Syntactic Phylogenetic Trees

Abstract

In light of recent controversies surrounding the use of computational methods for the reconstruction of phylogenetic trees of language families (especially the Indo-European family), a possible approach based on syntactic information, complementing other linguistic methods, appeared as a promising possibility, largely developed in recent years in Longobardi's Parametric Comparison Method. In this paper we identify several serious problems that arise in the use of syntactic data from the SSWL database for the purpose of computational phylogenetic reconstruction. We show that the most naive approach fails to produce reliable linguistic phylogenetic trees. We identify some of the sources of the observed problems and we discuss how they may be, at least partly, corrected by using additional information, such as prior subdivision into language families and subfamilies, and a better use of the information about ancient languages. We also describe how the use of phylogenetic algebraic geometry can help in estimating to what extent the probability distribution at the leaves of the phylogenetic tree obtained from the SSWL data can be considered reliable, by testing it on phylogenetic trees established by other forms of linguistic analysis. In simple examples, we find that, after restricting to smaller language subfamilies and considering only those SSWL parameters that are fully mapped for the whole subfamily, the SSWL data match extremely well reliable phylogenetic trees, according to the evaluation of phylogenetic invariants. This is a promising sign for the use of SSWL data for linguistic phylogenetics. We also argue how dependencies and nontrivial geometry/topology in the space of syntactic parameters would have to be taken into consideration in phylogenetic reconstructions based on syntactic data. A more detailed analysis of syntactic phylogenetic trees and their algebro-geometric invariants will appear elsewhere.

Topological Analysis of Syntactic Structures

“…There is some controversy over the use of computational methods for the reconstruction of language phylogenetic trees, especially for the Indo-European family, see [34]. In [39] the authors discuss the advantages of the algebro-geometric approach used in [37] over other distance based methods. It is shown that the Algebro-Geometric approach yields better results, especially when additional information, such is the subdivision into subfamilies and the position of the old languages is taken into account.…”

Phylogenetics of Indo-European Language Families via an Algebro-Geometric Analysis of Their Syntactic Structures

“…In [45] it was shown that, when using syntactic data of the SSWL database [51] with Hamming distances and neighborhood joining methods to construct linguistic phylogenetic trees, several kinds of errors typically occur. These are mostly due to a combination of two main factors:…”

“…Another possible source of problems is due to the structure of the SSWL database itself, where the syntactic binary variable recorded are not what linguists would consider to be actual syntactic parameter in the sense of the Principles and Parameters model [10], [12], see also [42]: there are conflations of deep and surface structures that make certain subsets of the syntactic variables of the SSWL data potentially problematic from the linguistic perspective. However, it was also shown in [45] that several of these problems that occur in a naive use of computational phylogenetic methods can be avoided by a more careful analysis. Namely, some preliminary evidence is given in [45] that, when a naive phylogenetic reconstruction applied simultaneously to the entire SSWL database is replaced by a more careful analysis applied to smaller groups of languages that are more uniformly mapped in the database, the phylogenetic invariants of Phylogenetic Algebraic Geometry can identify the correct phylogenetic tree, despite the imperfect nature of the SSWL data.…”

“…However, it was also shown in [45] that several of these problems that occur in a naive use of computational phylogenetic methods can be avoided by a more careful analysis. Namely, some preliminary evidence is given in [45] that, when a naive phylogenetic reconstruction applied simultaneously to the entire SSWL database is replaced by a more careful analysis applied to smaller groups of languages that are more uniformly mapped in the database, the phylogenetic invariants of Phylogenetic Algebraic Geometry can identify the correct phylogenetic tree, despite the imperfect nature of the SSWL data. The method of Phylogenetic Algebraic Geometry that we refer to here was developed in [34], [35] for applications to mathematical biology, see also a short survey in [5].…”

“…In terms of phylogenetic inference about Linguistics, however, it is important to locate more precisely where the root vertex should be. In the case of languages belonging to a subfamily of the Indo-European languages this can be done, as in the example we discussed in [45], by introducing the data of some of the ancient languages in the same subfamily as a new leaf of the tree, that will help locating more precisely the root vertex of the original tree based on the modern languages. For language families for which there are no data of ancient languages available, however, this kind of phylogenetic analysis will only identify a tree topology as an unrooted binary tree.…”