New horizons in evolutionary linguistics
演化语言学的新视野
Edited by Peng Gang 彭刚 and Wang Feng 汪锋
Abstract 摘要
The study of language evolution in Europe has an interesting history. In 1769 the Berlin Academy announced as the subject of a competitive prize essay the question of how language emerged. The winning essay by Johann Gottfried Herder, Ursprung der Sprache, has been thoughtfully discussed by Edward Sapir; it is available in English translation, together with another essay on the same theme by Jean-Jacques Rousseau. The question is a fascinating one, of course, but given the state of knowledge at that time, it must have attracted endless speculations and futile controversies which did little to advance linguistic scholarship. This in turn led the Société Linguistique de Paris to famously ban discussions of the question at their meetings, in the second article of their 1866 statutes, a ban that was echoed in London several years later. It took approximately another century before an article in the Scientific American by Charles F. Hockett in 1960 resurrected the question. By this time knowledge has advanced importantly on two critically relevant fronts: animal communication across species, and the diversity of linguistic structures. For the first time, with Hockett’s proposal of a set of design features comparing human language and animal communication, the discussion shifted away from philosophy and theology into the realm of empirical science. The pace for evolutionary linguistics picked up when in 1976 the New York Academy of Sciences sponsored a very successful conference that drew over a thousand participants, with experts interacting across a dozen disciplines, with anthropology and zoology literally spanning A to Z. The launching of a series of international conferences in 1996, called EVOLANG, by linguists at Edinburgh University, notably by James R. Hurford, meeting every two years, gave the study of this multi-disciplinary area a stable forum for interaction. Even though evolutionary theory entered China early in the 20th century, when YAN Fu 严复 translated some of Thomas Huxley’s lectures (Tian Yan Lun天演论), unfortunately this perspective has not had much influence on linguistics in China. Following EVOLANG, another series of annual conferences was then launched in 2009 in Guangzhou, called Conference in Evolutionary Linguistics (CIEL). Reports for most of these CIELs have been published in the Journal of Chinese Linguistics. The acronym CIEL is especially auspicious, since the meaning of the French word is ‘heaven’, calling to mind YAN Fu’s famous Chinese phrase for natural selection, which is wu jing tian ze 物竞天择 “organisms compete as heaven selects”. While CIEL meets primarily in China, one of its main goals is to promote international interaction, with participation of leading scholars from other parts of the world.
Keywords 关键词
Language evolution 语言演化
Abstract 摘要
The series of the international Conference in Evolutionary Linguistics (CIEL) have been successfully held for eight times in Guangzhou (1999), Tianjin, Shanghai, Beijing, Hong Kong, Xiamen, Tianjin, and this year again in Bloomington. The CIEL series have consolidated a platform for international interdisciplinary exchange on Evolutionary Linguistics, thereby promoting the development of this field of research. This monograph, New Horizons in Evolutionary Linguistics, includes selected revised papers from CIEL-4 held at Peking University in November 2012, and CIEL-5 held at The Chinese University of Hong Kong in August 2013. The 11 papers in this monograph all pertain to the broad spectrum of evolutionary linguistics.
Article 文章
Abstract 摘要
The Mirror System Hypothesis holds that the mechanisms that support language parity (the listener tends to understand what the speaker intended) evolved atop an ancestral mirror system for manual actions that supported both the generation and recognition of such actions. This paper1 summarizes research that has enriched our understanding of the adaptability of mirror neurons and their role in the evolution of the human language-ready brain.
镜像系统假说认为支撑语言奇偶性（听者趋于理解话者的意欲）的机制是从支撑行为动作的更早期的镜像系统演化而来，这个更早期的镜像系统支持动作的产生和识别这样的动作。这篇文章概括了那些丰富了我们对于镜像神经元适应性的了解以及镜像神经元在人类有语言能力的大脑的演化中的作用的研究。
Keywords 关键词
Mirror System Hypothesis 镜像系统假说 Mirror neurons 镜像神经元 Language evolution 语言演化 Language-readiness 语言就绪
Abstract 摘要
This paper mainly discusses two questions about language contact, first, what is the direction of language contact, and two, what is the change of linguistic typology caused by language contact. According to language contact between Dai and Southwest Mandarin, mother tongue interference and lexical borrowing are two important mechanisms of language contact, and they clearly have different directions. Words usually diffuse from the strong language to the weak one by borrowing, and the structures usually diffuse from the weak to the strong by interference. In consideration of the feature of language contact, we have proposed an inverse principle of lexicon-structure. According to our long-term research of language contact, both mother tongue interference and lexical borrowing can make the structures of the mother tongue and the target language isomorphic. Mother tongue interference can change language more easily, resulting in the disappearance of a large number of typological features.
