Suppr超能文献

语言、生物体、行为及其他组合系统中层次复杂性的通用标度律。

Universal scaling laws for hierarchical complexity in languages, organisms, behaviors and other combinatorial systems.

作者信息

Changizi M A

机构信息

Department of Psychological and Brain Sciences, Duke University, Durham, NC 27708, USA.

出版信息

J Theor Biol. 2001 Aug 7;211(3):277-95. doi: 10.1006/jtbi.2001.2346.

Abstract

There are many complex systems in nature where components, or "words", are combined together to make expressions, or "sentences". Such combinatorial systems include: (1) human language, where sentences are composed of words; (2) bird vocalization, where songs are built from syllables; (3) organisms, where organism-expressions (e.g. the tonsil) are made out of cells; (4) behavioral repertoire, where mammalian behavior consists of a temporal arrangement of muscle contractions; (5) universities, where student academic degrees are comprised of departmental concentrations; and (6) electronic devices, where the device's actions are implemented via strings of button-presses. My central aim here is to discover how combinatorial systems accommodate greater numbers of expressions; that is, what changes do combinatorial systems undergo when they "say more things?" Are there general laws characterizing the properties of combinatorial systems as the number of expressions increases? If so, what are they? My main result is that, in all the kinds of combinatorial system mentioned above, there appear to be general laws describing how combinatorial systems change as they become more expressive. In particular, in each of these cases, increase in expression complexity (i.e. number of expressions the combinatorial system allows) is achieved, at least in part, by increasing the number of component types. Each kind of system follows one of two kinds of scaling law. In the first kind of scaling law, expression complexity increase is carried out exclusively by increasing the number of component types; the number of components per expression (i.e. the expression length) remains invariant. This applies to human language over history, bird vocalization, organisms in phylogeny and ontogeny, and universities. In the second kind of scaling law, expression complexity is accomplished by increasing in a law-like manner both the number of component types and the expression length. This applies to two cases of the ontogeny of language-the development of words and sentences, and the development of phonemes and morphemes-and to mammalian behavior. By treating these diverse systems as combinatorial systems we, in addition to elucidating general principles underlying such systems, gain insight into each kind of system mentioned.

摘要

自然界中有许多复杂系统,其中各个组成部分(即“单词”)组合在一起形成表达形式(即“句子”)。这类组合系统包括:(1)人类语言,句子由单词组成;(2)鸟类发声,歌声由音节构成;(3)生物体,生物体的表达形式(如扁桃体)由细胞组成;(4)行为库,哺乳动物的行为由肌肉收缩的时间安排构成;(5)大学,学生的学术学位由各系的专业组成;(6)电子设备,设备的操作通过一连串的按键来实现。我这里的核心目标是探究组合系统如何容纳更多的表达形式;也就是说,组合系统在“说出更多内容”时会发生哪些变化?随着表达形式数量的增加,是否存在描述组合系统特性的一般规律?如果有,这些规律是什么?我的主要研究结果是,在上述所有类型的组合系统中,似乎都存在描述组合系统在变得更具表现力时如何变化的一般规律。具体而言,在上述每种情况下,表达复杂性的增加(即组合系统允许的表达形式数量的增加)至少部分是通过增加组成部分类型的数量来实现的。每种系统都遵循两种缩放定律中的一种。在第一种缩放定律中,表达复杂性的增加完全通过增加组成部分类型的数量来实现;每个表达形式中的组成部分数量(即表达形式的长度)保持不变。这适用于历史上的人类语言、鸟类发声、系统发育和个体发育过程中的生物体以及大学。在第二种缩放定律中,表达复杂性通过以一种规律的方式同时增加组成部分类型的数量和表达形式的长度来实现。这适用于语言个体发育的两种情况——单词和句子的发展,以及音素和语素的发展——以及哺乳动物的行为。通过将这些不同的系统视为组合系统,我们除了阐明此类系统的一般原理外,还能深入了解上述每种系统。

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