Abstract
The recent proliferation of deepfakes and other digitally produced deceptive representations has revived the debate on the epistemic robustness of photography and other mechanically produced images. Authors such as Rini (2020) and Fallis (2021) claim that the proliferation of deepfakes pose a serious threat to the reliability and the epistemic value of photographs and videos. In particular, Fallis adopts a Skyrmsian account of how signals carry information (Skyrms, 2010) to argue that the existence of deepfakes significantly reduces the information that images carry about the world, which undermines their reliability as a source of evidence. In this paper, we focus on Fallis’ version of the challenge, but our results can be generalized to address similar pessimistic views such as Rini’s. More generally, we offer an account of the epistemic robustness of photography and videos that allows us to understand these systems of representation as continuous with other means of information transmission we find in nature. This account will then give us the necessary tools to put Fallis’ claims into perspective: using a richer approach to animal signaling based on the signaling model of communication (Maynard-Smith and Harper, 2003), we will claim that, while it might be true that deepfake technology increases the probability of obtaining false positives, the dimension of the epistemic threat involved might still be negligible.
This is a preview of subscription content, log in via an institution to check access.
Similar content being viewed by others
Notes
This assumption seems to be at work in traditional theories of photography, since, as we mentioned before, they use the Gricean theory of communication to explain the epistemic mark they think photographs bear compared to other pictorial kinds–an epistemic mark that, according to these views, photographs share with other forms of natural meaning. Although we do not have the space to discuss the shortcomings of these views, we think they are partially derived from the sharp distinction Grice made between natural and non-natural meaning and a rather radical interpretation of Gricean views (see Atencia-Linares 2014). However, we think there are more nuanced models of communication that can help us to better account for the epistemic robustness of photography.
Advocates of the signaling model of communication use the terms ‘design’ or ‘designed’ to refer to features or mechanisms that developed or were selected with a particular purpose, but these terms are not associated with the suspicious idea of intelligent design. See Millikan (1989); Neander, (1991).
Mitchell Green helpfully explains three senses in which we can understand what Ekman means by involuntarily (Green 2007, pp. 121–1222).
The information model of animal communication is not uncontroversial (see Dawkins and Krebs, 1978; Owren et al., 2010). Nonetheless, for ease of exposition and taking into account that our main target (Fallis, 2021) heavily relies on the notion of information, we rely here on the standard approach. In any case, we think our arguments could probably be spelled out in manipulationist (non-informational) terms.
Receivers also develop the ability to understand signals. Actually, the selection and evolution of signals depend on the capacity of the signals to affect the receivers’ behavior.
Strategic cost is sometimes differentiated in the literature from efficacy cost. The latter refers to the minimal cost involved in sending the signal successfully, the former refers to an extra cost that has to be incurred in order to prevent fakes.
We will focus on the case of photography to make things simpler, but if our arguments are sound, they should apply equally well to videos. In fact, photographs are arguably easier to fake or manipulate than other mechanically produced pictorial media, so it’s the strongest case we can make. If our case works for photography, it should work better for more complex media involving sound and moving images.
The early instances of these devices produced inaccurate and blurry images, but since accuracy was desired and demanded alternative lenses were developed to correct this problem (Davenport 1999, pp. 4–6).
This does not mean that there have not been other aims partly driving the design of photographic devices, or that the design of photographic mechanisms has always tried to maximize accuracy at all costs. This is certainly not the case. Other aims that were crucial for important developments in the history of the photographic mechanism were, for example, the need to speed the exposure times or the aim of reproducibility or the possibility of obtaining multiple copies. On the other hand, other elements used in photography such as filters and different types of developing papers are partly designed to correct or embellish the scenes or allow different tonalities, grain, and contrast; that is, they are designed with the aim of contributing to the expressive quality of the final image which is frequently at the expense of maximal accuracy. Similarly, the aim of portability and ease of use, together with cost, is often at odds with obtaining maximal accuracy. However, it is plausible to claim that these motivations driving the design are complementary to the overall aim of capturing the appearance of real existing scenes and objects achieving a certain (marginal but high) degree of accuracy. This latter aim is a central aim of the design of most, if not all, photographic mechanisms.
For a detailed account of the development of lenses and how the technology evolved to keep optical aberrations to the minimum and to provide a sharp image that avoids flare and other optical defects see (Kingslake 1989). Kingslake’s book shows that the design of the lenses is the product of technical evolution driven by a social demand for accuracy.
For a general overview of these accounts and their problems see Costello & Phillips 2009.
This work was constructed from parts of more than fifty images, which were then digitally processed.
