As Matthieu Ricard notes in his rigorous book “Altruism,” if an 18-month-old sees a man drop a clothespin she will move to pick it up and hand it back to him within five seconds, about the same amount of time it takes an adult to offer assistance. If you reward a baby with a gift for being kind, the propensity to help will decrease, in some studies by up to 40 percent.

In 2001, the Boston fire commissioner ended his department’s policy of unlimited sick days and imposed a limit of 15 per year. Those who exceeded the limit had their pay docked. Suddenly what had been an ethic to serve the city was replaced by a utilitarian paid arrangement. The number of firefighters who called in sick on Christmas and New Year’s increased by tenfold over the previous year.

The firefighters example makes sense to me but this one does not make sense to me:

If you reward a baby with a gift for being kind, the propensity to help will decrease

"Ever since the concept of altruism was proposed in the 19th century, psychologists have debated whether or not people are born into the world preprogrammed to be nice to others. Now, a pair of Stanford psychologists has conducted experiments that indicate altruism has environmental triggers, and is not something we are simply born with.

In 2006, a study involving toddlers found that the 18-month-olds were willing to provide a helping hand to the experimenters without being prompted. This expression of altruistic behavior in such young children aligned with what many scientists believed to be an expression of innate altruism, and the findings have served as the basis for dozens of studies since.

Rodolfo Cortes Barragan, a psychology graduate student at Stanford, and Carol Dweck, the Lewis and Virginia Eaton Professor of Psychology, suspected there might be more to the story. As with most experiments involving toddlers, the researchers behind the 2006 study engaged in a few minutes of play with the children, in order to make them comfortable with new people in a new setting.

But this interaction, however brief, might have primed the toddler subjects toward altruistic behavior, and affected the outcome of the experiment.

"Kids are always on the lookout for social cues, and this is a very prominent one," said Barragan, the lead author on the research paper. "Does the person's play indicate that they'll care for me? These actions communicate a mutuality, and the child responds in kind."

Barragan and Dweck designed a novel experiment to isolate the effect of the pretest warm-up period. They enlisted 34 one- and two-year-olds and split them into two groups. In the first group, the experimenter would roll a ball back and forth with the child and chat. After a few minutes, the experimenter would "accidentally" knock an object off the table, and observe whether the child would help pick it up, exactly as in the 2006 study.

The difference was in the second group. Here, the experimenter and the child would each play with their own ball, known as "parallel play," while the experimenter engaged in the same kind of chitchat. Again, after a few minutes, the experimenter would knock an object off the table.

The children who engaged in reciprocal play were three times more likely to help pick up the items as the children who had engaged in only parallel play. When the scientists repeated the experiment under slightly different conditions with older children, the reciprocal-play group was two times more likely to lend a hand.

The results suggest that altruistic behavior may be governed more by relationships, even brief ones, than instincts."

http://news.stanford.edu/news/2014/december/altruism-triggers-innate-121814.html

Woah. So altruism is learned through social cues, not by instinct.

I used "gift" as the search word in both papers. Nothing came up to explain, "If you reward a baby with a gift for being kind, the propensity to help will decrease, in some studies by up to 40 percent."

Altruistic Helping in Human Infants and Young Chimpanzees 

Felix Warneken and Michael Tomasello; 2006

http://www.eva.mpg.de/psycho/pdf/Publications_2006_PDF/Altruistic_Helping_in_Human_06.pdf

I still believe some people are naturally kind. Others need to learn compassion. 

Playing with the child first would seem to be an obvious influencer. And the 2006 researchers ignored this?

"... with most experiments involving toddlers, the researchers behind the 2006 study engaged in a few minutes of play with the children, in order to make them comfortable with new people in a new setting.

But this interaction, however brief, might have primed the toddler subjects toward altruistic behavior, and affected the outcome of the experiment."