The Expressive Pattern of Laughter

Laughter as a vocal expressive-communicative signal is one of the least understood and most frequently overlooked human behaviors. The chapter provides an overview of what we know about laughter in terms of respiration, vocalization, facial action, and body movement and attempts to illustrate the mechanisms of laughter and to define its elements. The importance of discriminating between spontaneous and contrived laughter is pointed out and it is argued that unrestrained spontaneous laughter involves inarticulate vocalization. It is argued that we need research integrating the different systems in laughter including experience. Laughter is a conspicuous but frequently overlooked human phenomenon. In ontogenetic development it emerges, later than smiling, around the fourth month; however, cases of gelastic epilepsy (from Greek; gelos = laughter) among neonates demonstrate that all structures are there and functional on date of birth. Further evidence for its innateness comes from twin studies as well as from the fact that laughter is easily observable among deaf-blind children (even among deaf-blind thalidomide children, who could not "learn" laughter by touching people's faces). Man is not the only animal that laughs. Like smiling, laughter has its equivalent in the repertoire of some nonhuman primates. Beginning with Darwin (1872), many writers have been struck by the notable acoustic, orofacial, and contextual similarities between chimpanzee and human laughter. Especially among juvenile chimpanzees a "play-face" with associated vocalization was noted to accompany actions such as play, tickling, or play-biting (Preuschoft, 1995; van Hooff, 1972). 1. Laughter as an Inarticulate Utterance Laughter is estimated to be about 7 million years old (Niemitz, 1990). There is disagreement on how human speech developed phylogenetically. It could have originated from non-verbal vocal utterances, a prelingual gestural system, or sounds initially used to supplement the facial channel. However, it is safe to assume that laughter –like other utterances such as moan, sigh, cry, groan, etc.– was there before man developed speech and served as an expressive-communicative social signal. Considering the stage in the evolution of voice when laughter emerged allows us to make several deductions about the nature of the sound, its generation and the cerebral organization of laughter. If laughter was part of the human vocal repertoire before the speech centers were developed, it is likely that the sound was generated almost exclusively by laryngeal modulations, modified to some degree by supralaryngeal activity but not by articulation. This is because articulation requires voluntary control over the vocal system. Thus the production of speech sounds needs the coordination of respiration, phonation, resonance, and articulation, an analysis of laughter will involve mainly the consideration of the first three. But is it consistent with current knowledge to hypothesize that a laugh-pulse –a vocalization segment initiated by an aspirated "h" type sound followed by the utterance of one of several vowel sounds that is then abruptly terminated– is an inarticulate sound? Indeed, the /h/, an voiceless fricative glottal sound, is the only consonant produced at the level of the larynx; but as laughter is not "speech" we should not expect phonemes of a language to be uttered. Likewise, the assumption of resting articulators would suggest that the laughter "vowel" (but also the sound for other prelingual utterances, such as moaning or utterances of astonishment) is the central vowel schwa, or /e/ (as in about). For the production of this "neutral vowel" one needs to open the mouth and lower the jaw, but keep all other articulators passive. We know from everyday experience that there is a tremendous variety in the quality of the laughing sounds, so obviously deviations from the schwa occur. This does not contradict the hypothesis of an inarticulate vocalization per se but puts forward the task of defining the deviations from the resting articulation position that are due to altered emotional states and separate them from those supralaryngeal conditions modulating the sound that are due to voluntary actions. A second factor that we need to take into account is that there are different ways to generate laughing sounds and it is possible to intentionally modulate the emotion-driven spontaneous laughter. As the cerebral organization of vocal behavior progressed phylogenetically after these early facio-vocal signals became hardwired, higher centers obtained control over the laughter "timer". Thus, in addition to laughing spontaneously (emotional laughter), we can laugh voluntarily or on command (contrived or faked laughter), and we can even speak or sing the laughing sounds, which typically are phonetically represented as "ha ha ha." These forms of utterances differ in degree of volitional control and –inversely– emotionality, and imply different neural pathways and systems involved (for reviews of neuronal control of vocalization see, for example, Jürgens, 1998; Ploog, 1986; for reviews or models