本文主要讨论了两个方面的内容，一是语言接触的有向性机制；二是语言接触导致语言类型的改变。从傣语与汉语接触的现象来看，母语干扰与词汇借贷是语言接触的两种重要机制，它们都有极强的方向性，词汇通常是以借贷的方式从强势语言向弱势语言扩散，结构通常是以干扰的方式从弱势语言向强势语言扩散，这两种过程是逆向的，可称为接触的逆向对立原则。母语干扰和词汇借贷能使母语与目标语形成同构，尤其是语言接触中的母语干扰最能够改变语言的类型特征，导致大量少数民族语言及独特语言类型特征的消失，这些现象在语言深度接触中可以观察到。
Keywords 关键词
Language contact 语言接触 Mother tongue interference 母语干扰 Lexical borrowing 词汇借贷 Direction of languages contact [语言接触]有向性 Typological feature 类型特征
Abstract 摘要
This study makes a distinction between word and lexicon. A word is an individual lexical item. The lexicon is an aggregate of words. Past discussions of differences among lexicons could only list individual words for comparison. In this study, words of digital texts in Old Chinese, Pre-Modern Chinese, Modern Chinese, ancient poetic writings and modern press releases were examined and a small number of significant lexical elements were extracted to capture the distinct nature of the lexicons. The crucial point of distinction is how words are used and not whether particular words exit. Therefore, the lexical attribute discussed here is called dynamic attribute. The words in the text streams were tabulated for their frequency and percentage of the occurrence with respect to the entire texts. The cumulative percentage of occurrences of the 15 highest frequency words was also tabulated. It was then used as the concentration level of high-frequency words in use. This concentration level clearly differentiates the types of texts used in Old Chinese, Pre-Modern Chinese, Modern Chinese, ancient poetic writings and modern press releases. It is hoped that this quantitative lexical attribute may be of some use in future lexicological research.
本文区分词语和词汇，词语是个别词，词汇是词语的集合体。过去讨论不同词汇的差异都只能列举词语的异同，本文考查上古汉语、近代汉语、现代汉语、古代韵文以及现代汉语新闻稿文字，从中提炼出少数词汇要素。区别的关键在于词语的使用而不在于词语的有无。因此，本文所提出的从最高词频往下累积，得出词频在全部词语数目中的百分比。我们以词频统计中最高的15个词语的词频累积百分比作为高频词集中度，以高频词集中度当作词汇动态特性。从文本计算出来的词汇动态特性能清楚划分出所考查的古代、近代、现代、古代韵文和现代新闻稿的词汇属性。希望这个计量性质的词汇属性对今后的词汇研究有些助益。
Keywords 关键词
Word and lexicon 词语与词汇 Lexical dynamic attribute 词汇动态属性 Cumulative frequency percentage 词频累积百分比 Concentration level of high-frequency words 高频词集中度
Abstract 摘要
Construction of phylogenetic trees from observations is a fundamental challenge in both evolutionary biology and evolutionary linguistics. Here we approach the problem from a new perspective by adopting algebraic invariants associated with topological structures of phylogenetic trees. Our key development is based on machine learning to optimize the power of phylogenetic invariants for the construction of phylogenetic tree quartets, the building blocks of general evolutionary trees. Phylogenetic invariants are polynomials in the joint probabilities which vanish under a model of evolution on a phylogenetic tree. We give algorithms for selecting a good set of invariants and for learning a metric on this set of invariants which optimally distinguishes the different models. Our learning algorithms involve linear and semidefinite programming on data simulated over a wide range of parameters. We provide extensive tests of the learned metrics on simulated data from phylogenetic trees with four leaves under the Jukes-Cantor and Kimura 3-parameter models of DNA evolution. Our method greatly improves other uses of invariants and is competitive with or better than the popular neighbor-joining method. In particular, we obtain metrics trained on trees with short internal branches which perform much better than neighbor joining on this region of parameter space. These results exhibit potential advantages of applying the new methodology to evolutionary linguistics.
