<div dir="ltr"><div dir="ltr"><div dir="ltr"><div dir="ltr"><br></div><br><div class="gmail_quote"><blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex"><div><p><br>In linguistics, because of widespread horizontal transfer
(hybridization, i.e. contact-induced change), the genealogies are
typically quite uncertain at lower levels – Glottolog has many
cases of "rake-like" trees (e.g. Southern Bantoid [sout3152] has 9
branches, and Iranic [iran1269] has eight branches). So do the
methods adopted from computational biology add all that much? That
isn't clear to me – maybe also because the papers in the
general-science journals tend to be short and focused on
technicalities, not on explaining how the new methods relate to
earlier methods and why they are better in practice (not only in
theory).<br></p></div></blockquote><div><br></div><div>If we infer phylogenies using vocabulary, as opposed to morphology or structural traits, historical linguists have developed methods to identify layers of loanwords in a systematic way, and thus this issue can be solved to a certain degree. Using vocabulary has the advantage that you can then use this phylogeny to study how structural features evolve without a serious risk of circularity.</div><div><br></div><div>Bayesian phylogenies have many advantages: they don't simply produce one tree, but a tree distribution, and thus give information on which groupings are robust and which are less well supported. They are not just a topology: they can also be used to infer ages, and while in most cases we have large uncertainty, the results are surprisingly close to the intuition of historical linguists. We can learn a lot, even from seemingly rake-like phylogenies.</div><div><br></div><div>The consensus trees that are published in the articles on phylogeny is just the tip of the iceberg of the amount of information you can gain from these tree distributions, but for now there is no convenient interface to explore these data, and some knowledge of R or other languages is necessary. This forthcoming chapter presents a (hopefully) readable introduction to phylogenies for historical linguists: <a href="https://www.academia.edu/101656989/The_Family_Tree_model">(99+) The Family Tree model | Guillaume Jacques and Thomas Pellard - Academia.edu</a></div><div><br></div><div>In the end, what decides the reliability of these studies is the reliability of cognate coding, which means that historical linguistics specialized in meticulous etymologies and sound laws will play a crucial part, and should work collectively to produce better phylogenies, which typologists can then use to study the distribution of structural features through time and space.</div><div><br></div><div>Guillaume</div><div></div><div><br></div></div><span class="gmail_signature_prefix">-- </span><br><div dir="ltr" class="gmail_signature"><div dir="ltr"><div>Guillaume Jacques</div><div><br></div><div>Directeur de recherches<br>CNRS (CRLAO) - EPHE- INALCO <br></div><div><a href="https://scholar.google.fr/citations?user=1XCp2-oAAAAJ&hl=fr" target="_blank">https://scholar.google.fr/citations?user=1XCp2-oAAAAJ&hl=fr</a><br></div><div><a href="http://cnrs.academia.edu/GuillaumeJacques" target="_blank">https://langsci-press.org/catalog/book/295</a></div><div><div><a href="http://panchr.hypotheses.org/" target="_blank">http://panchr.hypotheses.org/</a></div></div></div></div></div></div></div>