of laughter: Arroyo et al., 1993; Fried et al., 1998). The distinction made between spontaneous and voluntary laughter is consistent with clinical observations; certain patients with degenerative brain disorders may be unable to move their abdomen voluntarily, but demonstrate vigorous expiratory movement of the abdomen when laughing spontaneously (Bright et al., 1986). Finally, it should be noted that there are also voluntary attempts to regulate spontaneous laughter; for sake of brevity the combination of voluntary and spontaneous in regulated laughter is only mentioned here. In spontaneous laughter we are following an impulse, an urge to laugh without restraining ourselves. There is no attempt to suppress the response or exert any control over its expression; the laughing person has been described as abandoning himself or herself to the body response (Plessner, 1941). The involuntary aspect can also be seen in the fact that during laughter our self-awareness and self-attention is diminished. Trying to direct attention during a laughter episode stops or reduces laughter. During laughter the state of consciousness is altered; as Hall and Allin (1897; p. 8-9) put it: "[t]he objective world has vanished and is forgotten, the proprieties and even the presence of others are lost, and the soul is all eye and ear to the one laughable object. Care, trouble, and even physical pain are forgotten, and the mind, as it were, falls back through unnumbered millennia and catches a glimpse of that primeval paradise where joy was intense and supreme." While descriptions of emotional experience are rather sparse and reports of subjects mainly refer to the awareness of the impulse to laugh, spontaneous laughter is clearly enjoyable. In voluntary laughter we may want to produce a sound pattern like that of natural laughter. A typical everyday situation demanding faked laughter is when we want to signal somebody that we enjoyed a humorous message and join the laughter of others but actually do not feel any enjoyment. While voluntary laughter may pass over into involuntary laughter, we can not voluntarily produce emotional laughter. Interestingly, research has shown that contrived and spontaneous laughter within a person are strikingly similar with respect to the respirational pattern (suggesting a "laughter signature"); however, it is doubtful whether this is also the case for supralaryngeal systems. The many ways in which they differ still need to be described; this will be especially important when studying individual differences as laughter gets socialized during ontogenetic development. Research has made use of the fact that we can voluntarily alter key features, and thus studies of laughter at different vowels, pitch, and voice quality exist (Citardi et al., 1996; Habermann, 1955). Laughing on command is embarrassing to some, and thus the results obtained for contrived laughter are of limited value for describing spontaneous laughter as it may merge voluntary movements with unintended affective states. Speaking or singing "ha ha" allows modulating the sound in many ways (e.g., stretching the length of vowels, emphasizing particular syllables), just as with other spoken words. As with voluntary laughter we intend to produce a particular sound pattern and we have even more control over the outcome; the major difference being that phonation is not based on forced breathing but on well dosed air supply resulting in less tracheal resonance, breathiness, and aspiration. Spoken laugh sounds may be devoid of emotional intonation; thus there will be less variability in melody, and the utterance will be more clearly articulated and often match the phonemes of a language. Emotional laughter and speech are rather independent; in fact, laughter and speech may co-occur, thereby lengthening the duration of laugh-pulses (Nwokah et al., in press). It remains to be verified that damage of the higher speech centers has little impact on the expression of tickle-induced emotional laughter. 2. Description of Laughter Laughter is not a term used consistently, nor is it precisely defined in research articles and even encyclopedias. In everyday life a smiling face is often referred to as "laughter" although the vocal element is entirely missing. Research studies usually focus on one or two systems of laughter rather than the integration of all the components involved. For example, a study of the acoustics of laughter typically restricts the investigation to the phonatory system and may arrive at the conclusion that the duration of laughter is, say, below two seconds. Sound production is contingent on air flow and the deviation from resting breathing exceeds the rhythmic forced exhalations underlying laughter by far; therefore duration estimates of eight seconds are not uncommon in respiration studies (e.g., Habermann, 1955). In studies of primate laughter the face gets most attention and also studies of joy in humans typically focus on the face, or even on the emotion-specific actions neglecting the mouth opening altogether. The analysis of videotaped laughter suggests a mean duration of five seconds while the study of facial muscle contractions exceeds this value due to the offset period where facial changes are barely visible (Ruch, 1990). 