从观测数据中构建演化树是生命演化和进化语言学的一个基础问题。本文试图从一个新角度来研究这个问题，即通过演化树的代数不变量来重建演化树的拓扑结构。我们关键的新发展是基于机器学习来优化选择演化树的代数不变量，针对四元演化树发展了一种新的构造方法。演化树代数不变量是指关于联合分布的一种特殊的代数多项式，其在树上的演化模型下恒等于零。本文主要贡献在于发展了一类算法，用于选择一组更好区分不同演化树模型拓扑结构的代数不变量以及相应的度量结构。我们的算法基于给定演化模型下的广泛参数变化而产生的仿真数据，采用线性规划和半正定规划来学习。文中对于DNA 演化的Jukes-Cantor 模型和Kimura 三参数模型进行了广泛的仿真试验测试。试验表明：本文方法整体上同目前广泛使用的Neighbor－Joining 算法相比，具有相似或者更好的性能；特别是对于四元树具有较短内部分支的Felsenstein 参数区，本文方法远远超过后者的性能。这些结果展示了将我们的新方法应用于进化语言学研究时可能具有的优势。
Keywords 关键词
Phylogenetic invariants 演化树代数不变量 Algebraic statistics 代数统计量 Semidefinite programming 半正定规划 Felsenstein zone Felsenstein参数区
Abstract 摘要
Human language is constantly changing. Many of the changes are achieved via diffusion of linguistic variants in a population of interacting individuals. We design a simple computer model following the Pόlya-urn dynamics to evaluate possible effects of linguistic, individual learning, and socio-cultural factors on linguistic diffusion. Simulation results and statistical analyses collectively reveal that: variant prestige, rather than individual bias, is a fundamental selective pressure for diffusion; given variant prestige, speaker’s and hearer’s preferences for prestigious variants show different degrees of efficiency on diffusion; and structural features of social network, such as average degree and level of centrality, also modulate the degree of diffusion. This work offers an interdisciplinary approach to study linguistic diffusion and modify available theories and simulations concerning individual learning or social factors.
人类语言不断变化。很多变化是通过在一组交流个体中扩散不同形式的语言变体完成的。我们设计了一个基于Pόlya-urn动态的简单计算机模型研究语言、个体学习和社会-文化因素对语言扩散可能起的作用。模拟结果和统计分析表明：（1）对变体的偏好，而不是对使用者的偏好，是决定扩散的基本选择要素；在存在变体偏好情况下，（2）说者和听者对变体的偏好具有不同程度的扩散效率；（3）网络结构因素，如平均度和集中度，可调节扩散程度。此工作为研究语言扩散提供了一个跨学科手段，并改进了已有的关于个体和社会因素对语言扩散影响的理论和模拟研究。
Keywords 关键词
Linguistic diffusion 语言扩散 Computational modeling 计算机模拟 Pólya-urn dynamics Pόlya-urn动态 Complex networks 复杂网络
Abstract 摘要
The production and origin of Tibetan tones were studied from multiple viewpoints. The following results and conclusion were obtained: 1) homophone is the internal impetus of the production and the origin of Tibetan tones, but the origin of Tibetan tones was restricted by the phoneme structure; 2) the distributions of homophone in all the Tibetan dialects are close to an exponential curve, which is the basic nature of Tibetan; 3) from the viewpoint of phoneme evolution, the phoneme load could transfer among the initials, finals and tones. The increase of initial loads led to the pitch patterns being transferred to the tone patterns; 4) this study proposed the concept of ‘quasi-phoneme’; 5) from the viewpoint of structure diffusion in the lexicon, the structure diffusion was caused by the decrease of initials and finals; 6) in the production and origin of Tibetan tones, the formation of tones was restricted physiologically and psychologically by phonation and perception. The devocalization of the initial is the basis for the origin of tones.
