Experimental Philosophy of Pain

The standard view of pains among philosophers today holds that their existence consists in being experienced, such that there can be no unfelt pains or pain hallucinations. The typical line of support offered for this view is that it corresponds with the ordinary or commonsense conception of pain. Despite this, a growing body of evidence from experimental philosophers indicates that the ordinary understanding of pain stands in contrast to the standard view among philosophers. In this paper, we will survey this literature and add to it, detailing the results of seven new studies on the ordinary understanding of pain using both corpus analysis and questionnaire methods. The standard view amongst philosophers today is that pains are mental states, and specifically that pains are phenomenally conscious mental states. Although researchers are not in complete agreement about how to understand the concept of phenomenal consciousness, the standard idea is that a being is phenomenally conscious if that being has phenomenally conscious mental states, and that a mental state is phenomenally conscious just in case there is “something it is like” (Nagel, 1974) to be in this state, where this vague phrase is understood to mean that the state has phenomenal qualities (or qualia for short). With regards to pains, the idea is that what makes a state a pain is the way it feels to the being experiencing it. As such, on the standard view there is no appearance–reality distinction to be drawn for pains. To have the appearance of pain (the phenomenal experience of pain) is for there to be a pain, and vice versa. In other words, the standard view of pain amongst philosophers today endorses the following two conditionals: If a person has a pain, then she feels that pain. If a person feels a pain, then she has that pain. 1 To appear in a special issue of the Journal of Indian Council of Philosophical Research on experimental philosophy edited by J. Knobe, E. Machery, and S. Stich. 2 These conditionals have a number of implications. Most importantly, the first conditional excludes the possibility of unfelt pains. Since having a pain implies feeling that pain, there could be no pain that is not felt. And, indeed, many prominent philosophers have asserted that this is the case, often treating the very notion of unfelt pains as absurd or oxymoronic. For instance, Murat Aydede (2005a, x) claims that “it is part of the commonsense conception of pains... that they can’t exist without someone’s feeling them.” He goes on to contend that “there is an air of paradox when someone talks about unfelt pains,” claiming that “one is naturally tempted to say that if a pain is not being felt by its owner then it does not exist” (2005b, 4).2 The second condition excludes the possibility of pain hallucinations. Since feeling pain implies that there is a pain that is felt, there could be no feelings of pain without a corresponding pain, and hence no pain hallucinations. And, again, many prominent philosophers have asserted that this is the case. To give but one example, Hilary Putnam (1963, 218) writes that “one cannot have a ‘pain hallucination’... simply because any situation that a person cannot discriminate from a situation in which he himself has a pain counts as a situation in which he has pain.”3 But, should we believe that there is no appearance–reality distinction for pains? Should we deny the possibility of unfelt pains and pain hallucinations? And, more generally, should we believe that pains are mental states (and phenomenally conscious mental states in particular)? The primary reasons that philosophers have offered for the standard view rest not so much on explicit philosophical arguments or empirical data, but rather on appeals to intuitions. Such an appeal is seen in the quotes from Aydede above, calling on “the commonsense conception of pains” and what “one is naturally tempted to say.” Elsewhere he writes that “common sense... 2 See Reuter and Sytsma (ms) for further examples and an extended discussion, as well as Reuter (2016) for a critical challenge to the idea that the standard view can be squared with data from developmental studies. 3 See Reuter, Phillips, and Sytsma (2014) for further examples and an extended discussion. 3 resists identifying a pain with any physical feature or condition instantiated in the body,” asserting that “a quick thought experiment should confirm this”: Suppose that we do in fact attribute a physical condition, call it PC, when we attribute pain to body parts, and that PC is the perceptual object of such experiences. So, for instance, John’s current excruciating experience (call this E) is caused by and represents a physical condition in his right thigh and our ordinary concept of pain applies in the first instance to this condition in his thigh. From this it would follow that... John would have pain if he had PC but no E (as would be the case, for instance, if he had taken absolutely effective painkillers or his thigh had been anesthetized). (2009, 4-5) Aydede then charges that this conclusion is “intuitively incorrect,” writing that it “appear[s] to clash with our ordinary or dominant concept of pain” (5). It is worth clearly noting, here, that Aydede and philosophers like him do not merely make a claim about their own intuitions. Instead, they take it to be part of the commonsense conception of pain, that unfelt pains and pain hallucinations are impossible. But is this correct? Is it true that the commonsense conception of pain corresponds with the standard view in philosophy, denying