When I first started in web development a decade ago, with slow dial-up connections prevalent, the web was a text-heavy place. Today, browsing the web is less like reading a book, and more like to wandering through a gallery. Text floats and wraps elegantly around pictures to tell stories in a magazine format. Images fluidly adapt and scale to whatever device a viewer is reading on. Web editors will rarely plonk an image they have found straight into their site, but instead stretch, shape and tweak it to reflect their design framework.
(NB Yes I know this site looks dreadful at the moment, as I haven't had time to redesign for the last couple of years.)
But as publishers tweak the colour balance, adjust the file format and above all crop images to scale beautifully, how many publishers stop to consider the ND – or “No Derivatives” – element of a Creative Commons licence, under which many images used by responsible publishers are licenced?
In this post I want to lay out the issues in relation to a question that seems to pop up in several places (see
), without ever having been satisfactorily answered: can you crop an image under a
Before proceeding further, two caveats apply.
Firstly, I am not an expert in this area, and as is evident below I don't think there is actually a clear answer to be had. Nevertheless, I hope laying out some of the issues will help contribute to the debate. If anyone with more expertise wants to wade in below the line or via twitter
, please do.
Secondly, this is a problem only in an abstract sense. On the ground, publishers almost invariably will
be manipulating images in order to fit the aspect ratios required for their sites. Additionally, as I explore at the end of this post, site content syndicated to social media platforms is automatically modified beyond the control of the publisher. As so often in the field of copyright – witness the historical illegality of digitising one's CD collection
– legislative or licencing issues are out of sync with the way things are being used in practice.
Nevertheless, when working over at READ towers
, I am as fastidious as I can be about not breaching copyright or licences, so for what it’s worth here’s the problem as I see it.
Derivation and adaptation in the Creative Commons licences
Here’s the human-readable version of the ND component in a Creative Commons BY-ND 4.0
No Derivatives — If you remix, transform, or build upon the material, you may not distribute the modified material.
This comes with the caveat that:
Merely changing the format never creates a derivative.
So it's fine to migrate an image to a different format (for instance, to save a large TIFF as a smaller JPEG). However, on face value of the above, even a minor crop is a transformation of sorts and presumably cannot be permitted. Nevertheless, this is a bit confusing, because the Attribution part of this same licence specifies that:
In 4.0, you must indicate if you modified the material and retain an indication of previous modifications. In 3.0 and earlier license versions, the indication of changes is only required if you create a derivative.
So we must indicate changes if we create a derivative, but apparently we're not allowed to create derivatives in the first place anyway. However, when we look at the ND element in its expanded, legal code
, we seem to have a little more room to manoeuvre, to iron out this inconsistency:
Adapted Material means material subject to Copyright and Similar Rights that is derived from or based upon the Licensed Material and in which the Licensed Material is translated, altered, arranged, transformed, or otherwise modified in a manner requiring permission under the Copyright and Similar Rights held by the Licensor. For purposes of this Public License, where the Licensed Material is a musical work, performance, or sound recording, Adapted Material is always produced where the Licensed Material is synched in timed relation with a moving image.
As ever the language is quite slippery. While the idea of adaptation seems qualitative, involving substantial and creative manipulation, as soon as the term “altered” is thrown into the mix our options seem to be narrowed, since an alteration is a more quantifiable measure. Hence the Creative Commons FAQ
falls back on national laws:
What is an adaptation?
An adaptation is a work based on one or more pre-existing works. What constitutes an adaptation depends on applicable law, however translating a work from one language to another or creating a film version of a novel are generally considered adaptations.
In order for an adaptation to be protected by copyright, most national laws require the creator of the adaptation to add original expression to the pre-existing work. However, there is no international standard for originality, and the definition differs depending on the jurisdiction. Civil law jurisdictions (such as Germany and France) tend to require that the work contain an imprint of the adapter's personality. Common law jurisdictions (such as the U.S. or Canada), on the other hand, tend to have a lower threshold for originality, requiring only a minimal level of creativity and “independent conception.” Some countries approach originality completely differently. For example, Brazil's copyright code protects all works of the mind that do not fall within the list of works that are expressly defined in the statue as “unprotected works.” Consult your jurisdiction's copyright law for more information.
We don’t have the space to go into the different conceptions of adaptation in different jurisdictions, and where the benchmark for adding “original expression” is placed. However, since the U.S. has a low threshold for originality, let's use Title 17 Section 101 of the Copyright Act
as one example. This defines a derivative work thus:
A “derivative work” is a work based upon one or more preexisting works, such as a translation, musical arrangement, dramatization, fictionalization, motion picture version, sound recording, art reproduction, abridgment, condensation, or any other form in which a work may be recast, transformed, or adapted. A work consisting of editorial revisions, annotations, elaborations, or other modifications which, as a whole, represent an original work of authorship, is a “derivative work”.
A crop certainly constitutes an “abridgment” or “condensation” in the literal sense. Nevertheless, these may not "as a whole" represent "an original work of authorship." So far, then, we seem to have got no definitive answer as to whether a crop constitutes an alteration, and whether an alteration of this kind breaks the no derivative clause.