本文从不同的视角对藏语声调产生和起源进行了研究，研究主要得到了以下结论：1）同音词的增加是导致藏语产生声调的主要内在动力。从结构的演化来看，声调的产生又受到藏语音位系统结构的制约；2）同音词的分布在藏语各方言中都是遵循指数曲线，这是藏语音位的基本内在性质；3）从音位负担的演化来看，在藏语信息量为常数的情况下，声、韵、调三者的音位负担可以相互转换，藏语声母负担的增加直接导致了声调的产生，因此，藏语的音调转换为声调；4）提出了“准音位”的感念；5）从语音结构在词汇中的扩散和语音演化看，藏语声、韵母的减少形成了藏语语音结构在词汇中的扩散；6）声调产生过程中，语音的演化主要受制于发音生理机制和感知机制,声母清化是声调产生的感知基础。
Keywords 关键词
Origin of tones 声调起源 Homophone rate 同音词率 Phoneme load 音位负担量 Lexical diffusion 词汇扩散 Speech perception 言语生理感知 [Tibetan language–Tone 藏语声调]
Abstract 摘要
This paper examines three kinds of coexistent human-denoting interrogative words in early Southern Min texts: ti7-lang5值人(<底儂), * si7-tsui7是誰 and mih4 lang5乜人(<物儂) representing different chronological strata, viz., Southern Dynasties period (5-6th century), Wei-Jin Period (3-5th century), and Tang-Song Period (8-10th century). If we assume that ti7-lang5 底儂 is the indigenous system, si7-tsui7是誰 and mih4 lang5物儂 must have been the implanted systems. As attested in the texts, si7-tsui7是誰 is most robust, whereas ti7-lang5值人 is the least robust, and mih4 lang5乜人stands in the middle. However, the fate of the three kinds of interrogative words differs in modern descendants of Southern Min. The relative robustness varies with varieties of Southern Min dialects. The Hainan dialect, an outlier Southern Min dialect, for example, still preserves the modern reflex of ti7-lang5底儂 in the fused form of ‘ʔdiang’. Other outlier dialects such as Yongchun (永春), Quemoy (金門), Penghu (澎湖), Dongshan (東山), Leizhou (雷州), and Chaozhou (潮州) dialects still keep the fusional form tiaŋ5. In this paper, we will also go beyond the pale of Southern Min dialects and examine how the interrogative words in question fare in other Min dialects. The related issues of the relative strength of the pseudo-variable lang5人儂 in the formation of plural personal pronouns in Min dialects will also be tackled and brought to bear on the distribution of the three kinds of interrogative words.
本文探讨早期闽南语文本中三种并存指人疑问代词: ti7-lang5 “值人”(<底侬)、si7-tsui7 “是谁”、mih4 lang5 “乜人”(<物侬)，分别代表不同的时代层次，即南朝(五至六世纪)、魏晋(三至五世纪)、唐宋(八至十世纪)层次。若假定 “底侬” 为固有层，“谁/谁人” 、“物侬”就是移借层。从明清戏文看来“谁/谁人”最有活力，“值人”最不具活力， “乜人”处于中间。但这三种指人疑问代词在现代闽南语中发展情况不一。 其活力视次方言而变。比如较边缘的海南方言还保存“底侬” 的现代反映，为合音形式/ʔdiang/ 。其他边缘的方言，如永春、金门, 澎湖, 东山, 雷州, 潮州等也保存合音词 tiaŋ5. 本文也将考察闽南语之外指人疑问代词在其他闽方言的分布情况。准变项“侬”在作为闽语人称代词复数形所显现的强度也会和三种指人疑问代词的分布放在一起讨论。
Keywords 关键词
Who 谁 Southern Min 闽南语 Chronological strata 时代层次 Evolution 演变
Abstract 摘要
It has usually been claimed that language is what makes humans uniquely human. However, I submit that it is the mind as a general problem-solving capacity for adaptations to current ecological pressures that makes humans uniquely humans. Non-linguistic communication may have provided part of the ecological niche for the emergence of languages, having predisposed hominines to pay attention to each other and to read each other’s mind. The same mind that produced all these new developments (in particular complex social life and a richer cognitive capacity) also enabled hominines to develop languages as communication technologies.
一个通常的说法是语言使得人成为独特的人类。然而，我认为是作为适应当前生态压力而具有的一般问题解决能力的心智才使得人成为独特的人类。非语言的交流也许为语言的涌现提供了部分的生态挈机，预先使得类人相互注意并能了解彼此的意图。心智导致所有这些新的发展（尤其是复杂的社交生活与更丰富的认知能力），进而使得类人把语言发展成为交际技术。
Keywords 关键词
Language evolution 语言演化 Exaptation 扩展适应 Culture 文化 Ontogeny 个体发生学 Phylogeny 群体发生学
Abstract 摘要
This paper discusses some of the most hotly debated topics over the past few years on syntactic-semantic change in a functional-cognitive perspective and proposes a new model of grammatical change in Chinese by providing more solidly-based definitions of such notions as “grammaticalization”, “lexicalization”, “degrammaticalization”, “exaptation”, “reanalysis”, and “analogy” with respect to internal processes of change as well as external ones, specifically, borrowing through language contact or contact-induced change. It will be proposed that this model is constituted by just two internal mechanisms: reanalysis and analogy. Consequently, it will be argued that grammaticalization – which has been by far the focus of most of the studies on historical grammar in recent decades – is secondary. The processes of grammaticalization, lexicalization and exaptation will thus be viewed as sub-classes of reanalysis, while some “degrammaticalization”/lexicalization processes will be more aptly viewed as a sub-class of analogy. The main motivations (if not genuine explanations) for grammatical change will also be discussed, i.e. semantic-pragmatic change, including mainly metaphorical extension, pragmatic inferencing or metonymization and (inter-) subjectification, as well as others, such as phonological change. Concerning the third – and external – mechanism of change, it will be shown that the several universals and principles of borrowing that have been proposed remain rather ill-defined.
