Silk: The stability of these children's responses is quite striking. One explanation for this could be that children are early on rewarded for this positive behaviour in a consistent fashion. Parents begin to treat them in a certain way and this becomes part of who they are. The stability is generated by the environment in which they are rewarded. What do the data tell us about the sources contributing to this stability?
Spinrad: I don't want to speak directly to the stability, because I don't know whether there is a relationship between socialization strategies and stability, per se. However, there is evidence of a positive link between parental warmth and prosocial behaviour, sympathy and empathy (Deater-Deckard et al 2001, Kiang et al 2004, Kochanska et al 1999, Laible & Carlo 2004, Strayer & Roberts 2004), as well as emotion regulation. Attachment security also has been found to relate to prosocial behaviour or sympathy (Waters et al 1986, Van der Mark et al 2002). Punitive parenting tends to be negatively related to these behaviours (Asbury et al 2003, Deater-Deckard et al 2001). We know that the parenting behaviours are likely to be stable over time, and this could be contributing to the stability.
Warneken: You said the longitudinal study starts with four to five year olds. Are there any longitudinal studies starting with children at a younger age? Hildy Ross claimed that prosocial behaviours might decrease from young into middle childhood, but that was based upon cross-sectional data.
Spinrad: Nancy Eisenberg and Richard Fabes conducted a meta-analysis on age changes and children's prosocial behaviours (Eisenberg & Fabes 1998). They found prosocial behaviour increases with age in general, but these vary by the context of the studies and with the age range. There are fewer studies in infancy and toddlerhood; however, Zahn-Waxler and her colleagues have some evidence that prosocial behaviour does increase in toddlerhood (Zahn-Waxler & Radke-Yarrow 1982, Zahn-Waxler et al 1992, 2001), as do some others (Van der Mark et al 2002, Lamb & Zekhireh 1997). In general, we see increases in prosocial behaviour with age, but the strength of these findings varies depending on the context and the methods used (Eisenberg & Fabes 1998).
Hauser: You find correlations between the sympathy/empathy measures and the prosocial behaviour. Could something else be driving this correlation? One variable could be Walter Mischel's delayed gratification or discounting as a parameter, which does show up much earlier in development. He has studied these children from the age of about two, and this is a remarkably stable characteristic. Could this be the variable driving the correlation you are finding, rather than it being anything to do with sympathy or empathy?
Spinrad: Yes, we are very interested in the role of emotion regulation in children's prosocial behaviour and empathy/sympathy. We have also recently distinguished between regulation that is effortful versus less voluntary (Eisenberg & Spinrad 2004). In terms of the delay task, we might expect that the ability to delay would at least partly tap effortful regulation, although it might partly tap children's low impulsivity as well (Spinrad et al 2006). Regardless of that issue, there have been quite a few data that tell us that children who are relatively well-regulated are more likely to behave prosocially (Eisenberg et al 1996, 1997) and to experience sympathy as opposed to personal distress reactions (Eisenberg & Fabes 1995, Eisenberg et al 1996, Eisenberg et al 2001). We believe that this behaviour is probably underlying some of the ability to be other-oriented as opposed to being self-focused.
Hauser: The nice thing about the species of monkey we work on is that they twin naturally, making DZ twins. We have begun to look at heritability of these kinds of behaviour, to see if there is consistency across different tasks. Animals that are patient in one task may well be patient in others. When you do Walter Michel's delayed gratification tasks, are the children who hold on longer the ones who are more prosocial? It could be the impulsivity level that is driving the prosocial behaviour and sympathy/empathy is just an intermediate variable, correlated because of the delayed gratification.
Spinrad: I would argue that sympathy is mediating a relation between emotional regulation and prosocial behaviours.
Sigman: Didn't the data you presented earlier about heritability speak to that?
Spinrad: There is an interesting study by Robinson et al (2001). Using a twin-design sample, the researchers examined toddlers' responses to feigned distress in several situations, when the victim was either a stranger or the mother. Their findings showed that the heritability of prosocial behaviour depended on whether the victim was the mother or the stranger (with a stronger heritability index toward the stranger).
