鹦鹉可以理解概率 ​《自然-通讯》

论文标题：Kea show three signatures of domain-general statistical inference

期刊：Nature Communications

作者：AmaliaP. M. Bastos,AlexH. Taylor

发表时间：2020/03/03

数字识别码：10.1038/s41467-020-14695-1

《自然-通讯》本周发表的一项研究Kea showthree signatures of domain-general statistical inference发现，新西兰的啄羊鹦鹉不仅能理解概率，还能据此采取行动。这一发现也是对大猿以外动物可以进行统计推断的首次报道。

版权:AmaliaBastos

新西兰奥克兰大学的Amalia Bastos和AlexTaylor设计并开展了一系列实验，对照统计理解的多个指标进行测试，呼应之前对灵长目和人类婴儿的研究。首先，他们训练了6只啄羊鹦鹉（名为Blofeld、Bruce、Loki、Neo、Plankton和Taz），让它们将黑色与有奖赏、橙色与无奖赏联系起来。作者按不同的相对频率在两个透明罐子里分装了橙色和黑色小棍，并从两个罐子中取小棍给啄羊鹦鹉选择，展示时把小棍攥在拳头里，不给啄羊鹦鹉看。

版权:AmaliaBastos

版权:AmaliaBastos

作者发现，啄羊鹦鹉更喜欢黑色比例更高的罐子里的小棍，但这种选择标准并不绝对。当罐子里放置一层水平夹板，改变了夹板上方可拿取黑色小棍的比例时，啄羊鹦鹉能发现这种物理限制，并选择可拿取黑色小棍概率最高的罐子。最后，啄羊鹦鹉还更喜欢选择“偏袒型”实验人员手里的牌子，即这些实验人员之前给出黑色小棍的概率更高。

版权:AmaliaBastos

之前的研究只在人类和大猿中证明了真正的统计推断能力。此次在鸟类中发现这种复杂高阶的认知过程，或有助于进一步认识统计推断的演化历史。

摘要：One key aspect of domain-general thought is the ability tointegrate information across different cognitive domains. Here, we testedwhether kea (Nestor notabilis) canuse relative quantities when predicting sampling outcomes, and then integrateboth physical information about the presence of a barrier, and socialinformation about the biased sampling of an experimenter, into theirpredictions. Our results show that kea exhibit three signatures of statisticalinference, and therefore can integrate knowledge across different cognitivedomains to flexibly adjust their predictions of sampling events. This resultprovides evidence that true statistical inference is found outside of the greatapes, and that aspects of domain-general thinking can convergently evolve inbrains with a highly different structure from primates. This has importantimplications not only for our understanding of how intelligence evolves, butalso for research focused on how to create artificial domain-general thoughtprocesses.