Can Grammar Define Similarity of Human Natural Languages?

Polyakov, Vladimir; Anisimov, Ivan; Makarova, Elena

doi:10.3844/ajassp.2016.1040.1052

Cited by 1 publication

(6 citation statements)

References 12 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Figure 1 shows a tree built for the set of 27 languages singled out in Polyakov et al (2016) and in Table 1 above. It is based on lexical (also implying phonetic) data from ASJP.…”

Section: Resultsmentioning

confidence: 99%

“…Finally, and closer to the spirit of the present paper, grammatical features have also sometimes been drawn upon for phylogenetic inferencing, e.g., Longobardi et al (2015). Moreover, such features have been used to find genealogical relationships among languages (Polyakov et al 2016), for identifying cases of language contact, and for dating language divergence (Solovyev 2009). Using grammatical features for these kinds of historical linguistic purposes is a recent endeavor, so it seems quite reasonable to expect some new results and further developments.…”

Section: Introductionmentioning

confidence: 83%

“…Concerning the third suggestion of accidental evolution, the probability of many features independently converging on acquiring the same values is very low-it is something which is more likely to take place when the data is sparse (Polyakov et al 2016). Figure 3 shows a tree that combines information on the divergence (ASJP tree) and convergence (WALS tree) of languages.…”

Section: Discussionmentioning

confidence: 99%

“…The fact that it represents a sparse data matrix requires users embarking on any kind of statistical analysis to implement a selection procedure. (Polyakov et al 2016) contains a detailed description of a procedure for cleaning WALS data and selecting languages based on their degree of attestation, and we follow the same procedure here.…”

Section: Methodsmentioning

confidence: 99%

“…Next, also following (Polyakov et al 2016), we select language pairs according to the following criteria, which should all be satisfied: (1) at least 26 features are attested for both languages; (2) at least 65% of the feature values match; (3) the pair should be above a regression line within the space of shared values as a function of overlapping features fitted such that all pairs involving unrelated languages are below the line; (4) one member of the pair is the closest matching language of the other language. Criterion (4) can cause a language to recur in different pairs if that language comes up when searching for language closest to A and then comes up again when searching for the language closest to language B. Swedish is a case where this happens.…”

Section: Methodsmentioning

confidence: 99%

See 4 more Smart Citations

Phylogenetic trees: Grammar versus vocabulary

Polyakov¹,

Makarova

Solovyev

2022

Russian Journal of Linguistics

View full text Add to dashboard Cite

Traditionally, genealogical relationships between languages are established on the basis of phonetic and lexical data. The question whether genealogical relationships among languages can be defined based on grammatical data remains unanswered. The objective of this article is to compare two phylogenetic trees: one built using the Automated Similarity Judgment Program (ASJP) project, and one using the World Atlas of Language Structures (WALS). We include data from WALS representing 27 languages from 5 language families of all continents that are deemed to be sufficiently well described. A Hamming distance matrix was calculated for all languages under study, and, based on the matrix, a phylogenetic tree was built. The trees built according to WALS and ASJP data are compared with each other and with a tree built by the classical comparative historical method. Both the ASJP-based tree and the WALS-based tree have their advantages and disadvantages. The ASJP-based tree is a good reflection of the evolutionary divergence of languages. Similarities of languages as calculated based on the typological database of WALS can provide information on the history of languages both in terms of genealogical descent and contact with other languages. The ASJP-based tree reflects genealogical relationship well at a relatively small time depth, while the WALS-based tree reflects genealogical relationship well at large time intervals. We suggest a new variant of a phylogenetic tree that includes information on both the divergence (ASJP project) and the convergence (WALS project) of languages, combining the benefits of both of these trees, although the problem of borrowings remains. The present research reveals prospects for future studies of genealogical relations among languages based on large-scale descriptions of their grammatical structures.

show abstract

“…Figure 1 shows a tree built for the set of 27 languages singled out in Polyakov et al (2016) and in Table 1 above. It is based on lexical (also implying phonetic) data from ASJP.…”

Section: Resultsmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 83%

Section: Discussionmentioning

confidence: 99%

Section: Methodsmentioning

confidence: 99%

Section: Methodsmentioning

confidence: 99%

See 3 more Smart Citations