本文从功能认知的角度讨论过去几年来争论最热烈的句法语意变化问题，并提出一个关于中文语法变化的模型。该模型为基于变化的内在处理和外部处理，尤其是语言接触导致的借用或者由接触引起的变化的语法化、词汇化、去语法化、扩展适应、再分析与类比等概念提供更具体的定义。
Keywords 关键词
Syntactic-Semantic change 句法语义变化 Grammaticalization 语法化 Exaptation 扩展适应 Lexicalization 词汇化 Degrammaticalization 去语法化 Metaphorical extension 隐喻扩展 Pragmatic inferencing 语用推理 Subjectification 主体化 Contact-induced change 接触引起的变化
Abstract 摘要
While noun and verb are two separate syntactic classes in opposition to each other in Indo-European languages, in Chinese they are a pair of pragmatic classes (reference and predication), and not in opposition to each other. Chinese nouns constitute a super-noun category with verbs as its sub-category. In other words, all verbs in Chinese are actually verbal nouns. This characteristic feature of Chinese can be explicated from the perspectives of cognition, language typology, grammaticalization, philosophical background, and experimental studies of word classes.
名词和动词在印欧语中已经分化为两个对立语法范畴，然而汉语的名词和动词间不存在类似的关系，二者分别对应指称语和陈述语，因而更切合语用范畴。此外，汉语的名词构成一个大类，包含动词。换句话说，汉语是一种名动包含的格局，动词实质上是兼有动词性的名词，即“动名词”。本文从认识方式、语言类型、语法化、哲学背景、词类的实验研究等诸方面论证汉语的这一特点。
Keywords 关键词
Super-noun 大名词 Verbal nouns 动名词 Predication 述谓语 Grammaticalization 语法化
Abstract 摘要
The Inexplicability Principle has been proposed as one part of the Distillation method in Wang (2006). It means the inability to describe a recipient language in terms of the phonological system of the donor language. Since the related morphemes originate either from borrowing or from inheritance, the inexplicable elements from borrowing are considered to be inherited from the ancestor language. In its application to Sino-Tai numerals, some puzzle appears. For instance, Tai *hɑC ‘five’ is related to Old Chinese (OC) *ŋɑɡx ‘five’, but the initial *h- in Tai cannot be borrowed from *ŋ- in OC. Therefore, it seems that they must be inherited from Proto-Sino-Tai **hŋ- and independently developed into *h- and *ŋ-. That means Sino-Tai related numeral ‘five’ is inherited. However, Chen (2007) and other scholars provided more evidence to confirm that Tai *hɑC ‘five’ has been borrowed from Chinese. To resolve this puzzle, we examine the phonological development related to ‘five’ and ‘six’ in Chinese and Tai, and suggest that the inexplicability principle is constrained by our knowledge of phonology of languages in comparison. In conclusion, the inexplicability principle is tenable, but it should be used with caution.
不可释原则是还原比较法的一个组成部分(Wang 2006)，指无法以施借语言的音韵系统来解释受借语言中关系语素的表现，这些不可解释的成分应该视为祖语的遗传，而不是借用的结果。在应用该原则到汉台数词时，出现了一些争议。例如，台语的 *hɑC ‘五’ 对应上古汉语的 *ŋɑɡx ‘五’，而台语的声母 *h- 不可能借自上古汉语的 *ŋ-。因此，它们应该都源自原始汉台语的 **hŋ-，各自发展出 *h-和 *ŋ-。这就是意味着汉台关系语素‘五’是同源词。但是，陈保亚(2007) 等举出更多的证据来说明台语的 *hɑC‘五’ 借自汉语。为了解决这一争议，本文检视了汉台关系语素‘五’和‘六’的语音发展，发现不可释原则的运用受制于我们对于所比较的语言史之了解。不可释原则在语源研究中可用，但要注意其限度。
Keywords 关键词
Inexplicability principle 不可释原则 Sino-Tai 汉台比较 Five 五 Six 六
ZH-HK 
EN 
ZH-CN