an appearance–reality distinction for pains?4 A growing body of literature in experimental philosophy has attempted to test this claim.5 The results suggest that the commonsense conception of pain is quite different from the standard view, allowing for the occurrence of unfelt pains and pain hallucinations, and locating pains in body parts rather than in the mind/brain. In this paper we will both survey the empirical work that has been done to date and expand on it. In the first section, we discuss the corpus analysis conducted by Reuter (2011) on the use of pain terms in English, then extend this to the use of 4 If this is correct, a further issue arises: Should we accept the deliverances of our intuitions with regard to the nature of pain? And to what extent should we treat these deliverances as being defeasible and open to revision in the light of theoretical considerations and/or empirical findings? 5 For a brief introduction to experimental philosophy see Sytsma and Machery (2013). For extended introductions see Alexander (2012) and Sytsma and Livengood (2015). For collections of articles see Knobe and Nichols (2008, 2013), as well as the volumes in the Advances in Experimental Philosophy series. For discussions of some recent disputes see Machery and O’Neil (2014) and Sytsma (forthcoming). And for an extensive survey of the state of the art of experimental philosophy see Sytsma and Buckwalter (2016). 4 pain terms in German. In the second section, we survey a wide range of studies using questionnaire methods, in addition to presenting the results of six new studies on judgments about the possibility of unfelt pains and pain hallucinations. 2. Experimental Investigation of the Use of Pain Language The most direct means to test whether people distinguish the appearance of a pain (feeling a pain) from the reality (having of a pain), would be to systematically listen to people speak about their pains. However, this is difficult to do for at least three reasons: first, it would be very time consuming; second, it would be difficult to gather a representative sample; and, third, people often describe their pains in ways that do not unambiguously bear on the appearance–reality distinction. For instance, people seldom make claims like “I don’t feel a pain, but I do have a pain in my stomach” or “I feel a pain in my stomach, but there is no pain.” The first two difficulties can be circumvented by using large linguistic corpora. Linguistic corpora are simply systematic collections of linguistic data—typically drawn from public sources such as newspapers, magazines, and the internet—that can be used for purposes of describing language use and testing hypotheses about languages.6 Since health issues are a central aspect of everybody’s lives and are talked about frequently in various outlets, data on the usage of pain language is available in existing linguistic corpora. To avoid the third problem, Reuter (2011) used an analogy with the traditional sense modalities. Most languages, perhaps all languages, allow people to express appearance statements (e.g., “the shirt looks blue”) and contrast such statements with factual claims (e.g., “the shirt looks blue but it really is black”). Very often, people’s choice of appearance language 6 For an introduction to corpus analysis and a discussion of its merits for and applications in philosophical research, see Bluhm (2016). 5 depends on how certain they feel about a given state of affairs (Quinton 1956, Sellars 1956). If a person feels very confident that the shirt is blue, then the person is more likely make a factual statement such as “the shirt is blue.” When they are less confident, however, such as when the lighting conditions are bad, people are more likely to make appearance statements such as “the shirt looks blue.” How confident people are about certain external affairs, in turn correlates with the perceived intensity of the property in question. If the shirt is perceived clearly in sunlight, people will be highly certain about what is and what is not the case. In a badly lit room at night, people’s confidence suffers from the poor conditions of perception (Lund 1926). Reuter then argues that while the intensity of certain stimuli has an effect on the use of appearance language, if people do not distinguish between the appearance and the reality of pains then no such effect should be found for the use of pain language. If, however, it turns out that people tend to state that they have a pain when the intensity of the pain is high, and to state that they feel a pain when the intensity of the pain is low, then this would indicate that they do in fact distinguish the appearance from the reality of pains. To test this, Reuter analyzed eight different pain attributes of the English language—four low-intensity ones (minor, small, slight, little) and four strong ones (severe, major, bad, big). The frequency of each of these attributes in conjunction with pain statements was established by entering phrases like “I feel a severe pain” into various search engines and comparing the number of hits with the contrasting phrase “I have a severe pain.” The aim of this method was to examine how strongly the intensity of pain influences people’s choice of language. The results show that whereas people use the feeling phrase slightly more often when talking about low-intensity pains, they use the having phrase around three to four times more often when talking about a strong pain. Thus, at least when it comes to the English language, the results support the hypothesis that the intensity of pain has a