Importantly, social dominance selection did not enhance or reduce the other forms of aggressiveness. Indeed, the degree of male red coloration accounted for most of the variation in dominance ability. This suggested that at least in the three-spined stickleback, adult intermale dominance is not closely related to either juvenile aggressiveness or adult territorial aggressiveness.
To summarize, while Bakker (1994) did not elucidate the underlying neurobiological mechanisms of aggressiveness in fish, he provided strong evidence that both juvenile aggressiveness and territorial aggressiveness were strongly influenced by genetic selection, with territorial aggressiveness perhaps being more closely linked to the RAGE/Anger system than juvenile aggression or dominance. While we are not aware of selection studies targeting fish SEEKING or FEAR systems, we would predict that such studies would confirm genetic foundations in fish for those temperament aspects of emotions as well, as has previously been well demonstrated in rodents.
ARE FISH SOCIAL ANIMALS?
Are fish personalities really just defined by SEEKING, RAGE, and FEAR sensitivities, or do fish exhibit social behavior that represents other possible personality dimensions, such as PLAY, as already noted, which may utilize many of the other emotional systems in a nonserious way? Most of the 24,000 species of teleost fish (the vast majority of bony fishes) swim together at times in groups commonly called schools, which biologists refer to as shoals, when the swimming becomes more synchronized (Faucher, Parmentier, Becco, Vandewalle, & Vandewalle, 2010). A key question is whether fish swim in groups because of the same social motivations as mammals—social bonding and separation distress associated with the PANIC/Sadness system—or for some other reason. Some biologists do refer to fish living in social groups (Colleter & Brown, 2011). However, the more common explanation is that shoaling reduces risks of predation, so perhaps reduction of FEAR is the likely emotion underlying shoaling behavior. This hypothesis seems especially reasonable, because shoaling is especially common in prey fish (Budaev, 1997; Ward, Thomas, Hart, & Krause, 2004).
A recent series of studies have argued that western mosquitofish, Gambusia affinis, are social, which may allow their shoaling tendencies to be used as a measure of sociability (Cote, Fogarty, & Sih, 2012). However, these authors measured sociability by placing an individual fish in the middle compartment of a tank with three compartments created by inserting two transparent glass partitions on the opposite ends of a large aquarium. One of the end compartments contained fourteen randomly selected mosquitofish. The compartment at the other end was empty. The authors defined sociability as the amount of time the individual fish spent near the compartment containing the other fish. However, in line with Burns (2008), we would argue this was more of an open field test generating FEAR rather than the separation distress of the PANIC/Sadness emotion, but more definitive conclusions may require the use of medications that are more selective for inhibiting these two systems, namely: benzodiazepines for the FEAR system (Panksepp, 1971), and low doses of opiates for the separation-distress PANIC system (Panksepp et al., 1978).
The Cote group reported that the individual experimental fish showed signs of stress when first introduced to the novel aquarium, as indicated by constant swimming along the sides of the tank, which may be an open-field fear response corresponding to Huntingford’s continuous-swimming fear measure. However, Cote et al. (2012) did not measure this behavior and also did not include any test intended to measure fear in their fish social personality test. So, their “social” fish may have spent more time near the other fish due to fear-induced shoaling, and the researchers provided no clear way of distinguishing shoaling due to social motivation from shoaling resulting from fear.
Cote et al. (2012) did report a weak positive relationship between sociability and boldness, suggesting a tendency for social fish to be less fearful, but unfortunately they used the emergence from a safe place to test for boldness, a test that has been criticized as confounded with exploratory behavior and a poor measure of fear (see Burns, 2008; Reale, Reader, Sol, McDougall, & Dingemanse, 2007).
It may be that mosquitofish exhibit a kind of group behavior that appears social, but the question for neuroevolutionary personality theory is whether this grouping behavior represents a precursor to the mammalian PANIC/Sadness system. Perhaps the most obvious argument against social bonding in fish is that they are so easy to raise in isolation. In fact, the Cote group housed their fish individually in small tanks during their experiments.