Frank: Some evidence suggests that criminals were overwhelmingly likely to have had impulse control problems as children. It could be that there is some independent competence called the ability to delay gratification, but causation could go in the other direction, too. If you think about the repeated prisoner's dilemma, a purely prudent person would want to cooperate on the first round to maintain a string of successful interactions. The difficulty is that the gain comes now if you defect, whereas the reward for cooperating comes only in the future. If you had some independent concern for the well-being of your trading partner, it would be easier to clear the impulse-control hurdle
Hauser: I find the discounting issue interesting, and am convinced by our tamarin results. The problem is not the understanding of the pay-offs of cooperation, but rather the inability to delay gratification. It could very well be that all animals are capable of perceiving the advantages of reciprocity, but simply fail to engage in such cooperative behaviour because they are incapable of waiting for returned rewards. Impatience causes reciprocity to crash.
Warneken: At the beginning you distinguished prosocial from altruistic behaviours, yet when you presented the data altruistic behaviours didn't show up any more. Is that because altruistic behaviour was so infrequent it was collapsed with prosocial behaviours in a broader sense?
Spinrad: The differentiation between prosocial and altruistic behaviours is in regard to motivation. Because it is difficult to assess children's motivation, we refer to all of the behaviours we measure as prosocial behaviours.
Warneken: But in the definitions it seems like a continuous variable on one dimension, with prosocial on one side and altruism on the other. There is no cutoff point where you would say now it is totally altruistic behaviour.
Spinrad: Again, I would argue that the difference is the motivation behind the behaviour, which is difficult to assess. Some behaviours are altruistic whereas some are not, but we simply have difficulty differentiating among them because we do not know the child's thoughts, goals and motives.
Warneken: Alright, according to what you just said, prosocial behaviour can be assessed by looking at the behaviour alone without looking at the motives. When you then want to find out whether it is altruistic or not you have to look at the motives. But it rather seemed to me that it is all about motives here, and only if the motives are identified and they are all about the other can we say it is altruistic.
Spinrad: Yes. Prosocial behaviour is a more general term, referring to behaviours such as volunteering, helping, sharing and comforting. Altruism is a specific case of this type of prosocial behaviour that is other-oriented (Eisenberg 1986).
Blair: I want to return to the delay of gratification and impulse control questions. The disadvantage of having this as the link between empathy and prosocial behaviour is that one of the measures was the emotional response to the pictures. This index of the basic emotional response is unlikely to relate to delay of gratification in the way that it is usually thought about. It would be difficult to link these two. Regarding the impulse control, we can be pretty sure that impulse control more generally is not the thing that is driving empathy because you have children with attention deficit hyperactivity disorder (ADHD), a classic impulse control disorder, and they are not showing indications of profound empathy impairment. We can be confident that this is not the explanation.
Van Lange: Research on the prisoner's dilemma confirms that the orientation with the future is quite independent of prosocial orientation. Some people with an individualistic orientation can take long-term orientation and be cooperative. One reason why pure reciprocity elicits a lot of cooperation is that people cooperate because they know in the long run it is a good thing to do, but these are not people who are inherently prosocial.
Gallese: Is there any relationship between sympathy, the development of prosocial behaviour and the specific type of attachment these children experience?
Spinrad: There is some work that suggests that children who are securely attached are more likely to be empathic, but much of work doesn't differentiate between empathy and sympathy (Van der Mark et al 2002). On a related issue, I have some data showing that maternal sensitivity/responsivity observed at 10 months of age predicts sympathy toward their mother and a stranger at 18 months of age (Spinrad & Stifter 2006). This study doesn't directly assess security of attachment, but given the existing links between maternal responsivity and attachment security, these findings are related to your question.
Montague: When you do an experiment like that, how do you tell which way it is going? I may be being more sensitive to my baby because it is eliciting this behaviour from me. The baby may be selecting for parental behaviour rather than the mother sending signals to the baby that induce a different state in the baby.