2.1. Laughter Segmentation Laughter bout. The term laughter bout is used here to refer to the whole behavioral-acoustic event, including the respiratory, vocal, and facial and skeletomuscular elements. Prototypically, a laughter bout may be subdivided into an onset (i.e., the pre-vocal facial part which in explosive laughter is very short and steep); an apex (i.e., the period where vocalization or forced exhalation occurs), which –in sustained laughter– might be interrupted by inhalations; and an offset (i.e., a post-vocalization part; usually a long-lasting smile fading out smoothly). Laugh cycle and laugh-pulse. The laughter vocalization period is composed of laugh cycles, i.e., repetitive laugh-pulses (Moore & von Leden, 1958) interspersed with pauses. There is laughter with only one or two pulses ("exclamation" laughter, "chuckle"; Nwokah et al., 1993), but studies typically report a mode of four pulses in a laugh cycle (Provine & Yong, 1991; Rothgänger et al., 1998). The upper number of pulses in a laugh cycle is limited by the lung volume, and different studies give numbers between 9 and 12 (Boeke, 1899; Provine & Yong, 1991); a laughter episode –two or more laughter bouts separated by inspirations– will have more. Table 1 presents an attempt to integrate empirical findings from studies of different systems involved in laughter. Obviously the same phenomenon was studied and converging information about the rhythmic pattern of laugh-pulses can be extracted from studies of air flow and pressure, respirational and laryngeal muscles, acoustics, and the oscillations in the pressure a finger puts on a plate. During intense laughter, the laugh-pulse would be extractable from different parts of the body, as the massive respiration movements cause vibrations of the trunk and limbs to occur which may be detected by an acceleration sensor on the body surface. Table 1. The laugh-pulse as studied in respiration, acoustics, and body movement Author(s) variable studied number of duration (in ms) pulses IPI pulse IPP Respiration Bloch, Lemeignan and Aguilera (1991) rectus abdominis muscle (EMG) 4.17/s 240 Hoit, Plassman, Lansing and Hixon (1988) rectus abdominis muscle (EMG) 5-6/s Air pressure Agostoni, Sant'Ambrogio and Portillo Carrasco (1960) gastric air pressure (swallowed balloon) 5-6 cycles/s thoracic air pressure (swallowed balloon) 5-6 cycles/s Schroetter (1925) exhaled air in closed system up to 5/s Laryngeal activity Moore and van Leden (1958) vocal fold vibrations (ultra high speed camera) 30-100 Luschei, Ramig, Baker and Smith (1997) thyroarytenoid muscle (EMG) 5 Hz cricothyroid muscle (EMG) 5 Hz posterior cricoarytenoid muscle (EMG) 5 Hz Acoustics Boeke (1899) "ha"-sounds/utterances 4.12/s 243 60 183 Bickley and Hunnicutt (1992) laughter syllables 204 97 167 laughter syllables 224 68 156 Nwokah, Davies, Islam, Hsu and Fogel (1993) laugh events 200 Nwokah, Hsu, Davies and Fogel (1998) events 110 events 130 Mowrer, LaPointe and Case (1987) laugh bursts 5.55/s 180 Provine and Yong (1991) laughter notes 4.76/s 210 75 135 Rothgänger, Hauser, Cappellini and Guidotti (1998) plosives 4.67/s 213 126 87 Body vibrations (sentics) Clynes (1980) transient finger pressure (sentograph) 5.03/s 199 Notes. Authors typically gave either frequency per second or the duration of utterances, or other information which was then transformed into one or both parameters by the author of the present chapter. IPI = interpulse interval; IPP = interpulse pause While so far there is no convincing evidence for Darwin's (1872) claim that the muscles of the limbs are thrown into rapid vibratory movements at the same time as the respiratory muscles, initial support can be found in a sample figure provided by Santibañez and Bloch (1986). In this illustration of the EMG-recordings from the brachioradialis (forearm muscle which bends the elbow) and rectus femoris (anterior muscle of the upper thigh) muscles, an increase in amplitude can be found, and the rhythmic pattern seems to match the one of the abdominal muscles. There are roughly five pulses per second, but the interpulse interval varies as a function of the position of the pulse in a sequence (Boeke, 1899; Provine & Yong, 1991; Rothgänger et al., 1998). Further segmentation of the laugh-pulse can be obtained in the different systems involved in laughter. Boeke, recording his own laughter on an Edison Sonograph, already discovered the existence of pauses between the laughter sound pulses whose length exceeded the duration of the ha-utterances by a factor of two. Due to the dynamics of respiration, the duration of sound pulses decreases sequentially from an earlier to later position, and the duration of the interpulse pause increases. Differences in operational definitions of pulse and interpulse pause add further variations to the estimations of length (see Table 1). Acoustic segmentation of the laugh-pulse. Acoustic analyses of laughter suggest a distinction between the vowel-like utterance nucleus and the preceding aspirated "h"-type sound initiating