Furthermore, touching and contact comfort are key features of the mammalian PANIC/Sadness system, with physical contact providing relief from separation distress (e.g., Panksepp et al., 1980), and fish generally do not touch each other in their shoals, but they may do this indirectly by feeling water-pressure movements from adjacent fish. Fish have a “lateral line” consisting of sensory receptors from head to tail that contain hair cell bundles, which detect water movement and vibrations and enable them to maintain appropriate distances from their shoaling neighbors. Faucher et al. (2010) has shown that after inactivating the lateral line fish cannot maintain a shoal: the distance between their closest neighbor doubles, they have difficulty maintaining their orientation relative to other fish, and they frequently bump into other fish. All shoaling behaviors returned to normal as the hair cell bundles regrew. In support of their findings, Faucher et al. (2010) also cite studies showing that cohesive schooling appears only after the developmental completion of the lateral-line system. Thus, it would seem that the maturation of the fish lateral-line system, which enables fish not to physically touch each other (with touching apparently disrupting their “unity”), sustains highly coordinated shoaling, but no clear social motive for this has yet been identified.
Regarding a potential social emotion underlying shoaling, a computer literature search revealed no reports of social attachments in fish. There were reports of reproductive monogamy in fish, but this seemed mostly driven by dominant female fish competing for limited resources and driving away other female fish (Wong, Munday, Buston, & Jones, 2008), although other reports (Harding, Almany, Houck, & Hixon, 2003) suggest sex-specific aggression by both genders can create similar functional monogamy that has nothing to do with mate loyalty.
A related question is whether fish play; that is, do they have affective rough-and-tumble social PLAY systems in their brains? Gordon Burghardt and Vladimir Dinets of the University of Tennessee, along with James Murphy of the Smithsonian’s National Zoological Park, recently published a report documenting three cichlid fish that independently acquired the behavior of repeatedly hitting a bottom-weighted thermometer sitting on the bottom of the fish’s tank and letting it bounce back, which they argued qualified as object play. The fish lived separately and clearly did not learn the behavior from watching each other (Burghardt et al., 2015). These fish were clearly not playing with each other, and there has yet to be a compelling report of social play in fish. In other words, these play behaviors could be envisioned to arise as examples of object play, perhaps arising from their SEEKING systems. In short, a distinct social PLAY system in fish remains to be demonstrated, that is, affectively positive rough-and-tumble social engagements that are highest in juvenile animals and that taper off dramatically following sexual maturation (Panksepp, et al., 1984).
What about parental CARE tendencies in fish? While most fish species do not provide parental care for their young, there are cases of mouthbrooding (see Grone, Carpenter, Lee, Maruska, & Fernald, 2012 for a biological analysis). There are also cases (mostly of male fish, e.g., sticklebacks and cichlids) defending their nests and guarding their offspring against predators (see Balshine & Sloman, 2011). Although rare, there are also cases of Central American cichlids in which parental fish provide food for their young in the form of skin mucus secretions, and of African cichlids stirring up the lake floor by rapidly beating their pectoral fins to expose bottom-dwelling prey for their young (Ota & Kohda, 2015).
Despite an amazing diversity of
parental care activities in fish, an evolutionary analysis of reproductive models in the very large group of ray-finned fishes (class Actinopterygii, a subclass of bony fishes that includes sticklebacks and cichlids) revealed no indication that female-only or biparental care was an outgrowth of a male-only care model or that biparental care has been an evolutionary stepping stone between paternal and maternal care potentially having some neuroevolutionary continuity with mammalian-style maternal parenting. However, the adaptation of females giving live birth (and thereby perhaps a more mammal-like postpartum parental reproductive pattern) has evolved independently at least eight times in ray-finned fish (Mank, Promislow, & Avise, 2005). Accordingly, such species may have antecedents of a primordial CARE system, but at present there are no researched instances of a mammalian-style CARE-type “family feeling” having emerged in fish.