Spinrad: It is easier to be sensitive to some babies than others, I agree. So, it is important to take the child's characteristics into account. However, I think that sensitivity is a code that takes the child's behaviour into account. For example, if the child is focusing on a particular toy, a sensitive mother would also focus on that toy and not move onto another toy until the child lost attention to that toy. Moreover, if a child is fussy, a mother can be sensitive by soothing and comforting that child.
Montague: A child's behaviour is difficult to characterize. It indexes a certain part of the parental behavioural space. This is why children with autism spectrum disorders are particularly hard on parents, because the parents are expecting a whole range of responses from the child which they don't get.
Spinrad: I don't think that there is any question that there are bidirectional effects. In fact, a study by Eisenberg and colleagues found that parental behaviour predicted children's emotion regulation over time (controlling for early emotion regulation). In turn, children's emotion regulation predicted parenting behaviours two years later (even after controlling for early levels of the behaviour). Thus, it is clear that parents influence their children as well as the reverse (i.e. children influence their parents).
Gergely: With respect to the question about attachment security, the data that are contradicting the idea that the quality of attachment would be solely infant- or temperament-induced (see Vaughn & Bost 1999) are the findings that indicate no (or only very low) correlation between the types of attachment (secure, avoidant, resistant or disorganized) that characterize an infant's relationship to different caregivers, respectively. A child can be securely attached to one attachment figure while showing insecure attachment in relation to the other. Also, the type of the parent's attachment status as measured by the Adult Attachment Interview (AAI) (George et al 1996) predicts rather well the infant's specific attachment classification with the parent at one year of age (as measured by the Ainsworth Strange Situation Test, see Ainsworth et al 1978) (Fonagy et al 1991). Temperament (child ^ parent effects) seems an inadequate account of this finding as the AAI of the parent is collected and coded before the birth of the child.
Singer: I'm trying to take a neuroscientific perspective on your research on empathy and fit your data into this account. The easy neuroscientific story would be to suggest that the better your action perception resonance mechanism in your brain, the more you share the feelings of others and the more empathic you are. This would imply, however, that sharing negative emotions with others lead automatically to own distress. Now you are showing negative correlations between personal distress and helping behaviour, the latter again associated with empathy. According to such a view, any resonance mechanism resulting in personal distress by the sight of someone suffering negative emotions is hindering empathy.
Now you are also saying that psychopaths don't respond as sensitively to emotional stimuli as normal controls, and this is why they aren't empathic. This seems to suggest in turn, that sharing affect with the others is a necessary condition for empathy to arise. In conclusion, I would suggest that you need both for empathy to arise: first a shared representation mechanism to allow for sharing affective experiences with others and, in addition, a top—down modulation mechanism which allows to suppress to strong empathic responses so that you can engage in helping behaviour.
Spinrad: I think there is an optimal level of distress. Eisenberg et al (1994) proposed that emotional overarousal would be associated with personal distress whereas moderate distress is associated with sympathy. Hoffman (1982) also has made this suggestion. Thus, if people can maintain arousal in a tolerable range, they should experience sympathy. There is empirical work to support this notion (Eisenberg et al 1991). Also, distress can be related to helping, when it is the easiest way to reduce one's own distress (for example, when a person cannot escape the distressed other).
Singer: Here neuroscience can make a distinction between pure cognitive perspective taking and empathizing. The first is based on different brain structures than the latter. It seemed that you were saying that sympathy is just engaging the perspective taking network or theory of mind network, which is totally different from the empathy system which is based on limbic brain structures rather than pre-frontal structures or temporoparietal junction (TPJ). But I didn't get the impression that you were really suggesting that. You said that one part of sympathy is actually based on an empathic response (that is, feeling the same thing as the other person does). That would clearly involve brain structures associated with the processing of bodily or sensory experiences rather than merely structures dealing with propositional attitudes.
Spinrad: It is feeling concern or sorrow for the other person. It may stem from feeling what the other person feels.