the sound pulse. As the glottis is open during the interpulse pauses, aspiration continues. Thus, not surprisingly, Provine and Yong (1991) report that laughter played backwards still sounds like "ha ha". Vibratory movements in a laugh-pulse. At the level of laryngeal movements a laugh-pulse can be further split up into the number and duration of vibratory cycles of the vocal cords, and even further into their contour; i.e., the phases of when vocal cords are opening, closing, or closed. Using ultra high speed motion picture photography (4000 exposures a second), Moore and van Leden (1958) found in their analysis of the vibration of the cords during a man's laughter a range of 5 to 15 cycles of opening and closing in laugh-pulses. Accordingly, the duration of laughpulses varied from 30 to 100 ms. Each vibratory cycle interrupts the air column ascending from the lungs and the rate of these cycles is the basis for the fundamental frequency of the sound. Further segmentation may be achieved by analyzing the time when vocal cords are opening, closing, or closed, and this vibratory curve contour reflects the dynamics of respiration as it undergoes a progressive change even within one laugh-pulse. These modifications in the vibratory pattern codetermine how laughter sounds and may be a key factor in distinguishing types of laughter. 2.2. Laughter Respiration A respiration cycle consists of inspiration, inspiration pause, expiration, and expiration pause. No matter where in a respiration cycle a person is, laughter typically begins with an initial forced exhalation, followed by a more or less sustained sequence of repeated expirations of high frequency and low amplitude, which may or may not be phonated as "ha ha ha;" i.e., the laugh cycles. While in the case of sustained laughter the expiratory phases will be interrupted by inhalations, there is no evidence for Darwin's (1872; p. 199) assertion that "... [t]he sound of laughter is produced by a deep inspiration followed by short, interrupted, spasmodic contraction of the chest, and especially of the diaphragm." No inspiration preceding the laugh is necessary as laughter is produced at a low lung volume (Bright et al., 1986). Normally the laugh cycles are initiated around functional residual capacity (FRC; i.e., at the lung volume after a normal expiration) and terminate close to residual volume (i.e., the air volume remaining in the lung after maximal expiration), or sometimes even exceed the level of maximal voluntary exhalation (Bright et al., 1986; Lloyd, 1938). Thus, most likely the initial forced exhalation is expelling the tidal volume, and the following sequence of laugh-pulses is based on the expiratory reserve volume. The increase in depth of respiration –the amplitude during laughter may be up to 2.5 times higher than during resting respiration– is therefore due to the stronger expiration; inspiration may add to the amplitude in case of laugher episodes, where single deep inhalations intersperse the expiratory sequences. The rhythmic laughter respiration pattern is produced by saccadic contractions of auxiliary expiration muscles; i.e., muscles that are typically passive during normal expiration, such as the diaphragm (Agostoni et al., 1960), the abdominal (rectus abdominis; Hoit et al., 1988; Santibañez & Bloch, 1986) and the rib cage muscles (triangularis sterni; De Troyer et al., 1987). Of the three muscles mentioned, only the diaphragm is involved in resting breathing in humans; its contraction causes inspiration. The role of the diaphragm is not entirely clear, however, as the discharges in the EMG recordings (albeit parallel to the air pressure) may be indicating reflex contractions due to the passive distention occurring as a function of the violent contractions of the ribcage and abdominal muscles. The triangularis sterni is passive during quiet breathing but involved in different active respiration maneuvers; i.e., respiration below FRC. It contributes to the deflation of the rib cage during active expiration such as in coughing and its neural activation is largely coupled with that of the abdominals. The relative contribution of rib cage and abdomen to the volume may vary even within one laughter cycle (Bright et al., 1986; Habermann, 1955). The respiratory muscles function in concert with the larynx; while without any closing of the glottis there may be single or a few forced exhalations, the adduction (closing) prevents the air to be exhaled too quickly, and allows the building up and maintaining of subglottal air pressure. The initial forced exhalation increases the transdiaphragmatic air pressure by about 5440 Pa (Agostoni et al., 1960) to about 6120 Pa (Schroetter, 1925); this pressure plateau is maintained and forms the basis for the sustained period of phonation of the laugh utterances. The heightened pressure makes the air stream up the airways through the larynx where the rhythmic closing and opening of the glottis interrupts the air stream. These vibrations are carried through the vocal tract whose shape amplifies or dampens certain frequency spectra, and finally the air escapes through the mouth or the nose.