Still, we suspect mammalian-type CARE in fish may have occurred, and if so, it will be most interesting to see if in those cases there are ancestral homologies in the underlying social neurochemistries that would suggest some mammalian-type continuities in fish (e.g., the ancestral variants of oxytocin, such as nine-amino-acid neuropeptides isotocin, mesotocin, and vasotocin. Indeed, it has been shown that isotocin does control paternal care in monogamous cichlid fish (O’Connell, Matthews, & Hofmann, 2012). There is a growing literature that there is some evolutionary CARE continuity from fish to mice and men—but by no means as strong as in women and mammalian mothers in general.
FISH SUMMARY
While neuroscience-type brain manipulations would be valuable in targeting specific emotions in fish, for purely observational research we would recommend more careful selection of tests to avoid the possible confusion of fearful behavior with various social, aggressive, or exploratory behaviors. Indeed, Reale et al. (2007) have recommended multiple tests of each target behavior (a la Huntingford and Burns) to clarify potential confusions.
In summary, curiosity, anger, and fear, which correspond well with the evolutionarily older SEEKING, RAGE/Anger, and FEAR brain systems, are consistently observed personality dimensions in bony fish, the most evolutionarily ancient class of vertebrates reviewed in this book. So far there is no compelling evidence for well-developed CARE, PANIC, or PLAY systems in fish that would support mammalian-like social behavior. Still, there are hints of evolutionary continuities that we suspect will get ever stronger as more neuroscience research is conducted.
HOMOLOGOUS ANIMAL AND HUMAN PERSONALITY TRAITS: A SUMMARY OVERVIEW
Here we pause to briefly summarize where we have been with this cross-species journey. Some may find it disagreeable to think that humans may share some personality traits with other animals, yet such deep homologies are no longer a surprise to psychiatrists, neuroscientists, and geneticists (e.g., Feinberg & Mallatt, 2016). That fish temperaments may be largely defined by the evolutionarily older SEEKING, RAGE/Anger, and FEAR emotion systems would also have come as no surprise to pioneers like Paul MacLean (1990), who was among the first to describe the numerous homologies across early vertebrate, mammalian, and human brains. With mammalian homologies in mind, MacLean (1990) and Panksepp (1998a) would predict that the more evolutionarily recent social emotion systems—CARE, PANIC/Sadness, and PLAY—would provide many commonalities between rats, dogs, foxes, primates, and humans with each of the six personality-focused primary-emotion BrainMind systems influencing mammalian temperaments, which is exactly what was observed. Rat, dog, and ape temperaments are incredibly complex: anticipating basic rewards, obtaining resources, and sometimes just exploring; defending those resources and life itself when necessary; avoiding danger and physical pain; caring for young and thereby unwittingly transmitting epigenetic adaptations (see Chapter 15); avoiding separation from bonded social mates and from familiar places; and learning to regulate key behaviors through socially interactive and physical play. Each of these systems was apparent in the temperament research of the three mammalian examples we summarized in the past chapters. There are echoes of certain homologies in fish, but relevant neuroscience remains scarce.
In addition, a Conscientiousness factor regulating emotional expression was identified in chimpanzees and brown capuchin monkeys (see Chapter 7), with additional demonstrations likely forthcoming as methodological issues are refined. There was also one more unanticipated possible addition to the temperament of the domestic dog: predatory behavior expressed toward humans (see Chapter 8). This is likely part of the food foraging system and just one of the many possible species-typical predatory expressions of a basic SEEKING urge shared by all vertebrates, for instance, the “quiet biting predatory attack” that can be provoked by deep brain stimulation in both rats and cats (Siegel et al., 1999; Panksepp, 1971; Siegel, 2004). Additional basic research involving the analysis of specific brain systems will be necessary to verify homologous systems in dogs.
