Lind Johan, Jon-And Anna
Biology Division, Department of Physics, Chemistry, and Biology (IFM), Linköping University, Linköping, Sweden.
Centre for Cultural Evolution, Department of Psychology, Stockholm University, Stockholm, Sweden.
Philos Trans R Soc Lond B Biol Sci. 2025 Jun 26;380(1929):20240116. doi: 10.1098/rstb.2024.0116.
Here, we explore some cognitive mechanisms that support and constrain sequential abilities in non-human animals (hereafter animals). By examining limits in memory for stimulus sequences and how behaviour sequences can be learned, we highlight the combinatorial costs that arise as sequences get increasingly longer, which may hinder the development of cognitive abilities that require faithful representation of sequences, like language. We discuss a trace memory model as a framework for understanding how animals represent stimulus sequences and suggest that animals represent sequences as unstructured collections of decaying memory traces rather than representing order faithfully. The implications of this model challenge traditional interpretations of declarative and rule-based learning in animals. In addition, we explore associative learning models that can account for how animals acquire behaviour sequences without precise memory of stimulus sequences. Current models have proven powerful in accounting for complex behaviour sequences. We end by asking what the value is of anthropocentric models in the study of animal intelligence, if other models provide more accurate predictions of animal behaviour.This article is part of the Theo Murphy meeting issue 'Selection shapes diverse animal minds'.
在此，我们探究了一些支持和限制非人类动物（以下简称动物）序列能力的认知机制。通过研究刺激序列记忆的局限性以及行为序列如何被学习，我们强调了随着序列变得越来越长而产生的组合成本，这可能会阻碍诸如语言等需要忠实表征序列的认知能力的发展。我们讨论了一种痕迹记忆模型，作为理解动物如何表征刺激序列的框架，并提出动物将序列表征为衰减记忆痕迹的无结构集合，而不是忠实地表征顺序。该模型的影响挑战了对动物陈述性学习和基于规则学习的传统解释。此外，我们还探讨了联想学习模型，这些模型可以解释动物如何在没有精确记忆刺激序列的情况下习得行为序列。目前的模型已被证明在解释复杂行为序列方面很强大。最后，我们提出，如果其他模型能对动物行为做出更准确的预测，那么以人类为中心的模型在动物智能研究中的价值何在。本文是西奥·墨菲会议议题“选择塑造多样的动物思维”的一部分。