Here we disagree with traditional theories which associate the nature of photography with the type of meaning (natural meaning) they purportedly convey. For an extensive analysis of why traditional theories of photography have taken a radical interpretation of Grice’s view to arrive at such conclusion see (Atencia-Linares 2014).
From this point onwards, for the sake of simplicity and consistency with the terminology of the signaling model, we will call photographs that reliably capture the appearance of real existing objects or events “honest signals.” In turn, we will call photographs that inaccurately depict objects or events or that depict non-existent objects “fakes”. Dominic Lopes (2016) also uses the term “honest signals” in reference to certain kind of photographs, but we think it originates from the literature of the signaling model of communication or at least, this is the original source from which we draw (see also Atencia-Linares 2014).
This is even more frequent in portable digital cameras.
Indeed, the examples and reasoning provided by Atencia-Linares (2012) as for how photographs can depict fictional entities by photographic means is consistent with this view: since fictional entities cannot reflect light onto the photographic surface–which is the easiest and default mode of depicting existing individuals–in order to depict ficta, the photographer must skillfully manipulate light and use other photographic resources.
Technical skills required are typically proportional to the amount of manipulation an image receives. Think, for example, about Jeff Wall’s A Sudden Gust of Wind (after Hokusai).
Note that, in the same way that the costs associated with producing a fake photograph vary significantly, they also greatly differ in the case of animal signals; just compare faking the tail of a healthy male peacock with producing a false alarm call or releasing a certain pheromone. Accordingly, we do not think this variance is a disanalogy between animal communication and photography. We would like to thank an anonymous reviewer for drawing our attention to this issue.
In 2021, for instance, Kendall Jenner received voiceful criticism of her Twitter followers for publishing an image of herself that was clearly manipulated.
Fallis’ uses ‘S’ to refer to a state of affairs and to the proposition stating that the state of affairs holds. Here it stands only for the former.
It is worth stressing that this is not exactly how Skyrms conceives of information and misinformation (for a discussion, see Godfrey-Smith, 2012). Nonetheless, we don’t think the details matter for present purposes.
The likelihood can be spelled out, as Catherine Abell suggests, in terms of ‘the distance between worlds in which the depictive content of pictures produced by that process is accurate, and worlds in which is not’ (Abell 2010, pp. 86–87).
Indeed, strictly speaking the current situation is not as bad as (a), since at the moment most deepfakes are probably low cost. In any case, for the sake of the argument, we’ll assume (a) as the initial scenario.
Of course, there can be intermediate cases where honest images outnumber deepfakes and the latter are not very costly.
Here we agree with Catharine Abell (2010).
Providing an account of the nature of photography is beyond the scope of this paper. However, nothing we say in this paper commits us to the view that the epistemic mark we describe is a necessary feature of photographs as a type of image.
Although we don’t think, as Abell does, that if digital techniques are used to undermine the epistemic value of photographs the resulting image will then be necessarily non-photographic.
In this sense, our view can be aligned with the moderate version of the New Theory of Photography (Atencia-Linares 2018).
Cohen and Meskin also suggest that the epistemic distinctiveness of photographs is merely apparent (2008).
Very early versions of this paper were presented at the Anglo- German Picture Theory Group Meeting at Dartmouth College, 2018 and at the London Aesthetics Forum Workshop on Photography 2015; a most recent version was delivered at the Philosophy of Digital Images Conference in Liverpool 2022. We are grateful to the participants of all these venues for their comments. Many thanks also to the two anonymous referees who provided insightful comments that helped us to improve the manuscript. Finally, both authors are grateful to each other for the excellent experience of writing this paper together. Part of the research conducive to the ideas developed in this paper was funded by the projects PGC2018-101,425-B- I00, PID2021-127046NA-I00 and PID2019-106420GA-I00 funded by the MCIN/AEI/10.13039/501100011033 and the project CIGE/2021/160 funded by the Generalitat Valenciana.
References
Abell, C. (2010). The epistemic value of photographs. In: Abell, C., & Bantinaki, K., eds., Philosophical perspectives on depiction. Oxford University Press.
Alonso-Alvarez, C., Bertrand, S., Devevey, G., Gaillard, M., Prost, J., Faivre, B., & Sorci, G. (2004). An experimental test of the dose-dependent effect of carotenoids and immune activation on sexual signals and antioxidant activity. The American Naturalist, 164(5), 651–659.
Atencia-Linares, P. (2012). Fiction, nonfiction, and deceptive photographic representation. Journal of Aesthetics and Art Criticism, 70(1), 19–30.