CFrith: Can't you say that it is all a matter of degree? There is social distress if your response is too great. At the beginning of the meeting I suggested that if you see someone sad and you feel sad, this initiates in you the need to take action to reduce your sadness. In the same way, this could lead you to act to reduce the other person's sadness, which will reduce your sadness because of the resonance.
Singer: I would say that sympathy arises because you have the ability for top— down control of your feelings. All the difference observed between sympathetic or non-sympathetic children should then be in whether these children have affective emotional regulation mechanisms in place.
Moll: It depends on the measure that you use. What was your measure of 'sympathy'?
Spinrad: We used three different measures: self report, facial measures and physiological measures. Facial concern is related to prosocial behaviour, but there is a level of distress that we consider an over-arousal, where children will be more focused on their own arousal and alleviating their own distress than being able to act on this and behave prosocially. In such studies, children can easily escape dealing with the other person who induced the empathy, so there is no need to help to reduce their own arousal.
Sigman: I have a question about gender differences. This area of research seems marked by very large gender differences. First, there is a big primary effect of femininity on the whole process, and then there are differences in stability over time, where it sounds like you are getting more prediction from males than females. How do you understand and interpret these gender differences.
Spinrad: Again, Eisenberg & Fabes (1998) conducted a meta-analysis on gender differences as well. Findings showed that gender differences seem to be stronger depending on the type of measure used. Self-report measures produce much stronger sex differences, whereas facial or physiological measures of empathy, sympathy or personal distress show weaker sex differences.
Sigman: What about stability? Does this continue to be stronger for males?
Spinrad: I don't know.
Blair: I have a comment about the personal distress/empathy differentiation. The potential differentiation isn't to do with the level of arousal to get to the negative affect, but the fact that one set of children have a set of strategies to deal with the problem facing them while another set don't. There's a good example of this, which I haven't seen published, where researchers took young children and put them in an empathy-inducing paradigm in which the researcher feigns hurting themselves. They taught one set of children to mop the experimenter with a cloth but not the other set. The non-taught children showed significantly higher levels of crying and confusion. The sad face is an aversive stimulus. The difference is, if you can deal with it and take away the sad face you are in a relief condition as opposed to a situation where you have no idea what to do. There is a way of characterizing the personal distress/empathy divide that is more compatible with the sorts of things you were going for.
Silk: I have a question about the distribution of these traits in children. Is this a bimodal distribution? How does this follow along these dimensions that you mentioned?
Spinrad: Children can experience both sympathy and personal distress. We show them sympathy-inducing films and some children will experience sympathy, and others experience more personal distress, but they can experience both, likely sequentially. We don't look at this as a distribution.
Silk: When you have your correlations, you must be putting something in for the different time periods for each child.
Spinrad: We look at the film, decide what the sympathy-inducing portion is, and decide what the distress-inducing portion is. We look to see whether children experience distress during the sympathy portion.
Silk: What does that distribution look like?
Spinrad: I don't know, but distress responses generally are more common than concern reactions, in the baseline and during the evocative portions of the film.
Frank: Some evidence suggests that criminals were overwhelmingly likely to have had impulse control problems as children. It could be that there is some independent competence called the ability to delay gratification, but causation could go in the other direction, too. If you think about the repeated prisoner's dilemma, a purely prudent person would want to cooperate on the first round to maintain a string of successful interactions. The difficulty is that the gain comes now if you defect, whereas the reward for cooperating comes only in the future. If you had some independent concern for the well-being of your trading partner, it would be easier to clear the impulse-control hurdle?
Spinrad: I think it has to do with the amount of cost. If we did control it, I think we'd find the same with both. In some of Nancy Eisenberg's work with children, helping was operationalized as low cost behaviours whereas sharing involved giving up territory or a possession.
Was this article helpful?
Has Your Child Been Diagnosed With ADHD Is Coping With Your Child's Behavior Wearing You Out Are You Tired of Searching For Answers An ADHD child does not have to have a dark cloud over his or her head. If You've Got Burning Questions About ADHD, I've Got Answers.