Suppr超能文献

自动控制:专家不假思索的行动方式。

Automatic control: How experts act without thinking.

机构信息

Vanderbilt University.

出版信息

Psychol Rev. 2018 Jul;125(4):453-485. doi: 10.1037/rev0000100.

Abstract

Experts act without thinking because their skill is hierarchical. A single conscious thought automatically produces a series of lower-level actions without top-down monitoring. This article presents a theory that explains how automatic control is possible in skilled typing, where thinking of a word automatically produces a rapid series of keystrokes. The theory assumes that keystrokes are selected by a context retrieval process that matches the current context to stored contexts and retrieves the key associated with the best match. The current context is generated by the typist's own actions. It represents the goal ("type DOG") and the motor commands for the keys struck so far. Top-down control is necessary to start typing. It sets the goal in the current context, which initiates the retrieval and updating processes, which continue without top-down control until the word is finished. The theory explains phenomena of hierarchical control in skilled typing, including differential loads on higher and lower levels of processing, the importance of words, and poor explicit knowledge of key locations and finger-to-key mappings. The theory is evaluated by fitting it to error corpora from 24 skilled typists and predicting error probabilities, magnitudes, and patterns. Some of the fits are quite good. The theory has implications beyond typing. It argues that control can be automatic and shows how it is possible. The theory extends to other sequential skills, like texting or playing music. It provides new insights into mechanisms of serial order in typing, speaking, and serial recall. (PsycINFO Database Record

摘要

专家无需思考即可行动,因为他们的技能是分层的。一个有意识的单一思维会自动产生一系列较低层次的行动,而无需自上而下的监控。本文提出了一种理论,解释了在熟练打字中,自动控制是如何成为可能的,在这种情况下,想到一个单词会自动产生一系列快速的按键。该理论假设按键是通过上下文检索过程选择的,该过程将当前上下文与存储的上下文进行匹配,并检索与最佳匹配相关的键。当前上下文是由打字员自己的动作生成的。它代表了目标(“打字 DOG”)和迄今为止敲击的键的运动命令。开始打字需要自上而下的控制。它在当前上下文中设置目标,从而启动检索和更新过程,这些过程在没有自上而下的控制的情况下继续进行,直到单词完成。该理论解释了熟练打字中分层控制的现象,包括处理高低层次的不同负荷、单词的重要性以及对键位和手指到键位映射的了解不足。该理论通过拟合 24 位熟练打字员的错误语料库并预测错误概率、幅度和模式来进行评估。一些拟合效果相当好。该理论的意义不仅限于打字。它认为控制可以是自动的,并展示了如何实现这一目标。该理论扩展到其他顺序技能,如发短信或演奏音乐。它为打字、说话和序列回忆中的序列顺序机制提供了新的见解。

文献AI研究员

20分钟写一篇综述,助力文献阅读效率提升50倍。

立即体验

用中文搜PubMed

大模型驱动的PubMed中文搜索引擎

马上搜索

文档翻译

学术文献翻译模型,支持多种主流文档格式。

立即体验