Atencia-Linares, P. (2014). Arts and Facts. Fiction, Non-fiction and the Photographic Medium (Doctoral Thesis. UCL, London, UK). Retrieved from https://discovery.ucl.ac.uk/id/eprint/1434513/
Atencia-Linares, P. (2018). Why grey is the New Black. Aisthesis: Pratiche, Linguaggi e Saperi Dell’estetico, 11(2), 216.
Bird, R. B., & Smith, E. A. (2005). Signaling theory, strategic interaction, and symbolic capital. Current Anthropology, 46(2), 221–238.
Cavedon-Taylor, D. (2013). Photographically based knowledge. Episteme, 10(3), 283–297.
Chalmers, D. J. (2022). Reality+: Virtual worlds and the problems of philosophy. W. W. Norton.
Cohen, J., & Meskin, A. (2004). On the epistemic value of photographs. Journal of Aesthetics and Art Criticism, 62(2), 197–210.
Connelly, B. L., Certo, S. T., Ireland, R. D., & Reutzel, C. R. (2011). Signaling theory: A review and assessment. Journal of Management, 37(1), 39–67.
Costello, Diarmuid (2017). On photography: A philosophical inquiry. Routledge.
Costello, D., & Phillips, D. M. (2009). Automatism, causality, and realism: Foundational problems in the philosophy of photography. Philosophy Compass, 4(1), 1–21.
Currie, G. (1999). Visible traces: Documentary and the contents of photographs. Journal of Aesthetics and Art Criticism, 57(3), 285–297.
Davenport, A. (1999). The history of photography: An overview. University of New Mexico Press.
Krebs, J. R. & Dawkins, R. (1978). Animal signals: Information or manipulation. In: J. R. Krebs & N. B. Davies, eds., Behavioural ecology: An evolutionary approach, first edition (pp. 282–309). Blackwell.
Ekman, P. (1997). Expression or communication about emotion. In: N. L. Segal, G. E. Weisfeld, C. C. Weisfeld, eds. Uniting psychology and biology: Integrative perspectives on human development, first edition (pp. 315–38). American Psychological Association (APA).
Fallis, D. (2021). The epistemic threat of deepfakes. Philosophy and Technology, 34(4), 623–643.
Fineman, M. (2012). Faking it: Manipulated photography before photoshop. New York: Metropolitan Museum of Art ; Distributed by Yale University Press.
Fromhage, L., & Henshaw, J. M. (2022). The balance model of honest sexual signaling. Evolution, 76(3), 445–454.
Godfrey-Smith, P. (1996). Complexity and the function of mind in nature. Cambridge University Press.
Green, M. S. (2007). Self-expression. Oxford University Press.
Godfrey-Smith, P. (2011). Signals: Evolution, learning, and information, by brian skyrms. Mind, 120(480), 1288–1297.
Greene, H. W., & McDiarmid, R. W. (1981). Coral snake mimicry: Does it occur? Science, 213(4513), 1.
Grice, H. P. (1989). Studies in the way of words. Harvard University Press.
Harris, K. R. (2021). Video on demand: What deepfakes do and how they harm. Synthese, 199(5–6), 13373–13391.
Hauser, M. D. (1992). Costs of deception: Cheaters are punished in rhesus monkeys (Macaca mulatta). Proceedings of the National Academy of Sciences United States of America, 89(24), 12137–12139.
Hicks, R. (2003). Darkroom basics and beyond. Collins & Brown.
Higham, J. P. (2014). How does honest costly signaling work? Behavioral Ecology, 25(1), 8–11.
Hopkins, R. (2012). Factive pictorial experience: What’s special about photographs? Noûs, 46(4), 709–731.
Hurd, P. L. (1995). Communication in discrete action-response games. Journal of Theoretical Biology, 174(2), 217–222.
KODAK. (2013). KODAK: Teaching basic darkroom techniques. Accessed July 5. http://www.kodak.com/global/en/consumer/education/lessonPlans/darkroom/combination.shtml
Langford, M. J. (1981). The Darkroom handbook. Ebury Press.
Lopes, D. M. (2015). Four arts of photography. Wiley.
Maynard-Smith, J. (1991). Honest signalling: The Philip Sidney game. Animal Behaviour, 1, 1.
Maynard-Smith, J. (1994). Must reliable signals always be costly? Animal Behaviour, 47(5), 1115–1120.
Smith, J. M., & & Harper, D. (2003). Animal signals. Oxford University Press.
Millikan, R. G. (1989). In defense of proper functions. Philosophy of Science, 56, 288–302.
Meskin, A., & Cohen, J. (2008). Photographs as evidence. In S. Walden (Ed.), Photography and philosophy: Essays on the pencil of nature. Blackwell.
Neander, K. (1991). Functions as selected effects: The conceptual analyst’s defense. Philosophy of Science, 58(2), 168–184.
