here, in a strange situation where the only propositions they can defend against copyright claims are false/incorrect results. It would not be a good look for an academic publisher to pursue a copyright claim on the basis that what they published was untrue. After all, their reputation is founded on the principle that their outputs (and science) produce meaningful, truthful descriptions of reality. Yet the types of statement that scientific results produce – “psilocybin helps to cure treatment-resistant depression” / “psilocybin does not help to cure treatment-resistant depression” – appear to have the straightforward descriptive characteristics of factual language, regardless of which of the sentences is actually true. Of course, many “factual” statements are actually just too straightforwardly descriptive and are, therefore, devoid of originality for copyright. That is to say that even shocking or straightforward sentences that are untrue may also not merit protection. The statement “elephants are grey” is almost certainly not copyrightable. However, likewise, the statement “elephants are orange” probably also does not, although the contextual situation will play a role here. Should the line “elephants are orange” form part of, say, a longer poem, it is possible that copyright protection would inhere in this false statement. The context matters because it provides an environment in which to read a statement. Although jurisdictions worldwide may interpret the definition of factuality differently, a good example in the US context is Custom Dynamics, LLC v. Radiantz Led Lighting, Inc. In this case, the court found that “photographs of aftermarket motorcycle lighting accessories ‘were meant to serve the purely utilitarian purpose of displaying examples of its products to potential customers, and do not merit copyright protection’.” That is, the use to which a set of photographs were being put determined whether they were accorded copyright protection. By contrast, had the same photographs been destined for an art gallery, they would have been protected. Hence, to understand the copyright status of scientific works that have been retracted for untruth, we must think about the context of scientific publishing and how a court is likely to understand it. On the one hand, academic publishing is an environment that purports to truth and integrity. The basic purpose of most scientific research is accurately to describe current or predict future reality. As such, it might follow that statements in academic papers are deemed factual/descriptive, even if they fail to remain true for all time. Simultaneously, academic papers (usually) possess originality as one of their key criteria. The crafting and execution of experiments – or the writing of, say, literary criticism in the humanities disciplines – must advance our understanding in ways that have not previously been broached. Yet there is, often, a distinction between the writing of a paper and the execution of an experiment. It may require significant originality and creativity to setup an experiment (say, for an example, to establish the existence of gravitational waves). However, this does not mean that the writing of the scientific paper replicates these characteristics. A short series of sequential factual statements that each merely describe a truthful situation may not meet the threshold for originality and creativity. On the other hand, the academic publishing industry relies on the copyrightability of various aspects of scientific articles. Often described as an “oligopoly” for the extreme profits made by this sector, in many academic publishing systems researchers are required to sign over their copyright at the time of acceptance. Norms of academic plagiarism are also an important consideration here. If someone else were to reproduce multiple statements from an academic paper, without citing the original author, it would be considered an academic offence. Copyright violation and plagiarism are not the same thing, even if the latter can entail the former. Yet the fact that plagiarism exists as a context would likely inform a court’s verdict on the copyrightability of academic statements. That said, even if one does not think that scientific statements, in sequence, in a paper, are individually copyrightable, it is possible that the design and layout of a typeset academic article possesses the characteristics of originality and creativity required for copyright. This is to say that there are various diegetic layers of copyright. A first layer may be the basic factuality of statements, which is not copyrightable (indeed, it is not necessarily even a tangible thing). The second layer is the expression of these facts in words. These statements may or may not be subject to copyright. Finally, there is the sum total of the article, which exists within various industry, financial, and market contexts. It is highly likely that this artifact as a whole is copyrightable, even if only for the layout characteristics. A similar feature pertains to metadata. Metadata, or data about data, are data that are descriptive of other forms. That is, they have the appearance of factual assertion. For this reason, it is sometimes said that metadata are "facts" that cannot be copyrighted and should always be open for maximum utility of the original object. However, we also know that metadata can be wrong. People can assert incorrect metadata. Different parties can assert metadata that conflict with one another. In short, metadata is a distributed set of assertions that triangulate an object. But what does this mean for the factual status of those metadata? Are they fact-like in their assertion structure, or is it the truthfulness of metadata that matters for factual status? Put simply: can we have factual statements that are false? Orange elephants all the way down... Also: does it matter _who_ asserts the metadata as to its factual status? Do some asserters have jurisdiction over the factuality of said data? Perhaps the owner of the original data? Yet even if they make a mistake, the truthfulness of said metadata may be in question. And we return to that same question: are there facts that are not true, or is it really the descriptive banality of a fact-like statement about which we care?