Does the ND (NoDerivatives) clause forbid the distribution of any form of text or data mining based on papers with such a license?

Question

I was reading "Your Guide to Publishing Open Access with Elsevier" (URL: https://www.elsevier.com/__data/assets/pdf_file/0020/181433/openaccessbooklet_May.pdf (mirror)) and saw this table on page 6:

Why does the CC-BY-NC-ND 4.0 license in this Elsevier only allow "Text & data min[ing]" for private use only and not for distribution? I understand one cannot make commercial use of it but does the ND (NoDerivatives) clause forbid the distribution of any form of text or data mining based on papers with such a license?

ND (NoDerivatives) is described by creativecommons.org as follows:

If you remix, transform, or build upon the material, you may not distribute the modified material.

It's not obvious to me that the output of a text mining or data mining program necessarily qualifies as a derivative.

Answer 1

According to the Guide to Mining CC-Licensed Material (mirror), the answer may depend on what you produce with the text/data mining. If you run some calculations and then write a paper describing what you learned, then that paper would not be a derivative work. If you produce a data set based on the mining, then that data set might be considered an adaptation of the original work. The guide does not propose a test for distinguishing between these possibilities. If it matters to you, I expect you would need to consult a lawyer.

Incidentally, I wouldn't take Elsevier's table too seriously. For example, it suggests that the CC BY-NC-ND license allows distributing a translation (since it doesn't have a "not for distribution" note there). However, that just isn't true: the license (mirror) explicitly lists translations as being adapted material that cannot be shared.