Henning, Michelle2024-03-012024-03-012021https://www.degruyter.com/document/doi/10.14361/dcs-2021-070203/htmlhttps://mediarep.org/handle/doc/23212This paper considers emotion recognition and sentiment analysis in relation to social media photographs. It addresses this as part of a larger regime of surveillance and control, in which photographs are treated as symptoms for a diagnosis, and are quantified as data. Auto- mated emotion recognition approaches are capable in principle of analysing the visual qualities of social photos insofar as these can be measured and represented numerically. In reducing the photograph to data, they select out features of the image, as a means to explain or describe a mental state that lies behind or beyond the image. To treat photographs as emotionally expressive goes against the historical idea of the photograph as objective recording. Originally, the idea that pho- tographs could move their viewers was linked to the sense of photog- raphy as detached documentation. Today, more and more people take and share photographs as part of a larger shift in emotional culture, which places a therapeutic sense of self at the heart of economy and governance. Yet while people use mobile phone photos as a means of expressive documentation and self-representation, emotion recogni- tion relies on a behaviourist and positivist model that is indifferent to their intentions and to culture, and which is premised on a myth of total knowledge.engCreative Commons Attribution Non Commercial No Derivatives 4.0 GenericPhotographySocial MediaEmotion RecognitionBehaviorismEmotional Capitalism70030010.14361/dcs-2021-07020310.25969/mediarep/218722364-2114