Research shows that significant discrepancies exist between self-report, non-verbal expressions of pain, and evidence of tissue damage, reflecting the impact of some of these criteria, in children (Doherty et al., 1993) and in adults (Craig et al., 1992). Prkachin et al. (1994) found that self-report, nonverbal expression and observers’ judgments were in agreement when the pain was severe, but that observers had difficulty judging accurately a sufferer’s inner state when the pain was submaximal, even though evidence was manifest in the face. There is no guarantee that pain expression will be detected by the observer or that the observer will be able to draw accurate conclusions about the state of the sufferer.

It has been suggested that this inconsistency reflects the multidimensional nature of pain and human ability to detect only selected features at any given time (Hadjistavropoulos & Craig, 2002). For example, it is common for clinicians to encounter patients who appear to complain unreasonably, but express minimal nonverbal pain, and the opposite. Prkachin and Craig (1995) propose that as the sufferer’s experience is expressed, there is a loss of information transfer, causing discrepancies between the different pain indices.

I propose to understand the relationships among the various components of pain by recalling Wittgenstein’s reminder that the word ‘pain’ refers to ‘patterns in the weave of our life’ (Wittgenstein, 1958), and the complexity of this weave explains why the different pain indices are not always in agreement. It also explains why third-person psychological judgments (e.g. clinical reports) are sometimes uncertain. This reflects an indeterminacy which is constitutive of our concept of pain. That indeterminacy in turn is due to communal patterns of behavior: the concept of pain must be flexible and elastic because human behavior, and our reaction to it, is diverse and unpredictable (Wittgenstein, 1980b; 1982; 1992). Given the complexities of the pain experience, it cannot be expected that simple criteria could capture the full range of the experience (Hadjistavropoulos & Craig, 2002; Wittgenstein, 1980b). Hence, the indeterminacy of the pain experience and the challenges associated with characterizing emotional states logically exclude either self-report or nonverbal behavior alone as capable of giving expression to the subtleties involved.

As Wittgenstein (1958) observes, connections between the sensation of pain and reporting pain are highly context-dependent and depend on the criteria used to judge it (e.g. self-report or observational criteria), who is expressing the self-report, the reasons for expressing the self-report, and the person’s understanding of the consequences of reporting pain. Moreover, there is considerable potential for response bias when self-report is used to communicate features of painful experience to others (Anand & Craig, 1996). Since self-report is a fallible source of data (Schwartz, 1999), nonverbal information is often needed and employed for pain assessment (Craig, 1993). Finally, even those who are closely acquainted with a person can make even the most subtle judgments with certainty, without being able to specify conclusive criteria, since their evidence is ‘imponderable’, that is, consists of a complex syndrome of behavior, context and prior events (Wittgenstein, 1958). The constitutional indeterminacy of our concept of pain means that the different pain indices are not typically connected in a rigid way.