Supplementary Material for: Long-Range Comparison between Genes and Languages Based on Syntactic Distances

<i>Objective:</i> To propose a new approach for comparing genetic and linguistic diversity in populations belonging to distantly related groups. <i>Background:</i> Comparisons of linguistic and genetic differences have proved powerful tools to reconstruct human demographic history. Current models assume on both sides that similarities reflect either descent from common ancestry or the balance between isolation and contact. Most linguistic phylogenies are ultimately based on lexical evidence (roughly, words and morphemes with their sounds and meanings). However, measures of lexical divergence are reliable only for closely related languages, thus large-scale comparisons of genetic and linguistic diversity have appeared problematic so far. <i>Methods:</i> Syntax (abstract rules to combine words into sentences) appears more measurable, universally comparable, and stable than the lexicon, and hence certain syntactic similarities might reflect deeper linguistic relationships, such as those between distant language families. In this study, we for the first time compared genetic data to a matrix of syntactic differences among selected populations of three continents. <i>Results:</i> Comparing two databases of microsatellite (Short Tandem Repeat) markers and Single Nucleotides Polymorphisms (SNPs), with a linguistic matrix based on the values of 62 grammatical parameters, we show that there is indeed a correlation of syntactic and genetic distances. We also identified a few outliers and suggest a possible interpretation of the overall pattern. <i>Conclusions:</i> These results strongly support the possibility of better investigating population history by combining genetic data with linguistic information of a new type, provided by a theoretically more sophisticated method to assess the relationships between distantly related languages and language families.

研究目标：提出一种用于比较远缘相关群体遗传多样性与语言多样性的新方法。研究背景：语言差异与遗传差异的比较研究，已被证明是重建人类人口史的有力工具。现有模型均假定，两类差异中的相似性要么源自共同祖先的遗传，要么反映了隔离与接触之间的平衡。多数语言系统发育研究最终均基于词汇证据（大致指带有语音与语义的词汇与词素），但词汇差异的量化方法仅对亲缘关系较近的语言可靠，因此迄今为止，大规模的遗传与语言多样性比较研究仍存在较大局限。研究方法：句法（用于将词汇组合为语句的抽象规则）相较于词汇更具可测性、跨语言可比性与稳定性，因此部分句法相似性或许能够反映更深层的语言亲缘关系，例如远缘语系之间的关联。本研究首次针对三大洲选定群体的遗传数据，与句法差异矩阵开展比较分析。研究结果：本研究结合基于62项语法参数值构建的语言差异矩阵，对微卫星（Short Tandem Repeat，短串联重复序列）标记与单核苷酸多态性（Single Nucleotide Polymorphisms，SNPs）两类数据库进行分析，结果证实句法距离与遗传距离确实存在相关性。本研究同时识别出若干异常值，并对整体分布模式提出了可行的解释方向。研究结论：上述结果有力支持了这一研究方向：通过将遗传数据与新型语言信息相结合，能够更精准地探究群体历史——这类新型语言信息由理论上更为完善的远缘语言与语系关系评估方法所提供。