Abstract Plants have developed toxic chemical and physical defenses as a consequence of their co-evolution with herbivores. Humans, like other animal species, have evolved strategies to protect themselves from such plant dangers. For example, recent studies have shown that human infants exhibit a reluctance to manually explore plants and use social learning (SL) to acquire knowledge about plants. However, SL can also be costly under certain circumstances and there is reason to suspect this may be the case for plants. Some plant species are difficult to distinguish from one another. For example, some plants have evolved an adaptive strategy to fight against herbivorous threats, called Batesian mimicry, in which an edible plant mimics features of a poisonous plant to minimize the probability that it is consumed. When SL is prevalent in a population, by proliferating the knowledge about an edible mimic, SL also spreads the risk of consuming its poisonous counterpart. Here we propose a model describing different scenarios where SL is (a) favored, (b) ecologically stable, and (c) expected to evolve. Results show that SL is selected when the proportion of poisonous plants is high. However, this is only true if the edible mimic population is below a certain threshold and its selection depends on the capacity to minimize errors when differentiating edible mimics from their poisonous counterparts.

中文翻译：

---

通过误差最小化植物社会学习的演变

摘要 作为与食草动物共同进化的结果，植物已经形成了有毒的化学和物理防御系统。与其他动物物种一样，人类已经进化出保护自己免受此类植物危害的策略。例如，最近的研究表明，人类婴儿不愿意手动探索植物并使用社会学习 (SL) 来获取有关植物的知识。然而，在某些情况下，SL 也可能是昂贵的，有理由怀疑植物可能是这种情况。一些植物物种很难相互区分。例如，一些植物已经进化出一种适应性策略来对抗食草动物的威胁，称为贝茨拟态，其中一种可食用植物模仿有毒植物的特征，以尽量减少它被食用的可能性。当 SL 在人群中普遍存在时，通过传播关于可食用模拟物的知识，SL 还传播了食用其有毒对应物的风险。在这里，我们提出了一个模型来描述不同的情景，其中 SL 是 (a) 有利的，(b) 生态稳定，以及 (c) 预计会发展。结果表明，当有毒植物比例较高时选择SL。然而，这仅在可食用拟物种群低于某个阈值时才是正确的，并且其选择取决于将可食用拟物与其有毒对应物区分开来时将错误降至最低的能力。结果表明，当有毒植物比例较高时选择SL。然而，这仅在可食用拟物种群低于某个阈值时才是正确的，并且其选择取决于将可食用拟物与其有毒对应物区分开来时将错误降至最低的能力。结果表明，当有毒植物比例较高时选择SL。然而，这仅在可食用拟物种群低于某个阈值时才是正确的，并且其选择取决于将可食用拟物与其有毒对应物区分开来时将错误降至最低的能力。