The problem seems, though, to be mainly one of extent. If we take US copyright as our baseline, what degree of manipulation or modification may "as a whole, represent an original work of authorship"?
When does a crop constitute original authorship?
Let’s take a look at this image, Raphael's rather lovely An Allegory ("Vision of a Knight").
This is made available by the National Gallery
under a CC BY-NC-ND 4.0 licence. Interestingly, the National Gallery states that its terms prohibit "derivatives or edits of images"; the word "edits" would seem to be a somewhat stricter term than that of "adaptation" implied by Creative Commons:
|An Allegory ('Vision of a Knight') about 1504, Raphael. (C) National Portrait Gallery. Reproduced under CC BY-NC-ND 4.0 licence.|
Now let’s crop the image to get rid of that tree at the top and to "zoom in" on the figures:
It's still recognisable as the original; it's an "abridgement" under the Creative Commons definition, but hardly constitutes "original authorship" under the US one. But what if we crop to to form a banner at 16:9 ratio, the scale used by Facebook or many Wordpress headers, for instance:
Now our Vision of a Knight
becomes Vision of Not Very Much at All
. We can also turn this image into Castle Landscape
This seems more like a new work in its own right. There’s a long tradition in art history of categorising (and canonising) works on the basis of its contents. As I write this some of my students
are working on Felebien’s hierarchy of genres
, and would know that a portrait stands (or stood) in higher regard than a landscape or nature painting. Our crop constitutes an adaptation that re-represents the original in a way different to that intended by the original creator.
This is an extreme case, admittedly, and with a canonical work like this many viewers would pick up on the fact that the original lies behind the scenes. However, it’s not hard to think of how less well-known works might be mistreated through cropping.
Consider the case of Flickr. Many amateur photographers who licence their work do so under a SA-NC-ND licence. This brief discussion
on the Flickr community is adamant that cropping constitutes a derivative. It’s not hard to appreciate why they might be defensive.
The size of the Flickr archive is so vast that the chance of someone searching for Creative Commons images stumbling upon any individual photo is small. So if that photo then gets picked up by a major media organisation, is cropped for size, and then “shared alike,” it is the cropped version which would become circulated on social media and appear more prominently under Google Images. The derivative work would become the canonical one by virtue of the power of the republisher.
While we might like to think that a gentle crop would not be viewed as a problem by most authors licencing their work under Creative Commons, and indeed many would no doubt be proud to see their work republished elsewhere in any form, we cannot get around the fact that some in the Flickr community would and do see this as a breach of the licence. They may be unlikely to take us to court and test the extent of adaptation under local laws, which is the arbiter according to the Creative Commons FAQ, but for publishers like READ we ignore these concerns at our peril: we are hoping to engage with the public, not alienate them.
Let’s now throw another problem into the mix. When publishing on the web, most publishers will syndicate their work elsewhere via social media. Wordpress’s publicise
feature, for instance, automatically republishes a snippet of a post to the likes of Twitter, Facebook, Linked In etc. In the process, the images attached to a post may be “transformed,” sometimes quite significantly.
On Facebook, for instance, the image will be cropped as per my first modification above, to 16:9. The original image can still be found by following the link back to the blog, but the original is not there on Facebook's servers.
Here it is on Twitter:
Again cropped, although unlike in Facebook it is still possible to view the original directly by clicking on the link, so the original itself has not been altered, only masked in some way (perhaps the analogy here is a viewer in a gallery walking around looking at images through an envelope sized slot).
Although not easily implemented at present, we can also imagine a future where blogs are syndicated to Instagram, run through automated filters, so the above scenarios could become even more radical. And we are on the cusp of virtual and augmented reality desktops, where the presentation of "our" content falls decreasingly under our control.
A further thing to consider is the use of responsive images
in efficient web design. In simple terms, this means that your website serves up the type of image appropriate to the device it is being viewed on. So to a visitor using a widescreen retina display on a broadband-connected Mac, your site might serve up a beautiful high-resolution landscape image. To a visitor browsing on a mobile with a 3G connection, your site might serve up a portrait image cropped to the salient elements.
This might be done by providing two separate images, and coding so that the rendering agent decides which image is most appropriate to load. In this case, a human designer will presumably have looked at and edited the images, and been able to take into consideration derivative rights. However, an automated approach can be provided by adaptive image programming which could have different and unexpected implications for the representation of the original image.
I think it’s highly unlikely that an image author or rights owner would challenge these cases since they are within the spirit, if not necessarily the letter, of the licence. Nevertheless, it is important for the integrity of Creative Commons that these nuances continue to be debated. Coming back to Molly Kleinman’s much referenced page
, it is notable that someone has asked exactly the question about cropping, yet nobody has responded.
Hopefully this offers some way of thinking through the issues, even if I have not the expertise to arrive at a firm conclusion. If you have any thoughts, please share below or via twitter