For humans, we thought it would be useful to have a personality test that attempts to evaluate the primary-process emotional reactivities of our own species—based upon brain emotional systems that have long been evident from cross-species affective neuroscience research (Panksepp, 1982, 1998a). As already summarized, our initial attempt to move in that direction generated a set of evolutionarily defined psychological test scales to tap into each of the primary affective brain systems except LUST (as already noted, omitted to avoid introducing unwanted “guarded” response biases) in the hope of promoting neuroscientifically anchored systematics in the field (Davis et al., 2003). Indeed, each of the six scales monitoring these basic emotional factors—from primordial SEEKING, RAGE/Anger, and FEAR to more socially sophisticated CARE, PANIC/Sadness, and PLAY—correlated highly with various Big Five dimensions except for Conscientiousness. Conscientiousness did not clearly represent any of the primary emotions, suggesting that it may be better envisioned as a cognitive rather than an affective factor, perhaps one representing higher cognitive brain functions regulating the expression of emotions, increasing behavioral sophistication, and adaptability. This idea could shed light on why Scott and Fuller (1965) found that cocker spaniels were so capable of inhibiting their emotional reactions to stressful events and suggests cockers might provide a reasonable animal model for studying an elementary form of Conscientiousness. That Conscientiousness so far has been psychometrically measured only in chimpanzees and brown capuchin monkeys suggests that Conscientiousness requires a well-developed neocortex (for an alternate view on rat prefrontal cortex, see Uylings & van Eden, 1990; for a rat model of orbital prefrontal cortex inhibition of aggressive behavior discussed in Chapter 7, see de Bruin, 1990). Even more dramatically, Frans de Waal (2009) has illuminated complex social emotions, including empathy and compassion, in our primate cousins, which may suggest other novel scales for higher anthropoid social emotions will eventually be needed.
Davis et al. (2003) showed that in humans PLAY correlated positively with Extraversion, CARE correlated positively with Agreeableness but negatively with RAGE/Anger, and SEEKING correlated positively with Openness to Experience. However, they also pointed out a basic flaw with the factor-analytically derived Big Five: RAGE/Anger, FEAR, and PANIC/Sadness all correlated negatively with Emotional Stability/low Neuroticism. Indeed, half a century ago, Walter Hess (1957b) recognized that RAGE and FEAR were two distinct primal emotions (although he did not use our primary-process terminology). MacLean (1990) also recognized that RAGE and FEAR had been major temperament dimensions for probably well over 500 million years (i.e., before the Cambrian explosion of new species), predating the evolutionary appearance of fish and agreeing with RAGE/Anger and FEAR being observable distinct personality dimensions in Huntingford’s (1976) fish research. However, along with PANIC/Sadness, the Big Five lumps together all three negatively valenced emotions, which figure so prominently in psychopathology, or at best relegates them to blends or facets rather than according them the primary temperament status that is apparent in the brain of every mammal. So using the six primary emotion brain systems as a template, Da
vis et al. (2003; refined in Davis & Panksepp, 2011) constructed the Affective Neuroscience Personality Scales (ANPS), which was conceptualized largely as a human research tool capable of situating human subjects in primary-process affective space but that could also usefully be adapted for the study of other species. However, such a jump would be too large for many academic psychologists.
Thus, the scientific study of emotions remains a contentious topic, and the existence of emotional experiences in animals is still more controversial than it should be (for full review, see Panksepp & Biven, 2012; Panksepp, Lane, Solms, & Smith, 2017). There is also a disconnect in what scientists who work with animals claim and what intelligent nonscientists who live with animals believe. To evaluate where the latter stand on such issues, Paul Morris and colleagues from the University of Portsmouth in the United Kingdom surveyed 907 animal owners’ beliefs about the existence of potential emotions in the animals they deal with on a daily basis (Morris, Doe, & Godsell, 2008). The results were striking. The vast majority of these respondents, including 337 dog owners and 272 cat owners, believed their companion animals experienced primary emotions such as anger, fear, surprise, joy/happiness, sadness, anxiety, and curiosity, and a smaller but substantial number of owners also believed the animals had what might be deemed derived emotions, such empathy, shame, pride, grief, guilt, jealousy, and embarrassment. In support of their view, consider that artificial activation of all subcortical emotional systems discussed in this book can serve as rewards and punishments in various learning tasks. Indeed, it is highly likely that the time-honored behavioristic concepts such as “reward” and “punishment” derive the capacity to mold learned behaviors because the various affective changes of brains were designed, in evolution, to guide learning (Panksepp, 1998a, 2005, 2010b, 2011a).
The Emotional Foundations of Personality Page 22