User:Johan (WMF)/Images


Personal notes on images in articles. Draft.

Images belong in a context[edit]

Just looking at the main or only image when searching around on Swedish Wikipedia, a quick, not very structured investigation indicated that around 20 percent had significant information issues when the image caption in the article was taken away. This seems very high, and a more thorough process would probably yield a lower number, but it says something about the problem. This could be for a variety of reasons.

  • One image of a castle and its surroundings was in colour, but the photo was from the 1910s, and the surrounding structures looked very dated. Without that information, this photo will give a directly false impression about the area, because the reader would assume it’s from a later date.
  • An opera was illustrated by a photo of the composer: without the caption, it would not have been obvious what the connection between the subject and the photo was.
  • The article on Arminianism was illustrated by the synod that rejected it.

Images are often picked because they can be contexualised; without that contextualisation, some of them lose most or all of their information value and might become problems rather than additions. Another problem is that an image might illustrate a specific part of an article, one which is not present in an article in another language.

Images and neutrality[edit]

Speaking for myself, not for the larger Wikimedia community.

A much used photo of Seljalandsfoss. What is the information value of the focus on the colours of the surroundings in the sunset? What does it say?

There’s a subjectivity in an image beyond neutrally explain the dimensions of a setting. It says so many things we never intended to say. An image of a place always asks: Does this look nice? Does it look pleasant? A photograph of a person always asks us for judgements, based on looks. On an encyclopedia with neutrality it its core, photographs on Wikipedia has not quite managed to free themselves of the demands for beauty. In the choice between two otherwise equivalent photos, we tend to go with the one that looks best because it’s part of the concept ”good picture”. There’s a connection between the photo and its creator that’s, in most cases, far stronger than between the article and its writers. An article isn’t my article when fifty other persons have added their pieces; an image remains mine. Where there’s a personal connection there’s personal pride and where there’s pride you want to show something that looks well. And as long as photograph hasn’t been too obviously enhanced – if the photographer just knows their equipment and was lucky with light and angle and then did some editing within the Wikipedia norms – it will never be edited for neutrality, even if the textual equivalent had been changed to give a far more muted expression.

And our wish to illustrate will tell stories. To mention just one way this can interfere with the information we want to give, when choosing pictures for places, we tend to go with the specific rather than the typical, the thing that sets it apart rather than how it looks. There’s a good argument for this – most Swedish small towns look very similar – yet it creates a false impression of a place, not least for readers who don’t know what the typical architecture of that area might look like even if they make a conscious effort to envision it. Go to Uppsala and it’s completely dominated by the cathedral, with a smattering of other historical buildings: the castle, the old university main building, the anatomical theatre, the botanical gardens, the old train station and so on. Very little from the 20th or 21th century. Or the way the Swedish article on Beijing used to have one modern photo – of the Tianmen, a historical building, giving the distinct impression that this city has no modern characteristics.

We do similar things for a lot of articles on people and human activity. And we don’t think too much about it, because the rhetoric of the image – neither alone nor together with other photos – is not taken into full consideration the same way we do with text.

Further reading and watching[edit]

  • Jönsson, Johan. Images and NPOV: A Shaky Relationship, presentation at Wikimania 2016
  • Jönsson, Johan. "Staden bilden och uppslagsverket", KRITIK #35 (2017)
  • Jönsson, Johan. "Wikipedia och skönheten", conference proceedings, Ad pulchritudinem 2019