Studies using the Facial Action Coding System (FACS; Ekman & Friesen, 1978) reveal a combination of facial expressions that appear to be specific to acute pain and that reveal the connection between pain facial expression and pain (Craig, 1980; Prkachin, 1992; Prkachin and Solomon, 2008). Core action units (AUs) in adults are brow lowering (corrugator: AU 4), cheek raise and lid tighten (both parts of orbicularis oculi: AU6 and 7), nose wrinkle and upper lip-raise (both parts of levator labii: AU 9 and 10), and eye closing (AU43) (Craig et al., 1992). It has been suggested that pain facial expression is a combination of the core actions along with a limited range of other actions; an expression that is better described as an indeterminate set than a fixed prototype (Prkachin & Craig, 1995). In naturalistic settings, we find it easier to describe a person as ‘in pain’ than to describe his facial features or behavior in precise physical terms, and we do not typically infer psychologically relevant descriptions of pain from austere physical ones (Wittgenstein, 1980a, 1992). For we often know the conclusions of such alleged inferences without knowing their premises (Wittgenstein, 1980a).

Objections to the specificity of the pain face suggest that it forms part of a whole body defensive response to pain, an effort to withdraw from the stimulus and to protect the body. Salzen (2002) claims that pain facial expression consisting of tension in the facial flexor muscles, contraction of the eyes, retraction of the lips, and clenched teeth is part of a general body flexor reaction. This is based on the idea that pain facial expression is a ‘sensory-motor feeling state’, which is ‘aversive’, but separate from secondary distress due to the continuation of pain despite the initial response. Pain facial expression seems to be part of a global flexor contraction of writhing or squirming, which may spill over into the facial musculature (Frijda, 2002; Panksepp and Pasqualini, 2002). One difficulty facing this challenge is that no such whole body behaviors have been described that are specific to pain in humans or in domestic or laboratory animals (Williams, 2002).

Moreover, given that the face is the primary target of visual attention between humans and provides a dynamic, embodied expression of pain and emotion in human interactions (Cole, 1997, 2001), some justification is required for collapsing pain facial expression into gross motor activity in connection to stimuli (Williams, 2002). Wittgenstein (1958b) makes a related point that if a person has hurt his hand, he may nurse his hand, but we comfort him: ‘…if someone has a pain in his hand, then the hand does not say so (unless it writes it) and one does not comfort the hand, but the sufferer: one looks into his face’ (Wittgenstein, 1958b, §286). It is not the body that is in pain, but the human being. Available evidence shows that the face in pain is highly salient for observers who ranked the eyes the most important feature, followed by brows, eyelids, mouth, head, forehead, and then other body parts (Prkachin et al., 1983).

Closely related to the question of a specific pain facial expression in the individual is its detection and interpretation by observers. Judgment of pain in another person relies heavily on facial cues: brow lowering, eye blinking, cheek raise, and upper lip raise are used consistently by observers to judge pain in adults and in children (Craig et al., 1991). There is evidence of reasonably accurate identification of pain expression in adults and infants employing these facial cues (e.g., Prkachin et al. 1994).

The overlap between pain and other emotions concerning the activated facial action units, however, challenges the evidence for the detection of the pain face. After all, when people are in pain, their faces may express a blend of emotions including fear, anger, disgust, surprise, and so on, reflecting a ‘fuzzy’ emotional state of distress or discomfort, rather than expressing the specific experience of pain. It does not follow, however, that sharing of AUs across expressions renders observers unable to distinguish anger from fear or sadness. After all, detection and interpretation only have to be effective, not perfect (cf. Wittgenstein, 1958b, §79d). This may support the claim, made earlier, that the pain face expression is accurately described as an indeterminate or ‘fuzzy’ set than a fixed prototype.