Ortwein, A. (2013). Guide to film photography - Basic film photography guide for beginner film photography tips, techniques, and information. Accessed July 5. http://www.guidetofilmphotography.com/
Owren, M., Rendal, D., & Ryan, M. J. (2010). Redefining animal signaling: Influence versus information in communication. Biology and Philosophy., 25, 755–780.
Palmer, D. (2014). Mobile media photography. In The Routledge Companion to Mobile Media (pp. 245–255). Routledge.
Penn, D. J., & Számadó, S. (2020). The Handicap Principle: How an erroneous hypothesis became a scientific principle. Biological Reviews, 95(1), 267–290.
Riechert, S. E. (1978). Games spiders play: Behavioral variability in territorial disputes. Behavioral Ecology and Sociobiology, 3(2), 135–162. https://doi.org/10.2307/4599163
Rini, R. (2020). Deepfakes and the epistemic backstop. Philosophers’ Imprint, 20(24), 1–16.
Rinn, W. E. (1984). The neuropsychology of facial expression: A review of the neurological and psychological mechanisms for producing facial expressions. Psychological Bulletin, 95(1), 52–77.
Savedoff, B. (2008). Documentary authority and the art of photography. In S. Walden (Ed.), Photography and philosophy (pp. 111–137). Blackwell.
Searcy, W. A., & Nowicki, S. (2005). The evolution of animal communication: Reliability and deception in signaling systems. Princeton University Press.
Spence, M. (1973). Job market signaling. Quarterly Journal of Economics, 87, 355–374.
Spence, M. (2002). Signaling in retrospect and the informational structure of markets. American Economic Review, 92, 434–459.
Svensson & Wong. (2011). Carotenoid-based signals in behavioral ecology: A review. Behaviour, 148(2), 131–189.
Számadó, S. (2011). The cost of honesty and the fallacy of the Handicap Principle. Animal Behaviour, 81(1), 3–10.
Skyrms, B. (2010). Signals: Evolution, learning, and information. Oxford University Press.
Walden, S. (2005). Objectivity in photography. British Journal of Aesthetics, 45(3), 258–272.
Walton, K. L. (1984). Transparent pictures: On the nature of photographic realism. Critical Inquiry, 11(2), 246–277.
Wiley, R. H. (1994). Errors, exaggeration, and deception in animal communication. In L. Real (Ed.), Behavioral mechanisms in ecology (pp. 157–189). Chicago: University of Chicago Press.
Zahavi, A., & Zahavi, A. (1997). The Handicap Principle: A missing piece of Darwin’s puzzle. Oxford University Press.
Acknowledgements
Very early versions of this paper were presented at the Anglo- German Picture Theory Group Meeting at Dartmouth College, 2018 and at the London Aesthetics Forum Workshop on Photography; a most recent version was delivered at the Philosophy of Digital Images Conference in Liverpool 2022. We are grateful to the participants of all these venues for their comments. Many thanks also to the two anonymous referees who provided insightful comments that helped us to improve the manuscript. Finally, both authors are grateful to each other for the excellent experience of writing this paper together. Part of the research conducive to the ideas developed in this paper was funded by the projects PGC2018-101425-B- I00, PID2021-127046NA-I00 and PID2019-106420GA-I00 funded by the MCIN/AEI/10.13029/501100011033 and the project CIGE/2021/160 funded by the Generalitat Valenciana.
Funding
Partial financial support was received from the following funding bodies/research projects Ministerio de Ciencia e Innovación, Spain Spanish Government.
Author information
Authors and Affiliations
Contributions
Both authors listed made substantial contributions to the conception or design of the work. PA-L, conceived and designed the main idea of this paper. However, both authors contributed to the overall research and analysis. The first draft of the manuscript was written by PA-L but Marc Artiga developed and wrote additional parts of subsequent drafts. Both authors commented on previous versions of the manuscript. All authors read and approved the final manuscript,
Corresponding author
Ethics declarations
Competing interest
We confirm that none of the authors of this paper have any financial or non-financial interests that are directly or indirectly related to the work submitted for publication.
Consent for publication
Both authors agree with the content of this publication and give explicit consent to its submission. Both authors obtained consent from the responsible authorities at the institute/organization where the work has been carried out, before the work is submitted. Both authors agree to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved.
Additional information
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Rights and permissions
Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.
About this article
Cite this article
Atencia-Linares, P., Artiga, M. Deepfakes, shallow epistemic graves. Synthese 200, 518 (2022). https://doi.org/10.1007/s11229-022-04003-3
Received:
Accepted:
Published:
DOI: https://doi.org/10.1007/s11229-022-04003-3